By Tek Sutikno, Duke Fluor Daniel
As the ambient temperature increases, a gas turbine’s output decreases. Since spot market prices for power invariably increase in the summer, the reduced gas turbine output at high temperatures can dramatically affect the financial success of a power plant. One solution to this problem is to install inlet air cooling to the gas turbine.
The three most common options for inlet air cooling systems are evaporative cooling, refrigeration chillers and inlet fogging. However, there are benefits and drawbacks to all these options. An evaporative cooler, or an inlet fogger, typically exhibits a low capital cost per marginal increase in power output. Unfortunately, these technologies become less effective as the relative humidity of the inlet air increases. Although chillers have higher capital costs, they can increase a gas turbine’s output when operated at high relative humidity levels. One of the drawbacks, however, is at higher humidity levels the chillers use more energy.
Chillers generally employ one of two types of refrigeration cycles: mechanical vapor compression or absorption. The efficiency of a chiller is expressed as the “coefficient of performance (COP),” which is the ratio of net refrigeration produced to the energy provided to the chillers. Absorption chillers typically have lower COPs (up to 1.2 ) than mechanical chillers. However, absorption chillers can use waste heat at temperatures as low as 250 F.1
Because waste heat is generally unavailable in merchant power plants, mechanical chillers are often used. Mechanical chillers have much higher COPs (as high as 6.0) than absorption chillers. Since this article focuses on merchant plant applications, only the performance characteristics of mechanical chillers are considered in evaluating net output gains.
Figure 1 shows a schematic diagram of a typical mechanical vapor compression chiller. In operation, the compressor pressurizes the refrigerant vapor to a discharge pressure high enough for complete condensation of the refrigerant vapor in the condenser.
From the condenser, the refrigerant flows into an accumulator and then through a pressure reduction valve, where it is partially vaporized.
The latent heat of vaporization cools the mixed phase refrigerant to a lower temperature. For the refrigeration cycle shown in Figure 1, the COP is equal to the refrigeration capacity divided by the power input of the compressor. To eliminate the potential risk of refrigerant leaking into the combustion turbine compressor, a secondary cooling water loop may be installed. However, a secondary loop increases the capital cost and decreases the COP.
The design parameters affecting a chiller’s capacity and its COP are primarily :
- Target chiller temperature
- Design ambient conditions
- Condensing temperature
Because the chilled air is near, or at saturation point, the target chiller temperature is normally specified at no lower than 40 F. Temperatures below 40 F can cause ice formation at the compressor inlet. For this article, a target chilled temperature of 45 F is used for evaluating the chiller’s performance.
Once the target chiller temperature is specified, the design ambient air temperature and relative humidity must be considered for sizing the chiller’s capacity. An excessively high design ambient parameter will result in an oversized chilling system. As a result, at partial loads, there will be higher capital and operation costs.
To maintain a minimum target chilled temperature of 45 F, a chiller rated to cool inlet air from 85 F wet bulb to 45 F will have to operate at partial loads when ambient wet bulb temperatures are less than 85 F. The efficiency (COP) of the cycle is reduced at partial load conditions, although this reduction may be minimized by employing multiple train chillers. Conversely, the opportunity to gain the financial benefit from the chiller cannot be fully taken when the specified ambient design parameters are too low.
One common approach for specifying the design ambient parameters is to use the historic, statistical weather climate data relevant to the plant’s location. This statistical information, plus the expected future spot market prices for power, should be reviewed before finalizing the design ambient condition. Except for locations with extreme weather conditions, the design ambient condition typically corresponds to a wet bulb temperature below 80 F. Wet bulb temperatures exceeding 85 F are rare.
For a chiller using a cooling tower, the cooling water supply temperature will dictate the condensing temperature of the refrigerant and the corresponding discharge pressure of the refrigeration compressor. Lower condensing temperatures give lower compressor discharge pressures, improved refrigeration cycle efficiency and reduced compressor cost. However, lower cooling water supply temperatures increases the size of the cooling tower and/or the associated refrigerant condenser. Again, the historic statistical weather data should be consulted to determine the optimum condensing temperature for a specific location.
Output gains from a 2×1 GE 7FA combined cycle plant configuration with a chiller are displayed in Figure 2. The chiller is sized to cool the gas turbine inlet air from 100 F at 37 percent relative humidity down to 45 F.
Figure 2 shows output gains from the chiller operating in a range of ambient dry bulb temperatures at a constant relative humidity of 37 percent. The chiller has a COP of approximately 4.17 and uses an arbitrarily selected R-290 as the base refrigerant with a condensing temperature of 120 F. The 4.17 COP is used to calculate the net power output gains from the chilling system. Obviously, if a selected chiller system has a higher COP, the net output gain will be higher.
The differences in output levels shown in Figure 2 represent the net gains relative to both non-chilled operations with and without evaporative coolers. Figure 2 also shows that the net output gain is reduced at lower ambient temperatures. The reduced gains are due to the target chilled temperature being specified at no lower than 45 F. In this case, the chiller would need to be operated at partial load.
