Q You specialise in predictive maintenance which gives you a unique insight into equipment performance. How have you seen the maintenance trends change for fleets of autonomous trucks in Australia based on the data for TRAKKA?
A A key source of work identification has always been an equipment operator, from an unusual sound or vibration to a machine responding differently (harder steering, less breakout force, sluggish hydraulics). The introduction of autonomous equipment has removed this form of work identification, so equipment owners/operators need to rely even more on condition and performance monitoring technologies such as vehicle sensors, equipment inspections, and fluid analysis, than they have in the past. The data captured from an autonomous truck is no different to a regular manned truck, aside from the additional sensors required for navigation and collision avoidance. So far, the predictive maintenance processes are effectively the same, minus anything an equipment operator would log during his shift.
Q How does TRAKKA differentiate itself from onboard maintenance monitoring packages offered by the equipment OEMs themselves in terms of capability and functionality?
A TRAKKA® works in conjunction with OEM onboard monitoring systems to help miners get the full benefit from them. A key advantage of TRAKKA is that it can ingest data from any OEM system, allowing a maintenance team to collect and collate the information from all their machines, regardless of OEM or equipment type, into one easy-to-use system. It seamlessly captures data from a variety of sources, including onboard sensors, oil analysis & equipment inspections, enabling our customers to analyse, manage and act on this information using one centralised platform. A big reason that miners approach DINGO is that they are frustrated with their inability to maximise on their data due to all the disparate systems in play. Moreover, most OEM systems are not integrated with a customer’s ERP system, so we are seeing more and more clients take data from systems such as Minecare, Minestar and Komtrax and feed it into Trakka. Once it’s in Trakka, impending issues are identified and quickly turned into work orders in systems such as SAP, Oracle, Ellipse and Dynamics AX through DINGO’s proprietary integration platform. End users don’t want to waste time transposing data from one system to another. They want to see all their condition monitoring data and work history in one platform.
Q For mining trucks specifically, what are the key maintenance parameters that mines want to keep on top of?
A Our larger customers are consistently tracking data from onboard sensors, oil analysis and equipment inspections. The main focus is always on the costliest systems such as engines, wheel motors, differentials/final drives, transmissions and suspension. The most critical parameters to track vary by system:
- For engines, our customers tend to focus on sensors measuring temperatures (exhaust, turbo, coolant) and pressures (filter, oil, boost), as well as oil analysis parameters related to wear and abnormal combustion.
- For differential/final drives and wheel motors, most of the attention is on oil analysis for wear and ferrous materials, and visual inspection (photos) of magnetic plugs. Wheel motors sensors reading temperatures, speed and power have proven effective in identifying issues.
- For transmissions, miners are looking at wear and particle counting from oil analysis, filter and screen inspections for clutch material and measurement of clutch shift/slip times
- The suspension system is measured via suspension cylinder pressures and the overall truck rack and pitch as it moves down the haul road.
Q How popular is your DINGO Managed Solution offering in terms of mining operations and for truck fleets in particular?
A A meaningful portion of DINGO’s customer base utilises our Condition Intelligence expert service because of their deep functional and subject matter expertise, and trucks tend to be a focal point due to fleet sizes and the potential for significant cost savings. However, very few clients currently use the full managed solution. It was popular as recently as a few years ago as the mining industry was evolving in the realm of condition monitoring, but today, most mid to large scale mines have a skilled team of resources who use TRAKKA to run their predictive maintenance programs. In our experience, junior to low level mid-tier miners typically don’t have access to the same level of resources, so these mines stand to benefit the most from DINGO’s managed solution. We have helped a number of these smaller mines transition away from costly maintenance and repair contracts.
Q More and more mines are installing LTE networks; does this mean customers will be able to get greater real time capability from TRAKKA?
A One of the limiting factors in getting real time data from equipment has been the communication systems at mines. As they improve the coverage and bandwidth of these networks, it will allow the vast array of data being collected and stored on the machines to be communicated to TRAKKA. While we don’t intend to replace real-time monitoring systems, TRAKKA will use this information to predict potential issues even earlier and recommend corrective actions faster. Our data scientists will also be able to use this information to update TRAKKA’s predictive analytics models more rapidly than we are able to now, which will further improve decision-making.
Q While you have been a trailblazer in predictive maintenance, the market does not stand still, what is the “next level” in terms of TRAKKA capability and functionality?
A Most of our customers have invested heavily in fleet management systems, so they can make decisions about immediate equipment issues (alarms going off in the cab) or dispatching equipment to optimal loading face. However, there is still a significant opportunity to harvest and use all a mine’s condition monitoring information to create the proverbial crystal ball for maintenance. DINGO has over 25 years’ worth of condition monitoring data on mining equipment, along with the corresponding maintenance recommendations, work performed, and outcomes. We are working with data scientists to tap into the power of this data and develop machine learning models that will automatically deliver the optimal corrective actions, detailed work instructions, and associated confidence levels. The beauty of these models is that they will get smarter and smarter as we continue to feed them new information. If we can get this right, miners will have the power to anticipate the future and proactively manage equipment maintenance, while reaping huge productivity, planning and safety benefits in the process.
Paul Moore │International Mining │ May 2019