Trains.com

Citizen hotbox detection? Is it possible?

2879 views
41 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    January 2002
  • From: Canterlot
  • 9,575 posts
Posted by zugmann on Wednesday, July 17, 2024 11:25 AM

jeffhergert
Inspections like that would take a bit more time and require actually walking instead of driving a set and/or release. Time is something the remaining forces don't really have.  They're under the gun to get things done ad quick as possible, too.

You'd also have to inpect cars, and not just block swap them and let them ride on the old airslips. 

But that goes against psr....

It's been fun.  But it isn't much fun anymore.   Signing off for now. 


  

The opinions expressed here represent my own and not those of my employer, any other railroad, company, or person.t fun any

  • Member since
    May 2003
  • From: US
  • 25,292 posts
Posted by BaltACD on Tuesday, July 16, 2024 8:02 PM

jeffhergert
Until relatively recently, you didn't hear off a bearing burning (figuratively speaking) very often. Even now it doesn't happen a lot. Still,  it seems it's happening more than it did.

It seems, anecdotally - as all experiences are, that the rise in number seems to corespond with the cutting off of mechanical staff and the rise in the use of hot bearing desks. Kind of like instead of good mechanical inspections they'll gamble that the detectors will catch a bearing in the early stages of failing, before it causes headlines in the news.

One man who was once a Carman, said they used to inspect the inside seal of the roller bearing (where it's close to the wheel) with a tool they devised. If the seal was loose, the wheelset was changed out. The loose seal would start leaking grease and eventually fail. He said they found about 8 cars a day at the yard he worked at that needed the change out.

Inspections like that would take a bit more time and require actually walking instead of driving a set and/or release. Time is something the remaining forces don't really have.  They're under the gun to get things done ad quick as possible, too.

Jeff

Simple Fix - Inspection Pit with a carman to visually look at the seals and other under car potential defects - Bad item is seen and carman pushes a button that will Paint an identity mark on the failed axle seal axle and afix a Shop Ticket.  Wasn't this a accepted procedure half a century or so in the past?

In the 21st Century an automated camera system with AI could do the function minus the manpower.

Never too old to have a happy childhood!

              

  • Member since
    March 2003
  • From: Central Iowa
  • 6,901 posts
Posted by jeffhergert on Tuesday, July 16, 2024 7:16 PM

Until relatively recently, you didn't hear off a bearing burning (figuratively speaking) very often. Even now it doesn't happen a lot. Still,  it seems it's happening more than it did.

It seems, anecdotally - as all experiences are, that the rise in number seems to corespond with the cutting off of mechanical staff and the rise in the use of hot bearing desks. Kind of like instead of good mechanical inspections they'll gamble that the detectors will catch a bearing in the early stages of failing, before it causes headlines in the news.

One man who was once a Carman, said they used to inspect the inside seal of the roller bearing (where it's close to the wheel) with a tool they devised. If the seal was loose, the wheelset was changed out. The loose seal would start leaking grease and eventually fail. He said they found about 8 cars a day at the yard he worked at that needed the change out.

Inspections like that would take a bit more time and require actually walking instead of driving a set and/or release. Time is something the remaining forces don't really have.  They're under the gun to get things done ad quick as possible, too.

Jeff

  • Member since
    May 2003
  • From: US
  • 25,292 posts
Posted by BaltACD on Tuesday, July 16, 2024 1:15 PM

I don't think any of us retirees are qualified to know what the present day 21st Century Defect Detector actually is.  Technology changes - rapidly.  What technology was when we retired, in my case 7 years ago, leave a wide gulf on what was in place then and what is in place now.

Never too old to have a happy childhood!

              

  • Member since
    December 2001
  • From: Northern New York
  • 25,020 posts
Posted by tree68 on Tuesday, July 16, 2024 12:36 PM

Overmod
Now you have a detector set up to record hot end caps.  Suppose the truck moves about 18" -- that detector is now pointing right at a piece of the wheelrim.  And if it is hot, either in active friction or because it's heatsoaked... BING!

