Research On Adaptive Control Method Of Small Deep Hole Drilling

Hole machining plays an irreplaceable role in mechanical manufacturing and processing technology as one of the fundamental machining processes.Micro-deep hole drilling is an essential branch of hole machining,and it is also the difficulty of hole processing technology.Especially,the drill needs to drill in a closed environment in the process of micro-deep hole processing.The drilling process is a complex dynamic deformation process involving fracture mechanics,thermodynamics,material mechanics,and many other theories.It is difficult to establish an accurate mathematical model of the process,which makes it difficult for the classical control method to control the complex and changeable drilling state.Therefore,the state inspection of the micro-deep hole drilling process is studied.The measurement method and adaptive control strategy have far-reaching significance.Cortex-A9 microprocessor and embedded Linux system are used to establish a drilling control system based on a high-speed steel twist drill through obtaining the state parameters,such as the drilling force and the drilling current during the drilling of micro-deep holes in this project.The adaptive control of the drilling process is realized based on a classical PID control algorithm,and the control algorithm is optimized by using a neural network algorithm.Firstly,the micro-deep hole drilling control platform was built,and the kernel and root file system of the Linux operating system was tailored and modified.At the same time,the related drivers were written to meet the functional requirements of the drilling control platform.Then use Qt to design a multi-window GUI control program and realize the manmachine information interaction of the drilling control platform through the touchable LCD,which ensures the simplicity and operability of the drilling control system.Furthermore,the drilling platform was used to collect sample data.The real-time drilling force and drilling current data were obtained through signal amplification,filtering,and discrete quantization.The expert PID control algorithm takes the axial drilling force as the optimization control objective,which is used to control the drilling feed speed to suppress the random fluctuation of the axial force in the processing process.The BP neural network algorithm is combined with the expert PID control algorithm to realize the online tuning of PID control parameters,to improve the universality of the drilling control system.Finally,the feature parameters are extracted to construct training samples by analyzing the signal characteristics of axial force,tangential force,and drilling driving current,and the drilling force identification model of deep hole drilling is established by PSO-BP neural network optimization algorithm,and the trained drilling force identification model parameters are transplanted to the Linux control system.In the drilling process,the real-time drilling driving parameters are used.Based on the dynamic current signal,the drilling force is identified online,and then the feed speed is adaptively controlled based on the identification value of drilling force,to realize the online optimization control in the process of micro-deep hole machining.Drilling experiment and test results show that the functional modules of the drilling control system can work typically;the expert PID control algorithm given in this thesis can keep the axial force in a small floating range and avoid the damage of the workpiece and drilling tool caused by the sudden change of the axial drilling force;combined with the online parameter setting of BP neural network,the universality of the drilling control system is improved;through the analysis of the simulation results,the simulation results show that the expert PID control algorithm can keep the axial force in a small floating range and avoid the damage of the workpiece and drilling tool caused by the sudden change of the axial drilling force The identification model of drilling force obtained by PSO-BP network algorithm has an excellent fitting effect with the real axial force and realizes the online adaptive control of minor deep hole drilling.