Presentation: BTT: Online Process Monitoring using Historical Process Data

As part of ongoing efforts to reduce process variability in semi-automatic and manual reactors, a long term project was designed to move from the traditional process monitoring to a multivariate approach. As part of this, an intermediate solution was developed to improve the process monitoring strategy and at the same time provide a smooth transition for the operators. After discarding the golden-batch approach, the concept involves using historical batch data stored in PI, together with some simple data analytics, to provide the operators with additional guidance and basic alert and alarming capabilities. The solution can also be used as a training tool to improve process understanding, and also helps the process engineers with process upset investigations. The approach has proved useful and has resulted in significant improvements in reduction of process variability and cycle time, delay of need to invest, increased skilled operators availability, and faster scale-up process.