Complex terrain wind field numerical simulation method considering turbulence anisotropy

Abstract

The invention discloses a complex terrain wind field numerical simulation method considering turbulence anisotropy, and the method comprises the steps: considering the turbulence anisotropy of an inflow boundary, and determining the condition type and parameters of the inflow boundary; acquiring ground roughness information of the complex terrain, and correcting a wall surface function; obtaining high-precision topographic data, determining a computational domain range, and generating a computational domain grid; on the basis of an inflow boundary turbulence retentivity principle and a leeward slope flow separation characteristic, a turbulence model closing coefficient is corrected; and the corrected linear pressure strain RSM model is adopted to simulate a complex terrain wind field value, and complex terrain wind field flow field information is obtained. According to the method, the inlet boundary and leeward slope turbulence anisotropy are considered, the wall function is corrected, meanwhile, the complex terrain modeling and grid generation mode is simplified, and the simulation precision is effectively improved.

Description

Technical Field

[0001] The invention relates to the technical field of wind power generation, and in particular to a numerical simulation method for a wind field in a complex terrain taking into account turbulence anisotropy. Background Art

[0002] With the continuous establishment of large-scale wind farms, the abundant wind resources on flat terrain are becoming scarce. Although many wind turbines are also installed offshore, complex terrain wind farms are gaining more and more attention due to the expensive cable and foundation costs (the cost of offshore installation is twice that of onshore installation). When selecting a suitable wind turbine location, it is crucial to correctly predict the wind speed and turbulence level on complex terrain, because wind turbines require high wind speeds and low turbulence levels to ensure high power generation and low fatigue loads. However, wind farms on complex terrain are always difficult to accurately predict because wind direction often changes, and flow separation and high turbulence may occur on the leeward side of hills. Accurately assessing the wind farm where the wind turbine is located is the first task before designing a high-efficiency wind farm, and it is also a huge challenge. Wind turbines and wind farms are located at the bottom of the atmospheric boundary layer, so studying the interaction between terrain flow, wind turbine flow, and future and atmospheric boundary layer flow fields is crucial to selecting a suitable wind turbine location and designing the best wind farm with high power generation efficiency and long life.

[0003] Many researchers have conducted extensive research on the characteristics of the atmospheric boundary layer and the wake of wind turbines over complex terrain. The main research methods include numerical simulation (including microscale, mesoscale and meso-microscale coupling), experiments (including wind tunnel experiments and field measurements) and theoretical models (including complex terrain wind field models and wake and terrain superposition models). The experimental method can only reflect limited measurement locations, and the anemometer must be calibrated before the experiment, which is more difficult to meet the requirements in complex terrain wind field tests (previously calibrated wind speed sensors will fail due to variable wind direction and high turbulence intensity); theoretical analysis is based on many assumptions, and its accuracy is still limited when dealing with complex flows. In contrast, numerical methods can fully reflect complex terrain wind fields in a shorter time and at a more economical cost. The biggest problem facing numerical methods is the solution of turbulence, which usually includes models such as RANS (Reynolds Averaged Navier-Stokes), LES (Large Eddy Simulation), DNS (Direct Numerical Simulation) and DES (Detached Eddy Simulation). Among these methods, the most commonly used method for simulating wind fields over complex terrain is RANS.

