基于机理-实测-辨识综合建模的大型水电机组调速系统控制参数优化方法

The large hydropower units are important equipment for the grid stability and high quality power supply due to their quick and flexible regulation ability of power. However, these advantages are not fully utilized because of the unreasonable control parameters of their governing systems. There are often the cases that they make a grid accident deteriorate. The fundamental reason is that the existing methods of the parameters optimization do not consider the engineering constraints as well as time-varying and nonlinear properties in the governing system. Therefore a system of the control parameters optimization is built, covering the theory, methods, tools and application for hydropower governing systems. The combined modeling method based on mechanism-field measurement-identification to build the precise models is proposed, and the novel intelligent algorithm with the new update strategy for penalty factors and the cautious mechanism for the optimization process is explored. It not only can handle the engineering constraints but also possesses the capability to precisely search global optimal solution. In addition, the new functional relationship between key performance indicators and control parameters is established to guarantee the capabilities of the setpoint tracking and regulation, disturbance rejection, noise reduction and robust stability of the systems. Finally the closed-loop research method is proposed, which is from the theory to the field experiments and then vice versa. According to the method, the optimum parameters are applied to the real-world unit and then the optimization theory and methods are improved with the actual effect. The aim of the project is to propose the advanced theory and methods of the modeling and optimization of the governing systems for the large hydropower units. Therefore, the project is theoretically significant and highly valuable in engineering.

大型水电机组因其能快速灵活调节功率而成为保障电网安全稳定运行和高质量供电的重要设备。但因其调速系统控制参数不合理而未能发挥优势，有时甚至恶化电网事故。根本原因是现有参数优化方法没有考虑工程约束条件和调速系统的时变与非线性。为此，本项目建立一套从理论、方法、工具到应用的水电机组调速系统控制参数优化体系。提出基于机理-实测-辨识的综合建模新方法，以建立反映工程实际的系统模型；研究基于新的惩罚因子变更策略和优化预警机制的新型智能优化算法，它不仅能处理工程约束条件，还具有全局搜索能力；建立各性能指标与控制参数间的新型函数关系，确保系统跟踪调节能力、干扰抑制能力、噪声衰减能力和鲁棒稳定性；提出"理论-现场试验-理论"的"闭环"研究方法，将优化参数运用到实际机组，再根据实际效果改进优化理论和方法。本项目旨在提出适用于大型水电机组调速系统建模与优化的先进理论与方法，具有重要理论意义和显著工程应用价值。

大型水电机组调速系统控制参数不合理会影响电网安全稳定运行和高质量供电，有时甚至恶化电网事故。根本原因是现有参数优化方法没有考虑工程约束条件和模型过于简单。.因此，提出基于机理-实测-辨识相结合的大型水电机组调速系统建模方法，将调节系统根据其相互关系分成若干模块分别建模，对控制器、液压系统进行参数实测；对水轮机、发电机、负荷通过辨识获得模型参数。已建立多种模型，以适应不同应用目的。.提出了一种基于改进蚁狮优化算法（IALO）的参数辨识，用高阶状态空间方程来表示水轮机调速系统模型。相比于GA、PSO和ALO算法，IALO收敛更快、精确度更高。空载工况属于闭环辨识，其方法研究较少。提出了一种基于PSO参数优化的PARSIM-K闭环辨识方法，在带有输出噪声的模型辨识中具有优良精确性；算法复杂度低，易于实现。.提出了一种改进的基于增广拉格朗日和粒子群鲁棒PID参数最优整定方法，使控制器获得满意瞬态特性、鲁棒稳定性和干扰抑制能力。在确定出鲁棒PID参数的同时，提供了全域稳定性。.建立了某水电机组孤网系统模型，成功再现频率振荡现象；提出了配置建议，增强了孤网系统稳定性。开发了与RTDS互联的水轮机-引水道-调速器实时仿真器，其仿真高精度再现系统在宽广区域的动态、稳态特性。.开发激励信号发生、仿真、测试、辨识于一体的综合试验系统，由激励信号发生、仿真、数据采集等功能的子系统和分析、测试、辨识等功能的综合子系统构成。技术先进、性能优越。.形成一整套包含试验方法、试验装置、参数辨识和控制参数优化工具的综合系统，包括：调速系统建模实测方案和方法、建模试验装置、建模与辨识软件、仿真-优化软件，完成从建模到参数优化的整个过程。已用于与中国电力科学研究院、南网电力科学研究院、贵州电力科学研究院、云南电力试验研究院的建模、参数优化、电网稳定性科研项目。.本项目成果具有理论意义和显著工程应用价值。