What the Rail Industry Gets Wrong About Risk

There’s a topic that keeps coming up in conversations we have with customers that I don’t think most of the industry has grasped.? Like COVID upended many conventions and norms that we had taken for granted, East Palestine has done the same for the rail industry.? The largest change, I believe, centers on risk.? Wheel and bearing derailment risk, in particular.

?

For years, railroads and their customers have relied on wayside defect detectors to identify and remove derailment risks.? The impact of these detectors on reducing derailments is unquestionable, as the charts below indicate. Both bearing and wheel derailments have decreased considerably since the FRA first started recording derailment data in 1975.

?

?

This is why the AAR and others have pointed to rail safety statistics and said, “The system is working.”? There’s no doubt that it has.? The real question is whether the system is still working. ?Zooming in on these graphs tells a different story.?

?

Since 2016, wheel and bearing derailments have been on the rise, particularly in the last couple years.? I also shared in my last article that hazmat releases from bearing derailments have changed: From 2010 to Oct 2022, there were only 2 hazmat releases from bearing derailments.? From Oct 2022 to Nov 2023, a span of 13 months, there were 4.

?

?

Since I first started selling our predictive wheel and bearing sensor all the way back in 2019, I’ve heard, “Yeah, but bearing derailments aren’t that big of a deal.? What about track?? Those are more significant, right?”? Track derailments are significant, and the industry should continue the great work that’s been done with expanding the use of autonomous track inspection cars, among other things. ?However, I think NS would want a word on whether a bearing derailment can have a significant impact.

?

The FRA seems to be focused on track as well.? For the recently announced research projects funded by the FRA for 2024, 53 projects were related to track while only 2 involved wheels.? (One of the two wheel projects was $800k in funding for “Phase 2” of a project to simply produce a better tool for the manual inspection of railcar wheels in a shop).? There were no projects related to understanding, predicting, or preventing bearing derailments.

?

For bearings at least, I believe there is a compelling argument as to why the existing wayside defect detection system has proven unable to make any further progress.? From our white paper, “Bearing Derailment Prevention,” we came to the following conclusions:

1. Manufacturing/metallurgical changes in railcar bearings combined with more hot box detectors (HBDs) reduced derailments between the 1980’s and the 1990’s.
2. Warm bearing trending in the late 90’s, the introduction of the M-976 standard, the shorter Class K bearing, and replacing steel cages for polymer cages in bearings all contributed to the second drop in bearing derailments in the 2000’s.?
3. The root cause of bearing failure has changed.? In 1993, 51% of bearings had a temperature-indicated failure mode while 24% had a vibration-indicated failure mode.? In 2021, that split had changed to 7% for temperature and 83% for vibration.?
4. The HBD network reached its saturation point years ago. A 2017 study by TTCI (MxV) found the optimal spacing of HBDs was 15 miles.? CN, for one, was pushing 12 mile spacing in 2007.

?

Will the AAR’s continued focus on expanding the wayside defect detector network lower the risk of an East Palestine derailment occurring again?

?

The fact of the matter is that the rail industry, not just the railroads, has outsourced the risk of another East Palestine-type release to a safety system that appears increasingly unable to eliminate that risk.

?

A Different Kind of Crisis

I stand by what I said in my last article: There isn’t a safety crisis in the rail industry, at least as measured in general negative outcomes like mainline derailments.? The chart I posted in my last article shows the data don’t support that conclusion.? Instead, we have a different crisis.? This is a crisis of complacency.? I’ll share a few anecdotes and observations to try and explain why I believe that to be the case.

?

Now, I don’t believe that there are any malicious actors in the industry deliberately putting “profit over people.”? That’s a lazy argument.? The reality is that in an industry with historically horrible insurance losses, accidents are baked into the cake.? They’re expected.? As long as there’s some marginal improvement, everyone’s happy.

?

Managing to marginal improvement is what mature companies do, and fewer companies are more mature than those that started nearly 200 years ago like railroads and their industrial customers.? As one Fortune 50 CEO of an industrial company told someone I know in an off-the-record conversation, “Safety, Environmental…those kind of programs could suck up all the time and money you have chasing after some goals.? We try to allocate some amount to those efforts just to keep things going, but we’re not going to break the bank on them.”

?

I’ve stated many times that I haven’t found a correlation between the adoption of PSR and negative safety outcomes.? However, what I have heard from several folks working in Class 1 mechanical groups is that cuts from PSR left them with too few people to tackle everything.? Whoever was running the mechanical desk at NS during the East Palestine derailment was working on another issue at the same time.? Did PSR cut too deep?? Are the Class 1’s letting more wheelset maintenance go simply because they don’t have enough people to actually do the repairs that are flagged by their wayside system?

