Performance evaluation for tool wear prediction based on Bi-directional, Encoder–Decoder and Hybrid Long Short-Term Memory models

2022 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Satish Kumar ◽  
Tushar Kolekar ◽  
Ketan Kotecha ◽  
Shruti Patil ◽  
Arunkumar Bongale
Keyword(s):  
Feature Selection ◽  
Tool Wear ◽  
Hybrid Model ◽  
Short Term Memory ◽  
Short Term ◽  
Wear Prediction ◽  
Term Memory ◽  
Content Type ◽  
Tool Wear Prediction ◽  
Long Short Term Memory

Purpose Excessive tool wear is responsible for damage or breakage of the tool, workpiece, or machining center. Thus, it is crucial to examine tool conditions during the machining process to improve its useful functional life and the surface quality of the final product. AI-based tool wear prediction techniques have proven to be effective in estimating the Remaining Useful Life (RUL) of the cutting tool. However, the model prediction needs improvement in terms of accuracy.Design/methodology/approachThis paper represents a methodology of fusing a feature selection technique along with state-of-the-art deep learning models. The authors have used NASA milling data sets along with vibration signals for tool wear prediction and performance analysis in 15 different fault scenarios. Multiple steps are used for the feature selection and ranking. Different Long Short-Term Memory (LSTM) approaches are used to improve the overall prediction accuracy of the model for tool wear prediction. LSTM models' performance is evaluated using R-square, Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) parameters.FindingsThe R-square accuracy of the hybrid model is consistently high and has low MAE, MAPE and RMSE values. The average R-square score values for LSTM, Bidirection, Encoder–Decoder and Hybrid LSTM are 80.43, 84.74, 94.20 and 97.85%, respectively, and corresponding average MAPE values are 23.46, 22.200, 9.5739 and 6.2124%. The hybrid model shows high accuracy as compared to the remaining LSTM models.Originality/value The low variance, Spearman Correlation Coefficient and Random Forest Regression methods are used to select the most significant feature vectors for training the miscellaneous LSTM model versions and highlight the best approach. The selected features pass to different LSTM models like Bidirectional, Encoder–Decoder and Hybrid LSTM for tool wear prediction. The Hybrid LSTM approach shows a significant improvement in tool wear prediction.

Tool wear prediction in high-speed turning of a steel alloy using long short-term memory modelling

Measurement ◽  
2021 ◽  
Vol 177 ◽  
pp. 109329
Author(s):  
Mohsen Marani ◽  
Mohammadjavad Zeinali ◽  
Victor Songmene ◽  
Chris K. Mechefske
Keyword(s):  
Tool Wear ◽  
High Speed ◽  
Short Term Memory ◽  
Steel Alloy ◽  
Short Term ◽  
Term Memory ◽  
High Speed Turning ◽  
Tool Wear Prediction ◽  
Long Short Term Memory ◽  
Memory Modelling

Short-term prediction of remaining life for lithium-ion battery based on adaptive hybrid model with Long Short-Term Memory neural network and optimized particle filter

2021 ◽  
pp. 1-36
Author(s):  
Ning He ◽  
Cheng Qian ◽  
Lile He
Keyword(s):  
Neural Network ◽  
Hybrid Model ◽  
Life Prediction ◽  
Short Term Memory ◽  
Lithium Ion ◽  
Remaining Useful Life ◽  
Short Term ◽  
Term Memory ◽  
Useful Life ◽  
Long Short Term Memory

Abstract As an important energy storage device, lithium-ion batteries have vast applications in daily production and life. Therefore, the remaining useful life prediction of such batteries is of great significance, which can maintain the efficacy and reliability of the system powered by lithium-ion batteries. For predicting remaining useful life of lithium-ion batteries accurately, an adaptive hybrid battery model and an improved particle filter are developed. Firstly, the adaptive hybrid model is constructed, which is a combination of empirical model and long-short term memory neural network model such that it could characterize battery capacity degradation trend more effectively. In addition, the adaptive adjustment of the parameters for hybrid model is realized via optimization technique. Then, the beetle antennae search based particle filter is applied to update the battery states offline constructed by the proposed adaptive hybrid model, which can improve the estimation accuracy. Finally, remaining useful life short-term prediction is realized online based on long short-term memory neural network rolling prediction combined historical capacity with online measurements and latest offline states and model parameters. The battery data set published by NASA is used to verify the effectiveness of proposed strategy. The experimental results indicate that the proposed adaptive hybrid model can well represent the battery degradation characteristics, and have a higher accuracy compared with other models. The short-term remaining useful life prediction results have good performance with the errors of 1 cycle, 3 cycles, and 1 cycle, above results indicate proposed scheme has a good performance on short-term remaining useful life prediction.

Ensemble-Based Feature Selection With Long Short-Term Memory for Classification of Network Intrusion

2021 ◽  
pp. 228-245
Author(s):  
Preethi D. ◽  
Neelu Khare
Keyword(s):  
Feature Selection ◽  
Performance Metrics ◽  
Short Term Memory ◽  
Short Term ◽  
Chi Square ◽  
Term Memory ◽  
Network Intrusion ◽  
Proposed Model ◽  
Long Short Term Memory

This chapter presents an ensemble-based feature selection with long short-term memory (LSTM) model. A deep recurrent learning model is proposed for classifying network intrusion. This model uses ensemble-based feature selection (EFS) for selecting the appropriate features from the dataset and long short-term memory for the classification of network intrusions. The EFS combines five feature selection techniques, namely information gain, gain ratio, chi-square, correlation-based feature selection, and symmetric uncertainty-based feature selection. The experiments were conducted using the standard benchmark NSL-KDD dataset and implemented using tensor flow and python. The proposed model is evaluated using the classification performance metrics and also compared with all the 41 features without any feature selection as well as with each individual feature selection technique and classified using LSTM. The performance study showed that the proposed model performs better, with 99.8% accuracy, with a higher detection and lower false alarm rates.

Sign in / Sign up

Export Citation Format

Share Document