surging centrifugal chiller, lowering condenser water temp helps. Why 1

[TallCotton]

Our 2000 ton York centrifugal chiller will begin operating at very low load conditions now that the temps here are beginning to fall into the high 30's at night. We experienced surging last winter and my solution was to lower the condensor water temperature to 65F from 70F. I hope some kind soul here can answer the following questions in very simple terms (thanks.)

What is surging?

What exactly is happening in the chiller while it's surging?

Why does lowering the condensoer water temp eliminate the noise the chiller makes at very low loads?

What is changing in the operation of the chiller when the condensor water temperature is dropping?

I also noticed that the evaporator saturation temperature falls when the condenser water temperature is dropped. I haven't pushed it but I've seen less than 34F. Is there a chance that this temp could go below 32F and cause icing of the evaporator tubes?

Also, is this the right way to eliminate this surging problem?

Thanks, TC

 

[SAK9]

Surging is well explained in thread 403-110655 of this forum.

As for the noise elimination( a roaring noise due to surging) when condenser water temperature drops,it is due to the reduction in the "lift"* that needs to be created by the compressor under these conditions.If you look through the performance curves of the compressor you would find a a line sloping down from right to the left .This is called the surge line above which when operated compressor goes into surging.You would note from the graph at low loads,the lift should be less than the design value.So when you reduce the condenser water temperature at low loads,you are reducing the demand for lift thereby pulling away from the surge line.This is why the compressor stops making noise and there is nothing wrong with addressing surge by this method.But It should be done with in safe limits. I would not recommend going below a suction temperature of 36F as frost may deveop.

Anything which increases the lift will cause the compressor to go into the surge zone.These can be:

High head pressure due to fouled condensers,high condenser water temperature,insufficient flow,refrigerant overcharge, air in the system etc

Low evaporator pressure due to malfunctioning feed valves,choked strainers etc

While selecting the centrifugal chillers the maximum and minimum load should be considered.If low load operation is required,a smaller seperate chiller is recommended.Another option is to use multistage chillers(more than one impeller in one machine) which are less prone to surging

*Lift=Head pressure-Evaporator pressure

 

[imok2]

It's a matter of exceeding total lift, Centrifugal compressors all have a total lift capacity and when it's exceeded, surging will result. Surging is when the discharge gas cannot overcome the existing head and reverses it's direction. So if you can decrease the head pressure or increase the suction pressure and lower the total lift the surging will cease.

 

[tombmech]

sak9 and imok2 have explained the mechanical phenomenon much better than I can. I would add, however, that higher pressure machines are inherently more tolerant to these effects, because there's a much higher pressure gradient available within the refrigerant cycle.

Also, variable-speed technology in the centrifugal chiller industry has lagged behind the rest of the industry, with the notable exception of York's Turbo-Modulator (whether you preferred it or not). That is really the ultimate solution, and more mfrs are getting onboard.

 

[krb]

McQuay has a paper on centrifugal chiller fundamentals that have some graphics to go along with what has been said above. You can download it here:

http://www.mcquay.com/mcquaybiz/literature/lit_corporate/AppGuide/AG_31_002.pdf

Chillers are more prone to surge at lower loads, as the velocity of the refrigerant leaving the compressor is slower, and the pressure developed in the diffuser is less that the saturated condenser pressure resulting in surge. You can also increase the chilled water leaving temperature to reduce lift, and help with any icing problem you may have. Your process will dictate if raising the LWT is possible. If it is HVAC, lower humidity in the winter will probably allow you to do this.

KB

 

[spector]

Is this a low pressure chiller?

If so first check for low charge and if you have air in the chiller....If so, investigate the purge unit for proper operation....A faulty purge unit can cause both.

Your low evap temp has me concerned. Please post some more readings. Include chilled water supply and return temps, condensor water supply and return temps along with sat temps for both evap and condensor...Also compressor discharge temp....Also refrigerant pressures in evap and condensor.

Thanks

 

[TallCotton]

Thanks everyone for your input and leads to further reading. Spector: I will print the information out tomorrow morning at 7am. It would be better if I was there a couple hours earlier but I won't be there at 5am unless someone makes me I'll post the info tomorrow evening around 6pm PST. By the way, I have only seen 34.XF once last year, as I watched today saturated evep temp hovered just above 36F all day. I'm OK with that.
TC

 

[TallCotton]

As I promised:

...........................4am..............7am
leaving chw temp...........36.8.............37.5
entering chw temp..........41.9.............44.0
evaporator pressure........31.9.............32.3
condenser pressure.........73.2.............73.6
oil pressure...............35.3.............35.8
entering cw temp...........64.5.............63.9
leaving cw temp............69.9.............70.3
motor current..............56%..............59%
leaving chw setpoint.......37.0.............37.0
evap saturation temp.......36.4.............36.9
cond saturation temp.......71.3.............71.8
discharge temp.............98.0.............98.0
oil temp...................126..............124
OSA temp...................44...............44
plant gpm..................1491.............1754
plant chw temp.............36.7.............36.5

Sorry for my lack of coding skills, but here is the data from this morning. Maybe you folks can see some relationships in the above that can help.

