Derailment near Carlisle - 19/10/2022 - 1715 Clitheroe Castle Cement Gb to Carlisle N.Y. (6C00)

[GC class B1]

Surely if one wheel on a fixed axle were to stop rotating the opposite wheel would also. Or am I missing something?

The confusion may be my fault. A wheelset (2 wheels on one axle) rotates as one entity. The BFCB arrangement operates by pushing the brake blocks against opposite wheels on each bogie. Both wheelsets are braked and a force applied by the brake rigging applied roughly equally on both wheelsets of the bogie.

The absence of any damage to the other wheelset is strong evidence that there was no brake fault causing the brake to lock. I agree RAIB's suggested sequence of events relies on a certain amount of supposition, but they seem to rule out all other possible causes.

RAIB didn’t find any other cause of the initial brake drag than a normal brake application in low adhesion conditions. What I find not credible is that the wheelset with the initial flat didn’t recommence rotating when the brake application was released. I can’t accept that the initial flat prevented the wheelset rotating because of the lower centre of gravity of the wheelset with the flats. The dynamics of the bogie suspension and variations in the track geometry would have reduced the wheel loading sufficiently to allow the wheelset to start rotating again at some time during the long distance during which the flat reached the length found. I find this explanation incredible.

 

[14xxDave]

The confusion may be my fault. A wheelset (2 wheels on one axle) rotates as one entity. The BFCB arrangement operates by pushing the brake blocks against opposite wheels on each bogie. Both wheelsets are braked and a force applied by the brake rigging applied roughly equally on both wheelsets of the bogie.

No I mean, according to the report, one wheel on a fixed axle locked up and stopped rotating. So what happened to the other wheel on the other end?

 

[GC class B1]

I am sure that is not what was meant. Can you quote the relevant part of the report. I read the report to say that one wheelset stopped rotating but the other wheelset on the bogie continued rotating. That is what happened with the other incidents involving loaded wagons with this brake arrangement.

 

[Wyrleybart]

The axles on such trains are normally fully instrumented to detect hotboxes etc (it's a lot cheaper to stick a cheap sensor on the axle/axlebox than to fit a fancy trackside sensor). One of the other use cases cited for ECP is detecting wheelslide by using cheap sensors fitted to the axles to determine if they are rotating.
Using a hall effect sensor the cost of fitting will be measured in dollars per axle.

If you insist on bringing ECP as mandatory you can probably kiss goodbye to third of your freight traffic. Its too marginal to start fitting electronic gizmos to wagons and locos.

 

[HSTEd]

If you insist on bringing ECP as mandatory you can probably kiss goodbye to third of your freight traffic. Its too marginal to start fitting electronic gizmos to wagons and locos.

Given that almost all freight traffic only exists because of huge indirect state subsidies anyway, I somehow doubt this.

The reality is that the savings to the railway industry (and thus society) from the adoption of ECP brake technology would far outstrip the costs of fitment.

But a wider discussion of this is very much off-topic for this thread.

 

[edwin_m]

What I find not credible is that the wheelset with the initial flat didn’t recommence rotating when the brake application was released. I can’t accept that the initial flat prevented the wheelset rotating because of the lower centre of gravity of the wheelset with the flats. The dynamics of the bogie suspension and variations in the track geometry would have reduced the wheel loading sufficiently to allow the wheelset to start rotating again at some time during the long distance during which the flat reached the length found. I find this explanation incredible.

It's not too clear what calculation was done to come up with the friction needed to re-start rotation - if I find the time and focus I may try to reverse engineer it but I can't right now. I agree it's questionable if it has just taken the axle load without any accounting for vehicle dynamics, although I think the steady state spring unloading would be counteracted by the bogie frame tilting to equalise the load on the two axles. However, if that's not the explanation, then what is?

 

[GC class B1]

It's not too clear what calculation was done to come up with the friction needed to re-start rotation - if I find the time and focus I may try to reverse engineer it but I can't right now. I agree it's questionable if it has just taken the axle load without any accounting for vehicle dynamics, although I think the steady state spring unloading would be counteracted by the bogie frame tilting to equalise the load on the two axles. However, if that's not the explanation, then what is?

