Railway Preservation News

Hydro testing issues
Page 1 of 4

Author:  Overmod [ Wed Dec 09, 2020 6:19 am ]
Post subject:  Hydro testing issues

Since a couple of technical discussions got cut off along with certain threads, I want to reopen the subject of 'best hydro practices'.

It is my opinion that the definition of 'service day' requires physical pressurization by steam above atmospheric in order to count -- merely heating water with a small fire on the grates to the stated ~120F maximum will not accomplish this no matter how much hydrostatic pressure is subsequently applied to the water.

In fact this ought to be true for the hypothetical 'desirable' case I stated earlier, that the waterspace temperature should be brought as close to 212F as does not begin 'overcoming latent heat of vaporization' as subsequent hydrostatic pressurization is applied. This would not reach what the statute indicates: pressure developed by firing.

As I recall -- I'm not diving into a cesspool to check -- someone indicated the idea of "raising hydrostatic pressure by firing alone". Aside from completely defeating the purposes of an actual hydro -- that, of course, would count as a service day.

Meanwhile, Robby Peartree raised some issues regarding Bourdon gauges used as pressure indication during a hydro test. He pointed out (nominally correctly) that these gauges are said to read most accurately at the middle of their mechanical range, and therefore if a hydro is to be pressurized, say, to 150psi, the gauge itself should read that in the middle of the range, not near one end of the indicated scale. He also raises a concern that sharp or excessive pressure excursions experienced near the outer scale limit of a 270-degree Bourdon movement might damage the gauge, either mechanically or by deforming the shape of the tube. He said in part:

"Remember that these gauges work by one or more bourdon [note sp.] tubes being deflected to create the needle motion. The straightening of the tube or tubes is the most consistent in the mid range of the gauge. This consistent behavior creates a better accuracy in the middle of the readings."

My opinion on this is that there isn't a real guarantee that the gauge will be more linear in the center of its range and 'more nonlinear' toward both extremes, any more than the most consistent driving range of a watch spring is 'precisely' in the middle of expansion. It would of course be comparatively simple to determine this for any particular gauge: you'd have to make a map of tube deflection resulting in pointer movement for a set of pressures determined by a more accurate gauge to be sure.

A more important criterion, I think, is the accuracy class of the gauge as measured on liquid -- this specifically relates to a guarantee of percentage deviation for the 'worst' case in the range. Note that as this is commonly expressed as a percentage of full-scale deflection, the "accuracy" of a gauge with a particular class reading, say, 150psi may be greater than one of the same class registering 300psi. (This is a consequence of how the gauge was made, and to an extent how well it has been treated during its service life, but I think the point can be seen.) There is also a certain amount of reading precision lost on a higher-reading gauge unless its diameter is also 'increased' so that distance between scale marks is comparable.

The first problem when we apply this stuff to hydro tests is that true accuracy is not as relevant to precise measurement as you might think -- elevating the gauge one foot will produce a pressure excursion of nearly 1.5psi, which is outside the 1% accuracy class guaranteed error for the lower-reading gauge already. Likewise tilt off vertical orientation, or incomplete liquid filling to the very tip of the Bourdon tube, affect accuracy ... perhaps more than nonlinearity in the tube deflection as manufactured.

The second problem is that, for a hydro, there is little concern with the precise measurement of pressure (given that the accuracy class and calibration of the gauge are reasonably correct). A couple of extra pounds over isn't a critical thing here, as it might be for setting multiple pop safeties to be a precise 2psi or so apart. You're already using so coarse a measure of overpressure above nominal that technical range linearity is a minor concern. In my opinion at least, it's certainly not worth paying for a larger-diameter gauge and need to spec a higher nominal accuracy class in a 300psi gauge that will never be exercised above 150psi for the purpose.

If you actually are concerned with calibrated accuracy across the displayed range of a gauge, there are technologies for pressure gauges that are better, and probably cheaper, than accurate metal-tube mechanical Bourdon gauges. A simple manometer setup is more linear and easier to determine corrections for. At least theoretically you could use something like a force-balanced quartz tube gauge (if you're reasonably sure you can apply the hydrostatic pressure smoothly and relatively slowly, and there is little likelihood of catastrophically rapid pressure relief at full hydrostatic pressure which might crack the tube) but oh, brother! is that overkill!

Incidentally I see at least one report that people use a pressure-washer pump to pressurize for a hydro. This is an interesting labor-saving method PROVIDED that there is very good 'automatic' safety relief not far above the nominal hydro pressure that can handle the mass flow produced by the pump at full rate. Is there an advantage to pulsing flow in pressurizing the boiler?

