R-134 Chilling Unit

Greetings,

I'm spending my summer as an engineering intern in a large chemical plant and have been tasked with modeling 5 large R-134a chilling units. One of the units has presented quite a challenge and I would welcome your expert opinions!

The unit in question is 172 tons with Howden double screw compressor. Both the condensor and evaporator are shell and tube design. The condensor is cooled with 90F water which is controled to maintain a head pressure of 160 psia. The material being cooled is monochlorobenzene (MCB). The plant was designed to provide 1111 gpm of MCB cooled to 65F from 75F. Suction pressure is set a 65 psia. Normal operations are 400 gpm. When flow is increased to 800 gpm, performance drops dramatically, the MCB is cooled to around 70F. I pulled all the data into a spreadsheet and the heat flux on the evaporator actually drops under the higher load condtions. I have speculated that we may be getting in to the area of film boiling under the higher flows.

Before I persue that line of reasoning (and the time to research it) I would welcome any thoughts yall may have.

Thanks
B H Spencer

Hi, Enginerous:)
I think you are getting way ahead of youself, film boiling is something I know little about, but fridge plants I can deal with.
Check the basics, is the fluid entering the chiller above design conditions, that would ofcourse mean higher leaving temps. Also things like refrigerant charge and the general state of the plant, worn compressors and fouled heat exchangers will not give as new capacities.
Also where is your commisioning data this will prove if the plant was ever able to preform to design spec.
You also mention that the flow rate increased to 800gpm, normally on chillers the recirculation pump is fixed flow rate accross the evaporator, is the increased flow rate to the process you are refering to.
What design of a plant is this, ie is the evaporator flooded, I would assume so at the rated capacity given and most inportantly what way is the compressors controlled, you seem to indicate a constant suction pressure, chillers are normally controlled on leaving temp.
This is a snap conclusion based on little facts, but if the plant was controlled on suction pressure and the refrigerant charge was short, capacity would be greatly diminished as the compressors would just off load to match the required suction pressure.
Send us more info, such as original comissioning data and some new pressure temperature and motor current readings at the load condition you are experiencing problems at.

Regards. Andy:cool:

MCB has a specific heat of about 150 J/mol-K, a MW of 112.56, and a specific gravity of 1.11, per a reference the Prof found on the internet. Since he prefers IP units, specific heat becomes: 150 / (112.56 * 1.8 * 1055) * 459.6 = 0.323 Btu/lb-R. A gallon of MCB should be 1.11 * 8.33 = 9.25 lbs/gal. So we have 1111 * 9.25 = 10,277 lbs/min of this stuff flowing at design, giving us a load of 10,277 * 0.323 * 10 = 33,195 Btu/min or 166 tons. Mmmmm, numbers seem to work out if the 172 ton figure is for the full flow rate. So lets call 400 gpm a 60 ton load, and 800 gpm 120 tons. Ok, need some operating pressures, superheats, etc. to understand why the chiller isn't functioning at the 800 gpm flow rate.

Andy,

Thanks for you quick reply. Some of your questions I can answer right off. Yes the compressor is controled on sutcion pressure. A slide valve is used to unload the compressor. The flow rate I refered to is the process flow of MCB. A hot gas bypass valve engages at 15% slide valve opening to keep some load on the compressor. The position of the slide valve along wih the current drawn by the compressor motor are all recored every 5 seconds in our process data base. I'll have to retreive those values for you later (today is my day off).

The chiller is indeed a flooded model with a level control. Condensed refrigerant flow into a reciver and from the receiver to the chiller. One problem I have not been able to resolve is the level in the chiller. When I first started looking at this system it was at 9 inches on an 18" chiller. I have had the level raised to 11 inches but the controls will not allow it to go any higher. I'm have the valve and controls check out. But at even this reduced area, the machine should be able to handle the load, at least thermodynamically.

I unfortunatley do not have any commission data only the TEMA sheets on the exchangers. I gather what I can and get back with you.

