1999 Gasification Technologies Conference San Francisco, California, October 17-20, 1999 Progress Report: Varnamo Biomass Gasification Plant Krister Ståhl Sydkraft AB SE-205 09 Malmö, Sweden Magnus Neergaard Sycon Energikonsult AB SE-205 09 Malmö, Sweden Jorma Nieminen Foster Wheeler Energia Oy P.O. Box 201, FI-78201 Varkaus, Finland SUMMARY Sydkraft AB has built the world's first complete IGCC Power Plant which utilises wood as fuel. The plant is located in Värnamo, Sweden, and the technology used in the power plant is based on gasification in a pressurised circulating fluidised bed gasifier. The gasification technology is developed in co-operation between Sydkraft AB and Foster Wheeler Energy International Inc. The plant at Värnamo produces about 6 MWe electricity to the grid, as well as 9 MWth heat to the district heating system in the city of Värnamo, from a total fuel input equivalent to 18 MW. The Värnamo plant is an important step forward in developing highly efficient and environmentally acceptable technologies based on biomass fuels. Experiences gained in this co-generation plant will be utilised in design of new and larger plants which will generate twice as much electricity as a conventional steam cycle plant with the same amount of heat demand. The start-up phase was completed during spring 1996 and the plant is now available for research and development work. A demonstration programme was launched in 1996, which will continue until June, 2000, and during this period advantages and possible limitations of the new technology are evaluated. Specific areas of interest includes environmental issues, fuel flexibility and production costs in future facilities in addition to the technical development and improvements of the plant. The accumulated operating experience amounts to about 8500 hours of gasification runs and about 3500 hours of operation as a fully integrated plant as per August, 1999. The test runs have so far been very successful and the plant has been operated on different wood fuels as well as on mixes of straw and wood. Tests of refuse derived fuels (RDF) are planned during fall 1999. 1 (16)
INTRODUCTION Increasingly heavy demands are expected on future power plants in terms of efficiency, impact on the environment, fuel flexibility, power production costs etc. Biomass fuels are in most countries a domestic source of fuel and are often found as waste products in different kinds of industries e.g. agriculture, forestry, pulp- and paper. Further, a lot of biomass waste such as e.g. packaging material is presently landfilled and in the future there will probably be more stringent requirements to recycle garbage, thereby reducing landfilling. Carbon dioxide, methane and freons are examples of greenhouse gases that absorb infrared radiation from the earth and contribute to the net increase of radiation energy in the atmosphere, which results in a temperature increase. By utilising biomass fuels as feedstock there will be no net increase in carbon dioxide levels from power production, in contrast to a situation where fossil fuels are used. Integrated gasification combined cycles (IGCC) have been developed and demonstrated for power generation using fossil fuels as feedstock. The main features are the possibility of cleaning the gas from impurities, such as particulates, sulphur, etc. under pressure before the gas enters the combustor of the gas turbine, and also the relatively high electrical efficiency. Higher efficiencies also means relatively lower emissions. On the basis of these considerations, Foster Wheeler Energy International, Inc. and Sydkraft AB have been developing the pressurised IGCC for biomass fuels since 1991. In June 1991, Sydkraft took the final decision to build a co-generation plant at Värnamo, Sweden, to demonstrate the technology. The plant generates 6 MW of electricity and 9 MW of heat for district heating. The Värnamo Demonstration Plant is the first of its kind in the world. The plant is aimed at demonstrating the complete integration of a gasification plant and a combined cycle plant, fuelled by biomass. The basic idea is to demonstrate the technology rather than to run a fully optimised plant. Flexible and conservative solutions were chosen for the plant layout and design, to ensure the success of the project and to make the plant suitable for R & D activities. The plant is now approaching the end of the demonstration phase and has been successfully operated with full integration. An aerial view of the plant is shown below in Figure 1. 2 (16)
THE PROCESS Figure 1: Aerial view of the Värnamo Demonstration Plant The normal wood fuel is dried in a separate fuel preparation plant, using a flue gas dryer, to a moisture content of 5-20 %. A simplified process diagram and a cross section of the gasification plant are shown in (Figure 2 and 3). Figure 2: Process diagram 3 (16)
