CN114742363A

CN114742363A - Energy efficiency state evaluation method, system and medium for wind turbine generator

Info

Publication number: CN114742363A
Application number: CN202210262342.5A
Authority: CN
Inventors: 朱俊杰; 任鑫; 杨玉中; 杨和康; 牛耘; 钟清; 郝龙; 彭安栋; 罗朝军
Original assignee: Huaneng Clean Energy Research Institute; Huaneng Dali Wind Power Co Ltd Eryuan Branch
Current assignee: Huaneng Clean Energy Research Institute; Huaneng Dali Wind Power Co Ltd Eryuan Branch
Priority date: 2022-03-16
Filing date: 2022-03-16
Publication date: 2022-07-12

Abstract

The application provides a method, a system and a medium for evaluating the energy efficiency state of a wind turbine generator, wherein the method comprises the following steps: analyzing an energy efficiency mechanism of the wind turbine generator, and establishing an energy efficiency state evaluation index system; acquiring historical data of a unit, and dividing the historical data into working conditions through a multi-step K mean value clustering algorithm; acquiring a sample set containing comprehensive indexes and characteristic parameters influencing the comprehensive indexes from historical data aiming at each working condition, and calculating the weight of the characteristic parameters through a support vector machine regression technology; determining a reference value and a threshold value of the characteristic parameters under each working condition according to historical data; and calculating the deviation of the characteristic parameters according to the actual value, the reference value and the threshold value of each characteristic parameter under the current working condition, and generating an energy efficiency state evaluation result of the unit according to the deviation and the weight of the characteristic parameters. The method can accurately evaluate the energy efficiency state of the wind turbine generator, is favorable for objectively and comprehensively knowing the current running condition of the wind turbine generator, and avoids faults.

Description

Energy efficiency state evaluation method, system and medium for wind turbine generator

Technical Field

The application relates to the technical field of wind power generation, in particular to a method, a system and a medium for evaluating the energy efficiency state of a wind turbine generator.

Background

Under the current situation of reducing carbon emission, wind power generation is attracting more and more attention as a main new energy technology. At present, the installed capacity of wind power is rapidly increased, the wind power accounts for continuously increasing in an electric power energy structure, the single-machine capacity of a wind power generation set is continuously increased, and along with the relative maturity of onshore wind power technology, in view of the huge development value of offshore wind resources, the construction of a large-capacity wind power generation field is being developed from the land to the offshore and even the open sea, and the offshore wind power generation set faces higher construction and installation cost, worse operation environment and higher operation and maintenance cost. Therefore, in order to improve the power generation capacity and reduce the operation and maintenance cost while ensuring the stable and healthy operation of the wind turbine generator, the evaluation of the energy efficiency state of the wind turbine generator becomes one of the key steps of the long-term healthy development of wind energy.

In the related art, an energy efficiency threshold is generally preset in a subjective assignment mode or an objective assignment mode combined with historical data, then the energy efficiency state of the wind turbine generator is monitored, and the monitored energy efficiency state data is compared with the preset threshold to evaluate the energy efficiency. However, the energy efficiency evaluation method in the related art does not comprehensively consider the energy efficiency loss mechanism of the wind turbine generator, is relatively comprehensive in evaluation method, and is not suitable for the complex operation condition of the wind turbine generator, so that the accuracy of the current energy efficiency evaluation is low.

Disclosure of Invention

The present application is directed to solving, at least to some extent, one of the technical problems in the related art.

Therefore, a first object of the present application is to provide an energy efficiency state evaluation method for a wind turbine generator, where a data mining technology is used to obtain a reference value, a threshold value and a weight of each energy efficiency characteristic parameter from historical data, and an energy efficiency evaluation result is obtained according to a deviation degree of real-time data from the reference value and the weight. Therefore, the energy efficiency state of the wind turbine generator can be accurately evaluated, the current operation state of the wind turbine generator can be objectively and comprehensively known, the operation state can be timely adjusted, and the probability of the fault of the wind turbine generator is reduced.

The second purpose of the application is to provide an energy efficiency state evaluation system of the wind turbine generator;

a third object of the present application is to propose a non-transitory computer-readable storage medium.

In order to achieve the above object, an embodiment of a first aspect of the present application is to provide a method for evaluating an energy efficiency state of a wind turbine, where the method includes the following steps:

comprehensively analyzing the energy efficiency mechanism of the wind turbine generator, and establishing an energy efficiency state evaluation index system containing multi-stage economic indexes;

acquiring historical data of a wind turbine generator to be evaluated, establishing a working condition characteristic set according to the energy efficiency state evaluation index system, and performing working condition division on the historical data through a multi-step K mean value clustering algorithm;

acquiring comprehensive indexes containing each sample from the historical data according to each working condition, and a sample set which is determined according to the energy efficiency state evaluation index system and influences all characteristic parameters of each comprehensive index, calculating the sensitivity coefficient of each characteristic parameter corresponding to each comprehensive index through a support vector machine regression technology, and determining the weight of the characteristic parameters according to the sensitivity coefficient;

determining a reference value and a threshold value of each characteristic parameter under each working condition according to the historical data;

the method comprises the steps of obtaining real-time data of the wind turbine generator to be evaluated, calculating the deviation degree of each characteristic parameter according to the actual value, the reference value and the threshold value of each characteristic parameter under the current working condition, and generating the energy efficiency state evaluation result of the wind turbine generator to be evaluated according to the deviation degree and the weight of each characteristic parameter under the current working condition.

