Fireside corrosion in coal-fired utility boilers (CORROSION)

This result concerns the outcome of short-term corrosion tests in a wood-fired circulating fluidised bed (CFB) boiler. The experiments were carried out in connection with the use of fuel additives for corrosion inhibition. The additives selected were kaolin, sulphur and an experimental additive. The chemical composition of the flue gases was investigated at two corrosion sites located at the final superheater in the convective pass and in a novel position in the cyclone loop seal. The flue gases were analysed by means of several measurement techniques, including Fourier Transformed Infra-Red (FTIR) spectroscopy and In situ Alkali Chloride Monitoring (IACM). The latter is a novel and unique technique for measuring alkali chlorides levels in a flue gas. Deposits were collected on deposit probes analysed by means of SEM/EDX. Sulphur and a sulphur containing additive, called 'ChlorOut', were found to be effective in reducing the amount of alkali chloride in the flue gas and in superheater deposits. This is expected to lead to reduced corrosion. However, the use of sulphur resulted in an increased amount of SO(2) in the flue gas and a lower pH value in the flue gas condensate. These effects were only marginal during the use of the experimental additive. The use of the experimental additive is therefore preferred.

During this project CESI was devoted to the development and qualification of a monitoring device able to measure the corrosion behaviour of the boiler materials on-line. CESI has contributed to this task with the design, construction and testing of a corrosion probe based on the electric measure of the real thickness of the tube specimen under testing. The description of the probe characteristics, construction and calibration is contained in the report of these results. In this document the experimental campaigns performed in the Ansaldo Caldaie boiler are described together with the results obtained and the conclusions that can be drawn. The results obtained in this project have contributed to increase the know-how about the material corrosion resistance and the attack mechanisms for the materials/coatings selection. The results have also allowed the development of a reliable on-line corrosion monitoring probe to be used in order to improve the plant management. The useful information for the material constructors consists of a good selection of high temperature materials or coatings under the designed operating conditions. The support for the plant operating personnel consists of predicting the effect of a combustion mode (sub-stoichiometric combustion with low NOx technologies) and of operating conditions (biomass or waste cofiring, low cost fuels burning, higher operating temperatures) on the waterwall and superheaters/reheaters tubes. The value for the customers of this up-to-date know-how can be measured in terms of life increase of the plant components and increase in plant availability, reliability and efficiency.

Two co-firing combustion test series were carried out in a 0,5MW combustion rig: The first one is: - Co-firing lignite with high ash paper sludge and compost and - Co-firing bituminous coal with olive husk. The possibility of fireside corrosion by co-firing was predicted for the above fuels. This was done through the investigation of interactions of corrosion related minerals from different fuel sources. In the case of co-firing lignite with fuels containing high ash and corrosive elements, the high content of inert ash in such fuel can dilute corrosive elements. As a result, the deposit ash is not very corrosive. In this case, not only the corrosive element in fuel, but also their concentration in fuel ash should be considered to judge the overall corrosivity of a fuel. Low sulphur level in the fuel ash could increase Cl concentration in the deposit ash, which is not desired. The characteristics of co-firing Ca, Fe, K rich biomass i.e. olive husk, with bituminous coal can be generalized as following: - Ca from biomass can bind S in coal and forms fine particle CaSO(4), which trends to deposit on the heat surface. - Ca and Fe in fuel ash are effective to capture K to form low-melting temperature ash particle, which could be entrapped on the probe surface or deposit on hot surface as mass deposit ash, so that not especially high K concentration is found in the cooled probe surface. - Co-firing of straw, wood and other biomasses especially in bituminous coal fired boilers with a thermal share up to 15% seams technically feasible. Higher shares have to be individually proved.

