基于智能优化方法的永磁电机驱动液压动力源控制策略研究

发布时间：2018-01-05 09:01

本文关键词：基于智能优化方法的永磁电机驱动液压动力源控制策略研究　出处：《西安建筑科技大学》2014年博士论文　论文类型：学位论文

【摘要】：本文研究的液压动力系统采用永磁伺服电机带动齿轮泵作为液压动力源，不仅具备结构简单、高可靠性、较宽调速范围、高节能率的优点，而且克服了传统液压系统结构复杂、高能耗等诸多不足，容易实现闭环控制。目前在控制策略上仍然广泛采用PID控制，由于液压系统在负载变化时流量和压力的强耦合特性，控制对象仍然具有不确定、时变和高度非线性[129]，采用简单PID线性控制器往往不能得到较好的控制性能。为此出现了多种先进控制技术结合智能控制运用到液压系统中都取得好的控制效果。目前，由于智能控制的基础理论发展仍不完善，所以各种智能控制方法的综合应用还存在许多需要我们改进的地方。因此，本论文结合模糊逻辑、神经网络、遗传算法、粒子群等优化算法，旨在设计出能提高液压动力源控制品质的控制器，并研究其参数优化方法。具体创新点和研究工作包括以下内容： (1)利用解析法对永磁交流伺服电机驱动定量泵（液压动力源）进行了数学建模。在分析永磁电机物理方程、转矩方程及其基于坐标变换的三环调节矢量控制系统的基础上，用MATLAB的simulink搭建了永磁伺服电机驱动定量泵的系统仿真模型，为后续章节控制系统的设计提供仿真平台，为研究控制参数优化算法提供理论支持。 (2)遗传算法中如果交叉率及变异率保持不变，极易引起过早收敛、陷入局部极值等问题，针对上述问题提出了利用模糊控制器调整交叉率及变异率的遗传参数自适应调整算法，提高算法的收敛速度和获得全局最优解的能力。通过对永磁电机驱动的液压系统流量进行常规优化方法和改进遗传优化方法控制的对比，仿真和实验结果表明：改进遗传优化方法，可使系统在复杂工况下，保持良好的控制性能,并且具有较高的控制精度和鲁棒性。 (3)由于实际的液压系统参数存在时变性，，系统易受外界载荷的干扰，具有非线性、强耦合的特征，难以建立准确的数学模型，针对上述问题采用了粒子群结合BP混合优化算法，优化前向神经网络PID控制系统。该控制系统PID控制器参数可通过神经网络自学习调整，该控制策略较好的结合了粒子群优化算法和BP算法的优点，先用粒子群算法离线优化后用BP算法在线优化控制器参数。并将其运用于永磁伺服电机驱动的液压系统中，仿真结果验证了该系统在各种典型工况下良好的动静态性能。 (4)智能控制方法的综合应用可以扬长避短、相得益彰。针对液压系统的非线性、强耦合特征，本文采用了一种新的神经网络控制方法，融合了模糊控制、神经网络及PID控制各自的特征。将专家推理和神经网络的自学习功能相结合，使神经网络的性能更加完善，同时采用RBF网络在线辨识，向神经网络控制器提供变化的梯度信息，进一步提高系统的控制性能。对液压动力系统进行了典型工况下的流量跟踪控制仿真。仿真结果验证了本文的综合控制方案优于单一控制方法，系统的各项控制指标均得到提高。（5）在深入研究传统PID控制和模糊控制原理的基础上，分别实现了变频调速液压动力源流量的实时在线控制。并结合具体的工况分析了传统PID和模糊控制各自的特点，得出模糊控制在液压系统正弦加载的工况下具有比PID控制更强的鲁棒性的结论，较适合于载荷频率变化较快的场合应用。（6）由于PID控制算法简单易行，大多数工业控制仍采用传统PID控制，在具体应用时存在一定缺陷,如：响应快速和超调小很难同时达到最优，所以在要求较高的场合PID控制不能满足要求。针对上述问题论文提出了模糊PID串联复合控制策略，充分将模糊控制的快速性与PID控制精度高的特点相结合，实现了液压动力源流量的实时在线控制，实验结果表明：复合控制响应快速、无超调、精度高，控制性能明显优于单一控制方法，适合控制要求较高的场合。
[Abstract]:The hydraulic power system is studied in this paper using permanent magnet servo motor drives the gear pump as the hydraulic power source, not only has the advantages of simple structure, high reliability, wide speed range, the advantages of high energy saving rate, and to overcome the traditional hydraulic system of complicated structure, high energy consumption and other shortcomings, to achieve closed-loop control. At present, the control strategy is still widely using PID control, due to the strong coupling characteristics of the hydraulic system flow and pressure changes in the load, control object is still uncertain, time-varying and nonlinear [129], using a simple PID linear controller often can not get good control performance. Therefore the emergence of a variety of advanced control technology combined with intelligent control theory is used to control the good effect get in the hydraulic system. At present, due to the development of the basic theory of intelligent control is still not perfect, the comprehensive application of intelligent control methods so there are various We need to improve the place. Therefore, this paper combines fuzzy logic, neural network, genetic algorithm, particle swarm optimization algorithm to design controller can improve the quality control of hydraulic power source, and to study the parameter optimization method. The innovation and research work includes the following contents:
(1) of the permanent magnet AC servo motor driven pump (hydraulic power source) using the analytic method of mathematical modeling. In the analysis of permanent magnet motor torque equations and physical equations, based on coordinate transformation of tricyclic regulation of vector control system, using MATLAB Simulink built the simulation model of constant pump permanent magnet servo motor drive the simulation platform provides for the design of control system for the following chapters, the parameters provide theoretical support for optimization control research.
