Wind driven generator structure high-cycle fatigue cross-platform simulation method based on multi-scale coupling modeling

Abstract

The invention discloses a wind driven generator structure high-cycle fatigue cross-platform simulation method based on multi-scale coupling modeling, and aims to accurately evaluate a fatigue damage state of a wind driven generator structure under the action of a high-cycle low-amplitude load. Firstly, a multi-scale model based on coupling of a fan overall structure finite element model unit and a local cracking area near-field dynamic model is established, and fan power time-history analysis is carried out based on a finite element analysis platform; then response data are transmitted to a processing platform, a rain flow counting method and a linear damage accumulation criterion are combined, the fatigue damage state of the connection key is determined, a fatigue life attenuation value is calculated, the fatigue cycle number needed by local crack propagation is determined, and the residual fatigue life of the unbroken key is updated; and finally, correcting the fan model, removing a newly added fracture key at the tip of the crack, and repeating dynamic time-history analysis and damage updating until the structure fails. According to the method, whole-process simulation and fatigue behavior analysis of fan fatigue crack propagation are realized, and important reference is provided for fan design optimization and maintenance.

Description

Technical Field

[0001] The present invention belongs to the field of structural crack propagation simulation and fatigue life prediction, and relates to a cross-platform simulation method for high-cycle fatigue of wind turbine structures based on multi-scale coupled modeling. Background Art

[0002] Wind turbines operate in complex and variable natural environments and are subject to various random and periodic loads such as wind loads, centrifugal forces, and wave loads (offshore wind turbines) for a long time. Long-term cyclic loading can induce microcracks in the stress concentration areas (such as key parts like welds) of the wind turbine structure, and with the continuous cyclic loading, fatigue damage gradually accumulates. The propagation of microcracks may ultimately lead to structural fracture or failure, seriously threatening the safe operation of the wind turbine. The common analysis methods in the fatigue research of wind turbine structures are mainly divided into two categories: 1. Fatigue assessment methods based on monitoring, which rely on real-time operation data of the wind turbine and can evaluate the fatigue state and predict the remaining life. However, limited by sensor layout and data integrity, the reliability of prediction results is relatively low in cases where data collection is insufficient or long-term monitoring is lacking; 2. Fatigue damage analysis and life prediction methods based on simulation, which usually adopt finite element models. By analyzing strain time history data, fatigue damage accumulation is calculated to provide a detailed analysis of the internal stress-strain state, making up for the deficiencies of the monitoring method. However, traditional finite elements still have certain limitations in simulating the processes of crack initiation, propagation, and final fracture.

[0003] Different from other indoor or small devices, as a large outdoor device, the wind turbine is long-term exposed to the natural environment and is subject to complex and variable climate conditions and external loads. Therefore, the crack derivation mechanism of the wind turbine is more complex, being affected by the long-term cumulative impact of environmental factors and being closely related to fatigue damage under dynamic loads. This unique operating environment and load conditions make the crack simulation and prediction technology of wind turbines face higher difficulties and greater challenges. The existing technology lacks a method that can accurately simulate the cracks of wind turbines.

[0004] For the simulation methods on other devices, there are also deficiencies in the methods themselves. Although the Chinese patents of CN116151071A, CN118350244A, CN113761760A, CN114970226A, and CN116895351A can simulate the crack derivation process to a certain extent by using a multi-scale modeling method of finite element-peridynamics coupling, the constraint methods of "shared nodes" or "regional coupling" are insufficient in terms of flexibility. Especially when refined modeling is required in a local area, the adaptability of this method is poor.

[0005] In addition, the criteria for judging the fracture of the connecting keys in the above patents are still limited to single parameters such as key elongation rate or energy density, making it difficult to accurately characterize the progressive damage process under complex load spectra and unable to reflect the cumulative effect of microscopic damage in materials. Although Chinese patents CN117709171A, CN118656986A, and CN17172025A introduce the "remaining life" damage variable as the fracture criterion for connecting keys, their technical solutions still conduct crack propagation simulations within the framework of a single peridynamics model, resulting in a sharp increase in computational cost as the analysis area expands, which restricts their application feasibility in actual large-scale projects.

