Although condition monitoring systems have been on the market and implemented for many years using a limited amount of data, it seems during this era of data overload generated by intelligent assets we get paralyzed about the quality and amount of data. It seems that it is holding off companies to implement new asset strategies because they feel they need more or better data.
The industry has been utilizing integrated systems for a long time but when it comes to Asset Performance improvement we are still waiting for more data to start with a company-wide digitalization project or even a Proof of Concept. We tend to forget the knowledge of our own people, but we also tend to forget that the right data could be residing within another department. Organizations who don’t have a joint asset management system in place still are confronted with silo’s which make the exchange of data more complex. Also we need to realize that a large portion of this data driven projects consist in data collection and cleansing.
As an industry we continue to ignore our people’s knowledge but also the knowledge that resides with our external contractors. Before we can fully rely on Machine learning or Artificial intelligence, we need to retain the knowledge within our organization and capture that knowledge in models and algorithms which should be retrained based on human interpretation. This will ensure that the confidence level of potential failures will increase and that the right actions are taken at an optimal moment in time ensuring the right balance between cost, performance and risk.
Successful asset performance management requires close cooperation between the maintenance, reliability, process engineering, and operations people in an industrial facility. And this is just a matter of fostering that collaboration culture.