【作者】 张彪；

【导师】 赵克定；

【作者基本信息】 哈尔滨工业大学 ， 机械电子工程， 2009， 博士

【摘要】 空气动力电液负载模拟器(Electro-hydraulic load simulator)是一种地面半实物仿真设备,用来模拟飞行器舵机所受的各种空气动力力矩载荷谱,从而检测舵机驱动系统的技术性能指标,其对国防军事工业具有重要意义。由于电液负载模拟器是典型的被动式力(力矩)伺服控制系统,仿真过程中,舵机的主动运动会对与其固连在一起的加载系统产生一种干扰作用,使加载系统两腔形成强迫流量,进而产生多余力矩。多余力矩具有强度大、时时存在、随舵机运动状态连续变化的特点,严重影响电液负载模拟器的加载性能。因此,如何抑制多余力矩,提高加载性能是研究电液负载模拟器的关键技术问题。本文在查阅大量国内外资料的基础上,综述了国内外电液负载模拟器的发展现状,分析了它们的优缺点,通过对国内外负载模拟器产品和相关资料的分析,确定了本文的主要研究方向。首先分别建立了电液负载模拟器的四阶和五阶数学模型,并通过系统辨识和实验数据验证,得出在一般加载工况下可以使用四阶模型代替五阶模型。并提出反步遗传算法的模型参数辨识方法,使模型辨识结果更接近实际物理系统,并对电液负载模拟器四阶系统进行了参数辨识,并验证了该辨识方法的有效性。根据所建数学模型,深入分析电液负载模拟器多余力矩形成的原因和特点,从解决系统强迫流量入手,提出使用可调节流阀来抑制多余力矩的方法,建立了含可调节流阀的电液负载模拟器数学模型并进行了仿真验证。得出使用可调节流阀可以达到排泄强迫流量从而有效抑制多余力矩的结论。为解决因舵机扰动而产生多余力矩的问题,提出一种神经网络补偿控制以补偿舵机端对加载系统扰动的方法。以舵机输出的各阶状态为神经网络输入,根据系统多余力矩大小神经网络调整补偿控制器输出,进而补偿给加载系统,从而抑制由于舵机运动产生的多余力矩。并将神经网络补偿控制器与传统PID控制器、结构不变性原理等进行仿真对比,得出神经网络补偿控制器能更好抑制多余力矩的结论。针对系统具有强耦合干扰的特点,提出考虑参数变化的电液负载模拟器反步自适应控制器。将复杂的系统模型分解成若干个相对简单的子系统模型,然后针对每个子系统模型设计控制器,系统的耦合项通过每个子系统控制器的设计而被迭代到系统的最终控制器中。在控制器设计时还考虑了系统参数变化的因素,设计了参数变化的自适应率。并将该控制器进行了各种工况下的仿真研究,得出反步控制器能有效抑制多余力矩,提高系统加载性能。最后在电液负载模拟器原理样机上进行了实验研究。验证了关于多余力矩形成机理及影响因素分析的正确性;进而分别将神经网络补偿、反步自适应控制应用在电液负载模拟器实验样机上,验证其在多余力矩抑制和力矩加载控制性能等方面的有效性;还验证了可调节流阀抑制多余力矩有效性。更多还原

【Abstract】 Electro-hydraulic load simulator(EHLS) is a Hardware-in-the-loop simulation equipment on ground. It is applied to check the qualification of rudder driving system by simulating aerodynamic load spectrum to the flight device rudder. Therefore, it is significant to national defense and military industry.As EHLS is a typical passive force (torque) servo control system, when it is running, the active motion of rudder will have a disturbance to loading system which is connected to the rudder. This disturbance results in forced flow between cavities of loading motor. As a result, the extra torque appears. The extra torque is known for its large power, existence at times and the variation caused by motion of rudder. These characteristics affect the performance of EHLS severely. Therefore, how to restrain extra torque and improve the loading performance is key technology in the study of EHLS.After synthesizing numerous related literatures and reference materials home and abroad, the development of EHLS system is summarized. The studying direction and content are decided by analyzing the merit and shortcoming of related products and studies.First, the fourth and fifth order mathematical model of EHLS system is built. Then the system identification and verification are made. The result shows that the fourth mathematical model has equal quality to fifth order model when system is working under the normal work condition. Then genetic algorithm identification based on backstepping theory is proposed, which makes the result of identification more close to physical system. Then the proposed method is used for the fourth order model. Verification result proves the effectivity of the method.The generating mechanics and characteristic of EHLS extra torque is analyzed according to the system model. Considering the forced flow of system, adjustable throttle to solve the extra torque is proposed, and the mathematical model with adjustable throttle is built. The simulation result shows using adjustable throttle can relieve forced flow, and the extra torque can be restrained as a result.To solve the extra torque which caused by movement of rudder system, the neural network compensation control strategy is proposed. It is used to compensate the disturbance of rudder system. Each order of states of the rudder output is taken as neural network inputs. Compensator changes its output according to extra torque and compensates the load system to restrain the extra torque. Then the simulation comparison among PID controller, structure invariability and the proposed controller is given. The result shows the proposed method’s effectivity. Aiming at the coupling between load system and rudder system, a adaptive backstepping controller with parameters variation adaptometer of EHLS is proposed. The complex system with high order is decomposed into a serial of simple low order subsystems. Then the controllers are designed for each subsystem. By doing these, the coupling relationship is passed to the system controller as a result. The variation of main parameters is considered at the same time, and parameters variation adaptometer is designed. The simulation results under different conditions show that the proposed controller can restrain extra torque effectively, and the load performance is finally improved.Experiments are made on the EHLS principle model machine. The extra torque generating mechanics and influencing factors are verified at first. Then experiments about neural network compensation controller and backstepping adaptive controller are made on the system. The performance about restraining extra torque and loading tracking of these two controllers are tested. The results show the controllers’effectivity. The experiments about adjustable throttle are made at last, the result shows its effectivity.更多还原