[0004] The most commonly used RANS turbulence models for simulating wind fields in complex terrain are the standard k-ε model and the standard SST k-ω model, both of which are based on the isotropic assumption. Previous studies have shown that the results of the k-ε turbulence model and the SST k-ω model on the windward side and the top of the mountain are consistent with the experimental measurement results; in terms of the prediction of velocity and shear stress in the separation zone, the SST k-ω model is more effective than the standard k-ε turbulence model. In order to solve the problem of difficulty in predicting the leeward wind field caused by high separation and turbulence, many researchers have also tried to modify the SSTk-ω model, such as modifying the closure coefficient of the turbulence model, adding nonlinear source terms, or modifying the wall function. The above modification methods have improved the accuracy of wind flow simulation in complex terrain to a certain extent, but these methods are based on the isotropic assumption and are inconsistent with the actual atmospheric boundary layer characteristics. Compared with the eddy viscosity model (EVM), the RSM model can accurately predict anisotropic turbulence, so it is more suitable for wind field simulation on complex terrain. Some scholars have also found that the RSM model can obtain better results when simulating wind flow over complex terrain, especially in terms of the generation of turbulence on the leeward side, but there is still a certain error between the numerical results and the experimental results. In addition, the rationality of the inlet boundary conditions is not considered enough (including the consistency of the inlet boundary profile and the influence of the inlet boundary on the accuracy of the downstream numerical simulation object flow field). Summary of the invention

[0005] The purpose of the present invention is to propose a numerical simulation method for complex terrain wind fields taking into account turbulence anisotropy, taking into account the turbulence anisotropy of the inlet boundary and the leeward slope, correcting the wall function, and at the same time correcting the turbulence model closure coefficient. It also simplifies the complex terrain modeling and grid generation methods, effectively improving the simulation accuracy.

[0006] In order to achieve the above technical objectives, the technical solution adopted by the present invention is: A method for numerically simulating a wind field in a complex terrain taking into account turbulence anisotropy, the method comprising the following steps: Step 1, considering the turbulence anisotropy of the inflow boundary, determine the type and parameters of the inflow boundary conditions; Step 2: Obtain the roughness information of the complex terrain and correct the wall function. The corrected wall function is: ; ; In the formula, and is the wall function constant, is the height of the center of the first-floor grid, is the surface roughness length; Step 3: Obtain high-precision terrain data, determine the scope of the calculation domain, and generate the calculation domain grid; Step 4: Based on the turbulence retention principle of the inflow boundary and the flow separation characteristics of the leeward slope, the closure coefficient of the turbulence model is modified and the eddy viscosity coefficient is The value of is corrected to 0.05, and the turbulent kinetic energy Prandtl number The value of is corrected to 0.82, and the turbulent dissipation rate Prandtl number The value of is corrected to 1.0, and the corrected linear pressure-strain RSM model is obtained; Step 5, based on the inflow boundary conditions in step 1, the wall function corrected in step 2, the computational domain grid generated in step 3, and the linear pressure-strain RSM model corrected in step 4, numerically simulate the complex terrain wind field to obtain the flow field information of the complex terrain wind field.

[0007] Step 1 further comprises: The types of wind speed and turbulent kinetic energy profiles at the inflow boundary, as well as parameter information including friction velocity, atmospheric boundary layer thickness, and roughness length are determined based on observational experiments or mesoscale numerical data.

[0008] In step 1, the following formula is used to obtain the logarithmic wind speed profile and ESDU turbulent kinetic energy profile: ; ; ; ; ; ; ; ; In the formula, , , are the wind speed, turbulent kinetic energy and dissipation rate profiles, , , , , , are the six Reynolds stress terms, is the friction speed, is the surface roughness length, is the atmospheric boundary layer thickness, is the height from the ground; At this time, the wind speed, turbulent kinetic energy and dissipation rate are constant and do not change with altitude.

[0009] As a preferred example, in step 2, the first layer grid height .

[0010] Step 3 further includes: Obtain high-precision terrain of the target wind farm area based on measured terrain data or digital elevation data; The calculation domain range is determined according to the mountain height and the location of the research target, and there is a distance between the leeward slope and the outlet boundary to ensure the stable development of turbulence on the leeward slope; Generate a computational domain mesh and refine it near the walls.

[0011] As a preferred example, in step 4, in the modified RSM model, the remaining closure coefficients are: , , , , and ; In the formula, is the coefficient of the generating term in the dissipation rate equation, is the coefficient of the dissipation term in the dissipation rate equation, is the slow pressure strain term coefficient, is the rapid pressure strain term coefficient, and is the wall reflection coefficient.

