(745h) Knowledge Sharing of an Optimizer Development and Sustainment

The knowledge and learning of developing and sustaining a chemical plant optimizer is shared. At plant product A is produced from five reactor trains which are controlled separately by two control rooms. The optimizer is a steady-state economic optimizer based on first principle fundamental models. It has fundamental kinetics and plant operation limits. It also has real prices which are used in optimization objective function which is maximized as the revenue generated from products minus the cost of the feed and utilities. The procedure of RTO model development is described with example of reaction system and steam system models. It also demonstrates how to do simplification without sacrificing the optimizer performance. The model has been validated with plant data at different operating conditions. The best practice on reactor model parameterization is shared along with sensitivity studies of feed conversion. The optimizer benefits come from both yield optimization and optimal reactor load balancing. Load management and shedding is achieved through plant operation providing setpoint of feed and RTO economically distributing the load to each reactor train. The optimizer performance is demonstrated with some case results analysis. The best practice to sustain the optimizer by effectively collaborating with operation teams is shared as well. At the same time interactive KPI dashboard plays a key role for monitoring the performance of the optimizer.