What happens when flow reverses in primary/secondary systems?

Figure 1. Click image for larger view.
Our industry has dealt with primary/secondary piping systems for years. Many hydronic heating pros know how these systems operate as well as how they are designed. Still, based on communications I've had with system designers, it's evident that questions remain. Some that come up quite often deal with flow reversal between the closely spaced tees that connect the secondary circuit to the primary loop. They include:
  • Can flow reversal even occur?
  • Would it cause damage or interfere with other parts of the system if it did occur?
    How might it affect the heat output of the secondary circuits?
  • Should I use a large circulator in the primary loop to make sure flow reversal cannot occur?

The concept of flow reversal runs counterintuitive to the idea of a single primary loop supplying heat to several secondary loops. It's easy to conceptualize a primary circulator as having to produce enough flow to “feed” all the operating secondary circulators.

Fortunately, this is not true. Although the primary loop does supply heat to each secondary circuit, it does not supply flow to each secondary circuit. Supplying heat to a secondary circuit is evidenced by a drop in temperature of the primary loop flow as it passes a pair of closely spaced tees. This is entirely different from supplying flow to the secondary circuit.

In animated terms, the primary circulator doesn't even know the secondary circuits exit and vice versa. Each circulator operates as if the circuit in which it is installed is completely isolated from the other circuits.

Let's look at a typical system in which a primary loop serves a secondary circuit having a design load of 40,000 Btu/hr. Assume the primary loop flow rate is 10 gpm and that the water arriving at the first tee (of the closely spaced pair of tees) is at 160 degrees F. Also assume that the circulator/piping design used for the secondary circuit allows a flow rate of 4 gpm to develop. Figure 1 shows the situation along with the calculations for temperature drops.

Following The Energy

The best method for predicting the output of a situation like this where multiple energy flows converge is to simply account for all energy carried in and out of the process. Just as the flow of water into and out of a tee must balance, the energy carried into a mixing point must equal the energy carried out of that point. This concept is based on the first law of thermodynamics, which states that energy cannot be created or destroyed - only exchanged.

Since the secondary circuit load removes 40,000 Btu/hr. of heat, the temperature drop in the secondary circuit is:

Hence the temperature returning from the load at design conditions is 160-20 = 140 degrees F.

The temperature drop along the primary loop can also be determined by accounting for the fact that heat is being removed at the primary/secondary interface at a rate of 40,000 Btu/hr. Hence the temperature drop in the primary loop is:

Hence the water temperature leaving the downstream tee is 160 - 8 = 152 degrees F. Some of you may be wondering why the temperature drop in the primary loop is different that the drop in the secondary loop. It's because of the different flow rates in these loops (10 gpm in the primary and 4 gpm in the secondary). Heat transfer into or out of a fluid stream is always based on the multiplication of flow rate and temperature drop. For a given rate of heat transfer, if flow is increased, temperature drop (e.g., ∆T) decreases and vice versa.

There is simply no way around these mathematical consequences of thermodynamics, and that's a good thing. Learn it, trust it and then use it to answer many of your hydronic design questions.

It's also possible to calculate the water temperature leaving the downstream tee by applying the “mixed-stream formula,” which again is based on conservation of thermal energy at a mixing point:

In this formula, f1 and f2 are the flow rates of the two streams being mixed, and T1 and T2 are the temperatures of these streams as they enter the mixing point. In the situation being analyzed, the mixing point is the downstream tee. There is a bypass flow from the upstream to the downstream tee of 6 gpm at 160 degrees F. So T1 = 160 degrees F and f1 = 6 gpm. T2 and f2 are the temperature and flow rate of the secondary circuit return (T2 = 140 degrees F and f2 = 4 gpm)(See figure 2). Putting these numbers into the formula yields:

Figure 2. Click image for larger view.

Overpumping Option

About now some of you are probably thinking, “That's all fine and dandy as long as the flow in the primary loop is higher than the flow in the secondary loop. But what happens if this is not the case?” Well, let's take a look.

