Overheated bearings are a common cause of train derailments.
The recent derailment of a Norfolk Southern train in East Palestine, Ohio, resulting in a toxic chemical spill was a rare but not unknown event on America’s railroads. Like every other wheeled vehicle in commercial service, rolling stock relies on anti-friction bearings. And like all ball and roller bearings, heat is the major cause of failure. For safety-critical applications in railways, thermography is used to check passing rolling stock for hot bearings, setting an alarm to alert the engineer about the problem. Modern systems can determine the exact car where a hot bearing is located, but predicting catastrophic failure is very difficult. More technology is needed to prevent a repeat of this disaster
Access all episodes of This Week in Engineering on engineering.com TV along with all of our other series.
* * *
Episode Transcript:
There’s a reason why ancient historians hailed the invention of the wheel. The mobility offered by the wheel let humanity develop global cities and societies, and trade goods to create a global economy.
For most of human history, wheeled movement was limited to the speed that oxen or horses could walk, but the invention of the steam engine and low-cost steel meant that railways could increase those transport speeds by an order of magnitude. And while people like Watt, Newcomen and Stevenson are credited with the development of railway technology, the simple fact is that what makes it all work are antifriction bearings.
The original form of antifriction bearing, the ball bearing, was patented by Philip Vaughn in 1794, conveniently just in time for the steam engine that kicked off the Industrial Revolution. And as every mechanical engineer knows, then and now, for ball and roller bearings to survive, a few basic factors must be met: they must carry no more than design loads, exclude dirt and debris, be lubricated and critically, be kept at temperatures low enough to preserve rolling element and race metallurgy.
Heat has always been the enemy, and in the recent Norfolk Southern East Palestine train derailment, the National Transportation Safety Board has issued a preliminary report stating that Norfolk Southern train 32N, consisting of 149 cars and three locomotives, suffered a serious bearing overheat condition in the 32nd car, leading to the derailment of 38 cars including 11 tank cars carrying toxic vinyl chloride.
The measurement of heat rise in bearings is standard industrial MRO practice, and is common for bearing housings, pillow blocks and today even bearing races themselves can contain sensors that relay data to maintenance software.
For rolling stock, the heat buildup only occurs when the trains are in motion, and the number of bearings involved is large. With four axles and two bogeys on each railcar, the Norfolk Southern train was hauling 596 axles, with a heat-induced failure at any one of these representing a derailment risk.
Since the bearings can’t be inspected by crews while the train moves, automated thermography is used on the rail bed to scan trains passing overhead and report hot bearing conditions. This is a technology that railroads call wayside defects detectors, or hot bearing detectors (HBDs). The systems send audible, real-time warning alerts to train crews when bearing temperatures reach pre-set thresholds.
Norfolk Southern safety standards required train crews to stop and inspect bearings between 170 and 200°F, and also stop to inspect where the difference between bearings in the same axle is greater than or equal 215°F.
At temperatures greater than 200°F, the railroad requires train crews to uncouple and set out the defective railcar. The suspect car on train 32N passed three HBDs over a 30-mile stretch near East Palestine, with bearing temperatures on the 23rd car showing 38°F above ambient temperature, then 103°F above ambient, followed by 253°F above ambient. The alarm sounded and the crew stopped the train, but it was too late to prevent the derailment.
According to the preliminary report, the train crew operated per standard Norfolk Southern procedures, and the train was operating at 47 mph, less than the maximum authorized speed of 50 mph. At this stage of the investigation, bearing detection systems operated properly, the train crew were warned in a timely manner, and they took the correct measures in stopping the train.
So how can a derailment like this be prevented in the future?
If temperature is the sole criterion for bearing failure prediction, there are very few possibilities to improve the system. More data points, through a closer spacing of detectors—which at the incident site were 10 and 20 miles apart—could give enough data points for a smart system to project time or distance to critical bearing overheat and flag the crew to stop the train sooner. Or, Norfolk Southern procedures could be amended to force the crew to stop the train at lower temperature thresholds and inspect for damage. A long-term solution might involve sensor-equipped bearings, and detection systems that include other data such as sound and vibration to predict bearing failure.
All these solutions would increase operating costs and would have to be weighed against other safety improvements which might be more cost-effective and result in greater overall rail safety, with the occasional hot bearing caused derailment.
The U.S. freight rail network runs on 140,000 miles of track, and railroads reinvest a very high percentage of revenues, 19 percent, on their systems. Was the Norfolk Southern East Palestine derailment an unfortunate outlier, or a sign of an industry-wide safety issue?
According to the Association of American Railroads, the mainline accident rate for Class One railroads like Norfolk Southern, is at an all-time low and is down 49 percent since 2000. Measured by carload, the hazardous materials accident rate is down 70 percent since 2000 and is currently the lowest-ever rate based on preliminary Bureau of Explosives data. The statistics suggest that freight rail in America is generally safe, but it’s widely expected that this accident will spur regulatory changes.
This may be a golden opportunity for application of embedded, low power and connected Internet of Things sensors to monitor safety critical components like bearings in real time. We’ll be watching for the engineering developments.