[0006] Therefore, there is a lack in the art of an efficient and accurate simulation method that can overcome the above disadvantages and predict the high-cycle fatigue of wind turbines. Summary of the Invention

[0007] The present invention proposes a cross-platform simulation method for high-cycle fatigue of wind turbine structures based on multi-scale coupled modeling. Based on the "multi-master node - single slave node" multi-point constraint method, it combines the peridynamics fatigue model and the finite element method, comprehensively considering the complexity of the wind turbine structure and the characteristics of fatigue crack initiation and propagation. This method can accurately simulate the propagation process of fatigue cracks, providing a scientific basis for the safe operation, maintenance decision-making, and life cycle management of wind turbines.

[0008] Technical Solution: The cross-platform simulation method for high-cycle fatigue of wind turbine structures based on multi-scale coupled modeling of the present invention includes the following steps:

[0009] Step 1: Adopt the "multi-master node - single slave node" multi-point constraint method to establish a multi-scale model that couples the finite element and peridynamics of the wind turbine structure including the local cracking area.

[0010] Step 2: Determine the high-cycle fatigue load according to the operating environment of the wind turbine, conduct a dynamic time-history analysis of the wind turbine, calculate the responses of the overall wind turbine and the local cracking area, and obtain the simulation data.

[0011] Step 3: Input the simulation data in Step 2 into the data processing platform, and conduct cyclic counting analysis on the strain response data of the local cracking area of the wind turbine based on the rainflow counting method; combine the multi-scale model and the linear damage accumulation criterion to calculate the fatigue life attenuation value of the connecting keys in the wind turbine structure and the number of load cycles required for local crack propagation.

[0012] Step 4: Update the remaining fatigue life and failure parameters of the connecting keys in the multi-scale model according to the number of load cycles, and calculate the fatigue damage state of the peridynamics particles in the multi-scale model.

[0013] Step 5: Update the multi-scale model of the wind turbine, and repeat Steps 2 to 4 until the wind turbine structure fails.

[0014] Preferably, in the method for cross-platform simulation of high-cycle fatigue of a wind turbine structure based on multi-scale coupling modeling, in step three, the fatigue life attenuation value Δλ of the connection key is calculated. i j It is determined by formula (1); the number of load cycles N required for local crack propagation is calculated. cyc It is determined by formula (2):

[0015]

[0016]

[0017] In the formula: and are respectively the k-th strain amplitude and the corresponding number of cycles of the j-th key in the i-th step, A 2 and m 2 are fatigue parameters in the crack propagation stage.

[0018] Preferably, in the method for cross-platform simulation of high-cycle fatigue of a wind turbine structure based on multi-scale coupling modeling, in step four, the range of the near-field domain of the crack tip in the locally cracked area of the wind turbine is the crack propagation stage, and the remaining fatigue life λ of the connection key i j is determined by formula (3); if it is not within the near-field domain range, the connection key is in the crack nucleation stage, and the remaining fatigue life λ of the connection key i j is determined by formula (4); the failure parameter of the connection key is determined by formula (5); the fatigue damage D(x) of any peridynamic particle x in the multi-scale model is determined by formula (6):

[0019]

[0020] In the formula: A 1 and m 1 are fatigue parameters in the crack nucleation stage; is the neighborhood of the peridynamic particle x, V x′ is the volume of each peridynamic particle x′ in the neighborhood of x, ξ is the original length of the bond, η is the relative displacement vector of the bond, and μ is the failure parameter of the bond;

[0021] The near-field domain is represented as a spherical volume region centered on the particle x with a radius of δ in the three-dimensional multi-scale model and as a circular region centered on the particle x with a radius of δ in the two-dimensional multi-scale model. δ is called the near-field domain radius; the particle x only interacts with other particles within its near-field domain.

