| 초록 |
A hierarchical synthetic real-time optimization (RTO) system is proposed to obtain the more profits than the typical RTO system although there are prediction errors and model performance changes. The hierarchical synthetic RTO system is composed of a conventional RTO system, an optimal choice system and a hierarchical synthetic decision system. The conventional RTO system is to optimize some portion of the utility plant, update the model parameters, and reconcile raw data from the plant. The optimal choice systems are designed to select the startup/shutdown units by integer programming. The hierarchical synthetic decision system analyzes the demand changes, decides the subsystems to synthesize total optimization problem, and determines whether or not the optimal choice systems select the units. We compared the results by the proposed method with those by the typical method. The optimization results are compared when prediction errors and model performance changes exist. When demand predictions are different from the current energy requirements, the operational cost is reduced by 2% compared with that of obtained by the typical RTO system. When model parameters are changed compared with those used in the multiperiod planning, the operational cost is reduced by 1.3% compared with that of obtained by the typical RTO system.
|