From my understanding of vehicle dynamics there isn’t a steady state axle load when a vehicle is moving. Acceleration and deceleration with transfer weight between axles and as the track isn’t perfectly flat the primary suspension movement will increase and decrease the axle load. I think the explanation for one wheelset on a bogie locking up when the other one doesn’t is because of a significant difference between axle loading. It is also of interest that all the wagons in these incidents were loaded and part way along the train. I am convinced that vehicle dynamics play a large part in the explanation of this anomaly and this does not seem to have been considered in the investigation. I think there may also be design and maintenance implications in these incidents that haven’t been considered by RAIB.

 

[3RDGEN]

No I mean, according to the report, one wheel on a fixed axle locked up and stopped rotating. So what happened to the other wheel on the other end?

In the report on page 9 item 4 it confirms both wheels on the leading wheelset had large flats,

"Immediately after the derailment, it was identified that the leading wheelset of the ninth wagon of the train (number VTG12455) had a very large wheel flat (a flat spot on the wheel tread) on each wheel (figure 3)."

(However you do get wheelsets with minor flats/cavities on one wheel and the other wheel is visually perfect and clear of defects).


At some point that wheelset must of locked up for long enough to flat that badly it wouldn't rotate once released and then the flats grew and created the false flange, its the cause of that which appears unexplained at present. If the wheel has locked under normal braking then once released the movement of the wagon/bogie and general running vibration would surely have caused it to rotate again. The report mentions passenger stock with WSP which basically drops the Brake Cylinder Pressure until the wheelset starts to rotate again, there's nothing to force a trailer wheelset to rotate again so why don't you see the same situation on passenger stock?

I think the explanation for one wheelset on a bogie locking up when the other one doesn’t is because of a significant difference between axle loading.

Doesn't the "Gotcha" system give you readings of the axle loadings & impact figures so if a bogie is "twisted" and the axle/wheel loads are out of balance you would see it, don't newer systems give an alarm in such cases?

I think there may also be design and maintenance implications in these incidents that haven’t been considered by RAIB.

In the report it states the Brake Cylinders are mounted at the "inner end" of the bogie so the leading wheelset, which locked, was the "outer" one furthest from the BC's. It's not mentioned in the report but in the other incidents was it the same "outer" wheelset that had flats or have the other incidents occurred at both wheelset position?

 

[edwin_m]

From my understanding of vehicle dynamics there isn’t a steady state axle load when a vehicle is moving. Acceleration and deceleration with transfer weight between axles and as the track isn’t perfectly flat the primary suspension movement will increase and decrease the axle load. I think the explanation for one wheelset on a bogie locking up when the other one doesn’t is because of a significant difference between axle loading. It is also of interest that all the wagons in these incidents were loaded and part way along the train. I am convinced that vehicle dynamics play a large part in the explanation of this anomaly and this does not seem to have been considered in the investigation. I think there may also be design and maintenance implications in these incidents that haven’t been considered by RAIB.

I should perhaps have described it as the average axle load over time. It's basic mechanics that if the load coming down through the bogie pivot will be evenly split between the two axles equidistant from it. I agree the individual axle loads will vary according to vehicle dynamics and track geometry, but have no idea how much!

Bogie twist might cause a variation in average wheel loads - a factor in several RAIB reports that they don't appear to have checked for in this one. It would increase the load on one wheel of one axle and the other wheel of the other axle, although the other two wheels would be equivalently less loaded, so could possibly be relevant to the mechanism of not resuming rotation.

 

[alf]

RAIB didn’t find any other cause of the initial brake drag than a normal brake application in low adhesion conditions. What I find not credible is that the wheelset with the initial flat didn’t recommence rotating when the brake application was released.....I find this explanation incredible.

I agree. RAIB have not got to the bottom of this.
This wheelset was not picked up as having a flat before the axle locked. There were two wagons that had flats noisy enough to sound on a Railcam. Yet this silent axle alone locked up for good after a short brake application & proven full release.
Obviously after miles of being dragged locked up under full axle load there was then a mighty flat such that momentum could not lift the wheel & wagon vertically out of.

The flat, as RAIB says, caused the derailment but it did not cause the axle to lock up.

It is also bizarre that RAIB seem to imply that only one wheel ultimately had a massive flat but not the other wheel on the axle.

If this wheelset can lock up for no reason then any wagon axle could lock up & ultimately derail. It is alarming.