[0022] Preferably, in the method for cross-platform simulation of high-cycle fatigue of a wind turbine structure based on multi-scale coupled modeling, in step three, the rainflow counting method is used to identify the cyclic loading part in the strain-time history curve of each connecting key, and the complex strain-time history is decomposed into a series of strain amplitudes, means, and corresponding cycle numbers; the peridynamic fatigue model is used to determine the fatigue life of the connecting key under different strain amplitudes, and the linear damage accumulation criterion is used to estimate the damage degree of the connecting key.

[0023] Preferably, in the method for cross-platform simulation of high-cycle fatigue of a wind turbine structure based on multi-scale coupled modeling, in step four, when the fatigue damage degree D(x) of the peridynamic particles in the multi-scale model is greater than D limit (generally taken as 0.3 - 0.5), then this particle is the crack tip, and at the same time, the crack propagation length is updated.

[0024] Preferably, in the method for cross-platform simulation of high-cycle fatigue of a wind turbine structure based on multi-scale coupled modeling, in step five, after each analysis is completed, only the newly added broken bonds at the crack tip need to be removed to update the model, thus avoiding re-modeling.

[0025] This application is used for the full-process simulation of fatigue crack propagation of a fan structure under a high-cycle fatigue cyclic load environment, accurately analyzing the fatigue behavior and cumulative damage process, and providing a reasonable reference for the design optimization, operation and maintenance, and life cycle management of the fan structure.

[0026] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0027] (1) This application is based on a multi-scale model that couples finite element model units and peridynamic model units, conducts high-cycle fatigue cross-platform numerical simulation research on the wind turbine structure, and innovatively proposes a multi-master node - single slave node multi-point constraint method, which can significantly reduce the calculation cost while ensuring the calculation accuracy of the model, is applicable to the coupled connection between any type of solid element and peridynamic model unit, improves the simulation accuracy while reducing the calculation cost, enhances the flexibility of simulating the local cracking area, and provides a flexible and efficient solution for the multi-scale modeling of complex structures.

[0028] (2) This application constructs a full-process numerical simulation system for high-cycle fatigue of a fan based on a multi-platform collaborative architecture, breaking through the limitations of the traditional single-platform framework in terms of functional modularity and algorithm compatibility. A single platform is limited by its inherent algorithm library and module architecture, making it difficult to efficiently integrate cross-platform professional functional modules, and there are problems such as underlying numerical conflicts and energy transfer distortion in the multi-scale model.

[0029] (3) This application introduces the damage variable of "remaining fatigue life", realizes the dynamic update of the remaining fatigue life of the connecting key, can accurately describe the processes of crack initiation, propagation and final fracture, and improves the simulation accuracy and crack tracking ability.

[0030] (4) This application can not only accurately simulate the fatigue crack propagation process of the fan structure, but also accurately predict the fatigue life, providing effective support for the design, optimization, operation and maintenance, and life cycle management of the fan. Description of the Drawings

[0031] Figure 1 (a) Schematic diagram of the multi-scale coupling model of the fan and the loads it bears;

[0032] Figure 1 (b) Truss beam element and its cross-section;

[0033] Figure 1 (c) Coupling interface between beam element and peridynamics model element;

[0034] Figure 1 (d) Coupling interface between solid element and peridynamics model element;

[0035] Figure 2 (a) Local mesh division under the constraint of "multiple master nodes - single slave node" of finite element model element - peridynamics model element;

[0036] Figure 2 (b) Local mesh division under the "co-node" connection of finite element model element - peridynamics model element;

[0037] Figure 3 (a) Three-dimensional schematic diagram of the pre-crack;

[0038] Figure 3 (b) Multi-scale model of the local cracking area in actual modeling;

[0039] Figure 3 (c) Elevation views of three pre-crack modes;

[0040] Figure 4(a) shows the time history of the pulsating wind speed at the rotor;