[0012] As a preferred example, in step 5, when there is a wind turbine, the following source term is added to the momentum equation: ; In the formula, is the thickness of the actuator disc, for The corresponding wind rotor thrust coefficient is, , and axial induction factor The following relationship exists: .

[0013] Step 5 further includes: The velocity inlet is used as the inlet boundary condition, the fully developed boundary is used as the outlet boundary, the side and top variable normal gradients are 0, and the modified wall function is used as the bottom surface to numerically simulate the wind field in complex terrain; The SIMPLE algorithm is used to solve the pressure-velocity coupling, the Green-Gauss node method is used for gradient discretization, the second-order upwind format is used for numerical discretization of velocity and turbulence variables, and the PRESTO! format is used for discretization of pressure terms; the numerical simulation residual is reduced to 1×10 -6 When , the flow field reaches a steady state; The wind speed, wind direction and turbulence intensity in the calculation domain are extracted, and the distribution characteristics of wind speed, wind direction and turbulence intensity in the vertical and horizontal directions are analyzed.

[0014] Compared with the prior art, the present invention has the following beneficial effects: First, the numerical simulation method of complex terrain wind fields considering turbulence anisotropy of the present invention takes into account the anisotropic characteristics of turbulence in complex terrain wind fields, improves the turbulence simulation capability of leeward slopes, and effectively improves the simulation accuracy. The numerical simulation results can be used for the micro-site selection and operation control of complex terrain wind farms.

[0015] Secondly, the numerical simulation method of complex terrain wind field considering turbulence anisotropy of the present invention ensures the rationality of the inflow boundary conditions and enables them to have better maintainability.

[0016] Third, the numerical simulation method of complex terrain wind field considering turbulence anisotropy of the present invention has simple terrain modeling and grid generation, takes into account both calculation accuracy and efficiency, and the calculation result is better than the two-equation turbulence model.

[0017] Fourth, by comparing with the measured data of the wind field, it is found that the numerical simulation method of the complex terrain wind field considering turbulence anisotropy of the present invention is very consistent with the experimental data on the upstream and leeward sides of all simulated mountains, and can well simulate the wind speed and turbulent kinetic energy distribution on the windward slope, mountain top and leeward slope. The wind speed on the leeward slope is particularly sensitive to the turbulence model, and the maximum relative root mean square error (RRMSE) of the four models tested can reach 11%. The simulation accuracy of turbulent kinetic energy depends on the retention of upstream turbulent kinetic energy and the prediction ability of the turbulence model for separated flow. The maximum relative root mean square error (RRMSE) of the four models can reach 47%. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The present invention provides a numerical simulation method for complex terrain wind fields taking into account turbulence anisotropy; Figure 2 This is a schematic diagram of the observation tower distribution in Case 1; Figure 3 This is the schematic diagram of the computational domain and grid for the numerical simulation of Case 1; Figure 4 Schematic diagram of numerical simulation results of Case 1, where (a) is a schematic diagram of wind speed at RS point changing with height, (b) is a schematic diagram of turbulent kinetic energy at RS point changing with height; (c) is a schematic diagram of wind acceleration ratio distribution along line A, (d) is a schematic diagram of turbulent kinetic energy increase ratio distribution along line A; (e) is a schematic diagram of wind acceleration ratio distribution along line AA, (f) is a schematic diagram of turbulent kinetic energy increase ratio distribution along line AA; (g) is a schematic diagram of wind acceleration ratio distribution along line B; Figure 5 This is a schematic diagram of the observation tower distribution in Case 2; Figure 6 This is the schematic diagram of the computational domain and grid for the numerical simulation of Case 2; Figure 7 Figure 2 is a schematic diagram of the numerical simulation results of Case 2, where (a) is a schematic diagram of the variation of wind speed at point M0 with height, (b) is a schematic diagram of the variation of turbulent kinetic energy at point M0 with height; (c) is a schematic diagram of the vertical distribution of wind speed at the locations of wind towers M7, M6, M3 and M8 along line B, (d) is a schematic diagram of the vertical distribution of turbulent kinetic energy at the locations of wind towers M7, M6, M3 and M8 along line B; (e) is a schematic diagram of the vertical distribution of wind speed at the locations of wind towers M1, M2, M3 and M4 along line A, (f) is a schematic diagram of the vertical distribution of turbulent kinetic energy at the locations of wind towers M1, M2, M3 and M4 along line A; (g) is a schematic diagram of the height distribution of island terrain along line B; (h) is a schematic diagram of the height distribution of island terrain along line A. DETAILED DESCRIPTION

