It is important to understand that no matter how good the system design is, adequate controls are necessary for all the components to operate properly as a system. It is equally important to understand that you cannot “control your way out of a bad system design.
The chiller plant consists of chillers, pumps, pipes, coils, cooling towers, temperature sensors, control valves, and many other devices. It is similar to an orchestra with many instruments. The existence of these pieces does not guarantee that the system will work properly. There needs to be an orchestra conductor. In the case of a chilled-water system, that conductor is a chiller-plant control system. How well the plant works depends on how well the control system gets all the pieces to work together.
- Chilled-water temperature control
- Monitor and protect
- Adapt to unusual conditions
The largest change to chillers in the last decade has undoubtedly been in the area of controls. In the past, chillers were pneumatically controlled, and they were protected by turning them off if the flow rates or temperatures changed too quickly. Today’s microprocessor-based controls provide accurate chilled-water temperature control, as well as monitoring, protection, and adaptive limit functions.
These controls monitor chiller operation and prevent the chiller from operating outside its acceptable limits. They can also adapt to unusual operating conditions, keeping the chiller operating by modulating its components and sending a warning message, rather than doing nothing more than shutting it down when a safety setting is violated. Improved control accuracy allows chillers to be applied in systems and applications that were previously avoided. When problems occur, diagnostic messages aid troubleshooting. Modern chiller controls also interface with a chiller-plant control system for integrated system operation.
What Is Important?
There are primarily five issues to address in a chiller-plant control system.
- When should a chiller be turned on or off?
- After we know that a chiller must be turned on or off, which one should it be?
- If we attempt to turn on a chiller, pump, or cooling tower, and there is a malfunction, what do we do next?
- How can we minimize the energy cost of operating the system?
- How can the chiller-plant control system effectively communicate with the operator?
- Turning on an additional chiller
- Turning off a chiller
- Which chiller to turn on or off?
Chiller sequencing refers to making decisions about when to turn chillers on and off, and in what order. Typically, turning chillers on and off is performed with the goal of matching the capacity of the chiller plant to the system cooling load. In order to do this successfully, the design of the chilled-water system must provide the control system with variables that are good indicators of system load.
The hydraulic design and size of the chilled-water system will determine the possible method(s) for effectively monitoring system load. Typical methods for load monitoring include:
- In series- or parallel-piped systems, the supply- and return-water temperatures, and sometimes chiller current draw, are monitored.
- In a primary-secondary system, the system supply and chiller return-water temperatures and/or the direction and quantity of flow in the bypass pipe are typically measured.
- In a variable-primary-flow system, the system supply-water temperature and the system flow rate may be monitored.
- Direct measurement of the system load (in tons, kW, or amperes) has also been used in some systems.
Other methods are also possible. It is imperative that the chilled-water system be designed with the control variables in mind; otherwise, the result may be a system that is impossible to efficiently control.
Today’s chiller controls can very accurately control the chiller’s leaving-water temperatures over a wide range of loads. This is especially true of centrifugal and helical-rotary chillers. This fact allows constant-flow chilled-water systems, similar to the system shown in Figure above, to use the system supply- and return-water temperatures to determine system load.
By sensing a rise in the temperature of the water leaving the chiller plant, the control system can determine when the operating chillers can no longer maintain the desired temperature. Often, the supply-water temperature is allowed to drift a predetermined amount before an additional chiller is turned on, to ensure that there is enough load to keep an additional chiller operating.
Deciding when it is appropriate to turn a chiller off is more complex. The control system may monitor the system ∆T, that is, return-water temperature minus supply-water temperature. This information, along with the capacities of the operating chillers, allows the control system to determine when a chiller can be turned off. To help stabilize system operation, the control system should use logic to prevent load transients from causing unwarranted chiller cycling.
In constant-flow systems that are suffering from “low ∆T syndrome” (airside systems that return water to the plant at lower temperatures than desired), some of the load terminals may starve for flow before the capacity of the operating chiller is exceeded. To preserve system efficiency, this situation is best dealt with by solving the airside problem. Typical causes of low ∆T syndrome include: a poorly-balanced flow system, dirty filters or coils, poorly performing air-handler controls, incorrect coil control valves, or undersized air handlers.
