( PDF ) A labview based synchronized data acquisition system with integrated webcam for welding processes

Download PDF

ads:

FERNANDA MENDONÇA SILVEIRA

A LABVIEW BASED SYNCHRONIZED DATA

ACQUISITION SYSTEM WITH INTEGRATED

WEBCAM FOR WELDING PROCESSES

UNIVERSIDADE FEDERAL DE UBERLÂNDIA

FACULDADE DE ENGENHARIA MECÂNICA

2006

ads:

Livros Grátis

http://www.livrosgratis.com.br

Milhares de livros grátis para download.

FERNANDA MENDONÇA SILVEIRA

A LABVIEW BASED SYNCHRONIZED DATA ACQUISITION SYSTEM

WITH INTEGRATED WEBCAM FOR WELDING PROCESSES

Dissertation presented to the Post-graduation

Programme in Mechanical Engineering of Federal

University of Uberlândia, as part of the

requirements to obtain the title of MASTER IN

MECHANICAL ENGINEERING.

Concentration Area: Manufacturing Processes.

Supervisor: Prof. PhD. Américo Scotti

UBERLÂNDIA-MG, BRAZIL

2006

ads:

FICHA CATALOGRÁFICA

Elaborada pelo Sistema de Bibliotecas da UFU / Setor de Catalogação e Classificação

S587l

Silveira, Fernanda Mendonça, 1981-

A labview based synchronized data acquisition

system with integrated webcam for welding processes /

Fernanda Mendonça Silveira. - Uberlândia, 2006.

156f. : il.

Orientador: Américo Scotti.

Dissertação (mestrado) – Universidade Federal de

Uberlândia, Programa de Pós-Graduação em

Engenharia Mecânica.

Inclui bibliografia.

1. Soldagem - Teses. 2. Processos de fabricação -

Teses. I. Scotti, Américo. II. Universidade Federal de

Uberlândia. Programa de Pós-Graduação em

Engenharia Mecânica. III. Título.

621.791

FERNANDA MENDONÇA SILVEIRA

A LABVIEW BASED SYNCHRONIZED DATA ACQUISITION SYSTEM

WITH INTEGRATED WEBCAM FOR WELDING PROCESSES

Dissertation APPROVED by the Post-graduation

Programme in Mechanical Engineering of Federal

University of Uberlândia.

Concentration Area: Manufacturing Processes.

Evaluation Committee:

______________________________________

Prof. PhD. Américo Scotti (UFU) – Supervisor

______________________________________

Prof. Dr. Jair Carlos Dutra (UFSC-SC)

______________________________________

Prof

. Dra. Ing. Vera Lúcia D. S. Franco (UFU)

UBERLÂNDIA-MG, BRAZIL

2006

I dedicate this work to my family, especially to my mother Edna and my grandfather

Agenor, who much helped me to reach this point of my life. Thank you!

ACKNOWLEDGEMENT

I would like to thank the Federal University of Uberlândia and the Faculty of

Mechanical Engineering for the opportunity of doing this course.

Cranfield University for the opportunity of using its facilities and the funds provided.

My supervisor in Brazil, Américo Scotti, for spending his time giving me the

orientation I needed to execute this work.

My supervisor in England, David Yapp, for the permission I had to develop this work

at Cranfield University.

My colleagues in Brazil for the moments we spent together.

My colleagues in England, especially Gil and Harry, for the patience and help I

received.

Thanks to CNPq and the Programme AlBan for funding this work.

“This work was supported by the Programme AlBan, the European Union Programme

of High Level Scholarships for Latin America, scholarship no. E04M03892BR”

“This work was also supported by CNPq, from Brazil, scholarship no. 132497/2004-2”

SILVEIRA, F. M. A LabVIEW Based Synchronized Data Acquisition System with

Integrated Webcam for Welding Processes. 2006. p. 156. Master Dissertation, Federal

University of Uberlândia, Uberlândia.

Abstract

The development of data acquisition techniques able to monitor the performance of an

automated welding system could be, for example, applied to a system able to repair pipes full

of operating fluid. In this case, the system must precisely control and monitor the heat input

during the weld. For this purpose, it would be necessary to develop data acquisition

techniques able to monitor the performance of such a system. Digital oscilloscopes, which

are usually the kind of equipment used for this purpose, present high-cost as well as other

solutions offered by the market. Moreover, those systems do not offer a specific solution to

the welding area. Thus, the aim of this work was to develop a synchronized data acquisition

system with integrated webcam that was suitable for the study of the several types of welding

processes. It could be used for automation of pipeline repair systems, as well as, for other

types of welding operations, including experimental evaluations into laboratories. A software

package was developed using LabVIEW as programming language. It is able to do

synchronized acquisition of signals and image and post analysis of the acquired data. The

use of low-cost cameras like webcams was introduced aiming to determine arc

characteristics as arc length, for instance. The acquisition of the emitted sound by the arc

during the welding was also considered, what could increase the feeling of the professional

working away and could be used for studies related to welding quality. Using the existent

resources in the laboratory, a hardware prototype with enough flexibility and suitable for

several types of welding processes was designed. Afterwards, a system with a physical

configuration more appropriate to the work environment and lower cost was built. This

system was assessed with two different welding operations, Gas Metal Arc Welding and

Resistance Spot Welding, when its limitations and advantages were explored and identified.

Keywords: Data Acquisition. Data Analysis. Welding Processes. Image. Webcam.

SILVEIRA, F. M. Sistema de Aquisição de Dados Sincronizado Integrado com Webcam

para Processos de Soldagem Baseado em LabView. 2006. 156 p. Dissertação de

Mestrado, Universidade Federal de Uberlândia, Uberlândia.

Resumo

O desenvolvimento de técnicas de aquisição de dados capazes de monitorar o desempenho

de um sistema de soldagem automatizado pode ser aplicado, por exemplo, à um sistema

capaz de fazer reparos em dutos cheios de fluídos em operação. Nesse caso, o sistema

deve precisamente controlar e monitorar a entrada de calor enquanto a solda é feita. Os

osciloscópios digitais, que são geralmente os aparelhos utilizados para esse fim,

apresentam custo bastante elevado, assim como outras soluções existentes no mercado.

Além disso, esses sistemas não oferecem uma solução específica para a área de soldagem.

Assim, o objetivo desse trabalho foi desenvolver um sistema de aquisição de dados

sincronizado integrado com webcam para uso no estudo dos diversos tipos de processos de

soldagem. Esse sistema poderia tanto ser usado para automação de sistemas de reparo de

dutos assim como para outros tipos de operações de soldagem, inclusive avaliações

experimentais em laboratórios. Desenvolveu-se um programa em LabView capaz de realizar

a aquisição de sinais e imagem de forma sincronizada e ainda oferecer recursos para uma

análise posterior dos dados adquiridos. Introduziu-se o uso de câmeras de baixo custo,

como as webcams, objetivando a determinação de características do arco voltaico, como

por exemplo, o seu comprimento. Também se considerou a aquisição do som emitido pelo

arco durante a soldagem, o que poderia ajudar um profissional trabalhando à distância e

também ser utilizado para estudos relacionados à qualidade da solda. A partir dos recursos

oferecidos em laboratório, construiu-se um protótipo de equipamento flexível o bastante para

atender diversos tipos de processos de soldagem. Em seguida, construiu-se um sistema

com uma configuração física mais apropriada ao ambiente de trabalho. Este sistema foi

avaliado em duas diferentes operações de soldagem, soldagem MIG e soldagem a ponto

por resistência, quando suas limitações e vantagens foram exploradas e identificadas.

Palavras-chave: Aquisição de Dados. Análise de Dados. Processos de Soldagem. Imagem.

Webcam.

LIST OF SYMBOLS AND ABREVIATIONS

PIG Pipeline Inspector Gauge

WERC Welding Engineering Research Centre

LAPROSOLDA Laboratory for the Development of Welding Processes

AWS American Welding Society

AC Alternate Current

DC Direct Current

GMAW Gas Metal Arc Welding

RSW Resistance Spot Welding

CAPS Cranfield Automated Pipewelding System

CCD Charge Coupled Device

DAQ Data Acquisition

DMA Direct Memory Access

PCI Peripheral Component Interconnect

USB Universal Serial Bus

HS High-Speed

NI National Instruments

AIGND Analog Input Ground

CMRR Common-Mode Rejection Ratio

NRSE Non-Referenced Single-Ended

AISENSE Single-Node Analog Input Sense

ADC Analog-to-Digital Converter

LSB Least Significant Bit

DNL Differential Nonlinearity

DAC Digital-to-Analog Converter

I/O Input/Output

RF Radio Frequency

LAN Local Area Network

PCMCIA Personal Computer Memory Card Interface Adapter

GPS Global Positioning System

BNC British Naval Connector

INDEX

CHAPTER 1 - INTRODUCTION.............................................................................................. 1

CHAPTER 2 - LITERATURE REVIEW ................................................................................... 3

2.1

WELDING PROCESSES ..................................................................................................... 3

2.2

ARC WELDING MONITORING AND ANALYSIS ...................................................................... 7

2.3

DATA ACQUISITION ........................................................................................................ 10

2.3.1 Concepts............................................................................................................... 10

2.3.2 Systems................................................................................................................ 16

2.4

LABVIEW ....................................................................................................................... 21

CHAPTER 3 - METHODOLOGY AND DEVELOPMENT ..................................................... 23

3.1

HARDWARE PROJECT .................................................................................................... 24

3.2

SOFTWARE PROJECT..................................................................................................... 32

3.3

CALIBRATION............................................................................................................... 124

3.4

SYSTEM CONFIGURATION ............................................................................................. 125

CHAPTER 4- EVALUATION OF THE SYSTEM "SMART"................................................ 134

4.1

FIRST SET OF EXPERIMENTAL TRIALS – SINGLE WIRE GMAW......................................... 134

4.2

SECOND SET OF EXPERIMENTAL TRIALS – DOUBLE WIRE GMAW ................................... 139

4.3

THIRD SET OF EXPERIMENTAL TRIALS – RSW ............................................................... 144

CHAPTER 5 - DISCUSSION............................................................................................... 149

CHAPTER 6 - FUTURE DEVELOPMENTS........................................................................ 149

CHAPTER 7 - CONCLUSIONS........................................................................................... 154

CHAPTER 8 - REFERENCES............................................................................................. 155

CHAPTER 1

INTRODUCTION

Underground oil and gas pipelines are susceptible to corrosion and other damage on

the outside surface of the pipe. Many pipelines have now been in service for periods of up to

50 years and these lines will continue in use for a foreseeable future. It is common for

pipelines to be inspected by an internal Pipeline Inspection Gauge (PIG). This device travels

along inside the pipe propelled by the media flow and uses ultrasonic or other sensing

methods to determine areas where pipe wall thinning may have occurred, as shown by

Figure 1.1 and Figure 1.2. These areas are, then, excavated and the pipe is repaired, while

still carrying its operating fluid (commonly oil or natural gas). Methods have been developed

to restrict the possibility of burn-through while making a weld repair on a live pipeline, but

these methods depend on the accurate control of heat point. This can be most easily

achieved by application of an automated welding system, where heat input can be precisely

controlled and monitored. In addition, the use of an automated system offers the possibility of

remote operation of the equipment. So, that personnel would not be required do be in the

immediate vicinity of the pipeline while the repair takes place.

Figure 1.1 - UltraScan Smart PIG (Pipetronix Inc.)

Figure 1.2 – CalScan Smart Pig (Pipetronix Inc.)

Cranfield University - UK, through its Welding Engineering Research Centre (WERC),

has already examined the concept of automating pipeline repair, by using a repair welding

head mounted on two bands encircling the pipe, of the type often used for pipeline

construction. The welding head can, then, be driven round the pipe, and also along the

length of the pipe, with accurate control of the welding torch position at any point round the

pipeline. This system can be then used for making either of the two principal types of welding

repair employed in the field: build-up layers of weld metal to replace the steel lost in thinned

area, or to weld half-pipes round the pipes in a clamshell arrangement, to cover the area

where thinning has occurred.

In order to develop such a project, a microprocessor-based control system must be

also developed to operate the mechanical hardware. For this purpose, one of the required

tasks must be the development of rules from which welding procedures and parameters

could be adjusted to compensate for variations in welding conditions. This task requires

analysis of welding experiments and development of control algorithms to adjust welding

parameters. Hence, it must be also necessary to develop and integrate sensors with the

welding system, and to develop data acquisition techniques to monitor the performance of

welding.

Data acquisition systems have been improved along the years. Most of them acquire

signals while an on-line analysis of the weld performance is carried out. Then, it is possible to

act on the welding parameters in order to adjust a better work setting. The data can also be

recorded for a post-analysis. Since acquisition is usually done in different types of work

environments, these systems have also shown portability as an important characteristic.

However, studies of applications for low-cost cameras integrated to data acquisition systems

are still not very explored. Since they are much cheaper if compared to the high-speed

cameras, this kind of study becomes very worthwhile.

Then, a partnership between the Laboratory for the Development of Welding

Processes (LAPROSOLDA, Brazil) of Federal University of Uberlândia (UFU) and the

Welding Engineering Research Centre (WERC, UK) of Cranfield University was established.

This work is focused on the development of a portable data acquisition system, which could

be used for automation of pipeline repair systems. Besides electrical signals; the

synchronous acquisition of frames of cameras (low-cost and high-speed ones) and audio had

to be also offered, providing a more comprehensive analysis of the phenomenon. This

system should be also suitable for other types of welding operations, including experimental

evaluations into laboratories.

CHAPTER 2

LITERATURE REVIEW

2.1 Welding Processes

The American Welding Society (AWS) definition, according to The ARCON Welding

Inc. (2005)

, for a welding process is "a materials joining process which produces

coalescence of materials by heating them to suitable temperatures with or without the

application of pressure or by the application of pressure alone and with or without the use of

filler material". Following this definition, there are two classes of welding processes, fusion

welding (in which the coalescence happens during the fusion of the joint) and pressure

welding (in which the coalescence happens through heating below the fusion temperature

and pressuring over the parts).

According to The Lincoln Electric Inc. (1994), welding used to be done by heating

metals and pounding or ramming them together (pressure welding) until obtaining the

coalescence between them. In the early 1800’s was discovered that a voltaic arc could be

created with a high-voltage electric circuit by bringing two terminals near each other (fusion

welding). The heat produced by the arc could be used for joining the desirable parts. Figure

2.1 illustrates a basic arc welding circuit. Since then, the welding has been developed a lot,

but only in the beginning of the 19

century the process became commercially available.

Fusion welding is still an important way of joining materials, being the main heat source the

arc voltaic.

Still according to The Lincoln Electric Inc. (1994), an AC or DC power source, fitted

with whatever controls may be needed, is connected by a ground cable to the workpiece and

by a “hot” cable to an electrode holder of some type, which makes electrical contact with the

welding electrode. When the circuit is energized and the electrode tip touched to the

grounded workpiece, and then withdrawn and held close to the spot of contact, an arc is

created across the gap. The arc produces a temperature more than adequate for melting

most metals. The heat produced melts the base metal in the vicinity of the arc and any filler

metal supplied by the electrode or by a separately introduced rod or wire. A common pool of

molten metal is produced and solidifies behind the electrode as it is moved along the joint

being welding. The result is a fusion bond and the metallurgical unification of the workpieces.

Figure 2.1 - The basic arc-welding circuit (The Lincoln Electric Inc., 1994)

Many studies have been done in order to improve the heat efficiency, welding quality

and productivity, besides the adaptability to the work environment and to the many types of

materials. That is why several types of welding processes have shown up. AWS has grouped

them together according to the "mode of energy transfer" as the primary consideration. A

secondary factor is the "influence of capillary attraction in effecting distribution of filler metal"

in the joint. Capillary attraction distinguishes the welding processes grouped under "Brazing"

and "Soldering" from "Arc Welding", "Gas Welding", "Resistance Welding", "Solid State

Welding", and "Other Processes." The welding processes, in their official groupings, are

shown by Table 2.1. This table also shows the American letter designation for each process.

GMAW, illustrated in Figure 2.2, has currently been thought as the most suitable

welding process for large diameter transmission pipelines. The AWS defines GMAW as "an

arc welding process which produces coalescence of metals by heating them with an arc

between a continuous filler metal (consumable) electrode and the work piece. Shielding is

obtained entirely from an externally supplied gas or gas mixture." The electrode wire for

GMAW is continuously fed into the arc and deposited as weld metal. This process has many

variations depending on the type of shielding gas, the type of metal transfer, and the type of

metal welded. A number of recent welding procedure developments have improved

productivity using GMAW.

