【作者】 刘邦银；

【导师】 段善旭；

【作者基本信息】 华中科技大学 ， 电力电子与电力传动， 2004， 硕士

【摘要】 逆变器运行于开关状态,是一个非线性系统,其精确数学模型难于确定,而基于逆变模型设计的传统控制器的性能却取决于模型的精度,给高性能的传统控制器的设计带来困难。本文在逆变器控制中引入与被控对象模型无关、广泛适用于非线性系统控制的智能控制技术,研究了模糊控制、遗传算法和模糊神经网络在逆变器控制中的应用。逆变器中的扰动包括周期性扰动和非周期性扰动。基于内模原理的重复控制技术是补偿周期性扰动的理想方案,而对非周期性扰动的补偿效果却很差。文中提出了逆变器的几种智能控制与重复控制相结合的混合控制方案,用重复控制补偿周期性扰动,智能控制补偿非周期性扰动,以获得较好的动态和稳态性能。本文首先建立了逆变器的数学模型,设计了重复控制器,然后针对重复控制器动态性能差的问题,采用电压瞬时值反馈PD控制器改善系统动态特性,输出电压均值反馈PI控制器提高输出电压稳定度。 由于逆变器的精确数学模型不易获得,高性能的PD控制器设计困难,本文提出了一种直接模糊重复混合控制方案,介绍了控制器的设计方法,分析了控制器的性能。在模糊控制器设计过程中,控制器参数都需要通过经验和大量的仿真实验确定,给设计带来困难,本文研究了采用遗传算法来设计和优化控制器参数,给出了优化设计方法和仿真实验结果,并与PD重复控制和直接模糊重复控制进行了对比分析。经过遗传优化的模糊控制器性能良好,然而由于控制器的设计是离线进行,控制参数不能根据系统运行状态实时自适应调整来获得当前最优控制器结构,因此本文又尝试采用模糊神经网络控制来实时改善系统动静态特性。在文中最后根据仿真和实验结果,对比分析了这几种控制方案的优势和存在的问题,给出了未来的研究建议。更多还原

【Abstract】 It is well known that the inverter system is working in switching mode, and can be regarded as a nonlinear system, so the mathematical model of the inverter is difficult to acquire. Because the performance of the classic controller is restrained by the precision of the model, designing an excellent classic scheme of the inverter become harder. In this paper, the intelligent control technology is introduced into the inverter system, and fuzzy control technique, genetic algorithm and fuzzy neural network control technique, which widely applied in the control of nonlinear system, are studied.The disturbance of the inverter include periodic and non-periodic disturbance. Repetitive control technique based on the inter model principle is a strategy to compensate the periodic disturbance, however, it is incapable to compensate the non-periodic disturbance. Several hybrid control schemes of intelligent control and repetitive control are studied in the paper. The intelligent controller is used to improve transient performance, while repetitive controller is applied to generate high-quality sinusoidal output voltage in steady state. Based on the mathematical model, the repetitive controller is designed firstly. PD controller is introduced, due to the bad transient response of the repetitive controller, at the same time, the PI controller is designed to insure the precision of output voltage.Considered the nonlinear of the inverter, direct fuzzy control scheme is designed to improve the transient performance of the inverter and the performance of the hybrid controller is analyzed. In order to acquire the good parameters, an optimized design method based on the genetic algorithm is studied in this paper. The results of the simulation and are given to analyze and compare the performance of the PD repetitive controller and direct fuzzy repetitive controller.The parameters of fuzzy controller can not adaptively adjust to acquire an optimum structure online, because it is designed offline. A fuzzy neural network control strategy is presented to improve the adaptability of the inverter under the condition of sudden changes in the load. Simulation and experimental results proved the advantage of intelligent control technique in the real time control of the inverter, as well as,some questions of the application of intelligent control technique are discussed and analyzed. Finally, some suggestions for farther research are presented.更多还原