Experts say software and automation advances provide key tools to target longtime goals,
but don’t solve the problems of increasing complexity and the generation gap
by jesse morton, technical writer
Hauler maintenance as a discipline has not changed much in seven decades since Letourneau’s 32-ton haul truck debuted in 1952, Dr. Tim Joseph, JPI and professor, mining engineering, University of Alberta, told Coal Age. The goal continues to be optimal availability.
More recently, the advent of artificial intelligence-leveraging fleet management systems offer powerful tools toward that goal. Yet, almost paradoxically, “with the advances in artificial intelligent systems, there is a danger of losing sight on what is actually real,” Joseph said. “The trap of creating and believing artificial data through the myriad of data management, analysis and interpretive systems, where end-users are sold on the notion of predicting failure, has, in part, replaced good old common sense.”
That warning is echoed by other experts, who told Coal Age that as the discipline is becoming more digitized, it risks deprioritizing a critical component, the personal development of technicians and managers who ultimately must make the crux decisions and do the routine work. If the industry trades skill development programs, when they are now needed most, for automated solutions, a brain drain of sorts would be the result, and not without consequences, Joseph said.
That fate, experts say, should be avoided if possible. They say, in theory, solutions exist, that a middle path can be forged. Meanwhile, the latest news from the space reveals how the discipline is evolving, in practice, one overhaul at a time.
The Discipline, in Theory
Paul Tomlingson, in his 2014 book Maintenance in Transition-The Journey to World Class Maintenance, argued that there are three key elements to an efficient maintenance program: 1) materials, 2) space and 3) time. More broadly, 1) parts, or materials, must be 2) inventoried and warehoused properly. And 3) orders must be placed timely. The three elements must be managed “harmoniously” and “dependably,” Tomlingson reported. For that to happen, certain protocols should be followed.
One of the topmost protocols is the maintenance area must be subject to strict rules of accountability. Tomlingson warned that without it, compartmentalization creates a situation where one team is blind to the actions of another. One typical result is a void of key parts in a main warehouse, where they should be, while a pile of the same is squirrelled away in an on-site hidey hole. The intent may be good, but the effect is problematic. “Industrial plants must solve the long-standing troublesome “cultural pack-rat” problem as one of the initial steps of improving maintenance performance.”
Tomlingson wrote that good planning can improve labor productivity. The goal should be to plan routine maintenance to minimize the need for emergency repairs. Planned maintenance typically is efficient, increasing the labor production rate. Planned work, he reported, tends to require roughly 15% less labor than unplanned or emergency work.
A big part of planning is ensuring the right materials are on hand in advance of the planned work. This responsibility is often split between the planner and supervisors.
In turn, material control responsibilities for planners, supervisors and crew must be specified. “Where possible, set up a periodic maintenance program that prescribes standard bills of materials for the repetitive major jobs,” Tomlingson reported. Personnel should be trained in their specific material control responsibilities.
Material control personnel should be able to effectively audit planners, supervisors and crew on their adherence to those responsibilities. “No maintenance work of any sort should be a mystery,” he said.
As a part of planning, forecasting is key to preventing material control responsibility infractions. “By identifying the future time when major components might be replaced, an advance need for materials is announced enabling material control to respond more effectively,” Tomlingson wrote.
Predicting future needs is an exercise that centers on analyzing historical trends. “Maintenance repair history information is used to establish the forecast for the replacement of major components,” Tomlingson reported. That information can be culled from equipment monitoring systems.
The analysis should drill down to discover patterns related to the timing of preventative maintenance tasks. For example, condition monitoring and repair history data can be assessed to determine “a meantime before failure for each of the major components on each family of mobile equipment,” Tomlingson wrote. “Once this data is analyzed and confirmed, it is possible to create a forecast well into the future that identifies each major component on every unit of equipment with an estimate of when those components should be replaced.”
Forecasting is critical to ensuring materials that need assembly prior to installation are assembled in advance of the job. “When the bill of materials is shared with warehousing and purchasing, they can assemble necessary materials in advance and ensure on-time delivery,” Tomlingson wrote. “The forecast alerts all parties of the need to replace specific components and permits purchasing to establish timetables for the completion of component rebuilding to ensure that the rebuilt components are restocked at a rate that matches maintenance needs.”