[0019] The embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0020] The present invention discloses a numerical simulation method for a complex terrain wind field taking into account turbulence anisotropy, the method comprising the following steps: Step 1, considering the turbulence anisotropy of the inflow boundary, determine the type and parameters of the inflow boundary conditions; Step 2: Obtain the roughness information of the complex terrain and correct the wall function. The corrected wall function is: ; ; In the formula, and is the wall function constant, is the height of the center of the first-floor grid, is the surface roughness length; Step 3: Obtain high-precision terrain data, determine the scope of the calculation domain, and generate the calculation domain grid; Step 4: Based on the turbulence retention principle of the inflow boundary and the flow separation characteristics of the leeward slope, the closure coefficient of the turbulence model is modified and the eddy viscosity coefficient is The value of is corrected to 0.05, and the turbulent kinetic energy Prandtl number The value of is corrected to 0.82, and the turbulent dissipation rate Prandtl number The value of is corrected to 1.0, and the corrected linear pressure-strain RSM model is obtained; Step 5, based on the inflow boundary conditions considering anisotropic turbulence in step 1, the wall function corrected in step 2, the computational domain mesh generated in step 3, and the corrected linear pressure-strain RSM model in step 4, numerically simulate the complex terrain wind field to obtain the flow field information of the complex terrain wind field.

[0021] See also Figure 1 , the specific numerical simulation process is as follows: 1. Consider the turbulence anisotropy at the inflow boundary and determine the type and parameters of the inflow boundary conditions Determine the type of wind speed and turbulent kinetic energy profile at the inflow boundary based on observational experiments or mesoscale numerical data, and determine key information such as friction velocity, atmospheric boundary layer thickness, and roughness length. Taking the logarithmic wind speed profile and ESDU turbulent kinetic energy profile as examples, the calculation formula is as follows: ; ; ; ; ; ; ; ; In the formula, , , are the wind speed, turbulent kinetic energy and dissipation rate profiles, , , , , , are the six Reynolds stress terms, is the friction speed, is the surface roughness length, is the atmospheric boundary layer thickness, is the height from the ground; At this time, the wind speed, turbulent kinetic energy and dissipation rate are constant and do not change with altitude.

[0022] The above inflow boundary conditions can be imposed by UDF, and z at the inlet boundary is the height above the ground, which can also be extracted by UDF.

[0023] 2. Obtaining the surface roughness information of complex terrain and correcting the wall function Determine the ground roughness length of complex terrain based on observation data , the modified wall function calculation formula is as follows: ; ; In the formula, and is the wall function constant, The height of the center of the first-floor grid. The possible values ​​are .

[0024] The extraction of the first-layer grid height and the modified wall function can be completed by UDF.

[0025] 3. Obtain high-precision terrain data, determine the scope of the calculation domain, and generate the calculation domain grid The high-precision terrain of the target wind field area is obtained based on measured terrain data or digital elevation data; the calculation domain range is determined based on the mountain height and the research target location, and the calculation domain should ensure that the turbulence on the leeward slope is fully developed; the high-precision terrain does not need to be merged with the flat terrain. Generate the calculation domain grid and encrypt it near the wall.

