Hussain Syed Asad and Richard Kwok Kit Yuen and Gongsheng Huang
Hybrid Adaptive Modeling to Enhance Robustness of RealTime Optimization
361 - 365
2017
11
4
International Journal of Energy and Power Engineering
https://publications.waset.org/pdf/10006605
https://publications.waset.org/vol/124
World Academy of Science, Engineering and Technology
Realtime optimization has been considered an effective approach for improving energy efficient operation of heating, ventilation, and airconditioning (HVAC) systems. In modelbased realtime optimization, model mismatches cannot be avoided. When model mismatches are significant, the performance of the realtime optimization will be impaired and hence the expected energy saving will be reduced. In this paper, the model mismatches for chiller plant on realtime optimization are considered. In the realtime optimization of the chiller plant, simplified semiphysical or grey box model of chiller is always used, which should be identified using available operation data. To overcome the model mismatches associated with the chiller model, hybrid Genetic Algorithms (HGAs) method is used for online realtime training of the chiller model. HGAs combines Genetic Algorithms (GAs) method (for global search) and traditional optimization method (i.e. faster and more efficient for local search) to avoid conventional hit and trial process of GAs. The identification of model parameters is synthesized as an optimization problem; and the objective function is the Least Square Error between the output from the model and the actual output from the chiller plant. A case study is used to illustrate the implementation of the proposed method. It has been shown that the proposed approach is able to provide reliability in decision making, enhance the robustness of the realtime optimization strategy and improve on energy performance.
Open Science Index 124, 2017