Because of the need to control the stability of the electricity grid, the intermittent nature of solar and wind energy poses completely new challenges for the dynamics and mode of operation of solid fuel fired power plants. In the future, such plants will be expected to follow load changes faster and operate with a wider load range. There will also be the need to store energy at the plant in different ways.
The key task in the use of bio fuels and waste-derived fuels is to manage the possible risks related to corrosion and fouling of the boiler. The higher steam values also ‘pressurise’ the development of boiler material technology. In coal combustion, the goal is to decrease the emissions of carbon dioxide in electricity production through increased efficiency, achieved by supercritical steam values. The use of more difficult and often local fuels is increasing worldwide, while the future capability of co-combust coal, biomass and gases, such as natural gas, is evident. Investigation has also begun into different ways of combining solar energy and the traditional power plant.
Tightening of emissions limits has necessitated the retrofitting of many old boilers. Investment is needed in the near future, both in boiler technology and in the cleaning of flue gases. The combined production of heat and electricity (CHP), well known in Finland, is also in a state of flux. As the aforementioned solar energy and various air and earth pumps become more common, the demand for district heating will decrease, with the resulting drop in incomes becoming a consideration for plant owners.
Figure 2. CFD modelling results for a 175 MW (fuel power) BFB boiler.
CFD modelling in the design and process development of boiler furnaces
Computational Fluid Dynamics (CFD) in the design and process development of new boiler furnaces and retrofits of existing boilers is a key tool for decreasing emissions and enabling reliable operation with varying fuel mixtures. CFD modelling is capable to reveal the relevant phenomena occurring inside a boiler furnace such as flow field and residence times of gases and of fuel and ash particles, progress of combustion and pollutant reactions, the distributions of temperature and of concentrations of chemical species, and the heat transfer distribution on heat transfer surfaces. Combining CFD results with analysis of the conditions of the boiler material components makes it possible to estimate the availability of the boiler furnace with respect to e.g. slagging, fouling and corrosion.
VTT has 25 years of experience in the development and application of CFD to the study of furnace processes, including pulverised fuel, oil and gas fired furnaces, bubbling (BFB) and circulating (CFB) fluidised beds, grate fired furnaces and recovery boilers. In most studies the major focus has been on the decrease of emissions while maintaining sufficient burn out level and good furnace availability, notably the reduction of nitrogen oxides (NOx) through in-furnace methods such as proper air staging and fuel injection, advanced low-NOx burners or selective non-catalytic reduction (SNCR) technology.
The air staging principle for bubbling fluidised bed boilers (BFB), developed in cooperation between Fortum Oy and VTT, is presented as an example of the application of CFD modelling (ref: VGB Powertech 11/2013, 75-79).
A fluidised sand bed stands at the bottom of a BFB furnace. Fuel is fed into the furnace from fuel chutes located somewhat above it. Large-size fuel particles fall on to the bed, burning there, while small particles may be combusted during flight higher up in the gas space, i.e. in the freeboard area. Bed temperature should be kept between certain limits, around 800–850 °C, partly to ensure complete combustion but also to avoid bed agglomeration and sintering.
The NOx emitted originates mainly from fuel bound nitrogen due to relatively low combustion temperature also in the freeboard area . The reduction of nitrogen emissions in furnaces is based on the staging of the combustion air; this is because in oxygen-rich conditions the fuel nitrogen is easily oxidised to NO, whereas in oxygen-lean conditions the previously formed NO is reduced to molecular nitrogen.
The general idea of air staging in a BFB boiler is presented in Figure 1. Combustion air is typically introduced in three stages. Primary (1’ry) air and in many cases some amount of recirculated flue gas (FGR) is fed into the furnace from below through the bed, and used to fluidise the bed into the bubbling mode. The secondary (2’ry) air level is usually located above the fuel chutes, and the tertiary (3’ry) air level in the upper furnace.
In the developed air staging principle, the furnace can be divided into three combustion zones: zone I below the secondary air elevation, zone II between secondary and tertiary air levels, and zone III above the tertiary air elevation. Proper choice of the air staging levels, and of the division of the amount of air introduced between the primary, secondary and tertiary levels, can minimise the formation of nitrogen emissions and simultaneously maintain good furnace availability.
Development of the air staging principle was based on the results obtained from the simulation of several alternative firing cases. As an example, simulation results are presented for cofiring of forest-based biomass and peat in a BFB boiler with the capacity of 175 MW (fuel power), and with two tertiary air levels in addition to a secondary air level. The CFD modelling concept, and especially the NOx sub model, was roughly validated against available plant output data. Figure 2 shows the simulation results. The predicted NOx emission compares well with the measured average value and is inside the 10% fluctuation from the average value detected during operation.
