多机电力系统励磁控制的人工神经网络逆系统方法

基于各发电机的完整的实用三阶模型，通过将发电机内部的不可测变量转化为发电机端的可测变量，在未做任何近似或简化的情况下，实现了仅采用本地可测量的多机系统中各发电机励磁控制的精确线性化。在具体的励磁控制器设计中，选取机端电压作为被控量再附加电力系统稳定器(PSS)环节进行辅助控制，使得同时满足系统对电压稳定性和功角稳定性的控制要求；采用人工神经网络(ANN)逆系统方法实现精确线性化，避免了对系统精确数学模型的依赖，使文中方法更实用。针对一个2区域4机系统的仿真结果表明，安装ANN逆励磁控制器可以极大地提高被控机组的稳定性能，同时，全系统的稳定性也可以得到明显的提高。

ANN INVERSION EXCITATION CONTROL METHOD FOR MULTI-MACHINE POWER SYSTEMS

Based on the standard simplified synchronous machine model with 3rd order differential equations, and by substituting the measurable variables at the terminal bus of the generator for the immeasurable variables inside the generator without any approximation or simplification, this paper gives a strict linearization method for the excitation system of each generator in multi machine power systems using only local measurable variables. In the process of designing the decentralized excitation controller, the terminal voltage is chosen as the control objective, while the PSS is affiliated in order to meet the system's control demand of enhancing the power angle stability and terminal voltage regulation performance simultaneously. As the ANN inversion method is used to realize the strict linearization, the proposed control strategy is independent of the system's precise model parameters and states, which makes it more suitable for the practical implementation. The simulation results of a widely used two area four machine power system demonstrate that the proposed ANN inversion excitation controller can significantly enhance both the controlled generator's stability and the whole system's stability. This work is supported by National Natural Science Foundation of China (No. 59925718, No. 60174004).