[0041] Figure 4(b) shows the comparison between the simulated wind speed power spectrum and the Davenport spectrum;

[0042] Figure 5 is the time history of the wind load at the rotor;

[0043] Figure 6 is the strain time history curve of the key with the maximum deformation at the crack tip;

[0044] Figure 7Histogram of the strain amplitude and mean strain of the key with the largest deformation at the pre-crack tip;

[0045] Figure 8 Fatigue life attenuation value of the key in the crack propagation stage;

[0046] Figure 9 Remaining fatigue life of the key in the crack propagation stage;

[0047] Figure 10 Data flow of the multi-scale simulation method for high-cycle fatigue of the fan;

[0048] Figure 11 (a) Crack propagation in the peridynamic simulation region at crack propagation step i = 1;

[0049] Figure 11 (b) Crack propagation in the peridynamic simulation region at crack propagation step i = 15;

[0050] Figure 11 (c) Crack propagation in the peridynamic simulation region at crack propagation step i = 30;

[0051] Figure 12 Quantitative correspondence between fatigue life and crack length; Specific implementation manner

[0052] The following further illustrates the specific implementation manner of the present invention in conjunction with the preferred embodiments of the present invention and the accompanying drawings of the specification.

[0053] Step 1: Establish a multi-scale model that couples the finite element model unit of the overall structure of the fan and the peridynamic model unit of the local cracking area in OpenSees. In the preferred embodiment of the present invention, the height of the fan tower is 90 m, and the density, elastic modulus, and Poisson's ratio of the tower material are 8.5×10 5 kg / m 3 , 210 GPa, and 0.25, respectively. The top diameter of the tower is 3.87 m, the bottom diameter is 6 m, and the thicknesses are 0.019 m and 0.027 m, respectively. The masses of the rotor, nacelle, and tower are 1.1×10 5 kg, 2.4×10 5 kg, and 3.4746×10 5 kg, as shown in Figure 1 (a). The tower is simulated by beam elements (as shown in Figure 1 b), and the bottom flange and the local cracking area are modeled by solid elements and peridynamic model units, respectively. The coupling interfaces between the beam element - solid element and the solid element - peridynamic model unit are as shown in Figure 1 (c) and Figure 1(d) As shown, through the multi - master - node - single - slave - node multi - point constraint method, a bridge is formed between beam elements, solid elements and peridynamics to transfer forces and displacements. Each element acts synergistically and influences each other, forming a complete multi - scale model. The complete multi - scale model is used for analysis during computational analysis.

[0054] Figure 2 (a) shows the local mesh division under the "multi - master - node - single - slave - node" constraint of the finite - element model unit - peridynamics model unit. Figure 2 (b) shows the local mesh division under the "co - node" connection of the finite - element model unit - peridynamics model unit. On the premise of maintaining the consistency of the calculation accuracy of the peridynamics model unit, the "co - node" connection scheme needs to construct a mesh discretization scale of 169,614 solid elements, while the "multi - master - node - single - slave - node" constraint scheme only needs 8,154 solid elements to achieve equivalent modeling accuracy. Compared with the traditional interface treatment method, the method adopted in this application significantly reduces the mesh refinement requirements in unnecessary areas and improves the calculation efficiency of the multi - scale coupling model.

[0055] In the peridynamics model unit, the mesh spacing Δx is 0.0027 m, the peridynamic domain radius δ is 0.008 m, and the micro - modulus c is 1.2×10 21 N / m 6 . Three different initial cracks with lengths of 0.105 m, 0.157 m, and 0.209 m are set in the local cracking area ( Figure 3 ).

