A Game Changer for the Trucking Industry.

As Frost and Sullivan’s global director of commercial vehicle research said at the Commercial Vehicle Outlook Conference (see September 2 blog) one of the biggest opportunities on the horizon for trucking fleets and equipment operations is the ability to manage downtime. Some call it prognostics, some call it condition-based maintenance (CBM) but what it means is that problems that take equipment out of service can be identified early and managed. That minimizes surprises, manages costs and avoids service failures.

Roadside repair and towing are expensive and disruptive, result in service failure and cause driver dissatisfaction. Preventive maintenance is how fleets attempt to avoid breakdowns today.
But new maintenance thinking is moving towards predictive repairs, exchanging the traditional model of preventive maintenance, which is time- or mileage-based, with one that is based on evidence of need.

Traditional PM and repair done at the home shop minimizes expenses and allows for planned equipment downtime. But a better way is to use predictive tools to measure the life of components and predict failures so equipment isn’t over-maintained, is ready for dispatch and doesn’t break down unexpectedly.

Leading the charge in predictive, prognostic or condition-based maintenance is the military. Many successes have been realized in equipment readiness, fitness for duty and mission completion by using predictive tools to guide maintenance and repair practices. These tools are set to make their way into trucking as fleet managers realize the value of managing failure while gaining the maximum life from equipment and components.

Breakdowns happen. To avoid them, fleets pour resources into preventive measures to minimize the impact. And they pour money into preventive maintenance as well. But across America, fleets overspend on PM by $25 billion dollars a year.

That’s the number Duke Drinkard, industry consultant and ex-vice president of maintenance for Southeastern Freight Lines, told the fall 2008 meeting of the Technology and Maintenance Council (TMC) of the American Trucking Associations is his presentation, “Condition-Based Maintenance – The Next Step.”

Drinkard, who now is an active member of the 21st Century Alliance – Research and Testing – says he doesn’t recall exactly how the number was derived but emphasizes that over-maintenance of vehicles is costing the industry a fortune. “The object of maintenance should be to get 95% of the life from a component before it’s replaced,” he said in an interview recently.

Condition-Based Maintenance
CBM isn’t new. As far back as the ‘40s, the Denver and Rio Grande Railway was looking at cracks in components that presaged failure to schedule maintenance/repair.

Today, the military is having significant success with CBM and predictive maintenance. Currently, says Guy Rini, president of consulting firm GTR Development and chairman of TMC’s Tomorrow’s Truck CBM taskforce, there are two trucking programs going forward with army co-sponsorship.

Rini is excited about the potential of CBM. As a retired Mack electrical engineer – Rini was responsible for development of Mack’s V-MAC engine controller among many other successes – he knows the value of the data that is stored in the engine ECM and its ability to predict problems. He cites the starter as one.

“Fleets know a starter is good for 150,000 miles, so they routinely replace them as the truck reaches that mileage,” he says. “But what’s more important than vehicle miles is the cranking time the starter experiences. If the engine ECM can be programmed to store starter cranking time, a much more informed decision can be made when it has reached the end of its useful life.” This, then, goes a long way to address Drinkard’s objective of getting 95% of a component’s useful life.

However, while this is an improvement, it still falls short of condition-based system maintenance.

System Health Management
As Nick Frankle, director of systems health management at Frontier Technology (FTI), says, “Predictive technology doesn’t prevent failures, but it provides early warning of future failures that allows managers to decide where and when to repair before the failure occurs.“

FTI is a California-based supplier of predictive health management (PHM) technology to the military. The company’s NormNet PHM creates standard operating profiles for a piece of equipment, establishing the norm – hence the name – from the data stream taken from sensors on the vehicle. For FTI, the focus has been on aerospace applications and wheeled military vehicles.

Once established, streams of data are compared in real time to the established norm and any slight deviation is analyzed. In a complex system, like a vehicle engine, says Frankle, if one particular data stream is off the norm, others are too, and by comparing the variances, the system can predict with surprising accuracy the component responsible, the failure mode and the time to failure.

