I realize the first part of my title almost sounds like a nerdy version of Steal This Book, but it’s actually not. But, after writing the previous post, I thought it might be helpful to walk you through how I made the decisions to break my own rules and move away from a pure “ladder on its side” diagram. And, because I am talking about the reasons for breaking the rules, which are generally related to operation and commissioning issues, this post also will provide some insight into trouble shooting condenser water system problems.
For this example, I will use the condenser water system from a project in Berkeley California. Here is the physical arrangement of the piping as shown on the contract documents so you can contrast it with the system diagram variations that I will discuss.
And, here is the schematic of the system as presented on the contract documents. This is the designer’s version of a system diagram for the system in question and is very similar to what I encounter on other projects where a schematic is included along with the physical piping arrangement.
An interesting point is that there is a slight difference between the piping as depicted on the plan and the piping as depicted on the schematic. If you look closely, you will notice that on the piping diagram, each pump connects to the header between the cooling tower basins independently while in the diagram, there is a common section of pipe ahead of the pumps containing future taps for a side stream filtration system.
Minor order of connection details like this can lead to unanticipated operating problems as illustrated in my post titled System Diagrams: Order of Connection Matters. At the time I started to develop my system diagram, this issue had already been recognized and an RFI (Request For Information, a formal way to ask and answer a question in a construction process) had been submitted. The answer was that the configuration in the diagram was correct and that the piping submittal the contractor was to prepare and have approved prior to installation should have a common length of pipe between the collection basin headers and the pumps.
All of that said, there is nothing particularly wrong with the designer’s version of the system diagram. But for me personally, employing some of the concepts I have been discussing to further simplify it makes it easier for me to understand the system dynamics.
It is important to recognize that to some extent, this is personal preference and also could be somewhat of a self-fulfilling prophesy. In other words, since I was exposed to this approach quite literally from the first day I was exposed to HVAC systems and have used it ever since, I have a natural level of comfort with it relative to other approaches. But, having said that, I have also discovered some of the strategies I learn really seem to make a difference in terms of recognizing and diagnosing a problem, thus I have become convinced of their merit in the general case, personal preference aside.
If I were to have drawn the system diagram in a pure “ladder on its side” format, it probably would have come out looking something like the following. Bear in mind, I knew about the RFI and thus show the common section of pipe between the tower basins and the pumps which is not what I would have drawn if I worked only from the piping plan.
There is nothing particularly wrong with my representation of the system in the context of a system diagram. In fact, its probably what I would have drawn and been satisfied with earlier in my career because:
- The diagram embodies in the “ladder on its side” concept.
- You can clearly see which elements are in parallel with each other and which are in series with each other.
- If you were to compare the diagram with the physical arrangement of the piping as shown on the documents I developed it from, you would find that the order of connection is correct.
- Turns (elbows) are only made on the diagram for organizational purposes. In other words many of pipes that show up as straight lines on the diagram are actually contain multiple elbows.
However, because of some of the experiences I encountered when working with open systems like cooling towers, the first draft of my system diagram for this particular system, which was based on the contract documents, looked like this.
There are several subtle but significant differences between this diagram and the pure “ladder on its side” version that I consider to be important in terms of helping me understand the system dynamics, identify potential trouble spots, and troubleshoot operational problems.
Symmetrical vs. Unsymmetrical Cooling Tower Cell Piping
If you study the contract document piping plan, you will discover that the piping to and from the cooling tower cells is arranged symmetrically. In other words, from the point where the return line splits to go to the two cells, there are the same number of elbows and the same lineal feet of pipe on both sides of the tee. And, the tee has been employed in a manner that means the pressure drop from the inlet to either outlet is identical. A similar arrangement exists in the piping leaving the collection basins.
Lack of Symmetry in the Cooling Tower Leaving Water Piping
Asymmetrical piping equates to asymmetrical pressure drop under flow. And, while the pipe length is short and thus the corresponding pressure drop difference is relatively small, when you are dealing with cooling tower basins, where the difference between full make up and an overflow condition can be in the range of 3 inches, a minor pressure drop difference between two basins piped in parallel can lead to operating problems.
Think of it this way; a pressure drop of 1 to 2 ft.w.c. which seems quite modest in the context of a system that requires 50 or 60 ft.w.c. of pump head equates to 12 to 24 inches w.c.. Thus a pressure difference of 1/4 to 1/2 ft.w.c. in the piping leaving two cooling tower basins that were piped in parallel could cause the make up system on one basin to be wide open if it was the basin with the shorter piping run, while the other basin, with the longer run could be over-flowing.
