First Kalina combined-cycle
plant tested successfully
By Hank Leibowitz and
Mark Mirolli, Exergy Inc.
California plant demonstrates the practicality of the Kalina cycle
In November 1996 the world`s first Kalina combined-cycle plant went into operation. This 6 MW plant, located at a U.S. Department of Energy (DOE) facility near Los Angeles, Calif., had just completed four years of operation and extensive testing as a bottoming cycle using an external heat source. Those tests have given the developers confidence that the Kalina combined-cycle unit will operate reliably and prove the efficiency of the process.
The Kalina cycle has been the subject of numerous technical articles (Power Engineering, July, 1996) since the original patents were issued in the early 1980s. Its significant performance improvement over state-of-the-art Rankine cycle technology is attributable to its use of variable composition ammonia-water mixtures as the cycle`s working fluid.
California demonstration plant
Approximately 10 years ago, Exergy Inc., the technology`s owner, raised private capital to construct a 3 MW demonstration plant at the Energy Technology and Engineering Center (ETEC), a DOE facility located near Canoga Park, Calif., approximately 40 miles northwest of Los Angeles. Exergy believed that the power industry needed to witness the Kalina cycle`s practical implementation before the market would accept it as a functional replacement for the Rankine cycle.
The demonstration plant was constructed in 1991 and became fully operational in August 1992. A summary of the plant`s operating conditions follows. Turbine inlet: 960 F, 1,600 psia, 31,450 lb/hr; turbine exhaust: 21.1 psia; cooling water: 52 F; and output: 3,050 kW at the generator terminal.
Exergy configured the plant as a bottoming cycle, using a gas-fired, waste heat stream from which to produce 3 MW of electrical output. Figure 1 is a schematic of the process flow. This figure shows four different ammonia-water compositions, which range in ammonia content from 95 percent down to 33 percent. The main working fluid composition, the one used in the boiler and turbine, is 70 percent ammonia, while the condenser directly downstream of the turbine operates with a 42 percent ammonia mixture. The leaner fluid has a lower vapor pressure, which allows for additional turbine expansion and greater work output. The additional degree of freedom provided by the ability to vary the working fluid`s composition, not found in the Rankine cycle, gives the Kalina cycle its significant performance advantage (10 to 15 percent or more in high temperature applications).
From August 1992 to November 1995 the plant accumulated approximately 5,200 hours of fired operation. Tests confirmed that the Kalina cycle process is fundamentally sound and can be successfully applied to large-scale, commercial power plants. Test campaigns conducted between 1992 and 1995 achieved the following results:
z -Performance data proved that the Kalina process is more efficient than the Rankine cycle. The thermodynamic property data of ammonia-water mixtures used in the plant were confirmed.
z –Operators could easily change and control the composition of the working fluid.
z —The plant operated safely, stably and reliably. There were no unscheduled shutdowns due to the Kalina process.
z -All materials and equipment were proven to be compatible with the ammonia working fluid.
z -Operation and maintenance were performed routinely.
As is the case in most demonstration projects, plant operators encountered some problems and learned several lessons during the test programs. The problems that were encountered were associated with individual components, not the process. Examples of this are as follows:
z -Low vapor turbine performance. Approximately 200 kW were lost due to the under performance of the turbine`s control stage.
z -Low condenser performance. Approximately 120 kW were lost due to the improper specification of the condensers. Engineers learned that the condensers should be designed as vertical, counterflow units instead of horizontal, cross-flow units.
z -Seals. The labyrinth seals on the turbine shaft and the packing rings in the feed pump both suffered from excessive leaks that have since been corrected.
At the end of 1995, the Kalina plant`s heat source was removed from service as a result of DOE budget reductions. Consequently, Exergy purchased a Solar Centaur 50 combustion turbine to replicate the waste heat source. An in-line auxiliary heater (duct burner) was installed to supplement the Centaur`s exhaust temperature and flowrate, which were less than that of the original waste heat stream. Nominal output is 6.5 MW. In December 1996, the plant resumed operation as the world`s first Kalina combined-cycle plant.
1996 to 1997 test campaign
The 1996 test campaign included testing in three major areas in order to provide detailed engineering data to support commercialization of the Kalina cycle for combined-cycle, geothermal and direct fired systems. These areas included thermal performance testing at 100 percent of maximum continuous rating (MCR), as well as off-design steady-state and transient tests, noncondensable gas tests and component materials testing. The plant logged approximately 1,500 operating hours during the test campaign from November 1996 through February 1997.
The primary objectives of the performance testing in the 1996 campaign were to provide detailed engineering information to document the Kalina bottoming cycle performance at the plant boundaries and to provide engineering data throughout the plant for individual components. Operators took data over a range of operating conditions to document the performance of individual heat exchangers within the distillation condensation subsystem and other plant components. As a secondary objective, the testing evaluated the stability of operation over a range of bottoming cycle loads and ammonia compositions. These data will form a basis for engineering model development for future Kalina cycle plants.