Thanks,

Hi, Enginerous:)
After reading your latest post thinngs are a little clearer.
The refrigerant level in the chiller would be a due to the design of the chiller, if the chiller has not got a separation vessel fitted above it to eliminate refrigerant liquid carry-over then the top 40% aprox of the surface area would be free space with no tubes in it for liquid separation.
Are you definately using R134a as the primary refrigerant? This would seem a strange choice to me especially given the use of Howden screw compressors, which are a heavy industrial compressor, where R134a is a good choice for domestic refrigerators.
If the system uses R143a oil rectification would be essential from the chiller barrel, this would usually involve taking a small amount of oil rich refrigerant from the liquid level on the chiller and returning this mixture through a heatexchanger to boil of the liquid refrigerant leaving only oil to be returned to the compressor suction line.
If the refrigerant level is low in the chiller the dip tube for oil return will be above the refrigerant level rendering the oil rectification useless. An oil logged evaporator would give very poor performance just as you have indicated. The oil lost to the chiller body would be noticed by maintenance personell, they should be keeping a record of oil and refrigerant added and removed.
Another possible cause of you problem would be a reduced flow on the evaporator pump circuit, ie a faulty evaporator pump or just simply a blocked/partially blocked pump inlet strainer, again a maintenance issue.
Don't forget to post us more info and in the mean time I hope these pointers help you.
Regards. Andy::cool:

enginerous,

I'll throw this in for consideration. Your chiller @172 tons would require 412 gpm for 10 degree delta T chiller barrel given clean tubes and good water. Anyway, at 800 gpm the velocity of the water is most likely out of the realm of engineering design for this chiller. As velocity increases, Delta T decreases as the water is flowing just too fast for heat exchange. The Manufacturer should have water flow charts for the chiller barrel if you can get the barrel info to give them. I don't know how high their charts would go. 800 gpm is pretty high. But, if 400 gpm water is 10 degrees delta T, then 800 would be 5. Anyway............

Hi, twcpipes:)
I think from reading an earlier post the flow rate doesnot change across the evaporator, the change in flow is on the process pump side, i.e. the recirculation pump which services the evaporator is a separate fixed flow rate pump, pumping from a baffled holding tank or a separate cold suply and warm return tanks. The process pump(s) then pump the fluid from the cold suply tank to the process(es) and return the fluid to the warm return side.
Regards. Andy





:)

Andy,

>The plant was designed to provide 1111 gpm of MCB cooled to 65F from 75F. Suction pressure is set a 65 psia. Normal operations are 400 gpm. When flow is increased to 800 gpm,<

I suppose I read that wrong by taking it out of some context. Hmm. I "assume" (what a dangerous word, LOL) the other four chillers are not having a problem. By the way, are they all 172 ton chillers I wonder. I will monitor the future conversations and it may come clear to me. I lost my head, but I will be alright in a minute. :>)

Hi, twcpipes:)
Don't worry about it, I had to read the post several times then ask for clarification. You have put me thinking, I wonder how the other plant are performing and as you say are all the plants alike, I suspect not.:confused:
Regards. Andy:)

Andy and enginerous,
It would be great to have a single line drawing posted (if possible). The 800 and 400gpm are two values somewhere but they are not the 1100gpm posted. I guess that is what has me wondering. Maybe some more info for this old guy to help understand as it is an interesting subject. Not just the chiller but the whole system. It may help others also. I started my first screw chiller in '73. It was the fifth Dunham/Bush (flooded) ever made. Screws are a real trip anyway.

Gentlemen,

Thanks for all your interest. I have a bit more data and the answers to some of your questions. This unit is acutally the smallest of the five. The other 134 unit is 250 Tons. The other machines are R22 units and are each in excess of 1000 Tons. In those areas we cool the MCB to about -23C.

Again, these all run at or near capacity constantly. The problem unit is more of a batch process. Andy, I did jump the gun abit with my film boiling conjecture. A subsequent calculation revealed that it was not possible in this case. I am fairly confident there is not a thermodynamic reason for the malperfomance of this unit. It must therefore be a system problem.

