New Benchmarks for Steam Turbine Efficiency

Issue 8 and Volume 106.

By Dr. Alexander S. Leyzerovich,

In the 20th century, steam turbines became the most powerful electric power generators available, accounting for more than 50 percent of the world’s installed power generation capacity. However, many people, even some power engineering professionals, had come to view steam turbines as a mature technology that would not experience any remarkable achievements in the near future. Indeed, by the late 1980s, the thermal efficiency of new steam turbines had practically stabilized. But the 1990s brought new breakthroughs in steam turbine technology, and technology progress continues today. This progress is primarily the result of two main factors. The first is the development of new heat-resistant high-chromium-percentage ferritic-class steels that enable steam turbines to reach elevated steam temperatures without resorting to austenitic steels. The second is implementation of new advanced approaches to steam path design. Noteworthy as well are advances in developing longer last-stage rotating blades that further decrease exit losses.

The leading producers of large power steam turbines in the world today are European-based multinationals ALSTOM and Siemens AG; GE Power Systems (GE) in the U.S.; Mitsubishi Heavy Industries (MHI), Hitachi and Toshiba Corp. in Japan; Leningrad Metallic Works (LMZ) in Russia; Ansaldo Energia in Italy; Turboatom in Ukraine; and Skoda in Czech Republic.

Latest and Greatest

The advances in steam turbine technology can be better understood by reviewing the design, installation, and commissioning results of several power plants that have recently come on-line. By late 2001, the first operating year’s and acceptance tests’ data had been processed and partially published for two of the newest power units commissioned in Germany and Japan. These two units are the 907 MW Unit 1 of the Boxberg power plant, operated by Eastern-Germany utility VEAG, and the 1050 MW Unit 2 of the Tachibana-wan power plant, located in Tokushima Prefecture (Shikoku Island) and operated by Electric Power Development Co. (EPDC).

Figure 1. Siemens steam turbine in use at Boxberg power plant in Germany. Photo courtesy of Siemens.
Click here to enlarge image

Boxberg Unit 1 went on-line in June 2000 and passed acceptance tests in October 2000. The unit’s net efficiency was 42.7 percent and the gross efficiency of the Siemens steam turbine was 48.5 percent. Its steam conditions of 3860 psi and 1013/1078 F do not practically differ from those of other recent-vintage turbines at German power plants.

Tachibana-wan Unit 2 entered commercial operation in mid-December 2000. With a gross efficiency of 49 percent, its MHI steam turbine has been acclaimed the most efficient worldwide. The unit’s steam conditions, at 3636 psi and 1112/1130 F, represent the next step in the Japanese steam temperature staircase: 1000/1051 F for Matsuura Unit 1 (1000 MW, 1990); 1000/1099 F for Hekinan Unit 3 (700 MW, 1993); 1051/1099 F for Nanao Ohta (500 MW, 1995); 1099/1099 F for Matsuura Unit 2 (1000 MW, 1997); and 1112/1112 F for Misumi Unit 1 and Haramachi Unit 2 (1000 MW, 1998).

The cited steam turbine efficiency figures provide a benchmark for new units. Meaningful is that both power plants burn solid fuel. It is worth recalling that the best steam turbines put into operation in the 1990s had already reached comparable gross efficiency values:

  • 47.4%-Japan’s Hekinan Unit 3, with a rated output of 700 MW and steam conditions of 3480 psi, 1000/1099 F (MHI turbine);
  • 47.6%-Germany’s Hessler plant, with a rated output of 720 MW and steam conditions of 3990 psi, 1072/1112 F (ALSTOM turbine); and
  • 48.4%-Japan’s Kawagoe Units 1 and 2, with a rated output of 700 MW and steam conditions of 4496 psi, 1051/1051/1051 F (Toshiba turbines).