[0056] Step 2: Determine the high - cycle fatigue load according to the operating environment of the fan. The present invention preferably selects the wind load as the high - cycle fatigue cyclic load for the dynamic time - history analysis of the fan. On the Matlab platform, the harmonic superposition method is used to numerically simulate the pulsating wind speeds at different heights of the fan tower barrel ( Figure 1 nodes in a). At the rated wind speed (11.4 m / s), a total of 12 segments of pulsating wind speed time - histories are simulated, with each segment having a simulation duration of 600 seconds and a time step of 0.1 second. Figure 4(a) shows the pulsating wind speed time - history at the rotor, and Figure 4(b) shows the comparison between the simulated wind speed power spectrum and the Davenport spectrum. Subsequently, calculate the wind load time - history at each node ( Figure 5 ), and call OpenSees for dynamic time - history analysis to simulate the overall and local responses of the fan structure with initial damage. Figure 6 shows the strain time - history curve of the key with the largest deformation at the crack tip;

[0057] Step 3: Import the strain time - history data obtained in OpenSees into Matlab, and use the rain - flow counting method to extract the strain amplitude, mean value, and number of cycles ( Figure 7As shown. Meanwhile, the fatigue life decay value of the connection bond in the crack propagation stage is calculated by combining the multi-scale model and the linear damage accumulation criterion (Formula 1), Figure 8 The fatigue life decay value of the connection bond in the crack propagation stage is shown. The number of fatigue cycles N required for local crack propagation is calculated by Formula (2) cyc :[[]]

[0058]

[0059] In the formula: and are respectively the k-th strain amplitude and the corresponding number of cycles of the j-th bond in the i-th step, A 2 and m 2 are fatigue parameters in the crack propagation stage.

[0060] Step 4: According to the position of the crack tip, judge the crack stage of the connection bond and calculate the remaining fatigue life. The remaining fatigue life λ of the connection bond in the crack propagation stage i j is calculated by Formula (3), and the remaining fatigue life of the connection bond in this stage is as Figure 9 shown. The remaining fatigue life λ of the connection bond in the crack nucleation stage i j is determined by Formula (4). The failure parameter μ of the bond is calculated according to Formula (5), and the fatigue damage D(x) of the peridynamic particles is calculated by Formula (6).

[0061]

[0062] In the formula: A 1 and m 1 are fatigue parameters in the crack nucleation stage; is the neighborhood of the peridynamic particle x, V x′ is the volume of each peridynamic particle x′ in the neighborhood of x, ξ is the original length of the bond, η is the relative displacement vector of the bond, and μ is the failure parameter of the bond.

[0063] If the fatigue damage D(x) of the particles in the peridynamic model element is greater than D limit (generally taken as 0.3 - 0.5), then the position of this particle is considered as the crack tip, and the crack length L is calculated according to Formula (7):

[0064]

[0065] In the formula: x 1 and y 1 are the abscissa and ordinate of the left crack tip, x 2 and y 2are the abscissa and ordinate of the right tip of the crack, and R is the radius of the bottom of the tower barrel.

[0066] Step Five: Update the multi-scale model of the wind turbine, and modify the failure parameters of the newly added fracture bonds in the peridynamic model elements. Repeat the dynamic time history analysis and fatigue damage analysis of the wind turbine until the crack length L reaches the critical crack length L c , thereby determining the failure of the wind turbine structure. The critical crack length L c is calculated by formula (8):

[0067]

[0068] In the formula: K IC is the fracture toughness of the deformation of the tower barrel, f is the geometric correction factor, and σ max is the maximum cyclic stress on the tower barrel.

[0069] Figure 10 This is the flow chart of the high-cycle fatigue cross-platform simulation method for the wind turbine structure based on multi-scale coupling modeling in the preferred embodiment of the present invention. Figure 11 shows the crack propagation of the peridynamic simulation area of the wind turbine structure under different crack propagation steps. Figure 12 It shows the quantitative corresponding relationship between the fatigue life and the crack length. According to the quantitative relationship between the fatigue life and the crack length growth and the prediction of the crack propagation path, the maintenance cycle and key inspection areas of the wind turbine structure can be scientifically formulated, providing a theoretical basis and technical support for the fatigue crack monitoring, repair and reinforcement strategies of the wind turbine structure.