To investigate different air staging alternatives for this reference case, several cases were conducted varying the level of secondary air injection and distribution of air among the primary, secondary and tertiary air levels. Forest-based biomass, peat, or the combination of biomass and peat in the energy ratio 30%/70% were considered for combustion. The total air ratio was kept at the constant value 1.2.
Figure 3 presents some simulation results showing the influence of the air ratio of zone I on the simulated furnace exit gas temperature (FEGT), the furnace exit concentration of CO, and NOx emission.
According to the results, by increasing the amount of zone I air it is possible in the combustion of biomass and the co-combustion of biomass and peat to simultaneously decrease the NOx emission and the furnace exit CO concentration, and to lower the furnace exit gas temperature and, correspondingly, increase the heat transfer rate from the furnace.
The predicted furnace exit CO level becomes much higher in peat combustion, and even increases along with the zone I stoichiometric ratio. This is due to small particle size leading to enhanced heterogeneous combustion in the freeboard and shorter residence time for CO burnout. This issue could be tackled by improving mixing at the secondary and tertiary air levels, as shown in Figure 4.
Through the CFD simulation of various alternatives the following recommendations were obtained:
• It is beneficial to set the 2’ry air feed at a lower vertical position when operating the furnace without additional air introduction because of a particular need for bed temperature control in case of e.g. very dry fuel or heavy volumetric loading of the furnace. This, however, leads to somewhat compromised NOx performance. NOx reduction largely takes place in the combustion zone II between 2’ry and 3’ry air elevations.
• With additional air introduction into the low furnace the combustion zone I can be more efficiently used for NOx reduction together with zone II. In this case, from the NOx point of view it is beneficial to set 2’ry air elevation at higher vertical position.
• Minimum NOx emission is achieved by using additional air introduction to optimise the zone I stoichiometric ratio combined with a higher position for 2’ry air feed. In addition, furnace exit CO and furnace exit gas temperature can be reduced simultaneously, also lessening the tendency for upper furnace fouling and corrosion. The optimal zone I stoichiometric ratio depends on fuel type.
• Burnout and furnace heat transfer can be enhanced, and upper furnace temperature distribution smoothed, by paying attention to 2’ry and 3’ry air system design to improve mixing conditions. According to the CFD results no drawback in NOx is expected.
Observations and feedback from practical applications of the air staging principles support the modelling results.
Figure 3. Effect of zone I stoichiometry with additional air feed on simulated furnace exit gas temperature (FEGT), on furnace exit CO concentration, and on NOx emission. Combustion of forest-based biomass or peat or a mixture of these in a BFB boiler. Stoichiometry of zone I: SR1-A < SR1-B < SR1-C.
Dynamic process simulation in power plant design
Dynamic process simulation is a computational method for mimicking process plant behaviour in various transient situations. Modelling typically includes a clearly larger process area than in the case of CFD, thus being also coarser in accuracy. Typically, phenomena are described in one dimension, for example flow in a pipeline is discretised in the axial direction only. Modelling creates a virtual plant that is operated in a similar way to the actual plant, thus requiring the inclusion of the relevant parts of automation and electric systems. In practice, this means at least the main control loops; sequences and interlockings are also modelled when the targeted transients demand. Electric system modelling typically covers the generator and the connection to the grid, as well as the power supply of process equipment if power failure studies are of interest.
Typical applications for dynamic simulation are evaluation of process and control design, automation testing, training simulators, and different types of analysis. Training simulators are the largest of the applications, and the project may include testing of the automation application before commissioning of the plant. Another use for automation testing relates to automation system renewal projects in nuclear power plants, where old analogue systems are updated to digital. Use for the purpose of engineering and analysis is common in both industry and research institutes, and among nuclear safety authorities.
Apros is commercial simulation software (www.apros.fi), which has been jointly developed for almost 30 years by VTT and Fortum. It is used for dynamic studies of conventional and nuclear power plant processes, as well as other industrial processes in almost 30 countries worldwide. A recently published Simantics-based (www.simantics.org) new Apros user environment facilitates fast and intuitive model-building and improves connectivity with other engineering tools. Dozens of reported successful applications have proven that a reliable system model can be built based solely on design information. This enables evaluation of the design, and observation of potential problems before the engineering project proceeds to implementation. The modelling resolution is selected depending on the project needs.
Dynamic process simulation can be applied in both process and control engineering, but the highlight is its ability to enable simultaneous evaluation of the results of these disciplines. This is of utmost importance in the case of novel process concepts where a new process design is automated for the first time. In commissioning and operation the process and automation must co-operate seamlessly. Typical critical design topics are dimensioning of control devices, limit switch values, size of buffers, timing of sequence steps, interlockings, synchronisation of operational mode changes, the need for check valves and tuning of control loops. The more complicated the system, the more difficult it is to discover different functional chains and estimate their meaning for safety, operability and economy. Relevant operations can be studied by simulation, and repeated with different design alternatives, which enables comparison and eases decision-making.