Corrosion mechanism and the influence of gas atmosphere and ash deposits on fire-side corrosion rate were investigated. Thin plates of more than 10 different metal specimens with chromium content between 1 and 24% were exposed in oxidising and reducing atmosphere containing HCl or SO(2) for max. 300 hours from 400-700°C. Ash deposits from firing of straw, wood and coal were applied on the metal specimens. The results obtained include: In oxidising atmosphere - Ash deposit containing alkali chloride, such as straw ash, is more corrosive than ash deposit without alkali chlorides such as wood ash and coal ash. - Under ash deposit containing alkali chlorides, low and middle Cr steels have considerably higher corrosion rate than stainless steel, however, the low Cr austenitic steel Esshete 1250 has no better corrosion resistance than low and middle Cr steels. - Stainless steels suffer more serious internal corrosion than 9-12% Cr steels. - The maximum metal loss is at 650°C, greater than that at 550°C and at 700°C. SO(2) accelerate chlorine induced corrosive. - NF 709 shows highest corrosion resistance in HCL and H(2)O vapour atmosphere Gas environment depleted of oxygen. - Alkali chlorides in ash deposits are minor activated. Sulphates are more corrosive than chlorides. - In N(2) atmosphere synthetic alkali chlorides are inactive in comparison to real ashes (straw, wood). - Chlorine induced corrosion shows no clear time depended corrosion rate. In general deposits formed at real combustion trials are much more reliable as predicting tool to identify corrosion mechanisms and rates under synthetic gas and temperature atmospheres.

The suitability of co-firing bituminous coal or lignite with straw is predicted, and furthermore, up to which thermal share of straw in the co-combustion is acceptable against serious fireside corrosion. Alumino-silca rich coals such as bituminous coals are more suitable to be co-fired with straw than calcium and magnesium rich coals such as Rhenish brown coal. More share of straw (till 25% thermal share) is allowed to be mixed with bituminous coal. This is based on the low formation potential of gaseous and solid alkali chlorides in flue gas by bituminous coal firing. The high concentration of Al and Si minerals in bituminous coal can capture the potassium from straw very effectively, as consequent, chlorine is present as HCl instead of gaseous potassium chloride. Rhenish brown coal contains very little Si and Al, therefore, potassium is present as potassium chloride and potassium sulphate in flue gas. This is not desired since both of them can deposit on superheater or waterwall tube surface and result fireside corrosion.

The aim of the CORROSION project was to develop an on-line electrochemical fireside corrosion monitor and demonstrate it at the Hemweg 8 coal fired power plant. The on-line electrochemical corrosion probes was successfully demonstrated in the boiler of the Hemweg 8 power plant of Reliant Energy between 26 October and 1 December 2000. During the measurements a porous mass formed on a regular basis. The porous mass was most probably formed due to the high temperate on the ceramic heat shields surrounding the probe face. Fortunately the obstruction could easily be removed from the probe face. A minimum probe temperature of 500°C is advised for linear polarisation resistance (LPR).

