基于神经网络的盾构推进液压系统控制策略研究

发布时间：2019-03-04 18:02

【摘要】：盾构施工技术以其快速高效、质优环保、安全可靠等优势,逐渐得到推广及应用。近几年,国产盾构的成功下线以及海外隧道项目的顺利验收,表明我国在盾构自主研发领域取得了历史性的突破。但在盾构自适应设计、系统集成技术等领域的关键技术仍被欧美、日本等发达国家所掌握。推进系统是盾构自适应设计的关键部分,盾构的直线前进、姿态调整以及方向改变都依赖于推进系统各区油缸的协调配合,推进系统控制的优劣对盾构掘进机的性能和隧道质量显得至关重要。因此,开展盾构推进系统控制策略的研究,对推动我国盾构产业的发展具有深远意义。本文以盾构推进系统为对象,研究推进系统在不同负载工况下推进压力和速度的响应特性,以及多组推进油缸的同步控制性能。首先介绍了盾构推进系统的机械结构、液压系统、以及实验台的工作原理;然后,利用AMESim软件建立液压系统的物理模型,采用常规PID控制对推进速度和压力响应特性以及多缸同步性进行分析,仿真结果表明推进压力对速度影响很大,常规PID控制下速度控制效果不理想。为此,本文针对盾构推进系统在变负载、变流量工况下,推进压力和速度之间相互干扰的问题,提出BP神经网络PID的控制策略,并结合Simulink软件进行联合仿真,仿真结果表明BP神经网络PID的稳定性好,很大程度上消减了盾构推进中流量、负载变化时推进油缸压力和速度的超调;为提高盾构刀盘在不均衡负载下的多缸同步控制性能,提出单神经元PID控制策略,经比较分析推进系统在不同控制策略下的响应特性,得出采用单神经元PID能够大大提高推进油缸在偏载时运动的同步性。最后,在盾构实验台上对以上控制策略进行试验验证,试验数据与仿真结果基本一致,证明了神经网络结合经典PID控制抑制推进系统不确定性和非线性的有效性。本课题的理论研究和试验分析结论,为盾构控制系统设计和优化提供了理论参考。
[Abstract]:Shield construction technology has been gradually popularized and applied because of its advantages of fast and high efficiency, high quality and environmental protection, safety and reliability. In recent years, the successful disembarkation of domestic shield and the successful acceptance of overseas tunnel projects show that China has made a historic breakthrough in the field of autonomous research and development of shield. However, the key technologies in shield adaptive design, system integration technology and other fields are still mastered by developed countries such as Europe, America, Japan and so on. Propulsion system is the key part of the adaptive design of shield. The straight forward, attitude adjustment and direction change of shield depend on the coordination of various cylinders in the propulsion system. The pros and cons of propulsion system control are very important to the performance of shield tunneling machine and the quality of tunnel. Therefore, the research on the control strategy of shield propulsion system is of great significance to the development of shield industry in China. Taking shield propulsion system as an object, the response characteristics of propulsion pressure and velocity under different load conditions and the synchronous control performance of multi-group propulsion cylinders are studied in this paper. Firstly, the mechanical structure, hydraulic system and working principle of the shield propulsion system are introduced. Then, the physical model of hydraulic system is established by using AMESim software, and the characteristics of propulsion speed and pressure response and multi-cylinder synchronization are analyzed by using conventional PID control. The simulation results show that the propulsion pressure has a great influence on the velocity. Under the control of conventional PID, the effect of speed control is not satisfactory. In order to solve the problem of mutual interference between propulsion pressure and velocity in shield propulsion system under variable load and variable flow conditions, the control strategy of BP neural network PID is put forward in this paper, and combined with Simulink software, the simulation is carried out. The simulation results show that the stability of BP neural network PID is good, which greatly reduces the overshoot of the pressure and speed of the cylinder during the shield propulsion, and the pressure and speed of the cylinder are overshoot when the load changes. In order to improve the performance of multi-cylinder synchronous control of shield cutter under unbalanced load, a single neuron PID control strategy is proposed. The response characteristics of propulsion system under different control strategies are compared and analyzed. It is concluded that using single neuron PID can greatly improve the synchronization of propulsion cylinder motion under partial load. Finally, the above control strategy is tested and verified on the shield experimental platform, and the experimental data are basically consistent with the simulation results. It is proved that the neural network combined with classical PID control is effective in suppressing the uncertainty and nonlinearity of propulsion system. The theoretical research and experimental analysis of this subject provide a theoretical reference for the design and optimization of shield control system.
【学位授予单位】：河南科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U455.39

【参考文献】