Fan load shedding method based on multi-damper optimization control

Abstract

The invention provides a fan load shedding method based on multi-damper optimization control, and the method comprises the steps: building a multi-damper fan model containing stroke limitation through a Lagrange equation based on a key mode of a fan; unknown parameters of the fan model are identified based on an LM algorithm, and the verification accuracy of the FAST model is compared; structural parameters of the multi-damping system are used as optimization variables, a multi-population genetic algorithm is selected to be matched with a multi-damper fan model containing stroke limitation, and parameter optimization is achieved; and on the basis of fan load simulation software, the effectiveness of the fan load shedding method based on multi-damper optimization control is verified. The problem that a single damper can only tune specific frequency can be solved, the multiple dampers are arranged in the fan cabin and the tower, the frequency tuning range is increased, the vibration peak value caused by structural resonance is reduced, the vibration control efficiency is improved, and the service life of the fan structure is prolonged.

Description

Technical Field

[0001] The invention relates to the technical field of wind power generation, and in particular to a wind turbine load reduction method based on multi-damper optimization control. Background Art

[0002] Wind turbines are operating in complex environments for a long time, and are subject to the influence of harsh factors such as strong winds and earthquakes. The tower is prone to severe vibration and deformation. With the development trend of large-scale wind turbines, the increase in tower height has increased the overall center of gravity, and the structural stability of the wind turbine itself has been significantly reduced. It is more likely to generate large fatigue loads and limit loads at the tower bottom and blade roots, reducing the service life of the wind turbine. Therefore, how to suppress tower vibration has become one of the key issues that need to be solved in the current wind power industry.

[0003] In terms of wind turbine structure control, domestic and foreign scholars have conducted many studies on single dampers and achieved good control effects. Most studies use intelligent algorithms and FAST models to optimize damper structure parameters.

[0004] However, a single damper is too sensitive to the vibration frequency of the fan structure. Once the frequency changes, the vibration reduction effect will drop significantly. In addition, the load reduction effect of the damper depends on its structural parameters. During the optimization of the damper structural parameters, the optimal configuration of the damper structural parameters cannot be quickly found due to the long simulation cycle of the FAST model and the strong nonlinear characteristics of the optimization area. Summary of the invention

[0005] The purpose of the present invention is to provide a wind turbine load reduction method based on multi-damper optimization control, which can solve the problem that a single damper can only be tuned to a specific frequency. Multiple dampers are configured in the wind turbine nacelle and tower to increase the frequency tuning range, reduce the vibration peak value caused by structural resonance, improve the vibration control efficiency, and extend the service life of the wind turbine structure.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is: a fan load reduction method based on multi-damper optimization control, comprising the following steps: Step 1: Based on the key modes of the fan, a fan model with multiple dampers and stroke restrictions is established using the Lagrange equation; Step 2: Identify the unknown parameters of the wind turbine model based on the LM algorithm and compare the accuracy of the FAST model verification; Step 3: Taking the structural parameters of the multi-damping system as optimization variables, select a multi-population genetic algorithm and a fan model with multiple dampers containing stroke constraints to achieve parameter optimization; Step 4: Based on the fan load simulation software, verify the effectiveness of the fan load reduction method based on multi-damper optimization control.

[0007] Preferably, the fan model with multiple dampers containing stroke restrictions in step 1 is expressed as follows: Among them, I t ,I p Represents the moment of inertia of the tower and platform; m t 、m p 、m T1 、m T2 Respectively represent the masses of the tower, platform, cabin damper and tower damper; g represents the acceleration of gravity; θ t ,θ p Respectively represent the angles of the tower and platform deviating from the vertical plane; x T1 、x T2 Respectively represent the horizontal distances from the nacelle damper and the tower damper to the initial point; k t , k p , k T1 , k T2 denote the stiffness coefficients of the tower, platform, cabin damper and tower damper, respectively, d t d p d T1 d T2 are the damping coefficients of the tower, platform, cabin damper and tower damper respectively; R t , R p , R T1 , R T2 They represent the distances from the tower mass center, platform mass center, cabin damper mass center and tower damper mass center to the vertical plane where the tower bottom is located; k T1lim , k T2lim Respectively represent the rigidity coefficients of the nacelle travel limiter and the tower travel limiter; d T1lim d T2lim are the damping coefficients of the nacelle travel limiter and the tower travel limiter respectively; x T1lim 、x T2lim Represent the distances from the nacelle damper and the tower damper to the travel limit point respectively.