The objective of Mitsui Babcock was the development and site demonstration of a corrosion probe-sniffer port assembly for the investigation of furnace wall corrosion in large coal-fired boilers. The Mitsui Babcock corrosion probe is a simple and inexpensive device, which is designed for insertion through the membrane wall of large utility boiler furnaces. The basic preparations for the probe installation are made during a short boiler outage but thereafter the probe tips can be inserted or replaced at any time. The probe tip contains an internal thermocouple, and the temperature of the tip is normally recorded continuously throughout the period of exposure. The local furnace gas composition is sampled for chemical analysis through the sniffer ports that are situated local to the corrosion probes. This is normally carried out periodically. The design of the corrosion probe-sniffer port assembly has been finalised, and the suitability of the probe for utility boiler application has been demonstrated during seven probing exercises of around 5,000-8,000 hours in duration, in two large pulverised coal-fired boilers in two different British power stations. The procedure for measuring the probes prior to and following exposure has also been finalised. Overall, the technique permits the measurement of the metal wastage rates with an accuracy better than +/- 0.01mm (10 microns). The corrosion probe metal losses, the probe tip temperature histories and the furnace gas analysis data for seven probe exposure exercises are described in the report, viz: - Ratcliffe 1, during April to December 2001, - Ratcliffe 2, during December 2001 to August 2002 - Ratcliffe 3, during August 2002 to July 2003 - Drax 1 and 2, during December 2001 to August 2002 and - Drax 3 and 4, during August 2002 to July 2003. The measured metal thickness loss values for the probes, during the seven probe exposure exercises, varied between 0.023mm and 0.570mm, depending on the conditions. The probe tip temperature histories were complex and reflected both the boiler load patterns, and the extent of ash coverage of the probes, in a way analogous to those of the furnace wall tubes. The furnace gas analysis data also showed wide variability, however the furnace wall conditions for all seven probing exercises were predominantly oxidising in nature. There was no significant evidence of flame impingement or of severely reducing conditions at any of the probing locations. A very good correlation (R2=0.85) was found between the measured metal losses from the probe tips, from all of the probing exercises, and the cumulative time that the probe tips were exposed to temperatures in excess of 500?C. This result indicates that the rates of metal loss from the probes were controlled, to a first approximation, by the probe tip temperatures. Other factors, such as the furnace/combustion conditions, the range of fuels fired and the operating regimes of the boilers were clearly secondary effects. This correlation is very encouraging and provides a useful baseline, reflecting the corrosion rates corresponding to predominantly oxidising conditions at the walls, against which the results of future corrosion probing data can be compared. The majority of the work was carried out using carbon steel probe tips, since this is the most common material employed for furnace wall tubes. Some test data are available for probe tips manufactured from HCM2S, a higher grade steel. These probes were all exposed to maximum temperatures in the range 360-450?C. All of the measured metal losses were relatively low, as would have been expected for this grade of material under the conditions of the probe exposure tests. The first corrosion probe campaign at Ratcliffe indicated that the design needed revising as the exposure temperatures of the probes were mostly above 450?C and consequently the metal loss and rate of metal loss values were high. The majority of revised design probes, which were exposed in the following tests all operated at temperatures of 450 ?C and below, and experienced low metal loss values of 0.028 -0.171mm, which correspond to average rates of metal loss of 4 - 23nm h-1. This is as would be expected for functional corrosion probes operating at these metal temperatures in mainly oxidising conditions.

Coal and biomass (olive husk and meat & bone meal) have been burnt in 'co-firing mode'. A burner able to operate with coal alone and with different percentages of biomass replacing coal heat input was set-up and experimented. The most important combustion parameters were investigated and optimised for achieving the best combustion efficiency and pollutant control.

The KEMCOP corrosion and fouling monitoring system (patent pending) consists of small probes, which are installed, in the strips between the tubes of the membrane walls. A total number of 64 KEMCOPs were installed at the Maasvlakte powerplant in the Netherlands in order to determine the influence of co-firing conditions on the performance of several materials. At first glance a minor amount of corrosion probes are showing significant (>0.22mm/yr) corrosion. The majority of probes are well within the range of normal corrosion rates (<0.22mm/yr). Based on these results the influence of co-firing a mixture of secondary fuels on a large scale on corrosion seems to be negligible. From the corrosion rate distinctions can be made in corrosiveness of different locations in the boiler, whereas the highest location in the superheater is generally showing the highest corrosion rates. The material testing data is showing to be very site specific. However, it showed to be very valuable information for the E.ON Maasvlakte power plant. No profound relationship was found between constituents in the corrosion layers, the corrosion rates and the location in the boiler or superheater. Probably longer exposure periods are required to measure higher concentrations in order to make a clear distinction between the processes taking place.

