The Rail Safety Super Issue

My apologies for not posting last week. I've wanted to dig into the two-man crew mandate for some time as it seems to be the key sticking point in the Rail Safety Act. Once I got into it, however, I found a lot of material. This article is going to be long and there are many charts, but hopefully it can help rail safety pros and policy makers think about rail safety in a new way.

But first, let's talk Quantum.

BNSF and JB Hunt Go Quantum

BNSF and JB Hunt announced a new “premium” intermodal offering this week.? From the Journal of Commerce ,

?

"BNSF Railway and J. B. Hunt Transport Services Tuesday said they have launched a new premium service called Quantum Intermodal, designed for shippers who have time-sensitive cargo that has traditionally been hauled via trucks because of inconsistent rail service.?

"Quantum will offer shippers priority loading and unloading of containers, priority drayage, and a team of J. B. Hunt and BNSF operations staff in Fort Worth, Texas, who will pay special attention to all loads. The Fort Worth office is housed within BNSF’s corporate headquarters.?

"Tom Williams, BNSF’s group vice president of consumer products, told the?Journal of Commerce?that 'eyeballs will be on Quantum loads' at all times, which will ride on BNSF’s standard and expedited train service.?

"'We’re going after shippers with customer-direct freight where the cost of failure is high and they’ve not wanted the risk of failure within intermodal,' he said. 'We know shippers who have wanted 95%-plus reliability have avoided intermodal, so the new teams will manage loads where the cost of failure is high and manage out the variability to deliver truck-like on-time performance levels.'"?

BNSF and J. B. Hunt believe that Quantum loads will have a 95% on-time performance standard. Pricing will likely be more expensive than standard intermodal service but cheaper than truckload in most cases.?

?

The move appears to be a no-brainer, as long as the underlying assumptions are correct.? The first assumption is that inconsistent, sub-95+% on-time reliability has removed a number of potential carloads from the market.? BNSF and JB Hunt believe that number is 7-11 million truckloads that would convert to Quantum if they can get that service reliability correct.

?

The second assumption is that pricing discounts alone are not enough incentive to attract business to intermodal.? That seems self-evident.? Volumes have steadily decreased since the 2018 peak, with the exception of the one-time COVID surge in 2021.? Average revenue per car increased dramatically in 2022, which indicates a possible mix of more aggressive pricing and/or a dropoff in the lower-priced intermodal relative to the other products lumped into BNSF’s “Consumer Products” segment. ?

?

(I didn’t include 2023 volumes, but whoo buddy, they’re feeling the Schneider effect and general intermodal recession with volumes down double digits y/y).

?

?

A more deeply integrated BNSF-JB Hunt service has all the right ingredients for success.? What I gather from the various articles, Quantum will be a combination of (1) High touch customer service where BNSF and JB Hunt employees will be co-located to address any issues in real-time, (2) Priority intermodal yard service, and (3) Priority drayage.

?

The first point fits my thesis that railroads need to grow their service stack, as I wrote about in "The Consistent Inconsistency of Rail Shipping ." Simply having a cheaper product than trucking doesn’t get the job done.? It never has.? We have the volume today that just being cheaper unlocks.

Quantum fixes this issue by offering high touch customer service where a dedicated team functions like white-on-rice whenever an issue arises.? Although not mentioned specifically, one would hope the Quantum team would incorporate an analytical element to obsessively analyze and eliminate any and all inefficiencies from the service once they become known.

?

The other two points will be interesting to see how they play out.? I’ve not operated an intermodal yard so I’m curious as to how those guys think about priority units. ?Will cherry picking out the priority boxes to fill the consist increase the time to load a train? ?Priority units also fly in the face of the PSR no-one-is-special handling of carloads.? I’m assuming you can solve this partially with a deeper integration with JBH that results not only in better visibility but also in better pre-consist planning, if that’s even a thing.? I have no idea as this isn’t my area of expertise – maybe it’s all a non-issue.

?

Deeper integration also solves the main issue with delays: interchange.? Quantum works great moving LA-Chicago, but doesn’t really work LA-NY.? From the Trains article ,

?

"No interline service is planned, but he wouldn’t rule it out if there were customer interest and a way to make it meet tight on-time performance standards. “We can certainly serve eastern markets with the drayage resources we have today, and we’re going to be constantly looking for ways to solve for our customers,” he says.

"BNSF directly serves Atlanta and Northwest Ohio via haulage rights agreements with CSX Transportation, and shipments to those terminals are not considered interline moves."

?