 

[spector]

Apparently you are running a low temp chilled water system where you have reduced AH/fan capacity and also low capacity CW distribution....The idea of that type system is to reduce the energy required to circ 44 degree water...
The theory is lower supply temps of 36-38 degrees require smaller CW pumps and piping and smaller AH's with smaller and slower fans.

In that case, then 36 degree CW is right on the money and so is your 37 degree setpoint.

Your readings do not indicate a problem. I see no evidence of air or a low charge.

 

[HVAC68]

Another reason why surging occurs :

While selecting a chiller - one needs to check for part load operation. If your part load operation requirement is down to (say) 30%, the general tendency is to select the same based on ARI reducing condenser water temperatures. In reality, when the load on the chiller is 30%, it's not necessary that the condenser water temperature is that low, thereby surging may occur, since the compressor cannot perform.

I have always selected my chillers at constant condenser entering water temperatures at part loads as well - some people say it's an overkill, but then you are sure that the machine will perform at part loads.

HVAC68

 

[TallCotton]

Thanks all for your responses. To Spector: I don't know that this system was designed for these chilled water temperatures or not, but my feelings are that it was not. I feel this way because our contracted plant supply temperature is 42F. However, we have been supplying these low teperatures (36-38F) since the day the plant began operation late 1999.

Over the weekend I looked for a performance curve/surge line chart as was mentioned by SAK9, but could find nothing in the data I have available to me. So, I spoke with the York technician who services our account on Monday and asked about a chart with a surge line indicated on it and he acted like there was no such thing. HMMM.

The York tech also doesn't think the chiller is surging at low loads either. The chiller will get noisy and and rumble (how can I possibly describe the noise) when it is at somewhat less than 50% total motor current at these OSA tmps. The York tech feels that this is the compresser "gulping" refrigerant vapor through the hole in the center of the closed vanes. He also says that the chiller can do this indefinately without any damage, no matter how noisey it gets. Maybe he's right, I ask you. Any unusual noise is a bad noise to me. Maybe I'm wrong, again, I ask you.

I'm just a plant operator, labled a stationary engineer. Unfortunately, I have to learn by my mistakes, but I'm hoping to make as few as possible and that's why I'm here picking brains.

These low supply temperatures have kept us from going into free cooling mode, as we have a difficult time getting our tower water cold enough to meet and then maintain our regular supply temps.

We supply water to several flat plate heat exchangers on our various customers properties, but we don't have total control of the entire process. They monitor the discharge temps of their HX's which then control butterfly valves on the supply side of the primaries of those HX's. I've been told this control scenario is troublesome, your opinion of why this is, and possibly how you would control those exchangers if you could do it any way you wanted would help me understand the big picture here.

Chiller model YKWFVDJ4-DHDS, Compresser model YDHA-90DD

Thanks for reading such a long post. TC

 

[mfi2000]

Why are the chilled water temps so low. I've run many plants in NYC and never supplied this low water. Even when supplying 7 million sq. ft. of building space we supplied 42 water. These machines are not really designed for this, although they can produce it. If the original design specs where for 42, then this might be the problem. It is also why you cannot switch over to plates. It's tough to pull down the condenser water to these temps. As far as the HX control scheme. If I understand it right. The secondary leaving ch water temp controls the inlet valves on the primary entering side. If this is so, and you have no bypass then you're system is a variable flow. In other words the amount of chilled water entering your chiller varies according to what the plate exchanger is calling for. This can play havoc with the chiller as it is always adjusting itself to the varying load.

 

[spector]

TallCotton,

I am very familar with the noise you just described, now that you described it...I listen to it daily each winter.

Only my #2 chiller sounds like that tho, and it is fersure not caused by surging...It, like yours, is a York high capacity open drive...It has been operating flawlessly for 10 years but always makes that same noise you described whenever it pulls down to 45% or less load. My York tech says it is normal at low loads. 10 years of perfect operation says he's right.

My other two chillers are older LP(low pressure) types and when operated out of the envelope they will announce it with a classic loud surging sound...But even then, each have very distinctive sounds when they are happy or sad.

My Carrier hermetic LP when loaded to 100% is so silent that you have to place your hand on it to see if it's running. But My York LP open drive screams at 100% due to the motor fan noise.

As for your CHW set points, I won't venture a guess. Who knows who decided or why it was decided that you supply 36-38 CW with that chiller. It is usually only done when the CHW system, fans and AH's are "downsized" to reduce the energy required to circ the CHW and operate the AH's where even the piping and HX's are smaller.