The formation of ash, vapour phase species causes various problems to the boiler utility ranging from ash deposition and corrosion in the furnace walls/boiler tubes to unplanned shut down. The work presented here focused on developing a mathematical model to predict surface fouling and corrosion due to ash and various vapour phase/gas phase chemical species generated during combustion. Formation of ash and vapour/gas phase chemical species during coal combustion is a complex phycial-chemical process. These products of combustion cause various problems like slagging, fouling and corrosion, which are the most severe causes of unplanned plant shutdowns. Predicting slagging, fouling type deposit formation and corrosion is very important. In this study an effort has been made to predict the amount of vapour/gas phase chemical species release, by predicting NaOH, NaO(2) and NaCl, followed by ash formation prediction and an ash deposition model for slagging and fouling type deposits. The alkali gas phases species release model and an ash deposition model have been validated against experimental results (40MW ANSALDO boiler). Predicted results show good agreement with experimental measurements. Further refinement is needed to include chemical interaction on the tube metal surfaces leading to corrosion. The developed ash deposition and corrosion model is relevant to all height temperature solid fuel fired combustors. The model will be able to predict in advance the consequences of high alkali content fuel blends utilization on the plant heat transfer surfaces. Its abilities will be disseminated through conference presentations, workshops and combustions courses, where CINAR has been involved for the last ten years.

The report concerns a detailed and critical examination of the corrosion trial work conducted in laboratory, small-scale combustion test rig and industrial scale facilities. It attempts to develop correlations between the results of the work performed at different scales, and with relevant information available in the technical literature. During the laboratory and combustion test rig experiments, reasonably close control over the experimental conditions is possible, however a number of simplifications and compromises were made, and there is a danger that the experiments do not provide an acceptable simulation of plant conditions. Conversely, it is not possible in the context of boiler corrosion trials to exercise as much control over the experimental conditions as may be desired, however the flue gas conditions, gas temperatures and heat fluxes are, as far as is possible, representative of those in operating boiler plants. The team tried to resolve these issues and to reach a technical consensus on the corrosion mechanisms, which apply in three situations: - The operation of superheater and reheater elements at advanced steam conditions. - The exposure of furnace wall tubes to the conditions associated with advanced primary NOx reduction technologies, i.e. highly reducing local gas atmospheres and ash deposits containing reduced sulphur species and perhaps carbonaceous material. - The effect of the combustion of biomass and waste materials either alone or co-fired with coal, i.e. the formation of fireside deposits rich in K and Cl compounds.

Using all of the information and experiences gained from the CORROSION-project, a design and operational guideline document was produced. The report details the current state of knowledge of fireside corrosion, and covers the following subject areas: - Advice on the impact of fuel characteristics and measures on boiler tube corrosion, with particular reference to biomass and waste materials. - Recommendations on good practices in the design and operation of advanced primary NOx reduction technologies, to minimise risks of accelerated furnace wall wastage. - Estimates of the corrosion rates of conventional and novel, boiler tube materials under the relevant conditions of metal temperature and ash chemistry. - Recommendations on the location and design of superheater and reheater elements to minimise high temperature corrosion rates. - Advice on the impact of the design and operation of on-line cleaning equipment on boiler tube metal wastage, and on good design and operational practice. - Recommendations on good engineering practices in the use of materials solutions to corrosion problems, i.e. the use of shields and bandages, weld overlays and coatings, and co-extruded tubing in sensitive areas. - Guidelines on the availability and use of on-line corrosion monitoring and diagnostic techniques, and on the interpretation of diagnostic data. - Guidelines on the use of the predictive model of ash deposition/corrosion developed in the course of the project. The document provides authoritative and quantitative information and guidance on the interpretation and use of this information by boiler designers and operators.

This result involved the completion of a series of laboratory-scale experiments that were concerned with the measurement of corrosion rates of a number of boiler tube materials, both conventional and novel, under strictly controlled conditions. The laboratory-scale corrosion experimental procedure was developed by Mitsui Babcock to permit the estimation of metal wastage rates to a precision that is relevant to industrial applications, i.e. to an accuracy of +/-10nm h-1, over a test duration of 1,000 hours. The following six materials were selected for the study: - HCM 2S 2 and a quarter % Cr boiler tube material. - T 91 9% Cr (modified), superheater and reheater tube material. - HCM 12 A 12% Cr superheater and reheater tube material. - Esshete 1250 15% Cr, austenitic superheater material. - NF709 high strength, high temperature austenitic materials. - 40:20 Ni base development material for use in ultra super critical plant. These materials were exposed to 6 different surface deposits for a period of 1,000 hours, at temperatures up to 800 degrees C. The surface chemical treatments applied were the following: - No treatment, i.e. baseline surface oxidation test, - With a loading of Na(2)SO(4), - With a loading of K(2)SO(4), - With a loading of NaCl, - With a loading of KCl, - With a loading of synthetic ash mixture, containing Na(2)SO(4)/K(2)SO(4)/Fe(2)SO(4) on a 1.5 mole/1.5mole/1 mole basis. Full metal wastage-temperature curves were determined for each combination of material and chemical treatment, and in each case, the temperature at which the measured metal wastage rate reached 25nm h-1 was determined. This specific temperature was used as a point of comparison between the materials under the different chemical treatments, and was known as the threshold temperature. The threshold value is equivalent to the temperature at which a material reaches the maximum acceptable rate of corrosion for boiler tube materials.

