By Jörg Stahlhut and Dirk Band, Balcke-Dürr GmbH, Germany
For many of the new European supercritical coal-fired plants, builders are turning their attention to a heat exchanger technology that offers convincing technical and economic advantages over conventional U-type feedwater heaters. The header-type feedwater heater, also known as the SNAKE heater or header-type heater, has traditionally been used in German high-efficiency coal-fired power plants but is not widely used in the U.S., although adoption of the technology is starting to gain speed.
The world’s largest header-type heaters weigh more than 270 tons when empty. Three header-type heaters and a separate desuperheater normally form a complete high-pressure heater train. The heaters are usually installed upright and normally have three or four water passes, depending on the piping system.
The U-type heater is used for low tube-side pressures and flow rates, while the header-type heaters are predominantly used for high pressures and flow rates. A heater design with header pipes instead of tubesheets allows a much better thermoelastic operating behavior during start-up, turbine trip or cycling power plant operation. Because of these advantages, header-type heaters have established themselves in the high-pressure range of German power plants.
Popular Everywhere but the U.S.
In Germany, high-efficiency coal-fired power plants are traditionally fitted with header-type heaters. Header-type heaters are used in both 1,100 MW units F and G of RWE Power’s lignite-fired plant in Neurath, Germany (BoA 2 and 3).
Approximately 30 coal-fired power plant units are currently being built in Germany because of the National Allocation Plan 2008-2012. Each has a nominal output of 600 MW to 1,100 MW and the total planned output is 26 GW. These plants will set new standards, with efficiencies of more than 45 percent, flexible power plant operation and high availabilityand all of them will be fitted with header-type heaters.
High-pressure header-type heaters have also been used for some time in Denmark. Coal-fired power plants with unit outputs of 300 MW to 600 MW such as Avedöre, Ensted or Nefo have been operated successfully with header-type heaters for decades.
In South Africa, many conventional power plants are traditionally equipped with header-type heaters. Unlike the normal arrangement of one high-pressure heater train, the South African header-type heaters are installed in two parallel piping trains with two high-pressure heaters each. One of the first large-scale power plants to be equipped with header-type heaters was the Duvha Power Station, which went into operation in 1975 with an overall capacity of 6 x 600 MW.
Because of severe power shortages, South Africa plans to build new coal-fired power plants in the next 10 years that will provide an overall capacity of more than 20 GW. These new plants will be designed with header-type heaters because of the convincing advantages and good operational qualities of the technology.
Historically, China and South Korea have been greatly influenced by U.S. power plant technology but interest in header technology is growing there. The Korea Power Engineering Co. (KOPEC), the planning company for the South Korean power plant operator KEPCO, described the technical and economic advantages of header-type heaters over U-type heaters in great detail in its 2006 study, “Next Generation Thermal Power Plant.”
In the U.S, though, high-pressure header-type heaters were largely unknown. American coal-fired power plants were operated exclusively with U-type feedwater heaters until the 1990s, when serious damage occurred in some U-type heaters.
Damage occurred in several U-type heaters after 12 to 15 years of operation, benefited by a cycling operation of the power plant. This was caused by stress-induced cracks in components with thick walls, at least at the connection between the tubesheet and waterbox. The damage forced reduced output operation of the power plants, numerous unscheduled downtimes for repairs and quite often the replacement of the heaters. The Electric Power Research Institute (EPRI) launched a study in 1990 to identify options for avoiding damage and improving the availability of American coal-fired power plants, with a simultaneous reduction of operating costs.
The EPRI study underlined the technical and long-term economic advantages of header technology and described the benefits of a conversion to header-type heaters. U-type heaters in the 100 MW Glenwood 5 power plant in New York were replaced by header-type heaters in 1991. Other heater replacements followed, such as the 500 MW Sioux 2 power plant in Missouri, in 1996.
Some of the planned coal-fired power plant projects in the U.S. include header technology, but the concept may not catch on because American feedwater heater manufacturers do not have extensive header technology experience. Support for the traditional U-type heater may continue.
U-type and header-type heaters differ fundamentally in the separation of the tube and shell sides and the design of the heat exchanger tubes.
In the U-type heater, feedwater and heating steam are separated by the tubesheet and bundle of U-tubes. In header-type heaters, they are separated by headers and a bundle of snake-shaped tubes.
