【作者】 张光辉；

【导师】 刘军； 李强；

【作者基本信息】 西安理工大学 ， 检测技术与自动化装置， 2002， 硕士

【摘要】 输油泵系统是一个具有较强非线性、滞后性、强耦合和多种不确定因素的复杂被控系统，采用传统的控制方法难以实现对输油泵系统的良好控制。本文分析了原PID控制方案导致系统剧烈振荡的主要原因，并且针对系统的非线性特点和系统控制要求，提出了将控制量基值与调节值相结合实现对输油泵系统控制的控制策略。在这种控制策略中转速控制量的基值根据修正后的相似公式进行计算，而转速控制量的调节值为模糊控制器的输出，这样做既保证了系统的稳态精度；也避免了系统振荡。为了实现方便，本文第三章中分别设计了两个二维模糊控制器，采用两个模糊控制器融合后的结果作为转速控制量的调节值，但是这种控制方法的稳态性能和抗扰动能力不是很理想。 由于在模糊控制器的设计过程中存在较多的人为因素，为了实现根据系统特性对模糊规则和隶属函数进行自动修正、完善和调整，本文将遗传算法和模糊控制结合起来，并针对前面设计的模糊控制器中所存在的问题进行了详细分析，提出了两种改进方案： 1．在简单模糊控制器的输入变量中加入了变量变化率的信息，即将输入变量和变量的变化率融合为一个输入量，并在模糊控制器的输出端加入比例、积分环节，然后分析了这种改进后的模糊控制器的解析结构，最后采用改进后的自适应遗传算子的遗传算法对模糊控制器中的隶属函数和融合因子进行优化，并将优化前后的结果作了比较和分析。 2．利用模糊逻辑强大的结构知识表达能力和神经网络强大的自学习能力，构成模糊神经网络控制器，通过训练学习，自动调节隶属函数的结构和参数，从数据中自动提取有效的控制规则。本文应用Sugeno模糊模型构造了模糊神经网络控制器，并采用改进后的遗传算法对模糊神经网络控制器当中的参数和隶属函数进行了学习和优化。 通过对上述两种方案的仿真研究，结果表明改进后的模糊控制器改善了系统的控制性能，达到预期的控制效果。 最后采用第一种改进方案，实现了输油泵系统的模糊优化控制，并且消除系统切换时的扰动。现场调试结果表明采用模糊优化控制方法实现了输油泵系统的流量、入口压力、出站压力的协调控制，达到了设计要求。系统具有响应速度快、稳态精度高、稳定性好等特点，并且具有较强的自适应能力和鲁棒性。更多还原

【Abstract】 Oil Feeding Pump System is a complex system, which exits serious non-linearity > time-delay,strong couple and some unascertained factors, so it can’t achieve better results by traditional control methods. The paper analyzes the main reason of causing system oscillation when it is adjusted by original PID control method, and proposes control strategy of combining basic value of rotation rate with adjustment value to control pump rotation speed according to non-linear characteristic of the system and control request. In the control strategy, the basic value of rotation rate is calculated by corrected similar formulation and the adjustment value is ascertained by output of fuzzy controllers, the strategy can assure steady precision and to eliminate vibration. In order to realize simply and conveniently, two two-dimensional fuzzy controllers are designed, and the outputs of the controllers is used as speed adjustment, but the’steady precision and anti-interference ability of this fuzzy control method are not desirable.There exits many subjective factors during designing the fuzzy control, in order to make fuzzy control rules and membership functions correction ., perfection and adjustment according to system properties, it combines genetic algorithms with fuzzy control, detailed analyzes the problem of designing fuzzy controller and proposes two advanced schemes:First scheme: The change-of-variables are emerged into input variables of the simple fuzzy controllers of Oil Feeding Pump System as one variable, and one PI block is connected after output of fuzzy controllers, consequently the structure of the improved fuzzy controller is analyzed, finally genetic algorithms with adaptive probabilities of crossover and mutation is applied to optimize membership functions and fusing factors of the fuzzy controllers , and the simulation results of before and after optimization are compared.Second scheme: Fuzzy neural networks controller is constructed by employing the great structure knowledge description ability of fuzzy logic and the great studying ability of neural networks to tune structure and parameters of membership functions, and extracts effective fuzzy control rules from data. The paper designs fuzzy neural networks controller based on Sugeno fuzzy model and applies the improved genetic algorithms to tune the networks parameters and membership functions of the designed fuzzy neural networks controller.According to simulation research of two control strategies, the simulation results show that the improved fuzzy controllers have improved control and achieve desired control results of Oil Feeding Pump System. The paper adopts the first advance scheme to realize the fuzzy optimization control of Oil Feeding Pump System and eliminates the turbulence generated by changing control ways. The experiment results demonstrate that the fuzzy control method realizes coordination optimization control of technique operation parameters of the system, what’s more , the system has the characteristics of fast adjusting speed, high steady precise, good steady ability and has the better adaptive ability and robustness.更多还原