[0026] 4. Based on the turbulence retention principle at the inlet boundary and the flow separation characteristics on the leeward slope, the closure coefficient of the turbulence model is modified.

[0027] The modified RSM turbulence model is used for numerical simulation of wind fields in complex terrain. When solving numerically, the expressions of the continuity equation and momentum equation are as follows: ; ; Where p is the average pressure, is the viscosity, is the source term, which can represent the effect of wind turbine on the flow field, is the Reynolds stress.

[0028] The RSM model is not based on the Boussinesq hypothesis. The Reynolds stress in each direction is solved by an equation, plus a dissipation rate equation, a total of seven additional equations are required. The classic RSM model Reynolds stress expression is as follows: ; In the formula, is the time derivative term, is the convection term, is the turbulent diffusion transport term, is the molecular diffusion term, is the pressure generating term, is the buoyancy generating term, is the pressure strain term, is the dissipation term, is the system rotation response, User-defined source items.

[0029] In the modified linear pressure-strain RSM model, Corrected from 0.09 to 0.05. In the model, is based on and The transport equation of RSM is modified, but the Reynolds stress transport equation in RSM is different from The model is different, so we cannot Modify the model as In the modified RSM model, we will From 1 to 0.82, Adjusted from 1.3 to 1.0. The other closure coefficients are not adjusted, which are: , , , , and .

[0030] If there is a wind turbine, the following source term can be added to the momentum equation: ; In the formula, is the thickness of the actuator disc, for The corresponding wind rotor thrust coefficient is, , and axial induction factor The following relationship exists: .

[0031] 5. Based on the inflow boundary conditions considering anisotropic turbulence in step 1, the wall function corrected in step 2, the computational domain mesh generated in step 3, and the corrected linear pressure-strain RSM model in step 4, the complex terrain wind field is numerically simulated to obtain the flow field information of the complex terrain wind field. When simulating the wind field in complex terrain, the inlet boundary condition is the velocity inlet; the outlet boundary is the fully developed boundary (Outflow), the normal gradient of the side and top variables is 0, and the bottom surface is the modified wall function.

[0032] When solving numerically, the pressure-velocity coupling is solved using the SIMPLE algorithm, the gradient discretization uses the Green-Gauss node-based method, the velocity and turbulence variables are discretized using the second-order upwind format, and the pressure term is discretized using the PRESTO! format. The numerical simulation residual is reduced to 1×10 -6 The flow field reaches a steady state when . The information of wind speed, wind direction and turbulence intensity in the computational domain is extracted, and the distribution characteristics of wind speed, wind direction and turbulence intensity in the vertical and horizontal directions are analyzed.

[0033] 6. Method Validation Case 1 Case 1 selects a mountain area to verify the effectiveness of the simulation method. In this area, southwest winds blow all year round, that is, the wind direction is roughly along the short axis. At the same time, the upstream terrain in the short axis direction is relatively flat, consistent with the coast, and the mountain and its surrounding surface are covered with short grass and low shrubs, so the surface roughness is relatively small.

[0034] In the field observation experiment, about 40 wind towers were distributed along Line A, Line AA and Line B. The intersection points of these three lines were located at the top of the mountain (HT) and the center point (CP), as shown in Figure 2. Figure 2 The observation data at the wind tower location can be used to verify the accuracy of the simulation results of the present invention. In this case, the inlet boundary condition is Figure 2 The measured data of the middle RS point is fitted, where , , The computational domain and grid are as follows: Figure 3 . Roughness length The computational domain size is 6 km (length) × 6 km (width) × 2 km (height), the horizontal grid size near the mountain is 7 m, and the minimum vertical grid size is 0.24 m. There are no wind turbines in this case, so no source items need to be added.