As I envision it, you wouldn't be looking for hot end caps - you'd be reading RFID-type information from a AEI-type transducer in/on the end cap.  That data would come from internal sensors - measuring the inside race temperature.   Blend this with pure IR readings from the rest of the wheelset to come up with a picture of what's going on. 

LarryWhistling
Resident Microferroequinologist (at least at my house) 
Everyone goes home; Safety begins with you
My Opinion. Standard Disclaimers Apply. No Expiration Date
Come ride the rails with me!
There's one thing about humility - the moment you think you've got it, you've lost it...

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Tuesday, July 16, 2024 7:24 AM

Euclid
There will also be background heat production from routine air braking that heats up all metal in the vicinity, including the bearing housings and components.  How in the world can you get a meaningful measure of bearing temperature with all of this background clutter?  Whoever thought this methodology would work?

This is actually easy even with the older 'spot' type of IR detector.  The usual reason for 'hot wheels' is some kind of sticking brake, and that usually gets reported right at the first detector out of the train's point of origin.  Remember that the lion's share of these wheelsets are in three-piece freight trucks, where a considerable arc of the wheel is exposed to 'lateral view'.  The sticking brakeshoe(s) heat the wheel from the tread in, and there is a considerable thermal mass in the rim and then a constrained volume through the plate to slow down heating by the time you get anywhere near the mass of the hub and then the axle fit.

Now you have a detector set up to record hot end caps.  Suppose the truck moves about 18" -- that detector is now pointing right at a piece of the wheelrim.  And if it is hot, either in active friction or because it's heatsoaked... BING!

Later mosaic or 'staring array' detectors actually produce a moving image in IR wavelengths, so you can distinguish a hot curved wheelrim from a hot cap or other bearing-related issue.  We had AI/ES almost half a century ago that could distinguish this even without sensor fusion.

The problem remains as it was observed at East Palestine: if all the visible fire and heat is out of the focus of the detector you're using, the observed heat rise at 'axle-end level' may not trend upward fast enough, or reach a high enough level, to give the 'right' warning in time... or permit wayside alerts or 'service desks' to get the right advice to give train crews.

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Monday, July 15, 2024 12:15 PM

Overmod

I think I have mentioned this before.  When an inner bearing starts to fail, the heat produced is transferred by conduction, outward from the outer race of the bearing and casing into the sideframe, and inward to the axle, thence in toward the wheel and out toward the end cap.  

I can see how that could be a problem.  You have heat of a failing bearing being produced inside of the bearing and transferring outward by conduction and radiation, through a variety of other structure of the truck, and into the car frame. On the way out, heat is lost by conduction and radiation from the heated railcar parts, at various rates, to the unheated air and railcar parts.
 
Then the wayside detector takes a reading only on the radiant component of this flow of escaping heat being transferred by a blend of radiation and conduction. 
 
There will also be background heat production from routine air braking that heats up all metal in the vicinity, including the bearing housings and components.  How in the world can you get a meaningful measure of bearing temperature with all of this background clutter?  Whoever thought this methodology would work?
  • Member since
    March 2003
  • From: Central Iowa
  • 6,901 posts
Posted by jeffhergert on Friday, July 12, 2024 9:39 PM

SD60MAC9500

 

 
jeffhergert

 

From the explanation given on another forum, I think more acoustic detectors are what's needed. They can find a bearing that's beginning to fail before there's any heat.

Jeff

 

 

 

The problem with ABD's is they give off the same, if not more false positive's than HBD's. Train 32N that derailed in EP, OH passed 3 ABD's, giving off no warning. The other issue is that a bad bearing can also be silent. Onboard bearing detection (located on the bearing adapter) would detect vibrations. In fact a company called HUM is trialing this tech as we speak.

https://www.humindustrial.com/products

 

What was explained elsewhere is that a bearing starting to fail start making a sound that is not heard by the human ear.  (Maybe we can station K9 detectors every do often.*)  This begins 100 or 150 miles before the bearing starts really getting bad.