[0008] Preferably, the parameters of the fan model to be identified in step 2 are: U = [d p d t k p k t I p I t ] T .

[0009] Preferably, in step 2, the sum of square errors S(U) between the fan model and the FAST model response is established, and the LM algorithm is used to identify unknown parameters.

[0010] Preferably, the error sum of squares S(U) is defined as: Where m and n represent the number and data length of output parameters respectively; y j (t i ) indicates t i The j-th output of the FAST model at time f j (t i ,U) represents t i The jth output of the wind turbine model at time Y j (t i ), F j (t i ,U) are y j (t i ), f j (t i ,U) in vector form; W represents the weight matrix.

[0011] Preferably, the identification of unknown parameters of the wind turbine model in step 2 comprises the following steps: Step 21, close the multi-damper system and open the tower degree of freedom of the FAST model; Step 22, set the initial pitch angle of the platform to 8 degrees, the initial setting of the wind rotor to the rated speed of 11.4 rpm, and other initial parameters to zero; Step 23, using the FAST model to obtain the output response of the tower front and rear displacement; Step 24: Use the LM algorithm to find the optimal values ​​of the fan model structural parameters.

[0012] Preferably, the parameter optimization in step 3 comprises the following steps: Step 31, set the parameters of the multi-population genetic algorithm: the number of populations, the number of individuals in each population, the probability range of random crossover and mutation, the minimum number of generations to keep the optimal individual, and initialize the population; Step 32: Calculate the fitness values ​​of all individuals in each population according to the fitness function; Step 33, the number of iterations is increased by 1, the population is updated through selection, crossover, mutation, and recombination functions, and the fitness values ​​of all individuals in each population are calculated; Step 34: exchange the best individuals of all sub-populations through the immigration function, and select the best individuals of the best population; Step 35, record the best individual in the current best population, and determine whether the fitness value of the current best individual is the same as the fitness value of the best individual of the previous generation. If they are the same, the best individual is kept algebraically plus 1. If they are different, the best individual is kept algebraically cleared and the best individual is updated; Determine whether the stopping condition is met: if the stopping condition is met, then exit the optimization process and use the global optimal solution as the optimal parameter; if the stopping condition is not met, then return to step 3; The stopping condition includes: whether the optimal individual retention generation number reaches a set value.

[0013] Preferably, the fitness function of the multi-population genetic algorithm is expressed as: Among them, σt represents the standard deviation of the longitudinal displacement of the wind turbine tower when there is a multi-damping system; σs represents the standard deviation of the longitudinal displacement of the wind turbine tower when there is no multi-damping system; η represents the standard deviation suppression rate of the longitudinal displacement of the wind turbine tower.

[0014] The beneficial effects of the present invention are:

[0015] This solution solves the problem that a single damper can only be tuned to a specific frequency. Multiple dampers are configured in the wind turbine nacelle and tower to increase the frequency tuning range, reduce the vibration peak caused by structural resonance, improve vibration control efficiency, and extend the service life of the wind turbine structure.

[0016] This solution overcomes the defects of the FAST model, such as long simulation time and complex calculation process. Based on the key modes of the fan, a fan model with multiple dampers containing stroke restrictions is established, which accurately simulates the output response of the FAST model, simplifies the fan modeling process, and improves the efficiency of load calculation.

[0017] This solution realizes the rapid optimization of the structural parameters of the multi-damping system. It uses a multi-population genetic algorithm and a fan model with multiple dampers with stroke constraints to quickly find the optimal configuration of the structural parameters of the multi-damping system, improve the efficiency of parameter optimization, and maximize the load reduction effect of the multi-damping system.

