| Title |
A Study on Power-Based Modeling for Predicting Nonlinear Compressed Air Flow Patterns |
| Authors |
임채영(Chae-Young Lim) ; 여채은(Chae-Eun Yeo) ; 김경호(Kyung-Ho Kim) |
| DOI |
https://doi.org/10.5370/KIEE.2025.74.3.500 |
| Keywords |
Smart Factory; Air Compressor; Flow pattern prediction; Deep Learning; Convolutional neural network; LSTM; Mixed Model |
| Abstract |
In smart factory manufacturing processes, irregular and uncertain environmental factors can lead to unexpected equipment failures or abnormal operations, resulting in product defects, safety incidents, and energy waste, which ultimately reduce productivity. To prevent these issues and ensure stable equipment operation, it is essential to accurately predict not only equipment failures but also energy consumption. Maximizing energy efficiency requires a predictive model that comprehensively considers both equipment status and energy usage patterns. This study proposes a machine learning-based flow pattern prediction model using power data from smart factory manufacturing processes. The model utilizes Multi-Layer Perceptron (MLP), Convolutional Neural Network (CNN), and Long Short-Term Memory (LSTM) algorithms to predict equipment power consumption and flow patterns. By applying these prediction models to a digital twin framework, the study demonstrates the potential to optimize energy consumption. |