Table 2.1 - Welding and allied processes with the correspondent letter designation by AWS

(according to ARCON Welding Inc., 2005)

Group Welding Process Letter Designation

Arc welding Carbon Arc CAW

Flux Cored Arc FCAW

Gas Metal Arc GMAW

Gas Tungsten Arc GTAW

Plasma Arc PAW

Shielded Metal Arc SMAW

Stud Arc SW

Submerged Arc SAW

Brazing Diffusion Brazing DFB

Dip Brazing DB

Furnace Brazing FB

Induction Brazing IB

Infrared Brazing IRB

Resistance Brazing RB

Torch Brazing TB

Oxyfuel Gas Welding Oxyacetylene Welding OAW

Oxyhydrogen Welding OHW

Pressure Gas Welding PGW

Resistance Welding Flash Welding FW

High Frequency Resistance HFRW

Percussion Welding PEW

Projection Welding RPW

Resistance-Seam Welding RSEW

Resistance-Spot Welding RSW

Upset Welding UW

Solid State Welding Cold Welding CW

Diffusion Welding DFW

Explosion Welding EXW

Forge Welding FOW

Friction Welding FRW

Hot Pressure Welding HPW

Roll Welding ROW

Ultrasonic Welding USW

Soldering Dip Soldering DS

Furnace Soldering FS

Induction Soldering IS

Infrared Soldering IRS

Iron Soldering INS

Resistance Soldering RS

Torch Soldering TS

Wave Soldering WS

Other Welding Processes Electron Beam EBW

Electroslag ESW

Induction IW

Laser Beam LBW

Thermit TW

Figure 2.2 - GMAW – Diagram process (CARY, 1995)

This welding process is the most popular method for automated systems because the

electrode wire is continuous (CARY, 1995). As a continuous-wire process, it has a high

operator factor and as due to the high current density capability (short electrodes), this

process provides high deposition rate. Furthermore, it is an all-position welding process,

what allows an orbital welding execution. Those are the main reasons why this welding

process is used by all industrial manufacturing operations. It is also used for field

construction, including pipelines, and for maintenance and repair work.

GMAW can be classified into Single and Double Wire. Single GMAW uses only one

wire through one torch. Double Wire class, in turn, can be divided into two other classes, i.e,

Single Potential (also miscalled twin arc), in which two wires are fed through one contact tip,

and Isolated Potential (so called Tandem), in which there are two contact tips into the same

torch, as shown by Figure 2.3. There is a combination called Dual Tandem GMAW, which

uses two torches, having two wires each one, as Figure 2.4 shows. This is the most

productive GMAW method and many studies about it have been done. That is why the data

acquisition system developed in this work will be based on Dual Tandem GMAW limits.

According to Widgery and Blackman (2001), Cranfield University developed the

concept of Dual Tandem GMAW for pipewelding and received funding from BP Exploration

Operating Company and TransCanada Pipelines to develop the Cranfield Automated

Pipewelding System (CAPS). CAPS involves the use of two tandem welding torches fitted on

one pipe welding bug so that four arcs operate simultaneously. The dual tandem head was

fitted with a sensor based control system, removing the need for a skilled operator to

continuously monitor the weld. Eventually, this will be used for adaptive control of the welding

process.

Figure 2.3 - Cranfield Tandem GMAW (WIDGERY; BLACKMAN, 2001)

Figure 2.4 - CAPS Dual Tandem welding torches (WIDGERY; BLACKMAN, 2001)

2.2 Arc Welding Monitoring and Analysis

A low-cost camera, such as a webcam, could be used for capturing the arc welding

image while the weld is done. A recent work (GILSINN et al., 1999) used this type of camera,

which has an embedded web server and can accept aiming and zooming commands, to

export images at about 1Hz. A web-based interface is used for communicating the server

with the camera. Anyone on the internet, running a general purpose browser, can view

images and control the camera. Figure 2.5 shows a top view of a welding rig and a robot arm

with torch. Since discrete part manufacturers using robotic arc welding cells often have

several more geographically distributed plants than welding experts, this kind of control

through the internet allows a remote expert to view the cell in operation and inspect parts

after a weld.

Figure 2.5 - Live Video Feed from Weld Camera (GILSINN et al., 1999)

A more recent work (MASON et al., 2005) presents another application for the low-

cost cameras. A great deal of past and present research overlooks one of the most important

components of the resistance spot welding process, the electrode, of which its profile (tip)

has a direct effect on the quality of the weld. As the number of welds increase, the electrode

tip wears down and therefore the applied current is increased in order to maintain the same

current density. Mason et al. (2005) introduced a way to monitor the electrode profile, by

using a compact low-cost camera integrated into a PC.

Another application of a camera monitoring system is arc length measurements. Arc

voltage can give an idea about the arc behaviour; it means to check if the arc is longer or

shorter, but not its length as a measurement. Through arc welding images, the arc length

could be calculated. Using a compact arc light sensor, Li and Zhang (2001) managed to

measure the arc length with adequate accuracy. According to them, the arc length

determines the distribution of the arc energy and, thus, the heat input and width of the weld.

In their work, they aimed at improving the measurement accuracy of arc length by using the

spectrum of arc light at a particular wavelength during gas tungsten arc welding (GTAW) with

argon shield.

Maeda and Ichiyama (1999) developed an adaptive control of arc welding by image

processing. A CCD camera with an electronic trigger shutter was used for acquiring images

of a molten weld pool. For this purpose, an image acquisition algorithm was also developed

aiming to detect the pool edge and the centre of the electrode wire through the image. They

managed to develop a system less susceptible to the external influences as spattering during

welding.

The researches related to welding analysis use high-speed cameras able to show

much more details of the arc welding. This type of camera is mainly used by researchers for

metal transfer mode analysis, as shown in Figure 2.6. Bálsamo et. al (2000) shows that the

use of high-speed cameras allows a more precise measurement of both the metal transfer

mode and metal transfer size. Thus, in their work, they correlate the droplet and dimensions

with current and voltage signals, on a time base, by the development of a data acquisition

technique suitable for GMAW metal transfer analysis. Alfaro et. al (2005) also developed

such a system, but suitable for RSW parameters setup.

Figure 2.6 - Photo sequence of transfer in the globular/short circuit mode (FERRARESI;

FIGUEIREDO; ONG, 2003)

Another research trend is the analysis of the sound emitted by the arc welding, which

has been used more and more in studies related to welding quality. A high-speed data

acquisition system was developed by Mansoor and Huissoon (1999) to record and analyse

the arc sound produced during GMAW. The recorded data was processed to obtain time

domain, frequency domain and time-frequency domain descriptors. Relationships between

these descriptors and the originating weld parameter levels and metal transfer mode were

investigated, as there were relationships between the electrical power supplied to the weld

and the arc sound. Results indicate that the arc sound exhibits distinct characteristics for

each metal transfer mode. The occurrence of spatter and short circuits was also found to be

clearly detectable in the arc sound record.

Drouet and Nadeau (1982) proposed another application for acoustic analysis. Based

on measurement of the sound wave produced by the arc welding, their technique determines

the time evolution of the voltage in the column of an electric arc, being possible to control the

arc length. The technique is based on the property of the current modulated arc to generate

a sound wave, whose amplitude is proportional to the value of the arc column voltage. The

measuring device is not connected to the arc power supply, it is well shielded from

electromagnetic fields and the signal detected is related only to the voltage drop across the

arc itself.

2.3 Data Acquisition

2.3.1 Concepts

In nowadays concepts, data acquisition involves a lot more than a few sensors for

electrical signals. There are ever-increasing demands on the systems to record other types

of data as well. More data being acquired at the same time requires more powerful

computers and acquisition boards. It means that the capabilities of the computer can

significantly affect the performance of the Data AcQuisition (DAQ) system.

Twenty years ago, PCs were capable of transferring at rates around 5MHz, although

today’s computers can transfer significantly faster (National Instruments Inc. - Application

Note 007). They are capable of DMA (Direct Memory Access) and interrupt data transfers.

DMA transfers increase the system throughput by using dedicated hardware to transfer data

directly into system memory. Using this method, the processor is not burdened with moving

data and it is, therefore, free to engage in more complex processing tasks. To reap the

benefits of DMA or interrupt transfers, the DAQ device must be capable of these transfer

types. While PCI and FireWire devices offer both DMA and interrupt-based transfers,

PCMCIA and USB devices only use interrupt-based transfers. Depending on how much

processing is needed during data transfer, the rate at which the data is transferred from the

DAQ device to PC memory may be affected by the data transfer mechanism.

Besides these capability concerns, the choice of sensors it is also very important in a

DAQ system’s development. Also named transducers, they are devices that convert one type

of physical phenomenon, such as temperature, strain, pressure, or light into another

(National Instruments Inc. - Application Note 048). The most common transducers convert

physical quantities to electrical quantities, such as voltage or resistance. Transducer

characteristics define many of the signal conditioning requirements of the measurement

system. Table 2.2 summarizes the basic characteristics and conditioning requirements of

some common transducers.

The data acquisition devices can have analog input/output, digital input/output or

counters/timers. If it has all of them, it is called multifunction device. The number of channels,

the sampling rate, the resolution and the input range are parameters that specify the analog

inputs.

The number of analog channel inputs is specified for both single-ended and

differential inputs for devices with both input types. Single-ended inputs are all referenced to

a common ground reference. These inputs are typically used when the input signals are high

level (greater than 1V), the leads from the signal source to the analog input hardware are

short (less than 15ft), and all input signals share a common ground reference. If the signals

do not meet these criteria, it is necessary to use differential inputs. With differential inputs,

each input has its own ground reference; noise errors are reduced because the common-

mode noise picked up by the leads is cancelled out.

Table 2.2 - Transducers and signal conditioning requirements (National Instruments Inc.)

Sensor Electrical Characteristics Signal Conditioning Requirement

Thermocouple Low-voltage output

Low sensitivity

Nonlinear output

Reference temperature sensor (for

cold-junction compensation)

High amplification

Linearization

RTD Low resistance (100 ohms

typical)

Low sensitivity

Nonlinear output

Current excitation

Four-wire/three-wire configuration

Linearization

Strain gauge Low resistance device

Low sensitivity

Nonlinear output

Voltage or current excitation

High amplification

Bridge completion

Linearization

Shunt calibration

Current output

device

Current loop output (4 -- 20mA

typical)

Precision resistor

Thermistor Resistive device

High resistance and sensitivity

Very nonlinear output

Current excitation or voltage

excitation with reference resistor

Linearization

Active

Accelerometers

High-level voltage or current

output

Linear output

Power source

Moderate amplification

AC Linear

Variable

Differential

Transformer

(LVDT)

AC voltage output AC excitation

Demodulation

Linearization

Differential measurement systems are similar to floating signal sources in that the

measurement is made with respect to a floating ground that is different from the

measurement system ground. Neither of the inputs of a differential measurement system is

tied to a fixed reference, such as the earth or a building ground. Analog multiplexers in the

signal path are used in order to increase the number of measurement channels when only

one instrumentation amplifier exists. In Figure 2.7, the AIGND (analog input ground) pin is

the measurement system ground.

Figure 2.7 - Differential measurement system (National Instruments Inc.)

An ideal differential measurement system responds only to the potential difference

between its two terminals - the positive (+) and negative (-) inputs. A common-mode voltage

is any voltage measured with respect to the instrumentation amplifier ground present at both

amplifier inputs. An ideal differential measurement system completely rejects, or does not

measure, common-mode voltage. Rejecting common-mode voltage is useful because

unwanted noise often is introduced as common-mode voltage in the circuit that makes up the

cabling system of a measurement system. However, several factors, such as the common-

mode voltage range and the common-mode rejection ratio (CMRR) parameters, limit the

ability of practical, real-world differential measurement systems to reject the common-mode

voltage.

Referenced and non-referenced single-ended measurement systems are similar to

grounded sources in that the measurement is made with respect to a ground. A referenced

single-ended measurement system measures voltage with respect to the ground, AIGND,

which is directly connected to the measurement system ground. Figure 2.8 shows a 16-

channel referenced single-ended measurement system.

Figure 2.8 - Referenced single-ended measurement system (National Instruments Inc.)

DAQ devices often use a non-referenced single-ended (NRSE) measurement

technique, or pseudo-differential measurement, which is a variant of the referenced single-

ended measurement technique. Figure 2.9 shows a NRSE system.

Figure 2.9 - NRSE measurement system (National Instruments Inc.)

In a NRSE measurement system, all measurements are still made with respect to a

single-node analog input sense (AISENSE), but the potential at this node can vary with

respect to the measurement system ground (AIGND). A single-channel NRSE measurement

system works as a single-channel differential measurement system. Figure 2.10 summarizes

ways to connect a signal source to a measurement system.

Figure 2.10 - Connecting a Signal Source to a Measurement System (National Instruments

Inc.)

The sampling rate determines how often conversions can take place. A faster

sampling rate acquires more data in a given time and can therefore often form a better

representation of the original signal. Data can be sampled simultaneously, with multiple

converters, or it can be multiplexed, where the analog-to-digital converter (ADC) samples

one channel, switches to the next channel, samples it, switches to the next channel, and so

on. Multiplexing is a common technique for measuring several signals with a single ADC.

The range, resolution, and gain available on a DAQ device determine the smallest

detectable change in voltage. This change in voltage represents 1 least significant bit (LSB)

of the digital value and is often called the code width. The ideal code width is found by

dividing the voltage range by the gain times two raised to the order of bits in the resolution.

It is also important to consider the differential nonlinearity (DNL), relative accuracy, settling

time of the instrumentation amplifier and noise. As the level of voltage applied to a DAQ

device is increased, the digital codes from the ADC should also increase linearly (Application

Note 007 - National Instruments Inc.). If voltage versus the output code from an ideal ADC is

plotted, the result would be a straight line. Deviations from this ideal straight line are

specified as nonlinearity. DNL is a measure in LSB (Least Significant Bit) of the worst-case

deviation of code widths from their ideal value of 1 LSB. An ideal DAQ device has a DNL of 0

LSB. In practical terms, a good DAQ device will have a DNL of ±0.5 LSB. Poor DNL reduces

the resolution of the device.

Relative accuracy is a measure in LSBs of the worst-case deviation from the ideal

DAQ device transfer function, a straight line. Relative accuracy is determined on a DAQ

device by connecting a voltage at negative full scale, digitizing the voltage, increasing the

voltage, and repeating the steps until the input range of the device has been covered. When

the digitized points are plotted, the result must be a straight line. Obtaining good relative

accuracy requires that both the Analog-to-Digital Converter (ADC) and the surrounding

analog circuitry are properly designed.

Settling time is the time required for an amplifier, relays, or other circuits to reach a

stable mode of operation. This parameter has to be as lower as possible, even when working

with high gains and sampling rates, in order to avoid delays.

Any unwanted signal that appears in the digitized signal of the DAQ device is noise.

Because the PC is a noisy digital environment, acquiring data on a plug-in device takes a

careful layout on multiple-layer DAQ boards by skilled analog designers. Simply placing an

ADC, instrumentation amplifier, and bus interface circuitry on a one or two-layer board will

likely result in a very noisy device. Designers can use metal shielding on a DAQ device to

help reduce noise. Proper shielding not only should be added around sensitive analog

sections on a DAQ device, but also must be built into the layers of the device with ground

planes.

Analog output circuitry is often required to provide stimuli for a DAQ system. The

settling time, slew rate and output resolution determine the quality of the output signal

produced for the Digital-to-Analog Converter (DAC). Slew rate means the maximum rate of

change that the DAC can produce on the output signal. Settling time and slew rate work

together in determining how quickly the DAC changes the output signal level. Therefore, a

DAC with a small settling time and a high slew rate can generate high-frequency signals

because little time is needed to accurately change the output to a new voltage level.

An example of an application that requires high performance in these parameters is

the generation of audio signals. The DAC requires a high slew rate and small settling time to

generate the high frequencies necessary to cover the audio range. In contrast, an example of

an application that does not require fast D/A conversion is a voltage source that controls a

heater. Because the heater cannot respond quickly to a voltage change, fast D/A conversion

is unnecessary.

Digital I/O (DIO) interfaces are often used on PC DAQ systems to control processes,

generate patterns for testing, and communicate with peripheral equipment. In each case, the

important parameters include the number of digital lines available, the rate at which the

system can accept and source digital data on these lines, and the drive capability of the

lines. The number of digital lines, of course, should match the number of processes to be

controlled. The amount of current required to turn the devices on and off must be less than

the available drive current from the device. DIO can also be used in industrial applications, to

verify that a switch is open or closed and to check the voltage levels as high or low. It can

also be used for high-speed handshaking or simple communication methods. In addition,

some devices with digital capabilities will have handshaking circuitry for communication-

synchronization purposes. The number of channels, data rate, and handshaking capabilities

are all important specifications that should be understood and matched to the application

needs.

