Machine learning has great applicability for risk modeling and management of the modern mine. Mining companies continue to make extensive technological investments in monitoring hardware and software on their mobile and fixed assets and human resources. Many mobile equipment fleets have a fleet management system and sensors tracking equipment health. These sensors provide the location, activity, usage, and many other indicators representing the machines operational state. Much of this data can be used to infer the human behavior of the operator and, to some extent, the maintainers of the equipment. Fixed equipment, commonly found in mineral processing / comminution circuits, also produce large amounts of data. Process data can pinpoint the equipment state, whether a piece of equipment is in a “trouble” mode, down, running, or standby. These states provide information on where equipment is being maintained, time of the day, and type of equipment. Thus, they can provide key insight into the operator and maintainer activities of the plant. While data sets do not track the entire mine site, they can in some cases represent a significant proportion of the work force. Many companies are seeking to extend digitization of their sites through deploying industrial internet of things (IOT) technology. Location, fatigue monitoring and fitness for duty tracking are a common IOT objective of human resource management in mining. Formal health and safety management systems track vast amount of activities and safety events. For instance, when a health and safety event occur (injury, property damage, incident, accident, near miss, violation, etc.) detailed information is captured depending on the severity of the event. Regulatory events such as a reportable injury, lost time incident, or violation are investigated following a standard process. Many mines track none regulatory safety events such as property damages, near-misses, and environmental. In addition, many localized & macro risk assessments are done daily through Field Level Risk Assessment, Take 5, Safe work observations, etc. These tools can provide valuable insight into the risk behavior & environment of a workforce. These siloed data sources can provide key business intelligence of a mine site and are highly underutilized. Machine learning offers a tool which enhances the knowledge discovery potential of the modern mine. A machine learning model is presented built using key operational and safety data of an operating mine. Results of the modeling are presented along with a process to incorporate these tools into health and safety management.