Author:  Dave [ Wed Dec 09, 2020 10:39 am ]
Post subject:  Re: Hydro testing issues

I'm lately seeing - with some amusement - a lot of discussion of theoretical "perfections" being discussed here. Rest assured there aren't any in this business. No weld will be 100% perfect, no staybolt entirely perpendicular..... no pressure exact and constant. The question of hydro test pressures rising and falling is based on the idea that in the real world we can in some way prevent it. The backshop isn't a clean laboratory with delicate gauges made to measure in scientific standards...... I think the writers of the standards are full well aware of this having done a lot of it themselves. Old boilers won't stay at hydro pressure because they all leak somewhere, and with water which can't be compressed a tiny seep will drop pressure rapidly. All diesel prime movers leak oil..... railroading need to be robust and forgiving about tolerances and margins (despite Porta's despair about what he called good enough engineering not being good enough). Pressure washers make great hydro pumps as long as somebody is there to squeeze the trigger when needed to jack it up a bit more again, and let off when it slightly exceeds desireable spec. I've seen air compressor regualtors used to regulate water fed for hydro testing with limited accuracy and robustness. We do the best testing we can consistent with the spirit of the regulations with what we have to do it with. If your 250 PSI test ends up running in cycles between nominal 245 and 255, it's still a good test, and will show you any areas of concern.

Author:  Robby Peartree [ Wed Dec 09, 2020 3:27 pm ]
Post subject:  Re: Hydro testing issues

It has always been my understanding that any test fire prior to the completion of the 1472 day inspection, and that meets the definition as outlined in the regulation, is counted as a service day and should be recorded as such.

MD Ramsey

To All

I would take this to say a hydro performed with both pressure boiler and a fire in the firebox would count as a service day. I think it is important to remember the reason to remove the boiler tubes is to inspect the interior of the boiler.

To that end the definition of a service day for a fire locomotive requires fire in the firebox. The Santa Fe Rules Governing Care of Steam Boilers, Air Reservoirs and Appurtenances Thereto, The Atchison, Topeka & Santa Fe Railway System Revised January 1, 1923 discusses using the water in the boiler if the locomotive is coming off the road or a warm water source of either a warm water tank or a live locomotive or other boiler.

On the subject of hydrostatic testing of "old boilers". Charlie Williams who inspected steam locomotives for the I.C.C. once stated that the best way to impress an inspector is to do the test with a hand pump. We were able to do it on a 1904 boiler with a vent valve in the steam dome.

On the subject of welding. If your welding has inclusions or does not fully fill the area intended or has other defects then why allow it into service? If the welder you have can not do the job then find another they are out there. When I first went to work for Grand Canyon Railway there were several welders working there who came out of the power plant industry. Their full penetration welds were incredible! They also were great artisans at creating items from a flat piece of steel. Jack's C. work on the firebox door of 4960 is just one example. Other welders include Marcus P, Kilroy, Ron, and there are others who names I have unfortunately forgotten. The late welding foreman a the copper refinery in El Paso could do some incredible welds and I won't even begin to describe the great work I have seen out of gas pipe line welders where they are used to every weld being x-rayed.
If you really want to get nasty about any piece of metal, they all have cracks or other defects including missing atoms in atomic structures and other issues that are upper level study in college. Good welding is an art. Not all welders are the top artist. It is worth they money to find the top artist for certain work. I am not a welder, I do not have steady enough hands for it and I never have. But my metallurgy degree gives me the skills to understand the science behind the art and help those who want to improve their skills.

Robby Peartree

Author:  Boilermaker [ Wed Dec 09, 2020 6:34 pm ]
Post subject:  Re: Hydro testing issues

In my experience hydrostatically testing vessels, a pool heater is the go to for warming the water. In case that's not an option, I have used wood fires and propane/diesel heaters for the same. The fire is put out once the water is sufficiently heated, prior to performing the hydro. Once put out, some time is given for the firebox to cool, so that it and the smokebox can be entered and inspected during the test.

I have never seen or heard of raising the pressure in the boiler -while- a warming fire is going, and I would not recomend trying, it would also be impossible to enter and inspect the smokebox and firebox during the test.


Author:  Overmod [ Wed Dec 09, 2020 8:52 pm ]
Post subject:  Re: Hydro testing issues

"I have never seen or heard of raising the pressure in the boiler -while- a warming fire is going, and I would not recommend trying. Doing so would technically count as a service day under FRA"

The thing is that I think it technically doesn't, and the rule is written to cover this.

