Reliability and Safety Assessment of Dumper Operations in Opencast Metal Mines Using Risk Analysis and Machine Learning
Dumpers are critical components of Heavy Earth Moving Machinery (HEMM) used in opencast mining operations for transporting overburden and ore. These devices are subject to frequent breakdowns, operational inefficiencies, and safety risks owing to their continuous operation in challenging environmental and operating circumstances. In addition to decreasing production efficiency, unplanned breakdowns increase maintenance costs and jeopardize the safety of adjacent workers and operators.’ Thus, enhancing the dependability and safety of dumper operations is crucial for environmentally friendly mining methods. This study investigated the reliability and safety performance of dumpers operating in two opencast metal mines in Odisha, India. A comprehensive framework integrating Hazard Identification and Risk Analysis (HIRA), Failure Mode and Effects Analysis (FMEA), and machine learning techniques was adopted. Operational data from 20 dumpers, spanning approximately 2.5 years, were analyzed, including failure frequency, breakdown duration, equipment utilization, and maintenance records. Exploratory data analysis revealed that tyre issues (15.9%), suspension failures (11.1%), and oil leakage (10.6%) were the most frequent failure categories, which aligns with similar failure trends reported in mining equipment studies. This study also evaluated existing functional safety systems, such as proximity sensors, fatigue monitoring systems, tyre pressure monitoring systems, and automatic fire detection systems, installed in modern dumpers. This study also proposes a machine-learningbased approach to enhance risk prioritization and predictive maintenance, which has been widely recognized as effective in modern industrial systems. Additionally, failure statistics such as frequency, downtime, and criticality are utilized for datadriven manpower planning, ensuring optimal allocation of skilled personnel and improved maintenance response. The results highlight the impor