[0035] See also Figure 4 , the numerical wind speeds of the four RANS turbulence models are basically consistent with the initial values ​​given at the inlet, while the LES results are significantly larger above a height of 10 m. In terms of turbulent kinetic energy simulation, there are obvious differences in the numerical results of the four turbulence models. Below a height of 100 m at the RS point, the error of the turbulent kinetic energy simulated by the modified RSM model is smaller than that of the standard model. The average absolute errors of wind speed and turbulent kinetic energy of the modified RSM model proposed in the present invention are relatively small, both less than 0.12, while the average absolute error of the standard model is much larger, and the relative root mean square error of the modified RSM is also smaller than that of other models. Comprehensively comparing the errors of wind speed and turbulent kinetic energy values, the standard SST The results of the model are poor, while the results of the modified RSM model are the best, that is, the modified seven-equation turbulence model based on the anisotropy assumption has greater advantages in predicting the wind speed and turbulent kinetic energy of wind flow over complex terrain.

[0036] Case 2 Case 2 selects an island to verify the effectiveness of the simulation method. The island is about 100 meters in circumference and about 12 meters high. The top of the island is relatively flat, and there is a steep cliff on the west side. The existence of this cliff poses a great challenge to most flow solvers. The slope of the terrain around the island gradually decreases from west to east, and there is a gentle slope on the east side of the island. Therefore, although the island looks simple, the slope of the surrounding terrain changes from gentle slope to steep, which is very complex. There is no vegetation coverage on the cliff wall on the west side of the island, and other surfaces are covered with short grass. During the wind measurement period, 8 wind towers (M2 to M8) were installed on the island, located at Figure 5 For Line A and Line B, a wind tower is installed on the southwest side of the island and on the east side of the island.

[0037] In this case, the inlet boundary condition is given by Figure 5 The measured data of point M0 is fitted, where , , (On the island), (Sea surface). Roughness length (On the island and on the shore ( Figure 5 The green diagonal area represents the shore area). (sea surface), using UDF to achieve two roughness distributions. The computational domain and mesh are as follows Figure 6 The computational domain size is 0.9 km (length) × 0.8 km (width) × 0.2 km (height), the horizontal grid size near the mountain is 0.3 m, and the minimum vertical grid size is 0.12 m. There are no wind turbines in this case, so no source items need to be added.

[0038] See also Figure 7 The numerical wind speeds of the four RANS turbulence models are also very consistent with the experimental results. In terms of turbulent kinetic energy, at M0, the ESDU turbulent kinetic energy curve is in good agreement with the experimental data. The turbulent kinetic energy of the model and the standard RSM model decays significantly near the ground. The modified RSM model agrees well with the experimental values, while the turbulent kinetic energy of the LES model is smaller than the experimental values ​​above 5 m.

[0039] The RRMSE values ​​of the four RANS models are all within 0.25, indicating that the wind speed simulation accuracy is good. Combined with other wind speed error statistical indicators, the modified RSM model has the best performance. Through a comprehensive analysis of the turbulent kinetic energy statistical errors of all anemometers, the modified RSM model is better than the standard RSM model, and the worst performance is the standard SST model.

[0040] After comparing with measured data from multiple wind fields, the present invention found that the newly proposed modified model can well simulate the wind speed and turbulent kinetic energy at the front of the mountain, the top of the mountain and the leeward slope, and its accuracy is close to the experimental results. It has great advantages in maintaining the inlet boundary conditions and predicting the separated flow on the leeward slope. The model is simple in form, easy to encode, and efficient and accurate in calculation.

