accurate measurement of steam temperature

That's true for saturated steam, but not for superheated steam, which is what the OP is trying to measure.

Andrew

Dusting off some ancient grey cells now. And then there is Gay-Lussac's law. And some Googling around shows if you combine all three, you get the combined gas law, PV/T = k. Still not clear on exactly how it applies though.

Then looking at how a steam pressure reducing valve works, it turns out that yes, there is some reduction in temp but relatively small, so the steam coming out of the PRV is lower temp than the input but still much higher than saturated steam at that temp, therefore it technically superheated. That's why if you are having trouble with water carrying over into your model steam engine, you can throttle back the steam valve on the boiler and the steam in the line to the engine is then superheated, ie its pressure drops but its temp does not drop as much, relatively. Has to do with enthalpy, steam tables and someting called isenthalpic expansion. Interesting.

Having read the various replies, I think I was on the right track suggesting a thermometer of the type sold by Amazon. That particular model can read two temperatures at the same time and is supplied with a pair of thermocouples. (If more thermocouples are needed they cost about £5)

Using a thermocouple to take truly accurate measurements is quite complicated involving hot and cold junctions and other fuss. The device sold by Amazon is relatively crude (±1.5%), but all you need to do is position the end of the thermocouple wire at the point the temperature is taken. I guess ±1.5% accurate enough for fizzy's needs.

I'd suggest fizzy position one thermocouple in the steam dome and the other close to the output of the super-heater. As others have pointed out, the measurement needs to be taken under realistic conditions. I suppose that means with the pistons connected so that the pressures at the super-heater output are 'normal'.

It would be very dangerous for me to mess with a boiler to get the thermocouple wires inside but fizzy clearly knows about boilers and fittings. What's needed are boiler fittings that allow a pair of wires to pass safely into the steam space. For me, that's the hard part. Once the wires are inside the boiler, you just plug in the meter and take the measurements.

The other methods discussed are all more or less indirect. Using an IR meter seems the best, but the other options seem to involve corrections requiring a rather deep understanding of steam, for example you can't just take the pressure and easily infer the temperature from that. (Though the technique would be useful in other circumstances).

I wonder if anyone has ever comprehensively fitted a model loco with thermocouples? Comparing steam temperature in the boiler, super-heater, cylinder inputs, and cylinder outputs might suggest improvements.

Dave

Edited By SillyOldDuffer on 07/02/2017 12:37:38

Yes, the late Jim Ewins did it on his O-6-2T loco and measured the temperatures at many points in the boiler, flues, etc. under various load conditions. His results were written up in ME for 18th March, 1st April, and 20th May 1966.

It is interesting to note that the maximum superheat temperature measured was 378°F (192°C) giving a final steam temperature of 620°F (327°C). The maximum superheater temperature recorded was 1570°F (854°C!)

John

Thanks John, I bought a box of old ME's last week and many of them are from the 60's. I wonder if I'll get lucky for once and have the right ones...

Dave

re Jim Ewins results:

It is very difficult to measure gas temperature accurately when you are close to a cooled wall, or a heated one for that matter, the thermocouple will take up some temperature where the heat received mainly by conduction from the gas to the thermocouple will be balanced by the heat lost (or gained) by radiation from the thermocouple to the surface. To do it properly you need a suction pyrometer **LINK**

There are other inconsistencies in JE's results, nice try but buyer beware.

With regard to heat lost by conduction down the thermocouple wires, the heat supply from burning coal in a medium size 5"g loco can be estimated at about 11kW. Taking a boiler efficiency of 70%, the energy flow rate along the steam pipe is 7.8 kW. I don't think a bit of conduction along thermocouple wires need trouble us.

John - did he report what the autoignition temperature of his steam oil was?

As you'll know, the Superheater Co. and petrochemical companies did lots of work to ensure high superheat, without turning the lubricating oils into carbon deposits, and thus having NO lubrication.

Just asking... JohnS.

Steady on people, original query did state 'cheaply' in the sentence.

My IR thermometer was about £8 form Lidl.

Neil

I have come back to this thread after a day at "work" (indexing old technical reports).