A hydro isn't done with any space in the boiler at all: it is completely full aside from some unavoidable bubbles and dissolved gases. The FRA Part 230 rules describe a service day as requiring above the ambient atmospheric pressure of steam, which you don't have except perhaps in ridiculously small amount due to small-scale Eisenhoffer effect if you've built your fire or impinged your heater wrong, and that will preferentially collapse as you increase the pressure to the range required by the hydro anyway at a mere 120 degrees F final temperature.

All bets would indeed be off if you kept firing until enough water was above the saturation temperature as pressurized to start producing actual steam generation. I'd like to think readers here are not that stupid.

We haven't had the technical discussion of whether it is better to preheat the gas path on the fireside as opposed to preheat the waterside mass through pumped circulation. I theoretically have always favored the latter, but would surely entertain heating through the gas space or in some part of it in conjunction with circulation.

Author:  ZOBEX [ Wed Dec 09, 2020 11:32 pm ]
Post subject:  Re: Hydro testing issues

I'd like to hear from those whom have actually run a hydrotest on working steam boilers as we were looking at a locally available steam traction engine, a buddies project. I am not concerned over rules but just at what point this thing is considered safe or not for nominal use while considering what he wants to do with it. I have and do "hydro test" compressor and other pressure tanks here for years. yes, using a pressure washer and many pressure relief valves. For compressor tanks up to and even over 50 gallon, we simply fill with water, install pressure relief valves which we have already capacity calibrated so we know what they can release, use an in line flow meter so we really know what is coming out. The pressure washer line out to the tank has a flow restrictor in it so even though it can pump out 1.5 gallon at 1200 psi, it truth it is about 0.4 gpm coming out. The adjustable pop offs can handle about 2.5 gpm. Water is not really going to compress any. So we pick an operating pressure and then a test pressure for static load for a given time at that over load pressure. For our purpose it suits us well. I am asking, what is the operating pressure and then the test pressure ratio or is it a fixed number of pounds above pressure that is used. Or, is it unique to a boiler type ?

Some basic numbers would be appreciated. Yes I know a hot boiler is different than a cold boiler but any basic numbers here would be appreciated.


Author:  Overmod [ Thu Dec 10, 2020 1:01 am ]
Post subject:  Re: Hydro testing issues

The short answer to this is remarkably quick. There is a more detailed answer which is not materially different but I think deserves a little discussion.

The simple answer is to adopt the Part 230 definition in the CFR, which requires one piece of information you may have to measure if you're not required to file a Form 4 on your equipment. That definition is 25% above maximum allowed working pressure. The thing that starts getting complicated is that the general definition of MAWP for pressure vessels is elsewhere in the CFR, and while it's worth reading it makes a :"Federal case" out of fairly going to the boiler and measuring what it ought to be.


What the law says for Part 230 purposes is that the MAWP is that pressure specified according to Appendix C, the Form 4 documentation. In practice, consider it the maximum pressure reached in practice when all the pop safeties have activated on overpressure. An approximation is the specified working pressure at which the first safety lifts to prevent further rise.

The complication is that, when you read through the boiler jargon, the folks defining MAWP for pressure vessels wanted the most conservative possible definition, e.g. the lowest pressure that counted as 'working pressure' under a range of conditions. You can see why this is a sort of unintended-consequence problem when another regulation makes a lowball number the criterion for a fixed-percentage excursion: the 25% is also the most conservative rating of test pressure ... but now in the wrong sense.

My personal opinion is that you'll never go broke following the letter of Federal law in Federally-concerned matters, though. If generally-accepted safety for large steam locomotives involves a nominal 25% over working pressure, you won't go wrong applying it as a standard for less-regulated or smaller-size steam power.

On the other hand there are all sorts of state laws that apply to different kinds of power boilers 'not on the railroad system of transportation' and, for all I know, some of these might require different hydro standards. If that is a concern, either search code for the particular state or consult a lawyer licensed to practice there...

Author:  ZOBEX [ Thu Dec 10, 2020 1:41 am ]
Post subject:  Re: Hydro testing issues

Thanks, I was not wanting to look at the CFR, rather play ostrich. And, 25% over is what we always use and then after that I still down grade. I cringe over looking at high pressure tanks. Goes back to my younger days and scuba tanks. I once saw an O2 tank blow, it was amazing and that was just compressed gas, not water-steam flash. For anything the size of locomotive boiler, I would be and will be very conservative and for that reason what we are haggling over, if and when, I most certainly want a restoration shop, with a proven track record, to rebuild it. Regardless what it surveys out at.

Will be on the lookout for old technical books on boiler construction and design. Seen them in the past but never till now had a reason to scoop them up.