[0041] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A numerical simulation method for complex terrain wind field considering turbulence anisotropy, characterized in that: The method comprises the following steps: Step 1, considering the turbulence anisotropy of the inflow boundary, determine the type and parameters of the inflow boundary conditions; Step 2: Obtain the roughness information of the complex terrain and correct the wall function. The corrected wall function is: ； ； In the formula, and is the wall function constant, is the height of the center of the first-floor grid, is the surface roughness length; Step 3: Obtain high-precision terrain data, determine the scope of the calculation domain, and generate the calculation domain grid; Step 4: Based on the turbulence retention principle of the inflow boundary and the flow separation characteristics of the leeward slope, the closure coefficient of the turbulence model is modified and the eddy viscosity coefficient is The value of is corrected to 0.05, and the turbulent kinetic energy Prandtl number The value of is corrected to 0.82, and the turbulent dissipation rate Prandtl number The value of is corrected to 1.0, and the corrected linear pressure-strain RSM model is obtained; Step 5, based on the inflow boundary conditions in step 1, the wall function corrected in step 2, the computational domain grid generated in step 3, and the linear pressure-strain RSM model corrected in step 4, numerically simulate the complex terrain wind field to obtain the flow field information of the complex terrain wind field.

2. The complex terrain wind field numerical simulation method considering turbulence anisotropy according to claim 1 is characterized in that: Step 1 further comprises: The types of wind speed and turbulent kinetic energy profiles at the inflow boundary, as well as parameter information including friction velocity, atmospheric boundary layer thickness, and roughness length are determined based on observational experiments or mesoscale numerical data.

3. The method for numerical simulation of wind field in complex terrain considering turbulence anisotropy according to claim 1 is characterized in that: In step 1, the following formula is used to obtain the logarithmic wind speed profile and ESDU turbulent kinetic energy profile: ； ； ； ； ； ； ； ； In the formula, , , are the wind speed, turbulent kinetic energy and dissipation rate profiles, , , , , , are the six Reynolds stress terms, is the friction speed, is the surface roughness length, is the atmospheric boundary layer thickness, is the height from the ground; At this time, the wind speed, turbulent kinetic energy and dissipation rate are constant and do not change with altitude.

4. The complex terrain wind field numerical simulation method considering turbulence anisotropy according to claim 1 is characterized in that: In step 2, the first layer grid height .

5. The complex terrain wind field numerical simulation method considering turbulence anisotropy according to claim 1 is characterized in that: Step 3 further includes: Obtain high-precision terrain of the target wind farm area based on measured terrain data or digital elevation data; The calculation domain range is determined according to the mountain height and the location of the research target, and there is a distance between the leeward slope and the outlet boundary to ensure the stable development of turbulence on the leeward slope; Generate a computational domain mesh and refine it near the walls.

6. The complex terrain wind field numerical simulation method considering turbulence anisotropy according to claim 1 is characterized in that: In step 4, in the modified RSM model, the remaining closure coefficients are: , , , , and ; In the formula, is the coefficient of the generating term in the dissipation rate equation, is the coefficient of the dissipation term in the dissipation rate equation, is the slow pressure strain term coefficient, is the rapid pressure strain term coefficient, and is the wall reflection coefficient.

7. The complex terrain wind field numerical simulation method considering turbulence anisotropy according to claim 1 is characterized in that: In step 5, when there is a wind turbine, add the following source term to the momentum equation: ； In the formula, is the thickness of the actuator disc, for The corresponding wind rotor thrust coefficient is, , and axial induction factor The following relationship exists: .

8. The complex terrain wind field numerical simulation method considering turbulence anisotropy according to claim 1 is characterized in that: Step 5 further includes: The velocity inlet is used as the inlet boundary condition, the fully developed boundary is used as the outlet boundary, the side and top variable normal gradients are 0, and the modified wall function is used as the bottom surface to numerically simulate the wind field in complex terrain; The SIMPLE algorithm is used to solve the pressure-velocity coupling, the Green-Gauss node method is used for gradient discretization, the second-order upwind format is used for numerical discretization of velocity and turbulence variables, and the PRESTO! format is used for discretization of pressure terms; the numerical simulation residual is reduced to 1×10 -6 When , the flow field reaches a steady state; The wind speed, wind direction and turbulence intensity in the calculation domain are extracted, and the distribution characteristics of wind speed, wind direction and turbulence intensity in the vertical and horizontal directions are analyzed.

Images