Research On Status Monitoring And Intelligent Control Of Micro-deep Hole Drilling Process

Hole drilling is an important branch of machinery manufacture,and deep hole drilling is a most important and difficult issue in the hole drilling.Since deep hole drilling bits work in closed or semi-closed,high temperature and high pressure conditions,which leads to difficulties in lubrication and chip removal for the drilling process,working life and efficiency of the drilling bit are all reduced seriously,and the deep hole drilling process is difficult to be monitored in real time as well as its accurate mathematical model is also difficult to be established,which results in the failure of classical and modern control theories to achieve optimal control.Aiming at the difficulties in the control of deep hole drilling process,this paper uses emerging neural network and classical measurement methods to establish a non-linear mapping model between driving drilling current and drilling force,achieving the working condition monitoring and adaptive control of the deep hole drilling process.Firstly,a platform of drilling force measurement was built to research the regular variation of drilling force during micro-deep hole drilling.The platform converts drilling force signals into analog signals with resistance strain gauge sensor,and then converts the analog signals into digital signals with signal processing technologies such as analog filtering,analog-to-digital conversion,and digital filtering,to achieve online measurement of the thrust force,tangential force and radial force for researching the regular variation of deep hole drilling force and sample labeling of neural networks.Secondly,based on the BP neural network of single hidden layer,a nonlinear mapping model was built,which maps the driving current into drilling force during deep hole drilling.Driving current is non-contact detected by Hall sensor,and then is collected online and digitally quantified for the input sample of the neural network.Synchronously,drilling force is detected by the platform of drilling force measurement for the output sample of the neural network.And the non-linear mapping relationship between driving current and drilling force is obtained through the supervised learning of the neural network.Finally,an adaptive control method for deep hole drilling was proposed.Driving current is recognized as drilling force in real time by trained BP neural network model,and the recognized drilling force is sent to the adaptive control module,and then the adaptive control module judges the status of the drill bit immediately and outputs the corresponding drilling control parameters to determine the subsequent actions of the driven system,including changing the feed mode,adjusting the feed speed and spindle speed of drill bit,etc.As a result,the whole system achieves the optimal control of drilling force and the adaptive closed-loop control of deep hole drilling process.Experimental simulation and test results show that the platform of drilling force measurement with independent intellectual property rights developed by this paper accurately detects the four-dimensional drilling force of micro-deep hole drilling in real time,and the measurement indicators meet the actual engineering needs;the model of recognizing drilling force constructed by neural network maps driving current into drilling force nonlinearly,and the matching degree between the recognized drilling force curve and the actual drilling force curve meets the control requirements;the adaptive control algorithm of drilling process has a favorable effect,can quickly adjust the drilling feed parameters,suppress the drilling force fluctuations during the deep hole drilling process,and protect the deep hole drill bit effectively and ensure the deep hole drilling efficiency.