The shop or garage where maintenance is done must be both transparent and organized for efficiency. “As field planners commit to schedules, they must know the up-to-date status of shop work in order to schedule the field work that will incorporate completed shop work,” Tomlingson wrote. “Once the weekly schedule has been launched, daily coordination meetings are held to adjust the timing of support services to accommodate operational delays.”
The shop must be prepared and maintained to handle numerous mobile equipment configurations. “As major components are removed and replaced, sufficient lifting capability must exist often in the form of overhead cranes,” Tomlingson wrote. “Both the conduct of the mobile equipment maintenance program and the configuration of the garage in which maintenance is being carried out should include consideration of the special configurations.”
Shop customers should be able to track job status and progress so they can schedule related field work. “There must also be shop procedures that preclude emergency work from hindering on-going shop work,” Tomlingson wrote. “Service activities that directly affect field work must be especially well-coordinated.” The work priority system in the shop should be fair to help the flow of critical work and to make sure resources are doled out equitably.
The shop should have key performance metrics that are designed to identify problems related to control and resource utilization.
Regarding mobile equipment, such as haulers, increasingly “operations often see units as a chassis on which components are removed and replaced as determined by accumulated operating hours,” Tomlingson wrote. “Therefore, major component replacements are seen as a significant aspect of mobile equipment maintenance.”
When major components like engines or transmissions wear out in multiples, overhauls are considered. “Generally, costs and repair history will dictate when this time has arrived,” Tomlingson wrote. Typically, “decisions are made whether to replace the equipment if the overhaul cost exceeds 60% of the cost of replacement equipment.”
The Discipline, in Practice
On paper, harmonizing the basic elements of maintenance seems logical and simple enough. Applying the discipline, especially when it comes to maintaining haul trucks, presents challenges that vary in degree depending on the mine.
Gerard Wood, managing director of Blufield Asset Management and the author of the Simplifying Mining Maintenance, told Coal Age the biggest maintenance challenge for miners with large haul fleets is, was, and likely will continue to be “just getting the basic trade work completed consistently with the right quality.”
Wood advises that a major need now becoming apparent is sufficient training in what he calls basic skills. “I think that there is a significant need in the industry for articles on the basic technical trade/craft skills,” he said. “These are being forgotten in the world today.”
The numbers show that a site culture that values personal technical development will outperform others that don’t, Wood said. “We have data to prove this,” he said. “It may not be sexy or even trigger the interest of people, but if the CEOs and CFOs knew the value that it can bring they would all be reading and learning about it.”
The need for personal technical development initiatives has never been higher, Dr. Joseph said. What the latest round of technological advances in the space has done for increasing efficiency and productivity has to be balanced against the complications introduced to the area of maintenance. “Maintainers skill sets are not changing but getting far more varied,” Joseph said. “We still need the mechanics, electricians and welders, but we now need a new breed of technologist that understands heavy equipment communications systems, computerized control systems, artificial intelligence, systems integration and the list goes on.”
Technology schools and colleges are well-attended and cranking out graduates, but opportunities for journeyman training in the field are relatively rare. “We see computational consulting groups stepping up to the fray with digital skills, but no experience in our industry,” Joseph said. “What is missing are the opportunities to bring this next generation into our world, able to understand the operating conditions and complex interactions by fleet and field service interactions.”
Of particular concern, both said, is AI and automation being pitched as the patch for the emerging generation gap. For example, AI cannot yet replace the understanding of an experienced operator when it comes to monitoring truck condition in real time, Wood said. “This has not yet been solved.”
Autonomous haulers have changed maintenance by making it more predictable, Wood said. “However, I have not seen the automation make any other differences as yet.” The data is simply not available to compare maintenance costs of partially or fully autonomous fleets to manually operated fleets, he said.
Predictable can but doesn’t necessarily always translate to meaningful cost savings, Joseph said. Systems designed to automate truck condition monitoring “inherently have so many assumptions, compounded by probability theories, the outcome rarely presents reality and even rarer is able to predict beyond common sense understanding of the likelihood of failure,” Joseph said. “Mine operating conditions change so quickly that AI cannot keep up, but real-time monitoring just keeps on indicating exactly what the hauler is exposed to as it happens.”
