Bob Piggott, MIMechE, CEng, is a Senior Research Technician in the Fatigue Department.
There has been no other strain transducer that has been more successful or invaluable to those involved in stress analysis and data collection than the ERS gauge. The bonded electrical resistance strain gauge came into use in 1938 and this year celebrates 50 years of service to engineers in research and industry, not least to those actively engaged in welded constructions. It was at about the same time that welded fabrications were also beginning to emerge and take over the role of castings and riveting in many structures. But to exploit fully this joining technique a greater understanding of the performance of welded joints under service loading was needed. Because the vast majority of welded structures are subjected to variable amplitude loading one of the most important design considerations is that of fatigue. It is therefore essential to be aware of the stresses due to service loading and this has been made possible with the use of the ERS gauge.
While the concept of the ERS gauge has not changed over the years its performance has benefited by the use of new material and production techniques ( Fig.1). In contrast we have seen vast strides in the advancement of instrumentation used to monitor and collect strain data. In a short time the reduction of strain data for fatigue analysis has progressed from the manual analysis of traces from a chart to the compact multichannel unit commonly called the 'on board real time analyser'. This can carry out tasks of collection and processing, and can be interrogated at any time to obtain the current status of measured stress ranges and cumulative fatigue damage.
Although the microprocessor has made the 'on board' data analyser possible it is not the only technique for gathering and processing service loading data. To illustrate some of these it would be useful to make reference to an exercise carried out by The Welding Institute some years ago and discuss the semi-manual techniques used then and the options we could use today. There are advantages and disadvantages in all the methods.
The task
As the demand for electrical power in this country increased the power station designer became well aware of the increased efficiencies he could obtain from his plant by increasing the physical size of individual components. These components would weigh between 300 and 500 tonnes but before designs of this weight could be manufactured designers had to ensure that movement of such large loads was possible and more importantly that hauliers had trailers capable of carrying them.
At that time the maximum load capable of being carried was around 300t using sophisticated trailers which met the legal requirements of gross vehicle weight and overall dimensions. The designer achieved this by manufacturing the trailer frames from a weldable high yield steel to accommodate the high service stresses. However, if the payload was to be increased still further it was essential that the trailer weight itself should be kept low so that the allowable gross weight was not exceeded. Under the circumstances this could only be achieved by the use of even higher strength steels. However it is known that the fatigue behaviour of steels, in the as-welded condition, is independent of the strength of the steel. It was therefore essential to show that fatigue strength was not a limiting design criterion for these trailers.
To examine this question strain gauges were bonded to a trailer structure ( Fig.2) at strategic locations and the strains were recorded throughout a 96km journey from Stafford to Pomona Dock, Manchester. At a road speed of some 8km/hr this journey took three working days.
Recording the strains
At the time this exercise was undertaken, multichannel direct writing mirror galvanometers and oscillographs were being widely used to record continuous strain data. If the frequency of the signal to be captured was sufficiently low (not greater than 25Hz), sensitive galvanometers without signal amplification could be used. Hence a compact uncomplicated recording system, producing an instant trace of the strain behaviour, was achieved. Such a system was used to record signals from the trailer structure and was mounted inside the cab of the leading tractor unit, taking power from the vehicle supply. This could be described as an early 'on board' data collection unit. The chart speed was extremely low so that continuous traces recording the strains could be gathered throughout the journey. To correlate these data with route conditions timing marks at five minute intervals were automatically written on the chart.
Analysing the traces
The first task in analysing the recorded data was to identify manually the maximum and minimum stresses which occurred during typical manoeuvres and hazards during the journey. This was followed by extracting all the turning points (peaks and troughs), from each trace, recorded throughout the journey. This was carried out by a semi-automatic process on a digitising table, measuring the peaks and troughs in sequence to produce a paper tape of the data points which could be used in the subsequent computer analysis.
Predicting the fatigue life of the trailer
In order to get a realistic prediction of the fatigue life of the structure it was necessary to use a computer to analyse the random stress distribution which had been obtained from the traces as described above. A computer program based on the cumulative damage clauses given in BS 153, 'Steel Girder Bridges' which incorporates the widely accepted Miner's cumulative damage rule, was used for the analysis.