?

In response to the East Palestine derailment, one Class 1 executive told me, “What NS messed up is how they managed the PR of the incident.? When I heard about the incident my first thought was ‘meh.’”? Expected.? Just another derailment.? Most derailments may just be a couple wheels off the track, but have we become desensitized to them?

?

There’s always going to be risk in anything we do.? By virtue of putting that railcar on the track, we invite risk.? By trying to make a buck on whatever we ship in that railcar, we operate under a tension between doing it faster/better/cheaper and doing it safer/more carefully/more responsibly.? That will never go away.? The risk will always be there, but it’s clear the nature of that risk has changed.

?

How Risk Has Changed

AAR Rule 1.a states, “Each handling line is responsible for the condition of all cars on its line.“? Shippers have long pointed to this rule and said, “I don’t carry the risk here, the Railroad does.”? That is true to an extent, at least until a major derailment like East Palestine occurs.? Yes, NS is cutting checks to car owners for damaged railcars, but the liability from East Palestine goes far beyond that.? (Despite Rule 1, NS is going to try and find any way to get money out of the car owners and shippers involved in East Palestine).

?

Car owners like GATX and UTLX are being sued by NS, which will really put to the test exactly how far Rule 1 will go in protecting car owners and shippers from liability.? Those car owners and Oxy, the shipper of the vinyl chloride, are the subject of the NTSB investigation.? This has resulted in the disclosure of a lot of their practices, such as how they handle maintenance, inspection, and more.?

?

I don’t know Oxy, but I do know GATX and UTLX run a pretty tight ship with well respected operations.? I doubt the investigation is going to dig up a whole lot on them, but you never know once everything gets put under the microscope.? Even the most disciplined operation can have some errors that lead to some pretty nasty outcomes in front of a jury, if it were to get to that point.? Those personal injury lawyers are circling like sharks right now.? It didn’t take them long to start filing lawsuits against CSX after their Nov 2023 bearing derailment that resulted in the burning of sulfur and the release of sulfur dioxide gas. ?Count on more lawsuits from future derailments, especially if they can get a big verdict out of it.

?

More specific to shippers, Oxy’s SDS (safety data sheet) for vinyl chloride monomer, which they wrote, is being called into question.? This incident has even stirred up consumer advocacy groups calling for the elimination of PVC, the end product of vinyl chloride monomer.? The Wall Street Journal recently reported that pressure from well-funded advocacy groups was a big factor for the Biden Administration in terminating LNG export terminal permits.? Like NS, just having your name in the papers tied to those pictures of big black clouds destroys a lot of goodwill.

?

If East Palestine hadn’t happened, few would have even heard of vinyl chloride monomer, let alone try to organize regulatory action seeking to ban it from the market.? From my read of the business, these few tank cars being shipped from the Gulf Coast to PA was likely a legacy business that, while no barn burner, was filling a small market need.? Did Oxy realize how much risk they were carrying by continue to ship by rail, rather than migrate this business to an integrated site?

These are all risks for the industry.? They’ve always been there, but now there’s going to be a picture and a dollar amount attached to them.? Personal injury lawyers are salivating at potential verdicts in their favor.? Ignore them at your own risk.

?

How Big is the Problem?? A Tale of Two Bearings

I’ll finish with one example from a wheelset we were monitoring that a customer pulled.? The pictures below show what we found when we cracked open the first bearing.? You can see from the blackened rollers and scorch marks on the cup raceway that this bearing got hot enough to bake the grease, yet not hot enough to raise a high temperature alert, such as from an HBD.?

?

The other bearing on the same axle shows a bearing with a repaired spall.? This is a reconditioned bearing.? It was crazy to me coming from the plant floor into the rail industry, but the AAR allows bearing defects up to a certain size (0.14 in^2).? Our system was able to detect this and another spall at a total defect size of 0.09 in^2.? (For comparison, the earliest an acoustic bearing detector can see is anywhere from 4-30 in^2, so long as there isn’t a bad wheel, or a braking event, or any other loud ambient noise to throw it off, which was likely the case in the East Palestine railcar.

?

Fortunately, the bad bearing we found was on a shortline that rarely exceeded 25 mph.? Had it changed hands or made its way onto a mainline…

There’s a lot more risk out there than I think the rail industry understands. Based on the way the bearing data is trending, the next East Palestine may not be that far away.