When the ambient conditions exceed the design wet bulb temperature of 77.6 F (100 F, 37 percent relative humidity), the output from chilling starts to drop as shown in Figure 2. As indicated earlier, the chiller is sized to chill the 7FA gas turbine inlet air from a design wet bulb temperature of 77.6 F to 45 F. At wet bulb temperatures above 77.6 F the refrigeration load required for chilling the inlet air to 45 F will exceed the mechanical capacity of the chiller.
The net gains from chillers, relative to evaporative cooling, are significantly higher at high humidity levels. The downside is that the chiller load increases with humidity. At high relative humidity levels evaporative cooling becomes less effective.
To optimize the net output gains from chillers, the following should be considered during the design stage:
- Improving the refrigeration cycle to increase COP.
- Using heat integration design (HID) to maximize the net output gain.
COP improvement of a refrigeration cycle typically requires additional capital costs and may result in longer payback periods. HID, on the other hand, increases the net output gain from the chiller without necessarily improving its COP.
Increasing the COP of the chiller cycle decreases the power input to the cycle and increases the net output gain from the chiller. The COP of a mechanical chiller is determined by a number of factors:
- The phase equilibrium characteristics of the refrigerant
- Target chilled temperature
- The temperature range of the cooling medium (heat sink)
- The mechanical efficiency of the compressor
- The design of the cycle.
However, only the cycle design offers a number of design options for maximizing COP and reducing the chiller’s power consumption. The other factors are determined by site conditions, project requirements, the physical properties of the selected refrigerants and the type of compressors available.
Common design options for improving COP include compression staging, mixing refrigerants and installing energy recovery devices such as hydraulic turbines. All these features improve the “Second Law” efficiency of the refrigeration cycle. Although systems designed with the features mentioned have been widely used in the hydrocarbon industry2 their use in a 2×1 7FA combined-cycle configuration is less common.
Table 1 shows the percent increases of net MW output from COP improvement features relative to the base case refrigeration cycle design. The two-stage, multi-component refrigeration cycle with energy recovery schemes gives the highest COP.
The percent increases in Table 1 may be higher or lower depending on the extent of optimization. For example, the cycle with multi-component refrigerants can give a higher COP if the mixture composition is optimized to improve the second law efficiency.3 However, these COP improvement features increase the complexity of the chiller system. Any additional capital costs associated with COP improvement features should be considered before making a final design decision.
Using HID does not necessarily change the COP of the refrigeration cycle. However, it does increase the plant’s net output by integrating the heat in all three thermodynamic cycles (Brayton, Rankine, and vapor compression refrigeration) of the combined cycle plant with a mechanical chiller. The gain in a chiller’s output from HID will depend upon the specific design features for each of the cycles. For example, duct firing the heat recovery steam generator tends to increase the gains from an HID. Various references 4,5,6,7 discuss HID and its implementation in industrial applications.
An HID system can further increase the percent gains from COP improvement features in Table 1. As an example, an additional 1.5 percent net output gain may be obtained in an evaluated HID for a 2×1 7FA plant with a chiller having a COP of 4.36 and a target chilled temperature of 55 F. This is a significant increase as the COP maximizing features in Table 1 result in a maximum of 1.3 percent (8 MW). HID does not change the COP of the refrigeration cycle; it only increases the magnitude of the output gain.
For combined-cycle plants, HID optimizes heat utilization in the Brayton, Rankine and refrigeration cycles. The output gains from HID will depend on the following:
- Power output ratio of the gas turbine to the steam turbine.
- The design approach temperature of the cooling tower.
- The type of refrigerant used.
- The target inlet temperature to the gas turbines.
- The design of the heat recovery steam generator.
Compared to the COP improvement features typically requiring higher capital costs, an HID system does not necessarily increase the capital cost. The capital cost of the chiller in the evaluated HID with 1.5 percent net output gain is substantially less than the costs of other chillers without heat integration. For a combined-cycle plant with mechanical chiller, an HID scheme should be seriously evaluated and considered.
Dr. Tec Sutikno is with Duke-Fluor Daniel and is responsible for the front end design and engineering reviews of combined-cycle plants. Dr. Sutikno graduated from the University of Kansas with B.S., M..S., D.E. and MBA degrees. He is also a registered professional engineer.
- Dhamadikari, S. “Consider trigeneration techniques for process plants, Hydrocarbon Processing, July 1997.
- Kaiser, V., Becdelievre. C., and Gilbourne, D.M. “Mixed Refrigerant for Ethylene, ” Hydrocarbon Processing, Oct. 1978, p. 129.
- Mathiprakasam, B. and T. Sutikno, “Theoretical Analysis of the Use of Binary Refrigerant Mixtures in Vapor Compression Cycles,” Proceedings of the 22nd National Heat Transfer Conference, August 1984.
- Rossiter, A.P, M.A. Rutkowski, and A.S. McMullan, “Pinch Technology Identifies Process Improvement, ” Hydrocarbon Processing, January 1991.
- Morgan, S.W. “Use Process Integration to Improve Process Designs and the Design Process, ” Chemical Engineering Progress, September 1992.
- Buehner, F.W. and A.P. Rossiter, “Minimizing Waste by Managing Process Design,” CHEMTECH, April 1996.
- Sutikno, T., and W. Coons, “Utility Design Cuts Cost,” Chemical Engineering, April, 1998.