Journal of Wuhan University of Hydraulic and Electric Engineering, article number 1006, Sliding Mode Variable Structure Control of Hydraulic Servo System Based on Neuron Adaptive Adjustment Yang Yongwen Lian Hongwen Wen 21. Changsha Electric Power College, Changsha 410077, China; 2 School of Electrical and Information Engineering, Wuhan University of Hydraulic and Electric Engineering. Wuhan Wuhan 430072, China 1 Based on the learning characteristics of sliding model points and neurons, a sliding mode variable structure control strategy based on adaptive adjustment of neurons is proposed. And it is applied to the position control of the hydraulic nonlinear servo system. The simulation results are clear. The proposed control strategy is feasible, and it has a positive effect on improving the dynamic characteristics and robustness of the sliding mode variable structure control system and weakening the sliding mode control chattering phenomenon. It is an effective control method.
The hydraulic servo system is a typical nonlinear system, which is widely used in industrial fields such as modern aerospace power metallurgy robots. In practical applications, it is required to have the characteristics of fast response control, high anti-interference ability and strong performance.
Sisheng. Because the kinetic mode is completely invariant to the internal parameter changes and external perturbations of the system, the system is fully adaptive, which makes it solve the problem such as automatic adjustment of motion tracking model tracking adaptive control and uncertain system control. The aspect has achieved good results. Its basic principle is to change the system of the switching surface in the system state space, the structure on both sides, according to the fixed control rules to ensure that the system state moves along the sliding mode. The modal modal sub-index is optimal to guarantee. However, in the sliding mode variable structure control, there is a jitter phenomenon due to frequent switching of the control action, and the dynamic quality of the control system and the selection of the switching function and the control action parameter are flawed.
Recently, the neural network has attracted extensive attention from the control community. 1 The theory of neural network theory shows that artificial neural networks have essential nonlinear and parallel structures; multi-layer neural networks can approximate arbitrary continuous and discontinuous functions. Modern control research is advancing the hybrid technology of the fading network technology. Expert system control and other, system technology; the body's intelligent control technology development direction 34, which reflects the intelligent control in solving the non-line. male. lecturer. From the automatic control surface of the mortar. Ut+Il system control has a bright future. Aiming at the characteristics of computer-controlled hydraulic servo system combined with sliding mode and neural network technology, this paper proposes a sliding mode variable structure control method based on adaptive adjustment of neurons. It is clear that the adaptive control of this variable structure control system is good. The rod is strong, and the system response jitter is effectively suppressed. And the dynamic performance is full.
1 station ten neuron, adaptive sliding mode variable structure controller design 1.1 sliding mode variable structure control Consider the following linear system. This sliding switching surface function is to use the pole configuration to determine the switching function matrix Zhuang sliding mode switching in equation (2) Partial phase change proportional control strategy is adopted in the neighborhood of the face, the equivalent sliding mode control is taken as the sum, and the selection is constant.
Since the sliding mode is present and extended, and the time constant is neglected, the variable structure control has a jitter phenomenon. In order to avoid too frequent switching of the die-casting technology, a thin layer region can be introduced near the sliding mode switching surface to improve the approach to the sliding surface condition, thereby alleviating the discontinuity of the sliding mode control. Therefore, the improved sliding surface condition is improved.è…½ structure, system, ten 570 is the approach rate constant.
1.2 Design of single neuron adaptive regulator The sliding mode variable structure control is based on the linearization of the system model. Although the influence of some nonlinear factors is considered, it is difficult to accurately define the real model of the actual system. After establishing the sliding mode variable structure control, an adaptive adjustment mechanism for introducing the process may be adopted. Since neurons have self-learning and fault-tolerant properties, a neuron adaptive adjustment mechanism is constructed for this purpose.
Neurons should be regulated by a single neuron. The difference between the output selection output and the neuron network regulator output is to improve the training process. The data can be standardized and buried. For neuron weights, online adaptive correction is used. Let the performance index number be the gradient of the sakis as it is available, approximate (1) the symbol function å² to replace, and then pay the 8 ç–” standardization, when asked according to the system output and the reference input value change, select the proportional regulator parameter The range selection should be appropriate, and the dynamic adjustment of the escape on the output of the controller will help speed up the system response. Therefore, the composition of the variable-mode control system based on neuron adaptive adjustment is 2.
2 simulation experiment results and discussion 1 + moxibustion 2, the output is the control signal adjustment amount, which is the neuron gain.
In order to improve the initial response characteristics of the servo system, the initial parameters of the adaptive regulator neurons are trained offline. In order to verify the effectiveness of the proposed control method, the off-line training is performed on the computer-controlled hydraulic system; Simulation experiment results 3 and 4.
When the reference input is 30507090 and the system load mass is =106.5, the hydraulic servo system pure sliding mode variable structure control step response curve 3 does not. It can be seen from the cabinet 3 that the pure sliding mode variable structure control response speed is fast, the control precision is high, the system loose type precision injection requirement is high, the parameter selection dependence of the sliding mode variable structure screwing is strong, and it is necessary to take measures to solve Good sliding mode variable structure control jitter problem.
Under the same experimental conditions as 3, 4 is a sliding mode variable structure control step response curve of hydraulic servo system based on neuron adaptive adjustment. The proposed sliding mode variable structure control strategy based on neuron adaptive adjustment based on neuron adaptive adjustment effectively absorbs the characteristics of sliding mode variable structure control and neuron technology, and compensates for the sliding mode variable structure control parameters. And the system control effect + ideal deficiency. The system responds to the jitter phenomenon to effectively suppress, and the dynamic performance of the system is good. The system is not sensitive to the parameter change. The adaptive robustness of the system is good. The self-learning adaptive information synthesis of the neuron as a modulo variable structure control adaptive regulator The ability and tolerance are effectively utilized here to better compensate for the main defects of the sliding mode variable structure control, and the control results are full.
The control step response curve 3 concludes the variable structure control strategy. For the sliding model, effective measures are taken to weaken the buffeting phenomenon caused by the frequent switching of the control function on the switching plane. It is applied to the simulation of hydraulic servo system. The results show that the proposed control strategy effectively utilizes the learning characteristics of neurons and enhances and improves the performance of sliding mode variable structure control. The dynamic characteristics and robustness of the control system are significantly improved, the jitter phenomenon is effectively weakened, and the method design is simple to control the step response curve 1 to be high. Variable Structure Control 1 Theory and Design Method 1. North North Science Press. 1996.
2 Hu Shouren. Neural network application technology. Changsha National University of Defense Technology Press. 1995.
3 èµ€ é” sharp and other internal energy control methods of the text fork six and its should look for 4 Shu Di before. Single Neuron Adaptive with Sub-Performance Indicators, Controller and Its Applications 1. Electrical Automation, 1997, 19147.
5 Yang Yong, Huang Wenmei. A new type of electro-hydraulic servo system neural network adaptive control. Journal of Hunan University, 1998, 2555559.
Carbon Steel Boiler Tube,Alloy P22 Steel Pipe,Astm Carbon Steel Boiler Tube
YUHONG GROUP CO.,LIMITED , https://www.jessro-pipefittings.com