Tool Wear Prediction For High-precision Machining Process Of Single-piece Small Batch Injection Mold

The high-precision machining process,as the last machining process of the workpiece forming,directly determines the machining quality of the workpiece.Therefore,monitoring the tool wear in the high-precision machining process is an effective measure to improve the machining accuracy and improve the tool utilization rate.This paper combines the high-precision machining process of single-piece small-volume injection molds to carry out research on tool wear monitoring methods for flat-end end mills in high-precision machining with fixed parameters and variable parameters.Combined with the characteristics of single-piece small batch,small tool size,small tool wear variation,variable mold structure and high product precision requirements during the mold processing,the research in this paper mainly involves the following aspects:Firstly,the processing process,processing characteristics and quality change characteristics of single-piece small batch injection molds are analyzed.Through the comparative analysis of mold quality change trend range and tool wear value,the wear value of injection molds during high-precision machining is quantified.method.On this basis,combined with the analysis of the problems in the process of factory mold processing,the data acquisition scheme of tool wear detection in the high-precision machining process of injection molds was completed.The noise reduction method and effective feature extraction and feature selection methods for injection molding high precision machining signals are determined.Combined with the extremely high cutting process of the injection molding die and the noise of the machining process,the processing method based on wavelet threshold denoising is determined.The wavelet decomposition and time-frequency statistical features are combined to extract the processing signal characteristics.Feature selection based on grayscale analysis method is completed.The tool wear prediction modeling method based on BP and LSTM neural network algorithms is studied.The generalization of the two modeling methods is studied under the condition of changing the cutting allowance and mold draft angle in high-precision machining.And the comparison of the two methods in modeling accuracy and model generalization is completed.Finally,through the analysis of the mold processing data collected at different times,the stability of the tool wear prediction at different time is verified.In addition,by applying the model to the actual mold production and processing of the factory,it is found that the actual prediction error of the model is maintained at 1-2 ?m.The effect is good.