In a primary-secondary system, the direction and quantity of flow in the bypass pipe is an excellent indicator of when to turn a chiller on or off . As discussed in Period Two, the water flow in the bypass pipe can be measured directly using a flow meter, or indirectly by measuring system water temperatures and applying flow-mixing equations. The rules applied to the bypass flow to determine when to turn a chiller on and off are:
- When there is a deficit flow, a chiller may be added.
- When there is excess flow greater than that of the next chiller to be turned off, plus a 10 to 15 percent safety factor, that chiller may be turned off.
- If neither of the above conditions exists, do nothing.
As an alternative to measuring flow in a primary-secondary system with four or less chillers, system supply and chiller-plant return-water temperatures may be used to decide when to turn a chiller on or off. This is similar to the logic applied to constant-flow systems. It is simple and has a low installed cost, but it is less accurate than flow determination, especially as the number of chillers increases.
“Low ∆T syndrome” can also affect the operation of primary-secondary systems. Unlike constant-flow systems, the primary-secondary system will maintain the required system flow and supply-water temperature, and therefore maintain occupant comfort. However, it accomplishes this by turning on additional chillers before all operating chillers are fully loaded. This may reduce overall system efficiency.
Although some have proposed solutions such as putting a valve in the bypass line, lowering the supply-water temperature, or controlling the system differently, these are only band-aids that mask the actual problem and often cause other operational difficulties. Fixing the root cause of low ∆T syndrome in the distribution system is the best course of action for proper and efficient system operation.
Another method of monitoring system cooling load is to measure the system water flow rate and temperatures directly, and then calculate the load. Although it would appear that direct measurement of the actual system load would be an excellent way to determine when to turn chillers on and off, this method has several drawbacks. It requires the use of flow meters with high accuracy and high turndown capacities. Although flow meters have become more accurate and less expensive, they require special installation conditions for reliable accuracy—conditions seldom achievable in real installations. Also, the equipment typically requires regular calibration. For these reasons, direct measurement of load has not been used as much as the simple and reliable methods discussed previously.
An alternate way to monitor chiller load is by measuring the current draw of the chiller motor. By itself, this does not provide an adequate control indicator, but when used in conjunction with other information, such as system supply-water temperature, it can be effective. System supply-water temperature is used to determine when to turn an additional chiller on, and operating chiller compressor-motor current draw is used to determine when a chiller can be turned off.
The most effective load indicator for any chilled-water system is dependent on the design of that system. Creative designers have used the control strategies as described here and in various combinations to effectively control a wide variety of chiller plants. It is highly recommended that one of the first tasks undertaken in the design process is to create a simplified flow diagram and a load model of the system that allows for the evaluation of various control strategies and sensor placements. This will help to ensure that effective chiller-plant control can be implemented.
When the system has determined that a chiller needs to be turned on or off, the next issue is to determine the sequence in which to turn chillers on and off. It is assumed that the first chiller in the sequence will always be turned on whenever cooling is required.
When the system consists of identical chillers, the choice of which chiller is turned on or off next has little impact on system efficiency. Some design engineers and operators prefer to equalize the run time and the number of starts for all chillers in the system. This is typically done by rotating the sequence of chillers on a periodic basis, often every few days or weeks. This method generally keeps the run time equalized reasonably well, and the operator knows exactly when to expect the rotation to occur. An alternative approach is to total the actual run hours on each chiller, in an attempt to rotate the chillers when a significant imbalance in the run time or the number of starts occurs. Rotation that is based on actual run time has the disadvantage of the operator not knowing when rotation will occur. In some installations, operating personnel prefer to manually initiate rotation.
On the other hand, some design engineers and operators believe that equalizing run times will result in all of the chillers needing to be overhauled or replaced at the same time. They tend to operate the most-efficient chiller first, followed by the next-most-efficient, and so on. With this approach, all chillers are turned on at least once a month to ensure that they will be able to start when required.
When the system consists of chillers with different capacities, efficiencies, or fuel types, the question of which chiller to turn on or off next becomes more complex. Although each system requires a complete analysis, there are some general principles that apply to most systems.
In systems with chillers of different capacities, such as the “swing” chiller concept introduced in Period Three, the goal is to operate the least number of chillers and the smallest chiller possible. This typically minimizes overall system energy consumption by closely matching the capacity of the plant to the system load, thus reducing the energy used by ancillary equipment.