Suppose a circulator capable of generating a flow rate of 15 gpm was installed in the secondary circuit, and the piping was scaled up accordingly. Figure 3 illustrates the resulting flows.

Figure 3. Click image for larger view.
Notice that flow between the closely spaced tees has indeed reversed. This reversal is the only possibility given the flow rates in the surrounding pipes, and rest assured it does not damage the system in any way.

Also notice that the flow rate in the primary loop for all practical purposes does not change. Why? Because the insignificant pressure drop between the closely spaced tees effectively prevents the primary loop from “feeling” any change in its hydraulic resistance. Remember, the primary loop doesn't even “know” the secondary circuit exits.

An even more interesting result is what happens to temperatures and heat transfer in the secondary circuit. Again, we have to follow the energy into and out of the tees to predict what will happen.

The flow reversal between the tees now creates a mixing point at the upstream tee. Just the opposite of what occurred in the first scenario where mixing occurred at the downstream tee.

This mixing point lowers the water temperature supplied to the secondary circuit. How much? Just use the mixed-stream formula again.


Tmix= supply temperature to the secondary circuit (degrees F)

TR = return temperature from the secondary circuit (degrees F)

Unfortunately, we don't currently know the value of the return temperature from the secondary circuit (TR), and without this, we can't calculate Tmix. We can, however, make a couple more observations that get us out of this situation.

One is that the heat output from the load will increase slightly, based on the increased flow rate through the heat emitters. For fin-tube baseboard, the increased heat transfer can be estimated as follows:

The heat output from most hydronic heat emitters increases very slowly as the flow rate increases beyond typical operating conditions (e.g., conditions that yield temperature drops of 15 degrees to 20 degrees across the circuit). The higher the flow rate, the smaller the increase in heat output. It's definitely a situation with very diminishing returns. This fact strongly supports the argument against using oversized circulators based on the assumption that they significantly improve heat transfer.

Anyway, lets conservatively assume the load is now 5.4 percent higher (42,160 Btu/hr. rather than 40,000 Btu/hr.) due to the flow rate increase. It's a conservative assumption because we already realize that the supply temperature to the load will be reduced because of mixing at the upstream tee.

The temperature drop across the secondary load is now:

Hence the return temperature from the secondary circuit (TR) is:

This can be combined with the mixing equation and solved for the mixed supply temperature:

I'll spare you the algebra: The result is Tmix= 157.2 degrees F. Hence, we've estimated that the supply to the secondary circuit is about 2.8 degrees lower due to mixing at the upstream tee.

Knowing this, we further recognize that the lower supply temperature also lowers heat output from the secondary circuit. How much? Here's an estimate, again assuming the heat emitters are fin-tube baseboard, and a room air temperature of 70 degrees F.

Figure 4. Click image for larger view.
Interesting. The increased flow rate (by itself) would increase heat output about 5.4 percent, while the decreased supply temperature (by itself) would lower heat output about 4.6 percent. In other words, it's essentially a wash. One effect compensates for the other, and any net change in heat output is insignificant.

Figure 4 shows the final picture of the flow reversal situation.

So where does this leave us? Well, we've conceptually installed a larger secondary circulator and managed to reverse flow between the tees. However, for all practical purposes, we've not changed the heat output of the secondary circuit, nor have we changed the temperature of the water leaving the downstream tee. Had this scenario been installed, we would have wasted money on a larger secondary circulator as well as larger secondary piping. We would also have created a system that requires substantially more electricity to operate the larger secondary circulator for years to come. Over a typical 20-year design life, this added operating cost could approach or exceed $1,000. Is this a smart move? Not in my book. By the way, if you installed a larger primary circulator under the premise that it would prevent flow reversal, you are just wasting even more pumping energy.

We've seen that flow reversal is possible in a P/S system, but when it occurs, the heat transfer from the secondary circuit is mostly unaffected, as are the other secondary loads in the system. The only significant difference is a needless increase in both installation and operating cost for the oversized secondary circulator and piping. This is something our industry can do without. As I've stated in the past - head is a terrible thing to waste.