Thermal error modeling of slant bed CNC lathe spindle based on BiLSTM with data augmentation and grey wolf optimizer algorithm
The BL20 slant bed CNC lathe is widely recognized for its compact design and precision; however, it suffers from significant thermal errors due to heat generated during operation. This study analyzes and models the thermal errors of the BL20 lathe, identifying the heat produced by the main spindle a...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-06-01
|
| Series: | Case Studies in Thermal Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2214157X25003508 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The BL20 slant bed CNC lathe is widely recognized for its compact design and precision; however, it suffers from significant thermal errors due to heat generated during operation. This study analyzes and models the thermal errors of the BL20 lathe, identifying the heat produced by the main spindle as the primary source. Utilizing finite element simulation in ANSYS, the thermal characteristics of the lathe are examined, yielding insights into the dynamics of heat transfer. Experimental data are processed using fuzzy c-means clustering and grey relational analysis, which leads to the identification of four critical temperature-sensitive points. To mitigate the challenges associated with small datasets, the Synthetic Minority Over-Sampling Technique (SMOTE) is employed for data augmentation. Furthermore, a predictive model that integrates Bidirectional Long Short-Term Memory (BiLSTM) with hyperparameter optimization via the Grey Wolf Optimizer (GWO) is developed. In comparison to traditional methods, including Convolutional Neural Networks (CNN), Support Vector Machines (SVM), and standalone BiLSTM, the GWO-BiLSTM-SMOTE model demonstrates predictive accuracy, achieving R2 of 0.95384 and 0.95004 at various rotation speeds. The proposed approach showcases enhanced robustness and generalization. This study establishes a comprehensive framework for predicting thermal errors in CNC machine tools, offering valuable insights for precision machining. |
|---|---|
| ISSN: | 2214-157X |