I shall dig out my schoolboy Newtonian mechanics book to check if it is possible for the resistance of a sliding wheel to be less less than the resistance of the wheel’s roller bearings.

If that is the case why are we wasting money on millions of wheel bearings; trains would have less resistance & fuel consumption if they were dragged on skids!

 

[Ediswan]

I shall dig out my schoolboy Newtonian mechanics book to check if it is possible for the resistance of a sliding wheel to be less less than the resistance of the wheel’s roller bearings.

A school level text iis unlikely to cover the non-Newtonian behaviour of the grease lubricating those bearings.

 

[ac6000cw]

Re. bearing stiction/friction, on page 16 of this (research paper) PDF is a graph (Fig. 10) for a wagon with tapered roller bearings of friction versus speed - https://www.mdpi.com/2075-4442/10/2/26/pdf

Basically the friction at very low speed or stationary is about twice that at 20 km/h.

Abstract: The frictional issues during motion of the axis-wheels assembly occurring in contact wheel–
rail and in bogie bearing were studied. The influence of greases upon friction therein was also
considered. The lateral dynamic behavior of the four-axle freight wagon model with two-axle Y25
bogies equipped with swing bolster was analyzed. Simulation models of such a wagon with bogies
with and without swing bolsters were elaborated for calculations considering the nonlinearities
of wheel–rail contact geometry and nonlinear methods of bogie stability. In these two options,
the cases of empty and fully loaded wagon bodies were considered. The lateral dynamic models
with 22 and 24 degrees of freedom were considered to determine the nonlinear critical speeds of a
freight wagon. It was found that the resistive torque in bearings of the assembly studied varied
nonlinearly with wagon speed. During motion along the curve track, values of such a torque can be
higher by 50% in case of the wheel under overloading and lower by 50% in case of the wheel under
underloading, respectively, compared to those obtained during motion along straight track.

 

[edwin_m]

The stiction in the bearings is unlikely to be significant relative to the upward force that needs to be exerted at the end of the flat to lift the wagon by the depth of the flat, so that it is back onto the rolling surface. This force to do so is the friction at the sliding surface multiplied by (I think) the sine of half the angle at the centre of the wheel subtended by the wheel flat. If someone can tell me the wheel diameter I'll go find the back of an envelope see if I draw a diagram to convince myself of this, and work out the number.

If the springs were extended by some random upward bounce of that part of the wagon, or by a dip in the rails, then the necessary upward force would be less. However the friction force that generates the turning moment would also reduce by the same proportion, because it is depends on the coefficient of friction multiplied by proportional to the weight on the axle, and the latter has just reduced. I think this means it is no easier for the friction force to lift the wheel onto its rotating surface.

Even if I'm wrong here, the other issue is time - the reduction in downforce will only be momentary, and the wheel has to lift up and rotate far enough that it is no longer sitting on the flat before the downforce increases, temporarily to more than the average, as the bounce or the rail dip ends.

 

[ac6000cw]

If someone can tell me the wheel diameter

For a TF25 bogie, it's between 840 mm (nominal) and 790 mm according to this document - https://www.rssb.co.uk/-/media/Proj...ed/Deviations/2020/09/16/10/13/12-222-DGN.PDF

(As a sanity check, the older Y25 is nominally 920mm and the newer LN25 is nominally 840mm)

 

[GC class B1]

The stiction in the bearings is unlikely to be significant relative to the upward force that needs to be exerted at the end of the flat to lift the wagon by the depth of the flat, so that it is back onto the rolling surface. This force to do so is the friction at the sliding surface multiplied by (I think) the sine of half the angle at the centre of the wheel subtended by the wheel flat. If someone can tell me the wheel diameter I'll go find the back of an envelope see if I draw a diagram to convince myself of this, and work out the number.

If the springs were extended by some random upward bounce of that part of the wagon, or by a dip in the rails, then the necessary upward force would be less. However the friction force that generates the turning moment would also reduce by the same proportion, because it is depends on the coefficient of friction multiplied by proportional to the weight on the axle, and the latter has just reduced. I think this means it is no easier for the friction force to lift the wheel onto its rotating surface.

Even if I'm wrong here, the other issue is time - the reduction in downforce will only be momentary, and the wheel has to lift up and rotate far enough that it is no longer sitting on the flat before the downforce increases, temporarily to more than the average, as the bounce or the rail dip ends.