Counter/timer circuitry is useful for many applications, including counting the

occurrences of a digital event, digital pulse timing, and generating square waves and pulses.

It is possible to implement all these applications using three counter/timer signals – gate,

source, and output. The gate is a digital input used for enabling or disabling the function of

the counter. The source is a digital input that causes the counter to increment each time it

toggles, and therefore provides the time base for the operation of the counter. The output

generates digital square waves and pulses at the output line. The most significant

specifications for operation of a counter/timer are the resolution and clock frequency. The

resolution is the number of bits the counter uses. A higher resolution simply means that the

counter can count higher. The clock frequency determines how fast is possible to toggle the

digital source input. With higher frequency, the counter increments faster and therefore can

detect higher frequency signals on the input and generate higher frequency pulses and

square waves on the output.

2.3.2 Systems

Due to the different types of work environment, the data acquisition systems for

welding should be portable. According to Yapp (2004), many of them have been developed

an applied over the last twenty years, from relatively simple system which provide a printout

of average values to microprocessor based systems which incorporate sophisticated data

analysis. Nowadays, it is very common to use a laptop connected to the data acquisition

system for data processing. Since a laptop can be easily substituted if it gets obsolete,

insufficient hard disc capability for instance, this configuration provides flexibility for the

system. Moreover, all the acquired data and analysis can be stored into only one device, the

laptop, being easy to change of work environment. The data can also be transmitted through

a network, making possible a remote control. Figure 2.11 shows the “ArcSentry”, a system

which provides measurements at multiple cells, transferred to the server computer on either

a wire or RF based Ethernet LAN.

Figure 2.11 - ArcSentry System (N.A. Technologies Inc.)

This is a specific solution for welding. There are many other general purpose ones,

which only acquire, store and make the data available for processing. In this case, softwares

should be developed to analyse the acquired data. Figure 2.12 shows the OMB-DAQBOOK

portable data acquisition systems for laptop and desktop PCs which offer 12 or 16-bit and

100kHz acquisition rate. They also provide more than 700kbyte/s bidirectional data

communication to the PC via an enhanced parallel port (EPP) or PCMCIA link interfaces. A

Windows-based data logging application for setting up the acquisition applications and

saving acquired data directly to disk is also provided.

Figure 2.12 - OMB-DAQBOOK series (Omega Inc.)

Portable data acquisition systems can also have a briefcase shape. Figure 2.13

shows the DEWE-5000 an industrial PC integrated version and Figure 2.14 shows the

DEWE-BOOK series a notebook integrated version, both from DEWETRON. They are very

powerful data acquisition systems which record several types of signals such as video,

audio, CAN Bus data, position, speed, distance information from GPS satellites and so on,

simultaneously and in synch. Moreover, it is also offered A/D cards with 24-bit resolution,

simultaneous sampling on all channels (up to 25 MS/s sample rate per channel), robust anti-

aliasing filtering, and plug-in signal conditioning modules. Data can also be transmitted by

Ethernet or Wireless LAN.

The SCC DAQ system is another concept for data acquisition systems, proposed by

National Instruments (Signal Conditioning Overview). It consists of a shielded carrier Figure

2.15, signal conditioning modules (SCC modules - Figure 2.16 ), a DAQ device (Figure 2.17),

and a cable (Figure 2.18), providing conditioned analog signals directly passed to the inputs

of the DAQ device. The SCC modules can be connected according to the application offering

flexibility, customization, and ease of use in a single, low-profile package. They are either

single or dual-channel modules that condition analog or digital signals and are available for

thermocouples, RTDs, strain gauges, force/load/torque/pressure sensors, accelerometers,

voltage and current input, isolated voltage and current output frequency-to-voltage

conversion, low-pass filtering, isolated digital I/O, relay switching and bread-boarding for the

custom circuitry.

Figure 2.13 - DEWE-5000 series (DEWETRON Inc.)

Figure 2.14 - DEWE-BOOK series (DEWETRON Inc.)

They can also provide up to 300V of working isolation to voltage and current

input/output signals from the DAQ device. Optically isolated digital I/O modules can condition

digital lines from the DAQ device or be accessed directly using the 42-pin screw terminal

mounted inside the box. Relay modules add switching to the SCC DAQ system makes

possible to access Analog Input, Analog Output, Digital I/O, Counter/Timer signals as well as

timing and triggering signals from the DAQ device using feedthrough modules. It is also

possible to cascade two SCC analog input modules on a single analog input channel,

passing an analog input signal through both an attenuator module and a filter module, for

instance.

Figure 2.15 - Shielded Carriers (National Instruments Inc.)

Figure 2.16 - SCC Modules (National Instruments Inc.)

Figure 2.17 - DAQ Device (National Instruments Inc.)

Figure 2.18 - Shielded Cable (National Instruments Inc.)

2.4 LabView

LabVIEW is a graphical development environment by National Instruments for

creating flexible and scalable test, measurement and control applications rapidly and at

minimal cost. With LabVIEW, engineers and scientists interface with real-world signals,

analyze data for meaningful information and share results and applications. Regardless of

experience, LabVIEW makes development fast and easy for all users.

In Figure 2.19, the first panel is called Block Diagram. It is where the programming is

done. The blocks, which are functions or controls, are connected resulting in a graphical

code for the application. The second panel, called front panel, is where the controls are

placed. They can be graphs, displays, buttons, etc, composing the application interface.

Figure 2.19 - LabVIEW (National Instruments Inc.)

Data acquisition is simpler to be done with LabVIEW. National Instruments offers NI

DAQmx tool. This is a programming package installed on LabVIEW 7.0 and above for data

acquisition. With this package, it is possible to set the parameters of the physical channels of

the DAQ devices. It is done by creating virtual channels. It means that for each physical

channel is possible to have many different configurations. Thus, the same physical channel

of the DAQ device can be used by different applications. A virtual channel can be created by

using the DAQmx blocks or by the Measurement & Automation Explorer (MAX), which is an

application included in the LabVIEW package for device configuration. Besides the

configuration, it is also possible to calibrate the DAQ board for different environment

temperatures, what must be done every time the work environment is changed.

In the other hand, programming with LabVIEW is sometimes hard. The programmer

gets limited to its functions and if he or she needs a different one, it is necessary to write the

code in another programming environment and use it as a DLL or other format. Moreover,

the programmer spends a long time learning this new concept. The LabVIEW documentation

is still poor, making programming sometimes a nightmare. National Instruments offers many

training courses, but they are quite expensive. In spite of this, LabVIEW is much

recommended for research development because the code can be easily modified at any

time.

CHAPTER 3

METHODOLOGY AND DEVELOPMENT

As mentioned at the end of the Introduction (Chapter 1), the objective of this work

was the development of a data acquisition system projected for the study of the several types

of welding processes. This target, named here SMART, had to be able to do a synchronous

acquisition of electrical signals and frames of an USB camera, and trigger a high-speed (HS)

camera that would start to acquire frames, independently. The acquired data would be stored

for a frame-by-frame post-analysis. It means to correlate each frame of the USB or HS

camera images and the acquired signals (voltage, current, temperature, sound, etc), on a

time-base, characterizing a visual analysis of the arc welding.

Besides that, SMART had to offer portability, too. For this purpose, a PCMCIA

acquisition board and a laptop would be responsible for acquiring and storing data,

respectively. In order to make easy the transportation of the system, a briefcase shape had

to be adopted, where the cables, the signal conditioning module and the laptop could be

kept.

Flexibility had to be offered as well. Therefore, for each different type of signal to be

acquired, a suitable signal conditioning module would be plugged into a signal conditioning

unit. It means to make possible the acquisition of any kind of signal since it is used the

adequate module.

From the methodological point of view, the development of this system had to be

based on National Instruments (NI) resources (hardware and software). Cranfield University

had already bought most of the necessary equipment, which was NI brand, to develop this

system. Moreover, they also had an academic license of LabView, the programming

language used for developing the software. Besides that, the NI resources offer flexibility.

There are many types of signal conditioning modules which can be easily plugged to a signal

conditioning unit, all of them offered by National Instruments. In addition, future modifications

in a LabView based software are, usually, simple to be done.

The system had to be designed for up to the dual tandem welding process limits,

what means 4 voltages and 4 currents. Temperature measurement of at least 2 different

points had to be also available, as well as, an external trigger for the high-speed camera. By

using a laptop, these signals, the frames of the USB and the HS cameras had to be

synchronously acquired. Afterwards, it will be described the hardware and software projects

developed to compose this system.

3.1 Hardware Project

The standard configuration established for the hardware has connectors for 4

voltages, 4 currents, 2 temperatures and 1 trigger output. Connectors for digital I/Os, analog

outputs and additional analog inputs are also available. Since the development of the system

is based on National Instruments (NI) resources connected to a laptop, an ideal acquisition

board is the DAQCard-6062E, for PCMCIA bus. It has 16 single-ended analog inputs, an

input resolution of 12-bit and an input range from

0.05 to

10V. Taking these parameters,

it was possible to characterize the necessary signal conditioning modules for the intended

system configuration.

In order to use the whole scale background, it is necessary to amplify the input

signals making their maximum value equal to 10V. Since the welding signals can be DC

(Direct Current) or AC (Alternate Current), the taken minimum value for the scale background

would be -10V, but for the standard configuration was taken 0V, for DC signals only.

Equation 3.1 represents the gain for making this conditioning. Equation 3.2

represents the

minimum possible detectable value of the input signal after being conditioned, where

scale_gain is given by

Equation 3.3. If the input signal fits the whole scale background,

scale_gain is equal to 1, since conditioned_signal_range is equal to scale_range.

rangesignal

rangescale

gainngconditioni

_ =

2*_

gainscale

rangesignal

LSB =

rangescale

rangesignaldconditione

gainscale

_ =

Equation 3.1

Equation 3.2

Equation 3.3

Afterwards, the necessary conditioning for the standard signals was established.

Since NI-DAQmx coerces the input limits to fit a suitable scale range (see Table 3.1), it was

considered a different scale_range parameter for each signal.

Voltage

For signal_range from 0V to 100V the nearest scale_range goes from 0V to 10V.

1.0

0100

010

_ =

−

=gainngconditioni

mVLSB 25

0100

≅

−

It means that it is necessary a signal attenuator with gain equal to 0.1 and the lowest

detectable amplitude is 25mV. In this case, the NI SCC-A10 Voltage Attenuator Module is

the most adequate one.

Current

If signal_range goes from 0 to 1000A and current sensor sensibility is equal to

1mV/A, then signal_range goes from 0 to 1V and the nearest scale_range goes from 0 to 1V.

)01(

_ =

−

=gainngconditioni

mALSB 245

)01000(

≅

−

It means that amplification is not necessary and the lowest detectable amplitude is

245mA. Since it is not necessary any conditioning module and there is no any suitable signal

input for it in the NI SCC-2345 device (the signal conditioning unit chosen for this system),

devices for making possible these signal inputs were built, as Figure 3.1shows.

Figure 3.1: Input modules for current signals

Temperature

If signal_range goes from 0

C to 1600

C (considering that welding temperatures are

always positive) and R/S thermocouple sensibility is equal to 10µV/

C, then signal_range

Current Input

goes from 0 to 16mV and the nearest scale_range would go from 0 to 100mV. But

thermocouple signals must be amplified in order to increase the SNR (Signal-Noise Rate).

Establishing conditioning_gain equal to 100, signal_range goes from 0 to 1.6V and the

nearest scale_range goes from 0 to 2V.

8.0

6.1

_ ==gainscale

CLSB

5.0

2*8.0

)01600(

≅

−

It means that it is necessary a signal amplifier with gain equal to 100 and the lowest

detectable amplitude is 0.5

C. In this case, the NI SCC-TC02 Thermocouple Input Module is

the most adequate one.

Since the welding environment is very noisy, it is necessary to use a low-pass filter

for every analog input signal in order to avoid aliasing. For this purpose, a RC low-pass filter

(1 pole), as shown by Figure 3.2, was attached to each analog input connector. The cut-off

frequency is given by Equation 3.4.

Figure 3.2: Butterworth low-pass filter

Table 3.1: Input Ranges for NI 6020E, NI 6040E, NI 6052E, NI 6062E, and NI 6070E/6071E

acquisition boards

Range Configuration Gain Actual Input Range

0 to +10V

1.0

2.0

5.0

10.0

20.0

50.0

100.0

0 to +10V

0 to +5V

0 to +2V

0 to +1V

0 to +500mV

0 to +200mV

0 to +100mV

-5 to +5V

0.5

1.0

2.0

5.0

10.0

20.0

50.0

100.0

-10 to +10V

-5 to +5V

-2.5 to +2.5V

-1 to +1V

-500 to +500mV

-250 to +250mV

-100 to +100mV

-50 to +50mV

Equation 3.4

For

kHzfc 10=

and nFC 10= ,

Ω

kR 5.1 .

Actually, the ideal resistor would be R=1.6kΩ, but the nearest value existent in the

laboratory was 1.5kΩ, which produces fc=10.61kHz. This cut-off frequency was chosen in

order to permit high-frequencies acquisition, for instance the audio signals (in this case, an

amplifier module has to be used). The Nyquist Theorem says that the sampling rate has to

be at least 2 times the highest signal frequency to be acquired. It means that a minimum

sampling rate for each channel is 20kS/s. When a narrow frequency range is wanted, a

digital filter has to be used. For this purpose, a Butterworth low-pass digital filter was

implemented for the system. The cut-off frequency has always to be less than half sampling

rate. So, the cut-off frequency range for this filter goes from 1 to 10kHz and can have 1, 2, 3,

4 or 5 poles.

The acquisition board of the system is the DAQCard-6062E which has 16 single-

ended analog inputs, 2 analog outputs, 8 digital I/Os, 2 counter/timers and 1 analog trigger.

The sampling rate is 500kS/s. Since it is used only 12 analog inputs for the standard

configuration, the maximum sampling rate for each channel is around 40kS/s. The resolution

for the analog outputs is also 12-bit, the output range

10V and the output rate is 850kS/s.

Following the idea of portability, it was used a briefcase to compose the system, as

shown by Figure 3.3. An aluminium panel was built in order to place the I/O connectors. It

can be opened and the laptop can be kept inside the briefcase, on the lid of the signal

conditioning unit placed on the bottom of the briefcase. Eventually, the briefcase’s lid can be

used for keeping loose cables.

Figure 3.3: Briefcase used for composing the system

I/O connectors

briefcase’s lid

panel’s door

Since most of sensors have BNC connectors as standard, it was used 8 female BNC

panel connectors for the 4 voltage and 4 current input signals. In order to connect the voltage

signals to their related inputs, 4 cables (2 meters length each) were built, being one extremity

finished with a male BNC connector and the other one finished with two crocodile

connectors, according to Figure 3.4.

Figure 3.4: Cable built for voltage signals

The current sensors have already their cables. For the temperature signals it was

used 2 mini thermocouple panel connectors and 2 extension cables (2 meters length each),

as Figure 3.5 shows.

Figure 3.5: Cable for temperature signals

For the trigger of the high-speed camera, it was used 2 female jack banana

connectors and 1 cable, being one extremity finished with 2 male jack banana connectors

and the other one finished with 1 female BNC connector, according to Figure 3.6. This cable

has a circuitry as shown by Figure 3.7, which works as an on/off switch.

Figure 3.6: Cable for the trigger of the high-speed camera

Figure 3.7: Circuitry for the trigger cable

It was also used 1 female DB9 connector for digital I/Os, just in case of future

modifications in the system need to use it. For additional signals, a hole was left. Figure 3.8

shows the idealized final panel.

Figure 3.8: Panel built for the system’s briefcase

The signal conditioning unit lies on the bottom of the briefcase. A top door gives

access to this unit, as Figure 3.9 shows. On this door, the laptop can be placed, composing a

tidy work environment. Figure 3.10 illustrates how the input and output terminals can be used

for connecting signals, as the trigger output, for example.

Figure 3.9: Signal conditioning unit

In order to provide more details of the arc welding, it was attached a lens with an

adapter to the USB camera, according to Figure 3.11. This lens is a Helios 44M and its focal

distance is 58mm.