Of significance is that these units, as well as those at Boxberg and Tachibana-wan, achieved these close efficiency values with the turbines at materially different steam conditions. According to MHI, a steam temperature increase from 1000/1100 F to 1112/1112 F makes a turbine more efficient (heat rate) by about 2.2 percent, that is, its efficiency rises by approximately 1.1 percent. According to German power plant engineers, raising the steam parameters from 3625 psi, 1004/1040 F to 3915 psi, 1085/1112 F should increase turbine efficiency about 1.3 percent. So, closeness of the actual efficiency values for the same capacity class turbines with remarkably different steam temperatures, and with regard to differences in the condenser vacuum (722 mm Hg at Boxberg and 730 mm Hg at Tachibana-wan), feedwater heating, etc., says at least that some of these turbines have noticeable reserves to increase efficiency by reducing losses in the steam path and using more progressive designs.

Design Features

The turbines at Boxberg and Tachibana-wan significantly diverge in their design schemes. Siemens’ 907 MW 3000-rpm turbine at Boxberg is a tandem-compound (TC) five-cylinder machine (single-flow HP cylinder, double-flow IP cylinder, and three double-exhaust LP cylinders, Figure 1), whereas MHI’s 1050 MW turbine at Tachibana-wan (Figure 2) is a typical cross-compound machine with the HP and IP cylinders positioned on the high-speed shaft (3600 rpm) and two double-exhaust LP cylinders on the low-speed shaft (1800 rpm). Today, the largest west-European TC turbines in operation for fossil-fuel power plants have a single capacity of 933 MW (ALSTOM units at the German power plant Lippendorf), second only to the Soviet TC 3000-rpm turbine with a rated output of 1200 MW. The German Niederaussem plant will reach about 1000 MW when Unit K comes on-line in November 2002. The maximum single capacity of the “high-speed” TC steam turbines manufactured by Japanese producers has remained at about 840 MW.

Figure 2. MHI steam turbine in use at Tachibana-wan power plant in Japan. Photo courtesy of MHI.
Click here to enlarge image

It is well known that the ultimate capacity of a given turbine is, to a degree, determined by the length of the last-stage blades (LSB). Siemens’ recent-vintage 3000-rpm large steam turbines, including the Boxberg units, are furnished with free-standing 39-inch steel LSB that provide an annular area of 98.6 ft2 per flow. Under development are 45-inch steel LSB (first applied in the Niederaussem K turbine) and 53-inch titanium LSB. Table 1 compares LSB characteristics for several recent steam turbine models from various manufacturers.

For the time being, Japanese steam turbines in the 1000-1050 MW range are designed cross-compound with two LP cylinders on the low-speed shaft. Until recently, these had been furnished with 41- and 44-inch steel LSB. The turbine for Tachibana-wan Unit 2, however, uses an integrally shrouded 46-inch LSB. This significantly reduces the energy losses in the last stage and makes it possible to use the same turbine configuration for a single capacity of up to 1300 MW. Moreover, a few years ago Hitachi declared their readiness to produce high-speed TC four-cylinder (HP+IP+2xLP) 1000 MW turbines with new LSB. MHI has similar projects. In the future, such a turbine could even transform into a three-cylinder machine-with the use of an integral HP-IP cylinder. In this case, a two-cylinder (HP-IP+LP) scheme could be used up to a single capacity of 750 MW. MHI also has applied 3600 rpm, 40-inch steel LSB-the scale design of a 3000 rpm, 48-inch LSB -in a 700 MW TC three-cylinder unit that started commercial operation in April 2002.

MHI traditionally produces its steam turbines with reaction blading in the HP and IP sections. Above 1000 MW, both cylinders are designed double-flow, symmetrical about the steam admission plane. In particular, the HP cylinder has double nozzle boxes and a double-disc control stage with triple-pin blades. For the high-temperature (HP and IP) rotors of 1112/1200 F-class steam turbines, MHI uses a new ferritic 12 Cr steel with first-stage steam cooling of the rotating blade attachment zones. To reduce bearing wear, the rotor journals are overlaid with a lower Cr weld material.