The dried and crushed wood fuel is pressurised in a lock-hopper system to a level which basically is determined by the pressure ratio of the gas turbine, and is fed by screw feeders into the gasifier a few meters above the bottom. The operating temperature of the gasifier is 950-1000 C and the pressure is approximately 18 bar (g). The gasifier is of a circulating fluidized bed type and consists of the gasifier itself, cyclone and cyclone return leg. The three parts are totally refractory lined. The fuel is dried, gasified and pyrolized immediately on entering the gasifier. The gas transports the bed material and the remaining char towards the cyclone. In the cyclone, most of the solids are separated from the gas and are returned to the bottom of the gasifier through the return leg. The recirculated solids contain some char which is burned in the bottom zone where air is introduced into the gasifier. The combustion maintains the required temperature in the gasifier. After the cyclone, the gas produced flows to a gas cooler and a hot gas filter. The gas cooler is of a fire tube design and cools the gas to a temperature of 350-400 C. The gas enters the candle filter vessel where the particulate clean-up occurs. Ash is discharged from the candle filter and from the bottom of the gasifier, and is cooled before entering the depressurization system. 4 (16)
Figure 3: Cross section of the Värnamo gasification plant The gasifier is of an air-blown type. Thus about 10 % of the air is extracted from the gas turbine compressor, further compressed in a booster compressor, and finally injected into the bottom of the gasifier. The gas generated is burned in the combustion chambers and expands through the gas turbine, generating 4 MW of electricity. The gas turbine is a single-shaft industrial gas turbine. The fuel supply system, fuel injectors and the combustors have been redesigned to suit the low calorific value gas (5 MJ/nm 3 ). The hot flue gas from the gas turbine is ducted to the heat recovery steam generator (HRSG), where the steam generated, along with steam from the gas cooler, is super-heated and is then supplied to a steam turbine (40 bar, 455 C). 5 (16)
The plant is equipped with a flare on the roof of the gasification building, which is used during start-up procedure and when testing less well known conditions, in order to protect the gas turbine. The technical data is summarised in Table 1 below. Power / heat generation Fuel input Fuel 6 MWe /9MW th 18 MWfuel (85%ds) Wood chips Net electrical efficiency (LCV) 32% Total net efficiency (LCV) 83% Gasification pressure / temperature 18 bar (g) / 950 C Lower calorific value of Product Gas 5 MJ/m 3 n Steam pressure / temperature 40 bar (a) / 455 C Plant owner Sydkraft AB Suppliers: Engineering Gasifier Gas cooler Ceramic hot gas filter Metallic hot gas filter Gas turbine Booster compressor Heat recovery steam generator Steam turbine Plant control system Sycon / Foster Wheeler Foster Wheeler Foster Wheeler Schumacher GmbH Mott Corporation Alstom Gas Turbines Ltd. Ingersoll- Rand Foster Wheeler Turbinenfabrik Nadrowski GmbH Honeywell Table 1: Data of the Värnamo Plant 6 (16)
DEMONSTRATION / DEVELOPMENT PROGRAMME An extensive demonstration/development programme is being carried out during 1996-2000. The work, so far, has partly been performed in collaboration between Sydkraft, Foster Wheeler Electricité de France and Elkraft.. It has also been partly financed by Elforsk AB, Sweden, The Swedish National Energy Administration and the European Commission. The overall aim of the demonstration programme is, from a technical and economical point of view, to verify the status and future potential of the biomass IGCC concept, utilising the Bioflow technology. In order to achieve this it is important to identify and verify the status of different parameters e.g. operability, maintainability and availability. Of particular interest to the success of the gasification technology is to verify the quality of the gas produced in the gasifier as well as the operation of the gas turbine. Linked to these main issues are a number of technical evaluations that have to be done; How should the fuel be pressurised in the most cost efficient way? How should the gas be cooled and cleaned in highly realiable way? etc. In parallell to the test runs an extensive work to estimate investment, operational and maintenance cost for future plants, based on experiences from the Värnamo plant is performed. EXPERIENCE GAINED DURING TEST OPERATION General Commissioning of the plant started late in 1992 by start-up of the fuel preparation plant. Commissioning of the combined cycle was completed on liquid fuel during March 1993. The first gasification test on wood chips at low pressure was performed in June 1993, and combustible gas was produced and burned in the flare. It should be remembered that at the time for commissioning of the gasifier, no experience existed from any biomass gasifiers at this pressure level. Accordingly, during tests with different bed materials, temperatures and pressure levels, deposits sometimes occurred. Deposits and fouling have verified the importance of carefully controlling the process as well as ensuring a suitable design of components. For the moment, magnesite (MgO) is used as bed material in the Värnamo gasifier, and this has proved very successful. However, we still believe that it will be useful to continue testing different bed materials or mixtures of bed materials to further optimise the gasification process and achieve the best result i.e. minimum of deposits and best possible gas quality. On the other hand, deposits can also be handled with a suitable design of the gasifier and the down-stream components. Apart from this, already during the early design stage, in particular two areas were of great concern, namely the gas clean-up and the gas quality 7 (16)