Optionally, in an embodiment of the application, calculating a sensitivity coefficient of each feature parameter corresponding to the comprehensive index through a support vector machine regression technique includes: establishing a regression function representing the mapping relation between the comprehensive index and the characteristic parameters based on a regression model of a support vector machine; introducing an insensitive loss function to convert the regression function into a first function, and introducing a structure risk function to convert the first function into a second function; converting the second function into a third function having a first set of constraints by solving an optimal solution for the convex optimization dual of the second function; processing the first set of constraints by a Lagrangian multiplier to convert the third function to a fourth function; converting the fourth function into a fifth function through a support vector machine regression model algorithm, wherein the fifth function has a second constraint condition set, and a decision function of the support vector machine regression model is obtained by calculating an optimal solution of the fifth function; and solving a first-order partial derivative of any characteristic parameter by the decision function to obtain a sensitivity coefficient of any characteristic parameter.

Optionally, in an embodiment of the present application, determining the reference value and the threshold value of each of the characteristic parameters under each operating condition according to the historical data includes: dividing the historical data into training sample sets with corresponding numbers according to working condition division results; setting the energy efficiency index value of each characteristic parameter as a corresponding reference value when the energy efficiency level of the wind turbine generator to be evaluated is highest in each training sample set; calculating the average value and standard deviation of each characteristic parameter under each working condition, and calculating the threshold value of each characteristic parameter by the following formula:

wherein, X_iIs a threshold value for the characteristic parameter(s),

σ is the standard deviation, which is the mean of the characteristic parameters.

Optionally, in an embodiment of the present application, calculating a deviation degree of each of the characteristic parameters according to the actual value, the reference value, and the threshold value of each of the characteristic parameters under the current operating condition includes: calculating the deviation value of each characteristic parameter according to the actual value and the reference value of each characteristic parameter; calculating a degree of deviation of each of the characteristic parameters by the following formula:

where μ is the degree of deviation, d_iAs deviation values of characteristic parameters, X_iIs a threshold value for the characteristic parameter(s),

is a reference value of the characteristic parameter.

Optionally, in an embodiment of the present application, determining the weight of the feature parameter according to the sensitivity coefficient includes: and performing regularization processing on the sensitivity coefficient, and labeling the regularized sensitivity coefficient.

Optionally, in an embodiment of the present application, the energy efficiency state evaluation result of the wind turbine generator to be evaluated is calculated by the following formula:

wherein F is an energy efficiency state evaluation value, n is the number of characteristic parameters, and S_iIs the weight of the characteristic parameter, mu_iIs the degree of deviation of the characteristic parameter.

In order to achieve the above object, an embodiment of the second aspect of the present application further provides an energy efficiency state evaluation system for a wind turbine generator, including the following modules:

the establishing module is used for comprehensively analyzing the energy efficiency mechanism of the wind turbine generator and establishing an energy efficiency state evaluation index system containing multi-stage economic indexes;

the dividing module is used for acquiring historical data of the wind turbine generator to be evaluated, establishing a working condition characteristic set according to the energy efficiency state evaluation index system, and dividing the working condition of the historical data through a multi-step K mean value clustering algorithm;

the first determining module is used for acquiring comprehensive indexes containing each sample from the historical data according to each working condition, determining a sample set which influences all characteristic parameters of each comprehensive index according to the energy efficiency state evaluation index system, calculating the sensitivity coefficient of each characteristic parameter corresponding to each comprehensive index through a support vector machine regression technology, and determining the weight of the characteristic parameters according to the sensitivity coefficient;

the second determining module is used for determining a reference value and a threshold value of each characteristic parameter under each working condition according to the historical data;

the generating module is used for acquiring real-time data of the wind turbine generator to be evaluated, calculating the deviation degree of each characteristic parameter according to the actual value, the reference value and the threshold value of each characteristic parameter under the current working condition, and acquiring the energy efficiency state evaluation result of the wind turbine generator to be evaluated according to the deviation degree and the weight of each characteristic parameter under the current working condition.

Optionally, in an embodiment of the present application, the first determining module is specifically configured to: establishing a regression function representing the mapping relation between the comprehensive index and the characteristic parameters based on a regression model of a support vector machine; introducing an insensitive loss function to convert the regression function into a first function, and introducing a structure risk function to convert the first function into a second function; converting the second function to a third function having a first set of constraints based on a convex optimization and a dual algorithm; processing the first set of constraints by a Lagrangian multiplier to convert the third function to a fourth function; solving the fourth function to convert the fourth function into a fifth function representing the optimization problem of the support vector machine regression model, wherein the fifth function has a second constraint condition set, and a decision function of the support vector machine regression model is obtained by calculating an optimal solution of the fifth function; and solving a first-order partial derivative of any characteristic parameter by the decision function to obtain a sensitivity coefficient of any characteristic parameter.

Optionally, in an embodiment of the application, the second determining module is specifically configured to: dividing the historical data into training sample sets with corresponding numbers according to working condition division results; setting the energy efficiency index value of each characteristic parameter as a corresponding reference value when the energy efficiency level of the wind turbine generator to be evaluated is highest in each training sample set; calculating the average value and the standard deviation of each characteristic parameter under each working condition, and calculating the threshold value of each characteristic parameter by the following formula:

wherein, X_iIs a threshold value for the characteristic parameter(s),

In order to implement the foregoing embodiments, an embodiment of the third aspect of the present application further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the energy efficiency state evaluation method for a wind turbine generator in the foregoing embodiments.

