WHEN TO REPLACE YOUR AGING CHILLER PLANT

by Joseph J. Watson, P.E.
E3 Designs

Replacing a chiller plant is often one of the largest end of useful life expenditures facing building owners. There are many factors to consider when determining when to replace an aging chiller plant. Property managers should consider using a life-cycle analysis as a decision-making tool. Life cycle analysis considers the four components of owning and operating costs including annual utility costs, maintenance costs, replacement costs, and remaining useful life. All costs are adjusted based on the time value of money to allow equivalent comparisons. With this method, all potential system types can be evaluated and compared to determine the most cost effective alternative.

Despite the traditional emphasis on minimizing the installed cost of a plant, the most significant expenditure over the life of the plant is typically the annual operating costs. For example, one plant that I analyzed had an installed cost of $500,000, annual operating cost of $100,000 per year, and a maintenance cost of $20,000 per year. Assuming a useful life of 20 years, the present value of the installed cost was $500,000, the operating cost was $1,300,000, and the maintenance cost was $250,000. As you can see, the installed cost accounted for only 25% of the total expenditures over the life of the system. This relationship should be kept in mind when evaluating alternative systems.

As that previous example points out, operating costs are the most significant expenditure over time. Higher first costs that provide improved efficiency can often be justified over the life of the system. But there are many pitfalls that can undermine the decision making process. In my experience, starting out with more efficient equipment is always a good idea. But installing efficient equipment does not automatically assure that your plant will operate efficiently. I have seen many occasions where a plant with a low installed kw per ton ends up operating at a much lower efficiency than older or less efficient plants. The two main factors that I have found most significantly impact chiller plant operating costs are building load profile and chiller sequencing strategies. In both cases, plant efficiency is determined largely by the part load efficiency of the chillers.

The first step in evaluating the chiller plant is to develop the plant load profile. The easiest way to accomplish this is by using the trending function of your plant energy management system. Typically, the data I like to collect from a chiller plant includes entering and leaving water temps on both the chiller and the condenser, chiller amps, and status information on chilled water pumps, condenser water pumps, and cooling tower fans. This information should be collected hourly, and can be used to calculate a part load performance curve for the chiller and the plant. As you continue to gather and evaluate this data over the operating season, a very accurate plant load profile can be generated.

As you can see from the above graphs, part load operation can significantly affect the operating efficiency of a chiller. In this example, a machine with a design efficiency of .5 kw per ton can operate at 1.2 kw per ton or higher during periods of part load operation. The goal in operating the plant is to operate the chillers as close to design capacity as can be achieved. In the case of a plant with a single chiller, or two chillers with full redundancy, the chiller part load profile is the same as the plant part load profile. If you find your load profile is the typical bell curve centered around 50% load, you may want to consider installing two smaller chillers to improve part load performance. Another alternative may be to consider a chiller equipped with a variable speed drive, which is much more efficient under part load operation.

Larger plants with multiple chillers offer a better opportunity to operate the chillers closer to design capacity. However, it is not an easy task to properly sequence multiple chillers. In my experience, most of the owners of large chiller plants prefer to sequence their chillers manually. Much of the reasoning for this decision is the comfort level from knowing one of the operators is verifying the operation of the chillers with "hands-on" control. Most competent operators are very good at determining when chillers should be brought on line, based on load conditions. However, much of the low end part load operation in a large plant occurs after a high load day, when the plant load drifts down from the peak. Often, multiple chillers may be left idling for several hours before being taken off line. The ability to aggressively take chillers off line will have a larger impact on plant efficiency than the decisions to bring the chillers on-line.

Both of these factors can have a larger impact on operating efficiency than the initial design efficiency of a machine. When considering a chiller plant upgrade, be sure that the plant has a DDC control system, and a viable sequence of operation before spending additional dollars on a lower kw per ton machine. In all cases, the control system should have the ability to monitor, control, and verify plant operation. The feedback you receive from the plant energy management system should be used to continuously fine tune the operation of the plant. If you need to rely on your operators for average plant load or chiller sequencing information, you may be missing opportunities to significantly improve the system operation.