(2) if the genetic algorithm crossover rate and mutation rate remained unchanged, extremely easy to cause the premature convergence problem into local extremum, aiming at the above problems by using fuzzy controller to adjust the crossover rate and mutation rate adaptive genetic algorithm parameters is proposed to improve the convergence speed and the ability to obtain the global optimal solution. The improved contrast control of genetic optimization method based on the flow of the hydraulic system of permanent magnet motor driven by conventional methods and optimization, simulation and experimental results show that the improved genetic optimization method, can make the system under complicated conditions, maintaining good control performance, and has high control precision and robustness.
(3) due to the presence of the hydraulic system parameters of practical time-varying system, vulnerable to interference, the external load is nonlinear, strong coupling characteristics, it is difficult to establish accurate mathematical model, aiming at the above problems using particle swarm optimization combining BP hybrid optimization algorithm, feedforward neural network PID control system. The control parameters of PID controller the system can through the neural network self-learning adjustment, the control strategy is a good combination of the advantages of particle swarm optimization algorithm and BP algorithm, using particle swarm optimization algorithm with off-line optimization optimization of controller parameters online. BP algorithm and applied to the hydraulic system of permanent magnet servo motor drive, the simulation results show that the system in the a variety of typical conditions good dynamic and static performance.
(4) the comprehensive application of the intelligent control method can avoid weaknesses and complement each other. According to the nonlinear hydraulic system, strong coupling characteristics, this paper proposes a new control method of neural network, combines fuzzy control, neural network and PID control of their respective characteristics. The self-learning function of combining expert reasoning and neural network. The performance of the neural network is more perfect, and the online identification of RBF network, to provide gradient information changes to the neural network controller, to further improve the control performance of the system. The typical working conditions of flow tracking control simulation of the hydraulic system. The simulation results verify the comprehensive control scheme is better than that of the single control method in this paper, the control indexes of the system are improved.
(5) based on in-depth research of traditional PID control and fuzzy control principle, realize real-time control of VVVF hydraulic power flow. Combined with the specific condition and analyzed the characteristics of traditional PID and fuzzy control, the fuzzy control has the robustness of control is stronger than PID in the hydraulic system of sine the loading conditions, suitable for load frequency fast changing situations.
(6) the PID control algorithm is simple, most industrial control is still using the traditional PID control, there are some defects in specific applications such as fast response and small overshoot is difficult to achieve optimal at the same time, so can not meet the requirements in the occasions with higher requirements for the PID control. The proposed fuzzy PID series composite the control strategy, and PID features fully fast fuzzy control precision high combination realizes real-time control of hydraulic power flow, the experimental results show that the composite control of fast response, no overshoot, high control precision, better performance than that of single control method for the control requirements of the occasion.

【学位授予单位】：西安建筑科技大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TM351;TP271.31

【参考文献】