According to DINGO CEO Paul Higgins, big data had become a buzzword across many industries but thick data was where it was at.
Put simply thick data is big data with some human context added to it.
“Thick data is simply the idea that numbers alone aren’t enough,” Higgins said.
“Resources operators are gathering vast amounts of data on their sites but not always able to apply the analytics in practical, effective ways.
“It’s adding the layer of context that humans can bring.”
Thick data may be better than big data for solving maintenance conundrums.
He gave the example of a job the company did in Canada for a company that was having trouble with its Komatsu 930E trucks.
Their trucks were suffering wheel motor failures and despite lots of efforts and data analysis they could not get to the cause of the problem.
At that stage DINGO had been doing a lot of work on oil analysis. However, the data the company gave it was electrical data.
The DINGO team asked for the oil data and eventually received it.
Their analysis of the oil data showed signs of silicon in the wheel motor armatures and gearboxes of the wheel motors that had suffered major failures.
“We presented this data to the team,” Higgins said.
“The guy from the customer’s side who had been on this for three years was angry.
“He said we didn’t know what we were talking about and threw the report on the floor and stormed off.”
It was his reaction that made Higgins think they might be onto something.
On further investigation the DINGO team discovered that when the trucks were being brought in for programmed maintenance the maintenance crews would use high pressure hoses to wash the mud off the wheels before the trucks were taken into the heated workshop.
It was this washing that was forcing grit and water through the wheel motor seals and into the armatures and gearboxes and ultimately causing the engine failures when the trucks returned to duty.
Higgins said thick data was not something DINGO had come up with but he could see the value in the concept.
“One of my team sent me a TED talk from Tricia Wang,” he said. “She talked about it.”
Higgins believes the mining industry is not taking advantage of the data at its disposal.
“Less than 1% of available data in the mining industry is being used,” he said.
“By unlocking the value of this dormant data, the industry could save US$100 billion in cumulative maintenance costs by 2025.
“Mining operations have always understood the importance of collecting data but being measurable doesn’t make it valuable.
“By uniquely applying a layer of context, or thick data, DINGO is now giving them the power to access it more immediately, understand it, and apply it quickly to real-world situations.
“And this leads to real world results.”
DINGO recently launched Trakka 4.5, a new version of its predictive maintenance software.
Noel Dyson │ Australia’s Mining Monthly