Study On Recognition And Prediction Of Unsteady Characteristics In Micro-grinding Of Monocrystalline Silicon Based On Acoustic Emission

In recent years,the micro-components represented by silicon-based hard and brittle materials have a wide application prospect in the cutting-edge industries such as photovoltaic new energy,chip semiconductor,electronic communication and so on.Micro-grinding with diamond microabrasions is an important method for precision machining of micro-groove,micro-hole and micro-channel of such parts.Due to the hard and brittleness of monocrystalline silicon,it is very easy to cause edge damage and other machined damage as result in poor machining quality in micro-structure micro-grinding.Moreover,the serious wear of micro-abrasive tools also affect the machining quality and machining accuracy.This paper aims to explore the dynamic recognition ability for unsteady state characteristics based on acoustic emission(AE)technology in hard-brittle material micro-grinding process.In the thesis,an online AE monitoring system is established,and then the AE singal analysis method is discussed.In order to achieve efficient and high-precision machining of monocrystalline silicon microstructure,a prediction model is futher built and verified for unsteady state characteristics in micro-grinding by AE signals based on experimental results.The specific work of the paper is as follows:(1)The source of AE in the micro-grinding process of the microstructure of hard and brittle materials was analyzed,and a micro-groove array acoustic emission testing platform for micro-grinding was built.Based simulation analysis results the optimum overhang length of the micro-abrasive tool was selected.The grinding process parameters was confirmed through the pre-experiment.Taking the depth of micro-groove as a variable,micro-grinding process experiments were designed and AE signals were collected.(2)Wear of micro-abrasive tools,edge damage of micro-groove and cross-section of micro-groove were selected as evaluation parameters of unsteady characteristics in micro-grinding process of monocrystal silicon.Firstly,the relationship between micro-grinding tool's tip diameter loss and grinding length was analyzed.Micro-grinding tool radial wear evaluation method was established.The tool radial wear in all sampling interval was calculated.And then the mechanism of edge damage was analyzed,the edge damage evaluation method was established.Based on microscopic observation and image processing technology,the taper angle of the microgroove cross section was established and calculated.(3)The AE signal data processing scheme was developed to indentify unsteady state characteristics through time-domain and frequency-domain analysis.The time-domain characteristics of AE signals were analyzed by the root mean square value method,and the sampling interval AE_RMS was calculated.Fast Fourier Transform Algorithm(FFT)method was used to analyze the frequency domain characteristics of the sampling interval,and frequency domain bands are divided according to the signal characteristics.The wavelet packet decomposition method is used to decompose the original signal.The energy and energy ratio of each frequency band are calculated.There is a positive correlation between the tool radial wear and the energy proportion in the middle frequency band.Combined with the energy proportion in the frequency band of the AE signal,the time period can be predicted when the chip accumulation is more serious.The data comparison results showed that the average edge damage width of microgroove had the similar variation trend as AE_RMS.The characteristics of edge damage in frequency domain were analyzed.It was found that there was a strong correlation between the average edge damage width of microgroove and the proportion of energy in high frequency band.The characteristics of the microgroove taper angle in the frequency domain show that the energy of the low frequency band is smaller when the microgroove taper angle is larger.(4)A radial basis function(RBF)neural network model was established to predict the tool radial wear,average edge damage width and taper angle of micro-grooves with micro-grooves depth,grinding length,root mean square value of AE signal and frequency band energy ratio as input values.The results show that this model has a good prediction effect on the three factors.