I ran into this problem very early in my career, so its a kind of “hot button” for me. In fact, one of my earlier blog posts looks at the details of this problem if you are interested.
In any case, as a result, one of things that I do when I develop a system diagram for an open system is to include in my depiction the symmetry (or lack of symmetry) that exists in the piping to the distribution basins and collection basins. To do this, I break my rule about minimizing turns in the system diagram and add a few that I would not necessarily need in the pure “ladder on its side” format. The following illustration contrasts the two versions of the diagram.
Note how the modified version shows identical piping runs in terms of length and elbow count to both basins, just like the physical piping plan does.
Lack of Symmetry in the Cooling Tower Entering Water Piping
Its important to recognize that lack of symmetry can also be an issue for cooling towers on the entering water side too, especially if the cooling towers distribute water over the fill via an open basin with holes in the bottom instead of a pipe with spray nozzles. (Incidentally, if you want to understand that difference a bit more or cooling towers in general, Marley/SPX offers a very good handbook titled Cooling Tower Fundamentals for free on their website.)
The lack of symmetry on the entering side of the tower can result in a situation where the flow is not divided appropriately amongst the cooling tower cells. If one cell is short of flow, then it will not deliver the full measure of cooling it is rated for. In addition, there can be issues associated with icing up the fill during extreme weather.
A cell served by an open basin type distribution system that receives more than its share of flow can overflow. This not only results in a loss of capacity. It wastes water and water treatment chemicals and can create a mess on the roof. This issue is further explored, along with non-symmetric collection basin piping, in an article in the April 2008 issue of CSE magazine titled Commissioning on Campus.
Using Balancing to Resolve Lack of Symmetry Problems
At first blush, it would seem like you could solve the symmetry problem using balancing techniques like throttling a valve in the shorter run to make it harder for the water to go that way. In fact, the first time I ran into this problem, that is what I thought I could do. But there are a couple of problems that come up.
The first one is that there has to be a valve to throttle and that is not always the case. Related to this is that the pressure differences we need to adjust for are relatively minor while the pipes we are dealing with and the line size valves that might be installed in them are relatively large. The result is that for a number of reasons including the dynamics of the pressure drop created by flow through the restriction represented by the balance valve and hysteresis; i.e. “play” or “looseness” in the valve actuation system, it may be difficult or impossible to adjust a valve that might exist to create the desired, relatively minor resistance.
To gain some practical insight into this, go to your kitchen sink and try to adjust the faucet so that you get one drip per minute. If you manage to do that and still have any hair left, try adjusting it just enough to get two drips per minute. If you actually make it to that point, then try putting it back to the one drip per minute setting. My bet is that while it is probably possible in theory, you will run out of patience much faster than the time it takes to prove it. The same result will likely happen in the field if you are trying to make a very small adjustment with a very large valve.
Complicating the issue are a number of other factors.
- Many multiple cell towers operating with different combinations of cells active, depending on the load conditions.
- Sometimes, the different cells are of different sizes and, amazingly enough, are mounted with different basin elevations.
- Multiple tower systems often are also multiple pump systems and the pumps are operated in different combinations that interact with each other, the loads, and the tower cells. Frequently, part of this interaction happens because the pumps have different operating characteristics.
Piping Symmetry and System Diagrams; the Bottom Line
The bottom line on the changes I made is that my version of the diagram provided a “memory jogger” to remind me that non-symmetrical piping should not be an issue in this system. That, in turn, would help me develop my test plan and would aid me in trouble shooting if we had issues with basin level control.
Piping Above the Cooling Tower Collection Basin Level
If you study the plan view of the cooling towers and chillers, you will come to realize that the cooling towers and chillers are all on the roof level, and the condenser water piping to and from the chillers is run high enough to let you walk under it. You can tell this from the “BOP +11’6″ above roof” note below the pumps; BOP stands for Bottom Of Pipe. (I guess it could also mean you would get “bopped” on the head if you were 11’6″ tall and tried to walk under it with out bending over.)
This piping arrangement creates traps and inverted traps in the lines which have operating implications, especially for open systems. I touched on these issues in my previous post about elbows in system diagrams and the post that it references about a project where air binding became an issue at some inverted traps.
Because of experiences with problems of this type in open systems, I often break my own rules with regard to elbows and the ladder on its side concept when I develop a system diagram. The method to my madness is just like it was for the symmetrical piping; breaking the rules provides a memory jogger about a potential system operational challenge that I might want to target in my testing or need to troubleshoot at some point in the future.
