|Boiler and superheater corrosion and fouling can cause problems for biomass plants
In recent years the focus on converting coal plants to biomass-fired plants has increased as a part of national green energy plans in Denmark and the rest of the western world. This transition is not without its problems, as biomass fuel contains high levels of corrosive species such as alkali salts. Biomass is also associated with huge buildups of ash deposits, known as fouling, in the boilers and on the superheaters.
Minimizing boiler corrosion and/or fouling problems could therefore save power plants large sums of money in expenses, such as for new superheaters, as well as maintenance costs. Furthermore, it could yield higher boiler efficiency and fewer unintended shutdowns.
In order to find methods to reduce corrosion problems, detailed knowledge about fuel, flow and combustion processes is required. Force Technology has combined CFD modeling with newly developed models for prediction of corrosion risk in biomass-fired plants.
As part of this development, Force Technology supported a case study to predict risk areas for corrosion damages carried out in co-operation with Verdo Renewables and utilising the company’s combined heat and power plant in Randers, Denmark.
The Verdo plant was recently converted from a coal plant to an entirely biomass-fired power station. After the conversion, Verdo experienced a high increase in the corrosion rate of the superheater (SH3), located directly above the grate (see Figure 1). The high corrosion rate resulted in a significantly lower lifespan for the superheater. This corrosion was well-documented through measurements, and provided an excellent reference for validation of the corrosion model.
|Figure 1: Superheater and economizers|
To develop a model that would predict corrosion risk areas in biomass-fired power plants, focus was on the main topics: high-temperature corrosion in biomass ash deposits, chemical reactions in the deposits and flue gas, release of the critical species during biomass combustion, and deposition mechanisms.
This focus resulted in three main models for the simulation: a BED model for simulating the release of corrosive species from the grate, a coarse ash deposition model, and a corrosion model giving the risk of corrosion on the metal surfaces in the boiler.
To find the release of the corrosive species and the combustion of the wood chips on the grate, a BED model provided by Force Technology was used. The model was adapted from a stoker-fed straw furnace, in order to cope with the wood chips distributed using the spreader system. The BED model uses thermal radiation from the furnace onto the grate to calculate the fuel combustion, thereby providing a temperature profile for the grate and the release of flue gas species including the corrosive gases.
A corrosion risk model suitable for CFD modeling needed to be developed. The corrosion model is based on the chemical reactions between highly corrosive alkali salt deposits on the metal surfaces of the boiler, the surface temperature of the metal, and concentrations of oxygen as well as sulfates and alkali salts in the flue gas. Each of these has a propensity to cause corrosion. By evaluating them at the metal surfaces, the model yields a final probability of corrosion between 0 and 1, where 0 is low risk of corrosion and 1 is high risk.
Coarse ash deposition modelling
Areas with high deposition rates are associated with corrosion problems and should be included in the final corrosion assessment. The deposition of ash, and the corrosive alkali salts that follow, were modeled with a new coarse ash deposition model using Lagrangian particles. The model determines whether an ash particle will stick to the surface or rebound based on the particle’s chemical content and temperature, the temperature of the wall, and the impact angle.
In order to simulate the corrosion risk correctly, a detailed CFD model is necessary. This includes simulating the combustion of the fuel and the release of the corrosive species, correct air supply, thermal radiation and heat transfer through boiler walls and superheaters.
The firing method simulated at Verdo combined a spreader system for grate firing of wood chips and suspension firing of miscellaneous biomass. To find the correct fuel distribution on the grate a preliminary CFD analysis was conducted on the wood chips using Lagrangian particles.
Likewise, a preliminary analysis was conducted on the suspension firing to find a correct flame length into the furnace.
To find the correct surface temperature of the SH3 tubes where the corrosion was located, the tubes and the superheated steam inside them were fully geometrically and physically resolved. This was done using the Starccm+ generalized cylinder tool and a conducting baffle interface between the steam and flue gas regions (see Figure 2).
|Figure 2: Steam temperature in one tube slap in SH3|
The rest of the superheaters and economizers were simulated using porous media for subtracting heat and inducing a pressure drop.
CFD analysis of the boiler revealed areas with a high risk of high-temperature corrosion and some low-temperature areas with room for improvements for better boiler performance.
In Figure 3 the mid-plane temperature profile, where both the grate firing and suspension firing are visible as high-temperature areas, is shown. The plot indicates an uneven temperature in the boiler as the main flow is concentrated in a narrow band generated by the suspension firing flame and the secondary air jets from the back wall.
|Figure 3: Temperature profile, boiler mid-plane|
The uneven temperature distribution in the boiler was also present in its transverse direction. This affected the performance of the superheaters, as the thermal load on the centre tube slaps was up to 50 per cent higher than on the tube slaps near the side walls. This information was only available once the tubes of SH3 were fully resolved, giving detailed new information on the performance of a superheater. No such analyses have been done before, according to the literature.
The tendencies of a distinct region with high temperature found in the temperature plots were also found in the concentrations of corrosive species such as potassium chloride, KCl, shown in Figure 4. A high KCl concentration in the flue gas gives a higher corrosion risk.
|Figure 4: Concentration of KCl in mid-plane|
The detailed temperatures of SH3’s tube surfaces allowed the developed corrosion model to be applied, as it is dependent on surface temperatures. This information cannot be derived with the same accuracy by the normal approach of simulating superheaters with blocks of porous media. The corrosion risk profile on the SH3 is shown in Figure 5.
|Figure 5: Elevated corrosion risk in the centre of the boiler|
Figure 5 shows a higher risk profile in the centre of the superheater. This is associated with higher temperature and load. The corrosion measurements conducted on the tubes showed higher corrosion rates in the centre of the superheater, as well as good correlation between the model’s predicted profile and the actual corrosion profile.
To identify fouling areas, the coarse deposition model was applied to the system. A total of 100 000 Lagrangian particles were used to resolved the ash particles in the flue gas, giving results close to the actual fouling picture which Verdo has experienced (see Figures 6 and 7).
|Figure 6: Coarse ash deposition flux on the lower part of SH3|
|Figure 7: Ash buildup on SH3 tubes|
Findings and perspectives
This case study has shown that is possible to combine detailed information from CFD and models for corrosion to predict high-risk corrosion areas. For Verdo’s combined heat and power plant, the elevated corrosion rate at the centre of the superheater was also predicted by the CFD analysis. The cause of the high corrosion rate in the centre was found to be an uneven load on the superheater. A solution strategy for better utilization of the boiler was offered to Verdo, including different settings for their secondary air nozzles.
CFD is a strong tool for further investigation of possible improvements in geometrical alterations, different fuel utilization, optimization of combustion air distribution (for example, in secondary air nozzle penetration) and superheater placement, all of which affect the corrosion profile in the boiler.
Once the main simulation has been set up, optimization changes can be easily and quickly tested, saving expensive trial-and-error experience in full-scale tests in the field. The optimizations would result in better utilization of the boiler for better performance and fewer downtime hours for repair, resulting in an overall better economy for the power plants.
Svend Skovgaard Petersen is project engineer for thermal energy & fluid mechanics at Force Technology.
Power Engineering International Archives
View Power Generation Articles on PennEnergy.com