On this traction engine the owner claims he, himself, did a hydro test at 125psi but then goes ahead and runs it at 125psi. So what does that prove ? ? Told my buddie, since he is not plowing anything and it is only a toy, IF he purchases it, run it at 100psi and live with it. I am sure it would drive about quite well at 85psi.

Author:  Overmod [ Thu Dec 10, 2020 3:05 am ]
Post subject:  Re: Hydro testing issues

The moral of the story is that a properly conducted hydro test cannot cause any sort of explosion -- that is a principal point of doing a pressure test that way.

It also provides a controlled way to watch and measure any swelling or expansion of the boiler as pressure develops, in much the same way that raising comparable steam pressure would. The difference being that you can develop that pressure without having to put the energy of latent heat of vaporization into the water to get that elevated pressure ... and should something leak, you won't get the rocket effect or momentum shock that contribute to 'boiler explosions' or other danger. The amount ejected from a leak is related to any swelling that might have taken place, likely slight, and as soon as that swelling has relaxed -- much as you see a dribble from a garden hose when the faucet is turned off -- that will be the end of whatever seepage got started in the test.

Now, you could hydro to MAWP and have reasonable assurance that the safety will pop before anything in the boiler starts to seep. But if there is anything that might fail catastrophically in the boiler at *just* over that pressure, it might cause danger even for low nominal gauge pressure -- so we test safely to a higher pressure. The analogue of this in testing is the requirement that a vehicle design be capable of a speed 10% over its :"maximum" rated speed -- say, 110mph for a rated 100mph peak.

There's nothing that says you can't test higher -- or that you're going to bung something loose if you do. The old ASME locomotive code contained a 400% factor of safety; the new one 500% ... which in a sense means you could hydro a 150psi boiler to 750psi before leaking would be expected, assuming all construction was correct. In practice or the imperfect 'real world' of course you'd see something start to fail well below such a pressure, perhaps progressing to the point pumping wouldn't keep the thing pressurized. There's comparatively little point in actually testing that high unless testing a unit to failure -- an expensive thing to prove you did your calculations accurately! But even if you did, the risk you'd run during failure would be comparatively tiny.

There is an exception, though, which I think is an important one: you must not have an appreciable volume of empty space anywhere in the pressurized volume -- no little air spaces or bubbles. These would act very much like accumulators, building up what might be a considerable volume of fluid to 'fill the void' partially at higher pressure, but effectively isostatically compressing all the fluid at the containing 'periphery'. Were something to start leaking, you wouldn't have the dramatic volume excursion of overcritical heated water corresponding to saturation pressure, but you would have a volume of high-pressure water that might be a considerable portion of the original 'bubble' size shoot out until the bubble came back to equilibrium "at or near its original gas volume".

It can be fun getting gas out of a boiler that is to be tested. I think it's important to design things so that, when filling, any relative 'high point' has its own bleeder, and very carefully all of these are loosened until no bubbles come from them, and then retightened before any actual hydraulic overpressure is applied. It might be possible to use the same vacuum technique that is used to degas casting resins ... but there is a corresponding careful warning: don't pull even a comparatively gentle vacuum on a pressure vessel that still has a substantial volume of gas in it. Components in a boiler are designed to resist pressure: they may or may not be designed to take vacuum (or more precisely, resist an outside differential pressure of ~15psi...

Author:  ZOBEX [ Thu Dec 10, 2020 9:05 pm ]
Post subject:  Re: Hydro testing issues

Can't imagine vacuum pumping to evacuate air. In years passed I did a lot of RTV mold making for small part casting and we made up our own evacuation chamber for that. Also operated a large glass evacuation chamber for gold filtrate evacuate. I ran that one with a slight leak off and then the pump near by. BUT power on and off from about 20 feet away. Even though it was manufactured to handle a near vacuum, I never trusted being near that much Pyrex glass.

If we get to the point on something as big as a boiler to pressure test, I assume we would level the top of the tank and try to direct everything to the steam dome, that being the highest point and it has the pop-off valves anyway. Plus another one some place lower down. One of the target loco's has a 125psi operating pressure and the other 165psi. Not the monster 300psi. Both are original rivet construction and one is actually in quite good condition, I had someone go do a look over and it's track record. I guess I recall federal rules require an annual inspection if it is rivet construction vs welded. I recall reading, stuff I am going to have to learn up on. I have seen and have had air compressor tanks go through the bottom at 100 and 125psi, no catastrophic failure, just leak out but hot water is another issue. As I said earlier, have a very competent boiler maker rebuild and give his report with documentation. Let him put his name on it. My time left and liability reasons, put someone else's name on it.