* I seem to recall the experimentation of using dogs to "sniff" out bad bearings when doing inspections at yards on inbound trains.  This may have been back when there were more plain bearings in use.

Today, after my last post last night, we had to stop so the conductor could go back and check a bearing getting hot.  It didn't set off the detector, the bearing desk had the dispatcher stop us and relay the car and axle needing inspection. 

The condr went back 38 auto racks, checked the indicated axle and 20 axles each side and the same on the other side of the train.  He found nothing.  The bearing desk, through the dispatcher, allowed us to proceed at 30 mph to the next detector, about 6 miles away.

After clearing it, we were instructed the temperature was dropping.  The condr and I discussed if the bearing had been hot, which he didn't find, could it have cooled while he made his way back?  It probably took 20 to 25 minutes to get back to it.  They said they would continue to monitor the car.

Jeff 

  • Member since
    December 2001
  • From: Northern New York
  • 25,020 posts
Posted by tree68 on Friday, July 12, 2024 9:05 PM

Overmod
Tree, the 'transducer' IS one of the sensors; it is a device that converts a signal from one form (e.g. acoustic) to a waveform that can then be digitized, stored, transmitted and analyzed, etc.

I'll buy that.  Again, I was thinking of two things - the operation of the low air sensor on your road vehicle (almost always triggered by the first cold snap of the season), and the idea of having the RFID as part of the axle/wheel assembly, as opposed to having to connect to the car in some manner.  

LarryWhistling
Resident Microferroequinologist (at least at my house) 
Everyone goes home; Safety begins with you
My Opinion. Standard Disclaimers Apply. No Expiration Date
Come ride the rails with me!
There's one thing about humility - the moment you think you've got it, you've lost it...

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Friday, July 12, 2024 8:11 PM

tree68
I mention the end cap only as a location for the transducer, which would be connected to sensors within the bearing. This would allow a trackside transducer to receive data, as opposed to something under the train.

Tree, the 'transducer' IS one of the sensors; it is a device that converts a signal from one form (e.g. acoustic) to a waveform that can then be digitized, stored, transmitted and analyzed, etc.

The thing that does the bidirectional transmission is a transceiver, even if it uses polyspectral light instead of radio.  Still makes better sense to put it on the sideframe instead of the rotating endcap, as it doesn't matter if anything but the specific staring IR sensor is optimized via 'machine vision' to look for endcaps and their characteristics.

  • Member since
    March 2011
  • 188 posts
Posted by dpeltier on Friday, July 12, 2024 7:26 PM

charlie hebdo

What would be the cost of EIGHT of 'your' detectors per car? Multiplied by cars in use nationwide.

What would be the cost of another East Palestine or worse?

$1.7B is the cost Norfolk Southern put on their balance sheet last I heard. This includes government fines and a class-action lawsuit settlement, as well as ongoing environmental and health monitoring. We don't know what the actual damages are and won't know for some time. My personal guess is that, had they been willing to be dragged through the mud of public opinion and forced everyone to justify their damage claims, they could have cut hundreds of millions off that settlement.

For how this compares to the cost of equipping the fleet with on-board sensors, see my previous post:

https://cs.trains.com/trn/f/111/t/294832.aspx?page=2#3442527

You're welcome to check my math. I estimated the ANNUALIZED cost of rare mega-events like East Palestine at $10M / year (for bearing failure incidents only), which given the final figure of $1.7B is probably too low. So I'll make some revisions.