In systems with chillers of different efficiencies, it makes sense to operate the most efficient chillers first and the least efficient chillers last. If different fuel types are involved, the control system may receive data on the costs of natural gas and electricity and calculate the real-time cost of operating the electric- versus gas-driven chillers.
The system might also benefit from having a heat-recovery chiller fully loaded. As discussed in Period Three, to maximize the amount of heat recovered, it is often desirable to preferentially load that chiller, sequencing it as a base chiller—“first on” and “last off.” Other chillers can then be turned on when the heat-recovery chiller cannot handle the cooling load alone.
A variation on this idea is an absorption chiller fueled by waste heat. It is preferentially loaded to handle as much of the cooling load as possible before turning on other chillers. The absorption chiller would be sequenced as a base chiller to make use of the free energy operating this chiller.
The variable-primary-flow system, introduced in Period Three, is designed to operate with variable flow through the chiller evaporators. Sequencing chillers in this type of system cannot be based solely on temperature, because in a properly-operating system the supply- and return-water temperatures will be nearly constant. Determining when to turn chillers on or off is not a simple task. For control stability and chiller reliability, the flow rates through the chillers, and the rate of flow change, must be kept within allowable ranges.
Therefore, control of a variable-primary-flow system must:
- Include a method to determine system load. Many systems measure flow rates and temperatures.
- Ensure that flow rates through the chillers are within the allowable minimum and maximum limits. Modulation of a control valve in the bypass pipe is commonly used to ensure minimum flow rates through the chillers.
- Control the rate at which the system flow rate changes, to ensure that it does not change more rapidly than the chillers can adapt. This is especially critical when turning on additional chillers.
Adequate time must be spent designing the control sequence and commissioning the system after installation, to ensure proper operation of a variable-primary-flow system.
System Failure Recovery
- Maintain flow of chilled water
- Keep it simple
- Lock out failed equipment
- Turn on the next chiller in the sequence
- Notify the operator
- Allow the operator to intervene
In addition to normal chiller sequencing, the chiller-plant control system must react when a chiller or another piece of associated equipment fails. Failure recovery, or ensuring the reliable supply of chilled water, is a very important part of the chiller-plant control system, and is an area where many systems have fallen short. This is especially true in field-programmed systems because of the difficulty of thorough debugging.
During periods of equipment malfunction, it is important to focus on the primary goal of the system, which is to provide the required flow of chilled water to the system at the proper temperature. It seems reasonable that the simplest and most reliable failure-recovery sequence is to simply turn on the next chiller in the sequence, and not try to turn several chillers on and off in an attempt to re-optimize the system.
During an equipment failure, it is especially important to notify the operator of the status, as well as to help the operator understand where the problem is and what might be the cause. The control system must also allow the operator to easily analyze the situation and to intervene if the failure condition will exist for an extended period of time. A system that provides this information will ensure that the system itself will be maintained and operated in proper condition.
In addition to failure recovery, it is wise for the system design engineer to work with the building owner to develop a contingency plan for chilled water in the case of an emergency shutdown or an extended breakdown. Many organizations have contingency plans for critical areas of their business. Some deal with natural disasters and others with the loss of power in critical areas. However, few have taken the time to think about what a loss of cooling would mean to their facility. This is often especially critical for process-cooling applications.
Cooling contingency planning is intended to minimize the losses a facility may incur as a result of a total or partial loss of cooling capacity. It allows a building operator to act more quickly by having a plan in place and by proactively preparing the facility. Such a plan often includes working with suppliers to temporarily lease cooling equipment. During initial construction,it is easy and cost-effective to provide piping stubs, which are built into the chilled-water system for quick connection, and easily accessible electrical connections. When equipment leasing is combined with these simple additions to the system, a contingency plan can be put into action quickly and the system can produce chilled water again in a short period of time.
It is important to first identify the minimum, or critical, cooling capacity required. With multiple chillers in a facility, it may be acceptable to have less than full capacity in an emergency situation. For example, the chiller plant may consist of 1,800 tons [6,330 kW], but the minimum capacity required in an emergency situation may only be 1,200 tons [4,220 kW]. Therefore, it is also important to identify a contingency plan if Chiller 1 fails, if Chiller 2 fails, if Chillers 2 and 3 fail, and so on.
