An Investigation of Machine Learning Algorithms for Predictive Maintenance in High Pressure Processing Systems

In this research proposal to enhance the operational efficiency and prolong the usable life of industrial machinery and equipment through predictive maintenance leveraging machine-learning techniques. As manufacturers increasingly prioritize cost reduction, minimized downtime, and enhanced operational uptime, the adoption of proactive maintenance strategies becomes imperative. This study intends to gather historical datasets to train machine-learning models for predicting equipment failures and develop an algorithmic framework for proactive maintenance scheduling. The primary objective is to contribute to the development of an efficient predictive maintenance model, thereby reducing industrial maintenance costs and positively impacting product costs. This research will employ various machine-learning approaches, big data preprocessing techniques, and feature engineering methodologies. Data preprocessing will involve cleaning, conversion, and standardization of datasets before model training. Feature engineering will focus on selecting the most relevant features for accurate machine failure prediction. Multiple machine-learning algorithms, including Random Forest (RF), Support Vector Machines (SVM), and Long Short-Term Memory (LSTM), will be evaluated to determine the most effective model for precise predictions. The comparative predictive performance utilizing by Root Mean Square Error (RMSE), R-squared (R2), Mean Absolute Error (MAE) to measure performance metric. The best performing machine learning models in this study have been deployed into real operation in factory. The best model expect to archive successful result by 5-10% increase operation efficiency.