• ITVI.USA
    12,124.580
    -525.260
    -4.2%
  • OTRI.USA
    27.850
    -0.080
    -0.3%
  • OTVI.USA
    12,070.710
    -528.180
    -4.2%
  • TLT.USA
    3.080
    -0.150
    -4.6%
  • TSTOPVRPM.ATLPHL
    2.890
    0.260
    9.9%
  • TSTOPVRPM.CHIATL
    2.930
    -0.150
    -4.9%
  • TSTOPVRPM.DALLAX
    1.280
    0.100
    8.5%
  • TSTOPVRPM.LAXDAL
    3.000
    -0.210
    -6.5%
  • TSTOPVRPM.PHLCHI
    1.750
    0.120
    7.4%
  • TSTOPVRPM.LAXSEA
    3.280
    -0.080
    -2.4%
  • WAIT.USA
    126.000
    5.000
    4.1%
  • ITVI.USA
    12,124.580
    -525.260
    -4.2%
  • OTRI.USA
    27.850
    -0.080
    -0.3%
  • OTVI.USA
    12,070.710
    -528.180
    -4.2%
  • TLT.USA
    3.080
    -0.150
    -4.6%
  • TSTOPVRPM.ATLPHL
    2.890
    0.260
    9.9%
  • TSTOPVRPM.CHIATL
    2.930
    -0.150
    -4.9%
  • TSTOPVRPM.DALLAX
    1.280
    0.100
    8.5%
  • TSTOPVRPM.LAXDAL
    3.000
    -0.210
    -6.5%
  • TSTOPVRPM.PHLCHI
    1.750
    0.120
    7.4%
  • TSTOPVRPM.LAXSEA
    3.280
    -0.080
    -2.4%
  • WAIT.USA
    126.000
    5.000
    4.1%
NewsTechnology

Future view: United Road bets on Uptake predictive maintenance

AI and machine learning help car hauler avoid breakdowns

Predictive maintenance is comparatively new in commercial trucking. But one heavy-haul fleet booked a 400% return on investment compared to the costs of on-road breakdowns.

Using customer and industry data from telematics, repair orders and fluid analysis, Chicago-based Uptake uses artificial intelligence and machine learning to pinpoint when a component or system is going to break down or wear out. 

“This technology has held promise for a number of years,” Braden Pastalaniec, Uptake vice president of Transportation, told FreightWaves. “We’re here to say to the market that this technology is real. It’s being validated by a lot of the top fleets in the U.S. It’s generating an immediate value.”

Uptake integrates 143 predictive models for failure modes with transportation management systems from Trimble, Geotab and others. The forensics are available across a mixed fleet with a single login. They run in the background of telematics and TMS systems.

“We positioned ourselves as the layer of intelligence on the systems in place with our customers today,” Pastalaniec said. Maintenance insights and recommendations come from data already being collected.

Artificial thought process

AI is an oft-touted term with scant definition. In truck maintenance, it can mimic a technician.

“AI in its simplest form is training a computer to replicate a brain function with some level of intelligence, Pastalaniec said. 

A battery failure makes a good example.

“If I put a three-month trend of cranking voltage that’s captured every five seconds in front of a maintenance expert or a technician, they’re able to tell me that battery is going bad or that it’s healthy,” he said. “What we do is train our machine-learning models to do that at scale and in real time.

Predictive maintenance began in the airline industry. Potential maintenance issues logged in flight are addressed when the plane lands. In the vehicle space, General Motors’ OnStar unit began offering predictive maintenance for fuel pumps, batteries and starter motors on some models in 2016.

“We really see it as a mature offering,” Pastalaniec said. “Our customers can start receiving value within a few weeks time, as opposed to a few years ago, when it was more of a science project.”

United Road experience

United Road is a finished vehicle logistics business based in Plymouth, Michigan. It tried Uptake’s Radar product for two months on 100 of its 2,500 company-owned trucks in two locations. Uptake watched for more thhan a half-dozen failure modes in the diesel after-treatment system. 

That included the accumulation of carbon on the inlet side of the diesel oxidation catalyst, called face plugging. A buildup can restrict exhaust flow. That hinders the performance of the diesel particulate filter and selective catalyst reduction.

“The cost we incur as an organization when a truck breaks down extends well beyond the cost of the roadside repair,” Jason Walker, United Road chief operating officer, told FreightWaves. “The cost of sidelining a truck for maintenance affects the entire supply chain — from our drivers to the dealerships and ultimately to the end customer.”

Uptake identified a potential nitrogen oxides (NoX) sensor failure 21 days before a United Road truck was brought in for service.

“Once the truck was at the shop, the technicians did the full diagnostics and confirmed that the NOx sensor was failing and replaced the part, ultimately avoiding that road call,” Pastalaniec said.

Predicting the root cause

In another case, Uptake predicted a coolant leak in a United Road truck in which the driver repeatedly topped off the fluid, treating a symptom instead of a root cause.

“I think that speaks to the ability to monitor the condition of these components remotely and not having to rely on the driver or a mechanic at a truck stop,” Pastalaniec said. “Being able to have your finger on the pulse from a computer screen really creates benefit.”

Walker said Uptake Radar works with United Road’s TMS. He has ordered the software for 100 more trucks at two more locations.

“When we weigh all of the pertinent factors and stakeholders involved when a truck breaks down, the value of an offering such as Uptake may be in excess of our calculated 400% return,” Walker said. 

Uptake taps into Geotab sensors

Artificial intelligence is critical in today’s workflow solutions

Commentary: Optimizing a truck fleet using artificial intelligence

Click for more FreightWaves articles by Alan Adler.

Tags

Alan Adler

Alan Adler is a Detroit-based award-winning journalist who worked for The Associated Press, the Detroit Free Press and most recently as Detroit Bureau Chief for Trucks.com. He also spent two decades in domestic and international media relations and executive communications with General Motors.

Leave a Reply

Your email address will not be published. Required fields are marked *

Close