- Load-confirmation timer
- Avoids transient conditions
- Staging-interval timer
- Allows time for the system to respond to turning a chiller on
- Minimum-cycle timer
- Prevents excessive cycling
In addition to turning chillers on and off, there are other functions of the chiller-plant control system that help prevent system flow instability from disrupting chiller operation. Flow instability can often be caused by normal valve and pump operation. The first is time delays.
Excessive cycling can be detrimental to the life of a motor. For this reason, turning a large motor (such as those used in large chillers) on and off should be minimized. Chilled-water systems typically have a large thermal mass (water in the system) and benefit from the diversity and slow rate of change of the system cooling load. Fast reactions, therefore, are typically not required. In fact, a response that is too fast will often cause system instability, waste energy, and cause unnecessary wear on mechanical equipment. To achieve stable and accurate control, many chiller-plant control systems provide time delays that can be adjusted by the operator to help minimize chiller cycling.
The first time delay is the load-confirmation timer. Its purpose is to delay turning on an additional chiller for a period of time following the initial indication that an additional chiller is required. This confirms that the indicated load is not a transient condition that would cause the chiller to be turned on and then quickly turned off.
The second time delay, which works in conjunction with the first, is a staging-interval timer. Its purpose is to allow the system time to respond after a chiller has been turned on. This prevents more chillers from turning on than are actually required, particularly during periods of pull-down or rapid load variation.
The third time delay is a minimum-cycle timer. This timer should have the highest priority. It requires a fixed period of time between turning an individual chiller on and turning it back off. This ensures that the chiller is not cycled too frequently.
It is important to understand that these timers are lower priority than the safeties built into the individual chiller controls. At all times, the individual chiller safeties must be capable of shutting the chiller down to avoid equipment damage.
Unload Before Start
he next control function is to partially unload the operating chillers before an additional chiller and pump are turned on. Depending on the system configuration, there can be very rapid variations in water flow through the chiller evaporator when a pump is turned on or off, or when a control valve is opened or closed. Partially unloading the chiller prior to such variations allows the chiller to continue to operate without interruption.
This can be explained by looking at a flow diagram for a chilled-water system with multiple pumps. This diagram shows that, with one pump and chiller operating, the flow rate through the chiller is 610 gpm [38.5 L/s]. When the second, same-size pump and chiller are turned on, the flow rate through the system increases to 870 gpm [54.9 L/s], but the flow through each chiller drops to 435 gpm [27.4 L/s]. This is an instantaneous reduction of 175 gpm [11 L/s], or 30 percent, through the first chiller.
The temperature of the water leaving the chiller and the temperature of the refrigerant in the evaporator drop as a result of this drastic flow reduction. New, advanced chiller controls may allow the refrigerant temperature to drop below the fluid’s freezing point for a brief period of time while the compressor unloads. The evaporator low-temperature safety may, however, turn off the chiller if the controls and compressor cannot react quickly enough.
The “unload-before-start” function partially unloads the operating chillers, raising the refrigerant temperature in the evaporator, before the flow reduction occurs. The chillers are allowed to reload as soon as the additional chiller is turned on.
Another control function that is desirable is called soft loading. It is typically used when the system has been off for an extended period of time and the chilled-water temperature is the same as the ambient temperature inside the building.
Soft loading either delays turning on additional chillers or varies the chilled-water set point, allowing the operating chillers to gradually catch up to the building pull-down load. This results in a very smooth pull-down, prevents overshooting the set point, and operates only the equipment required to satisfy the actual system load.
Constant-flow chilled-water systems frequently require individual chiller set- point control. Its purpose is to help maintain the system supply-water temperature by compensating for the bypass of return water through non-operating chillers.
The chiller-plant control system adjusts the individual set points for the operating chiller to “overcool” the water before it mixes with the higher-temperature water that bypasses through the non-operating chiller. The result is that the chilled water supplied to the system is as close as possible to the desired temperature. There are limits to the amount of overcooling. Depending on the design of the chilled-water system, one of two situations may exist. Either the chiller may not have been selected to produce cold-enough water, or the temperature required may be below the freezing point of the water being cooled. In either case, the control system must be intelligent enough to limit overcooling in order to prevent damage to the chiller.