The highest-temperature rotating blades of the first stages are made of austenitic refractory alloy R26. For stationary parts, the nozzle chambers, inner casings and No. 1 blade rings are made of 12 Cr cast steel, and 9 Cr forged steel is used for the HP valve casings and steam admission pipes. The turbine steam path, designed using advanced three-dimensional (3D) flow analysis techniques, is equipped with twisted, tapered, bowed and inclined vanes and blades. All the rotating blades are integrally shrouded with improved labyrinth seals. It is estimated that the heightened steam parameters, decreased exhaust losses due to the increased LSB length, and three-dimensionally designed blading improved the turbine heat rate about four percent as compared with MHI’s steam turbines launched before 1993.

Relative to its contemporary Japanese counterparts, Siemens’ Boxberg turbine has rather modest steam parameters. However, while the MHI turbine is rather “conservative” in design, the Siemens 1000 MW-class turbine has many original design features. Because of the relatively moderate length of the LSB, the turbine is made in five cylinders, including three LP cylinders. The 180-foot turbine-generator set is mounted on a specially tuned, spring-supported reinforced-concrete foundation. To facilitate the turbine’s thermal expansion, the bearings are rigidly mounted on the foundation; the outer casings of the HP and IP cylinders and the inner casings of the LP cylinders rest on the adjacent bearing pedestals, and the cylinder casings are free to slide about them along the axial keys. The shaft line expands from the combined journal-and-thrust bearing located between the HP and IP cylinders. All the rotors are made of solid forgings (without a central bore) with forged-on coupling flanges and are joined with hydraulically tensioned bolts.

As is typical for Siemens’ large capacity machines, the turbine is designed with combined stop-and-control valves, throttle steam admission control, a single-flow two-shell HP cylinder without the control stage and nozzle boxes, and with a barrel-type outer casing (i.e., without a bolted horizontal joint). Despite its remarkable overall weight of 120 tons, the cylinder was delivered to Boxberg fully assembled. The IP cylinder is of dual-flow, two-shell design. The upper and lower halves of its outer and inner casings are bolted together along a horizontal flange joint. Extra ribs reinforce the inner casing, and a thermal shield counteracts the uneven temperature distribution caused by unidirectional flow of steam leaving the cylinder through the upward port in the outer casing.

Click here to enlarge image

Even though Siemens, as well as MHI, traditionally employed reactive blading in the HP and IP sections, the latest turbines enable stage reactivity to be varied over a wide range. Interestingly, GE, which traditionally employed impulse-type blading, came from the opposite end of the spectrum in developing its “Dense Pack Steam Path” with intermediate stage reactivity.

Boxberg’s three LP cylinders are dual-flow with multiple-shell casings. The entire weight of the outer casing with its reinforcement beams rests on the condenser dome, to which it is rigidly welded. The cast inner casing is likewise of two-shell design, with the inner shell centered in the outer casing so as to be free to slide axially in response to thermal expansion. The inner casing’s outer shell is provided with a special droplet shield. Besides mechanical protection, the shield promotes superheated steam formation between itself and the outer shell, which thermally insulates the latter and reduces heat losses.

The steam paths of all three (HP, IP and LP) turbine sections were designed with 3D technology that resulted in the use of twisted, tapered, bowed (curved) and inclined vanes and blades. All but the LSB are integrally shrouded with optimized labyrinth-type seals. These advances provide about a two percent efficiency increase over conventional blading. Advanced CFD computation methods were also used to upgrade the non-bladed turbine areas. Siemens developed a new geometry for steam admission and exhaust paths by widening their flow area and installing special baffles, razors and screens to avoid backflow and vortex formation and reduce energy losses. The power plant acceptance tests demonstrated internal efficiencies for the HP and IP cylinders of 94.2 percent and 96.1 percent, respectively.

Higher and Higher

Even though the efficiencies reached at Boxberg and Tachibana-wan are impressive, they likely represent only interim, temporary benchmarks. Niederaussem Unit K, for example, is a potential champion. The Siemens machine at Niederaussem will have more elevated steam conditions (3990 psi, 1076/1112 F), lower condenser pressure (28.9/35.5 mbar), and a 25 percent larger exhaust area due to the use of the longer LSB. The new unit is designed to reach 45.2 percent thermal efficiency. For comparison, the previous 600 MW units at Niederaussem (G/H), with steam conditions of 2525 psi and 986/986 F, have a net efficiency of 35.5 percent. An even higher net efficiency is targeted for Westfalen’s Unit D, with a single capacity of 350 MW, steam conditions of 4210 psi, 1112/1148 F, and condenser pressure of 35 mbar.

