Short Cycle Prevention For Double-Digit Savings (Part I)

Abstract:
Most hot water heating boilers short-cycle. This creates mechanical problems and wastes energy. Surprisingly, most of this short-cycling is designed into modern system. Numerous customers, including many engineers and contractors, have come to us looking for a deeper understanding of the phenomenon and for definitive solutions. This article addresses these concerns, develops theory, and presents bullet-proof solutions to the short cycling problem.

Whats Wrong With Short Cycling?

There are two problems with short cycling, one mechanical, one economic. The mechanical problem comes from the effects of rapid cycling on boiler components. The burner material, for instance, rapidly heats and cools, and sometimes cannot run long enough to dry out. This can create stress and corrosion failures. Gas valves can see decades of use in a few short months. There also tend to be nuisance shutdowns and unexplained flame failures with flame programmer fault codes that have no easily identifiable cause. If you want to make a thirty year boiler fail in five years, and drive you nuts in the interim, short-cycle it. The economic problem is less widely known and appreciated. There is an old rule-of-thumb which says that a short cycling boiler achieves fifteen efficiency points less than the lowest efficiency achieved in nonshort-cycling low fire. An atmospheric flex-tube boiler, for instance, that achieves 72% efficiency at low fire will see 57% efficiency in short-cycling mode. The loss of fuel efficiency is staggering. If you want an energy efficient boiler plant design, there is often more to be gained from short-cycle than from choosing an ultra-high efficiency boiler.

weather hours for a city in the countrys heating zone. A BIN-HOUR chart like this is based on counting the number of hours spent during the year at each outdoor temperature. The data is tabulated by creating bins which span the three degrees, e.g., from 62F to 64F, from 60F to 62F, from 58F to 60F, and so on, down to the minimum recorded outdoor temperature. Obviously, each location will have its own distinctive data array. (The data is available from the National Weather Service, and is normally part of the database furnished with most energy analysis software programs.) All the hours spent within each temperature range are entered for the month in which they occur, and the total for all months is shown in the TOTAL column. The result is a chart of how many hours are spent at each outdoor temperature by month and for a typical year. There are several things worth noting. First, note how few hours occur at the lowest temperatures, i.e., as you approach the outdoor temperature that is used to calculate the heating load. Design conditions are relatively rare, and in many years those conditions dont even occur. Second, note that the hour distribution is multi-modal, i.e., the most commonly occurring temperatures occur here and there across the data (lots of Patterson-Kelley White Paper #3 April 30, 2004 Short Cycle Prevention (Part I) Author: Jack McKeegan 2004 Jack McKeegan www.pkboilers.com

The Load Profile

The facts are these: design heating loads almost never occur, and boilers spend nearly all of their operating hours only partially loaded. Table 1 shows BIN-HOUR data for Detroit, showing a typical distribution of

Table 1. A BIN-HOUR chart for Detroit showing the average number of hours spent at each outdoor temperature in the range of temperature for which heating is normally required. A similar chart can be made for other locations. hours at 31F to 37F, another clump at 43F to 45F, more at 55F). The hour distribution is not linear. Table 2 repeats the first three columns of Table 1, then adds three more columns. 1. %TOT lists the percentage of total hours occurring within each bin. The 269 hours occurring in the 63F bin are, for instance, just over 4% of total hours; the 227 hours occurring in the 39F bin are just over 3.5% of total hours. 2. CUM% lists the cumulative percentage of hours as we move from the maximum outdoor temperature at which heating is required downward to the design temperature. Thus, for instance, the 63F bin covers 269 hours which are 4.17% of total hours. We add the 247 hours occurring in the 61F bin which represent 3.83% of total hours, and have covered 8% of total hours with these two bins. The process continues continued until we have accounted for 100% of total hours. Note that 95% of total hours are covered by the time we reach the 15F bin, well above the design temperature. Another way to say this is that the hours between 15F and the design temperature represent only 5% of total hours. 3. %LOAD converts outdoor temperature to percent of design heating load. These numbers are the result of an interesting thought experiment to which we now turn. Is it possible to directly relate outdoor temperature to percent of design heating load? The short answer is, no. Building loads differ widely because of variations in architecture, use of space, magnitude of internal heat

We can take this analysis further by graphing outdoor temperature against this maximum theoretical load, and creating a series of curves at 25%, 50% and 75% of this maximum. Figure 1 shows just such a graph. A simple curve like this can be a powerful design tool. If light loads represent only 25% of the maximum theoretical loads, for instance, then the 25% curve can help size boilers for good part load performance.
100 90 80 70 60 50 40 30 20 10 0 62 to 64 56 to 58 50 to 52 44 to 46 38 to 40 32 to 34 26 to 28 20 to 22 14 to 16 8 to 10 2 to 4 -4 to -2 -10 to -8 25% 50% 75% 100%

