Parameter identification method of wind power mixed tower structure based on multi-source test information

Abstract

The invention discloses a parameter identification method of a wind power mixed tower structure based on multi-source test information, and belongs to the technical field of wind power mixed tower structure monitoring. In order to accurately identify parameters of the wind power mixed tower structure, a finite element model of the wind power mixed tower structure is established, and a discrete state equation is obtained based on improved unscented Kalman filtering solution; the method comprises the following steps: acquiring reaction time history information of a wind power mixed tower structure, establishing an observation equation by taking actually measured data as observation data, and resolving based on improved unscented Kalman filtering to obtain a discrete observation equation; a data time window length t is given, an observed quantity reconstruction parameter is solved by adopting a Chebyshev decomposition highest order determination method based on acceleration signal spectrum characteristics and the highest frequency of the signal, and then observation data except acceleration are reconstructed by utilizing a least square method and a Chebyshev polynomial to obtain reconstructed observation data; and recognizing the structural parameters of the wind power mixed tower based on an improved adaptive unscented Kalman filtering method by using the reconstructed observation data.

Description

Technical Field

[0001] The invention belongs to the technical field of wind power hybrid tower structure monitoring, and in particular relates to a parameter identification method of a wind power hybrid tower structure based on multi-source test information. Background Art

[0002] Wind energy is the most promising renewable energy source. In recent years, the global wind power installed capacity has developed rapidly. my country's wind power installed capacity is increasing year by year. According to the characteristics of domestic terrain, my country's wind power development is mainly based on low wind speed and high shear. Hybrid towers above 140 meters will become the mainstream choice for developing low wind speed wind farms in my country with their height advantages, good economy and stable support. The current problem with the application of hybrid tower structures is that there is still a lack of sufficient data support for the degradation law of structural bearing capacity and concrete fatigue damage under complex time-varying external loads. At present, a complete hybrid tower technical standard system has not been established in China, and there are still blind spots in the reference hybrid tower structure monitoring standards. There are few complete monitoring data for completed hybrid tower projects, which brings great difficulties to the defect analysis, damage identification and tower collapse analysis of the hybrid tower structure system. Therefore, it is urgent to develop continuous online monitoring research on towers in actual projects to provide guarantee for the safe operation of the units. Summary of the invention

[0003] The problem to be solved by the present invention is to accurately identify the parameters of a wind power hybrid tower structure, and a method for identifying the parameters of a wind power hybrid tower structure based on multi-source test information is proposed.

[0004] To achieve the above object, the present invention is implemented through the following technical solutions:

[0005] A parameter identification method for a wind turbine hybrid tower structure based on multi-source test information comprises the following steps:

[0006] S1. Establish a finite element model of the wind turbine hybrid tower structure based on the drawings and field data, and then solve the finite element model based on the improved unscented Kalman filter to obtain the discretized state equation;

[0007] S2. The response time history information of the wind power tower structure is collected through displacement sensors, inclination sensors, static level settlement sensors, strain sensors, and acceleration sensors. The measured data is used as observation data to establish an observation equation, and then the observation equation is solved based on the improved unscented Kalman filter to obtain a discretized observation equation;

[0008] S3. Based on the observed data obtained in step S2, given the data time window length t, based on the acceleration signal spectrum characteristics and the highest frequency f of the signal max, the highest order determination method of Chebyshev decomposition is used to solve the observation reconstruction parameter α, and then the observation data other than acceleration are reconstructed using the least squares method and Chebyshev polynomials to obtain the reconstructed observation data;

[0009] S4. Using the reconstructed observation data, the structural parameters of the wind turbine hybrid tower are identified based on the improved adaptive unscented Kalman filter method.

[0010] Furthermore, the drawings and on-site data in step S1 include geometric dimensions, steel structure specifications, elastic modulus, concrete grade and elastic modulus of the wind power hybrid tower structure;

[0011] The expression of the state equation after discretization is solved as follows:

[0012] X k =f(X k-1 ,U k-1 )+w k-1

[0013] Among them, X k represents the state quantity of the kth recursive step, f represents the time-varying nonlinear function including the state quantities X and U, and U k-1 represents the system input matrix of the k-1th recursive step, w k-1 represents the noise of the k-1th recursive step.