The technical scheme provided by the embodiment of the application at least has the following beneficial effects: according to the method, the influence factors of the energy efficiency level of the wind turbine generator are determined through researching the related mechanism knowledge of the wind turbine generator, so that a set of complete energy efficiency evaluation index system is constructed, and the comprehensive evaluation of the energy efficiency state of the top-down scientific system can be performed in the subsequent evaluation based on the system. Then, a data mining technology is adopted to obtain a reference value, a threshold value and a weight of each energy efficiency characteristic parameter from historical data, and an energy efficiency evaluation result is obtained according to the deviation degree of the real-time data and the reference value and the weight. The method comprises the steps of dividing a plurality of working conditions according to the characteristics of complexity and multilateral performance of the operation working conditions of the wind turbine generator, evaluating energy efficiency states according to different working conditions, considering the change of the energy efficiency states of the wind turbine generator under different environmental factors, and improving the objectivity and practicability of energy efficiency evaluation. And moreover, the weight values of the characteristic parameters are obtained by adopting a sensitivity analysis method based on the support vector machine regression technology, so that the method can be suitable for accurate calculation of the weight values under different working conditions. Therefore, the energy efficiency state of the wind turbine generator can be accurately evaluated, the current operation state of the wind turbine generator can be objectively and comprehensively known, the operation state can be timely adjusted, and the probability of the fault of the wind turbine generator is reduced.

Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.

Drawings

The foregoing and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:

fig. 1 is a flowchart of an energy efficiency state evaluation method for a wind turbine generator set according to an embodiment of the present application;

fig. 2 is a schematic diagram of an energy efficiency state index system of a wind turbine generator set according to an embodiment of the present application;

fig. 3 is a flowchart of a method for calculating a sensitivity coefficient of a characteristic parameter according to an embodiment of the present application;

fig. 4 is a schematic flow chart of a specific method for evaluating the energy efficiency state of a wind turbine generator set according to an embodiment of the present application;

fig. 5 is a schematic structural diagram of an energy efficiency state evaluation system of a wind turbine generator set according to an embodiment of the present application.

Detailed Description

Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.

The energy efficiency evaluation and diagnosis method and system for the wind turbine generator set provided by the embodiment of the invention are described in detail below with reference to the accompanying drawings.

Fig. 1 is a flowchart of a method for evaluating an energy efficiency state of a wind turbine generator according to an embodiment of the present application, and as shown in fig. 1, the method includes the following steps:

and S101, comprehensively analyzing the energy efficiency mechanism of the wind turbine generator, and establishing an energy efficiency state evaluation index system containing multi-stage economic indexes.

It should be noted that the energy efficiency state of the wind turbine is the embodiment of the economy of the wind turbine in the operating state, and the economy of the current operating state of the wind turbine can be embodied by evaluating the energy efficiency of the operating state of the wind turbine, so that the influence factors influencing the energy efficiency level of the wind turbine can be determined as the economy index of the wind turbine.

Specifically, the research of the related mechanism knowledge of the wind turbine generator system is firstly carried out, the energy efficiency mechanism of the wind turbine generator system is comprehensively analyzed, so that the power loss index of the wind turbine generator system is determined, and a systematic and complete energy efficiency state evaluation index system is constructed according to the relation among all indexes.

In an embodiment of the present application, according to an analysis flow of "system-subsystem-device", mechanism analysis is sequentially performed on each system and device of a wind turbine generator from each system in the wind turbine generator, then to the subsystem in each system, and finally to each device in the subsystem, so as to determine a power loss index of each system, and determine a device index of the power loss index of each system. And then combing and inducing the logic relationship among different energy efficiency state evaluation indexes, and determining the energy efficiency state evaluation index contained in each level, thereby constructing an index system influencing the energy efficiency state of the wind turbine generator as shown in figure 2, wherein the energy efficiency state evaluation index system comprises three levels of economic indexes, namely a system level, a subsystem level and a device level.

For example, as shown in fig. 2, the system level includes a wind energy capture system, a transmission system, and the like, the wind energy capture system includes economic indicators of a wind wheel subsystem, an encoder subsystem, and a nacelle subsystem, the economic indicators of the subsystems are represented by a wind wheel loss, an encoder angle deviation, and a nacelle position deviation, and the equipment-level economic indicators include a blade angle of a blade device and an impeller rotation speed of the blade device under the wind wheel subsystem.

It should be further noted that the wind turbine generator energy efficiency state evaluation index system shown in fig. 2 is only an example, and as shown in fig. 2, the system may further include other systems, subsystems, and the like, and the specifically constructed energy efficiency state evaluation index system is determined according to an actual mechanism analysis result. Therefore, a systematic and complete energy efficiency evaluation index system is constructed, and the energy efficiency state can be comprehensively evaluated from top to bottom and scientifically systematically when evaluation is performed subsequently based on the system.

Step S102, historical data of the wind turbine generator to be evaluated are obtained, a working condition characteristic set is established according to an energy efficiency state evaluation index system, and working condition division is carried out on the historical data through a multi-step K mean value clustering algorithm.

The historical data is historical operation data of a target wind turbine generator to be subjected to energy efficiency evaluation at present, and comprises measured values of operation parameters of various systems and equipment in the wind turbine generator in different operation time periods.

It should be noted that, under the working conditions of environmental factors such as different wind speeds, wind directions, environmental temperatures and the like and coping strategies adopted by the units to adapt to the environment, the energy efficiency states of the wind turbine units are different, so that in order to further improve the objectivity and effectiveness of the energy efficiency state evaluation of the wind turbine units, the working conditions are divided before the energy efficiency evaluation is performed, and the energy efficiency state evaluation is performed on the following working conditions.

In specific implementation, in an embodiment of the application, historical Data of a wind turbine to be evaluated is obtained in different manners, as an example, equipment of the wind turbine running on site may be monitored through a preset Data Acquisition And monitoring Control System (SCADA), Data of each system And equipment in the wind turbine at different times is acquired in real time, for example, Data change conditions of a cabin position And an impeller rotation speed in the example are monitored, And meanwhile, environmental information such as a wind speed And a wind direction of an environment where the wind turbine is currently located may also be acquired through the SCADA, so that subsequent working condition division is facilitated. Then, historical data monitored by the SCADA are obtained, for example, after the SCADA sends the monitored historical data to a rear-end cloud platform or a local database of the system, when energy efficiency evaluation needs to be carried out, the historical data of the wind turbine generator is called from the corresponding database or cloud platform according to a mode of storing data by the SCADA.

