Evolving Connectionist System and Hidden Semi-Markov Model for Learning-Based Tool Wear Monitoring and Remaining Useful Life Prediction

Abstract

Tool wear can cause dimensional accuracy and poor surface quality in milling process. During the operation of tool wear, it can also cause breakage and damage of the workpieces. To prevent these conditions, it's important that the tool wear is monitored and the remaining useful life (RUL) is predicted in real time. In this paper, time domain and frequency domain statistical features are firstly extracted using multi-sensory fusion method, including the cutting force, vibration and acoustic emission sensor. Seven eigenvectors are selected as the input of the prediction model based on the distance correlation coefficient between 140 feature vectors and the wear value, which provide the most sensitive features to wear faults. The paper establishes a nonlinear relationship between high-dimensional feature vectors and tools wear based on the evolving connectionist system (ECoS), which uses the incremental learning algorithm to realize real-time prediction of the tools wear. Finally, using the wear value predicted by ECoS as hidden state sequence of Hidden Semi-Markov Model (HSMM), the RUL prediction of the tool based on HMM is established. The 2010 PHM challenge data were used to train the model. The experimental result shows that in comparison with artificial neural network, the ECoS model has higher prediction accuracy, and its mean RMSE error for three tools is 14.8. In comparison with the RUL prediction of HMM model, Probability-based RUL prediction of HSMM is more stable. INDEX TERMSMulti-sensor fusion; Evolving connectionist system; Incremental learning algorithm; HSMM; Remaining useful life prediction. Abbreviations RUL remaining useful life HMM hidden markov model ECoS Evolving connectionist system PHM prognostic and health management HSMM hidden semi-markov model RMSE Root mean square error

Tool wear prediction based on hybrid feature selection

“…Some researchers obtained better performance in wear measurements using images and Convolutional Neural Network (CNN). According to the tool wear measurement algorithm [1], a tool wear monitor system using only a single camera was developed based on images captured from the machining process through an image treatment block, and image comparison block. Commonly, the CNN combined transfer learning method was used to classify the wear state of milling tools [2].…”

“…Because of the complexity of the machining process and the difficulty of image acquisition, a tool state classification method was proposed based on the modified AlexNet network using cutting force images converted by the Gramina Angular Summation Fields method in which an experiment carried out on a CNC lathe showed good prediction accuracy [3]. Multi-domain statistical features were extracted, and the Hidden Semi-Markov Model (HSMM) was used to predict Remain Use Life (RUL) which proved higher performance than artificial neural networks [1]. Ensemble learning methods were commonly used when predicting tool wear using machine learning.…”

Hidden Semi-Markov Models-Based Visual Perceptual State Recognition for Pilots

“…In current research applications, most focus on fault diagnosis [ 22 ], risk perception [ 23 ], intent recognition [ 24 ], wear monitoring, and remaining life prediction [ 25 ]. This paper provides a solution to the task ambiguity arising from overlapping instruments, mainly using the HSMM.…”

Scalability, Explainability and Performance of Data-Driven Algorithms in Predicting the Remaining Useful Life: A Comprehensive Review

“…Presenting observations as a continuous mixture of Gaussians means that any continuous density function can be approximated [240]. As such, the generalized HSMM of [92,241] has the potential to be applied in many different fields. While multi-sensor data can be transformed (e.g., using the PCA) to be used in univariate HSMM models [152], the HSMMs are more complex and consequently more computationally intensive [76] than HMMs.…”