Estimating cooling tower energy savings using bin data

[bigAlittlee]

A very fundamental to HVAC, though not so simple problem I am trying to solve using a route I wouldn't normally take. Up until now my experience estimating building performance has largely been with using various energy modeling programs, with libraries for equipment etc. Well - I got a new job and part of the description is retro-commisioning, which I am very happy to dive into, but using bin data, trend data, and MS Excel has not been my forte. However, these calcs must be transparent and reviewable, so that's the way it has to be.

This project's cooling plant is around 40 years old. Installed tower capacity is 6 cells @ design capacity of 833 tons (which includes oversizing/redundancy as CHW plant rarely breaks 3000 tons, though all 6 cells run under normal operation). Towers are crossflow (Marley I believe) similar to BAC 3000 series or Marley NC with single speed 40hp fans. Design CWS setpoint is 80F with 90F CWR. Carrier has provided documentation on these old chillers, verifying I can take them down to 70F.

Dropping CWS to 70F and saving ~ 10% on chiller consumption is simple enough to calculate. I have plenty of trend data giving current performance, and original documentation to assist in estimating savings due to condenser water relief.

Now what isn't simple to calculate is the increase in fan energy consumption that must be deducted from chiller savings. (Because VFDs will not be considered at this time, we are talking about fan run hours, and the increase of those hours at a lower condenser water temperature.)

Even better - optimal condenser water temperature / balancing chiller savings with fan energy increase to determine best operating point(s) or optimal operating curve vs. wet bulb etc.

I have used one method to come up with results, but I am not satisfied and don't feel I have "fully" solved the problem.

Does anyone know of any good references that I may not have thought of yet for solving such a problem? Basically the goal is to come up with coefficients characteristic of the tower based on operating conditions that I know and can verify, that can be used to predict performance under other operating conditions. "Refrigeration and Air Conditioning" by W.F. Stoecker (1958) for example comes up with fgA/0.245 = constant for a given tower with constant flow rates - this is what I am using as reference for this 2nd attempt at a better solution, but would be happy to hear any further comments/recommendations!

 

[ChasBean1]

It’s a good question. Some rough estimations:

3,000 tons of chiller at an estimated average 0.7 KW per ton is 2,100 KW. Assumptions: 1) Your chiller plant runs at an average of 50% of its capacity during the cooling season. You are above the 55°F mark for half the year, so the plant runs for 50% of the year.

This would translate to an annual average chiller draw of 25% of 2,100 KW, or 525 KW. A 10% chiller energy savings due to 70°F condenser water versus 80°F would then be 53 KW (again, averaged over a year).

Cooling towers total 240 HP, or 179 KW. Using the same assumptions (typically half fans running, half a year in operation), average yearly draw is 45 KW. The weather bin data between 70-80°F can be used as the number of hours of extra operation of the cooling tower fans. I looked up a random location (Illinois, US) and there is about 1,530 hours in that bin, about 17% of the year.

So a 17% increase in cooling tower fan operation is about 8 KW. So back of the napkin estimation – 53 KW saved, 8 KW added. So a 6.6 : 1 benefit vs. cost.

Please note that this is an estimation from scratch and without looking up studies/white papers. There are plenty of people out there who have studied this more in depth and I’d look forward to their input as well.

 

[Shophound]

Another caveat is local wet bulb data and the approach temp of your towers. If you lock in 70 degrees entering condenser water temp (ECWT) 24/7, on certain days of the year, the towers may not provide 70 degrees when outdoor wet bulb temps exceed 70 degrees. So your fans run flat out and accomplish little.

Have you considered a reset strategy according to outdoor wet bulb and the tower's design approach? Granted you stated that VFDs are not in play on these towers at this time, but if that changes, this form of reset may prove beneficial.

 

[bigAlittlee]

Absolutely. The reset strategy is Twb+5F or 70F (whichever is greater). The out of whack humidity sensors will be replaced for this.

The method I ended up using was running numerous 3000 series cooling tower selections with the BAC web based selection software for 2.5F weather bins with CWT before and after and compared kW for each and multiplied by hours.

I am not crazy about the method, but knowing that the selection software is based on test data verified by CTI, it might be as good or better than what I would come up with doing my own math. My results show that the extra fan energy eats up about 25% of the savings on the chillers.

 

[DRWeig]

ChasBean gave a pencil-on-napkin-at-lunch answer that is probably as good as lots of further analysis might prove. It matches closely with simulations I have done with eQuest / DOE 2.2.

Some relevant research from Gil Avery is attached, though maybe not strictly applicable to your case. There is a chart in there of chiller + tower total power versus condenser water temperature for a given load, I think. Forgive me for not re-reading it first...

The Max Cap operation that Gil coined refers to the additional chiller capacity available when condenser water temperatures are modulated down whenever possible. He has some numbers in there for chiller/tower combined energy at different loads. Gil's model has been applied to several large central plants at universities, and the reduced energy consumption has proven his theory very well.

The reset strategy is Twb+5F or 70F (whichever is greater).

I bet that you also have a minimum for that strategy and just didn't tell us what it is. If not, be sure you are coordinating with the bypass valve(s) to keep it warm enough for proper operation.

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[MaxwellsDemon]

Can you guys recommended any reference materials for retro-commissioning, estimating building performance, and eQuest? I'm about to jump into all of that.

 

[DRWeig]

Go join eQuest faction with these folks, and post your question there:

http://lists.onebuilding.org/

Some basic stuff to build your eQuest library is here:

http://www.rmi.org/

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[cry22]

check this
it has a few references that can help

 

[urgross]

RetScreen may be helpful, and the price is right (free).