【作者】 谢苗；

【导师】 毛君；

【作者基本信息】 辽宁工程技术大学 ， 机械设计及理论， 2012， 博士

【摘要】 针对大型锻造设备工作中振动冲击大、精确控制难的问题，以液气锤为对象，研究了锻造设备系统的动力学特性、控制策略与控制技术。通过对液气锤原系统和工况分析，提出了以大流量三级插装阀组为执行元件控制液气锤进排油过程，实现液气锤控制中的快速响应；设计了一种定位悬锤、泄漏主动补偿的液压控制回路，实现了液气锤任意行程打击并保障了工作的安全性；设计了液气锤控制系统的检测回路，分别对液压系统的输出油液压力、工作腔的气体压力和锤头提升高度进行实时监控，实现了打击工况的安全检测；设计了基于打击数据样本和模糊自适应控制器的PLC控制策略，提出了液气锤的有效打击时间控制方法，实现液气锤打击的自动控制。根据液气锤系统构成、工作原理和工况特性，将液气锤整机动力学系统划分为三个子系统：机械系统（含气压系统）、液压系统和控制系统。基于不同工况分别研究了各子系统及其主要元器件的数学模型和动力学特征，并构建了相应工况下的各子系统及其主要元器件的动力学模型和液气锤整机动力学模型，研究了模型中参数的确定和计算方法。研究了系统动力学特性的数字仿真方法，利用仿真软件分别对机械系统的动力学特性、液压系统及其主要元器件大流量插装阀的动力学特性、控制系统主要元件的响应特性进行仿真。利用子系统仿真数据和液气锤设计参数，建立液气锤整机系统动力学特性模型并进行了仿真，研究了影响系统动态特性的因素，仿真结果与试验测试结果比较，证明了理论分析的正确性。研究了基于模糊神经网络的液气锤动态特性的控制理论及方法，提出了有效打击、有效打击时间、可控性和可控域概念，通过试验建立了不同打击工况下的有效打击数据样本。导出了控制系统的传递函数，建立了模糊神经网络的控制规则，并利用样本中数据对网络结构进行训练，通过仿真确定了模糊控制器的控制规则。通过实验，证明了这种控制方法具有良好的动态响应特性和稳定性。将上述研究成果应用到40t·m液气锤技术改造中，经过一个月的试生产后转入正式应用，到目前为止已经应用20个月，效果良好，误差均控制在0.03s以内，证明了理论研究结果的正确性。更多还原

【Abstract】 Aiming at the big vibration shocking and the difficulty to control accurately of the largeforging equipment, and taking the liquid-air hammer system as the research object, the dynamicsperformance, the control strategy and the control technology of the large forging equipment arestudied.By analyzing the original liquid-air hammer system and the working conditions, the methodof controlling the sucking and exhausting oil process by taking the three-stage cartridge valve oflarge flow level as the actuator is proposed, and the quick response of the liquid-air hammercontrol is obtained. A hydraulic control loop system by fixed hanging weight, activecompensation for divulging is designed, the random stroke hammering of the liquid-air hammeris realized and the working safety is guaranteed. The PLC control strategy based on hammeringdata sample and fuzzy self-adaptive controller is designed, the effective hammering time of thehammer is put forward and the automatic control is realized.According to the structure of the liquid-air hammer system, the working principle and theworking conditions, the dynamics system of the liquid-air hammer system is divided into threesubsystems: the mechanical system(including the pneumatic system), the hydraulic system andthe control system. Based on the different working conditions, the dynamics performances ofeach subsystem and the parts within it are studied, the dynamics models of each subsystem fordifferent working conditions are established, the dynamics model of the total liquid-air hammersystem is set up, and the parameter determination and calculation of the models are researched.The numerical simulation of the dynamics performance of the system is studied, thedynamics performances of the mechanical system, the hydraulic system, the cartridge valve oflarge flow level and the response performance of the major parts of the control system aresimulated. The dynamics performance model of the liquid-air hammer system is set up andsimulated according to the simulation data of the subsystems and the design parameter of theliquid-air hammer system. The elements influencing the system performance are researched, andthe validity of the theoretical analysis is proved by comparing the simulation results and theactual experiment results.The control theory and method based on the fuzzy neural networks of the liquid-air hammeris studied, the conceptions of the effective hammering, the effective hammering time, the controllability and controllable field are put forward. The effective hammering samples fordifferent hammering conditions are established by the actual experiment. The transferringfunction of the control system is derived, the fuzzy neural network control laws are established.The network structure is trained by using the data of the samples, and the control laws of thefuzzy controller are determined by simulation. By the experiment, the control method is provedto be of good dynamic response performance and good stability.The research mentioned above is employed in the technical upgrade of the40t.m liquid-airhammer system, and the system turns to formal running after one-month trial production. Thesystem has been running20months so far, the working effect of the system is good, and theerror is controlled to be within0.03s, which verifies the validity of the theory research.更多还原