Get fresh updates from Hortonworks by email

Once a month, receive latest insights, trends, analytics, offering information and knowledge of the Big Data.

cta

Get Started

cloud

Ready to Get Started?

Download sandbox

How can we help you?

closeClose button
Hortonworks Customer
Metro Transit of St. Louis

Metro Transit of St. Louis (MTL) operates the public transportation system for the St. Louis metropolitan region. The organization’s mission is “Meeting the region’s transit needs by providing safe, reliable, accessible, customer-focused service in a fiscally responsible manner.”

Meeting the Challenge to Provide Safe, Reliable Public Transport

To ensure the safety of passengers and the proper use of public funds, MTL has always performed regular maintenance on its bus fleet. But lacking detailed data on how bus components were actually performing, the agency maintained vehicles reactively. It replaced parts after they failed, or simply bought new buses.

This supported MTL’s objective of assuring the safety of its passengers and the reliability of its service, but the team suspected that it often replaced parts or discarded entire buses even when replacement wasn’t necessary.

A lack of ongoing, reliable machine data from its buses forced a tradeoff between the two parts of MTL’s mission. “Safe, reliable and accessible” goals conflicted with its mandate to be “fiscally responsible.”

MTL decided to address that conflict by building a proactive bus maintenance program, and in 2002 the city developed the “K Plan” to consolidate its various maintenance divisions. Prior to 2002, each division reported to a different facility.

The specific goals of for this K Plan consolidation included:


  • Reduced maintenance time for buses

  • Accurate predictions for the life spans of different components

  • Accurate inventories of required parts to have on hand when needed

  • Balanced maintenance workloads so that buses could come in for maintenance at preset, predictable intervals

  • Reduced costs for parts and labor

Combining all the historical data on bus component failures

With the K Plan’s new consolidated structure, the MTL also wanted to consolidate its historical data on component failures across all buses and all divisions. The goal was to identify predictable patterns amongst all the bus parts and then derive warning signals that might indicate likely failure and the need to replace something before it broke.

However, data collected from the distinct maintenance centers was managed manually in spreadsheets. This approach was inefficient and antiquated, and the MTL systems for failure prediction began to break down—to say nothing of the buses they were supposed to maintain.

Complying with changing emissions standards

Another challenge that the K Plan promised to address revolved around understanding constantly changing emissions standards and the controls necessary to meet them. For example, in 2007 stringent emissions specifications drove up the cost per mile driven to 92 cents/mile. Subsequent emission standards for 2010 and 2012 proved to be increasingly strict. MTL needed a way to control the increasing costs of driving its buses, while still complying with the emissions standards.

Adjusting to a smaller work force

Another key challenge MTL faced was the reduction in its available labor force. A smaller number of skilled laborers meant that MTL needed a way to do more maintenance with fewer people while maintaining the same service levels.

In addressing these three challenges, MTL wanted a strategy to fit all districts and maintenance centers – they needed a “cookie cutter” solution. That solution needed to provide a consolidated view of the maintenance centers across all districts with the ability to proactively report status to the Chief Mechanical Officer.

Partnering with LHP Telematics and Hortonworks for a Proactive Maintenance Solution



As MTL developed its proactive approach to bus maintenance, they focused first on performance enhancements. They wanted to reduce the labor needed to do costly in-field maintenance, with a particular emphasis on reducing the amount of time that mechanics needed to repair a bus in the field.

To meet this goal as a critical component of the K Plan, the MTL developed a new “Smart Bus”. The Smart Bus generates machine data as it operates. That data helps the Smart Bus predict and schedule its own maintenance work. After other Smart Bus approaches stalled, MTL chose to partner with LHP Telematics (LHPT) and Hortonworks to enhance their maintenance program and bring the Smart Bus idea to fruition.

MTL currently leverages the existing electronics on a bus as well as data from all of the main bus subsystems. The LHPT system uses the existing CANbus to gather over 330 data points taken at 10-second intervals from various components. The data is compressed, transferred into the LHPT backend system, and then stored on Hortonworks Data Platform (HDP).

