Research On Spindle Hydraulic Loading Control System Based On Neural Network Adaptive Robust Control

The mechanical spindle is the most core functional component of CNC machine tools,and its performance directly affects the reliability level of the whole machine.In order to provide more data support for spindle reliability research,it is necessary to stimulate the spindle prone to failures under working conditions more quickly.Under the condition of saving cost and not changing the failure mode and failure mechanism,the bench reliability acceleration test is usually chosen,while the reliability test for mechanical spindle needs to simulate its real working condition.According to the characteristics of mechanical spindles often work in the low speed and heavy load area,the hydraulic loading system is chosen as the driving equipment to simulate the spindle working conditions.However,because the hydraulic system is more complex and difficult to model,the traditional control methods have been difficult to meet the test requirements,and most of the current advanced control strategies are only at the simulation level.Therefore,in this paper,a mechanical spindle reliability test bench is built and the development of a hydraulic loading control system is realized based on neural network adaptive robust control algorithm.This paper takes JSX150 BA model mechanical spindle as the research object,builds a bench test system to meet the requirements of spindle reliability test,establishes a hydraulic cylinder simulation loading part displacement control and force control model,designs a hydraulic cylinder radial basis neural network adaptive robust controller,and develops the corresponding lower computer control and upper computer operation platform.The main contents of the paper are as follows.(1)A test stand was built to meet the requirements of spindle reliability testing.Firstly,we analyzed the force situation of mechanical spindle under load conditions,and determined the range of loading force amplitude and loading frequency of the test bench.Secondly,the hydraulic loading method was selected to simulate the spindle force conditions through scheme comparison,and based on this,the spindle reliability test bench was built.(2)Hydraulic cylinder displacement control and force control models were established.First,the mathematical model of the electro-hydraulic servo valve,hydraulic cylinder and other links of the hydraulic control system is established.Then,the transfer function of hydraulic cylinder displacement control is established according to the feedback relationship,and the equation of state of hydraulic cylinder force control is established in order to meet the design needs of subsequent force control.(3)A RBF neural network adaptive robust controller is designed.First,for the two main problems of model uncertainty in the modeling process(i.e.,unmodeled dynamics and parameter uncertainty),a neural network controller is introduced to achieve compensation for the unmodeled dynamic part of the hydraulic system,and the radial basis neural network robust control is selected to control the parameter uncertainty of the hydraulic system after comparative analysis.After that,the controller is designed based on the Liapunov stability theory and its stability is proved.Finally,the performance of the controller controlling the hydraulic system is verified using Simulink simulation method.(4)The lower computer and upper computer control software of the hydraulic loading system were developed.Firstly,Lab VIEW software and c RIO hardware controller were selected for control system development.c RIO has FPGA chip with parallel computing capability,which can process complex algorithms quickly and provide assurance for the realization of control algorithms.The developed humanmachine interface provides a simple and clear control platform for users.In this paper,the mechanical spindle reliability test bench and its control system are firstly built.To solve the parameter uncertainty and unmodeled dynamic problems of the control system,neural network adaptive control and robust control are applied to the hydraulic loading control system to improve the control accuracy of the hydraulic loading force.