I don't think anyone has mentioned steam charts yet. They may be of use but like most other posters it is years since I have used them. The gas equation, PV = mRT, is only valid for perfect gases. That is gases away from their liquification and ionisation temperatures and very high velocities. Highly superheated steam may be a perfect gas but wet and just superheated steam is not. The nice equation is of no use at all.

This is not an easy measurement especially at a small scale. I would be very wary of believing any results without a thorough understanding of the possible errors in the measuring system.

JA

Edited By JA on 07/02/2017 17:42:08

Having sat here and said that it is difficult I feel it is only right to say how we would have done this at work.

If the pipe taking the fluid was large the instrumentation is easy: A sensible sized shielded thermocouple through a boss on the pipe. When the pipe was small we would have measured the outside surface temperature of the pipe with small grounded thermocouples. The pipe and instrumentation would have been well lagged and the potential errors understood.

I will give further details tomorrow.

JA

I dont think there is any question that Jim Ewins results were correct in the late 1960s.

He had access to the then state of the art stuff at Imperial College to test his 5" 0-6-2T loco which many years later I had the privilege of driving.

Jim Scott of the Newcastle and Tyneside club carried out similar experiments a couple of years ago with far more up to date sensors and replicated Jim Ewin's results.

Cheers,

Julian

Fizz, what makes you think the superheater would be cooling the steam? Is this even possible? Got pics of your boiler and superheater set up?

Depending on how much superheat you are getting, it may be difficult to determine the benefit by temperature meausurement alone. If the steam coming out of your boiler is very wet (quite common), then much of the heat added in the superheater tubes will be used up as latent heat to turn water droplets into steam, and not raising the overall steam temperature in the process.

As promised last night.

The thermocouple was made from a suitable metal combination for the temperature to be measured. 1mm or 1.5mm diameter mineral insulated thermocouple cable would be used. The assembly would be as follows:

1. Cut back the stainless steel cover of the MI cable
2. Weld the two wires together to form a bead
3. Attach the cable to the pipe by a stainless clip spot welded to the pipe
4. Spot weld the bead to the surface of the pipe
5. Cover the bead and end of MI cable with ceramic cement
6. Cover the bead, end of MI cable and cement with a sheath of thin stainless steel sheet spot welded to the pipe.

Wrap the pipe including thermocouple with suitable insulation.

Other than melting ice was used for the cold junction I know nothing about the signal processing.

This whole assembly is fragile particular when the thermocouple wires are only about 0.005” to 0.010” in diameter. It was expected that some thermocouples would be lost before and during use.

To gain an idea of the errors in the measurement the heat transfer to and from the bead needs to be understood. The temperature distribution from the flowing fluid to the outside air will be something like this:

With the corresponding heat flows:

Cvf Convention from hot moving fluid

Cnr Conduction across the wall of the pipe

Cna1 & Cna2 Conduction along the pipe

Cni Conduction through the cement, sheath and insulation

Cnw Conduction along the thermocouple wires

Cvo Convection from the outside of the insulation to the air

R Radiation from the outside of the insulation.

There are three heat balance equations:

Cvf = Cnr,

Cni = Cvo + R

Cnr + Cna2 ≈ Cna1 + Cnw + Cn1

The first two are trivial, the third shows us what is really happening. The conduction along the thermocouple leads can be minimised by attaching the cable to the pipe. The conduction along the pipe is difficult to get a hold on but Cna1 will be slightly less than Cna2. Therefore if the conduction through the insulation is reduced by increasing the insulation thickness a reduction in heat flow from the fluid to the bead follows and the bead temperature approaches the fluid temperature. Conversely if there is very little or no insulation the bead temperature will be considerably lower than the fluid temperature.

[Some of those still reading this post will say that the above is obvious, put insulation over the pipe and the pipe outside surface temperature approaches the fluid temperature. Yes, but a few postings suggested measuring the pipe temperature by a pyrometer and I can put numbers in the equations, do the sums and quantify the errors.]

The one difficulty with the above is the heat flow along the pipe. It is unknown but could be of significance. Because of this one should put thermocouples along the pipe. Given this and the need to have some redundancy for the couples that will fail you will have more thermocouples than pipe. Typical.

