Don't be Stymied by Chiller Specs
When specifying chiller efficiency, some HVAC designers actually work against themselves by producing specifications that prevent them from getting what they need from the bidding process: chillers with equal annual energy consumption at the best capital cost.
The problem usually occurs when two measures of chiller efficiency—the design-efficiency rating and the non-standard part-load value (NPLV) rating—are both used in the specification. So, let's dispel the confusion about these ratings once and for all.
First, let's clear up some misconceptions about the NPLV rating. NPLV can be measured in kilowatt per ton refrigeration (kW/TR), coefficient of performance (COP) or energy-efficiency rating (EER). Here, we will use kW/TR.
Test yourself by taking this simple three-question quiz:
1. True or False : A chiller's NPLV rating measures only its off-design efficiency.
Answer: False . The NPLV rating includes both off-design efficiency and design efficiency.
2. True or False : The integrated part-load value (IPLV) rating is a subset of the NPLV rating.
Answer: True . The IPLV rating is a subset of the NPLV rating, targeted to a very specific situation: when the project's design conditions are equal to the ARI standard conditions. The NPLV rating allows for efficiency measurements at a wide range of conditions and is normally used in specifications.
3. True or False : A chiller with good efficiency at design conditions, which is the simultaneous occurrence of both design load and design cooling-tower-water temperature (or design air temperature, if air-cooled), will automatically have a good NPLV rating.
Answer. False .Two chillers can have the same design efficiency, and yet can have NPLV ratings that vary widely, depending on capital cost. For example, a chiller with a design efficiency of 0.58 kW/TR can have an NPLV rating anywhere from 0.55 to 0.35 kW/TR. That's because chillers can have different off-design efficiencies.
Of course, off-design performance is of paramount importance, because chillers—including multiple-chiller plants—operate most of the time at partial load or at off-design cooling-tower-water temperatures (or off-design air temperatures, if air-cooled). A chiller with both good design efficiency and a good NPLV rating is possible, but this exacts a cost premium. Fortunately, it's easy to avoid adding cost by changing the way chiller efficiency is specified.
The consequences of misconceptions
The reason some chiller-efficiency specifications result in unintended consequences is simple confusion over how design efficiency relates to NPLV rating. Naturally, HVAC system designers want to produce as complete a specification as possible. But they stymie themselves when they reason that “more is better” and specify both the NPLV and the design efficiency.
To understand why, let's look at an example of what happens when both efficiency values are specified. Let's begin by comparing two 1,000-TR chillers, as shown in Table 1 on p. 52.
The specified chiller has an NPLV rating of 0.478 kW/TR and a design efficiency of 0.576 kW/TR. “Option A” chiller also has an NPLV rating of 0.478 kW/TR but a design efficiency of 0.590 kW/TR, which is higher than the specified chiller. Because both chillers have equal NPLV ratings, they have equal annual energy consumption. Remember, the NPLV rating includes both off-design and design efficiency. They also have equal capital costs of $250,000.
However, if the specification requires that a chiller meets both the NPLV rating and the design-efficiency rating, Option A can't meet both ratings and, therefore, can't be bid. It may be a function of compressor size, impeller diameter or rotational tip-speed. Regardless, the manufacturer of Option A will usually need to modify it by adding more heat-exchanger surface to meet the design-efficiency rating. The performance of this new chiller is shown in Table 2 (p. 54) as Option B.
Because of the additional heat-exchanger surface, Option B has an improved NPLV rating of 0.459 kW/TR, resulting in annual energy that is 4% better than the specified chiller. But in meeting the design-efficiency specification, it has also become more expensive. In our example, it costs $31,000 more.
By specifying design efficiency, the designer has complicated matters. Instead of equalizing energy consumption as a basis for comparing costs, now both annual energy consumption and pricing are unequal. But what the specification meant to accomplish was equal annual energy consumption at best capital cost.
Suppose the specified chiller and Option B are bid by two different manufacturers. What is the impact of the specification on bid day? The unfortunate result is that the chiller price settles out at slightly less than the higher of the two prices—about $280,000. This is one of the unintended consequences of competitive bidding. Specifically, the owner in our example is likely to end up purchasing the specified chiller, but pays about $30,000 more and gets no additional energy savings, simply because Option A did not satisfy the design-efficiency specification and could not be bid.
This example demonstrates why specifying the design efficiency is not only unnecessary, it may also be counterproductive.
The myths of design efficiency
If including both design-efficiency and NPLV ratings is unnecessary, why do some specifications still contain both? There are three reasons. Some HVAC system designers believe that the design-efficiency rating impacts electric-demand charges. Others believe that it dictates power-wiring size. And lastly, some energy codes and utility rebates require the design-efficiency rating be included in the specification. Let's examine each of these reasons to see if using the design-efficiency rating is really warranted.
