Cooling Tower Flow Variation

Maintaining uniform flow over the fill of a cooling tower is crucial in terms of ensuring that it performs as intended as discussed in my blog post on the topic. But frequently, uniform flow distribution is not achieved in the field for a variety of reasons. These video clips illustrate how flow can vary over a cooling tower under different operating conditions and complement a blog post on the topic titled Cooling Tower Flow Distribution and Variable Flow in Condenser Water Systems.

The Tower

A combination of piping configuration and fluid mechanics results in a two cell cooling tower with a gravity feed distribution system spending a significant number of hours in the year with one cell that has no flow over it but has the fan running.

The Mysterious Case of the Unbalanced Flow

Water flows into a tee serving the two gravity feed distribution basins associated with a cooling tower cell. There are no valves on the outlets of the tee, but mysteriously, only one hot basin receives water while the other remains dry.

Pressurized Feed Cooling Tower Flow Distribution

This video looks at how a pressurized feed cooling tower distribution system works under high and low flow conditions.

Variable Condenser Water Flow Dynamics with Weirs and Non-Symmetric Piping

This video looks at how weirs are used to allow cooling tower flow to be varied with out compromising the performance of the tower.

Go There

The Tower

This first video is a full length motion picture case study sort of thing (you'll laugh, you'll cry, it will change your life).

Some of it is a bit silly because when I did it, I was messing around trying to understand what you could do with the Video Pad video editing software package that I had acquired. And, since I was in a learning mode, the production quality is not perfect.

For instance, one of the things I learned was that different head sets produce different sound levels and sound quality, even though the settings in the video editing software were the same for all of them. So, you may need to adjust your volume settings some as you go through the video.

But there is some (what I believe to be) useful and relevant technical content. To help you locate that, I am providing time hacks (points in time into the video) so you can fast forward to the point of interest and skip the other stuff if you want to.

1 minute, 42 seconds - End of silly opening credits part; At this point, I orient you to the plant associated with the cooling tower the tower piping and configuration, and the flow distribution issue issue.

4 minutes, 37 seconds - Transition to the visual clue that alerted me to the flow distribution issue.

6 minutes, 10 seconds - Transition to narrated video that I shot in the field showing non-uniform flow distribution shifting from one cell in the tower to a different cell in the tower as the condenser water system goes through an operating transition from using a plate and frame heat exchanger for a "free cooling cycle" to using a chiller as a source of cooling. At the beginning of the transition, one of the cells has no flow over it but has the fan running the entire time. This represents a potentially significant energy savings opportunity.

8 minutes, 48 seconds - Transition to a narrated slide set that illustrates what I believe are the physics behind the flow phenomenon in the previous video segment. At a fundamental level, I think the phenomenon can be explained by the difference between the characteristics of open channel flow and fully developed turbulent flow in a full conduit.

15 minutes, 12 seconds - Transition to an illustrate discussion showing how to use a TMY (Typical Meteorological Year) file imported into Excel to figure out how many hours a year one tower cell might have had a fan running with no flow over it.

29 minutes - Transition to a discussion about how to turn the hours identified in the preceding video clip into a ballpark savings projection that would be achieved if you solved the problem.

34 minutes, 5 seconds - Transition to a discussion of the opportunities based on the savings potential and the nature of the climate. This includes a discussion of what the issues are if you run a tower with the fill that is not completely wet.

37 minutes, 37 seconds - Transition to a discussion of the things working against an implementation of an improvement in the specific project that the video came from.

40 minutes, 3 seconds - Transition to conclusions and lessons learned.

41 minutes - Transition to somewhat silly ending credits and acknowledgements.

A couple of points regarding some of the time hacks.

The savings potential discussion at the 15 plus minute point uses the TMY files that are available from the Pro upgrade of the Pacific Energy Center psych chart. I am pretty sure those same files would also be available from any "Pro" version of any of the Hands Down Software psych charts, like the ASHRAE electronic psych chart or the Greenheck electronic psych chart or the electronic psych chart Marriott provides in the AEP program.