Let's say that $1.7B event happens every 40 years (since I'm fairly certain there hasn't been anything comparable in the last 40 years), the annualized cost of that is $43M / year. Based on FRA data from 2018 to 2022, I estimated "typical" damages from mainline bearing failure at $20M / year, so the total potential savings from improved bearing-failure detection is $63M / year. If on-board sensors can prevent 80% of those incidents, they would be worth $50M per year.

Following the same calculations as my previous post, a sensor system that lasted for 20 years, prevented 80% of all bearing-related derailments that are happening out there today, and created 0 false positives, could have a lifecycle cost of about $125 per bearing and still break even. That's similar to the cost of replacing a tire pressure sensor in your car, and those don't last anywhere close to 20 years.

That's just the sensor! Equipping each car with the equipment to read the sensors and report it to wayside equipment or the locomotive would add considerably more cost. But even if the car has some sort of a future ECP braking system that provided a ready-made com link to each car, on-board wheel bearing sensors would not be cost effective.

And with East Palestine likely to result in mandated improvements to wayside bearing sensors, the cost-effectiveness of onboard sensors will further decrease.

Dan

  • Member since
    December 2001
  • From: Northern New York
  • 25,020 posts
Posted by tree68 on Friday, July 12, 2024 6:29 PM

Overmod
The problem with monitoring the endcap is that the detector only monitors the temperature rise at the outer face, when the temperature close to the inner bearing in the axle metal may have reached dangerous elevation.

I mention the end cap only as a location for the transducer, which would be connected to sensors within the bearing.

This would allow a trackside transducer to receive data, as opposed to something under the train.

LarryWhistling
Resident Microferroequinologist (at least at my house) 
Everyone goes home; Safety begins with you
My Opinion. Standard Disclaimers Apply. No Expiration Date
Come ride the rails with me!
There's one thing about humility - the moment you think you've got it, you've lost it...

  • Member since
    December 2015
  • From: Sharon, PA
  • 47 posts
Posted by SAMUEL C WALKER on Friday, July 12, 2024 6:20 PM

RObert Reebie's single axle Roadrailer design had a thermocouple that when broken initiated braking. IS that something for consideration? Would it be more or les effective?

 

  • Member since
    September 2002
  • From: Sterling Heights, Michigan
  • 1,691 posts
Posted by SD60MAC9500 on Friday, July 12, 2024 4:48 PM

jeffhergert

 

From the explanation given on another forum, I think more acoustic detectors are what's needed. They can find a bearing that's beginning to fail before there's any heat.

Jeff

 

The problem with ABD's is they give off the same, if not more false positive's than HBD's. Train 32N that derailed in EP, OH passed 3 ABD's, giving off no warning. The other issue is that a bad bearing can also be silent. Onboard bearing detection (located on the bearing adapter) would detect vibrations. In fact a company called HUM is trialing this tech as we speak.

https://www.humindustrial.com/products

Rahhhhhhhhh!!!!
  • Member since
    May 2003
  • From: US
  • 25,292 posts
Posted by BaltACD on Friday, July 12, 2024 1:54 PM

Euclid
...
But bearings can overheat quickly, the FRA says, and can burn off in as little as 1 to 3 minutes. Hotbox detectors failed to diagnose 124 severely defective bearings in the U.S. and Canada from 2010 to 2018, 117 of which resulted in derailments.”
 
...

Those would tend to sound like 'big' numbers - until you plot those numbers against all the DD inspections that happened during those eight years. A 100 car train generates 800 individual axle inspections + the number of axles of the engine.  How many trains operated past how many defect detectors during the eight years mentioned to have those 124 failures?

Would 1000 trains per day operating past 10 Detectors a day be realistic?, would 1500 and 15 Detectors be realistic.  In either case the total number of inspections gets real high real fast and 124 failures is a very small fraction of 1%.  The ideal remains ZERO, but as we humans know, while perfection is the goal, humans and their machines rarely if ever attain it. 

Never too old to have a happy childhood!