In U-type heaters, the feedwater passes through the inlet side of the waterbox, the U-shaped tubes and finally the outlet side of the waterbox. In header-type heaters, the feedwater initially enters the inlet header, passes through the snake-shaped tubes in three or four passes, and leaves the unit via the outlet header. (See Fig. 1 on page164.)
In supercritical steam power plants, high-pressure heaters have to withstand an internal pressure of 300 bar to 400 bar on their tube side. The mass flow of the feedwater ranges from 400 kg/s to 800 kg/s. Under these boundary conditions, the pressure-bearing components of the heater may have thick walls that react quite sluggishly to short-term changes in temperature. This leads to secondary stresses, which increase with the wall thickness of the components.
The tubesheet wall is thickest in a U-type heater. In high-output power plants, the tubesheet thickness is between 400 mm and 800 mm depending on the design data. Tests have shown that wall thicknesses over 500 mm must be regarded as critical. When wall thickness goes above the critical value, high peak stresses are induced through transient temperature gradients at the connection points between the thick tubesheet and relatively thin waterbox shell. The consequences are mostly cracks in the areas of the transition between the tubesheet and waterbox.
This type of damage is shown in Figure 2, by metallography and FE analysis. The micrograph shows the relief groove of a tubesheet with clear cracks and critical stresses for the material with values of more than 730 N/mm2. The results of the metallographic tests and FE analysis confirm that the cracking is caused by “expansion-induced corrosion.”
Figure 2 Micrograph and stress distribution in the transition area tubesheet/waterbox
To control the stress, the thickness of the tubesheet wall has to be reduced. That’s why U-type heaters are arranged in two trains in the high-pressure areain other words, twice the number of units in two parallel piping trains. The feedwater mass-flow through the heater can thus be cut in half, and the necessary shell diameter and tubesheet thickness reduced.
Thermoelasticity of the U-type heater nevertheless remains relatively low. Maximum temperature gradients of 10 F to 15 F per minute allow a very slow start-up and shutdown of the power plant and a very limited number of possible load changes, so operation in the cycling mode should be avoided.
U-type heaters in supercritical power plants often have a service life of 10 to 15 years and would have to be replaced up to three times in the average 40-year power plant lifetime. In comparison, header-type heaters with their excellent thermoelasticity have a number of advantages. Unlike the tubesheets in U-type heaters, the inlet and outlet header of a header-type heater have walls that are only 70 mm to 120 mm thick. Accordingly, the necessary header wall thickness under the same basic conditions is only around 15 percent of the tubesheet thickness.
Geometric discontinuities in the transitional areas between the header and shell that favor unacceptable stresses are completely avoided with the low wall thicknesses and wall thickness differences, thus ruling out the risk of damage through thermal cracks. Much lower stresses occur at the transition header/shell of header-type heaters than in the critical transition area of the tubesheet in U-type heaters.
The thermoelastic design of high-pressure header-type heaters allows a flexible mode of operation for power plants with a large number of load changes. Maximum temperature gradients of up to 40 F per minute permit a fast start-up and shutdown of the power plant with a service life of 35 to 50 years, and the heater section does not have to be designed as a two train system.
Header-type heaters allow for more flexibility in plant operation, with higher load changes than U-type heaters. They are more cost-effective both in terms of the investment and operating and maintenance costs than the U-type heaters.
An investment in a header-type heater starts to pay off at an output of about 500 MW. Both header-type and U-type heaters are designed as single-train systems below 500 MW, but header-type heaters, unlike U-type heaters, can also be used with higher power plant outputscurrently up to 1,100 MWin a single-train arrangement, lowering costs.
The operating and maintenance costs of header-type heaters are also much lower above 500 MW because of the lower failure rate and longer service life of header-type heaters. The EPRI study, which investigated the frequency and causes of damage in almost 200 high-pressure heaters in 51 power plants (35 in the U.S. and 16 in Europe), confirmed that European power plants using header-type heaters get much better results.
Authors: Jörg Stahlhut is head of the business division, power plant components, at Balcke-Dürr GmbH, a leading supplier of heat transfer equipment for large power plants owned by U.S.-based SPX Corp. He served as sales and project manager, overseeing several international power plant heat exchanger projects. He holds a master’s degree in mechanical engineering.
Dirk Band is sales manager for feedwater heaters in Balcke-Dürr’s business division, power plant components. A mechanical engineer, he is in charge of feedwater heater projects for conventional and nuclear power plants.