Further, a set of operating condition characteristics is established. The operating condition feature set comprises operating condition distinguishing parameters for distinguishing different operating conditions, and the parameters can be regarded as boundary conditions for dividing the operating conditions. In order to ensure the accuracy of the working condition division, the operating working condition division parameters should be reasonably selected, and the selected parameters are ensured to be the operating parameters which have direct or indirect influence on the state characteristic parameters of the wind turbine generator. In the embodiment of the application, a working condition feature set is established according to an energy efficiency state evaluation index system, and specifically, each index of a device bottom layer in the energy efficiency state evaluation index system, which can distinguish working conditions, and an environmental factor parameter are used as operation working condition distinguishing parameters. For example, the operating condition feature set established by the present application may include: wind speed, wind direction, torque, rotational speed, ambient temperature, and the like.

Furthermore, based on the established working condition characteristic set, the historical operating data of the unit is divided by a multi-step K-means clustering algorithm. Specifically, the optimal clustering number is determined according to the working condition feature set, the historical operating data is firstly subjected to preliminary division through a K-means clustering algorithm, an initial clustering number K is determined, K clustering points are obtained through the preliminary division, corresponding clusters are respectively generated, then the similarity between the preliminarily determined K clustering points and other sample data in the corresponding clusters is evaluated, and the final clustering number is determined according to the similarity. In order to further improve the accuracy of the working condition division, the working conditions are further divided in a multi-step division mode, namely after the final clustering number is preliminarily determined, the characteristic variables and the sample set of the K-means clustering algorithm are changed, and the multi-step working condition division is carried out in the same mode of calculating the clustering number K.

Therefore, according to the working condition division results, historical operation data monitored by the SCADA can be divided correspondingly, training sample sets with corresponding numbers are generated, and the parameter values for determining energy efficiency state evaluation under different working conditions are convenient to evaluate.

Step S103, acquiring comprehensive indexes containing each sample from historical data according to each working condition, and a sample set which is determined according to an energy efficiency state evaluation index system and influences all characteristic parameters of each comprehensive index, calculating the sensitivity coefficient of each characteristic parameter corresponding to each comprehensive index through a support vector machine regression technology, and determining the weight of the characteristic parameters according to the sensitivity coefficient.

The comprehensive index refers to an overall energy efficiency state index of the wind turbine generator, as can be seen from fig. 2, the comprehensive index is influenced by parameters of each level of the middle and bottom layers of the system, namely bottom layer indexes, and the indexes of each level are complex in relation and generally have an aliasing effect. The bottom layer index is a characteristic parameter corresponding to the comprehensive index, and the influence of the bottom layer index on the comprehensive index can be approximately regarded as a sensitivity coefficient between the comprehensive index in the energy efficiency state and the characteristic parameter.

It should be noted that, in the energy efficiency state evaluation method of the present application, a related method is adopted to evaluate the energy efficiency state of the unit system according to the difference value, i.e., the deviation degree, of the selected energy efficiency characteristic parameter from the reference value and the weight of each characteristic parameter. For the weight of the characteristic parameter, the error of the weight value set by the subjective weighting method in the related technology is large, which causes the inaccuracy of the energy efficiency state evaluation. The modeling of the mapping relation between the comprehensive indexes of the energy efficiency states and the characteristic parameters belongs to a complex regression problem, the energy efficiency states are different under different operating conditions in the operation process of the wind turbine generator, and the data volume of the operating parameters under some operating conditions is small, so that the regression under the condition of a small sample must be met when the weight of the characteristic parameters is determined. The nonlinear problem of the original space is converted into the linear problem of the high-dimensional space through nonlinear mapping, and the method is one of the characteristics of a support vector machine regression (SVR) model. Therefore, according to the historical data of the wind turbine generator, the weight value of the characteristic parameter is obtained by adopting a sensitivity analysis method based on the regression technology of the support vector machine.

In the specific implementation process, the weight of each characteristic parameter is determined under each working condition, because the historical data is divided into the working conditions in step 102, a sample set in a certain operation time period is obtained from the historical data for each working condition, wherein the sample set comprises a plurality of samples, each sample comprises a comprehensive index and all characteristic parameters corresponding to the comprehensive index, in the process of specifically obtaining the sample set, after the monitored comprehensive index of the unit is obtained, a bottom-layer index influencing the comprehensive index can be combed through the established energy efficiency state evaluation index system, and then data of the corresponding characteristic parameters are screened from the historical data, so that the sample set is generated. Then, the SVR model can be trained through a massive training sample set, the sensitivity coefficient of each characteristic parameter corresponding to the comprehensive index is calculated through the trained SVR model, and then the weight value is calculated according to the sensitivity coefficient.

As an example, in a certain operation time period under a certain condition, a sample set T { (x) of the operation of the wind turbine generator is obtained₁,y₁),(x₂,y₂),…,(x_l,y_l)}∈(X,Y)^lWherein x is_i＝{x_i1,x_i2,…x_in}^T∈RⁿThe vector of influence factors of the comprehensive index of the ith sample is i 1,2, … l, and l is the number of samples. x is the number of_ijJ is 1,2, …, n, n is the number of influencing factors. y is_iAnd E R is an observed value of the energy efficiency comprehensive index of the ith sample. The following mapping relationship is assumed to exist between the comprehensive index sample space and the influence factor sample space:

y＝f(x₁,x₂,…x_n)

then, let Δ x_jAnd deltay is the increment of the jth influencing factor and the increment of the comprehensive index. From the theory of differential science, the partial derivative represents the function change rate corresponding to the change of each independent variable, namely the sensitivity of the change. The present application therefore defines the sensitivity coefficient of a characteristic parameter as follows:

wherein, the first and the second end of the pipe are connected with each other,

the partial derivative of the energy efficiency comprehensive index to any characteristic parameter is shown, and the magnitude of the partial derivative can intuitively reflect the change rate, namely the sensitivity coefficient, of the comprehensive index to any characteristic parameter.