The compressor is powered by a 200 HP Marathon motor (460VAC, 3 phase). Under high load conditions (800 gpm MCB) the following was observed:

Discharge pressure: 174 psia
Suction pressue: 65 psia
MCB Temp in: 24.68C
MCB Temp out: 21.31C
Suction Temp: 17.85 C
Comp. Motor Amperage: 127
Slide Valve: 98%

Under a low load of 400 gpm MCB

Discharge pressure: 172 psia
Suction pressue: 65 psia
MCB Temp in: 21.5C
MCB Temp out: 16.5C
Suction Temp: 16.14 C
Comp. Motor Amperage: 84
Slide Valve: 13%

I appreciate your help and look forward to your input.

Thanks

Hi, Enginerous:)
Thankyou for posting back, I have carried out a few basic calculation (all I able to do) and I can see that the compressor performance is not what it should be.
I have calculated that it would take two Howden 165/145 compressor to acheive the 172 tons of refrigeration you are looking for at the given design conditions. Absorbed power given for that capacity would be 108 KW, that equates to a line current of 162 Amperes, your compressor motor is only absorbing 127 Amperes. Looking at this further the 200hp motor is just under 150 kW, at present you are only abosorbing 84 KW, a motor would not have been fitted nearly twice too big, as this would lead to a very in-effecient plant.:(
So we have established the compressors are not performing, what is the problem, here is a few possibilities.

1/Incorrectly set V.I as from the factory new.
2/ A out of calibration suction transducer combined with a compressor not actually at 98% slide position, ie the LVDT or the micro switches are not correctly calibrated.
3/Worn compressor not pumping as new.

What model is the compressor and what hours has it run and more importantly at what hours was the compressor last inspected.:(

Also is the other compressor running at the same conditions that would also help us determine where the problem is.
If both compressors are running as above I would estimate that the new system capacity is around 134 tons not 172 tons how does that calculation compare to the duty you perceive you are getting.
Ofcourse the easy way to determine you actual system duty would be to fit a flow meter to the evaporator pump and to measure the flow and return temperatures, all this data could be used to calculate the system mass flow and from that the energy removed. :)
Regards. Andy

Enginerous, you are reporting pressures in terms of psia rather than psig, which is unusual. There is a considerable difference. It would be clearer to think (and report) in terms of saturated suction temperature (SST) and saturated condensing temperature (SCT).

First, you must be certain whether the pressures you are observing are psia or psig, then determine the corresponding saturation temperatures using a P/T chart.

Gary and Andy,

Thanks for you prompt replies. I'll answer Gary first. Absolute pressure is common in a chem plant environment since we often deal with vacumes and pump suction head. The saturation condenser temp at this pressure is 45 C and the evaporator temp at this presssure is 12.3 C (from data provided by DuPont).

Andy, I agree that is a small amt of power for such a large motor. There is a bit of confusion in our corresdondence. This unit has only one compressor. The slide valve position is correct, it was calibrated just a few weeks ago. I'll have to review maintence records to see when this unit was last overhauled. The four degrees of superheating we see on the suction side lends some creedence to your idea that the suction pressure could be incorrect. One of the ME's in my office also suggested that the motor my have a "dead leg". The current measurments are made on only one leg. I have an electrician checking it tonight.

Thanks for all your help and I'll keep you all posted

BHS

Given the conditions with 800GPM, there seems to be excessive temperature difference between the SST and the MCB Temp out (9K TD). This would generally indicate lack of heat transfer between the two. Possible oil logging? Is the piping correct (counterflow or crossflow)? Is the piping (both MCB and refrigerant) the same as the other units? Does the MCB enter the top and exit the bottom or vice versa?

Gary,

The design is essentially the same as the other units. The refrigerant flows cross flow and enters at the bottom and exits the top through a surge tank. The MCB enters the bottom and exit the top on the some end; there are two tube passes. That paticular exchanger is 20" OD with 308 0.75" OD x 192" plain copper tubes.

BHS

Has the system ever worked correctly?

Hi, Enginerous
Hi, Gary
I tend to have the same feeling as Gary, that the system might never have worked.
As for the compressor Amperes it get's worse in that we should be consumming about 134KW with amperes around 200Amps. Enginerous, try and get hold/hire a Kwh measurement instrument that has a true RMS reading, with this the absorbed power in KW can be read off directly and stick a gauge on the surge drum this will allow you to see the evaporation pressure and to read it off directly in saturated temperature, the same can be done with the condenser to give a Saturated Condensing Temperature.
Also have you been able to obtain flow rates, actual not designed accross the evaporator on the secondary fluid side, that will indicate is the problem on the primary refrigerant side or the MCB side.:confused:
Regards. Andy.