The conclusion of this exercise was to show that the fluctuating stress ranges were small compared to the high mean stresses induced by the dead weight of the payload and that fatigue would not be a limiting factor in future designs.
What were the advantages and disadvantages of the system used?
Having a galvanometer oscillograph for signal frequency capture has two useful advantages. First no additional signal amplification is needed and second, the traces can be immediately observed. It does however suffer from a limitation in dynamic range. This is the ability to provide the same resolution of small as well as large trace deflections. The same errors in trace measurement exist throughout the entire galvanometer swing as a result of electrical noise and image sharpness. In addition to this there is an increasing error as the galvanometer is deflected from its central ahead position. These errors would limit the resolution to about one part in 100, often considered acceptable for this type of measurement, but it will be seen to be poor when compared to today's digital systems.
Probably the greatest drawback that exists is that the analogue traces have to be digitised if subsequent analysis is to be carried out by computer. With long recordings, involving a large number of fluctuating stress cycles, a lengthy exercise is required to convert these data to a computer compatible form.
A reduction in the manual handling of the raw strain data can be achieved by replacing the galvanometer oscillograph with a multitrack magnetic instrumentation tape recorder. When operated in FM mode, analogue signals can be recorded and stored on magnetic tape to be replayed as required. It is necessary however to raise the output from strain gauge bridges, by high gain instrumentation amplifiers, to the level required at the tape record inputs. A voice track is also available enabling additional verbal or coded information to be recorded to help in correlating the data.
Once the raw strain data have been stored on tape they can be replayed whenever required and the output is in the form of an analogue voltage. In this form it is not suitable for data reduction by computer and must therefore be further processed. The method used at The Welding Institute to transfer the analogue tape signals to computer is to use a high speed scanning device. The outputs from the tape recorder channels are fed simultaneously into the scanner which converts the analogue traces into digital format and writes them to tape which is computer compatible.
The tape containing the digital information is then mounted on the tape deck of the Prime mainframe computer to transfer and file the data for further reduction by the appropriate analysis packages. Packages exist on the Prime for the extraction of turning points from the fluctuating strain histories, 'rainflow' analysis and fatigue life assessment based on the fatigue clauses of BS 5400.
Has this system any disadvantages?
It could be considered a disadvantage that amplification is required in order to produce high level signals for the tape recorder. The high gain amplifiers available today have good stability and ability to reject noise, but extra care must be taken with external strain gauge cables.
The FM tape recorder suffers to a somewhat lesser degree than chart recorders, to provide good resolution of greatly varying signals. Spikes and 'drop outs' can occur on tape and these must be filtered out on replay prior to analysis. In multichannel application tape recorders can be bulky because a separate recording channel and associated amplification are required for each data channel.
Although it is now advantageous to be able to output the stored data as electrical signals they are still in analogue form. They must therefore be converted to digital form for computer analysis. The conversion, however, can be undertaken without manual involvement and by increasing playback speed the processing time can be greatly reduced.
The digital approach
Two undesirable points emerge quite clearly from the methods of data collection and reduction described so far. First that a recorder channel must be dedicated to every signal being measured, and second that the recorded analogue signal must be converted into digital format for subsequent computer analysis. These have now been overcome by the use of Pulsed Coded Modulation (PCM) which uses digital techniques.
Once accepted that the facilities offered by PCM systems were ideally suited for the collection of service loading data for use in predicting fatigue life, a system designed and manufactured by a Germany company Volland & Kraus was acquired by The Welding Institute.
The benefits are that this unit allows analogue signals to be recorded on a standard instrumentation tape recorder in digital format, and this has a number of important advantages. The dynamic range is significantly improved (resolution 4096:1) over the other systems discussed. Since information is in digital form it is less susceptible to corruption and it is readily transferred to a digital computer for analysis. The PCM multiplexing technique allows a number of separate channels of information to be stored on a single track of a tape recorder or transmitted via fibre optic or single cable links directly to a computer. At present the Institute system can handle 16 channels of data but can be expanded up to 256 channels.