Table 2. This table adds percent of total hours, a cumulative percentage total, and percent of design load. gains, and so on. But consider this thought experiment. Imagine a building in which all heating zones are perimeter zones, where there is no solar heat gain, and which is heated with 100% outdoor air. In such a case the heating load would be proportional to the difference between the indoor and outdoor temperatures. The load on a typical building clearly cannot be larger than this, and would certainly be a mere fraction of it. This represents a worst case scenario by a large measure. It answers this question: what must the load be less than? The median outdoor temperature, the temperature at which there are as many hours above as below, occurs in the 26F to 28F bin. The heating load at that point is only 50% of the design load, which means that at least 50% of total heating hours see a load thats only 50% of design or less. 95% of total hours are over by the time the outdoor temperature falls into the 15F bin, at which point the heating load is only 67% of design. This means that 33% of total heating plant capacity is installed for only 5% of total hours. Remember that the load cannot be larger than this, and is certainly a lot less. No wonder boilers short-cycle. Even if its only part of the overall problem, boilers are just too big for the real world loads they serve most of the time.

Table 2. The maximum theoretical load (100%) along with some load curves drawn at fixed percentages. This leads to the first part of our strategy. Every system should have a jockey boiler, a boiler sized to operate when the load is only a few percent of the design load. This is intuitively obvious when you consider that fire pump systems, chiller plants and house pump packages usually contain a similar device for light load service. In fact, the boiler room is the only mechanical system that does not normally have a part-load device. The answer normally given in response to this observation is that boilers can be furnished with modulating burners. But the answer is inadequate: the loads today are so small that low fire is still too much. Actual operating loads today are that small and represent just a trickle of heat. You need a small boiler to handle them. Heres a way to approach the problem. From Table 1 you can see that one-half of total hours have a maximum theoretical load of roughly one-third of the design load (it happens at about 40F OAT). Figure 1 shows that if the actual load is only one-third of this theoretical maximum, a boiler or two that can handle somewhere between 10% and 15% of the design load should carry the buildings running load at least half the time. Two 450MBH modulating boilers with an output range of 82 to 828 MBH might well eliminate short-cycling during those hours in which it is most likely to occur. The rest

of the design load can be divided up among additional, larger modules. Its a very different looking boiler room, and experience says that fuel consumption will drop significantly. Once past the morning pickup load, owners are often shocked to find that one or two of the small boilers can carry the building on a 15F or 20F day. The control sequence is simple. The energy management system dispatches the two small boilers until they can no longer carry the load; the larger boilers then operate with the smaller boilers locked out. Changeover can occur at a predetermined outdoor air temperature. At conditions close to a design load, all boiler are used.

NOW! Even if it has a modulating burner, what heat there is is instantaneously produced and delivered into the system. If only one or two zone pumps are running, the system flow is not adequate to carry the heat away from the boiler fast enough to prevent the boilers own (small) water content from rising in temperature to where the operating limit is tripped off. Once the boiler trips off, this same small flow rate is adequate to quickly cool the boilers small water content to where the operating limit is tripped on. Loads from only one or two zones occur more frequently than any other type of load in these systems, so this short cycling condition becomes chronic.

The Boiler-System Interaction

Engineers often overlook the short cycling effects of their system and piping designs. Well consider five common piping arrangements to illustrate the point.

Figure 3. The use of three-way valves at the boiler outlet has bothered boiler manufacturers for decades. Now its worse.

Figure 2. There are thousands of these systems out there, and nearly all become problem jobs when new boilers are installed. 1.. Take the case of a boiler paired with thermostatically controlled zone pumps. There are thousands of these systems in the field, many of them in churches and apartment buildings. The arrangement is shown in Figure 2. The problem is that the original boiler was a high mass design with a large water content. If the boiler is now replaced with a low mass design, a good move from an energy standpoint, short-cycling occurs. When the boiler fires, its heat output is in the piping RIGHT

