Automatic Generation of Emergency Response Procedures for Batch Processes

Quick response to mitigate the damaging effects of accidents has always been an important operational issue for chemical plants. In the present study, a systematic automata-based procedure is developed to automatically generate fail-safe operation steps in a given batch process. Specifically, three distinct sets of automata are first built off-line to respectively characterize the plant behaviors, the control specifications and also auxiliary constraints. By applying the parallel composition operation on the first two types of automata, an admissible supervisor for a given failure-induced system state can be generated online to steer the system away from hazardous conditions while still maintaining sustained production. In order to identify an efficient procedure, the auxiliary automata can be augmented with this supervisor to set the operation target(s) and to limit the total number of actuator actions. Two examples will be presented to demonstrate the effectiveness and feasibility of the proposed method.