I think there are two factors that mean the RAIB explanation of why the wheelset didn’t commence rotating after the initial short duration wheel locking isn’t credible. One is that the initial wheel flat would be quite short and the second that any rail joint or common rail crossing is likely to input a rotational force to the wheel on impact. It is not credible that in the long distance that the wheelset didn’t rotate without any apparent retardation force on the wheelset that the wagon did not encounter any dipped rail joints or common crossings (at points or crossings). Any calculation of the requirement necessary for the wheelset to commence rotating with the length of flat found after derailment is not relevant as the initial flat would have been much shorter.

 

[alf]

I have reread the RAIB report since I wrote my post at 460.
An apology first. I was wrong to suggest that RAIB implied only one wheel of the axle had a massive wheel flat. They clearly say both wheels had very deep flats.

Much more importantly at their para 118 RAIB say,
“RAIB has not been able to find any definitive evidence to explain why only the leading wheelset on the ninth wagon stopped rotating, and why only one wagon experienced a significant wheelslide in the adhesion conditions present.”

This honesty is praiseworthy but as I & others have said it is alarming.

A train wagon literally takes the line of least resistance.
If the brake blocks apply so much pressure that the rolling resistance of the wheel becomes greater then the sliding resistance of a locked wheel then the wheel will slide. It will slide at quite low brake pressures if the sliding wheel coefficient of friction becomes low due to leaves or moisture on the rail.

Equally if the rolling resistance of a wagon’s wheel bearings which text books say is about 5lbs a ton rises 300% (because of defective grease or broken or ill fitting rollers) to a figure greater than the sliding resistance of the wheel which in poor adhesion can be as low as 15lbs a ton then the wheel set will take the least line of resistance & slide.

The Report says the makers examined their axle bearings on the locked wheel set & found no fault.

Indeed RAIB say the axle when tested after the accident rotated.

So It remains an alarming mystery. And one which could happen again with much worse consequences.

 

[edwin_m]

I think there are two factors that mean the RAIB explanation of why the wheelset didn’t commence rotating after the initial short duration wheel locking isn’t credible. One is that the initial wheel flat would be quite short and the second that any rail joint or common rail crossing is likely to input a rotational force to the wheel on impact. It is not credible that in the long distance that the wheelset didn’t rotate without any apparent retardation force on the wheelset that the wagon did not encounter any dipped rail joints or common crossings (at points or crossings). Any calculation of the requirement necessary for the wheelset to commence rotating with the length of flat found after derailment is not relevant as the initial flat would have been much shorter.

Paragraph 113 references a previous study which suggests that a slide similar to the one probably encountered at Dent would generate a wheelflat 60mm long. The next paragraph asserts that a wheelflat of this size (the initial size, not the size as found) would require a particular coefficient of friction to resume rotation.

I do see your point on the effect of rail joints and common crossings. The latter were certainly encountered as the earlier paragraphs report on the evidence of a false flange at switches (but by the time there was a false flange, the wheelflat would be that much bigger). But I come back to the question, if this wasn't the failure mechanism, then what was?

 

[GC class B1]

Paragraph 113 references a previous study which suggests that a slide similar to the one probably encountered at Dent would generate a wheelflat 60mm long. The next paragraph asserts that a wheelflat of this size (the initial size, not the size as found) would require a particular coefficient of friction to resume rotation.

I do see your point on the effect of rail joints and common crossings. The latter were certainly encountered as the earlier paragraphs report on the evidence of a false flange at switches (but by the time there was a false flange, the wheelflat would be that much bigger). But I come back to the question, if this wasn't the failure mechanism, then what was?

I have a theory for a reason the wheelset did not start rotating again after a fairly lengthy, but not Full Service, brake application was applied and released. My theory includes a number of assumptions about the specific circumstances of these four incidents involving loaded wagons with the BFCB brake arrangement. I do not have enough information about these four incidents to test my theory. However similar circumstances to the assumptions i have made have occurred previously, but not in the combination necessary for my suggested theory to occur.

 

[edwin_m]

I have a theory for a reason the wheelset did not start rotating again after a fairly lengthy, but not Full Service, brake application was applied and released. My theory includes a number of assumptions about the specific circumstances of these four incidents involving loaded wagons with the BFCB brake arrangement. I do not have enough information about these four incidents to test my theory. However similar circumstances to the assumptions i have made have occurred previously, but not in the combination necessary for my suggested theory to occur.