Filters are necessary to decrease the incident arc brightness and gives definition to

the arc image. They are Cokin brand and have the following specifications:

• Ultra-violet (U.V) P231

• Neutral Grey ND4 P153

• Neutral Grey ND2 P152

• Neutral Grey ND8 P154

• Dark

The computer used for performing this system is DELL brand with a Xeon ™ Intel®

processor, 2.80GHz, 1GB RAM and USB 2.0 ports to plug the webcam.

signal conditioning

unit’s door

input and output

terminals

Figure 3.10: Detail that schemes the input and output terminals of the signal conditioning unit

and shows the trigger output

Figure 3.11: Webcam with magnifying lens and filters attached

3.2 Software Project

The software was developed in LabVIEW 7.1 for Windows XP because it is an easy

programming language and the standard in the laboratory, since the university has an

academic license provided by National Instruments. Programming in LabView is relatively

simple. The objects which belong to the function are placed on the “Front Panel”. This

function can be configured to work as a window and will show up, when executed. The code

is built on the “Block Diagram”, where these objects are linked to other functions. Afterwards,

it will be described the software development. The following figures can be better visualized

on the computer, where zoom tools are available, than on the printed version of this thesis.

Figure 3.12 shows the main window. All the signals selected by using the button

“Channels” are plotted on the graph “Signal”. The three buttons on the left and under this

graph make possible to move the cursor and zoom the plots. The boxes on the right show

the current position of the cursor. The three other buttons can be used for configuring the

plots and move the cursor.

The webcam frames are placed on “USB Camera” and the high-speed camera

frames on “HS Camera”. There are five buttons under these image graphs: play, backward,

forward, pause and stop. Pressing play, the webcam frames play until pause or stop is

pressed, or the last frame is reached. Pressing pause, the current frame is shown and it is

the reference for next executions. Pressing stop, the current frame is also shown, but the first

frame (frame zero) will be the next one, when a new execution is done. Pressing the

backward and the forward buttons, the previous and the next frames are shown, respectively.

The box “Frame” displays the number of the frame.

Pressing “Start” the acquisition starts and the red LED is turned on. Pressing “Stop”

the acquisition stops and this LED is turned off. The box “Project” contains the path of the

folder where the project file is saved.

Figure 3.13 shows the code that builds the menus which appear on the top of the

main window when the software is executed. They are File (items: New File, Open File, Save

as xls), Acquisition (items: Setup Signal, Setup Image, Filter) and Camera (item: Settings).

After building the menus, the software becomes a main sequence divided into three

frames. The first one makes the initialization of all variables. The second one makes the

calibration of the system. And the third one, more complex, is the core of the system.

Figure 3.14 shows that first frame, which contains the default values of all variables

used by the software. If there is an USB camera plugged to the computer, a session for its

image is created. If there is no any USB camera plugged to the computer, it does not happen

and the window that sets the parameters of the USB camera cannot be opened.

Figure 3.12: Illustration of the main window, referred in the project as function “Main”

Figure 3.13: Illustration of the code that creates the menu of the main window

The value of the variable “Delay” is set as 0.0026 seconds. This is a constant value

used for calculating the frame of the HS camera that should be plotted, when a frame-by-

frame analysis is executed. This value was chosen after some trials, as chapter 4 explains.

The value of the variable “dt USB” is 0.1 seconds. This is a constant interval of time

used for acquiring each frame of the USB camera and read the correspondent buffer of data

coming from the acquisition board.

The other variables and objects receive default values, which can be modified during

the execution of the system.

Figure 3.14: Illustration of the code that initializes the variables and objects of the function

“Main”

Afterwards, the system is calibrated. First, it is verified if there is any acquisition board

plugged to the computer. In a negative case, the warning message “The system cannot be

calibrated because there is no acquisition board and/or USB camera connected” appears, as

Figure 3.15 shows. Otherwise, the acquisition board is calibrated and reset, as Figure 3.16

illustrates. It is necessary to reset the acquisition board in order to invalidate every immediate

task associated with it. Then, it becomes immediately available for any new task.

Figure 3.15: Illustration of the code that shows a warning message when no acquisition

board is plugged to the computer

Figure 3.16: Illustration of the code that calibrates and resets the acquisition board

After calibrating and resetting the acquisition board, it is also verified if any USB

camera is plugged to the computer. In a negative case, another warning message appears

“The system cannot be calibrated because there is no USB camera connected”, according to

Figure 3.17. Otherwise, a short acquisition is done (one second), as Figure 3.18 shows. This

was necessary because the system was presenting a delay before starting to acquire data,

when it was run for the first time. Probably this is due to its dependence on the webcam.

Figure 3.17: Illustration of the code that shows a warning message when no USB camera is

connected

Figure 3.18: Illustration of the code that calibrates the system

Figure 3.19 represents the function “Image Test”, which makes this short acquisition.

First, all the initializations are done, according to Figure 3.20. Afterwards, the data acquisition

is carried out, although no any sample is saved, as Figure 3.21 illustrates.

Figure 3.19: Illustration of the front panel of the function “ImageTest”

Figure 3.20: Illustration of the code that initializes the variables and objects of the function

“ImageTest”

Figure 3.21: Illustration of the code of the function “ImageTest” that does a short acquisition

for the calibration of the system

The third frame of the sequence is more complex. It is composed by event frames.

The first one is “Menu Selection”, which is a sequence of frames, the items of the menu.

Figure 3.22 shows the first one, “New File…”, where nothing happens if a new project file is

not chosen, otherwise those variables are reset and a new project file is created. This file

receives the extension *.prj and it is where the values of the variables used during the data

acquisition are saved. Thus, it is not necessary to reconfigure the system every time it runs.

Figure 3.23 illustrates the function responsible for creating new files. A window asking

the name of the new project file shows up. If it is not cancelled and the chosen name already

exists, a dialog box does the following question “Do you want to replace <file>?”. If “Yes”, this

file is deleted and a new one is created. If “No”, nothing happens. If the file does not exist, a

new one is straight created. Figure 3.24 and Figure 3.25 illustrate how it works.

The function “Create”, represented by Figure 3.26 is responsible for creating three

new blank files with the extensions *.txt (for signal samples), *.xls (for converting the txt file to

an excel file) and *.avi (for videos), with the same name of the project file. If these files

already exist, they are deleted. Figure 3.27 shows how it works.

Figure 3.22: Illustration of the code that shows the item “New File…” of the “Menu Selection”

Figure 3.23: Illustration of the front panel of the function “New File”

Figure 3.24: Illustration of the code that shows the function “New File” asking a name for the

files to be created

Figure 3.25: Illustration of the code that shows the function “New File” asking a name for the

file to be straight created

Figure 3.26: Illustration of the front panel of the function “Create”

Figure 3.27: Illustration of the code of the function “Create” that creates three new files with

the extensions txt, xls and avi

The next item of the “Menu Selection” is “Open File…”, as Figure 3.34 shows. The

function responsible for opening the project is illustrated by Figure 3.28. A windows asking its

file name shows up. If it is not cancelled and the file exists, the function “Read” (Figure 3.31)

is called and returns the TXT and USB file paths, as Figure 3.29 shows. Otherwise, these

paths are empty strings, as Figure 3.30 illustrates. Function “Read” reads the project file, as

Figure 3.32 shows, and searches the *.txt and *.avi files into the project file’s folder. If they

exist, their paths are also read, as Figure 3.33 shows.

Figure 3.28: Illustration of the front panel of the function “Open Project”

Figure 3.29: Illustration of the code of the function “Open Project” that asks the name of the

project to be opened and returns txt and avi paths

Figure 3.30: Illustration of the code of the function “Open Project” that asks the name of the

project to be opened and returns empty strings for txt and avi paths

Figure 3.31: Illustration of the front panel of the function "Read"

Figure 3.32: Illustration of the code of the function “Read” that opens the project and reads its

data

Figure 3.33: Illustration of the code of the function “Read” that looks for the txt and avi files of

the read project

Figure 3.34: Illustration of the code that shows the item “Open File…” of the “Menu Selection”

After opening the files and reading their contents, the variables and objects of the

system are set. If the txt path is not empty, the number of samples is read by function “NF

Signal” (Figure 3.37 and Figure 3.38) and the signals are plotted, as Figure 3.35 shows.

Otherwise, only the variable “dt Signal” is set, as shown by Figure 3.36.

Figure 3.35: Illustration of the code that calls the function “NF Signal” and plots the signals

Figure 3.36: Illustration of the code that sets a default value for the variable “dt Signal”

Figure 3.37: Illustration of the front panel of the function "NF Signal"

Figure 3.38: Illustration of the code of the function “NF Signal” that reads the txt file and its

number of samples

When a project file is opened, a window shows up asking for the correspondent HS

camera file. If it is not cancelled and the avi path is not empty, the function “NF HS” (Figure

3.42 and Figure 3.43) reads its number of frames. If the chosen file is the same of the USB

camera file, the error message “Error! Choose a different file for the High Speed Camera”

shows up. Figure 3.39, Figure 3.40 and Figure 3.41 illustrate how it works.

If the avi path of the USB camera file is not empty, the function “NF USB” (Figure

3.46 and Figure 3.47) reads its number of frames. Figure 3.44 and Figure 3.45 show how it

works.

Afterwards, some ratios are calculated. The variable “dt HS” is divided by “dt Signal”

and “dt HS” in order to find out how many signal samples and USB frames, respectively,

correspond to a HS frame. “Delay” is divided by “dt HS” in order to find out how many HS

frames have to be shifted. And “dt USB” is divided by “dt Signal” and “dt HS” in order to find

out how many signal samples and HS frames, respectively, correspond to an USB frame.

This is illustrated by Figure 3.48. If any of these ratios is less then 1, it is inverted, as shown

by Figure 3.49. Then, Depending on the case (direct or inverse), a different type of execution

is set. It can be 1 (direct), 0 (inverse) or -1 (no camera file is opened), Figure 3.50 illustrates

this case.

Figure 3.39: Illustration of the code that asks the HS camera file to be opened and calls the

function “NF HS”

Figure 3.40: Illustration of the code that shows an error message if the HS camera file is the

same of the USB camera file

Figure 3.41: Illustration of the code that sets default values for the variables “NF HS” and “dt

HS”, just in case no HS camera file is opened

Figure 3.42: Illustration of the front panel of the function "NF HS"

Figure 3.43: Illustration of the code of the function “NF HS” that reads the number of frames

of the HS camera file

Figure 3.44: Illustration of the code that calls the function “NF USB”

Figure 3.45: Illustration of the code that sets a default value for the variable “NF USB”

Figure 3.46: Illustration of the code of the function “NF USB”

Figure 3.47: Illustration of the code of the function “NF USB” that reads the number of frames

of the USB camera file

Figure 3.48: Illustration of the code that calculates the ratios among the frames of each

camera and the signal samples

Figure 3.49: Illustration of the code that calculates the inverse ratios among the frames of

each camera and the signal samples

Figure 3.50: Illustration of the code that sets the variables to be used during the frame-by-

frame analysis if no camera file is opened

The next “Menu Selection” frame is “Save as xls…”. This item converts a txt file into a

xls file. First, the message “Saving .xls File…” appears on the toolbar of the main window

and the progress bar is activated, as Figure 3.51 shows. Then, the function “Save XLS”

(Figure 3.56, Figure 3.57 and Figure 3.58) asks the txt file to be converted and the function

“Read txt File” (Figure 3.59, Figure 3.60 and Figure 3.61) creates an xls path. If it is not

cancelled, the progress bar is set to zero, as Figure 3.52 shows. If a file with the created xls

path already exists, it is deleted by function “Delete” (Figure 3.62 and Figure 3.63), as Figure

3.53 illustrates.

Afterwards, the conversion is done by function “xls” (Figure 3.64 and Figure 3.65), as

Figure 3.54 shows. First, Excel is opened by function “Open Excel and Make Visible” (Figure

3.66 and Figure 3.67). Then, a workbook and a worksheet are created, by functions “Open

Specific Workbook” (Figure 3.68 and Figure 3.69) and “Open Specific Worksheet” (Figure

3.70 and Figure 3.71), respectively. After that, function “Row Col To Range Format” (Figure

3.75 and Figure 3.76) converts the column and the row, which are numeric, into a range

format of excel, that identifies the cell where a value is placed. Function “Set Cell Value with

Range” (Figure 3.72, Figure 3.73 and Figure 3.74) places the values into the cells.

Eventually, the toolbar is cleaned and the progress bar is inactivated, as Figure 3.55

shows.

Figure 3.51: Illustration of the code that shows a message and a progress bar on the toolbar

of the main window, before starting to convert a txt file into a xls file

Figure 3.52: Illustration of the code that sets a progress bar on the toolbar of the main

window

Figure 3.53: Illustration of the code that calls the function “Delete XLS”

Figure 3.54: Illustration of the code that calls the function “XLS”

Figure 3.55: Illustration of the code that inactivates the progress bar and cleans the toolbar of

the main window

Figure 3.56: Illustration of the front panel of the function "Save XLS"

Figure 3.57: Illustration of the code of the function “Save XLS” that asks the txt file to be

converted and calls the function “Read txt File”

Figure 3.58: Illustration of the code of the function “Save XLS” that returns an empty string

for the xls and txt files, case no txt file is opened

Figure 3.59: Illustration of the front panel of the function "Read txt File"

Figure 3.60: Illustration of the code of the function “Read txt File” that returns the paths of the

xls and txt files

Figure 3.61: Illustration of the code of the function “Read txt File” that returns an empty string

if no any txt file exists

Figure 3.62: Illustration of the front panel of the function "Delete"

Figure 3.63: Illustration of the code of the function “Delete” that deletes an existent xls file

Figure 3.64: Illustration of the front panel of the function "xls"

Figure 3.65: Illustration of the code of the function "xls" that converts a txt file into a xls file

Figure 3.66: Illustration of the front panel of the function "Open Excel and Make Visible"

Figure 3.67: Illustration of the code of the function "Open Excel and Make Visible" that opens

an excel application

Figure 3.68: Illustration of the front panel of the function "Open Specific WorkBook"

Figure 3.69 Illustration of the code of the function "Open Specific WorkBook" that opens a

specific workbook of an excel application

Figure 3.70: Illustration of the front panel of the function "Open Specific WorkSheet"

Figure 3.71: Illustration of the code of the function "Open Specific WorkSheet" that opens a

specific worksheet of a workbook of an excel application

Figure 3.72: Illustration of the front panel of the function "Set Cell Value with Range"

Figure 3.73: Illustration of the code of the function "Set Cell Value with Range" that writes a

value into a excel cell

Figure 3.74: Illustration of the code of the function "Set Cell Value with Range" that does not

write a value into an excel cell, if an error occurs

Figure 3.75: Illustration of the front panel of the function "Row Col To Range Format"

Figure 3.76: Illustration of the code of the function "Row Col To Range Format" that converts

the column and the row, which are numeric, into a range format of excel

Figure 3.77 illustrates the item “Exit” of the “Menu Selection”. If the variable “Camera”

is true, means that an USB camera session was previously opened and must be closed.

Otherwise, this is not necessary. Then, the software is quit.

Figure 3.77: Illustration of the code that shows the item “Exit…” of the “Menu Selection” that

closes an USB camera session and quits the software

The item “Setup Signal…” of the “Menu Selection” is shown by Figure 3.78. The

window “Setup Signal” (Figure 3.79) shows up for setting the acquisition parameters:

Channel, Clock Settings, Acquisition Mode and Advanced Clock Settings (actually, these

ones are not used by the software). If the button OK is pressed, the global variables are

upgraded. If the button Cancel is pressed, it does not happen. Pressing any of these buttons,

this window is closed. Figure 3.80, Figure 3.81, Figure 3.82, Figure 3.83 and Figure 3.84

show how it works.

Figure 3.78: Illustration of the code that shows the item “Setup Signal…” of the “Menu

Selection” that calls the function “Setup Signal”

Figure 3.79: Illustration of the window "Acquisition Setup", referred in the project as function

"Setup Signal"

Figure 3.80: Illustration of the code of the function "Setup Signal" that sets the values of the

objects, which appear on this window, with the values kept by global variables

Figure 3.81: Illustration of the code of the function "Setup Signal" that disables the objects

“Clock Source” and “Active Edge”, case the object “Clock Type” is internal (value 0)

Figure 3.82: Illustration of the code of the function "Setup Signal" that upgrades the global

variables “Clock Source” and “Active Edge”, case the object “Clock Type” is internal (value 0)

Figure 3.83: Illustration of the code of the function "Setup Signal" that enables the objects

“Clock Source” and “Active Edge”, case the object “Clock Type” is external (value 1)

Figure 3.84: Illustration of the code of the function "Setup Signal" that upgrades the global

variables “Clock Source” and “Active Edge”, case the object “Clock Type” is external (value

Figure 3.85 shows the item “Setup Image…” of the “Menu Selection”. The window

“Setup Image” (Figure 3.86) is used for selecting the trigger channel and the USB camera.