If I read your post correctly, you state that you also get CHW from a contracted chiller @ 42 degrees...And I am assuming that the CHW from both chillers is blended and then sent to a common load. If that is the case, I can understand your confusion.

However, if the load is not common, then it is possible that one zone is designed to use low temp CHW supplied by your chiller, and the other designed to use higher temp CHW supplied by the contract chiller.

 

[krb]

Does this chiller have a microprocessor controller on it? If so, it should be monitoring chiller motor amps and if the amps are fluctuating, it would determine a surge condition. On Carriers, the amount of fluctuation is operator specified, it may be different on your machine. If it is surge, motor amps drop off quickly whenever the surge occurs. I would look in your operators manual for this chiller. Also look for alarms or alerts related to surges on the proper screen (I know carriers and tranes have these)
Your tech either does not know of the compressor map or is reluctant to give you one. I have found that the latter is usually the case.

KRB

 

[SAK9]

How do you know whether the noise is due to surging?

You only need to check the motor current.If it holds steady,it is not surging.If it is hunting,it could be due to surging.

As for the low chilled water temperature ,your chillers seem to be on a district cooling sysetm where supply temeperatures of 36 F are common.The temperature on the secondary side of the PHEXs will be around 40~42F.The primary flow on the PHEX is normally controlled by the secondary chilled water outlet temperature as you described in your last post.There is nothing wrong with such a system as far as I know .

 

[TallCotton]

Hey all,

This is getting more interesting. Let me try to explain the layout of our system and see if I can get us all on the same page.

We have six 2000 ton york chillers, all the same model numbers as described above. All chiller supply temp setpoints are equal at 37F. Fixed speed primary loop and variable speed secondary loop. Off of this secondary loop we feed 5 plate and frame heat exchangers in parallel. These heat exchangers, at our customers locations,have the inlet sides of the primaries controlled by the temperature setpoints measured on the secondary sides, as was described in my last post. There are no bypasses, and yes it is a variable flow. The chillers are all microprocesser controlled, but as far as I know I have no control over the surge amp setpoints. It doesn't apperar to be a surging issue anymore as I originally thought anyway.

We are currently controlling the staging of chillers using only Plant GPM measured in our secondary loop, allowing the chillers supply setpoints set plants supply temperature. This has been, to me, working well compared to the previous control situation. That's a complicated issue that I would be glad to try to explain if anyone is interested.

I'm glad to hear that this rumbling noise is OK, and unless someone here tells me otherwise, to save some electricity, I will raise the condenser water temperature back up to...you guys pick the number. Right now it is at 65F. I'll listen to the noise...

I have a question about our staging setpoints though that I'm someone here can answer. Each of our chilled water primary pumps is rated at 3700gpm. One of these pumps will come on whenever a chiller comes online, three chillers, three pumps. As an example our second chiller will come on when our plant gpm raises above approximatly 3800gpm, one primary pumps worth of flow, over which point we begin to loose our plant supply temperature (secondary loop temperature.) I would think that we would be able to maintain our secondary loop temperature untill we reached 7400gpm in the secondary loop, 2x's 3700. But this isn't the case. We stage our second chiller on at about 6150gpm. and this number gets farther away from a multiple of 3700 as more chillers are operated. in other words, the more primary pumps that come on the less each additional pump increases the total flow in the primary loop. Primary loop gpm isn't something we are set up to measure. I can't tell which direction water is flowing in the bypass. I can only go by the ability of the secondary loop to maintain temperature. so, why do additional fixed speed primary pumps have an increasingly smaller effect on overall primary loop gpm?

Thanks again. Let me know when I'm wearing you guys out with my questions.

Merry Christmas, TC

 

[mfi2000]

I'll try to take a stab at this as it is similar to the way my plant used to run. Our plant had 3 4000 ton chillers, 4 7000 gpm chw pumps, which fed riser pumps which then fed HX's on up to the higher floors. We also fed 10 other buildings. We also had 2 large bypass lines which we could moniter with flow meters and a forward reverse meter which would tell us how many gpm's we had in exccess or shortage. One chiller could be fed with 2 pumps but would only flow 10,000 gpms. We would try to keep a forward bias on the supply gpm's to keep a resrve if the chilled water valves started opening on a rising load. When it came time to shut down most of the building we would try to be at 0, supply = demand. As you keep adding pumps and machines to the system it can only take what it needs, bypassing the rest. The pumps are riding back on their curves and only supplying what the system can handle. You won't just keep adding 3700 gpm's every time a pump is started because the system cannot handle it. The valves on the HX's are opening and closing which is determining the amount of gpm's the system requires. In a reverse situation the plant is supplying less than the system demands but on a falling load it doesn't matter, the valves are closing so less water is needed. If you had a BMS system that showed all the flow rates through the system you could see all this happening. Hope this helps.