Maybe approach RAIB with your thoughts? They may have other evidence that was too detailed for the reports but might help to confirm or otherwise.

 

[GC class B1]

Maybe approach RAIB with your thoughts? They may have other evidence that was too detailed for the reports but might help to confirm or otherwise.

Thank you for your suggestion. Unfortunately my experience dealing with RAIB has not been very positive.

 

[nag67]

Thank you for your suggestion. Unfortunately my experience dealing with RAIB has not been very positive.

If you are unwilling to share your theory or evidence with RAIB, how about sharing it with either the wagon owner or the other parts of the freight industry that will be considering the risks as part of their response to the RAIB recommendation?

 

[GC class B1]

If you are unwilling to share your theory or evidence with RAIB, how about sharing it with either the wagon owner or the other parts of the freight industry that will be considering the risks as part of their response to the RAIB recommendation?

To clarify, I am willing to share my theory with RAIB and any industry stakeholders. I do not have evidence that is not in the report but if they are interested in my theory and contact me via the forum I will discuss my theory and underpinning assumptions.
I am not aware of any mechanism for offering technical opinions to RAIB only for making enquiries.

 

[AndrewE]

To clarify, I am willing to share my theory with RAIB and any industry stakeholders. I do not have evidence that is not in the report but if they are interested in my theory and contact me via the forum I will discuss my theory and underpinning assumptions.
I am not aware of any mechanism for offering technical opinions to RAIB only for making enquiries.

I suspect that RAIB and "industry stakeholders" are a (the only?) parts of the railway world who do not read this forum!

 

[Clarence Yard]

You would suspect wrong - the RAIB do read this forum!

 

[AndrewE]

You would suspect wrong - the RAIB do read this forum!

Well well well... Every day is a schoolday!

I guess that one (or more) of them must be an enthusias then, as a civil servant couldn't really justify reading it as "work," even if we have got some very-well qualified people here (we also have an awful lot of easily-wound-up tyros.)

 

[trebor79]

Well well well... Every day is a schoolday!

I guess that one (or more) of them must be an enthusias then, as a civil servant couldn't really justify reading it as "work," even if we have got some very-well qualified people here (we also have an awful lot of easily-wound-up tyros.)

I knew a guy whose in the DfT, his boss used to read every railway magazine going in the office, on work time.

 

[AndrewE]

I knew a guy whose in the DfT, his boss used to read every railway magazine going in the office, on work time.

from some of the DfT's policies and their management of our railway, I guess some of the top dogs would benefit from reading Thomas the Tank Engine!

 

[Bikeman78]

RAIB didn’t find any other cause of the initial brake drag than a normal brake application in low adhesion conditions. What I find not credible is that the wheelset with the initial flat didn’t recommence rotating when the brake application was released. I can’t accept that the initial flat prevented the wheelset rotating because of the lower centre of gravity of the wheelset with the flats. The dynamics of the bogie suspension and variations in the track geometry would have reduced the wheel loading sufficiently to allow the wheelset to start rotating again at some time during the long distance during which the flat reached the length found. I find this explanation incredible.

I'm struggling with this too. I've been on some trains with very bad flats over the years, including some that have slid a few times en route. The wheels have always started turning again when the brakes released.

 

[GC class B1]

I'm struggling with this too. I've been on some trains with very bad flats over the years, including some that have slid a few times en route. The wheels have always started turning again when the brakes released.

I was involved in the follow up to an incident a very long time ago where a piece of a wagon wheel broke off as a result of a fatigue crack. Even though the wheel was no longer round it kept rotating and there were impact marks on the rail where the end of the opening kept hitting the rail. The opening on the wheel where the piece was missing did not sit on the rail and slide.

 

[sprite]

I was involved in the follow up to an incident a very long time ago where a piece of a wagon wheel broke off as a result of a fatigue crack. Even though the wheel was no longer round it kept rotating and there were impact marks on the rail where the end of the opening kept hitting the rail. The opening on the wheel where the piece was missing did not sit on the rail and slide.

I suspect in that case, the other wheel on the same axle would have been a factor. It would still be round, and by virtue of being intact, larger. This would introduce additional forces into play to enable complete or intermittent rotation of the damaged wheel.