              

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Friday, July 12, 2024 1:51 PM

I think I have mentioned this before.  When an inner bearing starts to fail, the heat produced is transferred by conduction, outward from the outer race of the bearing and casing into the sideframe, and inward to the axle, thence in toward the wheel and out toward the end cap.  The problem with monitoring the endcap is that the detector only monitors the temperature rise at the outer face, when the temperature close to the inner bearing in the axle metal may have reached dangerous elevation.  This is not helped if low ambient temperature (as at East Palestine) is abstracting heat before it reaches the endcap.

This can be, and I personally suspect might have been, augmented by a failure that precludes continued rotation of the initially-failed bearing, for example if the axle end fails and the truck end falls down into dragging contact.  This produces plenty of 'fire' out at the railhead, or perhaps in the plate of a dragged wheel down at railhead level (not in line with the axle center where the detector is scanning for 'hot wheels' due to stuck or misapplied brakes), but this will not be where the detector is reading 'heat rise', and you will see a relatively slow and perhaps nonproportional heat rise instead of the catastrophic symptoms that, for example, were picked up by doorbell cams.

I expect the NTSB report to have a clear timeline of the reported temperature rise that culminated in the order to stop the train.  I also expect this to be far below the sort of temperature that would warn a train crew to use careful handling in initially setting up and slowing the consist -- this being a factor in the 'other thread' we were having regarding more careful initiation of dynamic braking as the first stage of deceleration of a train with a heat-detector alert.

  • Member since
    December 2001
  • From: Northern New York
  • 25,020 posts
Posted by tree68 on Friday, July 12, 2024 12:40 PM

Any inside sensor for bearing temperature will have to account for the fact that there is no physical connection between the axle and the car, aside from gravity.  

RFID technology can deal with that - it's already done in your car (remember the low tire pressure sensor?).  Odds are the technology can be added to the endcaps.

As noted, the sensors can be added to the wheelsets at any time - new, rebuilt, etc.  Eventually all wheelsets would be properly equipped.  Equipping all of the cars will be a challenge.  I'd opine that "fleet" cars (unit trains, IM) would be an initial focus.

The best part is that a car will roll equally well on a wheelset with, or without, the sensor.

The worst part is that this will cost money, which will detract from the bottom line - something many investors may take exception to.

Not to mention that it's one more part that can fail.

LarryWhistling
Resident Microferroequinologist (at least at my house) 
Everyone goes home; Safety begins with you
My Opinion. Standard Disclaimers Apply. No Expiration Date
Come ride the rails with me!
There's one thing about humility - the moment you think you've got it, you've lost it...

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Friday, July 12, 2024 12:11 PM

jeffhergert

I thought from all I've read that the bearing was trending hot.  Not enough to set off the detector in the field, until the last one that was too late. Even though the detector doesn't alarm in the field, it sends data to a central location where a person interprets it. I think in my career I or my conductor have checked more possible hot boxes because of the bearing desk noticing heat building up rather than a detector giving a direct alarm. 

It's possible for detectors to get fooled. It's possible, maybe probable, that whomever looked at the detector readings saw something to indicate the heat rise was an anomaly and decided to not have the train notified. 

Yes, the detectors were designed when much if the car fleet had plain bearings and heat was much more visible. But that doesn't mean the detectors are useless.

From the explanation given on another forum, I think more acoustic detectors are what's needed. They can find a bearing that's beginning to fail before there's any heat.

Jeff

 

With the East Palestine wreck, I heard what you are describing.  That is that there seemed to be a consensus that first two hotbox detectors were only indicating a slightly elevated temperature, but not high enough to require the train to be stopped and inspected. 
 
I assumed that meant that the actual bearing parts were not hot enough to call for stopping and inspecting.  I naturally assumed that this was a true analysis of the heat inside of the critical bearing parts.
 
So, I concluded that between the second detector and the third one; where the derailment occurred, the actual bearing temperature rose to a critical temperature where bearing failure was imminent.  I also concluded this meant that the detectors are too far apart.
 
