Extraction Of Dynamic Cutting Force Feature And Predition Of Surface Integrity In Hard Turning

Hardened steel is a typically hard processing material, its hard turning by adopting PCBN tool can improve machining efficiency, drop energy consumption, reduce investment and decrease pollution, moreover, obtain surface quality that is equivalment with grinding and even exceeds grinding. Therefore, hard turning technology has broad application prospect. However, for this new technology, aspects of dynamic cutting force feature extraction and surface integrity predition in hard turning are lack of systematic study, thus its extension in actual production is seriously impeded.In this paper, hard turning GCr15 hardened bearing steel by PCBN tool is regarded as research object. Cutting force in hard turning is measured accurately by using cutting force data acquisition and analysis system based on LabVIEW. Dynamic feature of cutting force signal in hard turning is extracted effectively. Comprehensive predition of surface integrity based on BP artificial neural network is realized. And thus, certain theoretical basis and technical support are provided for extension and application of hard turning technology.Cutting force data acquisition and analysis system is developed by LabVIEW software platform. This system is able to make data acquisition of cutting force signal on forms of dynamic waveform and real-time data, and also the collected cutting force original signal is conducted by data playback, data interception, data filtering and data statistical analysis.The cutting force data acquisition and analysis system is applied to the research of hard turning experiment, analysis acquired cutting force change law under different condition and hard turning force of traditional theory comparatively, and further validate reliability and feasibility of this system.According to limitations of time domain analysis and frequency domain analysis of signal, adopt wavelet transform method to analysis hard turning force signal comprehensively in time-frequency domain. Apply Matlab software to compile program and achieve autocorrelation analysis, self-power spectrum analysis, wavelet decomposition and wavelet de-noising of hard turning force, so it is capable of understanding the feature of hard turning force deeply, and can also extract the feature quantity relevant to aspects of tool vibration and tool breakage effectively.Using BP artificial neural network method, combined with the data obtained by hard turning experiment, taking tool geometric parameters and cutting parameters as input, taking eigenvalue of residual stress and white layer as output, predition models of residual stress and white layer in hard turning are established. Based on predition models that have been constructed, comprehensive predition system of surface integrity in hard turning is developed by utilizing Matlab software to compile program, predition of residual stress and white layer in hard turning are implemented.