In addition, the Excel techniques illustrated in the video could be applied to any hour by hour or bin data file, including the free, raw TMY files you can download from NREL (National Renewable Energy Laboratory) or a real time hourly weather data file you get from NOAA (National Oceanographic and Atmospheric Administration) or from an ASOS (Automated Surface Observation System) site. If you want more information about weather data resources, you may find the blog posts I have done on the topic to be helpful.

If you use the raw NREL files, you will also need the free kW Engineering Get Psyched plug-in for Excel, which will allow you to quickly create all the basic psychrometric properties from the sensible temperature and dew point temperature data that comes in the raw TMY files. In fact, even if you are not using the raw NREL files, If you are working in this business and doing calculations, I think you really will find the Get Psyched plug-in to be an invaluable resource. And you can't beat the price.

The Mysterious Case of the Unbalanced Flow

This case study is centered on a pair of cooling tower cells that looked like they should be totally free of issues in terms of flow distribution when we walked up to them because the piping arrangement was symmetrical. In other words the fitting count and linear feet of pipe leaving the tee in the mains to each cell was virtually the same, all the way to the two hot basins on each cell. But when we opened up the basin covers, we were in for a surprise.

Even though there were no valves in the pipes leaving the flow balance fitting that divided the flow to a cell between the two hot basins associated with it, one basin was running dry while all of the water was going to the other basin. That meant that half of the tower cell's fill was dry and thus, the air flow through it (and the fan energy it represented) was performing no useful function. And, since the dry fill had less resistance to air flow than the wet fill, 57% of the total air flow for the cell was being wasted.

Unlike my first video case study, this video is free of silly opening scenes and closing credits. But, since it includes both video of the problem and content explaining it, I am providing time hacks so you can fast forward to a topic of interest and skip the other stuff if you want to.

The video starts out with a discussion about why we thought there would not be flow distribution issues, illustrated with pictures of the tower piping configuration.

2 minutes, 8 seconds - Transition to video illustrating the piping symmetry and two hot basins served by a common pipe with not valves in it where one basin has flow over it and the other basin is running dry, a condition that exists in both tower cells.

2 minutes, 55 seconds - These particular basins have weirs in them to ensure that at low flow rates, the water will be distributed across the entering face of the fill before the fill deeper into the tower is exposed to water. This helps ensure that there will be no dry fill and this, air flow short circuits at low flow conditions. In the basins with water in them, you can see the weirs in action, performing their function.

3 minutes, 15 seconds - This is a point in the video where I point the camera at a basin that has water flowing in it and then attempt to rotate the camera to the basin directly opposite of it, served by the same pipe which you can then see is running dry. As I rotate, I attempt to follow the pipe so you can see that it really is serving both basins. You see it through the cooling tower fan guard, so it is not as clear as if the pipe was out in the open.

4 minutes, 50 seconds - I transition to some of the manufacturers information on the flow balance fitting - basically the tee in the pipe between the two hot basins that is supposed to uniformly split the flow to give you a better sense of what that looks like.

5 minutes, 30 seconds - I transition to an illustration based on a model I have built of a Marley cooling tower with a similar piping arrangement to explain our theory about why there is not flow to both cells, which is that the inverted trap created by the piping arrangement can air bind at flow rates where the velocity and pressure drops are too low to sweep the air out of it.

6 minutes, 50 seconds - I transition to a model of the condenser water system in a building where a similar air binding issue happened in the mains at roof level on their way to the tower and point you at a couple of blog posts that I did about that particular incident and why it happened.

7 minutes, 45 seconds - I revisit and contrast the cells with flow and no flow over them but give you a frame of reference to compare them using a plan view of the roof and tower cells from Google Earth. As a result, on the plan view, I can highlight the basin I am showing in the adjacent image and give you a frame of reference for where it is in the tower bank relative to the others in the sequence.

8 minutes, 10 seconds - I show some quick field measurements we too of the entering air velocity across the face of the tower fill on the wet and dry sides. This illustrates that a lot of air is going through fill with no water over it, meaning the air is not doing anything useful in terms of rejecting heat from the tower. It also illustrates that dry fill has less resistance to air flow than wet fill since the velocities are higher on the face of the cells with the dry fill.

