Author’s Note: I just added a page to our commissioning resources website that has additional data on this topic that Brian Clark, Jay Tulley, and Ryan Stroupe helped me gather on an air handling system serving a dental clinic. Of particular interest is a video Jay may with from some logger data and a carpet plot tool I made in Excel that visually shows how the temperature stratification patterns in the unit’s mixed air plenum vary over the course of a day’s operation. (2019-1-07)
Happy New Year!
I hope everyone had a nice holiday and now are getting back into the swing of things. Kathy and I had a great holiday and are just finishing putting away the Christmas decorations, which always seems to take longer than putting them up.
On the professional side of things, last week, I had a chance to do an economizer mixed air plenum stratification test as a field exercise with the current group enrolled in the Existing Building Commissioning Workshop Series at the Pacific Energy Center. Here are the results of one of our test runs.
This test was conducted with the system at approximately 83% outdoor air. As you can see, there is some pretty significant stratification both in terms of temperature (the red-purple-blue carpet plot) and flow (the dark green-olive-tan carpet plot) .
Why Test a Mixed Air Plenum for Stratification?
When I lived in the Midwest (vs. Portland, where it seldom freezes), I typically got to be a “hero” several times a year by using this technique to assess why coils froze, why units were having a hard time staying on line in cold weather, the impact of flow stratification on the measured mixed air temperature vs. the real mixed air temperature NOTE 1, and general operational issues with the economizer. I have also used a similar approach to understand the temperature distribution across the face of a coil under different load conditions.
And, while I often use this technique reactively to figure out why something happened, it is also possible to use it proactively to identify potential problems. For instance, if you look at the temperature plot in the preceding image, the air in the lower right corner of the plenum is at virtually the same temperature as the outdoor air at the time of the test. That means that in that corner of the plenum, in the current operating mode there is little if any mixing occurring.
If the outdoor air was below freezing, its not out of the question that a coil downstream of the filters would freeze in that area, even though the average temperature in the plenum might be well above freezing. If you have ever frozen a coil, then you know that it can be “a significant emotional event” as Jay Santos would say. (Actually, the significant emotional event happens when the coil thaws, but you get the idea.)
Not a Static Situation
What is really interesting and challenging about this phenomenon is that in a working economizer, the temperature and flow patterns will change not only as the outdoor air and return air temperatures change but also as the dampers modulate. This was illustrated by our test results last week because we actually tested the system with the economizer set at three different outdoor air flow rates; 8%, 31%, and 83%. During out test window, the outdoor and return air temperatures were nearly constant. This graphic puts the results side by side for comparison.
The temperature plots are the blue-purple-red plots with bright blue being the coldest points and bright red being the warmest points. The absolute coldest point is the white square with blue letters. The absolute warmest point is the white square with red letters. The various shades of blue blending towards red through purple are temperatures in between the maximum and the minimum. A similar color code applies to the velocity plots with green being the fastest air and tan being the slowest air.
If you want a better resolution version of the image, you can find it on my Google Documents drive by following this link. But even with-out that, if you study the color patterns, you can see that they change as the economizer moves from the minimum outdoor air flow (left set of data and images) towards the maximum outdoor air flow (right set of data and images, and also the previous test results image).
Specifically, the coldest spot in the plenum moves from the upper left corner to the left side to the right side. Meanwhile, the plenum has a pronounced low velocity area towards the middle at 8% outdoor air, becomes more uniform at 31% outdoor air, and then concentrates the flow towards the right side at 83% outdoor air.
Taking a Closer Look at the Test Results
Other items of note in our test results include the following:
Minimum Outdoor Air Flow
The current Minimum Outdoor Air (MOA) setting probably is delivering less than the original design MOA requirement (8% vs. 19%). This may not be an issue since things probably have changed since 1950, when the original design was developed.
But, if the changes were not coordinated with appropriate modifications in the exhaust flow rates, consideration for envelop leakage, and the impact of the building pressures in adjacent, interconnected structures, then the building could experience negative or excessively positive pressures under some operating modes.
The Constant Volume System is Not Necessarily Constant Volume
The system flow rate shifts as the economizer modulates, going from about 49,500 cfm at 8% Outdoor Air (OA) to about 52,600 cfm at 31% OA to about 57,100 cfm at 83% OA NOTE 2.