Concerning the hot gas filtration one of the ideas behind this is of course to allow gaseous tars to pass through the filter and other tars to stick to the filter cake and not pass into the fine pore structure of the filter itself. As the amount of benzene and tars is not insignificant from gas heating value point of view this is very important to achieve. The diagram below (Figure 4) shows a typical operating curve for the hot gas filters and clearly indicates that no continuous increase in pressure drop is taking place. I GCC V a r n a m o 1 1 0 1 0 0 9 0 8 0 1 2.0 5 1 2.1 0 1 2.1 5 1 2.2 0 1 2.2 5 1 2.3 0 1 2.3 5 1 2. 4 0 1 2.4 5 1 2. 5 0 1 2.5 5 RM A1 0 C P0 0 3, T r yck f a ll ö ver f i lt e r ( m bar ) Figure 4: Pressure drop in a hot gas filter cleaned by nitrogen pulsing Gas quality During the commissioning as well as the demonstration programme the gas quality has been checked regularly. The gas quality have regarding hydrogen content turned out to be slightly lower than predicted, but the heating value has been maintained by an increase in methane. A typical range of dry gas composition is specified below in Table 2. CO H 2 CH 4 CO 2 N 2 16-19 % 9.5-12 % 5.8-7.5 % 14.4-17.5 % 48-52 % Table 2 Percentage in Table 2 is by volume and gas heating value in the range of 5.3-6.3 MJ/ m 3 n have been recorded. 8 (16)
Different operating conditions in the gasifier as well as a change of fuel produce different amounts of light tars and benzene as can be seen in table 3 below. Bark tends to produce less both benzene and tars than ordinary wood chips. Fuel Benzene mg/m 3 n Light tars, mg/m 3 n Bark 60% and forest res. 40% 5000 6300 1500 2200 Pine chips 7000 9000 2500 3700 Table 3 Levels of ammonia and hydrogen cyanide recorded as a function of nitrogen content in the fuel is presented in diagram below. 1600 30 1400 25 1200 20 1000 800 15 HCN (ppm) NH3 (ppm) HCN (ppm) 600 10 400 200 5 0 0 0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 Nitrogen in fuel % Figure 5: Ammonia and hydrogen cyanide in product gas Due to the relatively low combustion temperatures in the gas turbine combustors when burning product gas thermal NO x is very low. Total NO x emissions can however be higher compared to operation on liquid fuel with steam injection due to the conversion of ammonia (and HCN) into NO x. From figure 6 below the influence of the amount of ammonia is evident. 9 (16)
diesel bark wood chips 140,00 120,00 100,00 80,00 60,00 40,00 20,00 0,00 20.00.00 02.00.00 08.00.00 14.00.00 20.00.00 02.00.00 Time CFB01FQ001, NO-concentration in turbine exhaust CFB01FQ002, NO2-concentration in turbine exhaust Figure 6: Gas turbine exhaust emissions The recorded levels of alkalines have been below 0.1 ppm wt. Hot gas filter performance Originally a hot gas filter of the ceramic type was installed. This filter consisted of ceramic filter candles arranged in six groups with separate backpulsing. The ceramic filter showed good filtration efficiency, with stable pressure drop. However, after more than 1200 hours of troublefree operation suddenly two ceramic candles broke. No serious damage was caused but it was noticed in practice the difficulties to detect a relatively small failure in a hot gas filter. The broken candles as well as other candles from the filter were analysed and the reason for the breakdown was never traced by the supplier. The complete set of candles was changed to a new design of ceramic candles and was installed in the plant. After less than 350 operating hours one of the new type of candles broke. The breakdown was established by the supplier to be caused by mechanical fatigue since micro cracking was found in all tested elements and a chemical attack was excluded. To find out possible reasons for the mechanical fatigue, measurements of vibrations inside the pipe (noise) as well as in pipework and steel structure have been performed and been verified to be low. Finally to avoid any risk of fatigue, grids supporting the candles have been 10 (16)
installed. However, since the installation of this grid a tendency of bridging has been observed, which we did not experience earlier. To protect the gas turbine in case of a hot gas filter breakdown a metallic police filter has been installed downstream of the main filter. Rather soon after this installation was completed another failure of the main filter occurred. From figure 7 can be seen that the pressure drop over the police filter increases rapidly, whereas it is hard to spot the broken candle from the pressure drop across the main filter. 500,00 250,00 400,00 200,00 300,00 200,00 100,00 150,00 100,00 50,00 RMA10CP006 [mbar] 0,00 15.15.00 15.25.00 15.35.00 15.45.00 15.55.00 0,00 RMA10CPA04, dp hot-gas filter Time RMA10CP006, dp police filter Figure7: Hot gas filter failure During the summer 1998 it was decided to install metal filter candles instead of the ceramic candles in the main hot gas filter. The metal filter candles are installed in the original filter vessels but with a new tube sheet and backpulsing arrangement. The metal filter has, like the ceramics, shown very good filtration efficiency, with stable pressure drop. This filter has now been in operation for more than 2200 hours without any filter breakage or other damage during normal operation. Gas turbine experience The gas turbine installed in the plant is an almost standard Typhoon from Alstom Gas Turbines in Lincoln, England. Modified components are the combustors, the burners and the addition of an air bleed from the compressor. Furthermore, has a special design gas control module been developed to control the product gas, steam and nitrogen to the unit. 11 (16)
