Anomaly Detection for a
Waste Conveyor Belt System
Anomalyse Ltd in conjunction with AMRC Cymru
Anomaly detection on a waste conveyor system can be crucial for maintaining efficiency, reducing downtime, and ensuring the quality of waste processing. Below we outline a recent use case on equipment at the Advanced Manufacturing Research Centre Cymru, part of the University of Sheffield.
The Problem:
In a recycling plant, waste materials are transported through a conveyor system where various sorting and processing tasks occur. The conveyor system might experience different issues, such as:
Mechanical Failures (e.g., motor breakdown, misalignment of belt)
Obstructions (e.g., large or irregular waste objects blocking the conveyor)
Performance Degradation (e.g., slow conveyor speeds, inconsistent material flow)
Overload or Underload (e.g., if too much waste is on the belt, or too little material is detected for sorting)
Some issues are less prominent or even impossible to detect or predict manually without the aid of AI-powered anomaly detection. If left undetected, these issues can cause significant downtime, damage to equipment, or even unsafe working conditions.
The Process:
Data Collection: Real-time data is collected from various sensors installed on the conveyor system, including vibration, temperature and speed.
Anomaly Detection: A machine learning model is trained on a representative sample of data from the conveyor system. The model learns to identify what is typical behaviour for the system, considering all parameters simultaneously to give a contextual view of the asset’s behaviour.
Real-Time Monitoring: The Anomalyse system continuously compares real-time data with the expected patterns and provides the user with a measure of how normal (or abnormal) the current behaviour is, on a scale of 0-100. These values are accompanied by explanatory factors which explain the designated value (e.g. current abnormally high).
Alerting and Response: If a notable pattern of anomalous behaviour is identified then the user is notified by the system. Alerts can be tailored to ensure only events of interest are flagged.
The Solution:
Sensors were installed on the motors and Modbus connections made to the existing variable speed drives (VSDs):
Vibration Sensors: Monitor the motors and conveyor belt to detect unusual vibrations that could indicate mechanical issues.
Temperature Sensors: Monitor temperatures to detect overheating components, such as the motor or the belt.
VSD Registers: Track the speed of the conveyor belt to identify slowdowns or inconsistent operation.
Anomaly detection models then monitor key parameters of the conveyor system in real time, flagging any irregularities that may indicate an issue.
The Benefits:
Early Fault Detection: Detecting mechanical issues like motor failures or belt misalignments early on can prevent breakdowns and costly repairs.
Optimised Operations: Real-time alerts allow operators to address issues before they escalate, keeping the conveyor system running smoothly.
Reduced Unplanned Downtime: Identifying and resolving issues quickly ensures minimal downtime for the conveyor system, increasing the plant’s throughput.
Improved Safety: Identifying potential obstructions or unusual behaviours can prevent accidents or injuries related to the conveyor system.
Cost Savings: Predicting issues before they happen can reduce the need for costly emergency repairs and extend the lifespan of equipment.
Example Anomalies
Unexpected Vibration Patterns: A sudden spike in vibration could indicate a misalignment or wear on a motor or belt.
Temperature Anomalies: An increase in temperature could signal overheating motors or conveyor belt issues.
Slow Conveyor Speed: A sudden reduction in speed might indicate an overload, malfunctioning motor, or slippage on the belt.
In this scenario, anomaly detection helps ensure that the machinery runs more efficiently, avoiding system failures that would otherwise disrupt operations.
Want to learn more, or talk about how anomalyse can help your business improve productivity through Anomaly Detection? Contact sales@anomalyse.io today.