NS Derailment, East Palestine, Ohio - 3 Feb, 2023 - Engineering Aspect

[Astronuc]

Here I want to focus on the technical matters of train operation, defect detection, the accident (derailment), which is currently under investigation (by NTSB and FRA) and the investigation itself.

https://www.ntsb.gov/investigations/Pages/RRD23MR005.aspx

The derailment became significant when tank cars of vinyl chloride released the flammable contents and caught fire. One of the tank cars heated to the point where an exothermic polymerization reaction was apparent and an explosion anticipated, which precipitated a 'controlled' release into the environment, and burning of the vinyl chloride, which produced HCl and phosgene in the air. Other freight cars, some carrying plastic, also caught fire.

Preliminary report - https://www.ntsb.gov/investigations/Documents/RRD23MR005 East Palestine OH Prelim.pdf

Summary:

On February 3, 2023, about 8:54 p.m. local time, eastbound Norfolk Southern Railway (NS) general merchandise freight train 32N derailed 38 railcars on main track 1 of the NS Fort Wayne Line of the Keystone Division in East Palestine, Ohio. (See figure.) The derailed equipment included 11 tank cars carrying hazardous materials that subsequently ignited, fueling fires that damaged an additional 12 non-derailed railcars. First responders implemented a 1-mile evacuation zone surrounding the derailment site that affected up to 2,000 residents. There were no reported fatalities or injuries. At the time of the accident, visibility conditions were dark and clear; the weather was 10°F with no precipitation.

I expect the report will eventually indicate the mile post (MP) at which the derailment occurred. It happened east of the town.

Train 32N comprised 2 head-end locomotives, 149 railcars, and 1 distributed power locomotive located between railcars 109 and 110. The consist included 20 placarded hazardous materials tank cars transporting combustible liquids, flammable liquids, and flammable gas, including vinyl chloride. Train 32N was traveling about 47 mph at the time of the derailment, which was less than the maximum authorized timetable speed of 50 mph.

The train was still traveling near maximum allowable speed at the time the 23rd car passed MP 49.81, so they hadn't really slowed the train at that point, and head end is probably more than 1200 ft (366 m) beyond that point. I'd estimate derailment about 2 miles from the detector. Hard to tell without knowing the deceleration. Freight car lengths vary, typically 53-65 ft (16-20 m), but some tank cars are shorter if they carry dense commodities like sulfuric acid. The length of the vinyl chloride tank cars were on the order of 53 ft (16 m) coupler to coupler.

So 38 cars derailed starting with the 23rd freight car through the 60th car (and may be the 61st as well). Five derailed DOT-105 specification tank cars (railcars 28–31 and 55) carrying 115,580 gallons of vinyl chloride (about 23,116 gal per tank car)

On the Fort Wayne Line of the Keystone Division, NS has equipped their rail network with HBD systems to assess the temperature conditions of wheel bearings while en route. The function of the HBD is to detect overheated bearings and provide audible real-time warnings to train crews. Train 32N passed three HBD systems on its trip before the derailment. At MP 79.9, the suspect bearing from the 23rd car had a recorded temperature of 38°F above ambient temperature. When train 32N passed the next HBD, at MP 69.01, the bearing’s recorded temperature was 103°F above ambient. The third HBD, at MP 49.81, recorded the suspect bearing’s temperature at 253°F above ambient. NS has established the following HBD alarm thresholds (above ambient temperature) and criteria for bearings:

• Between 170°F and 200°F, warm bearing (non-critical); stop and inspect
• A difference between bearings on the same axle greater than or equal to 115°F (non-critical); stop and inspect
• Greater than 200°F (critical); set out railcar

After the train stopped, the crew observed fire and smoke and notified the Cleveland East dispatcher of a possible derailment.

NTSB/FRA will have to look at these restrictions in light of an axle that failed catastrophically before the train was stopped. Tank car design will probably be considered as well.

A posted video shows sparks from the wheels (IIRC, the lead truck/bogie) of the 23rd car, a covered hopper car. One narration claimed the video was taken about 20 miles before the derailment, which would put the video near the HBD detector at MP 69.01. I don't know if that is true.

More details will come as information is verified.HBD = Hot Box Dectector (old term) or Hot Bearing Detector (modern term). Hot box originated when freight car trucks had journal (friction) bearings and the boxes formed in the cast steel frame held lubricant. Modern freight trucks (wheel sets) use sealed roller-bearings. I used to help my grandfather oil/lubricate journal bearings on freight cars back during the 1960s, before such equipment was phased out for roller bearings.