[0014] Further, the reaction time history information in step S2 includes the vibration of the mixed tower structure, the inclination of the hub, the foundation settlement, the strain of the mixed tower structure, and the cable force collected by the displacement sensor, the inclination sensor, the static level settlement sensor, the strain sensor, and the acceleration sensor;

[0015] Design and determine the monitoring plan for the mixed tower structure, including the fast acquisition time history curve corresponding to acceleration, and the slow acquisition time history curve corresponding to displacement transmission, inclination transmission, uneven settlement, and strain;

[0016] The expression of the discrete observation equation obtained by solving the observation equation based on the improved unscented Kalman filter is:

[0017] Y k =h(X k ,U k )+v k

[0018] Among them, Y k represents the observed value of the kth recursive step, h represents the functional relationship between the state quantity, external load and the observed value, and U k represents the system input matrix of the kth recursive step, v k represents the observed disturbance at the k-th recursive step.

[0019] Furthermore, the specific implementation method of step S3 includes the following steps:

[0020] S3.1. The acceleration signal in the observation data obtained in step S2 Decompose, the decomposition formula is:

[0021]

[0022] Where T represents the Chebyshev polynomial basis function, W represents the Chebyshev polynomial basis function coefficient, and N f represents the order of the highest Chebyshev polynomial;

[0023] The highest Chebyshev polynomial order; the highest frequency f of the acceleration signal passed max OK, the expression is:

[0024] N f =α×5×t

[0025] Among them, α is the observation reconstruction parameter, and t is the length of the data time window;

[0026] S3.2. The observation reconstruction parameter α is calculated by the least squares method and with the number of periods np in the signal, the expression is:

[0027]

[0028] np=f max ×t;

[0029] S3.3. Determine α based on acceleration, and for other fast or slow test signals, perform least squares fitting based on the acceleration sampling frequency, the test signal and the Chebyshev polynomial to reconstruct the observation data other than the acceleration to obtain the reconstructed observation data.

[0030] Furthermore, the specific implementation method of step S4 includes the following steps:

[0031] S4.1. Set the initial values ​​of the parameters to be identified and the initial covariance matrix, the expression is:

[0032]

[0033] in, represents the initial state quantity to be identified, X0 represents the initial actual state quantity, the actual state quantity is obtained by calculation, E represents the expected value, is the initial covariance matrix;

[0034] The state quantity of each time step is recursively identified by UKF. During the recursive identification process, each time step will generate (2n+1) sigma points for UT transformation. The transformation process is shown in the following formula:

[0035]

[0036] in, represents the sigma point of the k-1th recursive step, j = 1, 2, ..., n, n represents the dimension of the state quantity, λ represents the gain, and λ depends on the dimension of the state quantity and the statistical distribution law of the state quantity;

[0037] S4.2. Based on the state equation, each sigma point at each time point is solved according to the state equation to obtain the estimated value of the sigma point at the next time point, as shown in the following formula:

[0038]

[0039] Make a priori estimates and calculate the state quantity mean and covariance matrix by the following formula:

[0040]

[0041] in, is the predicted value of the sigma sampling point, is the first weighting coefficient;

[0042]

[0043] in, is the prediction covariance, is the second weighting coefficient, Q k is the expected value of process noise;

[0044]

[0045] Among them, a is the mean adjustment coefficient, and its value range is 10 -4 ≤a≤1, b is the coefficient of variation, b=2;

[0046] S4.3. Based on the observation equation, the observed quantities are first fused as observation vectors with a unified sampling frequency. In the update step based on the observed quantities, (2n+1) sigma points of the observed quantities are also generated. The prediction steps of the observed quantities are calculated according to the following formulas:

[0047]

[0048] in, is the test estimate corresponding to each sigma point, is an estimate of the test volume;

[0049] Then calculate the new covariance matrix of the kth recursive step, the expression is:

[0050]

[0051] Among them, P yy,k is the updated value of the covariance matrix, R k is the expected value of the observation noise;

[0052]

[0053] Among them, P xy,k is the cross covariance matrix estimate;

[0054] S4.4. Obtain the gain K of the kth recursive step by calculating the Kalman filter gain k , the expression is:

[0055]

[0056] Then based on the gain K of the kth recursive step k Calculate the sigma point and covariance matrix updated in the kth step, the expression is:

[0057]

[0058]

[0059] Beneficial effects of the present invention:

[0060] The present invention provides a parameter identification method for a wind turbine hybrid tower structure based on multi-source test information. The method has a convenient operation process and high precision. On the basis of ensuring that the acceleration signal is correct and stable, other signals can be resampled and reconstructed, and then used as observation data for structural parameter identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a flow chart of a method for parameter identification of a wind turbine hybrid tower structure based on multi-source test information according to the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the specific embodiments described are only part of the embodiments of the present invention, rather than all of the specific embodiments. The components of the specific embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations, and the present invention can also have other embodiments.