2. Consider the case of a boiler paired with a three-way control valve for outdoor reset control. There are thousands of these systems out there in buildings of all types. The arrangement is shown in Figure 3. When the load is small and the reset schedule is calling for a lower water temperature, the valve port connected to the boiler outlet is only slightly open. Most of the system water flows through the bypass port. Even a small rise in the temperature of the boiler water entering at the boiler water port might cause the systems water temperature to overshoot the setpoint. This causes the valve to bypass even more water. The loss of flow through the boiler causes a rapid rise in boiler temperature, and this causes the operating limit to trip off. At that point the loss of hot water at the hot port of the mixing valve makes the valve open more to the boiler, which cools the boiler to

where it turns on again. Again heat is added at the hot port of the mixing valve, but is too much. The valve starts closing and the boiler again turns off. The on and off cycles follow each other rapidly which is the same as saying that the boiler is now short-cycling. With an older, high mass boiler with a large water content, the control problem was not this serious. There were, however, two other problems. First, these old boilers were highly inefficient. Second, most designs required relatively high flow rates to prevent damage from unwanted thermal expansion and contraction. When energy is added in a boiler, it wants to do work. If it cant do the work of heating water, if there is not enough system flow and velocity across the boilers metal surfaces to cool them at the same rate at which they are being heated, the boiler can be damaged. Most manufacturers dealt with this problem by installing a blend pump between the discharge and inlet to create the flow that the system was not providing. This solves the mechanical/metallurgical problem, the need for cooling flow across metal surfaces, but did nothing to keep the boiler from short-cycling.

recirculate water from boiler discharge to boiler supply. When the boiler fires, its heat is added RIGHT NOW! The valve actuator tries to stroke toward the other end of its range, but barely gets moving when the boiler cycles off due to the recirculation of hot water. The water content inside the three-way valve is so small that it takes just a few seconds to either cool it down (starting the boiler) or heat it up (stopping the boiler).

Figure 5. This is probably the most common piping method for modern low mass boilers.

4. Primary-secondary designs are quite common as well, and Figure 5 shows a typical arrangement. These systems can work reasonably well in large, constant flow heating systems. In recent years, however, three-way control valves have been replaced with two-way valves at the terminal units, and constant speed system pumps have been replaced with variable speed pumps. The ASHRAE energy code now mandates variable speed pumping in many applications. Note how each boiler gets its own constant speed pump. Note also how the boilers are piped in parallel in a secondary piping circuit which ties into the system header. There are two issues here. First, if the boilers are too big for the partial loads, their minimum energy input might be more than is required to raise the temperature of the system water to its setpoint. Since the sensor for the boiler sequencing controls are in the system header upstream of the main system circulating pump, it doesnt take long for the excess heat to reach the sensor and cycle off the boiler. When this happens, the sensor sees the loss of heat and immediately calls for more. Second, if system flow is reduced by the action of terminal unit control devices and the main system pumps speed controller, the boiler pumps flow rate might exceed the system flow rate. Flow will reverse in the common piping between the two primary-secondary connections. This rapidly raises the boilers entering water temperature, and even a small increase in temperature might well trip the boilers operating limit. Engineers and contractors have often dealt with this problem by moving the boiler sequencing sensor to the return side of

Figure 4. A system developed by one of the temperature control companies to control entering water temperature. 3. Another troublesome system is shown in Figure 4. This arrangement is intended to control the temperature of the water entering the boiler to prevent condensation in non-condensing boilers. The problem is that the response rate of valve and boiler are not the same, and the rangeability of the control valve is greater than that of the burner. When the boiler is off, the valve strokes to

the system, which works well enough but has the disadvantage of giving up any hope of controlling the systems supply water temperature.

fifteen gallon bucket. Theres a trickle of water coming back from the system and a tickle of water going back into the system. Theres also a burner under our fifteen gallon bucket. If thats a 2,000 MBH boiler with a minimum input of 400 MBH, its still a lot of heat for a bucket of that size. Its not hard to imagine how quickly that energy input can overshoot the setpoint and shut down the boiler, and then how quickly the bucket can cool to where the burner would again fire. Thats why most contractors try to deal with the problem by widening out the boilers operating differential. It takes a lot of differential to really fix the problem when the bucket only contains fifteen gallons.

Figure 6. This is a recent solution that was instantly endorsed by several boiler manufacturers. It will short cycle. 5. A more recent design is shown in Figure 6. A number of boiler manufacturers quickly endorsed this arrangement, seeing in it a solution to their flow and return water temperature concerns in three-way valve systems. In fact, both of these concerns are dealt with by this design. The short cycling problem, though, remains, with a dynamic that mimics the one described above for the system shown in Figure 3.

Figure 7. Most modern systems mimic the operation of a bucket with a burner. Theres no home for the heat. How big does the bucket have to be to prevent short cycling, to create perhaps five minutes of run time? Take the minimum energy input in BTUs, divide by 60 to get BTUs per minute. Now multiply by five to get BTUs in five minutes. Divide by the weight of water (8.3 lbs./gal.), and divide again by some tolerable temperature difference, say 20F. The result is the required thermal mass the boiler must work on to get the required five minutes of run time. A modulating burner makes this volume smaller, but doesnt make it go away. And remember that many modulating boilers dont start at their lowest input setting, but operate at something higher for a brief period for flame stabilization before going to low fire.