Maps care of USA Today
https://www.usatoday.com/story/grap...railment-ohio-explained-graphics/11251530002/

For those interested in Code of Federal Regulations on tank cars
https://www.govinfo.gov/content/pkg/CFR-2012-title49-vol3/xml/CFR-2012-title49-vol3-part179.xml

Some vinyl chloride tank cars may carry nearly 26k gallons.
https://www.gbrx.com/railcars/25800-gallon-vcm-pressure-tank-car/

 

[Astronuc]

At MP 79.9, the suspect bearing from the 23rd car had a recorded temperature of 38°F above ambient temperature. When train 32N passed the next HBD, at MP 69.01, the bearing’s recorded temperature was 103°F above ambient. The third HBD, at MP 49.81, recorded the suspect bearing’s temperature at 253°F above ambient. NS has established the following HBD alarm thresholds (above ambient temperature) and criteria for bearings:

• Between 170°F and 200°F, warm bearing (non-critical); stop and inspect
• A difference between bearings on the same axle greater than or equal to 115°F (non-critical); stop and inspect
• Greater than 200°F (critical); set out railcar

NTSB/FRA will have to look at these restrictions in light of an axle that failed catastrophically before the train was stopped. Tank car design will probably be considered as well.

An electro-pneumatic braking system would not have prevented the derailment, unless the above guidelines were different, and/or the crew responded more cautiously to an apparent increase bearing temperature.

Looking at the timeline, the bearing was warm at MP 79.9. Temperature increased from 38°F to 103°F over the 10.1 miles to MP 69.01. That did not exceed the threshold of the guidelines "difference between bearings on the same axle greater than or equal to 115°F", but difference occurred over 10 miles; the distance to the next detector (MP 49.81) is about 20 miles. One could interpolate (103 - 38)/(79.9-69.01) and extrapolate multiply by distance (~20 miles) and add to 103°F (119+103 = 222°F) to show that the defective bearing would exceed 200°F. So they should have stopped the train well before MP 49.81, i.e., after passing the detector at MP 69.01. Only that course of action would have prevented the derailment.

At MP 49.81, the detector recorded a temperature of 253°F according to the preliminary NTSB report. However, 1) the bearing had already failed before it got to the wayside detector at MP 49.81 (as evidenced by a video of sparks flying), 2) train was still traveling at 47 mph (at the time of derailment), and 3) the crew started slowing the train in earnest after the 3rd HBD warning (which came after the train derailed). The train went into emergency braking (automatically) because the train had already derailed.

I'm still trying to reconcile the timeline with various statements from NTSB and other sources. I'm sure the crew feels pressure to keep the train going - regardless. Once the train derailed, they had no choice. Likely the damage to the axle/truck made safely stopping the train (without derailment) and the rapid deceleration (from the emergency braking) exacerbated the situation. Air brakes have a delayed effect, and the question remains concerning the third locomotive (DPU (distributed power unit) between cars 109 and 110); did it brake as the lead units did, or was there a delay? In other words, was the DPU still pushing the train at the time of the emergency application and initial derailment of car 23?

DPUs have long been used on mountain grades for at least 5 decades, and I know of DPUs from the 1960s (Harris Locotrol). The SP, DRGW and MILW definitely used DPUs.

 

[berkeman]

The temperature numbers in the bulleted items don't explicitly say "above ambient". Is it implied, or are they mixing temperature numbers?

 

[Astronuc]

The temperature numbers in the bulleted items don't explicitly say "above ambient". Is it implied, or are they mixing temperature numbers?

Don't know, but probably implied (I'd have to see the NS rule book), since the bearings would be hot during the summer and cold during winter. It's the difference that is important.

Note that the second (middle) criteria (A difference between bearings on the same axle greater than or equal to 115°F) is the differential temperature between bearings on the same axle, i.e., the other side, assuming both bearings don't fail simultaneously. As far as I know, the temperature of the other bearing has not been published, but I haven't looked in detail. Nevertheless, in my opinion, the crew waited too long to stop the train and inspect.

Apparently the FRA is issuing an advisory on HBD and data produced by HBDs.
https://railroads.dot.gov/sites/fra.dot.gov/files/2023-02/(Safety Advisory 2023-01 Posting on-line 022823).pdf

They are probably reviewing the detectors and whether or not they were functioning properly.
Railroads urged to examine track detectors after Ohio crash
https://apnews.com/article/ohio-train-derailment-railroad-safety-df7ea511b8cc4f34c5f602be69a14e39Edit/update: The sentence before the bulleted items states, "NS has established the following HBD alarm thresholds (above ambient temperature) . . . " and the FRA draft advisory repeats the phrase. It would be a challenge to set a limit on absolute temperature, so it's appropriate for a relative temperature above ambient. I imagine the bearings on the sunny side (south facing, or E/W depending on morning or afternoon) would be warmer that the other side. Summer time temperature will be hotter than winter temperatures. So, relative to ambient, or differential with respect to the colder side, is appropriate. Acoustical signalling monitoring would be nice, but that more difficult and prone to false positives.