The first trap-like situation is created by the pipes rising up from the pumps and then dropping back down into the condensers. There are two general concerns with regard to the inverted trap that is thus created; back flow out of the line to the basin when the pump shuts down and air venting.
Back Flow Upon Pump Shut Down
Since the water in the pipe is at an elevation higher than the basin, if the pump shuts down, it will try to drain back down into the cooling tower basin through the pump. In our case, the check valves on the pumps will prevent this from happening once they close.
But there can be instances where there are no check valves between the elevated pipe and the basin. And for larger systems, the check valves can be slow closing. In these instances, the volume of water that is contained in the elevated piping (relative to the basin level) will drain back to the basin, potentially causing overflow, especially if the issue has not been considered by the design of the basins and the set up of the level control system.
And, when the pump(s) are restarted, the pipe has to be refilled. This means that the basin level will be drawn down by the volume required to fill the pipe or pipes at the volumetric rate associated with the operation of the pumps. Since this will usually represent a level change that is much faster than what would be encountered once the system flow rates had established themselves, it can create operating challenges on a number of fronts, including venting the air, which will be discussed in the next section and basin level control. The level control problem is interactive with the level control problem associated with drain-down when the pumps are shut down and is best addressed at the time of design.
A more subtle issue associated with drain down is that the water can create a siphon effect as it drains back towards the basin. In at least one instance I am aware of, the negative pressure generated by the back flow was enough to collapse the large fiberglass mains that experienced it, leading to a catastrophic failure of the condenser water system serving the chiller plant at a hospital at the worst possible time in terms of needing cooling (summer in the mid west).
Any air that is captured in the piping will be difficult to get out with out some sort of venting operation, and if the pipe drains to the basin when the pumps shut down, then it will be full of air that will need a way out when the pumps restart. If the pump shut down and start-up is automated, then the venting mechanism probably needs to be automated too.
Its also important to recognize that air can accumulate in an inverted trap even if the water does not drain from it, especially in a condenser water system, where the circulating water is, by the nature of cooling towers, continuously provided with an opportunity to entrain air. The details of the mechanisms behind this are discussed in a blog post I did in early 2009.
Traps, Inverted Traps, and the System Diagram
There were two locations on in the system we are discussing where the physical arrangement of the piping tended to create an inverted trap and one location where the physical arrangement of the piping created a trap. The first inverted trap was created where the supply line to the chiller rose out of the pumps, crossed the roof, and then dropped into the chillers. To capture that nuance in my system diagram, I modified the ladder on its side diagram to show the pipe on the discharge of the pumps above the cooling tower basin level as illustrated below.
The difference is subtle, but for me at least, it works. When I look at the diagram on the right, I think Oh yeah, the piping above the basins could drain back to them when the pumps shut down, especially if a check valve didn’t hold.
The second inverted trap is in the piping leaving the chillers, where it rises to cross the roof, rises to the cooling towers, but, at the last minute, turns down through a specialized balance valve into the distribution basins.
The conventional trap occurs at the chillers themselves. The condensers represent the low point in the system and the pipe rises to 11’6″ above the roof level on both the incoming and outgoing connections. As a result, there will be a tendency for sediment to accumulate in them, especially when the head pressure controlled valve on the outlet has the flow throttled.
The Bottom Line Taking Symmetrical Piping and Traps into Consideration
The bottom line on all of this is that I broke my own rules and drew the system diagram as shown below instead of drawing it in its pure “ladder on a side” form as illustrated previously.
Reality; the Real Bottom Line
When we got out into the field to actually work on this system, we discovered that it was not exactly piped as drawn. Specifically, the piping into and out of the tower was not symmetric. That caused me to take two steps as a commissioning provider. One was that I modified my system diagram yet another time to provide some memory joggers regarding the potential operational issues.
The other was that it caused me to write a test sequence that targeted the potential issues I was concerned about which included:
- Level control under all operating modes.
- Drain-down and pump cavitation issues associated with pump start-up and shut down.
- Nuisance chiller flow safety trips at start-up due to air entrainment.
The testing, enlightened by the information gleaned by creating the system diagram, did in fact reveal some operational issues including very complex dynamics in the basin control system that can easily lead to simultaneous make up and overflow, the water conservation equivalent of simultaneous heating and cooling.
So there you have it, a guide to step by step guide to breaking the rules with the reasons why to do it.
Type to you soon in the next post, and until then, power to the people, or at least to the system diagrams.
Senior Engineer – Facility Dynamics Engineering