[0018] This solution effectively suppresses the structural vibration amplitude of the fan under complex working conditions, ensures the stability of the fan structure, and improves the output power quality of the fan. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 The figure is a flow chart of the wind turbine load reduction method of the present invention.

[0021] Figure 2 It is a schematic diagram of a fan model of the present invention.

[0022] Figure 3It is a comparison diagram of the front and rear displacement of the tower and the pitch angle of the platform of the present invention.

[0023] Figure 4 This is a multi-population genetic algorithm optimization flow chart of the present invention.

[0024] Figure 5 It is a time domain diagram of the simulated turbulent wind of the present invention.

[0025] Figure 6 It is a comparison diagram of the power spectrum density of the front and rear displacements of the tower of the present invention.

[0026] Figure 7 It is a time domain comparison diagram of the platform pitch angle of the present invention. DETAILED DESCRIPTION

[0027] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0028] The present invention discloses a fan load reduction method based on multi-damper optimization control. Figure 1 As shown, the embodiment includes the following steps: Step 1: Based on the key modes of the Barge type fan, a fan model with multiple dampers and stroke restrictions is established using the Lagrange equation; Step 2: Identify the unknown parameters of the wind turbine model based on the LM algorithm and compare the accuracy of the FAST model verification; Step 3: Taking the structural parameters of the multi-damping system as optimization variables, select a multi-population genetic algorithm and a fan model with multiple dampers containing stroke constraints to achieve parameter optimization; Step 4: Based on the fan load simulation software, verify the effectiveness of the fan load reduction method based on multi-damper optimization control.

[0029] The traditional wind turbine modeling process generally uses the Kane method, but this method requires the selection of appropriate generalized coordinates for the system, so the modeling process is more complicated and reduces the computational efficiency. The Lagrange equation has a unified form for arbitrary coordinates, which effectively simplifies the modeling process and can directly incorporate constraints into the modeling process without introducing additional constraint equations. Therefore, a wind turbine model is established based on the Lagrange equation. According to previous studies, the structural stability of the Barge type wind turbine mainly depends on the platform pitch mode, the tower front and rear bending mode and the multi-damping system. The blades, hub and other structures have little effect on the stability of the wind turbine. Therefore, a wind turbine model is established based on the above key degrees of freedom.

[0030] The schematic diagram of the fan model with multiple dampers and stroke restrictions is shown in Figure 2 As shown in the figure, the fan model with multiple dampers containing stroke limitation is expressed as follows: Among them, It ,I p Represents the moment of inertia of the tower and platform; m t 、m p 、m T1 、m T2 Respectively represent the masses of the tower, platform, cabin damper and tower damper; g represents the acceleration of gravity; θ t ,θ p Respectively represent the angles of the tower and platform deviating from the vertical plane; x T1 、x T2 Respectively represent the horizontal distances from the nacelle damper and the tower damper to the initial point; k t , k p , k T1 , k T2 denote the stiffness coefficients of the tower, platform, cabin damper and tower damper, respectively, d t d p d T1 d T2 are the damping coefficients of the tower, platform, cabin damper and tower damper respectively; R t , R p , R T1 , R T2 They represent the distances from the tower mass center, platform mass center, cabin damper mass center and tower damper mass center to the vertical plane where the tower bottom is located; k T1lim , k T2lim Respectively represent the rigidity coefficients of the nacelle travel limiter and the tower travel limiter; d T1lim d T2lim are the damping coefficients of the nacelle travel limiter and the tower travel limiter respectively; x T1lim 、x T2lim Represent the distances from the nacelle damper and the tower damper to the travel limit point respectively.