If BNSF and JBH can pull this off, that’s a win for the industry and a model for growth going forward.? It’s certainly refreshing to see a Class 1 take some concrete steps to grow volumes rather than cross their fingers and do the same thing they’ve been doing for years.

?

What Are We Trying To Accomplish in Rail Safety?

I’ve seen a few opinion pieces written both for and against the two-man crew mandate that’s a component of the Brown Vance Railway Safety Act.? Since this provision is the main sticking point in that bill, I was curious if there was any merit to either side of the argument.

?

As I dug into the issue, nearly all of what I found were just anecdotes.? That’s all well and good, but anecdotes are not a great basis for sound policy.? In fact, in the rush to "Take Action" after East Palestine, I think we started mixing up two different questions:

1. Does the rail industry have a safety problem and, if so, what needs to done from a legislative or regulatory perspective to correct the issue?
2. Why did East Palestine occur and how do we insure that never happens again?

?

Some may believe these two questions are linked, that the East Palestine derailment is indicative of a more systemic issue that came about with the PSR-driven headcount reduction.? That’s plausible, but what do the data say?? Let’s examine each question more closely.

?

Question 1: Is the rail industry safe?? The following series of charts show the relative change in total derailments since 2015.? I’ve also highlighted when some of the railroads began implementing PSR.

?

?

To better understand why derailments are occurring, let’s first break the data down into the FRA’s classification codes.? There are five classification categories: Equipment, Human Error, Miscellaneous, Track, and Signaling.

The chart above shows that Human Error is the largest root cause of derailments today at 36% followed by the Miscellaneous category at 30%.? (Miscellaneous is a catchall with the most frequent and severe derailments occurring from grade crossings).

?

Derailment data also provides some indication of the hazmat impact of a derailment.? The chart below shows the number of derailments that involved either at least one hazardous car or involved an evacuation.? Again, no real increase in frequency and some relative stagnation.

?

?

Now let’s unpack these numbers further to understand and prioritize the severity of these incidents.? In terms of severity, I took the total cost of each particular derailment cause and multiplied it by the incident frequency of that cause.? I added a common correction factor just to make the numbers nice and round for a final “severity score.”? The top 10 items are shown in the table below with an associated description.? I’ve also highlighted these points in orange on a plot of cost vs. frequency.

?

(An important note on costs: The FRA collects the track and equipment damage costs from the railroad incident report. ?In cases such as East Palestine, this only amounted to ~$6m.? NS expects to spend close to $1 billion on the derailment, largely driven by environmental cleanup costs.? A more comprehensive number would need to include some of those additional costs).

?

?

The most alarming aspect of this data is that the #1 cause contributing to the most frequent and costly damage to rail is related to grade crossings by the general public.? This speaks to the severity of that issue, and one that is largely outside the control of the railroads.

?

If you remove #1 and #8 from the list, that leaves you with four track-caused derailments, 3 human error-caused derailments, and 1 miscellaneous which includes anything from rockslides to cars fouling the track.? That largely leaves track and human error as the two primary means of derailments.

?

The human error incidents are largely from yard movements.? I don’t know what any particular railroad’s rules are, but at ADM, we were required to have someone “guarding the shove” (string of moving cars) and provide radio communication from a groundman watching the shove to the guy in the locomotive cab/Trackmobile moving the cars, all in an effort to reduce the impact of coupling events.? A failure to guard the shove along with not going out to ensure a switch is thrown correctly are the three human related causes in the top 10 above.?

?

These charts do not indicate there is any safety crisis, nor do the data support the premise that PSR in and of itself has led to deteriorated safety results.? In fact, in most cases derailments decreased after the implementation of PSR.? This also flies in the face of the notion that, while plausible, reducing headcount has a negative correlation with safety metrics.? I don’t see that in the data.

?

(I’ve chosen to only include derailment data as that’s what I have readily available.? It would be interesting to see if the same pattern holds for employee recordables, injuries, and fatalities).

?

Does that mean there’s no work to be done?? Not at all.? Grade crossings remain a big issue.? Track maintenance has to continually improve.? Yard movements and, likely, employee training and support are all critical.? In one of our railyards, we had a switch that got damaged at least weekly it seemed.? It always happened on 2nd shift, which was where the new trainees often broke in.

?

Safety has to be a journey of continuous improvement.? Few acknowledge that it is also a journey that becomes increasingly difficult, requiring more time and investment once the low hanging fruit have all been picked.? If the railroads are guilty of anything it’s perhaps that they have allowed themselves to settle into relative stagnation in certain areas that could be improved upon with a bit more money, time, and effort.

