Water Rheostat Cooling Issues

[MacMcMacmac]

Hello. I am currently in the middle of a repair on a pair of three-phase water rheostats. They are used as a soft-start for a pair of 1300hp, 2300V wound-rotor motors which drive two separate stages of a 2MW exhauster. As a routine maintenance task, I recently removed 4 mild steel tube bundles from within the rheostat tanks which provide cooling for the rheostats. Although they are not very old, they have suffered from quite a bit of corrosion, mainly due to infrequent use, as well as being a local low spot in the cooling loop, which has led to a lot of sediment build up. After 4 days of cleaning and reassembly, pressure testing of the coolers has revealed leaks at the tube to end plate interface. Each tube was individually tig welded to the plate, but corrosion of the plate itself has led to leaking underneath some of the welds, as far as I can determine. Re-welding the coolers will be very time consuming, and will not solve the original problem of corrosion in the tube nest. As the tubes contain cooling tower water treated with various microbicidal and anti-corrosion chemicals, and the rheostats themselves are filled with distilled water lightly dosed with sodium carbonate, it is imperative that the coolers do not leak at all.

I have been trying to come up with an alternate solution. One idea is to build up a manifold of soldered copper pipe, probably 2" diameter, in the shape of a ladder, or two ladders back to back. The pipes would be several times larger in diameter than the individual 1" tubes in the nest, which would greatly reduce the tendency to foul, since they would not rust inside, and the velocity of the water through them would be greater than through the smaller but more numerous steel tubes. They would also be far more thermally conductive than the mild steel tubes currently in use. I am concerned with the fact that dissimilar metals would be immersed inside a large steel tank, but since it is filled with distilled water, would this be an issue? The tube bundles themselves were pristine on the outside after being submerged in the rheostat tanks for over 5 years. I measured the resistance of the water to approximately 500K Ohms. I check the specific gravity of the water almost weekly and it does not vary.

Another solution I have considered is to cool the rheostat water directly by pumping it from the tank, through a fan cooled water to air heat exchanger, then back into the rheostat tank. There are two 1" valves on either side of the tanks for draining. I was thinking I could pump from one side, put the water through a heat exchanger, then put it back in through the other side. The big bonus would be the elimination of the need for process cooling water, and the risk of it entering the rheostat due to a leak. What I am concerned about here is the fact that I may be introducing contaminants in the form or metal ions (?) from the pump and piping into the distilled water, although there is some rust already in the water which has fallen in from parts of the tank cover. I am also concerned if water turbulence in the tank would upset the operation of the rheostat itself.

A few other quick questions:

Is there a way to calculate how much heat is being generated within the rheostat based on voltage/amperage conditions while the rheostat is not fully closed? We sometimes have to operate the rheostat at partial load when we use the 2MW exhauster as a soft start for a 5MW exhauster/compressor through a suction turbine attached to the rear of the 5MW machine. It may remain in the partially closed position for several minutes.

Does temperature affect the resistance value of the water?

Any opinions you can offer would be greatly appreciated.

I realize this is old technology, but it was installed in the late 40's or early 50's and it works extremely well. Frankly, I am amazed that you can run 2300V through a steel tank filled with water and not electrocute everyone within 20 feet of the things, but obviously I am missing out on some basic theory of how they operate.

 

[crshears]

I'm in no position to offer a lot of input on your situation, unfortunately...but that doesn't mean there won't be any.

I once worked in a retired thermal generating station where all of the units had been decoupled from their turbines; a motor generator set was installed for the purpose of running up said decoupled generators [some single-shaft and rated at 100 MW, some twin-shaft units rated at 200 MW] so they could be synchronized to the grid for use as synchronous condensers.

I had the privilege of using this equipment numerous times; it worked extremely well.

The run-up set driver was a WRIM rated @ 3300 HP, fed from the 4160 volt station service; its current was regulated by means of a liquid rheostat [I was one of the few operators who took any interest in the nuts-and-bolts functioning of this equipment; I photocopied the equipment description and operating instructions for myself, and I'll be quoting from them].

The liquid rheostat itself had three bushings mounted in a triangular configuration in the centre of the tank, and the tank diameter was large enough that were there any unbalanced current, its value would not be significant. The bushings were cabled to the load side of a 4160V metalclad type air circuit breaker, and the star point of the rheostat consisted of a movable yoke solidly connected to station ground, providing correct fault detection and relaying co-ordination. The entire yoke could be raised or lowered within the liquid via a manually controlled 600V reversible motor. I wish I had a picture...

Here's a quote: "The liquid regulator is cooled by means of cooling water circulating in cooling tubes at the top of the regulator. As the liquid [a solution of washing soda and distilled water] is heated it will rise and be cooled. A thermostat set at 90°F opens a cooling water valve to provide regulated cooling. Another thermostat set at 160°F brings in an alarm indicating high temperature. Two consective starts can be made before the unit reaches an abnormally high temperature."

I do recall hearing somewhere that the resistance of the solution in the tank varied inversely as its temperature, but I have no documentation to that effect.

