SIMoNET Seminar Report
Report of the SIMoNET meeting 'Structural Integrity Monitoring'
held at UCL on 20 November 2002
Dr J.V.Sharp and Dr R.Karé
- Introduction
Professor Bill Dover introduced the programme for the day and noted that the next meeting was planned for the 2nd April with a theme of advanced composites, seismic structural monitoring and marine technology, Speakers were invited for this meeting. - 'Condition Assessment of Ageing Pipelines', Malcolm Wayman , Advanced Engineering Solutions Ltd.
Malcolm outlined the work undertaken by his company, on condition management of ageing pipelines ranging in diameter from 3" to 72". Assessment of the smaller iron pipes involved the checking of critical sections, using external inspection. Samples were often cut out, shot blasted and inspected for defects. Pipeline coatings could be a problem, particularly for the older lines. Corrosion modelling was used to predict life.
For the larger diameter pipelines inspection was for both internal and external corrosion. Stress analysis using FE methods was used to assess critical sections, factors of safety and remaining life. A critical parameter was the remaining wall thickness. The company used a 'bath tub' curve approach, working in the design zone and attempting to predict the end of life section of the curve. - 'Non-contacting stress measurement on I beams', Russell Hall, TSC Inspection Systems
Russell presented the recent developments in the application of ACSM to measure stresses on an I beam and to determine the principal stress direction. The new probe was manufactured to look in two directions at once and have the capability to measure the principal stress direction. Russell described the measurement procedure used to obtain a calibration line for the I beam which was then used to predict the stress in the beam.
His results were obtained using a simple 3 point bend rig for calibration, and an Instron Servo-hydraulic machine to simulate the service component loading on an I beam. The graphs showed good agreement between the applied stress measured using a strain gauge and the predicted stress using the Stress Probe. Stress Probe can measure cyclic or dynamic stresses and can also be used to map the stress on a structure. Russell's presentation generated a number of questions asking if acsm is capable of measuring dead loads. Russel with support from Bill Dover explained that it was possible to measure the dead load.. It was also established that ACSM could measure hoop stresses. - 'Structural Integrity of Steel Framed Buildings under seismic loading', Adam Bannister , Swinden Technology Centre, Corus Ltd.
As a result of earthquakes at North Ridge and Kobe it had been found that the performance of many steel framed buildings was not as anticipated. High through thickness stresses in the columns had led to fractures. This has led to a new research area, involving design, material toughness and weld quality for the flanges which could be up to 125mm thick. Corus supplies heavy steel sections to the US market.
AB noted that current US codes had limited information on requirements.
First order reliability methods had been used, with definitions of failure probability derived as a function of the several inputs. Data from Charpy tests on large scale sections had been correlated to average toughness levels. The software used had been developed within the SINTAP programme, using distributions for yield stress, fracture toughness, flaw sizes etc. A target probability of failure for the joint had been assessed as 2 x 10-5 , associated with a maximum credible earthquake. FE modelling had been used to develop though thickness stresses due to joint rotation. This enabled the level of fracture toughness required to be defined, related to depth in the flange. Beam-column stress interactions were investigated, checking where plastic hinges (e.g. in welded joints )could form. In addition the IIW risk assessment procedure had been used, enabling a target probability of failure to be determined, taking account of the sensitivity of different factors - 'Stresses in rail axles', Prof. R.Smith , Mech. Engineering Department, Imperial College, London
Professor Smith gave an interesting background introduction to rail accidents, with several examples, including those from 1842 and 1884. These involved fatigue failure which were a catalyst for early research on fatigue, particularly on axles. In the early days of the railways manufacturing flaws in the axles dominated, which had more or less now been eliminated. Safety is a critical aspect for axles, with current failure rates in the region of 1.6 p.a. The cost of inspection is high and its effectiveness is doubtful. The rational use of FE methods is limited by lack of knowledge of service loads. In a 30-40 year lifetime the number of stress cycles could be up to 109. and hence fatigue limits in the range of 106 to 107 were not adequate. Key parameters were local stresses, material properties, size effects and surface condition.
The axle bending stress was the primary load, and with modern rolling stock travelling up to 400,000 Km/yr the number of cycles of this load was in the range of 2 x 108 cycles p.a.. Other loads came from travelling over crossings, from local flat spots on the wheels. The complexity of the modern bogie made inspection difficult for axles. Life management of axles involves visual, UT and MPI inspections. UT is usually used from the end of the axle. However in general only 1 in 104 axles have cracks during inspection, which places a heavy responsibility on the human operator to locate these. Unfortunately there is no database of cracks detected in axles, and hence an estimate of the probability of detecting a crack is difficult.
Service load measurements are being undertaken, particularly abroad, using a system of strain gauges and accelerometers to measure loads on the axles. It is linked to a GPS system and hence it is possible to locate problem areas with the track. Unfortunately the costs of preparing safety cases and insurance have been making it difficult to implement in the UK. In conclusion he noted that the axle and wheel combination are a challenge to fracture mechanics. The costs and unreliability of NDE methods are key problem areas. The new method of long term monitoring of loads is a step forward. - 'Crack monitoring in drill string threaded connections', Dr. Mathew Knight, UCL NDE Centre
Matthew Knight presented the results on the first attempt at using the ACFM technique to monitor crack growth in drill-string threaded connections. ACFM has successfully demonstrated the systems capabilities in recording crack growth data for defects as small as 0.6 mm depth and 12 mm length. There was good correlation in results with the optically measured beach marks for the unrolled and cold rolled new specimens.
However, results have indicated that the ACFM defect sizing capability may be compromised when the defect is in the vicinity of a large residual stress field, such as that found when the connections are cold rolled. If a connection is cold rolled without first being defect free, subsequent inspections may underestimate crack depth with the standard interpretation. Inspection before rolling is therefore necessary.
