Farmer Iron
Using data to predict equipment failure
KEEP IT RUNNING: Dependable equipment, no matter the color, still needs maintenance. One company has a tool that could use machine data to predict failures, or help with diagnosis for quicker service.

Using data to predict equipment failure

Software used in the auto industry could someday come to ag for maintenance management.

Newer pieces of farm equipment are also data collectors. The host of sensors in that tractor or combine gather information about temperatures, pressures, time at work and more as you move through the field. How is machine data of value? If one company has its way, that information will become a tool for predicting failures early for better maintenance in the future, and perhaps less expensive unscheduled down time.

Predii, pronounced Pre-DEE, is a repair intelligence platform that is currently imbedded in more than 100,000 hardware tools performing automotive fault diagnosis. What does that mean? Essentially Predii has a system that helps a technician diagnose, and repair, a problem based on data for the machine.

Tilak Kasturi, CEO, Predii, explained how that works in the automotive world. "If the air conditioner goes out on your Ford F-150 pickup truck, there is a standard set of checks the technician will perform," he said.

Those checks might include refrigerant level, and the compressor, which is standard procedure, but may not fix the problem. "If the compressor relay switch is an issue, then the system would not be fixed unless the technician ruled that out," he said. "But this relay is not usually a suspected failure." Kasturi explained that with Predii, data showing that the relay goes bad on F-150 pickups would be available to the technician so he or she would check it and fix it - making sure the truck was ready to run again. (Note, this is an example, there is no specific F-150 data about an air conditioner relay).

"If that were the problem the relay is a$50 component that takes 5 minutes to replace, and Predii would have helped the technician find it quickly," he said.

The Predii system can help diagnose auto problems with 94% accuracy, the company claims. But Kasturi is more excited about how the machine data available can be used in a different way - predicting failure.

"We do not work in the ag space yet, this would be a new area for us," he explained. "These companies are working through dealer networks and collecting data from the equipment…we go where the data is being collected."

Putting data to work

Once the data is together Kasturi explained that analyzing the information provides the potential for predictive maintenance. Knowing something will fail in a few working hours, giving you time to schedule downtime, is much more efficient. And if you can schedule downtime - if there's enough notice - no work will be missed. That has tremendous potential from an opportunity cost basis - downtime at planting is more expensive than downtime when it's raining during planting.

"When you have access to the data you can learn as a human is learning, and not make the same mistake. We make sense out of the data and extract the repair intelligence with our system," Kasturi said. "We put that back into a system for the technician creating a virtual assistant."

The rising amount of machine data being captured in the field would be solid fodder for Predii to create - using artificial intelligence - tools for predicting service needs. The system may also one day be used to help diagnose machines in the shop to make the right decision.

While Predii isn't in agriculture yet, it's an example of technology that's closer to the engine management systems that have been predicted in the past. The idea that an engine can tell you when it will fail, with time for repairs, may not be that far into the future.

It's a future where all the big data is actually doing you some good.

TAGS: Technology
Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish