Intelligent Monitoring of Thermodynamic Parameters in Compressor Operations and Development of a Fault Prediction Model Using Deep Learning

Abstract

Compressors, as essential industrial equipment, are widely utilized in air conditioning, refrigeration, energy, and chemical sectors. The operational stability of compressors directly impacts system efficiency and safety. Due to the complex thermodynamic processes and variable operating conditions involved in compressor operations, accurately monitoring their status and predicting potential faults are crucial for improving equipment reliability and optimizing maintenance strategies. Traditional compressor fault prediction methods usually rely on thermodynamic models and statistical analysis. While some progress has been made, existing methods often face limitations when addressing complex nonlinear, multivariable, and time-varying characteristics. Recently, machine learning and deep learning-based fault prediction methods have gained significant attention, but challenges remain in real-world applications, including data quality, model accuracy, and computational efficiency. To address these issues, this paper proposes an intelligent monitoring and fault prediction approach for compressors based on deep learning. First, a thermodynamic model of the compressor operation process is constructed, enabling real-time acquisition of key operational parameters. Subsequently, a fault prediction model is developed by integrating the sparrow search algorithm (SSA) with the long short-term memory (LSTM) model. By performing time series analysis on compressor operational data, the model achieves accurate fault prediction and early warning. This approach effectively enhances prediction accuracy and robustness, offering strong practical application value.