The rise of solar and wind power has increased the demand for rapid start-ups and shut-downs among traditional power plants. This calls for plant flexibility and automation, but also deliberated practices for operating plants in a safe and equipment-friendly way. The heat stresses significantly influence the expected life time of the system. Apros has recently received tools for assessing the heat stresses and life time of the equipment during the simulated transients.
Figure 4. Effect of 3’ry air injection on mixing and combustion of CO in the end part of a BFB boiler furnace.
Integration of different process systems introduces complexity in system operability. For example, a power plant capable of carbon dioxide (CO2) capture and storage (CCS) includes oxygen production and CO2 processing units whose operation must be synchronised with the boiler island. On the other hand, the use of pure oxygen introduces new safety issues in various failure or malfunction situations. This kind of process featuring clearly different operating modes, and requiring smooth operation between the modes, challenges the traditional design methods. Dynamic process simulation offers a tool for testing different approaches virtually. Figure 5 presents a sample of simulation results in a project using Apros to evaluate Foster Wheeler Energia Oy’s 300MWe CCS-capable power plant concept Flexi-Burn® CFB1) in transient operations. The plant is initially operated in the air firing mode as in conventional power plants. The firing mode is then switched to oxy firing enabling CO2 capture. Air as an oxidant is replaced by a mixture of recirculated flue gas (RFG) and oxygen (GOX) from the cryogenic distillation plant. The left picture presents the major gas flows; the right picture the composition of the flue gas. Water steam side, with supercritical steam conditions, and the turbine island were also included in the modelling scope.
Despite the recognised benefits, dynamic simulation is not yet routinely used in plant engineering projects, at least outside the nuclear power industry. One reason is the general incompatibility of simulation and other engineering tools. VTT is committed to develop co-use of simulation and engineering software tools. Software interfaces have been implemented between Apros and, for example, Siemens’ COMOS, and Intergraph’s SmartPlant. Examples of co-use include model data exchange between a P&ID tool and the simulator, automatic generation of models, and document management. The developed interfaces make it easier to join dynamic simulation as a natural part of the engineering workflow.
Figure 5. Apros simulation of firing mode switch from air to oxy mode with the CCS-capable Flexi-Burn CFB power plant concept.
Figure 6. CFD model of a mixing tank coupled with an Apros model of a pipeline system.
Coupled simulations using Apros and CFD codes
Co-simulation with Apros and CFD code enables detailed three-dimensional modelling of one process component that is coupled with a complicated system of pipelines and other process components. Two different types of coupling can be readily identified. In one-way coupling, Apros simulation provides boundary conditions for the CFD calculation, but no feedback from the CFD calculation to the Apros simulation occurs. In two-way coupling, the CFD code also provides boundary condition for the Apros simulation.
In Apros 6, two-way coupling of Apros with ANSYS Fluent CFD code has been implemented at VTT in cooperation with Fortum. The coupling of the codes is illustrated in Figure 6, where a model of a mixing vessel is coupled with an Apros model. Fluids are flowing into the mixing vessel via two inlets, and the mixture flows out from the vessel via two outlets. The codes are coupled at the inlets and outlets, where they exchange information on the flow rates, temperatures and species components of the mixture.
In Apros 6, a user interface for defining the coupling is available. Apros parses the Fluent case file and shows the inlet and outlet boundaries where coupling is possible. The chosen inlets and outlets are then connected to the special coupling nodes of the Apros model and the type of the coupling is defined. The codes run in parallel and exchange coupling information at each time step by using message-passing libraries.
Coupling one-dimensional code with three-dimensional code poses several challenges. When fluid flows from the Apros model to the CFD model, Apros only provides average flow velocity to the CFD code at the coupling interface. Suitable velocity and turbulence values need to be generated at every point on the coupling interface. Numerical stability of the co-simulation is also an issue. It is necessary that the codes exchange information several times within each time step. Semi-implicit coupling of the codes has been found to be a suitable method.
Co-simulation with Apros and CFD code has potential applications in the analysis of furnaces and boilers. Other types of coupling of the codes would also be useful in these applications. Coupling on heat transfer surfaces or in heat exchangers, e.g., CFD simulation of the furnace processes coupled to Apros simulation of the water/steam side would be important especially in load changing situations, and valuable for the design and optimisation of plant operation.
1 ) Flexi-Burn® is a trademark of Foster Wheeler AG, registered in the U.S., EU, Finland