10 minutes - I close with the fan power equation, which points out that if you could address the problem and get flow to both sides of the tower cell and uniformly wet the fill on both sides, then the savings would be a near cubic function of the reduction in air flow you achieved, so a very powerful energy savings relationship. This is followed by a video illustrating what the entering face of the fill on a tower cell that has good water flow distribution should look like.

A couple of additional thoughts that I forgot to include in the video.

I mention that by nature, condenser water systems will have air in them and that I will get back to the reason for that, but then I forgot. The fast answer there is that by design, a cooling tower endeavors to place the water going through it in intimate contact with air to promote the evaporative cooling effect it is levering to cool the water passing through it. Since air is soluble in water, during this part of the process, air is dissolved into the circulating water.

But, since the solubility of air in water decreases as the temperature goes up, when the cool, air-laden water goes through the heat exchangers in the system, it is driven back out of solution and will tend to accumulate in the high points. There is more information about this in the three blog posts I mention in the video:

In the video, I say that the fan power for a cooling tower cell will follow the cube rule because in the case of a cooling tower cell, the fill is the system and thus represents a fixed system curve. That generally will be true if the water flow distribution through the fill is uniform.

But, there would actually be at least one other system curve, that being the curve associated with the airflow produced by the fan if there was no water going over the fill; i.e. you were operating the fan with no water flow over the tower. This would be an abnormal condition since if there is no flow over the tower, the air flow is serving no useful purpose. But if you were trying to quantify tower fan power with partially wet fill, then a quick field test to establish this system curve would give you a bounding condition.

If you did that test, you probably should not run that way for long since at full fan speed, the velocity could be high enough to cause the fill to flutter and crack. Alternatively, you could test at a reduced fan speed, which would still give you a point on the system curve and you could use the square law to project the full flow operating point.

Having said that, if the tower has been designed for variable water flow and includes weirs (like the tower in the movie) or cups (Marley's approach to the issue), then the system curve for wet fill that has the full design water flow rate being distributed across it is probably a bit different from the system curve for wet fill that only has the entering portion of the fill provided with water due to the action of the weirs or cups. I have not had an opportunity to measure this difference in the field yet, but it is on my list.

And if the water flow rate has been reduced below that required to provide uniform distribution across the face of the fill with or with-out cups or weirs, they you probably would be operating on a family of system curves that were bounded by the totally dry fill system curve and the totally wet at design water flow system curve.

The cooling tower cell arrangement in the video creates a special case that we should be aware of. Specifically, since there are two hot basins and two sections of fill associated with the fan, then, as we discovered, there exists the potential for a system curve with the fill associated with one basin being wet while the fill served by the dry basin is dry.

Pressurized Feed Cooling Tower Flow Distribution

This relatively short video shows a pressurized feed cooling tower flow distribution system, which is a contrast to the gravity feed distribution systems in the first two videos. In it, you can see how the distribution pattern varies as a function of how far the distribution manifolds are from the end of the header feeding them where the condenser water main connects as well as how a reduction in flow impacts the spray pattern from the nozzles in each manifold.

If you enjoy this video, then you defiantly will want to keep an eye out for our video of paint drying, which is currently in production. (We had hoped to have it released by now but the film crew keeps falling asleep during filming.)

Variable Condenser Water Flow Dynamics with Weirs and Non-Symmetric Piping

These are pretty raw video sequences with no narration or any other edits. But I wanted to get them up here to support my blog post on cooling tower flow variation.

They show how weirs in a gravity flow hot water distribution basin work. Weirs are one of the techniques tower manufacturers use to allow a wider range of flow variation on a cooling tower.

There are a number of reasons for doing this, which are discussed in the other videos and in the blog post I mentioned. Eventually, I will clean this up and add some narration but until then, the blog post provides a pretty good narrative that will give you a sense of the arrangement of the tower and its piping and what we thought was going on to cause the phenomenon we observed.