[0031] The platform inclination angle of the Barge type wind turbine is no more than 10 degrees under normal working conditions. Therefore, the secθ, sinθ, cosθ and tanθ in formula (1) are approximated with a small angle to obtain formula (2):

[0032] The known parameters of the fan structure model with multiple dampers with stroke restrictions are consistent with the FAST model. Parameter identification is used to find the unknown parameters in the dynamic model. There are a total of 6 parameters to be identified, which are expressed as: U=[d p d t k p k t I p I t ]T (3)

[0033] The LM algorithm is an optimization algorithm for solving nonlinear least squares problems. It combines the advantages of the gradient descent algorithm and the Gauss-Newton algorithm. In the optimization process, the weight matrix is ​​introduced to balance the rapidity and stability of iteration. The present invention establishes the error sum of squares S(U) between the fan model and the FAST model response, and uses the LM algorithm to identify unknown parameters. S(U) is defined as:

[0034] Where m and n represent the number and data length of output parameters respectively; y j (t i ) indicates t i The j-th output of the FAST model at time f j (t i ,U) represents t i The jth output of the wind turbine model at time Y j (t i ), F j (t i ,U) are y j (t i ), f j (t i ,U) in vector form; W represents the weight matrix.

[0035] Due to the nonlinear characteristics of the wind turbine model and the wide solution space, the optimization result can easily fall into the local optimal solution, and if the initial value of the unknown parameter differs too much from the optimal value, it will significantly affect the iteration speed and the accuracy of the result. Therefore, the initial value needs to be tested.

[0036] The identification and verification process of structural parameters is as follows: Step 1: First, turn off the multi-damper system and turn on the tower degrees of freedom of the FAST model; Step 2: Next, set the initial pitch angle of the platform to 8 degrees, the initial rotor speed to 11.4 rpm, and other initial parameters to zero; Step 3: Use the FAST model to obtain the response data of the tower's front and rear displacements; Step 4: Use the LM algorithm to find the optimal values ​​of the fan model structural parameters.

[0037] The initial guess and identification results are shown in Table 1. Table 1 Initial values ​​of unknown parameters of the kinetic model and identification results parameter Initial Value Estimated value <![CDATA[d p / (N·s / m)]]> <![CDATA[2.5×10 6 ]]> <![CDATA[4.78×10 6 ]]> <![CDATA[d t / (N·s / m)]]> <![CDATA[2.5×10 6 ]]> <![CDATA[3.65×10 6 ]]> <![CDATA[k p / (N / m)]]> <![CDATA[5×10 8 ]]> <![CDATA[8.12×10 8 ]]> <![CDATA[k t / (N / m)]]> <![CDATA[5×10 8 ]]> <![CDATA[2.96×10 9 <!-- 5 -->]]> <![CDATA[I p / (kg·m 2 )]]> <![CDATA[1×10 8 ]]> <![CDATA[2.45×10 8 ]]> <![CDATA[I t / (kg·m 2 )]]> <![CDATA[1×10 8 ]]> <![CDATA[3.62×10 8 ]]>

[0038] Figure 3The time domain comparison of the tower front and rear displacement and platform pitch angle of the wind turbine model and the FAST model when the disturbance is set to an initial platform pitch angle of 8 degrees. It can be found that the wind turbine model and the FAST model response results fit well.

[0039] The configuration of structural parameters in the multi-damping system plays a decisive role in the load reduction effect. The fitness function is designed according to the fan structural model with multi-damper and stroke restriction established above. The structural parameters of the multi-damper system are used as optimization variables, and the intelligent algorithm and the fan model with multi-damper and stroke restriction are selected to achieve parameter optimization.

[0040] The optimization scheme here can be optimization algorithms such as genetic algorithm, particle swarm algorithm, grey wolf algorithm, etc., and those skilled in the art can select according to actual needs. The following is an explanation using a multi-population genetic algorithm as an example.

[0041] Reference Figure 4 , taking the structural parameters of the nacelle damper and the tower damper as optimization variables, the steps of optimizing the parameters using the multi-population genetic algorithm are as follows: Step 1: Set the parameters of the multi-population genetic algorithm: the number of populations, the number of individuals in each population, the probability range of random crossover and mutation, the minimum number of generations to keep the optimal individual, and initialize the population; Step 2: Calculate the fitness values ​​of all individuals in each population according to the fitness function; Step 3: The number of iterations is increased by 1, the population is updated through selection, crossover, mutation, and recombination functions, and the fitness values ​​of all individuals in each population are calculated; Step 4: Exchange the best individuals of all subpopulations through the immigration function, and select the best individuals of the best population; Step 5: Record the best individual in the current best population, and determine whether the fitness value of the current best individual is the same as the fitness value of the best individual in the previous generation. If they are the same, the best individual is kept algebraically plus 1. If they are different, the best individual is kept algebraically cleared and the best individual is updated; Determine whether the stopping condition is met: if the stopping condition is met, then exit the optimization process and use the global optimal solution as the optimal parameter for the optimization process; if the stopping condition is not met, return to step 3; The stopping condition includes: whether the optimal individual retention generation number reaches a set value.

