Crystal-balling valve problems
Published: 08 July, 2011
When it comes to maintaining valve populations, plant engineers often wish they could see into the future. Since valves touch on all major areas within a plant, when they don’t perform well it can have a negative impact on tier-one assets. A small glitch can quickly escalate into something more serious that compromises plant efficiency and functionality, or even results in an emergency (and costly) shutdown scenario. Euros Jones, primary services manager at Severn Unival reports.
While the industry is yet to develop a crystal ball for valve problems, a sound approach to performance data acquisition and interpretation can underpin the next best thing: effective predictive maintenance.
So what does best practice look like? At its core, it should involve a strategic and intelligent approach to valve testing and measurement. This can help engineers identify emergent problems at an early stage. In turn, this enables shrewd prioritisation of work-lists to ensure valves operate more efficiently, have a prolonged lifespan and contribute to better overall plant performance.
Testing and measurement for valves going into critical service applications should ideally begin in the factory. Establishing a valve’s performance ‘footprint’ provides a useful benchmark for future data to be plotted against throughout its life. As performance is monitored over time, a footprint helps to verify whether set criteria are being met, and enables more accurate measurement of any deterioration. This is especially important for valves destined for challenging environments. If more manufacturers and plant engineers insisted on footprinting as a standard shop floor procedure, it would ensure much tighter lifetime performance control. An extra hour or two in the factory could potentially save weeks of future downtime.
Today, it is widely recognised that control valves in particular should be fitted with a ‘smart’ positioner to monitor and store diagnostic performance data from the initial testing stage onwards. Smart positioners pave the way for optimum testing and measurement throughout a valve’s working life. They play a pivotal role in ensuring maintenance or re-specification requirements are flagged at the earliest possible stage. For older valves, there are bolt-on smart solutions available. These can be attached during service to capture advanced diagnostic information, despite the lack of historic information specific to that valve.
Once a valve is operational, the big question is how often it requires attention. Most manufacturers suggest a standard maintenance routine, and it’s true that for many valves a regular annual or bi-annual assessment can be adequate. At a plant level, this can be enhanced by simply identifying the most critical valves - or the top ten ‘bad actors’ renowned for causing problems – to give some direction to the testing and measurement strategy.
However, while this provides a good foundation, a solid, timetabled approach to testing and measurement across the whole population doesn’t facilitate true predictive maintenance. A more dynamic, intelligence-led approach can reap dividends and ultimately prove far more cost-effective than static methods.
With smart positioners delivering real-time information on all key assets, it becomes easier to evaluate and predict the need for more detailed testing and measurement procedures. But use of this equipment needs to go hand-in-hand with the knowledge and experience of engineers. Predictive maintenance work-lists should be directed by both human engineering expertise and digital data.
For instance, an engineer-led approach may involve ‘walking the course’ looking out for typical problem indicators such as tell tale whistling noises or juddering. If a valve isn’t performing at its best, inspecting it during operation can lead to more efficient, well-honed testing and measurement, both across the entire population and on individual valves.
Marrying this with live performance data overlaid on standard routines enables engineers to make risk-based decisions about which valves to focus their attention on. Strong and consistent use of condition monitoring and condition-based decisions can lead to a significant shift towards a continually reducing volume of operational problems.
Increasingly, it is becoming common for digital analysis to be used to corroborate engineers’ diagnosis of emergent valve problems or underlying causes of poor performance. At Severn Unival, this is taken one step further with the Repair Intelligence Circle. A range of asset management and diagnostic tools feed into this dynamic data source which draws on historical records of performance, as well as valve repair and upgrade activity, from numerous plants. Analysing this wealth of data allows trends to be established, enabling precise evaluation of testing and measurement requirements. Ultimately this means better prioritisation of work-lists with potential problems predicted and dealt with in good time, before they impact on performance.
Providing hard evidence to back up expert opinion can also speed up the decision making process when a period of unplanned shutdown is necessary to remedy a problem. With shutdowns potentially costing hundreds of thousands of pounds per day, accuracy of diagnosis is paramount.
The best engineers have always had an innate talent for predicting which valves are likely to need attention. Digital technologies have elevated this skill, bringing a more scientific approach to testing and measurement planning.
We’d all love a crystal ball, but engineering skill backed up by innovative tools comes a close second. This heightened engineering intelligence creates a virtuous circle where an ever-decreasing rate of poor performance leads to optimum productivity and profitability.
Valve predictive maintenance in a nutshell
Measuring values such as hysteresis, dead band and step response against the ideal can provide early indication of underlying problems like underpowered and/or inaccurate control. Conducting this electronically enables the rapid formation of data for individual valves and across valve populations. This facilitates trend analysis, which plays a vital role in predicting problems and pinpointing bad actors before there is a significant impact on efficiency and productivity. This data, combined with real-time access to photographs of valves via web platforms, enables valve performance engineers to work more proactively.
Problem indicators that engineers measure and observe include:
- Valves’ ability to control required parameters
- Changes in supply and actuator air pressures
- Juddering and oscillation
- Frictional increase or decrease and ‘stiction’
- How quickly a valve reaches its assigned position
- How well a valve maintains its assigned position
- Spikes in travel
For further information please visit: http://www.severnglocon.com/