In September 1995, the commissioning of the modified gas turbine on liquid fuel was completed, and the first test runs on product gas were performed in October. In order to minimise the risks, the first test runs were very short and product gas was introduced gradually with a corresponding reduction in liquid fuel, which eventually resulted in operation solely on product gas. Already prior to being supplied to Värnamo, the special combustors and burners were tested in a rig in England utilising synthetic gas. Combustion has always been reliable in the turbine whether operating on gas fuel or liquid. The relatively low heating value of the gas (about 1/10 of natural gas) proved to be of no problem to the gas turbine and a stable flame has always been established even when the heating value has been lower than normal. Not even initially was it necessary to maintain a pilot flame of liquid fuel and thus has all operation for 3500 hours been on 100% gas and no liquid. This is valid within the full operating range from 1/4 to full load. Internal inspection of the turbine and combustors is carried out after every test run and we can now say that the hot gas clean-up is operating most satisfactory and no damage to the delicate parts of the turbine etc has been observed. A thorough combustion of the hydrocarbons has always been registered with results between 1 and 4 ppm, whereas a slightly high figure of CO has been observed with figures up to and sometimes even above 200 ppm. Work is however going on to improve this. As has been mentioned before (Figure 6) levels of NOx around 130 ppm has been recorded when operating on gas produced from biomass with high nitrogen content (like bark) while the lower nitrogen content of hardwood considerably reduces the NOx down to a mere 40 ppm. The development of new combustors will reduce the formation of NOx and further development of selective catalytic oxidation (SCO) of NH 3 and HCN will most likely further reduce the emissions. Fuel flexibility The fuel feeding system with the lock-hopper, pressurised fuel silo and screw feeders is primarily designed for wood chips and fuels with equivalent density/heating value per m3n. During commissioning and the first years of testing, forest residue and wood chips was the fuel used in general. A number of different fuels have however been tested in the plant during the last couple of years. As fuels with considerably lower density are of interest from gasification point of view these low density fuels has to be pelletized due to the installed type of fuel feeding system. Accordingly we have used straw, willow, bark and sawdust pellets. For testing purposes it is also beneficial to be able to mix different types of fuels to a predetermined ratio and this is of course considerably easier with the fuels in pelletised form. The following fuels have been tested up to now: 12 (16)
Wood chips Forest residue (bark, branches etc.) Saw dust and bark pellets Willow (salix) Straw RDF All these fuels have proved to be good and easy to gasify without causing deposits or sinter in the systems. The gas produced when operating on wood chips made from hard wood contains more benzene and tars as shown in table 3 above than forest residue containing bark. Bark has actually proved to be an excellent fuel and even feed rates up to 100% bark is easily gasified and the gas is very good for filtration and gas turbine operation. The rather high levels of alkalines in willow (salix) has not caused any problems in any part of the system and the amount of sintered material in the bottom ash was very small. Tests have been performed with willow in pelletised form and from small amounts up to 100 % willow. The only clear and negative effect observed was a reduction in heating value of the gas. The change was however not so great, and the gas turbine operation could continue. Straw has always been considered a very difficult fuel to burn/gasify due to it s high levels of alkaline and great amount of ash in the fuel. Also the chlorine level is very high in comparison to wood fuels. Tests has so far been carried out with straw and bark pellets in a 50/50 % mix. (Tests with only straw are planned in the near future.) Also this test has been performed without any problems or sintering and a gas was produced with a hydrogen content slightly higher than normal. The tested RDF was in a pelletised form. The size of these pellets are though bigger than the size used for most other fuels (18 mm versus 8 mm). Up to now (August 99) only one test has