But recently, I have been hearing an alternative explanation saying that hot bearing detectors are simply incapable of correctly interpreting the actual temperature inside of the bearing.  With this claim, it seems to say that the detector sensors can only read the exterior of the bearing assembly, which is cooler than the interior.  So this raises the obvious question of how science could have made such a colossal mistake that has gone unrealized for so long. 
 
Yet maybe there was no mistake because here are other sources that say hotbox detectors do work at least most of the time. The FRA says they work, as they indicate as follows:
 
“There’s no question that hotbox detectors work. Train accident rates caused by axle and bearing-related factors have dropped 81% since 1980 and 59% since 1990 due to the use of hot bearing detectors, according to the FRA.
 
But bearings can overheat quickly, the FRA says, and can burn off in as little as 1 to 3 minutes. Hotbox detectors failed to diagnose 124 severely defective bearings in the U.S. and Canada from 2010 to 2018, 117 of which resulted in derailments.”
 
I also notice that the reason why it is claimed that detectors don’t work is because they are not sampling directly inside of the bearings.  To overcome this problem, the solution is said to be onboard sensors directly connected to each bearing.  These will detect the actual temperature inside of the bearing, and transmit the information to all parties to take action needed to prevent a bearing failure.  They will also detect vibration as the earliest indication of a bearing failure trend. 
 
There seems to be a strong advocacy for what is likely to be a highly lucrative conversion of the U.S. railcar fleet over to onboard sensors.  Naturally the railroad industry is leery of this advocacy pushing for a mandate.  So it is the same dynamic as the ECP brake advocacy. 
 
Maybe there is some middle ground where a new way is found to increase the density of gound based hotbox detectors.
  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Friday, July 12, 2024 10:55 AM

BaltACD
Overmod
...
The acoustic signature of some bearing failures is not a high prolonged screech or other obvious timbre and pattern of sound -- and ultrasonic frequencies may not make it 'out of the truck frame' and through 15' or more of air to the microphone array.  This, again, is why some form of onboard monitoring close to each bearing seat is important if you actually want to have systematic protection against catastrophic roller-bearing failure.

What would be the cost of EIGHT of 'your' detectors per car? Multiplied by cars in use nationwide.

Fair question.

My version puts the temperature, vibration, and 'microphone' sensors for various ranges of sound in a module that goes in a recess in the bearing seat elastomer pad -- stock module in a stock recess, field interchangeable.  This communicates with a small processor that mounts on a sideframe, and that 'registers' that sideframe as a unit.  Wayside equipment can then communicate via one of the flavors of RFID or wireless networking either to receive data or communicate requests for data or testing.

The system would be installed incrementally, probably as trucks are serviced to change out wheelseats and release sideframes for installation.  There are options for scale of rollout, priorities for installation, financing methods including tax setasides, etc. just as there are for ECP and enhanced private-crossing devices.

There is also some cost associated with continuous data transception from 'all the devices' to wayside detection.  Since most of the regular traffic that actually has to be transmitted is within available PTC bandwidth, we might consider the necessary changes to the SDR architecture in PTC a one-time expense if done right, with other synergistic potential benefits for train operation and handling.

I am highly interesting in seeing how the NTSB recounts the cause of the East Palestine accident.  There seems to be a considerable amount of "vested interest" in the problem being something that is amenable to increased hours of carman inspection (on the union and Government side) and to more closely spaced and intricately sensor-fused wayside detector suites (on the grudging railroad side).  While I would hope this wouldn't influence the structure of a fact-finding report, it's something everyone here should be carefully watching with critical BS detectors tuned and working.

  • Member since
    September 2017
  • 5,636 posts
Posted by charlie hebdo on Friday, July 12, 2024 8:08 AM

BaltACD

 

 
Overmod
...
The acoustic signature of some bearing failures is not a high prolonged screech or other obvious timbre and pattern of sound -- and ultrasonic frequencies may not make it 'out of the truck frame' and through 15' or more of air to the microphone array.  This, again, is why some form of onboard monitoring close to each bearing seat is important if you actually want to have systematic protection against catastrophic roller-bearing failure.