Exergy conducted performance tests of one to two hours` duration, over a range of plant operating conditions from 2.7 to 1.6 MW, at steady-state loads, after a pre-test stabilization period of several hours. It also performed tests during transient conditions including plant startup, shutdown and various load ramps. Throughout the test period, the plant data acquisition system logged data at 2-second intervals, and the measured parameters were averaged.
Table 1 summarizes the steady-state load tests. Data obtained during a 1992 test period at 2.6 MW are included in Table 1 for comparison. A summary of the plant predicted design performance for the 100 percent MCR condition of the plant is also included.
These data were obtained prior to the 1996 installation of the gas turbine. Test results proved the stability of the Kalina bottoming cycle over a range of plant loads and ammonia compositions. Exergy performed steady-state tests over a bottoming cycle load range of 2.7 to 1.6 MW. Throughout the 1996 test campaign, ammonia compositions varied over the range of 45 to 70 weight percent ammonia through the heat recovery vapor generator (HRVG) and the vapor turbine. In addition, Exergy performed plant transient tests, including cold startup and plant trips, to demonstrate the operational stability of the Kalina process.
Noncondensable gas testing
In designing the plant equipment, engineers must take into account the accumulation of noncondensable gases in the working fluid of thermal power cycles. Failure to do so will have a deleterious effect on the heat transfer performance of surface condensers. In Rankine cycle surface condenser designs, noncondensable gases are removed through the use of reciprocating vacuum pumps or steam-jet air ejectors. Most of the noncondensable gases enter the cycle as a result of air infiltration into the condenser, which is operated under high vacuum conditions.
The surface condensers in a Kalina cycle plant do not operate under vacuum conditions, so air infiltration is not an issue. However, at the operating conditions typical for superheaters and reheaters and in the high temperature areas of the turbine generator, ammonia may decompose to form hydrogen and nitrogen, which will accumulate in the cycle working fluid as noncondensable gases.
During normal operation of the Canoga Park facility, operators routinely vent the high pressure (HP) and low pressure (LP) condensers every day. However, the composition and volume of the vent gases were unknown prior to this test campaign. Plant operators established a stable operating point and vented both the HP and LP prior to the start of each noncondensable test, then operated the plant at a stable condition for 24 to 48 hours. Then, at the completion of the test period, they vented both the HP and LP condensers, measured the vent gases` volume and determined the composition of the major gaseous species by gas chromatography. The total of 12 tests covered a range of plant operating conditions, with the HRVG superheater outlet temperatures varying over a range from 900 to 980 F. Table 2 summarizes a portion of the test results.
Test results quantified the volumetric flow and composition of vent gases over a range of operating conditions. The decomposition rate column in Table 2 assumes that all of the hydrogen gas measured in the vent gases originated as ammonia. This assumption is conservative because some of the hydrogen is probably a result of water decomposition, or possibly an artifact resulting from hydrazine addition to the vapor turbine steam sealing system. The “yearly makeup” column is a calculation of the ammonia makeup required to refresh the working fluid, thereby accounting for the ammonia depletion as a result of decomposition. Based upon these data, a 400 MW Kalina cycle power station operating at similar conditions will require the addition of approximately 17,000 gallons of ammonia per year to account for noncondensable gas venting. This represents a minimal cost of approximately $14,000 per year.
The ammonia-water working fluid in a Kalina cycle plant presents different material issues than in a steam plant. Oxidation of plant components throughout the power cycle is less likely because oxygen levels within the working fluid are extremely low. However, nitridation of high temperature components is a concern which should be considered when selecting superheater, reheater and high temperature turbine parts. The Canoga Park facility design includes test sections in the superheater outlet and turbine wheel chest for evaluation of material samples. Multiple sample exposures have been part of each test campaign since 1992 to add data to support future commercial plant designs, as well as life assessment activities.
Since startup in 1992, there have been no corrosion-related failures of plant cycle equipment caused by the ammonia-water working fluid. However, material exposures in the test sections within the plant have contributed to help understand the mechanism of nitridation in a working plant environment.
Test specimens of various materials are prepared and installed in the superheater outlet test section (1,660 psia, 960 F) and turbine wheel chest (515 psia, 780 F). These coupons remain throughout the test campaign. At the completion of testing, the coupons are removed for laboratory evaluation to determine nitridation resistance or the impact of the exposure on the mechanical and chemical properties of the test specimen. Exergy will soon remove the material specimens from the 1996 test campaign at Canoga Park and subject them to laboratory testing. z
Hank Leibowitz joined Exergy`s predecessor company in 1986 as vice president of project engineering and business development and has served as vice president of special projects since December 1995. Leibowitz has also held positions with Fayette Manufacturing, Mechanical Technology Inc. and General Electric Co. He has a master`s degree in mechanical engineering from the University of Connecticut and a masters of business administration degree from Rensselaer Polytechnic Institute.
Mark Mirolli joined Exergy in 1995 as senior vice president of project development and is responsible for technology development and for providing engineering support for development of initial commercial power plants using the Kalina cycle. Mirolli also served in various capacities at ABB and has a master`s degree in mechanical engineering from Rensselaer Polytechnic Institute.