This is likely due to non-linear damper characteristics, the relative sizes of the dampers (the outdoor air damper is significantly larger than the return damper, thus less of a restriction as the flow shifts to it) and also because at mid-stroke there are two damper sections that have significant flow areas, while near minimum and maximum only the RA and OA respectively are primary contributors. Incidentally, this flow variation will tend to exacerbate the issue mentioned in the first bullet.
For the points we tested the system seems to be at or above the (vintage 1950) design air flow rate of 47,500 cfm.
More on the Shifting Patterns
The temperature stratification pattern shifts from warmer air at the right to warmer air at the left as you go from minimum OA towards maximum OA as mentioned previously.
This is not exactly surprising given that the return dampers are on the left. But the pattern at the lower OA settings is kind of interesting in that it implies the jets from the return dampers are penetrating the jets from the outdoor air dampers and putting more warm air on the other side of the plenum when the return flows are significant. It’s not until the return velocity and flow are reduced that the warm air stays on the return damper side of the plenum.
And if you study the pattern on MOA, you can sort of see the impact of the jets created by the opposed blade OA damper at minimum. On the left side of the plenum, there are indications of warmer bands at the 2nd and 4th row of filters from the top, which is kind of where the jets from the nearly closed OA damper that existed for that test mode are located.
The Temperature Stratification Spread is Significant Relative to the Limiting Conditions
The minimum observed temperature stratification is about 5°F at a high percentage of return air. It expands towards 10°F as the system moves towards higher outdoor air quantities. This stratification exists when the spread between outdoor air and return air temperature is in the range of 20°F, meaning it is pretty significant in the context of the maximum it could be (basically, the difference between outdoor air and return temperature).
That implies that a plenum configured like this in an extreme environment on an extreme day might have 20-30°F of stratification across it if not more. In plenum with “issues” I’ve seen 40-60°F under those conditions with subzero outdoor temperatures and return temperatures in the mid to upper 60’s°F.
And, as you can see, the cold spot can shift around, meaning that to really protect the coil, you would need to cover it pretty thoroughly with the freezestat sensing elements (you would likely need 3 or 4 or more to really provide protection). Similarly, to really reflect the mixed air temperature, you would need to have a lot of sensing element strung around on the coil face.
At the 83% OA position, temperatures at the right side of the plenum are virtually the ambient outdoor air temperature. Not a serious problem in Berkeley, but in Sacramento (or other more extreme environments where temperatures can drop below freezing), you probably would freeze the coil or have trouble keeping the unit running when it was below freezing outside due to freezestat trips.
The velocity patterns are less pronounced, partly due to how hard it is to measure the velocity with all of the turbulence that existed in the plenum. Plus, they are probably influenced by the inlet location of the fan and the configuration of the face and bypass damper type preheat coil downstream of the filters (just like water flowing in a stream starts to spread to move around a rock that is coming up vs. suddenly turning when it gets to the rock).
Having said that, as the system approaches 100% outdoor air, the flow tends to distribute towards the right and bottom. On minimum OA it tends to distribute towards the left and middle. At the intermediate position, it is probably the most uniform. Compared to some plenums I have measured (including the one for AHU1 at the PEC) the flow stratification in this plenum is not too bad. For instance, at the PEC, in some modes, I have seen the maximum velocity at the filter bank up over 1.200 fpm for the high reading and below 250 fpm for the low reading.
The Impact of Mass Flow Rate on Average Temperature
As you would expect, the difference between the average of the temperatures measured and the mass flow weighted average of the temperatures measured (the true mixed air temperature) increases as the flow stratification increases. On 31% outdoor air (the second test) when the flow distribution is fairly uniform, the two values are virtually the same. At 83% outdoor air (the last test), where we saw the most flow distortion, they start to spread out, but the difference is still in the decimal place side of the number.
Our test tends to say that in terms of control (and in freezing climates, protecting against actually freezing something) positioning the temperature sensing elements so that they see all of the variability in temperature patterns is more important that the fact that the temperature elements are not measuring the true mixed air temperature due to the variations in the flow pattern. But, as the flow stratification becomes more pronounced, the impact of that effect increases.
