Optimal Failure Recovery in Batch Processing Using Petri Net Models

The transition of a batch plant from an error state to a safe state at minimum cost is a very complex process due to large state and decision spaces caused by the ststem's flexibility. A framework for system modeling, problem analysis and hierarchical control which can be used to guide the system to safety fast and in a cost effective way is presented here. The modeling is based on high-level Petri nets with timed places. Forward dynamic programming, with embedded heuristics is used in the optimization. The applicability of the approach is demonstrated in typical multi-purpose batch processing system.