自适应动态规划在电力系统中的应用研究

发布时间：2018-08-02 17:48

【摘要】：电力系统是一个非线性、时变、高维的系统，如何进行有效的控制，以保证其更加安全稳定运行，一直是众多学者研究的课题，而且随着电力系统规模的不断扩大，以及随机性较强的新能源的接入，电力系统的安全稳定性面临的挑战越来越严峻。常规的最优控制方法存在一定的局限性，如变分法只能解决控制量不受约束的最优问题，极大/极小值原理只能解决常微分方程描述的最优控制问题，动态规划虽然能够解决比常微分方程所描述的更具有一般性的最优控制问题，但由于其是时间向后的算法，存在维数灾的问题。自适应动态规划（Adaptive Dynamic Programming, ADP）是人工智能和控制技术融合的产物，其实质是利用神经网络的函数近似结构来近似哈密顿-雅可比-贝尔曼方程的解。该方法可以不依赖被控对象的数学模型，也可以不需要精确定义性能指标，而且能够在线学习，将这种方法引入到电力系统中，能够为电力系统的非线性最优控制提供一种新思路。本文首先介绍了ADP的发展历程，并将国内外文献中有关ADP在电力系统中应用的研究进行了总结归类。然后介绍了ADP的理论来源、分类、优点、参数选择方法和归一化方法。接着推导出了单机无穷大系统的三阶状态方程，介绍了ADHDP算法的原理和PSO寻优PID参数的原理，将ADHDP应用于同步发电机的励磁控制，并与PSO寻优的PID励磁控制进行了比较，以验证其控制的有效性。最后介绍了STATCOM的数学模型和控制策略，以及GrHDP算法的原理，将GrHDP应用于STATCOM的电压无功控制，，并与MATLAB中可自动调节PID参数的PID控制进行了比较，以校核其控制效果。
[Abstract]:Power system is a nonlinear, time-varying, high-dimensional system. How to effectively control it to ensure its more safe and stable operation, has been a research topic of many scholars, and with the continuous expansion of the scale of power system, As well as the randomness of new energy access, the security and stability of power systems are facing more and more serious challenges. The conventional optimal control method has some limitations, such as the variational method can only solve the optimal problem with unconstrained control variables, and the maximum / minimum principle can only solve the optimal control problem described by ordinary differential equations. Dynamic programming can solve more general optimal control problems than those described by ordinary differential equations, but because it is a time backward algorithm, it has the problem of dimensionality disasters. Adaptive dynamic programming (Adaptive Dynamic Programming, ADP) is a product of the fusion of artificial intelligence and control technology. Its essence is to approximate the solution of Hamilton-Jacobi-Belman equation by using the function approximate structure of neural network. This method can not depend on the mathematical model of the controlled object, nor can it define the performance index accurately, and it can learn online, and introduce this method into the power system. It can provide a new idea for nonlinear optimal control of power system. In this paper, the development of ADP is introduced, and the research on the application of ADP in power system is summarized and classified. Then, the theoretical source, classification, advantages, parameter selection and normalization of ADP are introduced. Then the third-order state equation of single-machine infinite bus system is derived. The principle of ADHDP algorithm and PSO optimization PID parameter are introduced. The ADHDP is applied to the excitation control of synchronous generator and compared with the PID excitation control which is optimized by PSO. To verify the effectiveness of its control. Finally, the mathematical model and control strategy of STATCOM and the principle of GrHDP algorithm are introduced. GrHDP is applied to the voltage and reactive power control of STATCOM, and compared with PID control which can automatically adjust PID parameters in MATLAB to check its control effect.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM712

【相似文献】