Furthermore, the sensitivity coefficient of the characteristic parameter is determined through a regression model of a support vector machine. In order to more clearly and specifically describe a specific implementation process of the key step of calculating the sensitivity coefficient of each feature parameter corresponding to the comprehensive index by using the support vector machine regression technology, an exemplary description is given below by using a method for calculating the sensitivity coefficient of a feature parameter, which is provided in an embodiment of the present application, and as shown in fig. 3, the method includes the following steps:

step S301, a regression function representing the mapping relation between the comprehensive indexes and the characteristic parameters is established based on a regression model of the support vector machine.

Specifically, in the support vector machine regression model, the regression problem of the mapping relationship between the comprehensive index and the characteristic parameter for determining the energy efficiency state is expressed as a function shown in the following formula:

wherein, omega is a weight vector,

b is the intercept, which is a non-linear mapping function.

Step S302, an insensitive loss function is introduced to convert the regression function into a first function, and a structure risk function is introduced to convert the first function into a second function.

Specifically, taking the regression function established in step S301 as an example, an insensitive loss function is introduced to convert the regression function into a first function as shown in the following formula:

by introducing an insensitive loss function, the robustness and the sparsity of the energy efficiency state estimation can be ensured.

Then, introducing a structure risk function converts the first function to a second function as shown in the following equation:

wherein a junction is introducedThe risk-forming function can satisfy the principle of minimizing the structural risk of the support vector machine, and in the formulas, y is an observed value of the energy efficiency comprehensive index obtained from historical data, namely a true value of the comprehensive index obtained through monitoring. In the second function, | ω | | non-volatile memory in the first term²The effect of (a) is to make the fitting function flatter to improve generalization capability and the error can be reduced by the second term in the function.

Step S303, the second function is converted into a third function having the first constraint set by solving an optimal solution of the convex optimization dual of the second function.

Specifically, the regression problem in step S301 is converted into a convex optimization dual problem, and the dual problem of linear programming is solved to obtain an optimal solution, so as to convert the second function into a third function having a first constraint group as shown in the following formula:

wherein the first constraint condition group is

Wherein C is a penalty parameter and C>0, the generalization capability of the model is reflected, when the C value is larger, the training error is smaller, the larger the C value is, the larger the test error is, i.e. overfitting is performed, so that the generalization capability of the model is poorer, and when the C value is too small, the training error and the test error are increased along with the reduction of the C value, i.e. underfitting is performed, so that the precision of the model is poorer, therefore, in the embodiment of the application, the optimization is adjusted and searched in the training of the SVR model, so that the proper C value is selected to ensure the robustness of the model. Epsilon is an insensitive coefficient, the model controls the tolerance of the model to the noise amplitude in the data and the sparsity of the model solution of the support vector machine, when the value of epsilon is smaller, the number of the support vectors is larger, the model precision is higher, the solution sparsity is lower, but the complexity of the model training is increased; xi, xi^*More than or equal to 0 is relaxation factor.

Step S304, the first constraint condition set is processed by the lagrangian multiplier to convert the third function into a fourth function.

Specifically, a lagrange multiplier is introduced to solve an optimal solution of the convex optimization dual problem under constraint conditions, each constraint condition in the first constraint condition group is processed by introducing the lagrange multiplier, and the problem is solved, so that the third function is converted into a fourth function as shown in the following formula:

wherein alpha is_iIs a lagrange multiplier.

Step S305, converting the fourth function into a fifth function through a regression model algorithm of the support vector machine, wherein the fifth function has a second constraint condition group, and obtaining a decision function of the regression model of the support vector machine by calculating the optimal solution of the fifth function.

Specifically, the optimization problem is converted by an SVR algorithm, and the fourth function is converted into a fifth function having a second constraint set as shown in the following formula:

wherein the second constraint condition group is

Further, an optimal solution of the fifth function under the second constraint condition set is calculated, so that the decision function in the SVR model is represented by the following formula:

wherein, K (x)_iX) is a kernel function that can map a non-linear low-dimensional space to a high-dimensional linear feature spaceAnd solving the nonlinear regression is realized.

It should be noted that, in this example, a support vector regression algorithm is determined by introducing a loss function and a kernel function, and the kernel function maps an original nonlinear space to a high-dimensional linear space, so as to convert a nonlinear problem into a linear problem and solve the linear problem. The structure of the feature space depends on the form of the kernel function and its parameters, which determine the level of the nonlinear processing capability of the support vector machine and the accuracy of the constructed regression function, so the key to perform regression modeling with the support vector machine is to select a suitable kernel function, and in this example, the kernel function is the kernel function K in the application embodiment because the number of parameters to be determined by the linear kernel function is small and the linear kernel function can reduce the numerical computation difficulty for the feature space with higher dimension. The specific form of the linear kernel function is shown as the following formula:

furthermore, substituting the kernel function into the decision function can obtain a mathematical relationship model between the comprehensive index of the energy efficiency of the wind turbine generator and the characteristic parameter as follows:

step S306, solving a first-order partial derivative of any characteristic parameter by the decision function, and obtaining a sensitivity coefficient of any characteristic parameter.