Specifically, fleet management systems that leverage AI theoretically appear capable of improving availa-
bility and lower costs. In practice, if those systems are introduced in place of personal skill development initiatives, or, worse, without a sufficient training regime on their proper usage, whatever benefits they offer could be quickly offset, Wood said. “The problem is that the improvements cannot be realized or achieved if the basic maintenance trade/craft skills are not good, which starts with the leadership,” he said. “If you have the best data system in the world and it tells you to go and change a component or to take some action on a truck, if the mechanic or electrical technician does a bad quality job, it will easily and quickly create a failure that will negate all of the benefits of the system.”
Wood said data systems can theoretically achieve from 90% to 93% availability, with lower costs, “but if we are operating in the low-80s% then the data will not help.”
In practice, behind the scenes, hauler availability and maintenance costs continue to sway major decisions on fleet makeup, mine design or even whether to open a new op, experts told Coal Age.
Determining fleet makeup can hinge on what the going rate is for hauler maintenance technicians in the area, Wood said.
The maintenance of ultra-class trucks requires more man-hours, making them costlier than smaller haulers. “This is usually made up for on the cost-per-ton, as they move more,” he said. “The fleet availability should be almost the same as the smaller trucks if the maintenance is done correctly.”
Where labor costs are lower, however, ultra-class and autonomous haulers make less sense, Wood said. “There is a sweet spot,” he said. “The 300T-class truck, like the Komatsu 930, is hard to beat on a cost-per-ton basis when the labor costs are not in high-labor-cost regions of the world.”
Wood advised mine engineers to site logical locations and facilities for hauler maintenance to occur when designing a site. “Not only the workshop but areas in or near the pit are needed for some tasks and inspections,” Wood said. “They should also design the mine so that the material can be mined within the operating limits of the machine. If the ramps, operating angles, roads, blasts and dumps are not designed so the equipment can operate within its design limits, this causes damage to the machine, which will reduce its available time and significantly increase the costs.”
Worse, mines not designed with optimal hauler availability as a key priority will see cut-off grade impacted as the mine progresses, Dr. Anoush Ebrahimi, principal engineer, SRK Consulting, told Coal Age. “What usually people don’t get right is that ‘reserve’ directly is linked to operating costs,” he said.
“In open-pit mines, maintenance and availability are always major contributors of production cost,” Ebrahimi said.
To lower maintenance costs and increase availability, mine engineers should focus on roads, Joseph said. “Good roads means less hauler maintenance, guaranteed,” he said.
There are other fixes that amount to low-hanging fruit, Joseph said. “The biggest surprise is what is available on almost all haulers and is not being used,” he said.
For example, all suspension systems have linear pressure transducers. “Many have a range of 0 to 15,000 psi for a response between four and 20 milliamp,” he said. “These sensors are reliable and are not prone to drift in use.” Few ops capture even a simple real-time data sample or stream, he said.
“A few duty cycles raw data record can easily and simply be viewed and interpreted to reveal huge impacting knowledge about hauler, tire and road performance and the expectation of maintenance or impending failure,” Joseph said. “Anyone can be trained in a couple of days to do most of this for themselves.”
There are hidden maintenance costs to operating in locales that are dangerous due to climate, crime or politics, Wood said. “Only removing people will be a long-term solution but from a maintenance perspective we are a long way from removing people,” he said. “We can remove the operators and we can create maintenance robots, but even the robots will breakdown and need inspections.”
For now, Wood said, in most cases, the more viable solution appears to be “looking for more sustainable ways to mine current deposits.”
The Discipline, in the News
New Acland Mine used the local Hastings Deering workshop in Toowoomba for full component changeouts on its entire fleet of Caterpillar 793F haulers. The work spanned six months and cost $2.2 million for each of the trucks.
The fleet was six years old, and each unit had roughly 24,000 work hours on the site.
The trucks reportedly cost roughly $5.5 million new. The miner reported it was “worth the expense.”
Maintenance supervisor at the mine, Rob Trapp, said it was the biggest job it attempted for the miner. “It was a huge job, replacing basically every component on the truck,” he said.
The trucks each received a certified power train rebuild, or a full replacement of the drive-train components, to include the engine. They were worked on in succession. “Each truck took three weeks and we had a week break in between each truck,” Trapp said.
The trucks were shipped from New Acland to Toowoomba. “We just take off the cab, the tray and the wheels and load it onto another big truck,” Trapp added.
This article was adapted from an article published in the August 2019 edition of Engineering & Mining Journal (E&MJ).