What need did Rithmik Solutions (Rithmik) see in the market for its Asset Health Analyzer (AHA)?

My co-founder Kris Isfeld and I both used to work at Matrikon, which had a product that collected data from mobile mining equipment. In the 10+ years each of us worked with that data, the value in it was clear, but companies never seemed able to fully unlock that value. After leaving Matrikon and studying AI and machine learning, we realized that these technologies were the ticket to unlocking the potential of this data to drive better maintenance-related decision-making. And of course, this had to be done in a way that handled the complexity and extremes of mines. We didn’t see any other solutions out there doing that, and that’s why we started Rithmik.

Five years later, we’ve worked with mines in four commodities on three continents and validated with our customers that our product can save upwards of US$10 million per year for an average-sized fleet of haul trucks. At a mine in Zambia, for example, we showed that some of the trucks had brake issues that significantly impacted the handling of the trucks and were creating wear on the powertrains. They’d had no idea what was causing them until they worked with us.

How does AHA utilize AI to predict what might go wrong with machinery before it happens?

We build an optimal replication model using AI that captures the characteristics of a piece of equipment running at its best. We then run the model at the same time as the actual assets, enabling us to see which sensors are deviating – for instance low oil pressure with high exhaust temperatures – essentially capturing the failure mode signatures from the equipment. We tailor those models to the mine environment using our proprietary infrastructure, which is how we can continue to provide highly-accurate results even in very different environments. So far, we’ve worked with various haul trucks and dozers and have had significant interest in running the system on shovels.

What are the main benefits of AHA’s implementation?

One major benefit is the early prediction of problems down the line. The system can also identify operator behaviors that are causing issues to the equipment as well as instances in which mechanics are not addressing the root causes of broken equipment. AHA also identifies inefficiencies, thereby allowing an operator to improve fuel burn and ultimately decrease greenhouse gas emissions. So far, we’ve measured opportunities to save almost 8,000 t/y of CO2 at the site level, and that’s just the beginning.

How has the market been responding to AHA?

We are receiving more requests today as people are increasingly looking for predictive maintenance solutions. Customers still want to see proof that what we’re doing works because a lot of companies have tried things in the past that haven’t worked. The good news is that we’ve always been able to deliver that proof.

Looking into the future, are there any other applications of AI in mining that excite you?

Rithmik is covering a piece of the mining value chain. Fully integrating the information we’re providing with the processes around it will amplify the benefits even more. A future where full automation is possible is very exciting.

What would Rithmik like to achieve in the next two years?

I see the company having significantly more partners, all collaborating in optimizing the entire value chain through technology and AI. It is important for us to have a focus and to do what we do extremely well, rather than trying to do everything ourselves. We’re already making our insights more explicitly prescriptive, and we’ll also be covering more equipment types.