Specifically, a mathematical relationship model between the comprehensive index of the unit energy efficiency state and the characteristic parameters is solved through steps S301 to S305, and the dependent variable y in the mathematical relationship function between the comprehensive index of the unit energy efficiency state and the characteristic parameters is adjusted to any independent variable, i.e. any characteristic parameter x, by using the above partial derivative sensitivity analysis method based on the support vector machine in this step_k(k is 1,2, …, n) the first order partial derivative is obtained, i.e. the number in step S305Substituting the chemical model into the sensitivity coefficient definition formula of the characteristic parameters to calculate partial derivatives, so as to obtain the following formula:

in this equation, for a linear kernel function, the derivation is as follows:

substituting the partial derivatives of the linear kernel function into a formula for solving first-order partial derivatives, and finishing to obtain a calculation formula of the sensitivity coefficient of the characteristic parameter, wherein the calculation formula is as follows:

wherein the first partial derivative s of the dependent variable to the independent variable_kNamely the sensitivity coefficient of the comprehensive index of the energy efficiency state of the wind turbine generator to the kth characteristic parameter. The sensitivity coefficient is a unit number with positive and negative values, wherein the positive and negative values represent positive and negative correlations of the characteristic parameters to the decision attributes.

Therefore, the sensitivity coefficient of any characteristic parameter can be calculated by the calculation method of the sensitivity coefficient of the characteristic parameter, and the first-order partial derivative of each characteristic parameter is obtained by the decision function sequentially through the method, so that the sensitivity coefficient of each characteristic parameter of the comprehensive index can be obtained.

Furthermore, after the sensitivity coefficient of the characteristic parameter is calculated, the sensitivity coefficient has positive and negative values, but the weight value has no negative value, so that the sensitivity coefficient is converted, and the weight of the characteristic parameter is calculated according to the sensitivity coefficient.

In an embodiment of the present application, when determining the weight of the feature parameter according to the sensitivity coefficient, the sensitivity coefficient is regularized first, and then the regularized sensitivity coefficient is labeled. Specifically, the susceptibility coefficient is first normalized by the following equation:

then normalized by the following formula:

thus, the sensitivity coefficient S is obtained by sequentially performing the above-described processing on the sensitivity coefficient of each characteristic parameter_kSince (k — 1,2, …, n) reflects the relative degree of influence of each feature parameter on the overall performance indicator, a weight matrix as shown below can be finally generated as a weight value of the feature parameter:

W＝[S₁,S₂,…,S_n]^T。

and 104, determining a reference value and a threshold value of each characteristic parameter under each working condition according to the historical data.

The reference value is a parameter value of a characteristic parameter when the wind turbine generator runs in the optimal energy efficiency running state under the corresponding boundary condition of a specific working condition, and the reference value is a standard for performing energy efficiency evaluation. The threshold is the largest value after taking the absolute value from the historical maximum value or the historical minimum value of the characteristic parameter under the operating condition.

In an embodiment of the application, when the reference value and the threshold value of each characteristic parameter under each working condition are determined according to historical data, the method includes the following steps of dividing the historical data into training sample sets with corresponding numbers according to a working condition division result, setting an energy efficiency index value of each characteristic parameter as a corresponding reference value when the energy efficiency level of a wind turbine generator to be evaluated is the highest in each training sample set, calculating the average value and the standard deviation of each characteristic parameter under each working condition, and calculating the threshold value of each characteristic parameter through the following formula:

wherein, X_iIs a threshold value for the characteristic parameter(s),

Specifically, the reference value is calculated by mining historical data information, and the reference values of the characteristic parameters under different working conditions are changed, and the working conditions are divided in step 102, so that in a corresponding training sample set generated after the working conditions are divided, when the energy efficiency level of the wind turbine generator is determined to be the highest, namely the wind energy utilization rate is the highest, by comparing the size of each data value in the sample set, the energy efficiency index value of each characteristic parameter is used as the corresponding reference value of each characteristic parameter under the working condition, and the reference value determined by the application is the value when the energy efficiency level of the wind turbine generator is the highest. It should be noted that, according to the present application, a plurality of sample sets, for example, sample sets in a plurality of time periods under a certain working condition, are obtained, so as to obtain massive historical data, and the accuracy of the determined reference value is improved by enriching the number of samples. Further, according to the value of each characteristic parameter in the corresponding sample set under each working condition, the average value and the standard deviation of each characteristic parameter under each working condition are calculated and then are obtained

A threshold value for each characteristic parameter is calculated.

And 105, acquiring real-time data of the wind turbine generator to be evaluated, calculating the deviation degree of each characteristic parameter according to the actual value, the reference value and the threshold value of each characteristic parameter under the current working condition, and generating an energy efficiency state evaluation result of the wind turbine generator to be evaluated according to the deviation degree and the weight of each characteristic parameter under the current working condition.

Specifically, after the historical data is mined, the weight, the reference value and the threshold value of each characteristic parameter of the wind turbine generator to be evaluated are determined, the current real-time operation data of the wind turbine generator is acquired by monitoring the current operation state of the wind turbine generator through an SCADA (supervisory control and data acquisition) system, and the actual value of the characteristic parameter under the current working condition is acquired. And then calculating the deviation degree of each characteristic parameter according to the actual value, the reference value and the threshold value of each characteristic parameter under the current working condition, and reflecting the deviation of the bottom layer characteristic parameters through the deviation degree.

As one possible implementation manner, when calculating the deviation of each feature parameter, the deviation value of each feature parameter may be calculated according to the actual value and the reference value of each feature parameter, specifically, the deviation value of each feature parameter is obtained by subtracting the corresponding reference value from the actual value of each feature parameter. Then, the degree of deviation of each characteristic parameter is calculated by the following formula:

is a reference value of the characteristic parameter.

Further, after the deviation degree of each characteristic parameter is calculated, the energy efficiency state of the unit is evaluated according to the deviation degree and the weight of each characteristic parameter.