 

[Astronuc]

NTSB launches investigation into Norfolk Southern, urges them to improve safety ‘immediately’
https://www.yahoo.com/news/ntsb-launches-investigation-norfolk-southern-211732860.html

The derailment due to a failed bearing and NS response is one issue. The derailment near Springfield, OH, is yet another issue. I'm still waiting to review the preliminary NTSB report. From limited video, I'd key on the coil flat cars - and their wheels, bearings and trucks, and draft gear. It's hard to tell if the trailing of three coil flat cars had a failure, or if the preceding car had mechanical failure, as both cars lurched to right at the grade crossing. Since subsequent cars traveled over the crossing, seemingly without a problem, one of the flat cars had a problem, or they cleared whatever problem that caused them to leave the rails.

https://www.ntsb.gov/news/press-releases/Pages/NR20230323.aspxEdit/update (Thursday, 9 March)
ABC, GMA -Feds looking into Norfolk Southern's handling of additional reported hazmat concern weeks after East Palestine
https://www.yahoo.com/gma/feds-looking-norfolk-southerns-handling-100150129.html

Another train with hazardous/dangerous cargo (ethanol and propane) received an indication that a freight car (#32) was 'trending hot', i.e., a wayside detector detected a warm or hot bearing. Car #32 was a tank car carrying ethanol.

The complaint alleges that the conductor, now concerned that the "trending hot" warning could lead to an overheated wheel, radioed the desk back and suggested that they stop the train and inspect it. But the dispatcher overruled the crew and urged them onward.

Meanwhile, a maintenance worker in the train's vicinity allegedly overheard the radio chatter and offered to observe the train as it passed by. The complaint states that when the worker reported that he hadn't witnessed any smoke, the crew was told to keep going some 40 miles south to Winston-Salem, North Carolina.

Ultimately, the train was able to complete its trip without further incident. But the Federal Railroad Administration is now looking into the previously unreported Feb. 27 incident as part of a broader "safety assessment" of Norfolk Southern, a spokesperson confirmed.

A video associated with the article had some statistics on derailments: 818 in 2022 alone, and 414 rail cars carrying hazardous/dangerous cargo damaged or destroyed. I don't know where the data originated or from where.

When I checked the article after I inadvertently closed the page, a new video gives an update on the East Palestine derailment. Importantly, it gives a time line of NS Train 32N, which passed a security camera at 8:12 pm (20:12). The camera showed sparks and fire under the car that eventually derailed. That car has passed the 2nd wayside detector, which indicated a warm bearing. The train derailed at 8:54 pm (20:54) after passing a third wayside detector that indicated a temperature (254°F above ambient), which prompted the train to start slowing down. At that point however, the bearing had failed and that freight car had derailed. The train slowed, then went into emergency braking, which probably exacerbated the derailment (i.e., the locomotive and freight cars foward of the derailed car slowed/stopped, while the following cars continued to move ahead piling up). This kind of buffing happens when some cars brake harder than the following cars, as there is a delay in air line running the length of the train. Leading cars brake before following cars, unless a DPU locomotive or rear-end system releases the air upon an electronic signal from the lead locomotive.List of derailed freight cars
https://www.epa.gov/system/files/documents/2023-02/TRAIN 32N - EAST PALESTINE - derail list Norfolk Southern document.pdf

https://www.epa.gov/oh/east-palestine-ohio-train-derailment-emergency-response

 

[berkeman]

And now the investigations have found issues with certain train cars having "loose wheels" associated with certain wheelsets. Yikes.

https://www.cnn.com/2023/03/10/us/springfield-derailment-norfolk-southern-safety-issue/index.html

 

[gmax137]

Several of the news stories I've seen over the last weeks have mentioned around 1000 derailments per year. That would be 2 or 3 per day - yet the news stories don't dwell on this.

with the article had some statistics on derailments: 818 in 2022 alone, and 414 rail cars carrying hazardous...

This would seem in agreement with the "about 1000" number, but it makes me wonder if they're counting number of incidents or number of derailed cars?

 

[Astronuc]

This would seem in agreement with the "about 1000" number, but it makes me wonder if they're counting number of incidents or number of derailed cars?

I think the 1000 is a ballpark figure. As one mentioned, 818 derailments ~ 1000 with a coarse rounding up. Then again, I found different numbers, so I don't know the range of time for the 818 number.

So, I went to the FRA website, found the statistical tool and chose derailments, which could be one set of wheels on one car, or several cars and/or locomotives, from all causes. By calendar year, the total number of derailment incidents from all causes is:
CY 2019 1,340
CY 2020 1,114
CY 2021 1,095
CY 2022 1,164

https://railroads.dot.gov/accident-...ccident-reports/209-train-accidents-and-rates

 

[ChemAir]

Based on my experience working at a HHC manufacturer, I believe these are quite common.

I've personally had to deal with three derailments at a pretty small manufacturer over a fairly small number of years. I think I have secrecy agreements that protect any type of communication on the issue, scope or resolution for the crashes, but I will say....

Only one of the three made the news. And it wasn't a good thing, but mostly because the local Fire Department would not listen to handling requirements. They proposed something we said would be a bad idea and they did it anyway, and blamed us.

Generally, most transportation accidents seem to happen with solid materials at logistics terminals--broken bags, damaged pallets....

 

[berkeman]

NS seemed to be resisting compensating the victims in the toxic derailment, and now the Ohio state Attorney General has stepped in...

https://www.cnn.com/2023/03/14/us/ohio-norfolk-southern-lawsuit/index.html