CHP – System-Wide Optimization

CHP systems can become quite complex. Multiple fuels, multiple run modes, multiple assets make optimization of the system too complex to perform manually. As CHP systems become more complex and diversified, a new robust optimization framework must be employed to maximize profits. The core optimization and control models must be simplified and shared. This framework must be sufficiently decentralized to accommodate local support of individual models. By sharing local control models with the system-wide optimizer, all stake-holders can be satisfied. A combination of first-principle and empirical modeling techniques must be employed to reduce deployment time and increase maintainability.

Optimization has become such an over-used term that it can mean anything from changing a filter to the type of system-wide optimization presented here, one that maximizes profitability of the entire system. Earlier generations of system-wide optimization have been less than successful precisely because of some of the reasons discussed here. Schemes based solely on 1st principle modeling are incredibly expensive to implement and maintain. Having separate control models and optimization models provides two versions of the “truth”, leading to sub-optimization of components rather than optimizing the complete system.

Combing both Model Predictive Controls (MPC) and System-Wide Optimization in one framework provides the best benefits of all. MPC controllers implemented at the plant level can operate units safely closer to constraints, environmental and equipment related. Sharing these models and constraints with the optimization framework provides true system-wide optimization.