been performed with RDF and then with about 25 % RDF and 75 % bark pellets. The pellets are produced from waste paper, plastics, cardboard, alumina etc etc. Since the test was carried out lately we have no data available yet, but gas quality as well as filtration was without problems and no sintering or other disturbances occurred. SELECTIVE CATALYTIC OXIDATION AND THE SLIP-STREAM SYSTEM During the summer 1998 a new system, the Slip-stream system, was added to the Värnamo plant. A small flow of hot, uncleaned gas is taken from the gasifier and led to the slip-stream system, where the gas is cooled, cleaned in a hot gas filter and led to an analysis system for continuos, on-line, analysis of the gas composition. The system also includes a SCO-reactor for selective catalytic oxidation of nitrogen compounds in the gas. 13 (16)
The design and installation of the system was made by Sydkraft as a part in an international consortium consisting of VTT Energy (Finland), Åbo Academy University (Finland), ALSTOM Gas Turbines Ltd. (UK) and Foster Wheeler Energia (Finland). The consortium carries out research and development work aiming to substantially reduce the NO X emissions from Biomass IGCC plants originating from fuel-bound nitrogen compounds. The research is co-ordinated by VTT Energy and funded partly by the participants and partly by the European Commission through the Joule programme. The method is based on controlled and selective oxidation of fixed nitrogen species, primarily ammonia and hydrogen cyanide, of the gasification product gas to N 2. The central part of this research is the development of a new SCO (Selective Catalytic Oxidation) technology. The slip-stream system was installed in the Värnamo IGCC-plant in order to make it possible to perform experiments under actual conditions in a gasification plant. The slip-stream system is designed for a gas flow of 0,1 kg/s, which corresponds to abt. 3 % of the total gas flow from the gasifier. The operating pressure is about 18 bar and the temperature in the system can be freely chosen in the range 300 550 o C. The main components in the system are three gas coolers in series, a filter with two filter candles, an ash lock-hopper, the SCO-reactor, the OPSIS gas analysis system and a flow control valve. The main layout of the system is shown in the Figure 8 below, which is the actual screen picture used by the plant operators when operating the system. The data in the picture are typical operation data from the hot commissioning. Figure 8: Slip-stream system 14 (16)
A small amount of the cleaned gas, abt. 2 %, can be led to the SCO-reactor and the OPSIS gas analysis system, but the main gas flow is led to the flare. The flow through the slip-stream system is controlled by a regulating valve at the outlet to the flare. The OPSIS gas analysis system is a newly developed on-line analysis system, which continuously measures a number of gas components using IR and UV technology. The instrument measures the content of CO, CH 4, CO 2, H 2 0, NH 3 and a number of other nitrogen compounds in the gas. The slip-stream system was commissioned in October 1998 and by the end of November the first part of the SCO test programme could be carried out. After adopting the system for high temperature operation by removing two of the three gas coolers and replacing them with uncooled pipes, the remaining part of the test programme was successfully completed in mid- February 1999. No reports regarding the tests of the SCO in the slip-stream system have yet been published, but results are very promising and should significantly reduce the NO x emissions from future plants. Since the commissioning of the slip-stream system it has been in operation for more than 1300 hours (August 1999). The installation of the slip-stream system has increased the versatility of the Värnamo plant in a number of ways and it is now possible to: Operate the slip-stream system at temperatures above the system design temperature of the Värnamo plant which gives information on material and equipment performance at elevated temperatures. Make filtration tests with different types of filter cartridges without endangering the components in the main system downstream of the gasifier. Make full scale gasification tests at high risk conditions without using the main filtration system in the plant. The main gas stream is led to the flare via the by-pass cyclone, while the slip-stream system is used for filtration and gas analysis. REFERENCES 15 (16)
(1) Ståhl, K., Neergaard, M., Experiences from the Biomass Fuelled IGCC Plant at Värnamo, presented at Biomass for Energy and Industry, 10 th European Conference and Technology Exhibition, Wurzburg, Germany, June 1998. (2) Leppälahti J., Ståhl K., Kilpinen P., Hupa M., Cannon M., Nieminen J., Development of selective oxidation technology for the reduction of NOx emission in gasification power plants, poster at Biomass for Energy and Industry, 10th European Conference and Technology Exhibition, Wurzburg, Germany, June 1998. (3) Ståhl K., Neergaard M., Stratton P., Nieminen J., IGCC power plant for biomass utilisation Värnamo, Sweden, presented at Developments in Thermochemical Biomass Conversion, Banff, Canada, May 1996. 16 (16)