?

East Palestine

Now let’s turn to question 2.? What needs to be done to prevent East Palestine from ever happening again?? The Rail Safety Act offers the following requirements:

1. Mandated wayside detector usage and 15 mile spacing of hot box detectors.
2. Expansion of hazmat train classification.
3. Better communication of hazmat trains to state and municipal governments.
4. RRs must use trained mechanics and give them no minimum time for car inspections.
5. Increased penalties for safety violations.
6. Two-man crews mandated.
7. Reimbursement of local first responders for hazmat incidents.
8. Make it easier for first responders to purchase hazmat equipment.

Will these provisions eliminate East Palestine from happening again?? I’m going to set aside #2-3, 5, 7-8.? Those may be good or bad ideas depending on the details.? They may help reduce the severity of a hazmat release from a derailment, but they don’t eliminate the derailment from occurring.? That leaves #1, 4, and 6 as the corrective actions that would prevent another East Palestine.? Let’s examine these more closely.

?

Wayside Detectors – As we wrote extensively in our white paper, “Bearing Derailment Prevention ,” increased usage of wayside detectors, and hot box detectors in particular, would result in a marginal safety improvement at best.? Studies have shown that optimal hot box detector spacing is 15 miles, but the Class 1 railroads have already had most of their network at or below that density for more than a decade.

?

Once a bearing heats up above a certain threshold, it’s already in a failed state.? In many instances, these bearings can fail less than 1 mile after a hot box detector gave no indication of failure. ?Sometimes a bearing fails even without a temperature increase.? While the existing system of detectors has helped to reduce bearing derailments from about 250/yr in 1975 to 16/yr today, it hasn’t been able to further decrease derailments in over a decade.? We discuss all the reasons why and what could be done to decrease these derailments in our white paper.

?

Car Inspection Times – I’ve heard from multiple shippers, for years, that cars aren’t getting inspected.? NS’ handling of car inspections on the East Palestine train seems sketchy at best.? However, the data do not show any dramatic increase in equipment related derailments.? From the derailment frequency chart above, there were 273 equipment-caused derailments in 2010 and only 224 in 2022.? 206 are projected for 2023.?

?

Additionally, bearing condition cannot be discerned from an external visual inspection.? Do bearing sometimes weep grease, yes.? But failed grease seals occur in <1% of all bearing failures.? It’s very, very unlikely that a visual inspection of multiple hours of a railcar bearing would have resulted in identifying the damage, unless the carmen jacked the car, pulled the bearing, and did a teardown.

?

Two-Man Crew Mandate - The most controversial aspect of the bill is the two-man crew provision.? Unions say that it ensures the safety of trains going forward.? The railroads say it’s got nothing to do with safety, with the example of the East Palestine derailment having three in the cab.? Who’s right and who’s wrong?

?

Here’s what the data say about two-man crews.? A 2019 study by Duke University used a simulation in conjunction with what sounded like good old-fashioned time studies to evaluate (1) The cognitive workload of an engineer during a normal short haul run on a Canadian shortline and (2) The impact of assistance technology on the engineer for that run.?

?

Over the course of the study, the engineer’s cognitive workload was quantified at various intervals such as startup, putting the train into motion, the course of the journey, coming into some traffic, and then putting the train into a yard.? (There’s a theory that suggests work should be performed at a 30-70% cognitive load, as evidenced but the two horizontal dotted lines in the chart below.? Below 30% a worker is more prone to distraction and boredom, above 70% they’re overwhelmed and less likely to catch everything).? The study also notes that the yard phase is the source of the highest rates of accidents.

?

?

In total, the engineer spent 20 minutes above the 70% threshold and 80 minutes below the 30% threshold in this 5-hour period.? In other words, one-third of the time is spent in a “danger” zone.

?

Two graphs about the impact of a second crew member and assistive technology on the engineer produced by the study are shown below.? “None” is an engineer alone in the cab, “All” is an engineer with a conductor and assistive technology like cruise control and PTC, “Automation” is an isolated engineer with assistive technology, and “Conductor” is a traditional two-man crew without assistive technology.? A lower penalty value indicates the engineer is closer to the 30-70% optimal workload.

?

?

In the first graph, when the engineer was experiencing a “high” workload, the best outcome came from having technology in the cab.? Having a conductor in the cab in addition to the technology added a small benefit that likely falls within the study error margin (i.e., there’s not a significant benefit).?

?