The system is programmable, and allows different sampling rates to be selected on each channel to accommodate a wide range of signal bandwidths. It also includes automatic anti-alias filters to remove signal frequencies above those which can be accurately captured at the selected sampling rate. An analogue unit allows four channels to be reconstituted from the digital data as an aid to checking signal integrity or for plotting on an x-t recorder.
Because of the varied input requirements of strain gauges and transducers used in the field today a universal input adaptor has been constructed for use with this equipment. Arrangements have been made to provide strain gauge bridge completions for ¼, ½ or full bridges, 120 and 350 ohms, and transducer sockets to enable any mix of inputs to be made.
The system is illustrated in Fig.3 with the RACAL Store 7 instrumentation tape recorder which is used for data storage. In this illustration the PCM system is shown as one unit, but in this form is prevented from being used in the field at the same time as it is needed to transfer data to the mainframe computer. The system has now been split into the two main components, record and relay, so that tapes returned from the field can be mounted on a second RACAL Store 7 and transferred to the mainframe computer via the replay unit.
Once transferred to the mainframe analyses such as turning point extraction, cycle counting, fatigue damage calculations and frequency analysis may be performed.
The advantages far outweigh the disadvantages in this system and only a brief recap is necessary. Probably the major disadvantage is cost, and this is not insignificant, but this is offset by the benefits of digital recording which provide computer compatibility and selectable sampling rates on each channel. Also real time events and markers can be inserted into data which can be used during computer analysis. Probably the greatest benefit is the fact that data can be stored on the single tape track obviating the need for bulky recording machines.
The 'Black Box' era
With the advent of the microprocessor the on board or real time data collector and analyser has become a reality. Probably the first commercially available unit of this type was the 'Datamyte' which arose from the demands of the American automobile and earth moving industries for an unattended data collection unit which could be mounted on board a vehicle. When interrogated it would provide stress history and fatigue damage information ready processed from raw data.
Great strides have been made in this field, and these advances can largely be attributed to the electronic packages available to designers. These units carry out a cycle counting analysis of a strain history in real time. Thus, a major advantage is that long term monitoring is possible without the need for extensive data storage. A microprocessor continuously analyses the incoming signal and accumulates a cycle count, usually based on the 'rainflow' algorithm. With some systems the unit is interrogated using a microcomputer ( Fig.4) to transfer and print the data and to perform a simple fatigue damage calculation based on the methods prescribed in BS 5400: Part 10. Other systems differ in as much as they have the analysis programs built in and the data are accessed via a terminal.
Whilst these systems are commonly referred to as on board data collection systems they can be used in many other applications where stress data are required such as offshore structures, bridges, tall masts, plant and machinery. It does not require a great deal of imagination to visualise how well suited this system would have been to the trailer exercise. In this and many similar exercises which are undertaken the saving on attendance on site and hours spent in analysis would be considerable.
There are also occasions, over very rough terrain, when both operator and equipment fail to perform at their best and could be replaced by one of these units.
These units obviously have many advantages over the other systems mentioned and as they come into wider use further advantages will be discovered. Since only the cycle count is captured, for long sequences of recording one has to forfeit the information relating to the order of the cycles, their form and frequency. It is possible to capture data points in sequence but this is at present restricted by the memory available to store the information. This aspect of these units will prove to be a major disadvantage if future research enables the accuracy of fatigue life predictions to be increased by taking into account sequence effects, such as the delay in crack growth rate which can occur following a high tensile stress (i.e. crack growth retardation).
This has been a simple overview of the type of equipment which has been used by the Institute from 30 years ago to the present day and it should be added that all these systems have a role to play in our daily data retrieval and analysis exercises.
The case history chosen to illustrate the necessity for field service measurement and analysis is of course one of a number that could have been used. As well as moving structures associated with road vehicles, earth moving equipment, railway rolling stock and cranes there are fixed structures such as road and rail bridges, wind excited structures including television masts, offshore platforms, plant and machinery and unusual stuctures like floating roofs of large storage tanks.
