基于T-S模糊模型的燃气轮机系统负荷跟踪多目标预测控制

传统控制策略主要解决燃气轮机系统负荷跟踪问题，缺乏对经济性能的考虑。设计了基于T-S模糊模型的多目标预测控制器以同步提高系统跟踪性能和经济性能。首先，针对燃气轮机系统强非线性特性，选用T-S模糊模型增量结构，选取某联合循环机组历史数据进行模型辨识，所得模型能够根据运行工况实时更新模型参数，避免了模型失配问题。然后，基于此模型设计多目标预测控制器，定义负荷跟踪指标和经济性能指标，并对指标进行加权构成多目标代价函数。在多目标优化问题求解中，使用同步传热搜索算法优化代价函数获得控制量，提高控制系统快速性。仿真实验及评价指标结果表明，该多目标预测控制策略能够有效提高系统跟踪性能和经济性能。

Multi-objective Predictive Control of Gas Turbine System Based on T-S Fuzzy Model

The traditional control strategy of gas turbine system mainly focuses on the load tracking problem without taking the economic performance into full consideration. In this paper, a T-S fuzzy model-based multi-objective predictive controller is proposed to enhance both the tracking performance and economic performance together. First, regarding the strong non-linearities of gas turbine system, an incremental T-S fuzzy structure is applied and model identification is processed based on some historical data from a combined cycle unit. To avoid possible model mismatch, the parameters of the prediction model is updated in realtime according to the current operation conditions. Next, the multi-objective predictive controller is designed in which the load tracking index and economic index are defined and combined into a comprehensive multi-objective cost function. Then, in order to improve the settling speed of load tracking process, the simultaneous heat transfer search algorithm is employed to optimize the cost function and determine the control variables. Simulated experiment results have shown that this multi-objective predictive control scheme could enhance tracking performance and economic performance effectively.