Author:  NYCRRson [ Thu Dec 10, 2020 10:35 pm ]
Post subject:  Re: Hydro testing issues

I apologize if I missed some context, did not have the time to read all the detailed replies, but this caught my attention;

"elevating the gauge one foot will produce a pressure excursion of nearly 1.5psi,"

Not following the logic here. A Bourdon tube pressure gauge actually measures the differential pressure between the outside of the tube and the inside of the tube. It needs to be calibrated for the elevation it is used at. At Sea Level the outside psi is about 14.7 psi (depending on weather). At 50,000 feet the outside pressure is about 1.61 psi;

https://www.engineeringtoolbox.com/air- ... d_462.html

So, that is about 0.00026 psi per foot of elevation; (14.7-1.61)/50,000 = 0.00026.

If your hydro test gauge is reporting 1.5 psi higher when you elevate it a foot you have some serious errors in your test setup.

Doing a hydro test in Denver versus Miami with an improperly calibrated gauge would be about a 2.5 psi error. (14.7 psi - 12.2 psi).

Author:  Kelly Anderson [ Thu Dec 10, 2020 11:09 pm ]
Post subject:  Re: Hydro testing issues

NYCRRson wrote:
I apologize if I missed some context, did not have the time to read all the detailed replies, but this caught my attention;

"elevating the gauge one foot will produce a pressure excursion of nearly 1.5psi,"

Not following the logic here. A Bourdon tube pressure gauge actually measures the differential pressure between the outside of the tube and the inside of the tube. It needs to be calibrated for the elevation it is used at. At Sea Level the outside psi is about 14.7 psi (depending on weather). At 50,000 feet the outside pressure is about 1.61 psi;

https://www.engineeringtoolbox.com/air- ... d_462.html

So, that is about 0.00026 psi per foot of elevation; (14.7-1.61)/50,000 = 0.00026.

If your hydro test gauge is reporting 1.5 psi higher when you elevate it a foot you have some serious errors in your test setup.

Doing a hydro test in Denver versus Miami with an improperly calibrated gauge would be about a 2.5 psi error. (14.7 psi - 12.2 psi).
There is an error there, but not the one you are talking about. A column of water one inch square by a foot tall weighs .43 pounds. If you mount a pressure gauge 10' above the boiler, and the tube leading to the gauge is filled with water, then the gauge will read low by 4.3 pounds due to that part of the boiler pressure being used to push the water up the tube.

Author:  Overmod [ Thu Dec 10, 2020 11:46 pm ]
Post subject:  Re: Hydro testing issues

"Not following the logic here..."

Your example is for a Bourdon gauge reading gas pressure. That is not relevant to hydro testing, which fills the connecting line with much heavier material whose mass follows different physics from gases.

I got the figure I quoted from a gauge-calibration site; if they don't understand the factors that affect fluid-filled Bourdon-gauge test setups, it is difficult to imagine who might.

Author:  NYCRRson [ Thu Dec 10, 2020 11:49 pm ]
Post subject:  Re: Hydro testing issues

Kelly, with all due respect, when pressurized (above atmospheric = 14.7 psi) all the water in a pressure vessel pushes out with equal force in all directions regardless of the elevation.

"A column of water one inch square by a foot tall weighs .43 pounds. If you mount a pressure gauge 10' above the boiler, and the tube leading to the gauge is filled with water, then the gauge will read low by 4.3 pounds due to that part of the boiler pressure being used to push the water up the tube."

When a pressure vessel is unpressurized the PSI reading at different heights will of course vary. Like holes in the side of a bucket, lowest hole "spurts" the farthest because of the higher psi.

That condition does not hold for a pressurized vessel... In a pressurized vessel a PSI gauge installed at any elevation should read the same (ignoring calibration errors).

If this did not hold then all steam locomotives would require boiler pressure gauges at various elevations on the back head... Lowest elevation of the back-head (floor level) would have a higher pressure than the crown-sheet elevation.... And we know that is not the case.

Author:  NYCRRson [ Fri Dec 11, 2020 12:06 am ]
Post subject:  Re: Hydro testing issues

To add to my explanation;

Here is a thought experiment;

A 5000 foot high "Boiler"/"Pressure Vessel" filled with water and having PSI gauges spaced 500 feet apart.

When unpressurized the lowest gauge will report a much higher pressure than the highest gauge (area of the boiler times the weight of water times the height).

Once the pressure vessel is pressurized above atmospheric pressure each and every pressure gauge will report the same pressure at all elevations (assuming they are calibrated correctly for the elevation they are at).

This also holds for "most" gases (nitrogen, oxygen, etc) that are below the pressure where they become liquid (unlike Propane for example).

Page 1 of 4 All times are UTC - 5 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group