In the second graph, when the engineer was experiencing a “low” workload, the best outcome was the single engineer by themselves without the help of technology or a second person.? The worst outcome was having technology and a second person.? Why? The authors suggested that the engineer was more distracted from the various screens or by the other person, perhaps through conversation.

?

Before anyone takes the study too far, it’s important to note this is one isolated look at one particular aspect of train operations for a particular size of railroad.? It should not be taken as a hard and fast rule across all types of train operations.? There are a lot of factors that aren’t captured in this study.

?

But it does bring up some interesting points:

1. Technology does have some drawbacks and should not be considered a pure, across-the-board improvement over current operations.
2. Technology is better at handling higher volume tasks (yard operations) requiring more awareness.
3. Having a 2nd person in the cab can negatively impact performance and, therefore, safety.? Just like technology isn’t a cure-all, neither is a two-man crew.

?

If you have a hammer, everything looks like a nail.? It becomes very difficult to find the optimal situation across various geographies, demographics, consist makeup, local labor needs, etc. if the one-size-fits-all solution ties your hands.? I can certainly appreciate the perspective of the unions who feel the pressure of Wall Street drives railroads to do dumb things (it certainly does) like taking a hatchet to headcounts.? However, continuing to push the two-man crew mandate may even be detrimental to their members, even if it does secure jobs for the future.

?

From the 3 prescriptive requirements by the Brown Vance Railway Safety Act, it’s unlikely any of them or even all three together will be able to prevent another East Palestine derailment.

How To Fix It All?

At Hum we wrote “Bearing Derailment Prevention ” to educate and guide policy makers and regulators on what would actually help prevent another East Palestine.? It’s clear from the data that the wayside detection system has done all it can do, which was significant, and now a new approach is needed to move the needle towards zero on bearing derailments.? In the paper we recommend onboard monitoring as a proven technology that’s ready to scale to meet that need (noting my bias).? Onboard would prevent another East Palestine.

?

To answer the first question I posed above, no, the data do not support the notion that we have a major rail safety crisis in the United States.? Nor do I see any dramatic uptick or uncontrolled trend going in the wrong direction.? The rail industry has a lot of work to do to improve its safety performance, but is the general public best served by more regulation? ?I don’t see it.?

?

My idea of government may be quaint, but I believe regulation should exist only where the free market is unable or unwilling to address societal needs.? With a regulatory body (the FRA) that already exists to oversee rail safety, I’m not sure what more is gained by additional laws on the books?? Especially when the FRA is generally underfunded for enforcement of those rules? Unless your aim is to pad the lawyers’ pockets that is.

?

So where should we put the dollars and the elbow grease?? Here’s Byron’s 4-point plan:

First, Invest in grade crossing solutions.? I don’t know this space, but it’s Public Enemy #1 in terms of rail safety.? I’m guessing there aren’t any good solutions and you can’t prevent people from willfully choosing to ignore crossing guards.? But the severity of these incidents demands a serious plan to reduce their prevalence as much as possible.

?

Second, Invest in and reduce regulatory barriers that would prevent usage of automated track inspection technology.? I’m not talking about trying to replace bodies with machines.? I’m talking about enhancing the efforts and efficiency of the boots on the ground to be more predictive, more targeted, and more accurate in catching track failures before they occur.

?

Third, Make the Autonomous Railyard a reality.? The three human related derailments in the Severity Top 10, all came from yard operations.? One of the primary motivations for me to start Hum came because we had two fatalities from yard operations.? We need to get bodies out of the yard.? The faster we can go to fully or even semi-autonomous railyards, the better.

?

Fourth, Incentivize adoption of onboard monitoring technology.? Yes, I’m very biased on this point, but only because I had the conviction in this technology to sink all my savings into creating a company that could make this a reality.? Onboard monitoring, and vibration monitoring of wheels and bearings in particular, would move the industry from a reactive “oh shoot, stop the train the bearing is on fire!” to a proactive, “I’ve got 32,000 miles of life remaining on my bearing, I’m going to schedule maintenance for XXX day.”? Numerous additional benefits like reduced maintenance costs and GPS tracking on every railcar would come along as well.

?

Do we need Congress to legislate these solutions?? I don’t think so.? Regulations are never going to be perfect and often create as many problems as they do solutions.? Can the FRA and rail industry participants make this happen?? Absolutely, and there are programs in place in some of these areas.? If the industry could rally around these 4 areas, why can’t we start dropping derailments 10, 20, even 30% from where they are now?? I hope we do.