This allows for planned action. It also allows repair in a shop where the manager has control. By preventing downstream failure, predictive maintenance allows for fixes that cost hundreds of dollars instead of thousands.

The military is applying predictive technology to fighter aircraft, helicopters and ground combat vehicles. The objective is to ensure that before sending soldiers into harm’s way there is a high confidence level that the vehicle or weapons system is capable of completing the mission without an equipment failure.

This is an extreme cost of breakdown, but trucking companies have exposure too: there’s the cost of repair at a dealer compared with repair at home, towing charges, downtime, driver dissatisfaction, lost opportunity cost and even penalties for service failure that, depending on the shipper/receiver, could run into thousands of dollars per hour if manufacturing processes are interrupted.

To help people in the commercial environment understand the benefits of predictive maintenance, Joel Luna, FTI Senior Operations Research Analyst has constructed a simple calculator that attempts to quantify the payback of predictive maintenance, or prognostics.

Prognostics and Health Management
“At the Department of Defense, they’re looking for more than just CBM,” says Luna. “The interest is in management of the total asset. They’re looking for the system to predict when a failure is likely to occur so they can schedule equipment for service at a convenient time and fix the problem before it occurs.

“The requirements for predicting the time to failure isn’t the same; it varies based on the type of equipment and the operations supported by that equipment. Take, for example, a U.S. Marine Corps Light Armored Vehicle. The military wants to know if it’s going to fail before the next mission is completed. However, warnings farther into the future allow for optimum scheduling of parts, personnel and maintenance facilities. Predictive analysis, which includes failure detection, problem diagnosis, time to failure and recommended course of action, allows users to avoid bad consequences. And they find they can fix potential failure more efficiently and economically. Once a vehicle is in a maintenance facility, a wide range of detected and preventable problems can be corrected. Equally important, the data recorded after the repair will confirm that the problem has been completely fixed as the data stream reverts to the norm.”

In order to provide the very extensive data for such predictive analysis, the military reads from a lot more sensors than are available on commercial trucks. That’s a point that was raised by Duke Drinkard in his 2008 presentation: “FMCSA approached TMC years ago and wanted to know if it was possible to get predictive information or the condition of components, like steering and brakes. They were also looking for data to assist in accident reconstruction. But there weren’t and still aren’t, the necessary sensors on the truck,” says Drinkard. “There are a few sensors to tell the oil pressure, temperatures on components and system voltages. But we still can’t determine oil content or endplay in a draglink, for instance. We have to develop the sensors, decide where to set them and what to measure. Then collect the data and design the algorithms that give us results we can use.”

But FTI is pushing ahead, aligning the algorithms built for the more complex military systems with the ability to utilize the data currently available from the truck.

The electrical arena is likely the most accessible to a statistical condition-based approach. Certainly, there’s money to be saved. As Rini says, citing the starter motor example, the analysis of cranking hours to failure instead of miles allows for a far more accurate assessment of remaining life. The same approach can be made to the life of an alternator or windshield wiper motor.

In a collaborative project with Mississippi State University, researcher Roger King is looking at Vehicle Maintenance Reporting System (VMRS) data from three of the nation’s most progressive fleets to see if it’s possible to build a predictive model from the reams of data that VMRS presents.

Such programs hint at the value of condition-based or predictive techniques from systems such as NormNet in reducing the over-maintaining of vehicles that costs money in inventory, technician time, downtime, and lost opportunity.

Summary
Roadside repairs are expensive, disruptive and bad for driver morale. But traditional preventive maintenance frequently includes over-maintenance in parts replacement and lost opportunity for the trucks. So it’s also expensive and disruptive. Introducing predictive maintenance will avoid roadside repairs and at the same time get the maximum life from the truck components and systems, avoiding the expense of over-maintenance and maximizing the availability of trucks.

Predictive maintenance can lower costs and increase revenues, the goal of every maintenance manager. Prognostic technology will be a game changer for the trucking industry.