Key Technologies for Digital Transformation and Performance Optimization of Manufacturing Enterprises Based on Deep Learning

Abstract

With the rapid development of intelligent sensing, industrial wireless communication, and Electromagnetic Waves, Antennas and Propagation technologies, the efficient integration of heterogeneous manufacturing data has become essential for digital transformation and operational optimization. This study proposes a deep learning-driven technology framework for performance optimization in manufacturing enterprises by integrating multimodal data fusion, attention-enhanced long short-term memory (LSTM) networks, deep reinforcement learning, and edge deployment. Multi-source equipment signals, including vibration, current, and temperature data, are aligned and fused to construct representative health features for remaining useful life prediction and fault diagnosis. The predicted health states are further utilized to dynamically optimize maintenance scheduling through reinforcement learning, forming a closed-loop architecture from state perception to intelligent decision-making. Experimental validation demonstrates that the proposed framework achieves 96.8% fault prediction accuracy, improves Overall Equipment Effectiveness by 11.8%, and significantly reduces unplanned downtime while enabling low-latency edge inference for real-time industrial applications. The proposed methodology provides an effective solution for data-driven digital transformation and offers valuable engineering references for distributed sensing, industrial wireless monitoring, and intelligent communication systems associated with Electromagnetic Waves, Antennas and Propagation.