How To Fix It
A boiler of a given input must do its work on a minimum volume of water. Thats one of those obvious truths that become obvious once someone points it out. Whatever the minimum incremental energy input of a boiler, one thing must happen if the boiler is not to short-cycle: the heat must be carried away from the boiler fast enough to keep the operating limit from tripping off. One of our Patterson-Kelley representatives has a term for this: he calls it creating a home for the heat. Figure 7 shows this in an easy- to-understand format. Whats important is the water content of the boiler and the connected piping between the boiler and the boiler sequencing sensor. If the boiler contains ten gallons and the connected piping adds another five, its like having a

Figure 8. A buffer tank between system supply and return keeps variations in system flow from affecting the boiler.

Figure 10. This arrangement is even more powerful if one or more of the boilers is downsized to create a jockey boiler design.

Where a system provides the required mass and an adequate heat extracting flow rate, theres no problem. The fact is that few modern systems do this, and thats why it can truthfully be said that short-cycling is designed into most modern systems. If the system cant create a home for the heat, you, the designer, have to install it. This might add a small cost to your system, but remember that it will save 15% on the fuel bills. In fact, designing a system with a fool-proof home for the heat might be the cheapest energy strategy around: your payback will be immediate, and the investment will pay dividends for decades to come.

Figure 10 shows a multi-boiler system that easily replaces the primary-secondary systems in current use. Some boiler controllers have a return water limit feature to keep water from entering the boiler at too low a temperature. Figure 10 incorporates this feature for systems utilizing non-condensing boilers. Figure 11 shows a variable speed pumping system with a buffer tank. Note also the variable speed circulator that draws water from the outlet of the tank and injects it into the system return water header. This is optional, but a good way to keep the return water temperature above the condensing point when using non-condensing boilers.

Figure 9. This arrangement gives the boiler a mass of water to work on. The VSP can be used to control boiler EWT. Figures 8 through 11 show how each of these systems can be fixed, how short cycling can be eliminated once and for all, by designing the home for the heat into the system. In each of these systems the fix comes in the form of a small buffer tank. Figure 8 shows a system with thermostatically controlled zone pumps. Figure 9 shows the fix for three-way valve systems. The boiler or boilers are piped to the points shown.

Figure 11. The VSP from tank outlet to system return may not be necessary.

How To Size A Buffer Tank

There is a simple formula for sizing buffer tanks: Volume = [ t x (Qmax Qmin) ] / [ 500 x T], where, t = minimum desired run time in minutes Qmax = minimum boiler energy input (lowest firing rate) Qmin = minimum system energy extraction rate

500 = 8.3 lbs./gal. x 60 min./hr. T = the allowable tank temperature drop. Qmin for most systems under partial load is so close to zero, it might as well be zero. Zero is the most often used value for purposes of this calculation. In effect what youre doing is installing some working volume and using your boilers as sidearm heaters for the tank. Think of the boilers as gas fired heat exchangers. The tank becomes the control point for boiler sequencing, and the boiler sequencing sensor is located there, rather than in the header. Tank size determines cycle times and, therefore, cycle efficiency. The system designer is, therefore, in total control of what happens when the system is commissioned. Cycle times and efficiencies are as the designer makes them.

requirements, and can help you develop control strategies for your desired control platform.

Final Considerations
1. The best strategy combines jockey boilers with buffer tanks: the system energy efficiency gain is truly impressive. In fact, it may be the best system out there. 2. Some manufacturers say that burner turndown makes this all unnecessary. That position simply cannot be technically defended. While high turndown rates can make buffer tanks smaller, few systems today create a home for the heat that is adequate to prevent short cycling. Other manufacturers market what they call zero flow boilers which cannot be damaged by a loss of flow. That does not, however, mean that they wont short-cycle. Short-cycling wastes energy. (There is an entire tradecraft related to variable flow, variable temperature, low temperature and condensing boiler systems that is the subject of another paper.) 3. Many engineers, though persuaded by the arguments presented here, are reluctant to specify either strategy because they expect contractors to offer voluntary alternates based on a more conventional approach (all boilers the same size, no home for the heat). That is understandable given the structure of our industry today. But it must be remembered that the two strategies outlined in this paper, alone or (better) together, improve both equipment longevity and energy use. The added cost is minimal. The real question is this: how can you afford not to do it? 4. These systems are bullet-proof: they work no matter whose equipment gets purchased by the contractor. They have the added advantage of leaving the existing system balance undisturbed in retrofit applications. 5. Always contact your Patterson-Kelley representative for support in applying either of these concepts. They are experienced in applying them to a wide variety of building and system types. They can also support you with specifications that are sensitive to local code