 

What would be the cost of EIGHT of 'your' detectors per car? Multiplied by cars in use nationwide.

 

What would be the cost of another East Palestine or worse?

Corporate profits reign supreme over health, safety and property?

  • Member since
    May 2003
  • From: US
  • 25,292 posts
Posted by BaltACD on Friday, July 12, 2024 7:36 AM

Overmod
...
The acoustic signature of some bearing failures is not a high prolonged screech or other obvious timbre and pattern of sound -- and ultrasonic frequencies may not make it 'out of the truck frame' and through 15' or more of air to the microphone array.  This, again, is why some form of onboard monitoring close to each bearing seat is important if you actually want to have systematic protection against catastrophic roller-bearing failure.

What would be the cost of EIGHT of 'your' detectors per car? Multiplied by cars in use nationwide.

Never too old to have a happy childhood!

              

  • Member since
    September 2003
  • 21,669 posts
Posted by Overmod on Friday, July 12, 2024 7:19 AM

jeffhergert
But that doesn't mean the detectors are useless.

I don[t think anyone actually says that.  Indeed, just in your experience you can probably document many cases where detector information has, in fact, warned of a developing failure -- both in bearings and in the 'visible' portion of wheels with sticking or unreleased brakes.  Better resolution in 'camera suites' combined with proper (not fake) AI can only improve the quality of these functions.

The thing I keep harping on is essentially the same thing that makes the 'increased car inspections' in the House and Senate "post-East-Palestine" bills so idiotic: a thermal detector will almost never 'alert' on an inside-bearing failure, or indeed an axle failure; this adds to the realization that some types of failure can progress catastrophically within even the ten miles between detectors that the legislation would mandate.

From the explanation given on another forum, I think more acoustic detectors are what's needed. They can find a bearing that's beginning to fail before there's any heat.

These are certainly something I'd advocate for... but some of the 'magic bullet' capabilities may not actually be realizable for what's needed.  The acoustic signature of some bearing failures is not a high prolonged screech or other obvious timbre and pattern of sound -- and ultrasonic frequencies may not make it 'out of the truck frame' and through 15' or more of air to the microphone array.  This, again, is why some form of onboard monitoring close to each bearing seat is important if you actually want to have systematic protection against catastrophic roller-bearing failure.

  • Member since
    March 2003
  • From: Central Iowa
  • 6,901 posts
Posted by jeffhergert on Thursday, July 11, 2024 10:39 PM

I thought from all I've read that the bearing was trending hot.  Not enough to set off the detector in the field, until the last one that was too late. Even though the detector doesn't alarm in the field, it sends data to a central location where a person interprets it. I think in my career I or my conductor have checked more possible hot boxes because of the bearing desk noticing heat building up rather than a detector giving a direct alarm. 

It's possible for detectors to get fooled. It's possible, maybe probable, that whomever looked at the detector readings saw something to indicate the heat rise was an anomaly and decided to not have the train notified. 

Yes, the detectors were designed when much if the car fleet had plain bearings and heat was much more visible. But that doesn't mean the detectors are useless.

From the explanation given on another forum, I think more acoustic detectors are what's needed. They can find a bearing that's beginning to fail before there's any heat.

Jeff

  • Member since
    July 2014
  • 82 posts
Posted by Ajsik on Thursday, July 11, 2024 8:17 PM

tree68

I once reported a misplaced tarp (it was hanging off the car) to CSX on a train at Deshler - from my home in NY.  The operator was a bit incredulous, but took my report.

When getting the number off the blue placard at the crossing, make note of the crossing number as well.  "Smith Street in Podunk" isn't as much good to them as crossing 123456P.