Another factor to consider is the maintenance and repair cost for your chillers. Scheduled maintenance costs for chillers of similar type and size are usually consistent. However, new machines will have lower unscheduled repair costs than older machines. The historical repair data for an existing machine can be summarized and included in the analysis. It is important to review manufacturers recommendations for the required frequency for component overhauls, especially when comparing between different machine types. (turbine driven vs centrifugal, for example). An overhaul can be a significant expenditure, and may be required more frequently depending on hours of use. These expenditures should be identified, and included as part of the life-cycle analysis. Getting a full maintenance service contract price including required overhauls during the budgeting process is a good way to get a realistic estimate of these costs on an annualized basis.

This brings us to the most difficult task in a life cycle analysis, determining system useful life. In my role as chairman of the ASHRAE committee on Building Owning and Operating Costs, many of our discussions centered on how to determine the useful life of a component or system. There have been numerous studies and surveys performed in an effort to quantify this information, either through detailed life studies of individual components, or by anecdotal evidence gathered through operator surveys. Both methods have severe drawbacks in attempting to use the conclusions to make assumptions about the useful life of any given system.

Many factors can come into to play when determining when an existing system has reached the end of it's useful life, some relating to the component itself, and others relating to it's specific application. For example, there are currently many chillers operating on cfc-based refrigerants. The manufacture of these refrigerants were legislated out of existence in the mid 1990's, meaning that replacement refrigerant would only be available from existing stock. Many plant operators chose to retrofit or replace their chillers prior to the phase-out date to assure uninterrupted operations. Others made plans to replace or retrofit once the refrigerant became scarce. I can assure you that in my discussions with plant operators in the mid 1990's, no one expected their chillers to be operating on cfc-based refrigerants in 2002. However, with an increased emphasis on leak detection and control, adequate supplies of refrigerant are still available today, and probably will be for the foreseeable future. This is just one example of how external factors can affect estimates of useful life. Were the chillers that were changed out in the mid 1990's replaced prematurely? Not necessarily. In hospitals and other critical applications, the reliability of the equipment is paramount, and decisions are routinely made to minimize the risk of potential failure. In other cases, especially those with high annual operating hours, improvements in operating efficiency justified the expenditures. Still others may have been confronted with repair parts that were no longer available due to obsolescence, and the refrigerant issue was just another factor in the decision making process. So you can see that the useful life of a piece of equipment is determined not only by the component, but by it's application as well.

So how does an operator address useful life in the analysis? Unfortunately, there is no universal truth that is applicable to all situations. However, there are factors that can minimize it's impact. When performing a comparison between alternative systems, estimates of useful life must be made for each system type. Fortunately, due the nature of the life-cycle analysis, the difference between a system with a twenty year useful life, and one with a thirty year useful life may not be significant in terms of present value. Expenditures made after 20 years have a relatively small impact on the present value of a project.

However, this is not necessarily true in determining whether to replace an existing plant. In the absence of factors which would necessitate a system replacement, an existing chiller plant could continue to be operated indefinitely. Indeed, I have surveyed a number of chiller plants with equipment installed in the early 1960's that are continuing to function adequately for their specific application. The life cycle analysis is the best tool to determine if the system should be replaced from an financial standpoint. Beyond that, it is the responsibility of the facility manager to decide what level of reliability and risk they are willing to accept in the operation of their plants.

BIO

Joseph J. Watson, P.E., a senior project manager for E3 Designs in Kitty Hawk, North Carolina, has over 25 years of experience in all facets of the HVAC industry. He is currently responsible for overseeing numerous chiller plant optimization and monitoring projects throughout the United States, and was the recent chairman of the ASHRAE technical committee on Building Owning and Operating Costs. He would welcome any feedback on this article and can be reached at jwatson@e3designs.com.