[0042] In complex environmental conditions, the longitudinal displacement of the Barge-type wind turbine tower will increase significantly, aggravating the ultimate load and fatigue load at the connection between the tower and the platform. Therefore, the longitudinal deviation amplitude of the tower is used as the optimization evaluation criterion, and the fitness function of the multi-population genetic algorithm is expressed as follows:

[0043] Among them, σ tRepresents the standard deviation of the longitudinal displacement of the wind turbine tower when there are multiple damping systems; σ s It represents the standard deviation of the longitudinal displacement of the wind turbine tower when there is no multi-damping system; η represents the standard deviation suppression rate of the longitudinal displacement of the wind turbine tower.

[0044] The optimization results of the structural parameters of the multi-damping system are shown in Table 2. Table 2 Optimization results of structural parameters of multi-damping system

[0045] Based on the fan load simulation software, the effectiveness of the fan load reduction method based on multi-damper optimization control is verified.

[0046] In this step, the fan load simulation software openFAST (Fatigue, Aerodynamics, Structures, and Turbulence) is used for simulation verification.

[0047] Barge type wind turbines are mainly affected by wind loads and wave loads in the working environment. For wind loads, the Kalman spectrum is selected as the wind spectrum model, and 18m / s turbulent wind is generated according to the IEC standard. For wave loads, the JONSWAP spectrum is selected as the simulated wave spectrum model, the effective wave height is selected as 4m, and the peak period is 11.8s. The simulation time is set to 630s, the sampling time is set to 0.00625s, and the output data of the wind turbine with large fluctuations in the first 30s of the simulation are removed. The simulation results are as follows: Figures 5 to 7 shown.

[0048] in, Figure 5 is the time domain diagram of turbulent wind, Figure 6 : is a comparison diagram of the power spectrum density of the longitudinal displacement of the tower under the traditional control strategy and the method of this embodiment, from which it can be seen that the multi-damping system significantly reduces the maximum peak value near 0.08 Hz. Figure 7 : is a comparison diagram of the platform pitch angle under the traditional control strategy and the method of this embodiment. It can be seen from the figure that the pitch motion amplitude of the fan platform is reduced by 20% after the multi-damper is placed.

[0049] This embodiment verifies the effectiveness of the wind turbine load reduction method based on MTMD optimization control by comparing the load suppression effects of the traditional control strategy and the method proposed in this embodiment on key parts.

[0050] It should be noted that the parts not described in detail in the above embodiments are all prior art.

[0051] The above description is only a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any technician familiar with the profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention should be covered within the protection scope of the present invention.

Claims

1. A fan load reduction method based on multi-damper optimization control, Features: The steps include: Step 1: Based on the key modes of the fan, a fan model with multiple dampers and stroke restrictions is established using the Lagrange equation; Step 2: Identify the unknown parameters of the wind turbine model based on the LM algorithm and compare the accuracy of the FAST model verification; Step 3: Taking the structural parameters of the multi-damping system as optimization variables, select a multi-population genetic algorithm and a fan model with multiple dampers containing stroke constraints to achieve parameter optimization; Step 4: Based on the fan load simulation software, verify the effectiveness of the fan load reduction method based on multi-damper optimization control.

