An innovative method for predictive maintenance of cold stores in shrimp processing industry using machine learning and data trends

Abstract

This study presents a novel approach to predictive maintenance in cold storage systems by analyzing temperature trends and detecting anomalies. Utilizing threshold-based outlier detection and linear regression for trend prediction, the system forecasts maintenance needs and visualizes temperature dynamics. Results from a two-store case study demonstrate a 95% accuracy in detecting anomalies and predicting failures within operational thresholds. The proposed system offers a scalable and cost-effective solution for improving cold storage reliability. This paper also presents a novel approach to predicting the maintenance needs of cold stores in shrimp processing industries based solely on temperature trends. The study focuses on utilizing daily temperature data collected at four-time intervals and analyzing the data using a MATLAB-based linear regression model. This method aims to detect anomalies, analyze trends, and forecast future maintenance requirements. Results indicate that temperature-based predictions offer a reliable and cost-effective solution for preventive maintenance, ensuring operational efficiency and reduced downtime in cold storage systems.