Specifically, firstly, a characteristic parameter set x ═ x { x } composed of each characteristic parameter monitored by the unit to be tested under the current working condition is determined₁,x₂,…,x_k}^TThe current deviation degree of each characteristic parameter constitutes a deviation degree set of μ ═ μ₁,μ₂,…,μ_n}^TAnd a weight set W [ S ] composed of weights of the respective characteristic parameters previously calculated by the SVR model₁,S₂,…,S_n]^TAnd then calculating the energy efficiency state evaluation result of the wind turbine generator to be evaluated through the following formula:

The calculated F is a number in an interval from 0 to 1, the current energy efficiency state of the wind turbine generator is rated according to the value of the F, the poorer the energy efficiency state of the wind turbine generator system is represented as the value of the F is closer to 0, and the higher the energy efficiency state of the wind turbine generator system is represented as the value of the F is closer to 1. When the energy efficiency state of the unit is poor, the evaluation result can be generated into alarm information and sent to relevant operation and maintenance personnel, and the operation and maintenance personnel are reminded to pay attention to whether the failure occurs in the wind turbine unit or not so as to reduce the probability of the failure of the unit.

Therefore, the economy of the current running state of the wind turbine generator is reflected through the energy efficiency evaluation of the running state of the wind turbine generator. The method specifically researches related mechanism knowledge of the wind turbine generator, determines influence factors influencing the energy efficiency level of the wind turbine generator on the basis, and determines economic indexes. And constructing an energy efficiency evaluation index system according to a characteristic parameter selection principle, acquiring a reference value, a threshold value and a weight of each energy efficiency characteristic parameter from historical data by adopting a data mining technology, and acquiring an energy efficiency evaluation result according to the deviation degree of the real-time data and the reference value and the weight.

In summary, according to the energy efficiency state evaluation method for the wind turbine generator, influence factors of the energy efficiency level of the wind turbine generator are determined through research on related mechanism knowledge of the wind turbine generator, so that a set of complete energy efficiency evaluation index system is constructed, and comprehensive evaluation can be performed on energy efficiency states of a top-down scientific system during subsequent evaluation based on the system. Then, a data mining technology is adopted to obtain a reference value, a threshold value and a weight of each energy efficiency characteristic parameter from historical data, and an energy efficiency evaluation result is obtained according to the deviation degree of the real-time data and the reference value and the weight. The method comprises the steps of dividing a plurality of working conditions according to the characteristics of complexity and multilateral performance of the operation working conditions of the wind turbine generator, evaluating energy efficiency states according to different working conditions, considering the change of the energy efficiency states of the wind turbine generator under different environmental factors, and improving the objectivity and practicability of energy efficiency evaluation. And moreover, the weight value of the characteristic parameter is obtained by adopting a sensitivity analysis method based on the support vector machine regression technology, so that the method can be suitable for accurately calculating the weight values under different working conditions. Therefore, the method can accurately evaluate the energy efficiency state of the wind turbine generator, is favorable for objectively and comprehensively knowing the current running state of the wind turbine generator, timely adjusts the running state, and reduces the probability of the wind turbine generator failing.

In order to more clearly describe the flow of the method for evaluating the energy efficiency state of the wind turbine generator according to the embodiment of the present application, a specific embodiment for evaluating the energy efficiency state of the wind turbine generator is described in detail below. Fig. 4 is a schematic flow chart of a specific method for evaluating the energy efficiency state of a wind turbine generator set according to an embodiment of the present application. As shown in fig. 4, the method comprises the steps of:

step S401, obtaining historical data of the unit.

And step S402, dividing working conditions.

In this step, since the reference value is obtained by mining the historical data information, a basic condition is needed, that is, different working conditions are determined, and the weight of each index is determined under each working condition in this embodiment, so that the working conditions are divided first. Due to the influence of wind power, machinery, electricity, ambient temperature and the like, the operation condition of the wind turbine generator has the characteristics of complexity and multilaterality, and the reasonable selection of characteristic parameters of the operation condition is the premise of realizing the division of the condition. The selected parameters are operation parameters which directly or indirectly influence the state characteristic parameters of the fan, and the invention establishes a working condition characteristic set u (wind speed, wind direction, torque, rotating speed and environment temperature) by taking each index of the bottom layer of the energy efficiency state index system equipment as the characteristic parameters. And performing working condition division on the historical data of the operating working conditions by adopting a multi-step K mean value clustering algorithm.

Step S403, a regression model is established.

Step S404, determining the sensitivity coefficient.

In step S405, a weight coefficient is calculated.

In the above steps, in order to find out the individual influence of each bottom layer index on the comprehensive index, i.e., the overall energy efficiency level, a sensitivity coefficient analysis model between the comprehensive index of the energy efficiency state and its influence factor (bottom layer index) under each working condition needs to be constructed. In this embodiment, a feature parameter sensitivity coefficient analysis model is constructed according to a support vector machine regression model, and then the sensitivity coefficient of each feature parameter is determined according to the sensitivity coefficient analysis model. The specific process is as follows: training an SVR regression model, analyzing errors, solving sensitivity coefficients by a deviation method, normalizing by regularization, and obtaining weight coefficients. The specific implementation method of each step in the process may refer to the related description in the above embodiments, and is not described herein again.

In step S406, a reference value is determined.

In the step, the SCADA operation and monitoring data are correspondingly divided by using the working condition division result, and training sample sets with corresponding numbers are generated. And determining that the energy efficiency level of the wind power generation unit is the highest under all working conditions through mass historical data, namely, each energy efficiency index value is used as a reference value when the wind energy utilization rate is the highest, and the reference value is the value obtained when the energy efficiency level of the wind power generation unit is the highest.

It should be noted that the order of step S406 and step S403 is not limited here, that is, the weight coefficient and the reference value may be calculated at the same time.

And step S407, acquiring real-time operation data of the unit.

In step S408, the degree of deviation is calculated.

In this step, since the specific expression of the energy efficiency state is the deviation of the bottom layer characteristic parameters, the deviation values of the characteristic parameters can be calculated according to the previously calculated reference values of the characteristic parameters and the monitored real-time data of the characteristic parameters, and then the deviation degrees of the characteristic parameters are defined according to the deviation values. The specific implementation method of this step may refer to the related description in the above embodiments, and is not described herein again.