[0063] Therefore, the following detailed description of the specific embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected specific embodiments of the present invention. Based on the specific embodiments of the present invention, all other specific embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0064] In order to further understand the content, features and effects of the present invention, the following specific implementation methods are given as examples, and the attached Figure 1 The detailed instructions are as follows:

[0065] Embodiment 1:

[0066] A parameter identification method for a wind turbine hybrid tower structure based on multi-source test information comprises the following steps:

[0067] S1. Establish a finite element model of the wind turbine hybrid tower structure based on the drawings and field data, and then solve the finite element model based on the improved unscented Kalman filter to obtain the discretized state equation;

[0068] Furthermore, the drawings and on-site data in step S1 include geometric dimensions, steel structure specifications, elastic modulus, concrete grade and elastic modulus of the wind power hybrid tower structure;

[0069] The expression of the state equation after discretization is solved as follows:

[0070] X k =f(X k-1 ,U k-1 )+w k-1

[0071] Among them, X k represents the state quantity of the kth recursive step, f represents the time-varying nonlinear function including the state quantities X and U, and U k-1 represents the system input matrix of the k-1th recursive step, w k-1 represents the noise of the k-1th recursive step.

[0072] S2. The response time history information of the wind power tower structure is collected through displacement sensors, inclination sensors, static level settlement sensors, strain sensors, and acceleration sensors. The measured data is used as observation data to establish an observation equation, and then the observation equation is solved based on the improved unscented Kalman filter to obtain a discretized observation equation;

[0073] Further, the reaction time history information in step S2 includes the vibration of the mixed tower structure, the inclination of the hub, the foundation settlement, the strain of the mixed tower structure, and the cable force collected by the displacement sensor, the inclination sensor, the static level settlement sensor, the strain sensor, and the acceleration sensor;

[0074] Design and determine the monitoring plan for the mixed tower structure, including the fast acquisition time history curve corresponding to acceleration, and the slow acquisition time history curve corresponding to displacement transmission, inclination transmission, uneven settlement, and strain;

[0075] The expression of the discrete observation equation obtained by solving the observation equation based on the improved unscented Kalman filter is:

[0076] Y k =h(X k ,U k )+v k

[0077] Among them, Y k represents the observed value of the kth recursive step, h represents the functional relationship between the state quantity, external load and the observed value, and U k represents the system input matrix of the kth recursive step, v k represents the observed disturbance at the k-th recursive step.

[0078] S3. Based on the observed data obtained in step S2, given the data time window length t, based on the acceleration signal spectrum characteristics and the highest frequency f of the signal max , the highest order determination method of Chebyshev decomposition is used to solve the observation reconstruction parameter α, and then the observation data other than acceleration are reconstructed using the least squares method and Chebyshev polynomials to obtain the reconstructed observation data;

[0079] Furthermore, the specific implementation method of step S3 includes the following steps:

[0080] S3.1. The acceleration signal in the observation data obtained in step S2 Decompose, the decomposition formula is:

[0081]

[0082] Where T represents the Chebyshev polynomial basis function, W represents the Chebyshev polynomial basis function coefficient, and N f represents the order of the highest Chebyshev polynomial;

[0083] The highest Chebyshev polynomial order; the highest frequency f of the acceleration signal passed max OK, the expression is:

[0084] N f =α×5×t

[0085] Among them, α is the observation reconstruction parameter, and t is the length of the data time window;

[0086] S3.2. The observation reconstruction parameter α is calculated by the least squares method and with the number of periods np in the signal, the expression is:

[0087]

[0088] np=f max ×t;

[0089] S3.3. Determine α based on acceleration, and for other fast or slow test signals, perform least squares fitting based on the acceleration sampling frequency, the test signal and the Chebyshev polynomial to reconstruct the observation data other than the acceleration to obtain the reconstructed observation data.

