Estimating Air Cooler Capacity 1

[Aus02]

I need some advice on estimating the capacity of air coolers (in particular, depropaniser and debutaniser condensers). It is well known that the fin fans have a higher capacity (ie maximum duty in KW) at lower ambient temperatures. I need to estimate the actual capacity of the coolers through trials and then somehow use the trial data to extrapolate the maximum duty for different ambient temperatures.

I have had a go at taking measurements on the air side using an anemometer and thermometer, then using this data to estimate the duty. I have also used plant data from the process side to estimate the duty. One issue with this is how to scale up the duty if the pitch or speed controller of the fin fans does not have 100% output. Another issue is the accuracy of both the online temperature/flow instruments on the process side, and the air side measurements taken during trials.

I can also read the current drawn by each fan, and it is possible to use the amps and an assumed power factor to back calculate a duty and then an air flow rate. The problem is, the duty I calculate using the air side data, process side data and current drawn are all different and I am not sure of the best method to use to estimate the duty and then to adjust this duty for different ambient conditions.

Does anyone have any advise on how to do this?

 

[speco]

Dear Aus02,

The first question I would ask is, what do you mean by "capacity"? There are many simultaneous variables in operation here. Can the question be simplified (or is that even helpful)?

Here's one approach. I would presume that you have the manufacturer's data sheets for these coolers. Given that, you might be able to make a few assumptions:

1. The air flow in acfm remains fixed (not applicable if you have some kind of air flow control mechanism such as automatic variable-pitch fans, VFD's, or louvers).

2. The overall heat transfer coefficient remains constant. This may be adjusted if the process flow is widely variable.

3. The process inlet and outlet temperatures remain constant. If this assumption is not correct, then the problem is much more difficult, since you are now dealing with a non-linear condensing curve for the process (the heat load is not proportional to the temperature change in the process).

Now from the data sheets you know the air flow, the cooling surface, and the LMTD (effective temperature difference between the process fluid and the air) at the design condition.

For a small change in air temperature you can assume that the air mass flow remains unchanged. For a big change in air temperature, you should adjust the mass flow by the ratio of the absolute temperatures. Note: as the air temperature goes down, the mass flow goes up as well as the static pressure across the tube bundle. This results in a higher current draw on the fan motors.

Given the above assumptions, you now have enough data to estimate the new heat load (which is proportional to the process flow). You will probably have to reiterate a bit to hone in on an answer.

Given the new air inltet temperature you can make a stab at the new heat load. Using the fixed air mass flow, you can calculate a new air outlet temperature (use .24 as the specific heat of the air). Now you have four temperatures and can calculate a new LMTD. Check the total heat load by comparing your assumed load to U X A X LMTD. (U = overall transfer coefficient and A = surface) If the answers are about the same you have guessed right. If not, adjust your assumption and try again.

There are normally a few built-in safety factors in air-cooled exchanger design, usually in the form of fouling factors. Typically these provide a 10-20% safety factor, if the cooler manufacturer has not chosen to "cheat" on the rating, as some do. Usually these safety factors are enough to prevent performance problems. So, adjusting the heat transfer coefficients for small changes in mass flow should not be necessary.

Hope this helps.

Regards,

Speco

http://www.stoneprocess.com

 

[25362]

If it can help as an addition to what Speco has said, consider the following.

Simplified equations for the HTC by convection on the air side shows "h" as being proportional to ([Δ]T)^1/3 on turbulent flow, and to ([Δ]T)^1/4 on laminar flow.

[Δ]T being the temperature difference between the tube wall and air. It appears offhand the depropanizer, having a narrower LMTD, may profit more than the debutanizer from having cooler air temperatures.