I'd be very much in favour of providing an independent radiator and pump to eliminate risk of cross-contamination between the solution in the tank and the cooling water.

Hope this helps.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[waross]

I installed a set years ago. There were three verticle plastic cylinders with a common header across the top of all three. A pump took water from the header and forced it into the bottom of each of the vertical cylinders. We had six mills with two 3000 HP motors on each mill. We had three reostats, two working and one spare. The reostats were switched and could be used to start any pair of motors.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[jraef]

The tank is not filled with distilled water, it would not work. The sodium carbonate (soda ash) is there to create an electrolyte solution that should be specifically maintained at the design solution, as it is the resistance in the circuit. Be that as it may, I'm no chemist, but I know soda ash is used as a water softener, but if you have copper pipes, there can be issues with it requiring maintaining some other specific chemistry to avoid problems. The point is, adding copper pipes may cause you to have to make a choice of living with some new problem you created because yoiu don't want to change the more critical chemistry of the electrolyte used in the liquid resistance starter.

With it not working right now, this might be an opportune time to look for something different here. Instead of the LR control, you could consider adding what's called a Slip Energy Recovery system to these motors so that instead of figuring out what to do with the heat from slip losses, you convert it to electricity that you can feed back to the line source to be used elsewhere.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[crshears]

Both the OP and I stated there is/was washing soda in the mix to provide the proper conductivity...which I suppose means there's agreement that the tank is not filled with JUST distilled water.

As to modernizing to a slip energy recovery system, the OP would have to speak to that; depending on location, the cost of electricity and how often the rheostats are used, the capital cost of upgrading might not be justified.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[LionelHutz]

Use stainless tubing and plates?

To calculate the power, you can assume rotor voltage is inversely proportional to speed and rotor current is proportional to load.

 

[btrueblood]

The conductance of any ionic solution will increase with temperature, so the OP is correct (resistance will decrease with increasing temp.)

"Is there a way to calculate how much heat is being generated within the rheostat based on voltage/amperage conditions while the rheostat is not fully closed? "

Do you know the flow rate of the water in the coils? Heat = mass flow rate x delta-temperature x specific heat

 

[ScottyUK]

Cooling tower water is a horrible brew. Unless you have remarkably good control of the tower chemistry it will either be either a weak solution of strong acid (H₂SO₄) or weak solution of a strong alkali (NaClO) which are both aggressive to ferrous metals, plus some blend of anti-foam, inhibitor, de-flocculant and other stuff. It is rare to find a tower which is truely neutral for any length of time.

You might be better off asking in forum338 because this primarily a materials problem and not an electrical one. Most metals will be sufficiently conductive for this application, so solve the corrosion problem first.

 

[MacMcMacmac]

Ok, so I found a company in Oklahoma which does a lot of work with liquid rhostats, and looking at a few pictures, it would seem that they actually circulate the water from the rheostats through a shell/tube exchanger. I sent them an email and the response was that the exchanger, pump and impreller were bronze, and the pipe was ordinary black pipe. See this picture:

http://pumpandmotorworks.com/solutionpics/lrusa/Sample Shredder Photos_Page_3.jpg

Gneral Electric bulletin GEH3080 referencing thier IC9291 liquid rheostats also reference electrolyte pumps and heat exchangers.

So it would seem to be a fairly straightforward modification to provide cooling with an external heat exchanger without having to worry excessively about corrosion. If we go this route, and I hope we do after seeing how many leaking tubes we have on the old bundles, I have to figure out what heat removal capacity the cooler requires. I managed to find some documentation on the units we have.

Motor and Load Data: Primary 2300V, 60Hz, 3ph, 1350hp @ 1800 rpm. Secondary 1200V, 480A

Inertial Load 34000 lb-ft unit 1, 16000 lb-ft unit 2. Starting torque is 20% of full load torque. At full speed, the exhausters will be loaded at 500hp apiece, fully throttled, (which they are at start up). Start time unit 1, 2.9 minutes, Unit 2 1.1 minutes.

Quote

"Each speed control must be capable of continuously dissipating 350hp. This corresponds to a linear torque-speed load characteristic in case of the motors being used to drive other machines than blowers."

Rating: The rheostats are to have a minimum thermal capacity of 14000hp/min for 50C temperature rise and a minimum continuous disspating capacity of 23hp"

My apologies if this is a lot of extraneous information. The rheostats control the secondary winding voltage correct? If I assume a worst case scenario where 1200V X 480 A = 576000W of energy is completely turned to heat instead of energizing the rotor to do work, I converted this to BTU's/hr and found it comes out to just under 2,000,000 btu/hr. This is within the range of a reasonably sized shell and tube heat exchanger (about 4 ft long). If I have made any mistakes in reasoning, please let me know. I won't be signing off on any modifications, but I'd like to give a reasonably informaed opinion as to how we might proceed.

Thanks for your help.

 

[LionelHutz]

You forgot the sqrt(3) 3-phase factor. I did mention how to estimate the voltage and current in my last post. Using that voltage and current is worst case where the motor is producing rated torque without turning.