 

I did the same while observing vehicle traffic driving around the obviously malfunctioning gates on the Blue Island camera. I used Street View to confirm the crossing number.

The CSX operator was surprised by the situation, but once I guided him in, he confirmed the issue (also seemed to have access to a camera on RR property where he was viewing the same from the opposite angle).

Maintainers arrived within an hour, though the problem seemed to correct on its own after about 30 minutes.

I was surprised at the YouTube commenters who obviously saw the same thing I did but were content to chat about it without realizing they could actually help.

  • Member since
    July 2006
  • 1,530 posts
Posted by NKP guy on Thursday, July 11, 2024 7:53 PM

ChuckCobleigh
thought that bringing a camel on railfanning trips could provide a solution to the spitting strategy.

But if he's wrong and the bearings way overheat, would you walk a mile for a camel?

  • Member since
    December 2001
  • From: Sunny (mostly) San Diego
  • 1,920 posts
Posted by ChuckCobleigh on Thursday, July 11, 2024 3:48 PM

jeffhergert
I always liked what the old heads said to do. Spit on it. If it spits back,  it's hot.

Jeff

PS. Don't take that last part to mean railfans should be spitting at passing trains. 

Lots of common sense (remember that?) from the old heads.

I read the PS and thought that bringing a camel on railfanning trips could provide a solution to the spitting strategy.

  • Member since
    May 2003
  • From: US
  • 25,292 posts
Posted by BaltACD on Thursday, July 11, 2024 2:43 PM

The more you know, the more you find out what you don't know and need to find out.

Never too old to have a happy childhood!

              

  • Member since
    December 2001
  • From: Northern New York
  • 25,020 posts
Posted by tree68 on Thursday, July 11, 2024 1:15 PM

Euclid
Okay, thanks for that explanation.  Was this insufficiency of the detectors not known when they came into widespread use?

The original hotboxes were friction bearings - I would suspect that early detectors were made to detect those.   Only heat was the issue.

The earliest hotbox detectors were the crew's noses in the caboose.  IIRC, something was added to roller bearings to make them stink when they got hot.  With no caboose, this would be pointless now.

Roller bearings are a comparatively recent development.  As SD60 mentions, it's the inner races that need detection now, in addition to the outer races.  This may well be a case of not realizing this was also a problem early on.

LarryWhistling
Resident Microferroequinologist (at least at my house) 
Everyone goes home; Safety begins with you
My Opinion. Standard Disclaimers Apply. No Expiration Date
Come ride the rails with me!
There's one thing about humility - the moment you think you've got it, you've lost it...

  • Member since
    January 2014
  • 8,221 posts
Posted by Euclid on Wednesday, July 10, 2024 10:38 PM

SD60MAC9500

 

 
Euclid

 

 
Overmod

Keep in mind that, just as no amount of stationary car inspection would divulge the kind of bearing failure in accidents like the one at East Palestine, no 'civilian' addition of devices scanning the outer face of the truck for 'heating' would.

Only temperature detection that accesses the inside roller bearing area, or can be pointed to it during operation, would work.  It might be mentioned that no scanning or sensor-fusing (e.g. IR and high-frequency sound analysis) located outside the plate clearance will work -- you need something like an arrangement of mirrors or fiber-optics that can 'point' at the area of the inside bearing from below, subject to dust, weather, and other damage.

 

 

 

Why did the hot bearing detectors fail to prevent the East Palestine derailment?

 

 

 

Because HBD's do not detect the temperature of both bearing raceways...  Onboard bearing detection needs to become a thing.

 

Okay, thanks for that explanation.  Was this insufficiency of the detectors not known when they came into widespread use?

Join our Community!

Our community is FREE to join. To participate you must either login or register for an account.

Search the Community

Newsletter Sign-Up

By signing up you may also receive occasional reader surveys and special offers from Trains magazine.Please view our privacy policy