In a situation like the one at the PEC, it could mean you were off 1 to 2 °F based on past experience. Still not so bad when you consider that an averaging element may only be accurate to +/- 1 to 1.5°F out of the box unless you spend a lot of money.
The Temperature Pattern is Critical in Most Cases
So bottom line, understanding the temperature pattern is probably more important in practical terms, especially if you can create one where there are spots where the temperature is equal to the outdoor air temperature (i.e. no mixing, like to the bottom left in our third test) and it is below freezing outside and there is a coil with water in it down stream of that location.
There’s a Reason for All of This
If you look at the physical configuration of the plenum, the reasons for the pattern shifts start to become evident. In this picture, the outdoor air dampers are the dampers on the left, the return dampers are the dampers I was facing when I took the picture, and the filter bank I used as a grid for the tests is just off camera to the right (a higher resolution image is at this link). The plenum is about 10 feet long, wide and high and in plan, looks very similar to the schematic.
As a matter of fact, no matter if I am doing new construction commissioning, retrocommissioning or ongoing commissioning, the reasons I turn to the stratification test are often triggered by physical clues in the way the system and mixing plenum are configured. So, this provides a great transition from an overview of the mixed air plenum stratification issue to a discussion about how you identify the potential for an issue in it in the first place. So that is where I will pick up on my next post in this series.
An Opportunity to Learn More
As I close, it occurs to me that if the topic of HVAC economizers is of interest to you, Ryan Stroupe and I will be doing an all-day class titled Economizers; Design, Performance, and Commissioning Issues at the Pacific Energy Center on February 5th from 8:30 am to 4:30 pm. So, if you are in the Bay Area, consider dropping in. The class will also be offered as a webinar; either way there is no cost and they will qualify for continuing education units if you stay for the entire class. You can register by going to the Pacific Energy Center Classes and Seminars web page and selecting either the San Francisco location or the Internet location.
Note that if you go to that link as of the date of this posting (January 17, 2013), the spring calendar has not been made active so you won’t find the listing. But that is supposed to change any day now, so just check back later this week or early next week.
I should also mention for those who are reading this post at some point in the future that this is a class we repeat on an annual basis, along with others of a similar nature on Pumps, VAV Systems, Fans, Ducts and Air Handling Systems, Chilled and Condenser Water Systems, Steam and Hot Water Systems, Control Systems, Variable Speed Drives, and other field focused, technically oriented commissioning topics. In addition, the PEC offers a wide array of classes on other topics related to energy efficiency and resource conservation, all at no cost to the attendees, and many offered as webinars. So be sure to check the class and seminar web page if you are looking for educational opportunities related to commissioning, energy and resource efficiency, and building science.
Meanwhile, I hope that the information I have presented here provides some help for those who are dealing with challenges created by mixed air plenum stratification currently as we pass through the coldest part of the year in the Northern Hemisphere.
Senior Engineer – Facility Dynamics Engineering
Its important to realize that the temperature indicated by an averaging temperature sensor assumes a uniform flow profile over the sensor. If the flow is not uniform, then the indicated value will be biased as a result of the mass flow rate.
An easy way to understand this is to think of mixing 99 gallons of water at 100 °F with 1 gallon of water at 50°F. The average temperature is 75°F. But clearly, the result will be 100 gallons of water with a temperature very close to 100°F because you are mixing just a little bit of cold water with a whole lot of hot water.
The 99 gallons of 100°F water represents a lot of energy due to the amount of mass at the higher temperature. And the 1 gallon of 50°F water doesn’t represent much energy in contrast and thus, it does no have as big an impact on the resulting temperature of the mix.
The bottom line is that when you are mixing volumes of fluids, you are blending their energy content. And, while temperature is an indicator of the magnitude of the energy content, you need to also consider the mass to understand how much energy is actually there.
Its interesting to contemplate that system associated with this economizer is a “constant volume” system in terms of its design. But the operating reality is that its volume varies as things in the system move. How much things vary will be impacted by a number of factors including how well things like economizer dampers are sized and configured. But for most systems, everything affects everything else and saying they are constant “anything” is a nominal reference at best. This is a very important thing to be aware of both in design and in operation.