A recent pilot application has demonstrated our unique QCM-based Artificial Intuition capabilities in producing anomaly pre-alarms as well as diagnosing their sources. The case in question is that of a vast array of monitored pipes and valves (detail of valve is shown below). Temperature and pressure have been monitored for each valve.
QCM has diagnosed a malfunctioning valve 72 hours before its actual breakdown.
The figure below depicts the evolution of complexity along the monitored period of time. It is evident that there are two sharp increases of complexity, pointing to similar behavior that constitutes indication of an anomaly. As the QCM algorithm computes and analyzes the complexity contributors in each iteration, we can track contributors’ behavioral changes in order to produce recommendations for immediate diagnostic and maintenance actions, well before the actual malfunction occurs.
Such diagnostics capability has been demonstrated and validated by our partners in a recent pilot project. The complexity contribution charts shown below (4 monitored parameters for each valve) show the behavior of the faulty valve (red). It can be seen that it is one of five leading contributors which is deteriorating over the period before failure occurs.
Complexity contribution charts
- The pilot application demonstrates Ontonix QCM capabilities in monitoring complex systems and processes.
- By early identification of the development of minor changes we can avoid risking the State of Health (SoH) of a system or a process. This leads to an advance diagnostic capability in supporting of maintenance engineering teams and helps eliminate/reduce the probability of a severe event.
- Due to the versatility, agility and scalability of the Ontonix QCM technology, the described and demonstrated capabilities can be easily tailored to any complex industrial use case and is the next-gen AI tool for proactive maintenance and implementation in the industry 4.0 context.
- Remember: We are doing this without Machine or Deep Learning.