The box “fps” means frames per second acquired by the USB camera. Normally, the

webcams can acquire from 15 to 30fps. In order to avoid a repeated frame, the chosen value

for this parameter is 10fps. Another reason, it is the quantity of information to be plotted

during the acquisition. A bigger interval of time is better because the plotting function is

called less times. Figure 3.87, Figure 3.88, Figure 3.89, Figure 3.90 and Figure 3.91 show

how this setting of parameters is done.

Figure 3.85: Illustration of the code that shows the item “Setup Image…” of the “Menu

Selection” that calls the function “Setup Image”

Figure 3.86: Illustration of the window "Setup Image", referred in the project as function

"Setup Image"

Figure 3.87: Illustration of the code of the function "Setup Image" that sets the values of the

objects, which appear on this window, with the values kept by global variables, and verifies if

any USB camera is plugged to the computer and if any USB camera session was started

(true case)

Figure 3.88: Illustration of the code of the function "Setup Image" that verifies if any USB

camera is plugged to the computer and if any USB camera session was started (false case)

Figure 3.89: Illustration of the code of the function "Setup Image" that verifies if USB cameras

are plugged to the computer and lists them

Figure 3.90: Illustration of the code of the function "Setup Image" that upgrades the global

variables, if the button OK is pressed

Figure 3.91: Illustration of the code of the function "Setup Image" that does not upgrade any

global variable, if the button Cancel is pressed

Figure 3.92 shows the item “Filter…” of the “Menu Selection”. The window “Filter”

(Figure 3.93) asks the values for the cut-off frequency and the number of poles of the

Butterworth low-pass digital filter. Figure 3.94, Figure 3.95 and Figure 3.96 show how it

works.

Figure 3.92: Illustration of the code that shows the item “Filter…” of the “Menu Selection” that

calls the function “Filter”

Figure 3.93: Illustration of the window "Low- pass filter", referred in the project as function

"Filter"

Figure 3.94: Illustration of the code of the function "Filter" that sets the limit values for the

objects that appear on this window

Figure 3.95: Illustration of the code of the function "Filter" that sets the values of the objects

that appear on this window

Figure 3.96: Illustration of the code of the function "Filter" that upgrades the global variables,

if the button OK is pressed and does not upgrade anything if the button Cancel is pressed

The item “Settings…” of the “Menu Selection” (Figure 3.97) calls the window

“Camera” (Figure 3.98), responsible for setting the USB camera parameters. Each webcam

has different types of parameters which can be set. Thus, pushing the button “Image”, it is

possible to set the image resolution, for instance. And pushing the button “Video”, it is

possible to configure brightness, saturation, sharpness, gamma, mode and other

parameters.

Figure 3.97: Illustration of the code that shows the item “Settings…” of the “Menu Selection”

that calls the function “Camera”

Figure 3.98: Illustration of the window "Camera Settings", referred in the project as function

"Camera"

After doing this, those parameters are kept and the image acquisition of the USB

camera will be done with the last configuration. Figure 3.99, Figure 3.100, Figure 3.101,

Figure 3.102, Figure 3.103, Figure 3.104, Figure 3.105 and Figure 3.106 show how it works.

Figure 3.99: Illustration of the code of the function “Camera” that enables the buttons “Video”

and “Image”, and shows the name of the USB camera indicated by the global variable

“Camera”

Figure 3.100: Illustration of the code of the function “Camera” that initializes a new USB

camera session and shows the USB camera frames, if none of the objects on the window

“Camera Settings” is changed in 10ms

Figure 3.101: Illustration of the code of the function “Camera” that initializes a new USB

camera session if the object “USB Cameras” changes, meaning that a new USB camera was

chosen

Figure 3.102: Illustration of the code of the function “Camera” that stops the current USB

camera session, opens the “Image Settings” window of that camera and starts to play the

frames again using the new configuration

Figure 3.103: Illustration of the code of the function “Camera” that stops the current USB

camera session, opens the “Video Settings” window of that camera and starts to play the

frames again using the new configuration

Figure 3.104: Illustration of the code of the function “Camera” that upgrades the value of the

global variable “Camera” if the button OK is pressed

Figure 3.105: Illustration of the code of the function “Camera” that does not upgrade anything

if the button Cancel is pressed

Figure 3.106: Illustration of the code of the function “Camera” that does not execute anything

if there is no USB camera plugged to the computer

If none of the items of the “Menu Selection” is selected, a default value equal to

“false” is used for keeping the software waiting for an event, without executing any function,

as Figure 3.107 shows.

Figure 3.107: Illustration of the code that shows the item “Default” of the “Menu Selection”

that does not do anything if no other event happens

The next frame of the sequence of events is “Channels Button”, as Figure 3.108

shows. The window “Channel” (Figure 3.109) is used for selecting the channels of which

signals will be plotted during the acquisition. If it is cancelled and there is no signal already

acquired, nothing happens. Otherwise, the signals related to the selected channels (variable

“User Channels”) are extracted from the buffer of signals (variable “Signal”) and plotted.

Figure 3.108: Illustration of the code that shows the event “Channels Button” that calls the

function “Channel” and sets the colour of the plots on the main window

Figure 3.109: Illustration of the window "Channels", referred in the project as function

"Channels"

When the window “Channel” is opened, the last configuration has to be shown. For

this purpose, the window “Channel” is firstly cleaned and only the names of the channels are

kept, as shown by Figure 3.110. Then, the selected channels receive a “check” symbol

indicating that they are selected. A sequence of colours (size equal to the quantity of

available channels) is built. Each channel receive a colour which will be the same used for

plotting its acquired signal, as Figure 3.111 shows.

Figure 3.110: Illustration of the code of the function “Channel” that cleans the object “Local

User Channels”, responsible for listing the names of the set channels

Figure 3.111: Illustration of the code of the function “Channel” that sets the object “Local

User Channels”

Then, the current selected channels are saved just in case the selection of channels

is cancelled later, as Figure 3.112 illustrates. After that, an event is expected. If an item is

selected, a “check” symbol is used. If the “OK” button is pushed, the current “User Channels”

are updated. Otherwise, if the “Cancel” button is pushed, the previous configuration is

recovered. Figure 3.113, Figure 3.114 and Figure 3.115 show how it works.

Figure 3.112: Illustration of the code of the function “Channel” that saves the current selected

channels just in case the selection of channels is cancelled

Figure 3.113: Illustration of the code of the function “Channel” that upgrades the global

variable “User Channels”, if the button OK is pressed

Figure 3.114: Illustration of the code of the function “Channel” that rewrites the original

selection of channels on the object “Local User Channels”, if the button Cancel is pressed

Figure 3.115: Illustration of the code of the function “Channel” that puts a check mark in front

of a selected channel, if the object “Local User Channels” is clicked

The next event is “Start”, responsible for doing the data acquisition. First, the current

project is deleted, as shown by Figure 3.116. Then the function “Create” (Figure 3.117)

checks if txt, xls or avi files, which belong to the current project, already exist. In a positive

case, they are also deleted. Eventually, paths for these files are created, as Figure 3.118

shows.

Then, the items of the menu on the window “Main” are disabled, as well as the “Start”

button and the buttons related to the cameras. The other objects are reset and the variables

related to the acquisition configuration are saved into the project file following the order

below:

• order: it is the order of the low-pass digital filter;

• fl: it is the cut-off frequency of the low-pass digital filter;

• Camera: it is the position of the webcam in the listed USB cameras plugged to the

computer;

• Trigger: it is the name of the channel used for triggering the high-speed camera;

• Pulses: it is the number of pulses necessary to trigger the high-speed camera (it is

always equal to 1);

• Acquisition Mode: it can be continuous or finite, assuming the values 10123 or 10178,

respectively;

• Clock Type: it can be internal or external, assuming the values 0 or 1, respectively;

• Active Edge: it can be rising or falling, assuming the values 0 or 1, respectively.

• Sample Time: it is the time in seconds case the acquisition mode is finite; it assumes

discrete values between 1 and 60 seconds.

• Rate: it is the sampling rate per channel.

• Clock Source: it specifies the channel for the sample clock if an external clock is

used;

• Channels: these are the channels used for doing the acquisition;

• User Channels: these are the selected channels to be plotted;

• Color Graph: these are the colours used for the selected channels.

Figure 3.116: Illustration of the first frame of the event “Start” that deletes the current project

Figure 3.117: Illustration of the front panel of the function "Create"

Figure 3.118: Illustration of the code of the function “Create” that creates paths for txt, xls and

avi files of the current project

In order to avoid any kind of interference during the acquisition, all the possible

combinations of types of execution had to be previously done. Thus, the variable “Execution”

was created. If there is no webcam plugged, trigger for the high-speed camera is not solicited

and there is at least one selected channel, this variable becomes equal to 3, as Figure 3.119

shows. If there is a webcam plugged, “Execution” becomes equal to 4, as Figure 3.120

illustrates. If the USB camera is not plugged but the trigger is solicited, this variable assumes

value 1, as shown by Figure 3.121. If the webcam is plugged, the trigger is solicited and

there is at least one selected channel, this variable assumes value 2, as Figure 3.122 shows.

The default value for this variable is 0.

Figure 3.119: Illustration of the second frame of the event “Start” that sets all objects on the

main window, saves all global variables into the project file and sets the variable “Execution”

with value 3

Figure 3.120: Illustration of the second frame of the event “Start” that sets the variable

“Execution” with value 4

Figure 3.121: Illustration of the second frame of the event “Start” that sets the variable

“Execution” with value 1

Figure 3.122: Illustration of the second frame of the event “Start” that sets the variable

“Execution” with value 2

After setting the variable “Execution”, if its value is equal to 0, nothing happens, as

Figure 3.123 shows. Otherwise, the type of execution is selected and the acquisition starts.

The program is divided into 3 threads: one responsible for webcam images acquisition,

another one responsible for reading data from the acquisition board and a main one

responsible for synchronizing the two first threads and plotting the signals and images on the

window “Main”. The communication among them is done by notifiers. This procedure is

shown by Figure 3.124.

Figure 3.123: Illustration of the third frame of the event “Start” that does not do anything if the

variable “Execution” is equal to 0

Figure 3.124: Illustration of the third frame of the event “Start” that calls the function “Image

1” and plots the acquired data

Since a webcam frame is acquired every 0.1 second (10fps), a buffer of signals has

to be read into the same interval of time. Function “Image1” (Figure 3.125) is responsible for

doing the acquisition when the variable “Execution” is equal to 1. After setting the necessary

variables (Figure 3.126) and checking the acquisition mode, which can be finite or

continuous, the data acquisition starts, as Figure 3.127, Figure 3.128 and Figure 3.129 show.

If the acquisition is continuous, the button “Stop” has to be pushed in order to stop the

acquisition. If the acquisition is finite, the button “Stop” has to be pushed or the “Sample

Time” has to be reached.

Figure 3.125: Illustration of the front panel of the function "Image1"

Figure 3.126: Illustration of the code of the function "Image1" that sets the inputs to be used

by the functions during the data acquisition

Figure 3.127: Illustration of the code of the function "Image1" that calls functions “Start2” and

“Butterworth” and acquires data during a determined interval of time

Figure 3.128: Illustration of the code of the function "Image1" that calls functions “Start2” and

“Butterworth” and acquires data until the button Stop of the main window is pressed

Figure 3.129: Illustration of the code of the function "Image1" that waits for the fifth USB

camera frame to start the data acquisition

Function “Start2” (Figure 3.130) starts to acquire the signals and shots the trigger in

order to start the HS camera image acquisition, as shown by Figure 3.131 and Figure 3.132.

Being the delay between these two moments deterministic, it is possible to calculate the

correct frame to be plotted, during a frame-by-frame analysis.

The function “Butterworth” (Figure 3.133) is a Butterworth low-pass digital filter which

filters the acquired data. LabVIEW can filter only one signal each time. Thus, 16 filters were

put in parallel and can filter up to 16 channels at the same time, as shown by Figure 3.134.

Figure 3.130: Illustration of the front panel of the function "Start2"

Figure 3.131: Illustration of the code of the function "Start2" that starts reading the inputs of

the acquisition board

Figure 3.132: Illustration of the code of the function "Start2" that shots the trigger of the HS

camera

Figure 3.133: Illustration of the front panel of the function “Butterworth"

Figure 3.134: Illustration of the code of the function “Butterworth" that filters a data buffer of

each input channel

If the variable “Execution” is equal to 2 (Figure 3.135), the function “Image2” (Figure

3.136) does the acquisition. In this case, the USB camera frames are also acquired. Since

the webcam presented an irregular behaviour when acquiring the first frames, the acquisition

actually starts after disposing the first five ones, as shown by Figure 3.137, Figure 3.138,

Figure 3.139 and Figure 3.140. It assures a regular quantity of brightness and a rate of 10fps

for all frames. This procedure is done even when no webcam image is acquired, what

provides a similar behaviour for all types of executions.

If the variable “Execution” is equal to 3 (Figure 3.141), the acquisition is done by

function “Image 3” (Figure 3.142). In this case, only the signals are acquired and the HS

camera is not triggered, as Figure 3.143, Figure 3.144, Figure 3.145 and Figure 3.146 show.

Thus, the function “Start1” (Figure 3.147) is used for starting the data acquisition, as shown

by Figure 3.148 and Figure 3.149.

If the variable “Execution” is equal to 4 (Figure 3.150), the acquisition is done by

function “Image 4” (Figure 3.151). In this case, the signals are acquired as well as the USB

camera frames, but the HS camera is not triggered, as Figure 3.152, Figure 3.153, Figure

3.154 and Figure 3.155 show.

Figure 3.135: Illustration of the third frame of the event “Start” that calls the function “Image2”

and plots the acquired data

Figure 3.136: Illustration of the front panel of the function "Image2"

Figure 3.137: Illustration of the code of the function "Image2" that sets the inputs to be used

by the functions during the data acquisition

Figure 3.138: Illustration of the code of the function "Image2" that calls functions “Start2” and

“Butterworth” and acquires data during a determined interval of time

Figure 3.139: Illustration of the code of the function "Image2" that calls functions “Start2” and

“Butterworth” and acquires data until the button Stop of the main window is pressed

Figure 3.140: Illustration of the code of the function "Image2" that waits for the fifth USB

camera frame to start the data acquisition

Figure 3.141: Illustration of the forth frame of the event “Start” that calls the function “Image3”

and plots the acquired data

Figure 3.142: Illustration of the front panel of the function "Image3"

Figure 3.143: Illustration of the code of the function "Image3" that sets the inputs to be used

by the functions during the data acquisition

Figure 3.144: Illustration of the code of the function "Image3" that calls functions “Start1” and

“Butterworth” and acquires data during a determined interval of time

Figure 3.145: Illustration of the code of the function "Image3" that calls functions “Start1” and

“Butterworth” and acquires data until the button Stop of the main window is pressed

Figure 3.146: Illustration of the code of the function "Image3" that waits for the fifth USB

camera frame to start the data acquisition

Figure 3.147: Illustration of the front panel of the function "Start1"

Figure 3.148: Illustration of the code of the function "Start1" that starts reading the inputs of

the acquisition board (internal clock case)

Figure 3.149: Illustration of the code of the function "Start1" that starts reading the inputs of

the acquisition board (external clock case)

Figure 3.150: Illustration of the fifth frame of the event “Start” that calls the function “Image4”

and plots the acquired data

Figure 3.151: Illustration of the front panel of the function "Image4"

Figure 3.152: Illustration of the code of the function "Image4" that sets the inputs to be used

by the functions during the data acquisition

Figure 3.153: Illustration of the code of the function "Image4" that calls functions “Start1” and

“Butterworth” and acquires data during a determined interval of time

Figure 3.154: Illustration of the code of the function "Image4" that calls functions “Start1” and

“Butterworth” and acquires data until the button Stop of the main window is pressed

Figure 3.155: Illustration of the code of the function "Image4" that waits for the fifth USB

camera frame to start the data acquisition

After doing the data acquisition, the button “Stop” is disabled. Then, the system is

prepared to convert the acquired data, which is saved into a txt file, to an xls file. For this

purpose, a dialog box asks if this conversion is desirable. In a positive case, a toolbar, on the

main window, shows a progress bar indicating how the conversion goes. This procedure is

shown by Figure 3.156. If any xls file already exists it is deleted by function “Delete”, as

Figure 3.157 shows. Eventually the conversion starts, as Figure 3.158 illustrates. At the end,

the items of the menu and the other objects on the window “Main” are set with the previous

values, as shown by Figure 3.159.