Matthew had shown that even though ACFM had not been previously applied to a stitch rolled component, it did however successfully detect cracks in the stitch region. During questions Matthew was asked if he had tried acoustic emission on drillstrings. MK replied no but he did not think it would be successful because of problems with the geometry effect. - 'Eddy current modelling', Steve Jenkins
Steve Jenkins explained that he was responding to comments apparently made in an earlier Simonet meeting that eddy currents couldn't be modelled. Of course, as he stated, there are several methods which can be used to model eddy currents, with Green's functions being his favourite. Using Green's functions, Steve showed model predictions of eddy current phase plots from slots which got closer to real data the more cells that were used to model the slot. He suggested that, by restricting the problem a little, model results such as those could be used to extract more information on defects during inspection (such as depth sizing without calibration). Following Galilieo's example, Steve put himself up as the academic in a 3-way discussion, involving himself, an NDT customer and a technician. The academic understood the model, but had little experience of real inspection; the technician didn't understand the model but could see how the results could benefit his work. The problem, as Steve sees it, is for these two to persuade the customer to pay extra for model results to be included in his inspection. - 'Tunnel and track Instruments on the London Underground', Gerwyn Price, CMCS
Several projects were described associated with deterioration of the fabric of the infrastructure. Key factors were the changing capacity and safety standards of the Underground system. The first project involved improvements to safety of the Angel station on the Jubilee line, which involved building new 6m diameter tunnels. This could affect the ground stresses and to monitor these a system was set up to measure clearances for the trains. To manage the risk of undertaking the building work whilst operating trains the feedback was an essential requirement.
The second project was 'Effra' where a new tunnel was planned to pass close to existing tunnels. Pressure, water level and vibrating wire gauge monitors allowed the measurement of changes to the existing tunnels, which in practice showed only 5mm movement. - 'Asset management of rail embankments - inspection strategy', John Perry, Mott Macdonald Ltd
Many railway embankments are up to 150 years old, constructed manually with little compaction and with steep slopes at the side. These now need to be maintained to a safe standard. Several examples were shown of failure, including those as a result of pore water pressure from heavy rain. Serviceability failures were often associated with the need for high maintenance e.g. result of local subsidence. The scale of the problem was illustrated by Network Rail spending between £10-30M p.a. over a 10 year period, with London Underground having spent £70M on earth structures between 1994-99.
The objectives of the asset management were described, both at the strategic and tactical level. The purposes of inspection were to assess structural stability, detect incipient defects, classify the condition of the embankment, identify a need for any preventive action and collect data for developing future maintenance strategies. There were several stages to the assessment which included site survey, stability study and reporting and prioritisation. The results allow the development of a business plan for the local network, providing assurance to the owners, operators and the public. An example was shown of the embankment on the M23 Gatwick link, where slips had been identified.
Some automatic monitoring is in place, usually installed after a manual inspection. This is to measure ground movements and ground water. In conclusion it was noted that many existing railway embankments are old, exceeding their original design lives. A regular appraisal of their current condition provides good value for money. - New miniature stand-alone triaxial accelerometer alarm system, Ian Ramage, Techni Measure
Ian Ramage described a novel monitoring system (SeTAC) developed by an Italian University, now available in the UK from his company. It incorporates both a sensor to measure x,y and z accelerations using a low cost silicon sensor, and also an alarm system. Its bandwidth is DC to 4kHz; it also has an in-built temperature sensor. Full scale acceleration is 5g, with 2mg resolution and able to survive a 1000g shock wave. It is lightweight (55 g) in a aluminium box, priced at £650 each. The power supply required is 5V, at 50 ma. Four digital logic outputs are available. These can be programmed to set alarms. The device has some data storage capacity. Applications described included tool wear detection, tilt switch control for cranes, monitoring mechanical vibration, vibration control and feedback, monitoring repetitive machine cycles. - Engineering Integrity Society (EIS) Seminars for Practising Engineers', Richard Cawte, EIS Fatige Group
Richard Cawte explained that the EIS was now 25 years old. It had 3 working groups, one of which was durability and fatigue. It ran conferences, seminar and workshops, as well as an annual measurement exhibition. A journal was produced, with two editions each year. Examples were given of the various conferences, including a major event in 2000 in Cambridge. A further conference was planned for 2003, covering the fields of aerospace, ground vehicles, ceramics and coatings, life extension etc. Technical groups associated with the main group were on software benchmarking and smart structures. - 'Cost & Effect - remote sensing of underground rail track deterioration', Kevin Hamson, Frazer-Nash Consultancy Ltd.
Noise and vibration due to underground railways can be a problem to properties close to the underground, often associated with deterioration in rail performance over time. An example was described where people living in a property close to the underground system were affected by noise and vibration, some of which was considered extreme. The objective of the study was determine the baseline noise and vibration levels and determine the causes of the extreme events experienced and then propose a solution.
One of the usual causes of noise and vibration is rail corrugation, which can be reduced by replacing the rail. New rolling stock can accelerate wear of the track and can cause excessive noise and vibration.
Monitors were installed to measure the noise and vibration, correlated with the passage of trains and their speed. Equipment included accelerometers, noise and vibration sensors and proximity sensors to identify the trains. The spectral content of the vibration data was also examined. High vibration levels were found in one particular position, which were reduced by a factor of 6 by re-railing that section.
In summary it was noted that prior to re-railing noise levels were audible, but barely perceptible afterwards and acceptable to the people involved . No indication was found of the 'extreme events reported at the beginning. There is now also a good set of baseline data. - Closure
Professor Dover closed the meeting, thanking all those who had presented papers.
10.12.02