For the 800 gpm flow, power consumption (assuming a power factor of 0.8): 460 * 127 * 1.73 * 0.8 / 1000 = 80.8 kW. On the refrigeration side: 800 * 9.25 * 0.323 * (24.68 - 21.31) * 1.8 = 14,499 Btu/min = 255.0 kW which gives us an apparent COP of 255 / 80.8 = 3.16, which doesn't seem bad. For the 400 gpm flow, power consumption: 460 * 84 * 1.73 * 0.8 / 1000 = 53.5 kW. On the refrigeration side: 400 * 9.25 * 0.323 * (21.5 - 16.5) * 1.8 = 10,756 Btu/min = 189.1 kW which gives us an apparent COP of 189.1 / 53.5 = 3.54, an improvement of 12 percent, and backwards from what one would expect. Something interesting is going on here.. :)

Although we are lacking condenser outlet subcooling information, what I am seeing at this point is insufficient heat transfer between the MCB and the refrigerant. If we were getting good heat transfer, the SST and the MCB out temp would be much closer.

To ( I don't know at this time ):

Under a low load of 400 gpm MCB

Discharge pressure: 172 psia
Suction pressue: 65 psia
MCB Temp in: 21.5C
MCB Temp out: 16.5C
Suction Temp: 16.14 C
Comp. Motor Amperage: 84
Slide Valve: 13%

These readings, perhaps, need to be retaken. Also, still wish I could see a single line drawing if this is a primary/secondary loop or single loop. The MCB is almost equal to the cooler temp. That's zero approach in a single loop. I don't know what you would call it in a P/S loop. Maybe I just don't get it, but I am trying.

As I understand it, this is a single loop system, the "Suction Temp" is the suction line temp, and the suction pressure converted to saturation temp (SST) is 12.3C, for an approach of 4.2C.

As always, reporting SST and SCT makes for clearer communication and diagnosis than does suction and head pressure.

Here is the (minimum) list of temperatures I use when diagnosing any system:

Low side:

Evap air/water in temp
Evap air/water out temp
SST (saturated suction temp)
Suction line temp near compressor

High side:

Cond air/water in temp
Cond air/water out temp
SCT (saturated condensing temp)
Liquid line temp near receiver/condenser

Hi, Professor:)
I tend to agree the COP is out, why are we getting at good COP at 15% slide position (well better than at load) this would be un heard of that a screw would give good COP performance at these conditions. I would say that the performance at full load is about 76% of what I would expect it to be.:confused:
The selection program for the Howden compressor has given a COP of 4.5, on reflection that is probably a little high, I don't see an allowance for power factor correction anywhere in the selection. If I correct for power factor at say 0.85 the COP comes out at about 3.8, that I feel would be about correct for a syatem such as this.:)
What do you feel Professor?
Nice to see a post from Marc, take heart mate the summer is nearly over.:D
Regards. Andy.

I would pull the end caps and rod out both tube bundles, then see what the numbers look like. The condenser is running a little high, too. Especially considering it is running at light load. But then we don't have the cond water in and out temps. In any case, it wouldn't hurt to clean them both.

Hi, Marc, hi Gary:)
My gut instinct goes with Gary and the fouled heatexchanger. I once went to a job which was doing similar things to this except control was leaving temp not suction pressure. Anyway it was reduced flow due to a partially blocked evaporator pump strainer, cleaned it, fitted it and to coin a phrase she flew.
I have attached a basic flow diagram(very basic) Enginerous has posted that the flow varies on the process side, but that the evaporator pump is fixed. The attached drawing is my best guess at how the system is piped, the GPM meter must them be fitted on the common MCB return.
Or again Marc may be correct and we have just a closed loop system with the plant loading up on suction pressure as the flow rate increases, the only one who can say is Enginerous:)
Regards. Andy[