2. A fan load reduction method based on multi-damper optimization control according to claim 1, Features: The fan model with multiple dampers containing stroke restrictions in step 1 is expressed as follows: Among them, I t ,I p Represents the moment of inertia of the tower and platform; m t 、m p 、m T1 、m T2 Respectively represent the masses of the tower, platform, cabin damper and tower damper; g represents the acceleration of gravity; θ t ,θ p Respectively represent the angles of the tower and platform deviating from the vertical plane; x T1 、x T2 Respectively represent the horizontal distances from the nacelle damper and the tower damper to the initial point; k t , k p , k T1 , k T2 denote the stiffness coefficients of the tower, platform, cabin damper and tower damper, respectively, d t ,d p ,d T1 ,d T2 are the damping coefficients of the tower, platform, cabin damper and tower damper respectively; R t , R p , R T1 , R T2 They represent the distances from the tower mass center, platform mass center, cabin damper mass center and tower damper mass center to the vertical plane where the tower bottom is located; k T1lim , k T2lim Respectively represent the rigidity coefficients of the nacelle travel limiter and the tower travel limiter; d T1lim ,d T2lim are the damping coefficients of the nacelle travel limiter and the tower travel limiter respectively; x T1lim 、x T2lim Represent the distances from the nacelle damper and the tower damper to the travel limit point respectively.

3. A fan load reduction method based on multi-damper optimization control according to claim 2, Features: The parameters that need to be identified for the fan model in step 2 are: U = [d p d t k p k t I p I t ] T .

4. A fan load reduction method based on multi-damper optimization control according to claim 3, Features: In step 2, the error sum of squares S(U) between the fan model and the FAST model response is established, and the LM algorithm is used to identify unknown parameters.

5. A fan load reduction method based on multi-damper optimization control according to claim 4, Features: The error sum of squares S(U) is defined as: Where m and n represent the number and data length of output parameters respectively; y j (t i ) indicates t i The j-th output of the FAST model at time f j (t i ,U) represents t i The jth output of the wind turbine model at time Y j (t i ), F j (t i ,U) are y j (t i ), f j (t i ,U) in vector form; W represents the weight matrix.

6. A fan load reduction method based on multi-damper optimization control according to claim 3, Features: The identification of unknown parameters of the wind turbine model in step 2 includes the following steps: Step 21, close the multi-damper system and open the tower degree of freedom of the FAST model; Step 22, set the initial pitch angle of the platform to 8 degrees, the initial setting of the wind rotor to the rated speed of 11.4 rpm, and other initial parameters to zero; Step 23, using the FAST model to obtain the output response of the tower front and rear displacement; Step 24: Use the LM algorithm to find the optimal values ​​of the fan model structural parameters.

7. A fan load reduction method based on multi-damper optimization control according to claim 1, Features: The parameter optimization in step 3 includes the following steps: Step 31, set the parameters of the multi-population genetic algorithm: the number of populations, the number of individuals in each population, the probability range of random crossover and mutation, the minimum number of generations to keep the optimal individual, and initialize the population; Step 32: Calculate the fitness values ​​of all individuals in each population according to the fitness function; Step 33, the number of iterations is increased by 1, the population is updated through selection, crossover, mutation, and recombination functions, and the fitness values ​​of all individuals in each population are calculated; Step 34: exchange the best individuals of all sub-populations through the immigration function, and select the best individuals of the best population; Step 35, record the best individual in the current best population, and determine whether the fitness value of the current best individual is the same as the fitness value of the best individual of the previous generation. If they are the same, the best individual is kept algebraically plus 1. If they are different, the best individual is kept algebraically cleared and the best individual is updated; Determine whether the stopping condition is met: if the stopping condition is met, the optimization process is exited and the global optimal solution is used as the optimal parameter; If the stop condition is not met, return to step 3; The stopping condition includes: whether the optimal individual retention generation number reaches a set value.

8. A fan load reduction method based on multi-damper optimization control according to claim 7, Features: The fitness function of the multi-population genetic algorithm is expressed as: Among them, σ t Represents the standard deviation of the longitudinal displacement of the wind turbine tower when there are multiple damping systems; σ s It represents the standard deviation of the longitudinal displacement of the wind turbine tower when there is no multi-damping system; η represents the standard deviation suppression rate of the longitudinal displacement of the wind turbine tower.

Images