Mention the phrase ‘asset data collection’, or even the more modern activity of ‘remote condition monitoring’, and you tend to see a blank look and eyes glaze over. Somehow the whole business of monitoring and recording assets is a ‘turn off’ when compared to other elements of railway engineering.
It’s so much more exciting to be designing things at the forefront of technology, or project engineering the installation and commissioning of a new system. Being given the job of asset recording is seen as a mundane task where keeping records up to date is a ‘no win’ exercise and the multiplicity of information gathered is rarely considered by senior management and tends to be hidden away in some form of electronic record, only to be used if an incident occurs so that historic evidence can be trundled out for an enquiry.
Yet knowing what assets your company owns and the condition of these is a vitally important task that, if carried out properly, can make huge differences to the economics of the company and the performance of all that it does. So can the profile of this activity be raised and just what kind of deliverables can be expected?
A recent visit to Telent, the communications system integrator with its headquarters in Warwick, yielded some interesting opportunities.
Starting the initiative
With its many contracts for telecom projects and network provisioning, Telent became aware of condition monitoring applications in various transport systems involving ‘things that move’. Could it enter this arena and build-in better communications links for the movement of asset data? It could then improve the way the data is analysed and used to provide really meaningful information presented in such a way that senior management would take notice.
Telent itself has maintenance contracts that look after 250,000 assets across a multiplicity of industries, so just keeping track of all that was sufficient to stir management motivation.
Described as “driving intelligence into old unintelligent assets”, maybe the life of equipment could be usefully extended without suffering embarrassing failure rates at the same time.
Preliminary work began in 2010, but the real advances came in 2012 with the London Olympics. London Underground was acutely aware that the transport service to and from the Games and around London generally was a key part of making the Olympics a success. A particular concern would be the performance of lifts and escalators and especially anything involving the movement of physically impaired people.
For years, London Underground has been developing and using asset condition systems, but getting the right information from the masses of data obtained and interpreting this in a way that could be easily understood was something of a challenge.
Telent devised a solution that pulled together data from lifts and escalators and delivered it in an actionable way to the maintenance staff. As it turned out, the Olympic transport arrangements worked well, but the seed had been sown as to whether a more comprehensive package for asset monitoring could be devised that would be applicable to a multiplicity of assets owned and operated by many industries.
In 2013, Innovate UK (formerly the Technology Strategy Board) was promoting innovation in the railway and the idea of digitising railway systems. Telent responded to this and offered to put together a consortium that would combine existing expertise and some new thinking. The company approached London Underground to see if a predictive and preventative maintenance package could be devised for its many assets, all based around digital technology.
Those involved were Telent as leader and system integrator, Humaware (a company with contracts for health and usage monitoring in the aerospace industry) and both Loughborough and Nottingham universities utilising their PhD students and faculties. The overall objective was to use existing remote condition monitoring (RCM) data to manage asset degradation and give predictions on future performance and asset life. The project was co-funded by Innovate UK, RSSB and the consortium in a two-year timeframe for a budget of £750,000. Work began in January 2014 and was completed early in 2016.
Developing the idea
With a set timescale and limited budget, clear boundaries needed to be set as to what was realistically achievable. With its Olympic experience, London Underground proposed using escalator performance and maintenance as the test bed. It would have been impractical to incorporate the complete suite of escalators (over 400), so a chosen few from a handful of central London stations were selected.
London Underground provided a complete set of historical data from these escalators, many of which had experienced maintenance events.
`An early task was to drill down into this data to see if the events were identifiable; many reported failures are investigated by sending people to site, only to find that no actual failure occurred – itself an expensive exercise.
The measurement criteria would concentrate on vibration within the main components of motor, gearboxes and running gear. The existing data was analysed in a completely different way with the intent of using the information to deliver predictions on when failures are likely to occur and to predict the remaining useful life of the particular component.
Telent, Humaware and the London Underground asset engineers worked jointly to develop the predictive analytics software using an open architecture. One immediate analysis revealed that a problematic escalator fault could have been detected five months before a major failure actually happened. It became obvious that a major benefit would be a re-scheduling and management of the task to maintain assets that are in use for 20 hours a day.
Asset information by its very nature can amass a huge amount of data, and portraying these results in a format that can be easily interpreted by engineers and managers is itself a challenge.
Enter Nottingham University, which has expertise in this field. Many people are aware of using traffic light colours to give a basic indication on whether a piece of equipment or system is in good health or otherwise. Nottingham has taken this process further and developed a series of line sequence images that show the major components of a system each as a coloured dot. If the dots are all in line, it indicates that the complete system is fine. If any dot is out of line, with either an amber (warning) or red (imminent failure likely) colour, it becomes obvious that a potential problem exists and a deeper analysis is required.
Drilling down into the data will reveal the likelihood of that component failing within a time period and, by implication, the remaining useful life of that particular asset.
If several dots in the chain are out of line, the subsequent analysis will help decide which component needs to be changed first, an important factor for such as a London Underground escalator when only a limited time for maintenance is available each day. It must be stressed that this is a predictive tool that relies on real data from an existing asset monitoring system rather than a modelling exercise derived from statistical data.
It will give a percentage chance of the risk of a failure occurring but nothing can be absolutely certain. However, of the 20 escalators trialled, two alerts were found at King’s Cross and Piccadilly immediately, which were further monitored for failures before they actually occurred. It is, of course, up to asset managers as to whether or when action is taken to act on the predictions given.
Escalator replacement is a costly and time-consuming task, typically taking several months with considerable inconvenience to travellers in this period. If more effective monitoring can extend the machine’s life before a complete renewal becomes necessary, then it will be a valuable tool
Further monitoring work
Whilst the first application of the technique was on escalators, the process should be capable of deployment on any piece of moving equipment. A subsequent exercise, co- funded by RSSB and the consortium, is underway to examine the performance of bogies on the Southern Railway fleet of Class 377/8 EMUs. This will select two bogies – one having just been maintained, the other coming close to its routine maintenance cycle – to see how the two compare in day-to-day performance.
The fitting of sensors will be done at Selhurst depot and the same consortium has been engaged for the trial with Birmingham University providing the sensing instrumentation. Again the results could well lead to a different maintenance regime and maintenance cost savings for the train operator.
Telent has given this initiative a brand name, ACUMEN – Asset Condition Usage, Maintenance and Engineering – and has high hopes that this will have a big future. The ultimate test is whether this can tie in with other initiatives under the banner of customer performance improvement.
Thanks to Alastair Norman, project manager, Keith Wilson, business development manager and Peter Felton, account director for rail, for their help in preparing this article.