Predictive Maintenance of Machines Using AI & ML based Controllers

Abstract

The advent of Industry 4.0 has emphasized the significance of predictive maintenance in optimizing machine efficiency and minimizing downtime. This project proposes an innovative approach to predicting faults in a 5HP compressor driven by A800 series Variable Frequency Drive (VFD), utilizing FX 5U Programmable Logic Controller (PLC) and GOT 2000 Human-Machine Interface (HMI). Leveraging Machine Learning (ML) algorithms, specifically Random Forest Classifier, this system monitors and analyses real-time data from various sensors to forecast potential compressor failures. Historical data from sensors, such as vibration, temperature, pressure, and current, is collected and processed using FX 5U PLC. The data is then transmitted to GOT 2000 for visualization and further analysis. The Random Forest Classifier algorithm is trained on the dataset to identify patterns and correlations between sensor readings and compressor faults. This enables the system to predict potential failures, providing actionable insights for maintenance scheduling.

Specification

Description:The increasing complexity of modern industrial systems has led to a significant rise in equipment failures, resulting in costly downtime, reduced productivity, and compromised safety. Traditional maintenance strategies, such as reactive and preventive maintenance, have limitations in addressing these challenges. Reactive maintenance often leads to prolonged downtime and expensive repairs, while preventive maintenance may not effectively identify potential issues. The advent of Industry 4.0 and the Industrial Internet of Things (IIoT) has paved the way for Predictive Maintenance (PdM), a proactive approach leveraging advanced technologies to forecast equipment failures. PdM utilizes real-time data analytics, machine learning (ML), and artificial intelligence (AI) to identify patterns and anomalies, enabling proactive maintenance scheduling. This project focuses on developing a predictive maintenance system for a 5HP compressor driven by an A800 series Variable Frequency Drive (VFD). The compressor is a critical component in various industrial processes, and its failure can lead to significant production losses. Compressor failures can be attributed to various factors, including mechanical wear and tear, Thermal overload, Vibration-induced damage and Electrical faults. Conventional maintenance methods often rely on scheduled inspections and manual data analysis, which may not detect potential issues in a timely manner. Sensors are installed in the compressor to sense the temperature, vibration, current, voltage, oil conditions and other physical parameters. The values are fed to the machine model to predict the fault based on the behaviour of the machine. The AI model is taught with different failures by the recorded readings. , Claims:We claim that,
1. On designing this reduced downtime and increased productivity can be achieved.
2. Extended equipment lifespan and optimized overall equipment effectiveness (OEE) can be achieved

Documents