Additionally, the control system must know when to turn another chiller on to meet the system chilled-water-temperature set point. Turning an additional chiller on may be required to meet the system demand for flow, even though the operating chiller may not be fully loaded.
- Decrease condenser-water temperature
- Increase chilled-water temperature
- Chilled-water pump (variable-flow system)
- Increase chilled-water ∆T
- Cooling tower
- Increase condenser-water temperature
- Condenser-water pump (variable-flow system)
- Increase condenser-water ∆T
The chiller-plant control system can also be used for system optimization. For the purposes of this discussion, we will define optimization as minimizing the energy used by the chiller plant (including chillers, chilled-water pumps, condenser-water pumps, and cooling tower) while still maintaining comfort or satisfying process loads.
The first step is to examine the energy use of the major components of the chiller plant, to see what can be done to minimize each component individually
The chiller energy usage can be reduced by lowering the condenser-water temperature or by raising the chilled-water temperature.
In a variable-flow system, chilled-water pumping energy can be reduced by lowering the chilled-water temperature while increasing the system ∆T. With the lower water temperature and increased ∆T, the coil requires less water flow to handle the same load.
Cooling-tower energy can be reduced by increasing the condenser-water temperature. This allows the tower fans to cycle or slow down. Condenser- water pumping energy can be reduced by increasing the ∆T through the condenser side of the system, thereby pumping less water. This is achieved by reducing the water flow through the condenser.
Obviously, looking at only a single component presents a conflicting picture for energy reduction, and a change in one component has an impact on other components. To truly optimize the chiller plant, all components must be analyzed together.
Chilled Water Reset
- Reduces chiller energy
- Can work in constant-flow systems
- Increases pump energy in variable-flow systems
- Can cause loss of space humidity control
- Complicates chiller sequencing control
As previously stated, as the chilled-water temperature set point is reset upwards, the chiller will use less energy. In constant-flow systems, this chilled-water reset strategy is fairly simple to implement and can be controlled based on the drop in return-water temperature.
In a variable-flow system, however, as the chilled-water temperature increases, the pumping energy also increases. While the COP of the chiller is approximately 6.5, the COP of the pump is about 0.65. Often the increase in pump energy will be more than the amount of chiller energy saved, especially because the chiller will often operate at part-load conditions. Another potential problem with resetting the chilled-water temperature upward is that space humidity control can be compromised if the water gets too warm. Finally, the chiller-plant control system must account for the changing supply-water temperature.
ASHRAE/IESNA Standard 90.1–1999 (Section 188.8.131.52) requires the use of chilled-water temperature reset in systems larger than 25 tons [88 kW]. It does, however, exclude variable-flow systems and systems where space humidity control will be compromised.
Some engineers feel that designing the system for low flow rates and a lower supply-water temperature, thus minimizing pump energy use, might be a better answer than attempting to reset the temperature upward.
Lowering the temperature of the condenser water can also reduce the energy consumption of the chiller. Depending on the system load and outdoor conditions, cooling towers typically have the ability to supply colder condenser water than at design conditions. This, however, increases the energy consumption of the cooling tower fans. The key to maximizing energy savings is knowing the relationship of cooling-tower energy consumption to chiller energy consumption.
At design conditions, a chiller typically uses five to ten times more energy than a cooling tower. This would suggest that it might be beneficial to use more cooling-tower energy to save chiller energy. However, there is a point of diminishing return where the chiller energy savings is less than the additional energy used by the cooling tower. Figure 106 shows the combined annual energy consumption of a chiller and cooling tower in a system that is controlled to various condenser-water-temperature set points. The third column shows a system that attempts to supply 55°F [12.8°C] water from the cooling tower at all times. Of course, at design conditions, the cooling tower may not be able to supply this temperature, but it will supply the water at the coldest temperature possible.
The fourth column shows a system that uses a control system to dynamically determine the optimal condenser-water temperature that minimizes the combined energy use of the chiller plus cooling tower. It is obvious that this method of optimal control minimizes overall system energy consumption.