[0070] The implementation scheme of the present invention is not limited to the above preferred embodiments. All combinations and improvements of the above technical features shall be regarded as the protection scope of the present invention. For those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A cross-platform simulation method for high-cycle fatigue of wind turbine structures based on multi-scale coupling modeling, characterized in that: The following steps are involved: Step 1: Using the multi-point constraint method of "multiple master nodes-single slave node", a multi-scale model of the wind turbine structure coupled with finite element and peridynamics including the local cracking area is established; Step 2: Determine the high cycle fatigue load according to the wind turbine operating environment, carry out the wind turbine dynamic time history analysis, calculate the response of the wind turbine as a whole and the local cracking area, and obtain simulation data; Step 3: The simulation data in step 2 is transferred to the data processing platform, and the strain response data of the local cracking area of ​​the wind turbine is analyzed by cycle counting based on the rain flow counting method; the fatigue life attenuation value of the connecting key in the wind turbine structure and the number of load cycles required for the local crack to propagate are calculated by combining the multi-scale model and the linear damage accumulation criterion; Step 4: Update the remaining fatigue life and failure parameters of the connecting key in the multiscale model according to the number of load cycles, and calculate the fatigue damage state of the peridynamic particles in the multiscale model; Step 5: Update the wind turbine multiscale model and repeat steps 2 to 4 until the wind turbine structure fails.

2. The method according to claim 1, characterized in that In step 3, the fatigue life attenuation value of the connecting key is calculated. Determined by formula (1); Calculate the number of load cycles N required for local crack growth cyc Determined by formula (2): Where: and are the kth strain amplitude and the corresponding number of cycles of the jth root bond in the ith step, respectively. A2 and m2 are the fatigue parameters in the crack propagation stage.

3. The method according to claim 1, characterized in that In step 4, the local crack area of ​​the fan is within the near field of the crack tip, which is the crack extension stage, and the remaining fatigue life of the connecting key is Determined by formula (3); if it is not within the near field, the connecting key is in the crack nucleation stage, and the remaining fatigue life of the connecting key is is determined by formula (4); the failure parameter of the connecting key is determined by formula (5); the fatigue damage D(x) of any peridynamic particle x in the multiscale model is determined by formula (6): Where: A1 and m1 are fatigue parameters in the crack nucleation stage; is the neighborhood of the peridynamic particle x, V x′ is the volume of each peridynamic particle x′ in the neighborhood of x, ξ is the original length of the bond, η is the relative displacement vector of the bond, and μ is the failure parameter of the bond; The near field is represented as a spherical volume region with particle x as the center and a radius of δ in the three-dimensional multi-scale model, and as a circular region with particle x as the center and a radius of δ in the two-dimensional multi-scale model, where δ is called the near field radius; particle x only interacts with other particles in its near field.

4. The method according to claim 1, characterized in that: In the step 3, the rain flow counting method identifies the cyclic loading part in the strain time history curve of each connecting bond, and decomposes the complex strain time history into a series of strain amplitudes, means and corresponding cycle numbers; The peridynamic fatigue model is used to determine the fatigue life of the connector under different strain amplitudes, and the linear damage accumulation criterion is used to estimate the damage extent of the connector.

5. The method according to claim 3, characterized in that: In step 4, when the fatigue damage degree D(x) of the peridynamic particles in the multiscale model is greater than the limit value D limit , then the particle is the crack tip, and the length of crack extension is updated at the same time; D limit The value is 0.3-0.

5.

6. The method according to claim 1, characterized in that In the step 5, after each analysis of step 2 to step 4 is completed, the model can be updated by simply removing the newly added broken bonds at the crack tip, thereby avoiding re-modeling.

7. Application of the cross-platform simulation method for high-cycle fatigue of wind turbine structure based on multi-scale coupling modeling as described in any one of claims 1 to 6 in fatigue simulation of wind turbine.

Images