Hi, Marc:)
You are probably right able the system piping, I just added the common way to pipe such a system for discussion. I have posted the diagram again, hopefully a little clearer this time.:)
A compressor fully loaded at the design suction with only 75% flow rate would indicate a problem with the compressor, what about something as simle as VI, if the full length of the rotors was not being utilized as per design, that would cause a similar effect to what we are seeing. Also if the compressor was damaged the discharge temp would rise picking this up.
Evaporator superheat is also high, indicating something wrong there. If this was my screw I would be inclined to do an annual inspection, to determine slide end stop and what wear is on the slide v and the rotors nevermind the bearings.:(
Regards. Andy

Hi, Marc:)
I have read you post carefully and hopfully I have learnt something.
Most noticible is that the approach has risen beyond design even at 800GPM, in laymans terms the chiller is not able to grab the heat from the fluid, something very wrong in the heat-exchange from the MCB and the refrigerant.:(
Another possibility can be ruled out is that the evaporation is lower than stated (calabration of suction tranducer out) as this again would increase the LMTD. One possibility that would fit is that this compressor has an LVDT fitted and that in actual fact this is showing 100% when the compressor is at a lower capacity. Enginerous stated that this had been calabrated, If I was on site I would energise the load solinoide with a suply just to check anyway.
Another thought that can to me was the slide loading cylinder, but if this was a WRV Howden they are equiped with oil load and oil unload. A WRV on the other hand is spring return to unload, if the slide piston seal is worn oil would bypass at greater rates the further the slide was loaded, a point could be reached at which the spring would defeat the oil pressure before full load is obtained. The LVDT calabration is dependant on 100% and 0% as refererance you load to 100% and press a button, what if the slide had never reached 100% you would be given a false 100% load position, this could be our compressor problem.:o
But as we are agreeded on we also have an evaporator heat exchange problem. Seems this is indeed a sick plant:(
Regards. Andy.

Clearly there is a heat exchange problem. This should be first resolved, then a new set of measurements taken.

It has been my experience that eliminating one problem at a time works better than trying to diagnose and repair all of the problems at once. I would liken this to bringing a picture into focus one click at a time.

I agree, Marc. Accurate measurement is essential to accurate diagnosis. :)

Hi, Marc:)
I think the most important item I have seen posted yet was the line that you posted about the service mechanic and his instruments. Someone needs to take a look at this plant to determine is the problem a design or a maintenance issue or both.
It's just one of those plants that I would love to get my hands on, making something work that others have failed at is right up my street. Unfortunetly this plant is a little far away for me and my service van Enginerous would be better getting someone local to at least take the essential readings. Thinks I would like to see which we have not seen to date would be.

1/Discharge Temp
2/Liquid line subcooling
3/Flow/GPM accross the evaporator from another calabrated instrument.
4/Compressor absorbed power from a KWh meter.

Other information that would be essential is

1/Compressor Model
2/Type of oil cooling employed

Just another thought if this is an XRV Howden they are a different design to the old WRV compressor with spring capacity slide return, a leaking loading cylinder would mean the oil pressure might not be able to defeat the spring and 1005 load would not be obtained.
Also a failed XRV (not uncommon) would mean oil carry-over at excessive rates, fouling the evaporator, both problems accounted for.
Regards. Andy
:(

Hi, Marc:)
hmm well he might go for the flights and the lodgings, but as for the Big Whoppers, I think you could bankrupt the guy:D
only joking
Regards Andy:cool:

Andy, I too wish you could have taken a look at this machine. Well, my time is over at this plant and I left them when a complication of all of your suggetions. As an aside, there was definatlely a problem with the RTD on the suction side. The amount of superheating being measured was not physically possible.

Thanks for all of you for your help. It is a shame that every engineering problem does not have a forum of experienced professionals to go to for advice.

Thanks again,
Enginerous

Hi, Enginerous:)
Nice to hear from you again. I hope that your stint away from uni has helped you see the practical aplications for you new found skills. The only problem for us now is we have seen half the movie with the end cut out. If you hear what the problem was or have any other items to discuss let us know, I personally have found this interesting and rewarding.
Must go lawn to mow:D
Regards. Andy.