What are the pitfalls that cause predictive maintenance projects to fail and how do you avoid them? How do you ensure teams acquire the condition monitoring expertise and appropriate ‘best practice’ for predictive maintenance as part of wider digital transformation projects?
In this film, Senseye's Alexander Hill and Rob Russell are joined by Peter Gagg from MCP Consulting Group, to discuss what needs to be in place from the outset to give predictive maintenance initiatives the best possible chances of success, and how manufacturers and other industrial organisations can build on these initial stepping stones to continue driving sustained improvements.
Transcript
Alexander Hill, Senseye: The main causes of PdM projects failing fall into two main categories: technology and culture.
On the technology side, it may be that data is just not available - the project might start full of enthusiasm but you find the data you can get from these machines is just not suitable for condition monitoring purposes. There are things that can be done about that, sensors can be retrofitted, further data analysis work can be done in order to bring out better quality data.
The other side of things is cultural. Too often predictive maintenance is seen as a nice, interesting innovation project. But this is real-world maintenance, it has real consequences for machines and humans, and so the business case should be very clear.
Rob Russell, Senseye: There needs to be a direct pull from the end-users and that pull should be created by central teams explaining the benefits of these solutions. It’s really important that you have senior-level buy-in for the process, but also that there’s dedicated resources that are incentivized to push forward on predictive maintenance activity.
Peter Gagg, MCP Consulting Group: If you don’t get a buy-in then you’re not going to make the new techniques sustainable. To have that sustainability, you really got to demonstrate that there’s some very significant benefits to be had - whether it’s reducing downtime, making the engineer’s time more effective, or making their life easier.
You’ve got to recognize that during that initial part of the journey, you’ll make mistakes and failures will happen. You’ve got to learn from that. No system is going to be fool-proof from the early stages. You’ve got to have that repeat learning process.
Alexander Hill: Is there a strategic objective to implement predictive maintenance? If this is just a science project, then it’s probably not likely to get the full exposure, buy-in, and interaction that it needs. You need to be putting this in the hands of maintenance - the people whose day-to-day lives will be improved by it. If it’s sitting in an R&D lab, it’s much harder to do that.
Rob Russell: There definitely needs to be a plan of how you take these new technologies and you put them into those operating environments. It has to be recognized that, while that’s taking place, there will be periods of inefficiency as people are getting up to speed with the new technology.
Peter Gagg: It keeps coming back to training the people, checking they’re following the right procedures or processes, and having some KPIs or metrics to demonstrate that you’re getting benefits from the new techniques that have been introduced. Make sure that you have monetarized the predictive maintenance program so that it becomes a profit generator and not a cost.
Alexander Hill: When we’ve proven value - we are typically able to do that in two months - then it’s a case of finding where else can we expand, is the data available, is the need there? Then it’s as simple as putting in the data, switching it on, and making sure we can put it into more people's hands.