After acquiring the signals and images and opening the correlated files, a frame-by-

frame analysis can be done. Pressing the button “forward HS”, it is added 1 frame to the

variable “f HS”, as Figure 3.160 shows. If this value is greater than the quantity of acquired

frames, nothing happens. Otherwise, the next HS camera frame is shown as well as the

correspondent signal buffer and USB camera frame.

Figure 3.156: Illustration of the sixth frame of the event “Start” that sets objects on the main

window

Figure 3.157: Illustration of the sixth frame of the event “Start” that calls the function “Delete”

Figure 3.158: Illustration of the sixth frame of the event “Start” that asks if it is desirable to

create a xls file; in a positive case, the function “xls” is called

Figure 3.159: Illustration of the sixth frame of the event “Start” that sets the main window with

the previous values

Figure 3.160: Illustration of the code of the event “forward HS” that adds 1 frame to the

variable “f HS” (case -1)

100

The variables “dF HS Signal” and “dF HS USB” are the quantity of samples and USB

camera frames, respectively, that have to be considered for each HS camera frame. The

variables “P HS Signal” and “P HS USB” decide how to calculate which signal buffer and

USB camera frame have to be plotted, respectively. If these variables are equal to -1 (Figure

3.160), nothing happens. Otherwise, if they are equal to 0 (Figure 3.161), it means that the

sampling rate of the HS camera frames is less than the sampling rate of the signals and the

USB camera images. In this case, the variable “f HS” is divided by variables “dF HS Signal”

and “dF HS USB”. If they are equal to 1 (Figure 3.164), the sampling rate of the HS camera

frames is greater than the sampling rate of the signals and the USB camera images. In this

case, the variable “f HS” is multiplied by the variables “dF HS Signal” and “dF HS USB”.

The variable “DSignal” is always equal to zero, because the signals samples are not

delayed. Function “Camera USB” (Figure 3.162) is responsible for plotting the calculated

USB camera frame if it is equal or greater than zero and less than the quantity of acquired

frames, as Figure 3.163 shows. Eventually, function “Camera HS” (Figure 3.166 and Figure

3.167) plots the solicited HS camera frame, as shown by Figure 3.165.

Figure 3.161: Illustration of the code of the event “forward HS” that adds 1 frame to the

variable “f HS” (case 0)

101

Figure 3.162: Illustration of the front panel of the function "Camera USB"

Figure 3.163: Illustration of the code of the function "Camera USB" that plots an USB camera

frame

Figure 3.164: Illustration of the code of the event “forward HS” that adds 1 frame to the

variable “f HS” (case 1)

102

Figure 3.165: Illustration of the code of the of the event “forward HS” that calls the function

“Camera HS”

Figure 3.166: Illustration of the front panel of the function "Camera HS"

Figure 3.167: Illustration of the code of the function "Camera HS" that plots a HS camera

frame

Pressing the button “backward HS”, it is decreased 1 frame from the variable “f HS”. If

this value is equal or less than -1, nothing happens. Otherwise, the next HS camera frame is

shown as well as the correspondent signal buffer and USB camera frame. For this purpose,

103

the same procedure previously explained is executed, as illustrated by Figure 3.168 and

Figure 3.169.

Figure 3.168: Illustration of the code of the event “backward HS” that decreases 1 frame of

the variable “f HS” (case 0)

Figure 3.169: Illustration of the code of the event “backward HS” that calls the function

“Camera HS”

Pressing the button “forward USB”, it is added 1 frame to the variable “f USB”. If this

value is greater than the quantity of acquired frames, nothing happens. Otherwise, the next

104

USB camera frame is shown as well as the correspondent signal buffer and HS camera

frame.

The variables “dF USB Signal” and “dF USB HS” are the quantity of samples and HS

camera frames, respectively, that have to be considered for each USB camera frame. The

variables “P USB Signal” and “P USB HS” decide how to calculate which signal buffer and

HS camera frame have to be plotted, respectively. If these variables are equal to -1, nothing

happens. Otherwise, if they are equal to 0 (Figure 3.170), the sampling rate of the USB

camera images is less than the sampling rate of the signals and HS camera images. In this

case, the variable “f USB” is divided by variables “dF USB Signal” and “dF USB HS”. If they

are equal to 1 (Figure 3.171), the sampling rate of the USB camera images is greater than

the sampling rate of the signals and HS camera images. In this case, the variable “f USB” is

multiplied by variables “dF USB Signal” and “dF USB HS”.

Variable “DHS” represents the delay mentioned before, in terms of frames, and it is

used for shifting the HS camera frames. Function “Camera HS” is responsible for plotting the

calculated HS camera frame, if it is equal or greater than zero and less than the quantity of

acquired frames. Eventually, function “Camera USB” plots the USB camera frame.

Figure 3.170: Illustration of the code of the event “forward USB” that adds 1 frame to the

variable “f USB” (case 0)

105

Figure 3.171: Illustration of the code of the event “forward USB” that adds 1 frame to the

variable “f USB” (case 1)

Pressing the button “backward USB”, it is decreased 1 frame from the variable “f HS”.

If the result is equal or less than -1, nothing happens. Otherwise, the next USB camera frame

is shown as well as the correspondent signal buffer and HS camera frame. For this purpose,

the same procedure previously explained is executed, as illustrated by Figure 3.172.

Typing the desired USB frame in the box “f USB”, it is checked if the typed value is

into the acceptable limits. In a negative case, the previous value is shown in the box “f USB”,

as Figure 3.173 illustrates. Otherwise, it is executed the same procedure explained when the

button “forward USB” is pressed, as shown by Figure 3.174.

106

Figure 3.172: Illustration of the code of the event “backward USB” that decreases 1 frame of

the variable “f USB” (case 0)

Figure 3.173: Illustration of the code of the event “f USB” that keeps the same USB camera

frame if the typed frame is not under the acceptable limits

107

Figure 3.174: Illustration of the code of the event “f USB” that plots an USB camera frame if

the typed frame is under the acceptable limits (case 0)

Typing the desired HS frame in the box “f HS”, it is checked if the typed value is into

the acceptable limits. In a negative case, the previous value is shown in the box “f HS”,

Figure 3.175 illustrates. Otherwise, it is executed the same procedure explained when the

button “forward HS” is pressed, as shown by Figure 3.176.

When the button “play HS” is pressed, the current images are disposed and the other

buttons related to the cameras are disabled, except for the buttons “pause HS” and “stop

HS”, as Figure 3.177 shows. These buttons stop playing the frames. If the execution is

paused, the variables “f HS” and “f USB” receive the last played frames. If it is stopped, these

variables receive the value -1. In this case, zero will be the first frame of the next execution.

This procedure is shown by Figure 3.183.

Function “Play Camera HS” (Figure 3.181) plays the HS camera frames using the

playback rate defined when the files are opened for analysis, as Figure 3.182 shows. The

correspondent signal buffer and USB camera frame are calculated and plotted for each

played HS camera frame. In order to avoid any delay, the USB camera file is opened only

once, in the beginning, and closed when the execution is paused or stopped. This procedure

is shown by Figure 3.178, Figure 3.179 and Figure 3.180.

108

Figure 3.175: Illustration of the code of the event “f HS” that keeps the same HS camera

frame if the typed frame is not under the acceptable limits

Figure 3.176: Illustration of the code of the event “f HS” that plots an HS camera frame if the

typed frame is under the acceptable limits (case 0)

109

Figure 3.177: Illustration of the code of the event “play HS” that sets the buttons on the main

window

Figure 3.178: Illustration of the code of the event “play HS” that opens the USB camera file

110

Figure 3.179: Illustration of the code of the event “play HS” that calls the function “Play

Camera HS”

Figure 3.180: Illustration of the code of the event “play HS” that calls the function “Camera

loop”

111

Figure 3.181: Illustration of the front panel of the function "Play Camera HS"

Figure 3.182: Illustration of the code of the function “Play Camera HS” that plays the HS

camera frames

Figure 3.183: Illustration of the code of the event “play HS” that resets the objects on the

main window if the acceptable limits are reached

112

Function “Camera loop” (Figure 3.184) is responsible for playing the USB camera

frames, as Figure 3.185 shows. Eventually, all the buttons are enabled again, as illustrated

by Figure 3.186.

Figure 3.184: Illustration of the front panel of the function "Camera loop"

Figure 3.185: Illustration of the code of the function "Camera loop" that plays the USB

camera frames

When the button “play USB” is pressed, the current images are disposed and the

other buttons related to the cameras are disabled, except for the buttons “pause USB” and

“stop USB”, as shown by Figure 3.187. These buttons stop playing the frames. If the

execution is paused, the variables “f USB” and “f HS” receive the last played frames. If it is

stopped, these variables receive the value -1. In this case, zero will be the first frame of the

next execution. Figure 3.193 illustrates this procedure.

Function “Play Camera USB” (Figure 3.191) plays the USB camera frames at a rate

of 10fps, as Figure 3.192 shows. The correspondent signal buffer and HS camera frame are

113

calculated and plotted for each USB camera frame. In order to avoid any delay, HS camera

file is opened only once, in the beginning, and closed when the execution is paused or

stopped. Figure 3.188, Figure 3.189 and Figure 3.190 show how it works. Function “Camera

loop” (Figure 3.184) is responsible for playing the HS camera frames, as illustrated by Figure

3.185. Eventually, all the buttons are enabled again, as Figure 3.194 shows.

Figure 3.186: Illustration of the code of the event “play HS” that resets the buttons on the

main window

Figure 3.187: Illustration of the code of the event “play USB” that sets the buttons on the

main window

114

Figure 3.188: Illustration of the code of the event “play USB” that opens the HS camera file

Figure 3.189: Illustration of the code of the event “play USB” that calls the function “Play

Camera USB”

115

Figure 3.190: Illustration of the code of the event “play USB” that calls the function “Camera

loop”

Figure 3.191: Illustration of the front panel of the function "Play Camera USB"

Figure 3.192: Illustration of the code of the function “Play Camera USB” that plays the USB

camera frames

116

Figure 3.193: Illustration of the code of the event “play USB” that resets the objects on the

main window if the acceptable limits are reached

Figure 3.194: Illustration of the code of the event “play USB” that resets the buttons on the

main window

117

Moving the cursor and clicking on the graph, the cameras frames correspondent to

the clicked signal value are also shown. The x axis value taken by the cursor is divided by

the variables “dt USB” and “dt HS” (they are the inverse of the camera sampling rate) in

order to find out the correct frames to be plotted. This procedure is shown by Figure 3.195.

If the window “Main” is closed and an image session had been previously initialized, it

is closed as well, as shown by Figure 3.196. All the available images are also disposed,

making free the space they occupied in memory.

All the global variables used in the program are shown by Figure 3.197. They are

necessary because more than one function needs and modifies those variables.

Figure 3.195: Illustration of the code of the event “Graph Plots” that plots the correspondent

frames of the cameras on the main window

Figure 3.196: Illustration of the code of the event “Panel Close?” that closes an USB camera

session if the main window is closed

118

Figure 3.197: Illustration of the front panel that shows all the global variables

Each function created is a thread. It means sequences of code to be executed

following priorities, which can be: background (lowest), normal, above normal and time

critical (highest). The window “Main” and the functions “Image1”, “Image2”, “Image3”,

“Image4”, “Start1” and “Start2” have time critical priority. These functions are responsible for

the data acquisition and need to occupy more times the PC processor, in order to provide a

deterministic execution. The other functions have normal priority.

LabVIEW also permits to set the “Preferred Execution System”, which can be: user

interface, standard, instrument I/O, data acquisition, other1, other2, same as caller. This tool

helps to organize the software structure. All the created windows are set as “user interface”.

The functions previously cited are set as “data acquisition”. All the other ones are set as

“same as caller”.

Figure 3.198 shows how these parameters can be set by the windows “VI Properties”

of LabVIEW.

Figure 3.198: Illustration of the window "VI Properties" that sets the priority and type of

execution of the functions created in the software

119

The diagram shown by Figure 3.199 summarizes the software. It is divided into 3

parts: Menu, Acquisition an Analysis and contains the created functions and their hierarchy.

The full lines mean that the following function belongs to the previous one. And the traced

lines mean that the following function does not belong to the previous one, but it is executed

in that order. Figure 3.200 presents the software’s data flow.

After doing the code, it was necessary to build a stand alone application (the

executable version). LabVIEW offers this option and permits to create an installer as well.

For this purpose, it is necessary to add the files “ImaqDirectShowDLL.dll” and “NIVisSvc.dll”

which can be found into the folder C:\Windows\system32.

Before installing the created application, it is necessary to install the National

Instruments Vision Assistant 7.1 that has a separated license and contains the video

functions used by the application. Besides that, it is also necessary to install the NI-IMAQ for

USB Cameras (can be downloaded from www.ni.com

for free), which contains the USB

functions used for acquiring frames of USB cameras.

This final version was obtained after many trial versions. In order to have

synchronism among signals and images, an ideal situation would be to provide an

instantaneous trigger for the high-speed camera, although, this is not possible by software.

Thus, a suitable solution was to provide a deterministic delay when the trigger is produced.

Then, it would be possible to calculate the correct high-speed camera frame.

A real-time version was considered. For this purpose, it was installed the LabVIEW

Real Time 7.1 (RTX). This version of LabVIEW runs the applications using a real-time

extension for Windows XP. This extension is another operating system, which means

another kernel, it is called RTX and it is provided by Venturcom. An academic license of

LabView does not provide RTX, so it has to be purchased apart and it is quite expensive.

The found solution was to contact Venturcom and ask a trial version of RTX 5.5 (later

versions are not acceptable for LabVIEW Real Time 7.1). It is very important to uninstall

service pack 2 before installing RTX 5.5, otherwise nothing works. It is also necessary to

install DAQmx Base 1.5.1 (can be downloaded from www.ni.com

). This is the package used

for programming data acquisition in a real-time environment. The following steps have to be

executed in order to make LabVIEW Real-Time (RTX) works:

• Install RTX 5.5 (if the LabView license does not include it)

• Install LabView 7.1 (if it was not previously installed)

• Install LabView 7.1 Real Time (RTX)

• Install DAQmx Base 1.5.1 (it contains the functions for data acquisition in Real-

Time)

120

• Update device driver (the driver of the acquisition board driver has to be changed

to RTX supported)

Figure 3.199: Illustration of a diagram that summarizes how the software works

121

122

123

Figure 3.200: Illustration of the software's data flow

124

The problems in using this configuration are many. First of all, the DAQmx Base does

not have all the functions of the DAQmx and the entire application has to run by the same

kernel, what makes impossible to maintain all the previous code resources. There are no

available drivers for PCMCIA acquisition boards, which means that a laptop could not be

used for doing the acquisition. Finally, there are no available drivers for USB cameras, so the

webcam could not be integrated to the system. Those are the main reasons to not use the

real-time version of LabVIEW.

Another possibility was studied, the LabVIEW Real-Time (ETS). This version

executes the application on hardware targets running the real-time operating system of

Venturcom Phar Lap Embedded Tool Suite (ETS). It provides a real-time operating system

that runs on NI RT Series hardware to meet the requirements of embedded applications that

need to behave deterministically or have extended reliability requirements. This solution was

not suitable because one of the objectives of this work was to build a flexible application to

run on laptops, not an embedded one.

3.3 Calibration

Calibration is necessary in order to achieve higher measurement accuracy when the

work environment changes. There are two types of calibration, device calibration and

channel calibration. Device calibration consists of verifying the measurement accuracy of a

device and adjusting for any measurement error. It means to measure the performance of

the device and compare these measurements to the published specifications. During

calibration, voltage levels or other signals are supplied and read using external standards,

and then the device calibration constants are adjusted. The new calibration constants are

stored in the memory (EEPROM). These calibration constants are loaded from memory as

needed to adjust for the error in the measurements taken by the device.

One type of device calibration is the external calibration, which is typically performed

by a metrology lab, which uses a high-precision voltage source to verify and adjust

calibration constants. This procedure replaces all calibration constants in the EEPROM and

is equivalent to a factory calibration. Because the external calibration procedure changes all

EEPROM constants, it invalidates the original calibration certificate. If an external calibration

is done with a National Institute of Standards and Technology (NIST) certified voltage

source, a new NIST traceability certificate can be issued. It should be done at a regular

interval as defined by the measurement accuracy requirements of the application. National

125

Instruments recommends a complete calibration at least once every year. This interval can

shorten to 90 days or six months.

