AutomatedBuildings.com Article - Improve Chiller Plant Efficiency

In 1992, I read an article by researchers at the Energy Engineering Institute, a part of San Diego State University. The article reported results of a survey of 31 chiller plant sites across the US which found chillers that should be operating in a range of from 0.65 to 1.00 kW/ton according to their manufacturers specifications were often actually operating at 1.5 to 3.0 kW/ton. This correlated well with my experience in benchmarking existing commercial buildings for possible energy improvements. I often found it necessary to incorporate very low efficiencies in the simulation database for chiller plants in order to make the simulated building energy use profiles match the actual energy use data. The article also stated that the survey found most chiller plant operators had no idea how well their chiller plants were performing, and the plants did not have the instrumentation required to test or monitor efficiency. This was, and continues to be, in accord with my experiences. I don't think this situation has changed since this survey was conducted.

If these poor operating efficiencies are indicative of chiller plants in general, this excessive energy use constitutes one of the largest opportunities for energy reduction that exists in North America today since it is estimated that chiller plants alone consume about 20% of the total electrical power generated. However, to capture these substantial savings, new thinking about how chiller plants should be operated is required.

Current operating standards for chiller plants are simple; most recommend sequencing chillers and towers as necessary to ensure the cooling load is met with the least amount of equipment in operation and that on-line equipment operate as near full capacity as possible. Since most operators don't know how well their plants are performing, this mode of operation has become institutionalized over the last few decades with little thought toward plant efficiency. But improving performance demands a much better match of equipment and operations to meet the part load conditions that constitute most of a plant's operating hours. Lower outdoor wet bulb temperatures lead to lower entering condenser water temperatures which reduce the head on the chiller compressors and thus the power consumed. Reduced load on each chiller also decreases the temperature difference across the evaporator and condenser heat exchangers and thus further reduces the compressor head and power consumed. Therefore, overall chiller plant efficiency should increase dramatically at partial load conditions.

But plant designers and operators know that the situation is more complicated that this. Chiller compressor efficiency decreases as chiller capacity falls. Furthermore the power requirements of conventionally operated condenser pumps and tower fans can also reduce plant efficiency when chillers and towers are operated at low capacities. This is why the industry has adopted a sequencing strategy in which the on-line equipment is selected for full capacity operation whenever possible.

However, ignored in that decision is the fact that chiller plant efficiency also can suffer when chillers are operated up to full capacity at part load conditions. As entering condenser water temperature falls, chiller compressors "ride the curve" and the compressor capacity increases by 15% or more. As the chiller compressor capacity increases, the load transferred increases and the condenser and evaporator heat exchangers become undersized and less efficient in transferring energy between refrigerant and water. Compounding the problem are the reduced efficiencies of the chiller compressors and cooling towers at these conditions. The cooling towers, whose efficiencies are already reduced at low wet bulbs are further impacted by the higher heat rejection loads from the chillers. These factors all act to reduce the efficiency of a chiller plant at part load conditions under conventional operating procedures. The result is that overall plant operation is far less efficient than it should be at typical part load conditions.

Poor performance from plants that focus on minimizing the number chillers and towers operating at part loads is a good part of what leads to the poor efficiency noted in the chiller plant survey and is frequently confirmed by chiller plant efficiency monitoring and simulation. But it is also clear that simply allowing all the equipment to remain in operation as the load falls is not efficient either. The overall thermodynamics of a chiller plant requires operations that include some means of sequencing chillers in response to load changes. So, the question remains - "How should equipment be operated and sequenced in a chiller plant?"

My preferred solution is to design and implement a simple and economical chiller plant control network for the chillers, pumps and tower fans, that automatically operates and sequences all equipment to meet the load and optimize efficiency. For plants under a few thousand tons this control can be easily developed by analyzing equipment operating parameters at various load conditions and developing operating and sequencing strategies that provide the most efficient operation at those load points with smooth transitions to other load conditions. For more complex plants some numerical analysis may be required, and other parameters may have to be considered (such as the relative cost of energy sources for hybrid plants). I have found that implementing automatic operations to achieve high plant efficiency does not require complex algorithms or chiller sequencing. Instead, considerable chiller plant efficiency improvement can be obtained with simple and easily understood efficiency based sequencing and control. Of course the exact control required always depends on the particular configuration of equipment in the plant.

If plant management prefers to retain manual control of equipment operation, this can be accommodated, but as much of the system should be automated as possible (for example, tower fans and condenser pumps should operate automatically in response to the particular sequence of chillers and towers chosen by the operations staff). If manual sequencing is employed, it is essential that energy monitoring equipment be installed so that the operations staff can at all times see the real-time plant operating efficiency as a single kW/ton or COP value, and that this value is continually updated every few seconds. With this information, the operations staff can "learn" how to operate the plant most efficiently under various load conditions. Manual plant operation with real-time efficiency data does require attentive and committed operations staff and will cost more than automatic control, but it does offer some benefits, particularly for large plants. Manual operation ensures that the equipment will be closely watched and problems or failures may be less likely to occur in such plants.

Whether automatic or manual chiller plant efficiency control is employed, the change in operation from "conventional" equipment sequencing to an efficiency focused plant operations strategy will easily reduce annual chiller plant energy costs by $20 to $100 or more per installed ton, depending on climate, application and utility rates. Such an improvement will almost always be cost effective and at the same time contribute considerably to a greener environmental future. So, how efficient is your chiller plant? Now is a good time to find out and do something about it!

Additional information on technologies discussed in this article is available at www.hartmanco.com . Comments and questions about the article may be addressed to Mr. Hartman at tomh@hartmanco.com .

Those who wish to hear and participate in further discussion on the topic - Improve Chiller Plant Efficiency! presented by Tom Hartman in his commentary this month have a chance to do so at the ASHRAE Winter Meeting, January 27-31, 2001, in Atlanta, Georgia. Mr. Hartman will be making two presentations on the issues involved in improving chiller plant operating efficiency:

1. Symposium 2, at 8:00 AM on Sunday, January 28th is titled Central Energy Plant kW per ton and Demand Controls. Papers will be presented on the issue of monitoring chiller plant efficiency in order to guide operators to improve the efficiency of their plants. In this Symposium, Mr. Hartman will present a paper titled "Instrumentation Issues for Monitoring Chiller Plant Efficiency." The presentation will include a discussion of the value of real time efficiency monitoring for chiller plants and recommend methodology for employing real time plant efficiency measurement to improve plant efficiency.

2. Seminar 29 at 10:15 AM on Tuesday, January 30th. This Seminar is titled Adding New Life to Old Systems: Control Retrofit Case Studies and includes discussions concerning the use of controls to update existing systems. Mr. Hartman's presentation in this Seminar is titled "Integrated Controls Improve Chiller Plant Performance." This presentation will discuss and compare the relative value of retrofit options to an actual chiller plant. The retrofit options considered include 1) high-efficiency chiller replacement only, 2) a network based plant control upgrade only, and 3) a combined chiller and controls upgrade. This presentation is instructive as to the type of upgrade that is most cost effective for existing chiller plants.

Plan to attend these sessions or e-mail Mr. Hartman with your questions or comments about his commentary.

Events	Want Ads
Our Sponsors	Resources