[0090] S4. Using the reconstructed observation data, the structural parameters of the wind turbine hybrid tower are identified based on the improved adaptive unscented Kalman filter method.

[0091] Furthermore, the specific implementation method of step S4 includes the following steps:

[0092] S4.1. Set the initial values ​​of the parameters to be identified and the initial covariance matrix, the expression is:

[0093]

[0094] in, represents the initial state quantity to be identified, X0 represents the initial actual state quantity, the actual state quantity is obtained by calculation, E represents the expected value, is the initial covariance matrix;

[0095] The state quantity of each time step is recursively identified by UKF. During the recursive identification process, each time step will generate (2n+1) sigma points for UT transformation. The transformation process is shown in the following formula:

[0096]

[0097] in, represents the sigma point of the k-1th recursive step, j = 1, 2, ..., n, n represents the dimension of the state quantity, λ represents the gain, and λ depends on the dimension of the state quantity and the statistical distribution law of the state quantity;

[0098] S4.2. Based on the state equation, each sigma point at each time point is solved according to the state equation to obtain the estimated value of the sigma point at the next time point, as shown in the following formula:

[0099]

[0100] Make a priori estimates and calculate the state quantity mean and covariance matrix by the following formula:

[0101]

[0102] in, is the predicted value of the sigma sampling point, is the first weighting coefficient;

[0103]

[0104] in, is the prediction covariance, is the second weighting coefficient, Q k is the expected value of process noise;

[0105]

[0106] Among them, a is the mean adjustment coefficient, and its value range is 10 -4 ≤a≤1, b is the coefficient of variation, b=2;

[0107] S4.3. Based on the observation equation, the observed quantities are first fused as observation vectors with a unified sampling frequency. In the update step based on the observed quantities, (2n+1) sigma points of the observed quantities are also generated. The prediction steps of the observed quantities are calculated according to the following formulas:

[0108]

[0109] in, is the test estimate corresponding to each sigma point, is an estimate of the test volume;

[0110] Then calculate the new covariance matrix of the kth recursive step, the expression is:

[0111]

[0112] Among them, P yy,k is the updated value of the covariance matrix, R kis the expected value of the observation noise;

[0113]

[0114] Among them, P xy,k is the cross covariance matrix estimate;

[0115] S4.4. Obtain the gain K of the kth recursive step by calculating the Kalman filter gain k , the expression is:

[0116]

[0117] Then based on the gain K of the kth recursive step k Calculate the sigma point and covariance matrix updated in the kth step, the expression is:

[0118]

[0119]

[0120] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0121] Although the present application has been described above with reference to specific embodiments, various modifications may be made thereto and parts thereof may be replaced with equivalents without departing from the scope of the present application. In particular, as long as there is no structural conflict, the various features in the specific embodiments disclosed in the present application may be used in combination with each other in any manner, and the fact that these combinations are not exhaustively described in this specification is only for the sake of omitting space and saving resources. Therefore, the present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A parameter identification method for a wind turbine hybrid tower structure based on multi-source test information, characterized in that: The steps include: S1. Establish a finite element model of the wind turbine hybrid tower structure based on the drawings and field data, and then solve the finite element model based on the improved unscented Kalman filter to obtain the discretized state equation; S2. The response time history information of the wind power tower structure is collected through displacement sensors, inclination sensors, static level settlement sensors, strain sensors, and acceleration sensors. The measured data is used as observation data to establish an observation equation, and then the observation equation is solved based on the improved unscented Kalman filter to obtain a discretized observation equation; S3. Based on the observed data obtained in step S2, given the data time window length t, based on the acceleration signal spectrum characteristics and the highest frequency f of the signal max , the highest order determination method of Chebyshev decomposition is used to solve the observation reconstruction parameter α, and then the observation data other than acceleration are reconstructed using the least squares method and Chebyshev polynomials to obtain the reconstructed observation data; S4. Using the reconstructed observation data, the structural parameters of the wind turbine hybrid tower are identified based on the improved adaptive unscented Kalman filter method.

2. A parameter identification method for a wind turbine hybrid tower structure based on multi-source test information according to claim 1, characterized in that: The drawings and on-site data in step S1 include geometric dimensions, steel structure specifications, elastic modulus, concrete grade and elastic modulus of the wind power tower structure; The expression of the state equation after discretization is solved as follows: X k =f(X k-1 ,U k-1 )+w k-1 Among them, X k represents the state quantity of the kth recursive step, f represents the time-varying nonlinear function including the state quantities X and U, and U k-1 represents the system input matrix of the k-1th recursive step, w k-1 represents the noise of the k-1th recursive step.