Another type of device calibration is the self-calibration which does not require any

external connections. It adjusts the calibration constants with respect to an onboard

reference stored on the device. The new calibration constants are defined with respect to the

calibration constants created during the external calibration to ensure that the measurements

are traceable to these external standards. The new calibration constants do not affect the

constants created during an external calibration because they are stored in a different area of

the device memory. A self-calibration can be performed at any time to adjust the device for

use in environments other than those in which the device was externally calibrated.

Channel calibration is necessary in applications where the highest degree of

accuracy is critical, but it does not replace device calibration. Channel calibration

compensates for various errors, including those introduced by cabling, wiring, and sensors.

SMART does a self-calibration every time the system is initialized. The other types of

calibration have to be done if a high accuracy of the measurements is necessary. For

welding parameters measurement, the self-calibration is usually enough.

3.4 System configuration

The first step to configure the system is to create virtual channels through the

Measurement & Automation Explorer (MAX), which can be downloaded from www.ni.com

for free. This application also contains the drivers for the NI acquisition boards. Thus, it is

advisable to download its last version, which is supposed to be more complete. Figure 3.201

illustrates how to create the virtual channels.

126

Figure 3.201: Illustration of the panel “Measurement and Automation Explorer” that creates a

Virtual Channel

The measurement type has to be selected as Figure 3.202 shows. Voltage and

current signals are both analog inputs. Figure 3.203 shows how the standard configuration

for voltage signals was done. Since the output of the current sensor is a voltage signal,

current signals have to be configured as voltage inputs. In this case, a scale according to the

specifications of the sensor has to be build. Figure 3.204 and Figure 3.205 show how to

create a scale. Figure 3.206 shows how the standard configuration for current signals was

done.

The configuration of the voltage and current terminals are non-referenced single-

ended (NRSE) because the measurement of these signals is made with respect to the same

node, but the potential at this node can vary with respect to the measurement system ground

(since different types of sensors are used). This type of configuration makes possible to use

a greater number of inputs.

In order to measure temperature, virtual channels were configured as thermocouples

of R/S type, as Figure 3.207 illustrates. In this case, the configuration of the terminals is

differential because the thermocouple outputs are differential as well.

A virtual channel for the external trigger is actually a pulse output, as Figure 3.208

shows. Figure 3.209 illustrates its standard configuration.

127

When the software runs for the first time, it takes a while to calibrate the system and

reset the acquisition board. After that, the next step is to configure the software. First, it is

necessary to create a new file or to open an existent one, as Figure 3.210 illustrates. After

doing this, the name of the current file appears in the box “Project”.

Figure 3.202: Illustration of the panel “Measurement and Automation Explorer” that chooses

the measurement type for voltage and current signals

Figure 3.203: Illustration of the panel “Measurement and Automation Explorer” that

configures the voltage signals

128

Figure 3.204: Illustration of the panel “Measurement and Automation Explorer” that creates a

NI-DAQmx Scale

Figure 3.205: Illustration of the panel “Measurement and Automation Explorer” that selects a

type of scale

129

Figure 3.206: Illustration of the panel “Measurement and Automation Explorer” that

configures a scale to be used for current signals

Figure 3.207: Illustration of the panel “Measurement and Automation Explorer” that

configures the measurement type of the temperature signals

130

Figure 3.208: Illustration of the panel “Measurement and Automation Explorer” that

configures the measurement type of the external trigger

If after opening an existent file a new one is created, the previous parameters are

kept, otherwise the default ones last. It is possible to change signal, image and filter

parameters, as Figure 3.211 shows. In order to change the signal parameters, the desired

channels have to be selected among the created virtual channels. The sampling rate per

channels has also to be defined (it is recommended 20kHz or above, since there is a low-

pass filter with cut-off frequency equal to 10kHz attached to each input connector). If the

acquisition mode is Finite Samples, the Sample Time has to be set (the minimum value is 1

second and the maximum value is 60 seconds). If the acquisition mode is Continuous, it is

advisable to worry about the space available in the HD of the computer, what limits the

acquisition time.

131

Figure 3.209: Illustration of the panel “Measurement and Automation Explorer” that

configures the external trigger

Figure 3.210: Illustration of the menu used for creating or opening a file

In order to set the image parameters, it is necessary to select the channel for the

trigger of the high-speed camera among the created virtual channels and the desired

132

webcam. Eventually, it is necessary to set the cut-off frequency (it has to be less than half of

the sampling rate) and the number of poles for the digital filter.

Figure 3.211: Illustration of the menu used for configuring the data acquisition

After configuring the acquisition parameters, the next step is to configure the

webcam, as Figure 3.212 shows. Different webcams have different types of parameters. The

webcam used for performing the final tests is Logitech brand and was configured as

following:

• Brightness = 50%

• Saturation = 0%

• Sharpness = 50%

• Gamma = 0%

• Black and White mode

The resolution must be always equal to 640x480 and the diaphragm of the magnify

lens must be closed.

Before starting the acquisition, it is necessary to select the channels of which signals

will be shown on the signal graph. It can be done, pressing the “Channels” button and

selecting the desired channels. After that, pressing the “Start” button, the data acquisition

starts.

133

Figure 3.212: Illustration of the menu used for configuring the USB camera

134

CHAPTER 4

EVALUATION OF THE SYSTEM “SMART”

Two sets of experimental trials were done using the GMAW process. The first one

was performed with single wire and the second one with double wire (tandem). A third set of

experimental trials was performed with Resistance Spot Welding (RSW) process. In all of

them, current and voltage signals were collected, as well as, video camera images.

4.1 First set of experimental trials – single wire GMAW

The objective of these trials was to verify the possibility of the arc welding

visualization by using a webcam. Besides that, the efficiency of the synchronism among

electrical signals (welding current and arc voltage), USB camera frames and HS camera

frames had to be also verified. The HS cameras are usually used by researchers in order to

obtain more detailed images of the arc welding.

Figure 4.1 and Figure 4.2 show the rig configuration used for the first set of trials.

Two plain carbon steel plates were positioned (as illustrated by Figure 4.3) over the carriage,

which moves at a constant speed when it is turned on, by a controller, before starting the

data acquisition. The torch head was driven by a step motor, in such a way that it could be

oscillated in a horizontal movement when wished. The only possible travel angle of this torch

head was the right angle. Then, the first trial of this set was performed just to check any

possible image distortion.

The travel angle refers to the angle in which welding takes place. It may be a right

angle, a push angle or a drag angle, depending on the position of the torch. When the torch

is pointing back to the weld direction, it is known as pulling (or dragging) the weld. When the

torch is pointing forward to the weld direction, it is referred to as pushing the metal. At the

right angle, the torch is perpendicular to the weld direction. Once Figure 4.3 (a) does not

illustrate any of these cases, the denomination “plate angle” will be used for referring to the

angle on that illustration.

The high-speed camera was set with a sampling rate of 5kHz, at a resolution of

256X128 pixels and a playback rate of 5Hz. As explained in section 3.2, the USB camera

135

frames are acquired at 10fps, with a resolution of 640x480, by a Trust webcam. The image

parameters were set as following:

• Brightness = 50%

• Saturation = 0%

• Sharpness = 50%

• Gamma = 0%

• Black and White mode

The diaphragm of the magnify lens was closed and filters were used, as explained in

section 3.1. Concerning the electrical signals, different values of sampling rates were

adopted in order to verify if it would affect the synchronism between signal samples and HS

camera frames. Table 4.1 summarizes the signal sampling rate, plate angle and torch

oscillation values, used for doing the six experimental trials of this set.

The distance between the torch and the webcam was about 700mm, while the high-

speed camera was positioned about 2000mm away from the torch, both at the same

alignment in relation the longitudinal centreline of the weld bead. No digital filter, as

mentioned in section 3.2, was employed at first.

Figure 4.1: Illustration of the ring configuration used for doing the first set of trials (single wire

GMAW)

Webcam

High speed camera

136

Figure 4.2: (a) side view of the test plate, (b) test plate carriage, (c) torch head and the (d)

webcam, in details

Figure 4.3: Schematic arrangement of the plates (3mm - thick) for welding in relation to the

torch position: (a) at a plate angle of 135

; (b) at a plate angle of 90

; (c) and a side view of

the test plate

(a)

(b)

(c)

(d)

137

Table 4.1: Parameter setting for the first set of trials

Trial Signal Sampling Rate (kHz) Plate Angle (º) Torch Oscillation (Hz)

test1 5 135 0

test2 5 90 0

test3 5 90 5

test4 13 90 0

test5 4 90 4

test6 5 90 5

Results

After setting the system, bead-on-plate welds were carried out. In order to evaluate

the level of similarity between the waveforms provided by this device and the system, the

signal samples were also acquired by a digital oscilloscope at 5kHz. All of these waveforms

seemed to be very similar.

Good results related to the arc image, acquired by the webcam, were presented.

Since there was only one wire, the arc length could be detected, even if the torch oscillated.

However, by comparing Figure 4.4 and Figure 4.5, it is possible to see that the detection of

the arc length would be more difficult for a “plate angle” of 135

Figure 4.4: Illustration of a shot of the first

set of trials - Test1 (“plate angle” equal to

135

) - voltage traces (top), USB frames

(left bottom) and HS frames (right bottom)

Figure 4.5: Illustration of a shot of the first

set of trials - Test2 (“plate angle” equal to

) - voltage traces (top), USB frames

(left bottom) and HS frames (right bottom)

138

A synchronism between signals and webcam frames was observed. When the arc

welding is turned on, the webcam frame that shows this moment and the buffer of signal

samples, which represents its arc voltage value, are plotted, as shown by Figure 4.6 and

Figure 4.7. In a similar behaviour, when the arc welding is turned off, the webcam frame that

shows this moment and the its buffer of signal samples, which are equal to zero, are plotted,

as shown by Figure 4.8 and Figure 4.9.

Nevertheless, no synchronism between signals and high-speed camera frames was

observed. It is due to a delay existent between the moments that the acquisition board starts

the signal acquisition and triggers the high-speed camera. This delay was not even constant,

making impossible to synchronize those signals. Then, it was necessary to modify the

program in order to guarantee a more deterministic delay, according to section 3.2.

Figure 4.6: Illustration of a shot of the

second set of trials - Test2 (there is still no

arc welding) - voltage traces (top), USB

frames (left bottom) and HS frames (right

bottom)

Figure 4.7: Illustration of a shot of the

second set of trials - Test2 (there is

already an arc welding) - voltage traces

(top), USB frames (left bottom) and HS

frames (right bottom)

139

Figure 4.8: Illustration of a shot of the

second set of trials - Test2 (the arc

welding is being extinguished) - voltage

traces (top), USB frames (left bottom) and

HS frames (right bottom)

Figure 4.9: Illustration of a shot of the

second set of trials - Test2 (the arc

welding is already extinguished) - voltage

traces (top), USB frames (left bottom) and

HS frames (right bottom)

In order to analyse the effect of using different sampling rates for acquisition of signal

samples and high-speed camera frames, it was forced the synchronization of these data by

considering the delay occurred in each trial. When the sampling rate used for acquiring the

signals is not a multiple value of the sampling rate used for acquiring the high-speed camera

frames, the accurate of the frame-by-frame analysis decreases. It is because the system

calculates the quantity of signal measurements that should be considered at each frame, by

dividing the sampling rate used for acquiring signals by the sampling rates used for acquiring

frames of USB and HS cameras. If the results are not integer values, there will be an error

that propagates frame-by-frame.

4.2 Second set of experimental trials – double wire GMAW

Figure 4.10 shows the rig configuration for the second set of trials, which had the

same objective of the first one, although it was performed with double wire GMAW. As it is

140

possible to see, only one plain carbon steel plate (20mm - thick) was positioned over the

carriage (what provides a “plate angle” always equal to 90

), which also moves at a constant

speed when it is turned on, by a controller, before starting the data acquisition. The torch

head could not oscillate.

The high-speed camera setting did not change, that is, a sampling rate of 5kHz, at a

resolution of 256X128 pixels, and a playback rate of 5Hz. A Logitech webcam was used for

performing these trials and it was set as the webcam used for performing the first set of trials.

Its frames were also acquired at 10fps, with a resolution of 640x480. The diaphragm of the

magnify lens was also closed and the same filters were used, as explained in section 3.1.

Concerning the electrical signals, a sampling rate of 25kHz for all trials was used. The

low-pass digital filter had already been implemented (as item 3.2 explains) and the

configuration of its parameters was cut-off frequency equal to 1000kHz and number of poles

equal to one.

The cameras were in the same position but not at the same alignment, being the

distance between the torch and the cameras about 700mm.

Figure 4.10: (a) test plate, (b) test plate carriage, (c) torch head, (d) high-speed camera and

(e) the webcam, in details, illustrating the rig configuration used for doing the second set of

trials (tandem GMAW)

(a)

(b)

(c)

(e) (d)

141

Results

Four trials for this set were performed. In the first one, the arc image acquired by the

webcam was not satisfactory. Due to the presence of two wires under a out-of-phase pulsed

current, when one arc is turned brighter (pulse current) the other one is turned dimmed (base

current) and vice-versa, so the arc image is not very defined, as Figure 4.11 shows. Thus,

this first trial was disposed and the webcam was not used for performing the next ones.

Figure 4.11: Illustration of a shot of the second set of trials - Test 1 (disposed trial) - voltage

traces (top), USB frames (left bottom) and disabled HS frames (right bottom)

The modifications implemented in order to make constant the delay explained in

section 3.2 were positive. Table 4.2 shows the results obtained by connecting the output that

triggers the high-speed camera to one of the input channels of the acquisition board. Then,

by acquiring this signal, it would be possible to determine the delay, before going to a real

trial, as shown by Figure 4.12. It characterized a simulation of trials, which found a delay

value around 0.002650 seconds. After that, bead-on-plate welds were carried out. Table 4.3

shows the results obtained. It is possible to verify the determinism in the delay values. Once

the sampling rate normally used for acquiring frames of high-speed cameras in the welding

area is around 5kHz, a variation of 0.0002 seconds is acceptable. With this sampling rate,

0.0002 seconds represents 1 frame. Thus, 1 frame can be in advance or delayed, during a

frame-by-frame analysis.

142

Table 4.2: Trigger delay - Simulation of trials

Trial Continuous Samples

Delay (s)

Finite Samples

Delay (s)

1 0.002499 0.002650

2 0.002550 0.002650

3 0.002552 0.002649

4 0.002599 0.002652

5 0.002650 0.002798

6 0.002450 0.002599

Table 4.3: Trigger delay - Final tests

Trial Delay (s)

Test2 0.002509

Test3 0.002600

Test4 0.002597

Figure 4.12: Illustration of a shot of the second set of trials - Test 3 (determination of the

trigger delay, by simulation) - trigger pulse (top), disabled USB frames (left bottom) and

disabled HS frames (right bottom)

143

A frame-by-frame analysis of the high-speed camera frames show that these frames

are adequately synchronized with the electrical signals. When the arc welding of the torch on

the right is turned on (Figure 4.13), the cursor on the graph shows a pulse peak of the

voltage signal. When it is turned off (Figure 4.14), it is shown a pulse base of that signal.

Figure 4.15 and Figure 4.16 show that the same happens for the next peak.

Figure 4.13: Illustration of a shot of the

second set of trials - Test 1 (the arc

welding is turned on) - voltage traces (top),

disabled USB frames (left bottom) and HS

frames (right bottom)

Figure 4.14: Illustration of a shot of the

second set of trials - Test 1 (the arc

welding is turned off) - voltage traces (top),

disabled USB frames (left bottom) and HS

frames (right bottom)

144

Figure 4.15: Illustration of a shot of the

second set of trials - Test 1 (the arc

welding is turned on again) - voltage

traces (top), USB frames (left bottom) and

HS frames (right bottom)

Figure 4.16: Illustration of a shot of the

second set of trials - Test 1 (the arc

welding is turned off again) - voltage

traces (top), USB frames (left bottom) and

HS frames (right bottom)

4.3 Third set of experimental trials – RSW

The objective of these trials was to verify the applicability of the system for a different

type of welding process. In the RSW process, two metal sheets are clamped together and

the spot welding is carried out. There is no presence of arc welding. Thus, besides the

synchronization between electrical signals (arc voltage and welding current) and USB

camera frames, these trials aimed at to grab the movement of the welding clamp and a

circumstantial presence of spatters.

The high-speed camera was not used for performing this set of trials, although, a

Labtec webcam was used for acquiring frames at 10fps, with a resolution of 640x480. Since

there is no interest on arc length at this time, the default image setting of the webcam was

kept and no magnify lens was attached to it.

145

Concerning the electrical signals, it was used a sampling rate of 1kHz for all trials.

The parameters set for the low-pass digital filter were cut-off frequency equal to 500Hz and

number of poles equal to one.

