By Angela Kerr
The mining industry operates in some of the harshest and most remote parts of the world, requiring equipment that is as reliable and efficient as possible. Robotics and semi-autonomous equipment only add to the maintenance complexity. When breakdowns occur, whether for the latest technologies or for tried and tested tools, they are time-consuming and expensive, resulting in lowered productivity, lengthy delays and an increased risk of injury. To minimize such disruptions and expenses, mining operators are turning to predictive maintenance to prevent breakdowns before they happen, keeping equipment functioning at optimal performance levels.
Historically, mining companies have relied on scheduled maintenance to keep equipment operating properly. Predictive maintenance systems improve on this by monitoring equipment while it’s in use. Also called condition monitoring, this method relies on sensors that, once placed on multiple pieces of equipment, capture data needed to help maintenance teams anticipate failure. Knowledge of potential failure then allows them to make necessary repairs regardless of the maintenance schedule. As a result, the maintenance team can better focus its efforts, minimize unscheduled downtime, increase equipment effectiveness and enhance both personnel and environmental safety.
In 2018, a McKinsey Global Institute report listed predictive maintenance as one of the mining operations activities with the highest rate of returns. Its earlier 2015 report estimates the economic impact of sensors, those used for predictive maintenance on the worksite to better manage operations and equipment, will range from $200 billion to $900 billion by 2025.
Vibrations as Early Warnings
Some equipment vibrations are normal, but those outside the accepted range signal impending equipment failure. For example, if a bolt or tie-down has loosened or a belt, gear, bearing, drive motor or other component is worn, vibrations will be different than if all components are new. Measuring the frequency or amplitude of vibration provides an early warning that something is about to go wrong.
Previously, machine condition-monitoring programs were not able to detect and communicate vibrations and impacts. However, that standard is changing as vibration sensors specifically designed for operating equipment are being introduced and deployed. These advanced sensors detect low frequency vibrations and impacts that fall outside normal ranges and provide real-time alerts that help maintenance teams react before serious faults occur or equipment fails completely.
Vibration monitoring not only helps predict impending equipment failure, but also alerts maintenance teams to changes in operating parameters that may affect how the equipment operates. For example, bearings on rotating machinery, such as pumps and conveyor lines, gradually wear down, harming performance. Vibration monitors can detect the subtle variations that signal degradation so repairs can be made before operations are seriously affected. This is also true for draglines and other equipment, where changes in vibration frequency over time (when measured at the same loads and speeds) signal deteriorating gear conditions.
Better functioning equipment improves mining safety as well as efficiency. Studies show a direct correlation between equipment failures and lost-time accidents. For example, in Ghana, between 2004 and 2015, 85% of mining accidents and 90% of fatalities were equipment related. Globally, mining deaths cost mining companies $240 billion, according to estimates from the International Labour Organization.
Analyze the Data
Predictive maintenance programs generate a huge amount of data that must be analyzed to reveal trends. Sensors, therefore, go hand-in-hand with computerized maintenance management system (CMMS) software. Combined, sensors and CMMS applications can provide a performance baseline and analyze data to help managers identify equipment conditions that are outside normal operating parameters, monitor trends and identify the root causes of failure.
Some of the most forward-thinking operators are taking analysis another step forward by using machine learning and artificial intelligence (AI) to sift through the mountains of data from sensors, condition logs, maintenance records, geologic conditions, production trends, weather and other information that otherwise may not be correlated and, in fact, is often unused. Although AI and machine learning applications are still in their early stages, preliminary accounts suggest they will be used increasingly to help users identify probable causes of specific issues.
Currently, however, predictive maintenance relies on more standard methods to identify trends and predict failure patterns for specific pieces and types of equipment so maintenance teams can accurately predict when and where issues are likely to occur. With this advance warning, maintenance teams can schedule repairs and ensure they have the necessary parts, reducing equipment downtime.
Shifting Management’s Mindset
Getting the most from vibration and other sensors requires a management process to identify conditions that trend toward failure, and a cultural mindset that spurs prompt action. Knowing something is likely to break and then waiting for it to actually break defeats the purpose of monitoring. Instead, maintenance teams must respond to fix problems before they occur, to increase the uptime for specific equipment, and improve overall mining operations.
For example, the McKinsey Global Institute report cited earlier says one mining company used sensors and machine learning to predict equipment failures in its 20-ton heat exchangers. By doing so, it reduced maintenance from every 70 days to every 160 to 200 days, incurring substantial savings.
The global mining industry is among the industries most receptive to predictive maintenance even though implementation is not yet pervasive. Many managers realize that approximately 40% of mining (and other heavy industry) costs are related to asset management. If those assets, mainly equipment, run efficiently, operating costs go down.
This realization is changing the value proposition, according to Deloitte’s global mining leader, Phil Hopwood, quoted in the 2018 Future of Mining Survey. “The industry’s value proposition may be shifting to how well a company acts on information to optimize production, reduce costs, increase efficiency and improve safety,” Hopwood said. Data, and the ability to analyze and integrate it, is becoming a competitive differentiator.
The benefits of a data-based predictive maintenance approach can be substantial. The Plant Engineer’s Handbook (R. Keith Mobley, “Predictive Maintenance,” pp 869-870, Butterworth-Heinemann, 2001) notes that predictive maintenance decreased actual maintenance costs by 50% and catastrophic machine failures by 55%. It also increased equipment operating life by 30%, reduced spare parts inventories by 30%, and allowed engineers to predict mean time between failures and thus determine the best time to replace equipment. As a result, plant production output increased by 50%. Similar gains are possible in mining and other heavy industries.
Vibration Sensors are Step One
Vibration sensors are an important part of a comprehensive predictive maintenance program. Until relatively recently, vibration sensors were expensive and difficult to place on equipment. Steadily advancing technologies and wireless capabilities have overcome those barriers, making sensors more cost effective and easy to place, even in hard-to-access locations within machinery.
Mining companies have noticed. According to the same 2018 Future of Mining Survey, nearly 40% of responding companies plan to deploy condition monitoring sensors on their mobile fleets by 2023, and all expect the internet of things (IoT) to become ubiquitous throughout their operations.
Advanced monitoring technologies, combined with analytics, are transforming the mining industry, making it more efficient, more productive and, more importantly, safer. Onsite predictive maintenance, even in remote and hazardous conditions, is becoming an important component of today’s mining operations and will continue to grow in importance moving forward.