| In metal cutting,the cutting temperature directly affects the surface integrity,machining accuracy and tool life.However,due to factors such as time-varying local high temperature,narrow area and large temperature gradient in the tool-chip contact area,as well as strong interferences such as cutting fluid,oil and chips,in-situ online measurement of cutting temperature is still challenging.Fiber optic sensors have the advantages of small size,high temperature resistance,anti-electromagnetic interference,and high chemical stability.The in-situ online temperature measurement of cutting tools under harsh and complex working conditions can be realized by integrating fiber-optic sensors in cutting tools.In order to obtain the spatiotemporal temperature distribution of the cutting tool,this paper proposes in-situ online measurement methods of cutting zone temperature based on near-infrared fiber-optic sensing.Based on the measured cutting temperatures,the studies of temperature field reconstruction and tool wear status monitoring are carried out.The specific research contents are as follows:1.For the problems of existing cutting temperature measurement methods in stability under harsh conditions,fluid/chip interference,and in-situ measurement,an improved fiberoptic two-color temperature measurement method is proposed.The model considers the influence of filter bandwidth,and introduces a slope compensation term based on the spectral emissivity model.The theoretical calculation error of the improved model under the selected hardware parameters is only 48.12% of that of the traditional model.Based on the improved model,a near-infrared optical fiber two-color cutting temperature measurement system is developed.The calibration shows that the relative error of the system measurement is less than 1.5% in the commonly used temperature range of 300-700 °C.The stainless-steel cutting experiments are carried out.In-situ online measurement of tool temperatures in dry/wet cuttings and chip temperatures in dry cuttings are realized using the self-developed system.Continuous cuttings verify the positive correlation between cutting parameters and tool temperature,and the intermittent cutting shows that the response time is about 10 ms.This work improves the accuracy of fiber-optic two-color cutting temperature measurement,and provides a reference for in-situ online measurement of tool temperature under harsh conditions and strong interferences.2.For the problem of poor accuracy of the two-color method in measuring temperatures below 300 °C,a multi-spectral cutting temperature measurement method based on near-infrared optical fiber is proposed.Using thermal radiation spectrum for temperature measurement,the method optimizes the lower limit of temperature measurement to 150 °C while improving accuracy.The calibration shows that in the range of 150-200 °C,the average relative error of measurement is less than 2%;in the range of200-250 °C,the average relative error of measurement is less than 1%;in the range above250 °C,the relative measurement error is stable below 0.5%.The titanium alloy cutting experiments are carried out.In-situ online measurement of tool temperatures in dry/wet cuttings and chip temperatures in dry cuttings are realized using the self-developed system.The influence of cutting parameters on cutting temperature is studied,and the real-time response of the temperature measurement system to the cutting state is verified.This work broadens the cutting temperature measurement range with better measurement accuracy,and lays the groundwork for accurate measurement of cutting temperature.3.For the problem of existing inverse heat conduction problem solving algorithms in time lag,poor stability,and low computational efficiency,an online estimation method of cutting heat flux using long short-term memory based encoder-decoder is proposed.Under the premise of using no future information,the method establishes the mapping relationship between the temperature series and the heat flux series,and realizes end-to-end learning.The measured single-step calculation time of the model is only about 30 ms,which meets the requirements of online calculation.Compared with other classic models,the proposed model has the smallest relative root mean square errors at all tested noise levels,exhibiting excellent robustness and noise immunity.Using the fiber-optic multi-spectral temperature measurement information,the temperature field reconstruction of the titanium alloy cutting process is realized.The reconstructed temperature field is cross-validated with the thermocouple temperature measurement results,which proves the accuracy of the thermal model,the inverse heat conduction problem solving algorithm,and the temperature field reconstruction results.This work provides a new idea for on-line reconstruction of the tool temperature field with no time delay,nonlinear,and complex three-dimensional structures.4.Using the developed near-infrared fiber-optic cutting temperature measurement methods,the research on tool wear monitoring is carried out.The multi-spectral temperature measurement system is used to measure tool temperatures in the tool wear process of titanium alloy cuttings.The system shows excellent online monitoring capability of the tool status.Taking the original signals of temperature measurements as inputs,a machine learning model combining sparse autoencoder and k-means clustering is used to identify tool wear status.The model shows good online prediction performance and a classification prediction accuracy of 97.3%.The two-color temperature measurement system is used to measure tool temperatures in the tool wear process of stainless steel cuttings,proving the online monitoring capability of the system.Taking the temperature measurements as inputs,the prediction of tool flank wear is realized by a model based on short-time Fourier transform and the convolutional neural network.This work proves the application potential of the proposed tool temperature measurement methods,and opens up a new way for tool wear condition monitoring. |
PDF Full Text Request