The webcam was about 500mm from the welding rig, in such a way that was possible

to enclose the welding clamp and the metal sheets.

Eight trials were performed for this trial. It was applied a pressure of 86kgf (2.3kN) for

all of them. Table 4.4 shows the duration in cycles of each welding function and the current

used for each one. Higher values of current were used for performing the two last trials in

order to provoke the presence of spatters.

Table 4.4: Setting welding parameters for the third set of trials

Trial I

rms

(A) Pre-Pressure

(cycles)

Pressure

(cycles)

Weld 1

(cycles)

Pulse

(cycles)

Weld 2

(cycles)

test1 3500 30 30 20 0 0

test2 3500 30 30 40 0 0

test3 3500 15 15 40 0 0

test4 3500 15 15 40 10 40

test5 3500 15 30 40 20 40

test6 3500 15 30 40 20 40

test7 6500 10 5 40 0 0

test8 7500 10 5 30 0 0

Results

A frame-by-frame analysis of the USB camera frames shows that is possible to

observe that these frames are quite synchronized with the current (in red) and voltage (in

blue) signals. In the beginning, before executing the pre-programmed functions, the welding

clamp was still opened (Figure 4.17). After that, the clamp gets closed (Figure 4.18). Then,

when the metal sheets are clamped, the weld cycles start and it is possible to notice the

presence of spatters (Figure 4.19). The clamp is kept closed during this time (Figure 4.20).

After finishing these cycles the clamp starts to open again (Figure 4.21) and returns to the

initial state (Figure 4.22). By running the software, these shots are more defined and that

sequence can be better realized.

146

Figure 4.17: Test 8 - shot that shows the

clamp still opened - voltage traces (top),

USB frames (left bottom) and disabled HS

frames (right bottom)

Figure 4.18: Test 8 - shot that shows the

clamp getting closed - voltage traces (top),

USB frames (left bottom) and disabled HS

frames (right bottom)

147

Figure 4.19: Test 8 - shot that shows the

metal sheets getting clamped and the

presence of spatter - voltage traces (top),

USB frames (left bottom) and disabled HS

frames (right bottom)

Figure 4.20: Test 8 - shot that shows the

metal sheets clamped during the weld

cycles and the presence of spatter -

voltage traces (top), USB frames (left

bottom) and disabled HS frames (right

bottom)

148

Figure 4.21: Test 8 - shot that shows the

clamp getting opened after the weld cycles

- voltage traces (top), USB frames (left

bottom) and disabled HS frames (right

bottom)

Figure 4.22: Test 8 - shot that shows the

clamp back to the initial position - voltage

traces (top), USB frames (left bottom) and

disabled HS frames (right bottom)

149

CHAPTER 5

DISCUSSION

The focus of this work was on the synchronism between electrical signals and

images, which is not a trivial task. These signals are acquired by different devices which

work by different principles. Thus, to find a solution able to integrate all of them it was not

that simple. It is believed to be a first step for future developments.

In this work, the synchronism consists on acquiring a buffer of electrical signals, an

USB camera frame and a buffer of high-speed camera frames, at regular intervals of time.

Thus, it was projected a system responsible for acquiring signals and USB camera frames,

besides triggering the high-speed camera, which acquires frames independently.

A visual analysis of the signals and the USB camera frames can be done while the

data acquisition takes place, that is, an online analysis. After it finishes, the high-speed

camera frames can be downloaded into the computer. Then, these frames, the acquired

signal samples and the acquired USB camera frames can be used for performing an offline

analysis that relates all these data.

An ideal behaviour of the system is illustrated by Figure 5.1. The acquisition of frames

and signal samples had to start at the same time (0ms), which means no delay among the

acquired data. Besides, each buffer of signal samples and USB camera frame had to be

immediately read after this frame became available. Webcam is the type of USB camera

used for performing this work. It usually provides a sampling rate of up to 30fps. It means

that at every 33.33ms (1/30), a different frame would be, in theory, available for reading, but

this rate is not very accurate. According to the image resolution of the frames, for instance,

640x480, this rate decreases to 15fps.

For each frame read and plotted, a buffer of signal samples must also be read and

plotted. Plotting usually takes a considerable time and the more frequent it has to be done,

the slower the system becomes. Thus, in order to avoid bad performance of the system and

considering a desirable image resolution of 640x480, a sampling rate of 10fps was adopted.

It means that at every 100ms (1/10), a different webcam frame is read, performing the real

behaviour of the system, which is also illustrated by Figure 5.1.

150

Figure 5.1: Illustration of the behaviour of the system along the time

According to the real behaviour of the system, after starting the signal acquisition, the

high-speed camera is triggered. However, there is a delay (dHS) between these two

moments. Moreover, when an USB camera frame becomes available, it cannot be

immediately read, and another delay is presented (dUSB), as mentioned before.

If dHS is deterministic, it is possible to shift the HS camera frames and provide

synchronism when doing the offline analysis. In this case, determinism means that dHS has

the same value every time the acquisition runs, under any circumstance. It is not very simple

to do, since a multitasking operating system does not offer determinism. However,

establishing an acceptable limit for the variation of the delay and triggering the high-speed

camera immediately after starting the signal acquisition, it does not make a serious matter.

In the other hand, the effect of the delay dUSB cannot be treated. Since the webcam

tries to equilibrate the quantity of brightness when it starts to acquire the first frames, the

sampling rate is very irregular during this time and these first frames are very bright and not

33.33 66.66 100 133.3 166.66 200 233.33 266.66

USB camera

HS camera

Signals

33.33 66.66 100 133.3 166.66 200 233.33 266.66

USB camera

HS camera

Signals

dHS

Time

(

)

Ideal behaviour

Real behaviour

dUSB

Time

(

)

acquired frames

151

good for analysis. In order to avoid this situation, the system starts the webcam, waits for a

while, starts the signal acquisition and triggers the high-speed camera. After that, at every

100ms, the last available webcam frame and a buffer of signal samples are read and plotted.

Thus, dUSB does not present the same value every time the signal acquisition starts. A

smaller irregularity is also presented along the acquisition of the subsequent frames. Thus,

the acquired webcam frame cannot be related to only one signal sample, but to the entire

buffer of signal samples. A solution to increase the accurate of the webcam’s sampling rate

would be to transfer the responsibility of grabbing each frame, from the webcam to the

computer.

Then, the way as the system is currently working, makes possible to do an offline

analysis of signals and images. It can be done by visualization of buffers of signal samples,

USB camera frames and HS camera frames, all of them related one-by-one. However, more

improvements can be done in order to increase the possibilities of applications for this work.

152

CHAPTER 6

FUTURE DEVELOPMENTS

The arc length is directly related to the arc voltage. Then, an arc length control is

usually done by correcting variations in the arc voltage value. For this purpose, an arc

voltage reference that provides the desirable arc length must be set. It is done by visualizing

the arc length. An algorithm could be developed in order to detect the arc length by

processing the USB camera frames. Having the correspondent arc voltage at each frame, it

would be possible to set its reference value automatically.

If a data transmission was implemented, the adjustment of these welding parameters

could also be done by distance. In this case, even if the arc length could not be detected by

an algorithm, the user could remotely see the arc welding and adjust its length by changing

the arc voltage, also by distance.

Another improvement of the system would be to increase the sampling rate of the

webcam frames. As explained in Chapter 5, increasing this rate, the system becomes slower,

since the data plotting becomes more frequent. However, a compromise could be

established. During the acquisition, the data could be plotted using the current rate of 10fps,

but it could be stored at a higher frequency, of 30fps, for example. Thus, it would be possible

to do an online analysis in order to monitor a weld and immediately act on the system by

distance (moving the camera, for example), as well as, more details could be seen during an

offline analysis.

Different types of USB cameras could be performed. An infra-red camera, for

example, could be used for detecting spatters during a weld. Besides that, more than one

USB camera could be integrated to the system, offering a better visual analysis of the

welding phenomena.

The sound emitted by the arc would be another interesting aspect for remote

monitoring. Listening to this sound, the felling of the professional working away is increased.

Besides that, the metal transfer mode could be identified by audio signal analysis. The

acquisition of this signal would be possible if an amplifier was correctly built for it and

connected to the acquisition board of the system. The audio board of the computer could not

be used for acquiring that signal. Since this device is controlled by the operating system, its

synchronization with the other devices (acquisition board, USB camera, HS camera) would

153

not be possible and would decrease the performance of the system. Moreover, the sampling

rates provided by this device could not be suitable for audio signal analysis, since they are

fixed at few possible values.

The sampling rate of the signals also affects the performance of the system. Higher

rates means more points being plotted at each interval of time and more computer memory

has to be allocated for this purpose. Then, even if the acquisition board manages to provide

a high sampling rate, if the computer does not manage to read the data and plot it on the

graph more quickly, the performance of the system decreases. Thus, it is necessary to use

powerful computers for higher sampling rates.

In order to provide a more consistent synchronism among the signals, a master clock

could be used for all devices. Since every acquisition board of National Instruments is very

accurate, its clock could be used for the high-speed camera and the webcam as well.

However, it would be necessary to use a clock divider for the webcam and find a way to take

its control. Thus, the three devices would acquire under the same clock source.

A last concern is about LabVIEW, the programming language used for performing the

system. In spite of being quite easy to program with this language, the documentation about

its functions is still poor. National Instruments has headquarters around the world to help the

developers, but most of time, it is not enough for not ordinary developments. These

circumstances make longer the development of the system. In order to save time when

developing LabVIEW applications, it is advisable to attend appropriate courses for training

offered by National Instruments. In some cases (basic), it can be done on internet for free. In

other cases it has to be paid and it is expensive.

154

CHAPTER 7

CONCLUSIONS

This work shows that it is possible to acquire electrical signals and images

synchronously in an integrated system. The introduction of low-cost cameras was feasible,

by providing images suitable for image analysis. Using the existent resources in the

laboratory, it was possible to build a prototype for the hardware with enough flexibility and

suitable for different types of welding processes. Eventually, a portable format for the system

was developed, which is more appropriate to the different types of work environment.

The system was tried in single wire GMA welding with reasonable success. However,

the system is still incomplete and limited. Concerning the incompleteness, the arc length still

needs to be determined by an algorithm to be developed for this purpose. When tried with

double wire GMA welding the results were not so good, showing some limitations. Due to arc

energy alternation, the acquisition rate of the USB camera is not high enough to focus on

only one wire.

When tried in resistance spot welding, the system was quite successful, too. It was

possible to visualize the movement of the welding clamp and the presence of spatters during

the weld. Data transmission still needs to be integrated to the system in order to provide a

remote monitoring of such events. Thus, the monitoring could be centralized and the

presence of a specialist in the plant would not be necessary.

However, a list of possible improvements were carried out, pointing out the design of

a more comprehensive system, which can be used for enlarge the field of applications. In

order to integrate this work to an automated welding system for pipeline repair, for example,

it is still necessary to develop the control of the welding parameters that would take the data

provided by this work.

Therefore, the system developed in this work is believed to be a first step for future

developments. From this point, it is possible to integrate other resources and implement

different applications in the welding area.

155

CHAPTER 8

REFERENCES

1. ALFARO, S. C. A.; VARGAS, J. H.; VILARINHO, L. O.; WOLFF, M. A. High Speed

Images of Resistance Spot Welding Synchronized with Electrical Signals. 18th

International Congress of Mechanical Engineering. Brazil, p. 8, 2005.

2. ARCON WELDING INC. Welding Process. Available on <www.key-to-steel.com

Access date: 15/01/2005.

3. BÁLSAMO, P. S. S.; VILARINHO, L. O.; VILELA, M.; SCOTTI, A. Development of an

Experimental Technique for Studying Metal Transfer in Welding: Synchronized

Shadowgraphy, In: INT. J. FOR THE JOINING OF MATERIALS, 2000, The European

Institute for Joining of Materials (JOM), Denmark, v. 12, n. 1, pp. 1-12 (ISSN 0905-

6866).

4. CARY, H.B. Arc Welding Automation. New York: Marcel Dekker Inc, 1995. p. 527.

5. DROUET, M.G; NADEAU, F. Acoustic Measurement of the Arc Voltage Applicable to

Arc Welding and Arc Furnaces. Journal of Physics E: Scientific Instruments.

Canada, v. 15, n. 3, p. 2, 1982.

6. FERRARESI, V. A.; FIGUEIREDO, K. M.; ONG, T. H. Metal Transfer in the

Aluminium Gas Metal Arc Welding. Journal of the Brazilian Society of Mechanical

Science & Engineering. Brazil, v. 25, n. 3, p. 6, 2003.

7. GILSINN, J., et al. A Welding Cell That Supports Remote Collaboration. In: NINTH

INTERNATIONAL CONFERENCE ON COMPUTER TECHNOLOGY IN WELDING,

1999, Detroit. p. 7.

8. LI, P.; ZHANG, Y. Robust Sensing of Arc Length. IEEE Transactions on

Instrumentation and Measurement. v. 50, n. 3, p. 8, 2001.

156

9. MAEDA, T.; ICHIYAMA, Y. Development of Adaptive Control of Arc Welding by

Image Processing. In: NINTH INTERNATIONAL CONFERENCE ON COMPUTER

TECHNOLOGY IN WELDING, 1999, Detroit. p. 11.

10. MANSOOR, A.M.; HUISSOON, J. P. Acoustic Identification of the GMAW Process.

In: NINTH INTERNATIONAL CONFERENCE ON COMPUTER TECHNOLOGY IN

WELDING, 1999, Detroit. p. 12.

11. MASON, A., et al. Real Time Image Processing of Electrode Profile in Resistance

Spot Welding. In: GERI ANNUAL RESEARCH SYMPOSIUM 2005, 2005, Liverpool.

12. NATIONAL INSTRUMENTS INC. Data Acquisition Fundamentals, in

APPLICATION NOTE 007. Available on <www.ni.com

>. Access date: 08/02/2005.

13. NATIONAL INSTRUMENTS INC. Intuitive Graphical Programming Language for

Engineers and Scientists. Available on <www.ni.com

>. Access date: 06/03/2005.

14. NATIONAL INSTRUMENTS INC. SCC Signal Conditioning Overview, in

PORTABLE MODULAR DAQ SYSTEMS. Available on <www.ni.com

>. Access date:

20/02/2005.

15. NATIONAL INSTRUMENTS INC. Signal Conditioning Fundamentals for

Computer-Based Measurement Systems, in APPLICATION NOTE 048. Available

on <www.ni.com

>. Access date: 20/02/2005.

16. THE LINCOLN ELECTRIC COMPANY. The Procedure Handbook of Arc Welding.

Cleveland: Lincoln Electric, 1994. ISBN 9994925822.

17. WIDGERY, D.; BLACKMAN, S. Pipelines for Stranded Gas Reserves: Cutting the

Cost, 2001. Cranfield University, UK, 2001. p. 11.

18. YAPP, D. Trends in Data Acquisition for Weld Monitoring. Cranfield University,

UK, 2004. p. 14.

Livros Grátis
( http://www.livrosgratis.com.br )
 
Milhares de Livros para Download:
 
Baixar livros de Administração
Baixar livros de Agronomia
Baixar livros de Arquitetura
Baixar livros de Artes
Baixar livros de Astronomia
Baixar livros de Biologia Geral
Baixar livros de Ciência da Computação
Baixar livros de Ciência da Informação
Baixar livros de Ciência Política
Baixar livros de Ciências da Saúde
Baixar livros de Comunicação
Baixar livros do Conselho Nacional de Educação - CNE
Baixar livros de Defesa civil
Baixar livros de Direito
Baixar livros de Direitos humanos
Baixar livros de Economia
Baixar livros de Economia Doméstica
Baixar livros de Educação
Baixar livros de Educação - Trânsito
Baixar livros de Educação Física
Baixar livros de Engenharia Aeroespacial
Baixar livros de Farmácia
Baixar livros de Filosofia
Baixar livros de Física
Baixar livros de Geociências
Baixar livros de Geografia
Baixar livros de História
Baixar livros de Línguas

Baixar livros de Literatura
Baixar livros de Literatura de Cordel
Baixar livros de Literatura Infantil
Baixar livros de Matemática
Baixar livros de Medicina
Baixar livros de Medicina Veterinária
Baixar livros de Meio Ambiente
Baixar livros de Meteorologia
Baixar Monografias e TCC
Baixar livros Multidisciplinar
Baixar livros de Música
Baixar livros de Psicologia
Baixar livros de Química
Baixar livros de Saúde Coletiva
Baixar livros de Serviço Social
Baixar livros de Sociologia
Baixar livros de Teologia
Baixar livros de Trabalho
Baixar livros de Turismo
 
 