Higher performance efficiency is also expected from new Japanese power units, due to further heightened steam conditions (up to 4350 psi, 1166/1166 F by 2005) and turbine steam path advances.

For steam turbine retrofits, the most attractive component opportunities are typically the LP cylinders. Refurbishment usually comprises the use of longer LSB, advanced 3D design of the blading and steam path (including an improved meridional profiling), the use of integrally shrouded blades with multiridge seals, improved exhaust hood diffuser, and so on. An advanced LP cylinder proposed by ALSTOM features a special, “conical” meridional outline of the steam path with 3D, leaned and curved last-stage vanes that provide more uniform steam flow distribution along the height, with less vorticity. According to ALSTOM, modernizing the LP steam path of 20-year-old 300-700 MW turbines by applying all of these measures decreases the turbine heat rate by more than 3 percent.

There also exist some opportunities for raising steam turbine efficiency that have attracted less attention. Consider a typical 570 MW turbine with steam conditions of 2400 psi, 1000/1000 F. Steam leakage flow from the HP valve seals amounts to about 5000 lb/hr (based on the rated main steam flow of about 3,800,000 lb/hr). This loss is equivalent to about a 1 MW loss in turbine capacity. Naturally, the higher the main steam parameters, the greater this loss. At the same time, it could be avoided if the usual labyrinth seals at the valve stems are replaced by special, hermetic liquid-metallic seals (LMS) preventing steam from flowing along the stem. Due to a low friction coefficient (less than 0.05), the LMS does not hamper the valve motion freedom. Such seals were developed in the former Soviet Union and, after detailed bench tests, since 1987, have been installed on the HP control valves of about 20 supercritical 300 MW turbines. They have successfully operated without additional inspections or maintenance outside scheduled overhauls every six years. No forced outages have occurred because of the seals. According to power plant data, the turbine efficiency has increased about 0.2 percent.

Advances in meridional profiling, 3D shaping of the steam path, and optimizing the stage reactivity significantly reduce the profile, secondary and incidence energy losses in the steam turbine stages. There also exist design modifications to reduce leakage losses through the overshroud (tip) and diaphragm (or undershroud) glands. In addition, it is no good underestimating “root leakage”-the steam flow through the axial clearance between the nozzle vanes and rotating disc in the stage root section. Depending on the stage geometry and reactivity, this flow can be directed from the diaphragm gland seal into the rotating blade row or, by contrast, from the nozzle vanes into the disc’s balance holes. Experiments show that in the latter case, by optimizing the root leakage flow amount, removal of the root steam layer disturbed by the secondary (end) losses can increase the stage efficiency by up to 0.4-0.5 percent.

The increases in turbine efficiency gained due to individual measures are typically counted in fractions of a percent. However, implementation of all these design improvements could raise the turbine efficiency remarkably. A gross steam turbine efficiency level of 50 per cent is possible in the near future.

It is understandable that power units designed for the highest efficiency are primarily intended for baseload operation. At the same time, especially with their increased use in Germany and Japan, these steam turbines are designed such that their flexibility allows them to accommodate demand changes by deep unloading or shutdown.

It should be emphasized that the steam turbines advances discussed in this article primarily pertain to large-scale utility power units. It will take time to transfer their technical details into “ordinary” serial steam turbines of smaller capacity. This is especially true for the numerous steam turbines being used in combined-cycle units and cogeneration facilities, which require different approaches and encounter other specific problems.


Dr. Alexander Leyzerovich, an independent consultant in power engineering, earned his scientific degrees from All-Russia Thermal Engineering Research Institute (VTI), where he worked as a leading specialist in the operation and analysis of large steam turbines. He has authored numerous professional publications, including the two-volume book, “Large Power Steam Turbines: Design & Operation,” published by PennWell in 1997.