Boilers & Accidents

Just stumbled across another US tragedy.

 

There is one I recall where the gauge glasses were showing different water levels. The crew think that the one giving the lower water level is faulty, it is tested, changed, and still shows a low level. The crew decide that it must be wrong and so ignore it, including when the water vanishes, only it turns out the faulty one was the one showing the higher water level which wasn't clearing properly. The crew were lulled into thinking the boiler had more water than it actually had with inevitable results.

People thinking that what is right is wrong and vice versa.

I seem to recall that some of the explosions that involved the S160s were attributed to valves on the gauge glasses opening other way, so firemen thought they were open when they weren't and again with inevitable results.

 

I'm no footplate expert, but given the tendency most of us show towards confirmation bias, I don't find it difficult to believe that a fireman would read the one that looked "right" rather than take the sensible worst case view.

 

I’m sure it has been referenced on this thread. Happened on a Duchess in 1948. Due to an error in re-assembly of the gauge frames, the two glasses gave different readings. The driver erroneously shut off the frame that turned out to be giving the correct reading, and relied just on the other which turned out to be faulty (and reading high), with the result that the crown sheet collapsed.

http://www.railwaysarchive.co.uk/docsummary.php?docID=734

Tom

 

"On reaching his engine, the Driver should satisfy himself that the water level in the boiler is correct, and observe the condition of the firebox, tube plate, fusible plugs, etc., to ensure everything is satisfactory".​

"On joining the engine, [the fireman's] first duty must be to examine and test the water gauges."

"Water gauges (See Drawings No. 1 & No.2

These are fitted to show the height of water in the boiler which exists above the level of the crown. Various patterns adopted for general use on locomotives are very reliable, but false readings may occur if the water or steam passages become fouled with scale, or if there is any lack of pressure balance caused by a leaking fitting. For this reason two gauges are fitted so that one is a check on the other, and every gauge has a 'test cock' or 'drain cock,' by which the gauge can be tried."​

From "Practical hints for Footplate Men" issued by the Southern Railway; I'm sure other railways had more or less equivalent instructions. (*)

It's drummed into you from the beginning to test the gauges individually; and any significant discrepancy should be an immediate cause for concern.

(*) Except Swindon probably, they did things differently there in the manner of water gauges ...

Tom

 

I never understood why such doubt was cast on the linking of the upper and lower gauge cocks. They are much better in the event of a burst gauge glass, especially with the extended operating lever, which can be knocked closed with the shovel without the need to get your hands near any blowing steam. Yes, there is a problem if assembled wrongly. On the other hand, it's like criticising a car because it won't run very well if the plug leads are fitted to the wrong sparking plugs.

 

Its an interesting comment on how tricky safety regulation is. Because they had two gauges it was considered acceptable for the locomotive to leave the shed with only one working properly, so in this particular odd set of circumstances having two gauges actually reduced safety rather than increased it because in the event neither was working properly. And you'd say, well, what are the chances of both gauges having faults, but when you have thousands of locomotives then its a circumstance that is going to happen from time to time. Under modern style regulation I imagine it would be prohibited for a locomotive to leave the shed with either gauge considered suspect...

 

Except on the GW where they only had one gauge.

 

... but they also had test cocks. I believe that the in the US one glass and a set of test cocks was traditionally considered adequate but FRA regs now require 2 glasses.
Ray.

 

I think there are all too many cases where people have refused to believe the warnings, believing that the warning equipment is malfunctioning. Think about the accidents involving the Sykes key where signalmen concluded that the system must have failed, break the seal, use the release key and then end up with trains running into the back of the other.

In this case, it was the fact that at the start of the saga, both gauges were reported as faulty. When the fitters tested as there was water coming out of the drain cock on the right hand one, they concluded quickly that the boiler must be overly full and then moved onto the left hand one and concluded that it must be at fault and wrong because it was slow to clear. All the fitters (working against the pressure of time) became fixed on trying to sort out the correct reading but slow to clear left hand one. They ran out of time, failed it and sent the loco out with what they thought was a working gauge.

The crew were lulled into a false sense of security as the fireman said, light load, free steaming loco, experienced crew, the water level won't change much. The sound of steam was explained as a leaking stay.

 

I'll take your point about on-shed maintenance.

However, for the design in general - and this might just be irrationality on my part - I've always felt more comfortable with a normal gauge glass that you can blow down the steam and water ways independently of each other and thereby confirm both are free of obstruction. That isn't something that is possible with the LMS / BR standard design where the gauge cocks are linked.

Tom

 

Don’t forget it originated on the GWR. Says it all.

 

Some instrumentation systems use three measuring devices, and a two-out-of-three voting system in the event one becomes faulty. Was this ever tried?

I don't know what you would do if you get three different results.

 

The only problem here is that in creating an advantage you also create a disadvantage in that it now not possible to check individually if the waterways in the top and bottom fittings are clear.

 

Why not put slots rather than holes in the connecting link then you could close each cock individually or in an emergency both together

Sent from my SM-A105FN using Tapatalk

 

Possibly a good idea, but about a hundred years too late!

 

I think that the S160's had a 'remote' water column, which was connected to the boiler via copper tubes, the 'Steam' tube being curved, thus unable to be 'Rodded' clear, the 'fittings' themselves were essentially a Tee fitting fed/isolated by a globe valve, so unlike a plug cock there was no 'Visual' confirmation of being open or closed, so there were TWO possible problem areas, coupled with everything else to worry about. I've got in mind that they also had no fusible plugs.

 

I am surprised that no one came up with it before considering how many clever engineers there were on the railways.

Sent from my SM-A105FN using Tapatalk

 

Gauge glasses rarely malfunction in a random way.
If the lower boiler connection is partially obstructed, steam in the glass will condense and fill the glass with water (that can't flow back into the boiler).
If the upper connection is obstructed, the condensing steam in the glass cant't be replenished, thus creating a slightly lower pressure in the glass and the water level in the glass will thus rise.

Thus, malfunctioning gauge glasses almost always show a too high water lever. No guesswork needed.

And they should regularly be tested, water level coming back slowly is a clear indication of a malfunctioning glass.