3. The parameter identification method of a wind turbine hybrid tower structure based on multi-source test information according to claim 2 is characterized in that: The reaction time history information in step S2 includes the vibration of the mixed tower structure, the inclination angle of the hub, the foundation settlement, the strain of the mixed tower structure, and the cable force collected by the displacement sensor, the inclination sensor, the static level settlement sensor, the strain sensor, and the acceleration sensor; Design and determine the monitoring plan for the mixed tower structure, including the fast acquisition time history curve corresponding to acceleration, and the slow acquisition time history curve corresponding to displacement transmission, inclination transmission, uneven settlement, and strain; The expression of the discrete observation equation obtained by solving the observation equation based on the improved unscented Kalman filter is: Y k =h(X k ,U k )+v k Among them, Y k represents the observed value of the kth recursive step, h represents the functional relationship between the state quantity, external load and the observed value, U k represents the system input matrix of the kth recursive step, v k represents the observed disturbance at the k-th recursive step.

4. The method for parameter identification of a wind turbine hybrid tower structure based on multi-source test information according to claim 3 is characterized in that: The specific implementation method of step S3 includes the following steps: S3.

1. The acceleration signal in the observation data obtained in step S2 Decompose, the decomposition formula is: Where T represents the Chebyshev polynomial basis function, W represents the Chebyshev polynomial basis function coefficient, and N f represents the order of the highest Chebyshev polynomial; The highest Chebyshev polynomial order; the highest frequency f of the acceleration signal passed max OK, the expression is: N f =α×5×t Among them, α is the observation reconstruction parameter, and t is the length of the data time window; S3.

2. The observation reconstruction parameter α is calculated by the least squares method and with the number of periods np in the signal, the expression is: np=f max ×t; S3.

3. Determine α based on acceleration, and for other fast or slow test signals, perform least squares fitting based on the acceleration sampling frequency, the test signal and the Chebyshev polynomial to reconstruct the observation data other than the acceleration to obtain the reconstructed observation data.

5. The method for parameter identification of a wind turbine hybrid tower structure based on multi-source test information according to claim 4 is characterized in that: The specific implementation method of step S4 includes the following steps: S4.

1. Set the initial values ​​of the parameters to be identified and the initial covariance matrix, the expression is: in, represents the initial state quantity to be identified, X0 represents the initial actual state quantity, the actual state quantity is obtained by calculation, E represents the expected value, is the initial covariance matrix; The state quantity of each time step is recursively identified by UKF. During the recursive identification process, each time step will generate (2n+1) sigma points for UT transformation. The transformation process is shown in the following formula: in, represents the sigma point of the k-1th recursive step, j = 1, 2, ..., n, n represents the dimension of the state quantity, λ represents the gain, and λ depends on the dimension of the state quantity and the statistical distribution law of the state quantity; S4.

2. Based on the state equation, each sigma point at each time point is solved according to the state equation to obtain the estimated value of the sigma point at the next time point, as shown in the following formula: Make a priori estimates and calculate the state quantity mean and covariance matrix by the following formula: in, is the predicted value of the sigma sampling point, is the first weighting coefficient; in, is the prediction covariance, is the second weighting coefficient, Q k is the expected value of process noise; Among them, a is the mean adjustment coefficient, and its value range is 10 -4 ≤a≤1, b is the coefficient of variation, b=2; S4.

3. Based on the observation equation, the observed quantities are first fused as observation vectors with a unified sampling frequency. In the update step based on the observed quantities, (2n+1) sigma points of the observed quantities are also generated. The prediction steps of the observed quantities are calculated according to the following formulas: in, is the test estimate corresponding to each sigma point, is the estimated value of the test quantity; and then the new covariance matrix of the kth recursive step is calculated, and the expression is: Among them, P yy,k is the updated value of the covariance matrix, R k is the expected value of the observation noise; Among them, P xy,k is the cross covariance matrix estimate; S4.

4. Obtain the gain K of the kth recursive step by calculating the Kalman filter gain k , the expression is: Then based on the gain K of the kth recursive step k Calculate the sigma point and covariance matrix updated in the kth step, the expression is:

Images