Using HDP, the LHPT data science team runs algorithms for pattern recognition to correlate events to component failures and weed out maintenance inefficiencies. The data scientists look at fault codes collected from the bus and then match the codes to failure events on the bus. With this process to recognize recurring patterns among the day-to-day driving events, MTL is able to better predict when a component on a particular bus will fail, which allows them to proactively service the bus and avoid having to take it out of service for a prolonged period due to unscheduled maintenance events.

Currently, over two-thirds of the MTL buses have been equipped with the LHPT and Hortonworks solution, with complete implementation expected in 2018.



The Results: Time Between Bus Failures Extended by More Than a Factor of Five



“We did not expect to get as much information as we did with access to data from HDP,” says Darren Curry, Chief Mechanical Officer for MTL. “We are always finding something new, finding things that we didn’t know about.”

MTL is now able to run highly granular reports that give their maintenance centers an unprecedented level of visibility. MTL migrates more and more data into HDP, including telematics functionality and predictive analytics. The algorithms now run automatically and can look at much larger amounts of data than they ever had before.

According to Curry this access to machine data brings 3 important benefits to MTL:


  1. It allows them to make the best business decisions possible

  2. It reduces their operating and labor costs

  3. It improves safety for their customers




For example, before implementing this solution MTL faced an average Mean Time Between Failures (MTBF) of 4,000 miles. Today with the K Plan, the MTBF has been improved to 21,000 miles and MTL estimates that with the LHPT and Hortonworks solution on the Smart Bus, the MTBF can be further extended to 30,000-35,000 miles.

With these results, MTL soon realized that they could run the buses for much longer, thereby increasing their return on investment in their fleet. Previously, buses were being retired after 35,000 miles per year at 12 years and currently MTL is able to continue using the buses up to 60,000 to 70,000 miles per year at 15 years. This is a 2x improvement on mileage and 30% increase in bus lifespan.

Originally MTL thought the K Plan would cost more to install, because it involved retrofitting their existing buses with new hardware, but the reality has proven to be quite the opposite. MTL was able to cut its original operating, maintenance and fuel costs by 50%. With the newer solution from LHPT and Hortonworks on the Smart Bus, operating costs have decreased from 92 cents per mile to 43 cents per mile, and MTL estimates that it will be able to reduce costs by another 20-30%.

The K Plan and Smart Bus have not only enhanced MTL’s maintenance program, they have changed the way MTL does business.



Next Steps: Buses Submitting Their Own Work Orders, Then Ordering Their Own Parts and Labor



Curry sees these benefits continuing to grow. “I believe that we will not only sustain our current performance, I still believe that we will be able to increase our performance with LHPT and Hortonworks another 2-3 times on top of what we have achieved,” he says. “Everything is very positive and I am really excited about how we move forward from here. This solution will give us a lot of tools to analyze data and look at things differently than what we were doing before. This really changes everything.”

MTL envisions a series of next steps to help refine the Smart Bus implementation. Using machine-learning algorithms, MTL wants to isolate the important events on its buses and create a predictive model for these events.

These predictive capabilities will allow buses to:


  • Make their own work orders with detailed troubleshooting steps

  • Order required parts and have them shipped to where the bus is housed

  • Alert maintenance crews as to when a major failure is likely

  • Report on overall health of components on a regular, proactive basis to help avoid a catastrophic failure

  • Change fueling procedures based on how much time buses spend idling



Eventually MTL would like to extend their K Plan to other fleets for vehicles such as para-transport buses and service vehicles.

Banner

We did not expect to get as much information as we did with access to data from HDP... We are always finding something new, finding things that we didn’t know about.

Darren Curry, Chief Mechanical Officer for MTL

About Metro Transit of St. Louis

Metro Transit of St. Louis provides one of the nation’s top-performing transit systems with best rankings for on-time performance, vehicle reliability and safety. MTL prides itself on its improvements in vehicle maintenance that have saved taxpayers more than $2.5 million per year by doing more with less.

Metro Transit of St. Louis is committed to making its community a better place to live and work.

About LHP Telematics

LHP Telematics is an industry leader in creating custom telematics solutions for the heavy equipment OEM marketplace, transportation, service, and construction fleets. LHP Telematics offers the most configurable end-to-end monitoring platform on the market. A deep understanding of the equipment in the field, scalable cloud services and Big Data analytics capabilities sets LHP apart from the competition.