Industrial applications have typically used some combination of SCADA, PLC, DCS and Historians to monitor and control plant operations over the years and in the process have accumulated years worth of process data which if mined and analysed properly can deliver spectacular improvement in plant operations.
Until recently, analytic tools available within SCADA or Historian were limited to trends and some limits processing and computing basic statistics e.g maxima, minima, standard deviation etc. Engineers and technicians used these tools to identify issues and problems following a defect or failure happening in the process control. This approach can be best described as Descriptive Analytics.
As opposed to the descriptive analytics, we employ Predictive Analytics using machine learning and artificial intelligence techniques including quantum computing principles. The AI techniques when applied on the vast amount of stored data generate patterns that can be used as a signature to detect anomalies in real-time, perform root cause analysis and predict an outcome with a certain degree of confidence. Moreover, these analytics can then be used to trigger an automatic workflow e.g, notification that an asset is likely to fail. Thus, managers and engineers can decide whether to run the asset to failure or organise a preventative maintenance.