This sensor measures the thickness of a metal foil that is attacked by corrosion as a function of time by determining the heat transfer in the foil. This is realised by applying a heat pulse generated with optical means and by measuring the diffusion of this pulse applying pyrometry. The foil is be located at the top of a KEMCOP and gives information about the corrosion rate at the location of the KEMCOP, which is (like the foil) made of the same local material. A mathematical model of the heat transfer in the foil attached to a heat sink has been applied to calculate the pyrometry signals. An experimental arrangement has been realised consisting of the optical systems and of a metal foil in a calibrated furnace to verify the results of the model. The results of the experiments show, that: - Contactless optical determination of thickness is possible with pigtailed laser and fiber pyrometry at an accuracy level less than 1 % in a measuring time less than 5 minutes. - A KEMCOP with a thin area (order 102im) is required. This thickness determines the accuracy for corrosion rate determination (the thinner the more accurate) and the life time of the sensor (the thinner the shorter life time). Optimisation for this is needed. - Commercially higher diode laser power and better detector sensitivity than applied are available resulting in a signal improvement of a factor 100, but also in higher costs. - Emissivity measurements as function of time at high temperatures are required for applied KEMCOP materials for signal calibration. With this emissivity knowledge the sensor can also be applied as thermometer.

Corrosion rates have been measured for four austenitic steels of advanced characteristics and different Cr content (from 12,5 to 22%) operated at full scale and actual plant conditions in the 40MWth experimental boiler of Ansaldo Caldaies test facility in Gioia del Colle. The tube skin temperature was maintained at 650°C by adjustment of the high temperature steam (600°C) flowrate in the test section tube panel (last stage of boiler superheating pass). Different fuel combinations (co-firing of coal and biomass) were investigated in three successive experimental campaigns and for each fuel combination a complete set of corrosion rates was measured.

A corrosion probe positioned in the cyclone loop seal and heated by sand instead of flue gases showed lower metal loss values from erosion or corrosion than a probe in the convective pass (usual position of superheater). Some austenitic stainless steels containing 15-18% chromium showed very good corrosion resistance.

The aim of theses corrosion experiments was to verify the behaviour of materials for thermo-electric plants operated at higher temperatures and in low-NOx combustion mode. The following activities were carried out: - Laboratory experiments aiming at obtaining corrosion data for the development of predictive corrosion models, and providing improved understanding of attack mechanisms over reducing and oxidizing gas side conditions. - Development of a fire-side Corrosion Monitoring Probe for control of boiler superheater tubes. The results obtained have contributed in increasing the know-how about the material corrosion resistance and the attack mechanisms for the materials/coatings selection. The results have also allowed the development of a reliable on-line corrosion monitoring probe to be used in order to improve the plant management. The useful information for the material constructors consists of a good selection of high temperature materials or coatings under the designed operating conditions. The support for the plant operating personnel consists of predicting the effect of a combustion mode (sub-stoichiometric combustion with low NOx technologies) and of operating conditions (biomass or waste cofiring, low cost fuels burning, higher operating temperatures) on the waterwall and superheaters/reheaters tubes. The value for the customers of this up-to-date know-how can be measured in terms of life increase of the plant components and increase in plant availability, reliability and efficiency.

Deliverables

Share this page

Download