This ‘predictive maintenance’ market is set to be worth an incredible $24.7bn globally by 2019, and ways to streamline processes and increase cost efficiencies remain an area of continued interest.
Companies often talk about predictive maintenance but its real value is not commonly understood and has yet to be widely used. Using conventional condition-based maintenance is currently the norm and, although predictive maintenance remains an industry buzzword, there has been limited adoption to date, despite the vast potential to significantly reduce costs.
How it works
The process itself is fascinating and complex, with many varied approaches. Realtime anomaly detection algorithms can pick up the tell-tale signs of a fault in a huge mass of data. Text mining is used to extract information and statistical trends in maintenance reports, while machine learning allows us to make predictions based on representative datasets of past activity.
To achieve the optimal balance in maintenance, parts need to be changed at just the right time – before a fault or failure occurs, to avoid further damage and limit the impact of systems downtime and loss of production. Such ‘preventative’ repairs must also be scheduled to minimise disruption to production.
With all the instrumentation and sensors built into our factories and systems today, combined with the connectivity of the Internet of Things, we now have access to a huge amount of potentially useful data for maintenance. However, it still needs to be extracted and interpreted.
This is where the magic of data science comes in, providing a set of methods and tools for turning raw data into value-added, actionable information. Data science helps us find correlations in these datasets, detect weak signals and develop predictive models.
At Assystem Technologies we recognise the importance of developing tools and services that our clients are able to integrate into their engineering operations. As a result, we have implemented a predictive maintenance model for drilling units. The process entails carrying out the identification of key parameters for maintenance of the drilling units and the development of a unique machine learning algorithm, enabling the creation of an efficient predictive mode.
Strengths and limitations
Implementing predictive maintenance in your engineering operations has a whole host of benefits apart from saving on maintenance costs. Predictive maintenance reduces unplanned downtime and ensures a more efficient labour planning approach. It also has the ability to identify the root of the fault, by providing a set of methods and tools for turning raw data into value-added, actionable information.
As with any ground-breaking innovation, predictive maintenance also has its limitations. Owing to the nature and complexity of the process, the accuracy of the tool can at times be compromised. Typically, the number of actual faults is very low, compared to the number of healthy components. The risk, therefore, is that the inaccuracies inherent in these models will lead to a significant number of false positives, or false indications of faults. That would actually increase maintenance costs.
All in all, the growth of predictive maintenance offers unparalleled opportunities for the industry and businesses that choose to implement it in their engineering practices. Innovation needs to remain at the heart of our ongoing journey of developing revolutionary tools and applications that can be utilised across many sectors including energy, automotive, aerospace and robotics.
The profound progress of the sector and the widespread adoption of new tools will only speed up the rate of growth and the beneficial impact on the future of our engineering operations. We should take it upon ourselves, as an industry, to recognise the potential of these new technologies and embrace the opportunity they provide, propelling our industry even further forward.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.