I have been careful to use the term fluid. The above can be used for gases and liquids. Because of the latent heat of evaporation of steam the thermocouple(s) can actually be calibrated in situ: If the boiler is held at a fixed known pressure and the heat to the boiler slowly increased there will be a long period when the pipe will be flowing wet steam. This steam temperature is known from the boiling point for that pressure and can be compared with the thermocouple temperature.

JA

Edited By JA on 08/02/2017 16:15:49

You don't actually need a true cold junction (0 C) to compensate a thermocouple. You generally use a reference junction along with a temperature sensor(!) to determine the temperature of the reference. It may seem counterintuitive or questionable at first sight but of course the reference junction is generally somewhere close to room temperature, unlike the "hot" junction.

These days it is all taken care of in a modern digital thermometer.

Murray

As I said in my post of the 6th, weld the cold junction to the boiler (or a small plate firmly bolted to it), then what you measure is the temperature of the superheated steam above the temperature of the wet steam in the boiler.

Edited By duncan webster on 08/02/2017 16:45:39

This is interesting stuff, but I fear we are over-answering Fizzy's original question.

All he wants to know is 'can I tell if if the superheater making the steam hotter or colder by measuring eth teperature of the steam jet'.

The answer seems to be in two parts:

1 The behaviour of steam means that measuring the temperature of the expanding jet of steam will be unreliable.

2 a thermocouple or IR thermometer aimed at the pipe carrying the steam should be ample to judge if it's hotter or colder with a reasonable indication of the size of the difference.

Fizzy, have you measured it yet and if so, what's the result?

Neil

From the comfort of my armchair, I would say it's very possible that a superheater might be ineffective, particularly so in a small boiler.

Consider this diagram of Duffer's Patent Boiler, a design bursting with possibilities.

How does it work? The fire is very hot. Heat from the fire boils the water, and efficiency is lifted by passing combustion gases to the chimney through the boiler, giving it more time to extract heat.

Steam gathers in the steam dome at the top of the boiler . The temperature here is much lower than that in the firebox.

Steam is collected from the dome and taken out of the boiler into an even cooler smokebox. Here it loses some heat. As steam passes down the superheater tube towards the firebox it is reheated by the boiler, but only to a maximum of steam dome temperature. It might be less.

In the firebox the steam passing through the superheater is potentially made very hot, but this depends on how long it stays in the firebox. The hot part of the superheater might not be very effective.

Unfortunately, to get to the cylinders the steam passes back through the boiler again. As the boiler is much cooler than the firebox, the steam loses heat to the boiler. If the tube is long enough, all the superheat will be lost.

The steam still has to get to the cylinder. As this involves a journey through the cool smokebox yet more heat will disappear.

I think getting a superheater to work in a small boiler could be quite a challenge, and congratulate fizzy for wanting to check his work.

Hope this is sensible, believe it or not I've been known to get things wrong...

Dave

Edited By SillyOldDuffer on 08/02/2017 18:51:00

As a biologist, when I first saw the hairpin arrangement of superheaters I was immediately struck by the resemblance of the arrangement to the 'countercurrent multiplier' arrangement of tubules in the kidneys. if you wanted the most efficient way to move heat from the cold tube to the hot one it can only be improved on by using concentric tubes!

I realise it would be impractical in most boilers, but the best arrangement would be to have the 'in' and 'out' superheater tubes in separate boiler tubes.

Neil

SoD's drawing shows the superheater tubes in contact with boiler water, which is wrong. There is a big flue (or flues), typically 1" or so running from firebox to smokebox, and the steam pipes (known as elements) run down this flue, so no contact between elements and water in this area. The flue gas is always hotter than the boiler water, so whilst you might not get a lot of heat transfer from the flue gas to the elements within the flues, what there is will tend to heat the steam, not cool it. And before everyone jumps down my throat, the flow conditions in models means that the flue gas flow is likely to be laminar, not turbulent as in full size, which is why heat transfer in the tubes is less effective, which is why models tend to use radiant superheaters