Page 59 - Vol 28 Issue 29 2019

equipment maintenance mining

measured at the same loads and speeds) Currently, however, predictive The benefits of a data-based predictive
signal deteriorating gear conditions. maintenance relies on more standard maintenance approach can be substantial.
Better functioning equipment improves methods to identify trends and predict The Plant Engineer’s Handbook (R. Keith
mining safety as well as efficiency. failure patterns for specific pieces and Mobley, “Predictive Maintenance,”
Studies show a direct correlation types of equipment so maintenance pp 869-870, Butterworth-Heinemann,
between equipment failures and lost- teams can accurately predict when and 2001) notes that predictive maintenance
time accidents. For example, in Ghana, where issues are likely to occur. With this decreased actual maintenance costs by
between 2004 and 2015, 85% of mining advance warning, maintenance teams can 50% and catastrophic machine failures
accidents and 90% of fatalities were schedule repairs and ensure they have by 55%. It also increased equipment
equipment related. Globally, mining the necessary parts, reducing equipment operating life by 30%, reduced spare
deaths cost mining companies $240 downtime. parts inventories by 30%, and allowed
billion, according to estimates from the Shifting Management’s Mindset engineers to predict mean time between
International Labour Organization. failures and thus determine the best time
Getting the most from vibration and other to replace equipment. As a result, plant
Analyze the Data sensors requires a management process production output increased by 50%.
Predictive maintenance programs to identify conditions that trend toward Similar gains are possible in mining and
generate a huge amount of data that failure, and a cultural mindset that spurs other heavy industries.
must be analyzed to reveal trends. prompt action. Knowing something is Vibration Sensors are Step One
Sensors, therefore, go hand-in-hand with likely to break and then waiting for it to
computerized maintenance management actually break defeats the purpose of Vibration sensors are an important part of
system (CMMS) software. Combined, monitoring. Instead, maintenance teams a comprehensive predictive maintenance
sensors and CMMS applications can must respond to fix problems before they program. Until relatively recently,
provide a performance baseline and occur, to increase the uptime for specific vibration sensors were expensive and
analyze data to help managers identify equipment, and improve overall mining difficult to place on equipment. Steadily
equipment conditions that are outside operations. advancing technologies and wireless
normal operating parameters, monitor For example, the McKinsey Global capabilities have overcome those barriers,
trends and identify the root causes of Institute report cited earlier says one making sensors more cost effective and
failure. mining company used sensors and easy to place, even in hard-to-access
machine learning to predict equipment locations within machinery.
failures in its 20-ton heat exchangers. By Mining companies have noticed.
Some of the most forward- doing so, it reduced maintenance from According to the same 2018 Future of
thinking operators are taking every 70 days to every 160 to 200 days, Mining Survey, nearly 40% of responding
analysis another step forward incurring substantial savings. companies plan to deploy condition
monitoring sensors on their mobile fleets
by using machine learning The global mining industry is among by 2023, and all expect the internet
and artificial intelligence (AI) the industries most receptive to of things (IoT) to become ubiquitous
to sift through the mountains predictive maintenance even though throughout their operations.
implementation is not yet pervasive. Many
of data from sensors, managers realize that approximately 40% Advanced monitoring technologies,
condition logs, maintenance of mining (and other heavy industry) costs combined with analytics, are transforming
records, geologic conditions, are related to asset management. If those the mining industry, making it more
production trends, weather assets, mainly equipment, run efficiently, efficient, more productive and, more
and other information operating costs go down. importantly, safer. Onsite predictive
that otherwise may not be This realization is changing the value maintenance, even in remote and
correlated and, in fact, is proposition, according to Deloitte’s global hazardous conditions, is becoming an
important component of today’s mining
often unused. Although mining leader, Phil Hopwood, quoted operations and will continue to grow in
AI and machine learning in the 2018 Future of Mining Survey. importance moving forward.
“The industry’s value proposition may
applications are still in their be shifting to how well a company acts
early stages, preliminary on information to optimize production,
accounts suggest they will reduce costs, increase efficiency and
be used increasingly to help improve safety,” Hopwood said. Data, and
users identify probable the ability to analyze and integrate it, is
causes of specific issues. becoming a competitive differentiator.
Angela Kerr

AFRICAN POWER Mining & Oil Review Vol 28, Issue 29, 2019 | 59