Does a chiller's design-efficiency rating impact electric-demand charges?
Consider Option A, which has a design efficiency of 0.590 kW/TR. At first glance, that chiller would appear to cause higher electric-demand charges than the specified chiller, which has a design efficiency of 0.576 kW/TR. But is that really the case?
It's true that during the few peak-cooling months, Option A will have a slightly higher demand charge. However, the fact that both chillers have the same NPLV rating means that Option A must have a better off-design efficiency. So during the many months of off-design operation, Option A will have a lower demand charge, unless demand is ratcheted year-round.
However, electric-demand charges are becoming a non-issue because of increasing electrical power deregulation. These days, the electricity user often has a choice of providers. As a consequence, demand charges cannot be easily passed on to a non-captive marketplace. Another side effect of spreading deregulation is that ratchet clauses will likely die out. Although captive customers have grown accustomed to demand charges, ratchets have always been disliked and considered a double-dipping on demand. Now, many electricity users have the choice of using a different supplier rather than endure the ratchets.
Finally, chiller maximum kW has less impact on demand charges due to building-power patterns. The building's kW and the chiller's kW typically peak at different times of the day. This phenomenon can be referred to as the “flywheel effect” of the building's demand vs. the chiller's demand, which is illustrated in Figure 1.
Surprisingly, most chillers reach peak energy consumption between 3 p.m. and 7 p.m. Why so late? At about noon, the sun's rays strike the ground at the most direct angle. Through convection, the ground then heats the ambient air to its highest dry-bulb temperature about two hours later. Once the air temperature is at its maximum, around 2 p.m., the heat is slowly conducted through the building skin, a process that peaks building load around 4 p.m. In parallel, the wet-bulb temperature of the ambient air also reaches its maximum later in the day. The higher wet-bulb temperature raises the cooling-tower-water temperature. Both temperatures raise the head pressure against which chillers must work, hurting energy efficiency. When these factors are combined, the chiller sees its peak load, head and kW in late afternoon.
Now, looking at building kW, most air-conditioned buildings reach their peak electric demand between 10 a.m. and 3 p.m. That's when occupancy is usually at its highest, which maximizes the “people” load. Higher occupancy also translates into more heat generated by lights, elevators, cafeterias, office equipment, etc. When these factors are combined, the building encounters its peak kW draw in late morning to mid-afternoon, hours before the chiller's peak.
In summary, the electric demands of buildings and chillers rarely peak at the same time. In fact, when the building kW hits its maximum, the chiller kW has typically reached only 80% to 90% of its eventual maximum. These two events often occur on different days as well.
For all these reasons, small differences in the design efficiencies between chillers will have little impact on the demand charges incurred by the building.
Does design efficiency dictate the size of the power wiring?
The correlation between design efficiency and wiring size is usually a non-issue, because a given wire size can handle a range of amps. Thus, a chiller with a higher design kW/TR will not necessarily require larger wire. In fact, about 90% of the time it will not, because the wiring size can already handle slightly higher amperage.
In any case, a better way to ensure proper wiring size is to specify maximum full-load amps and minimum power factor at the chiller starter. Full-load amps should be treated like the chiller's physical dimensions. A designer can obtain data on amperage requirements from a number of manufacturers, and specify the largest appropriate wire size. On the rare occasion that a chiller selection can reduce the wire size, the owner can get a deduct.
Do energy codes or utility rebates require the design-efficiency rating?
When energy codes or utility rebates require inclusion of the design-efficiency rating in the specification, designers are advised to specify the maximum kW or kW/TR required by the code or rebate. As the earlier example demonstrated, specifying a lower value could result in higher capital costs with no reduction in annual energy costs.
At some point, utilities and code-writing agencies are likely to recognize that annual efficiency, as measured by the NPLV rating, is more important than the design-efficiency rating. In fact, some code writers have already made the change.
Chiller-efficiency specifications that specify both the NPLV rating and the design-efficiency rating are almost always unnecessary, if the goal of the specification is the lowest capital cost for equal annual energy. That's because the two ratings can create inequalities in annual energy-consumption comparisons, which also result in higher capital costs passed on to the owner. The design-efficiency rating also has little practical impact on electrical-demand charges and wiring size.
Instead of using both ratings, the best chiller-efficiency specification uses the NPLV rating by itself. The design kW/TR can be dropped entirely. For power-wire sizing, specifying the maximum full-load amps and the minimum power factor eliminates all ambiguity about actual size requirements. If energy codes or utility rebates require that the specification includes the design-efficiency rating, the maximum allowable kW should be specified.
Specifying the NPLV rating without the design-efficiency rating is the better way to write a chiller-efficiency specification.
Table 1 - 1,000-TR Chiller Compari sons
Option A Chiller
Table 2 - Impact of Specifying Both NPLV and Design Efficiency
Option A Chiller
Option B Chiller