Control of Condensing Pressure
Related to the issue of condenser-water-temperature control is the control of condensing pressure. Every chiller requires a minimum refrigerant-pressure difference between the evaporator and the condenser, in order to ensure that refrigerant and oil circulate properly inside the chiller. This pressure difference varies based on the chiller design and operating conditions. The chiller must develop the required pressure difference within a certain amount of time, as specified by the manufacturer, or the chiller controls will turn it off due to a safety limit. During some start-up conditions, this pressure difference may be difficult to achieve within the time required.
An example of such a condition is an office building that has been unoccupied during a cool autumn weekend. The temperature of the water in the sump of the cooling tower is 40°F [4.4°C]. Monday is sunny and warm, and the building cooling load requires a chiller to be started. Because the chiller is operating at part load and the tower sump is relatively large, the minimum pressure difference may not be reached before the chiller is turned off on a safety. If, however, the flow of water through the condenser is reduced, the minimum pressure difference can be obtained. The lower flow rate increases the temperature of the water leaving the condenser, which results in a higher refrigerant pressure inside the condenser. After the minimum pressure difference is reached, the flow may again be increased.
Either the refrigerant pressure in the condenser or the condenser-evaporator refrigerant-pressure differential can be monitored and used to control the temperature or flow rate of the condenser water, to prevent this pressure differential from dropping below the limit.
System-level communication and control is very important. Today, the amount of communication between the components (chillers, cooling towers, pumps, control valves, and so forth) has increased immensely, allowing many chilled-water systems to be fully automated.
In some facilities, however, the largest energy user in the HVAC system (the chiller plant) has not progressed beyond manual control. In some cases it was reduced to manual control shortly after the building was commissioned.
Why does this occur? Chillers are large, with very expensive pieces of equipment which, if damaged by incorrect operation, can cost the owner a substantial amount of money to repair or replace. Operators are, therefore, very sensitive to chiller plant operation. If the operator does not understand how the system is designed and controlled, it is likely that the system will be put into a manual control mode. Therefore, initial and ongoing operator training and support is critical.
There is an amazing amount of information available within a chilled-water system. Often the problem is not a lack of information, but how to interpret that information. Therefore, a clear and concise interface between the control system and the system operator is extremely important.
Information that should be communicated to the operator includes:
- Chiller-water system temperatures
- Chiller status (on or off)
- Information specified by ASHRAE Guideline 3
- Any pending control actions (chiller about to turn on or off)
- Status of system time delays
- Status of ancillary equipment (pumps, cooling towers, and so forth)
In addition, the chiller-plant control system should notify the operator of problems that are occurring, or are about to occur, in the system. These warning or diagnostic messages may point to a single piece of equipment malfunctioning, or be indicative of system changes that may cause problems. Diagnostics that occur at the chiller control panel should be communicated to the chiller-plant control system.
chiller operating log
ASHRAE Guideline 3
- Chilled-water inlet and outlet temperatures and pressures
- Chilled water flow
- Evaporator-refrigerant temperature and pressures
- Evaporator approach temperature
- Condenser-water inlet and outlet temperatures and pressures
- Condenser water flow
- Condenser-refrigerant temperature and pressures
- Condenser approach temperature
- Compressor-refrigerant suction and discharge temperatures
- Oil pressures, temperature, and levels
- Refrigerant level
- Vibration levels
- Addition of refrigerant or oil
ASHRAE Guideline 3, Reducing Emission of Halogenated Refrigerants in Refrigeration and Air Conditioning Equipment and Systems, includes a list of recommended data points to be logged daily for each chiller. Much of this data may be available from the display on the chiller control panel. It is also helpful to the operator if this information is available at the chiller-plant control system and presented in a clear format.
In addition to current status, historical operating information is valuable for keeping the equipment operating at peak efficiency and for identifying operating trends that signal either impending problems or a drop in system performance. For example, the condenser approach temperature is the temperature difference between the water leaving the condenser and the refrigerant inside the condenser. If there has been a problem with water treatment in the cooling tower, fouling may build up inside the tubes in the chiller condenser. This will cause the difference between the condenser water and refrigerant temperatures to increase, reducing chiller efficiency. By noting an increase in this approach temperature, the operator can schedule cleaning of the condenser tubes. By monitoring system and equipment trends, the operator has a chance to fix minor issues before they cause operational problems.