In step S409, an energy efficiency evaluation is generated.

Therefore, in the embodiment, the energy efficiency state of the system is evaluated according to the difference value of the selected energy efficiency characteristic parameters from the reference value and the weight of each characteristic parameter. In this method, two basic elements required for energy efficiency evaluation are a degree of deviation and a weight of a characteristic parameter, and the degree of deviation is calculated based on a reference value and a threshold value of the characteristic parameter, and mining is usually performed using historical data. For the weight, in this embodiment, a sensitivity analysis method based on a support vector machine regression technique is adopted to obtain the weight value of the characteristic parameter according to the historical data of the wind turbine. And finally, obtaining an energy efficiency state evaluation result of the wind turbine generator by combining the characteristic parameter deviation value and the weight, and providing visual reference for the operation optimization of the wind turbine generator.

In order to implement the foregoing embodiments, the present application further provides an energy efficiency state evaluation system for a wind turbine generator, and fig. 5 is a schematic structural diagram of the energy efficiency state evaluation system for a wind turbine generator provided in the embodiments of the present application, and as shown in fig. 5, the system includes an establishing module 100, a dividing module 200, a first determining module 300, a second determining module 400, and a generating module 500.

The establishing module 100 is used for comprehensively analyzing the energy efficiency mechanism of the wind turbine generator and establishing an energy efficiency state evaluation index system containing multi-stage economic indexes.

The dividing module 200 is used for acquiring historical data of the wind turbine generator to be evaluated, establishing a working condition characteristic set according to an energy efficiency state evaluation index system, and dividing the working conditions of the historical data through a multi-step K-means clustering algorithm.

The first determining module 300 is configured to obtain, for each operating condition, a comprehensive index including each sample from the historical data, and a sample set that affects all feature parameters of each comprehensive index and is determined according to the energy efficiency state evaluation index system, calculate a sensitivity coefficient of each feature parameter corresponding to the comprehensive index by using a support vector machine regression technique, and determine a weight of the feature parameter according to the sensitivity coefficient.

The second determination module 400 is configured to determine a reference value and a threshold value of each characteristic parameter under each condition according to the historical data.

The generation module 500 is configured to obtain real-time data of the wind turbine generator to be evaluated, calculate a deviation degree of each characteristic parameter according to an actual value, a reference value, and a threshold of each characteristic parameter under a current working condition, and obtain an energy efficiency state evaluation result of the wind turbine generator to be evaluated according to the deviation degree and the weight of each characteristic parameter under the current working condition.

Optionally, in an embodiment of the present application, the first determining module 300 is specifically configured to: establishing a regression function representing the mapping relation between the comprehensive indexes and the characteristic parameters based on a regression model of a support vector machine; introducing an insensitive loss function to convert the regression function into a first function, and introducing a structure risk function to convert the first function into a second function; converting the second function into a third function having a first set of constraints by solving an optimal solution for the convex optimization dual of the second function; processing the first set of constraint conditions by a lagrange multiplier to convert the third function to a fourth function; converting the fourth function into a fifth function through a regression model algorithm of the support vector machine, wherein the fifth function has a second constraint condition group, and a decision function of the regression model of the support vector machine is obtained by calculating the optimal solution of the fifth function; and solving a first-order partial derivative of any characteristic parameter by the decision function to obtain a sensitivity coefficient of any characteristic parameter.

Optionally, in an embodiment of the present application, the second determining module 400 is specifically configured to: dividing historical data into training sample sets with corresponding numbers according to working condition division results; setting the energy efficiency index value of each characteristic parameter as a corresponding reference value when the energy efficiency level of the wind turbine generator to be evaluated is highest in each training sample set; calculating the average value and standard deviation of each characteristic parameter under each working condition, and calculating the threshold value of each characteristic parameter by the following formula:

wherein, X_iIs a threshold value for the characteristic parameter(s),

is a special oneThe mean values of the parameters were characterized, σ is the standard deviation.

Optionally, in an embodiment of the present application, the generating module 500 is specifically configured to: calculating the deviation value of each characteristic parameter according to the actual value and the reference value of each characteristic parameter; the degree of deviation of each characteristic parameter is calculated by the following formula:

where μ is the degree of deviation, d_iIs a deviation value, X, of a characteristic parameter_iIs a threshold value for the characteristic parameter(s),

is a reference value of the characteristic parameter.

Optionally, in an embodiment of the present application, the first determining module 300 is further configured to: and carrying out regularization processing on the sensitivity coefficient, and carrying out labeling processing on the sensitivity coefficient after regularization processing.

Optionally, in an embodiment of the present application, the generating module 500 is specifically configured to calculate an energy efficiency state evaluation result of the wind turbine generator to be evaluated according to the following formula:

It should be noted that the explanation of the embodiment of the energy efficiency state evaluation method for the wind turbine generator is also applicable to the system of the embodiment, and details are not repeated here

To sum up, the energy efficiency state evaluation system of the wind turbine generator system in the embodiment of the application adopts a data mining technology to obtain the reference value, the threshold value and the weight of each energy efficiency characteristic parameter from historical data, and obtains the energy efficiency evaluation result according to the deviation degree of the real-time data and the reference value and the weight. The method comprises the steps of dividing a plurality of working conditions according to the characteristics of complexity and multilateral performance of the operation working conditions of the wind turbine generator, evaluating energy efficiency states according to different working conditions, considering the change of the energy efficiency states of the wind turbine generator under different environmental factors, and improving the objectivity and practicability of energy efficiency evaluation. And moreover, the weight value of the characteristic parameter is obtained by adopting a sensitivity analysis method based on the support vector machine regression technology, so that the method can be suitable for accurately calculating the weight values under different working conditions. Therefore, the system can accurately evaluate the energy efficiency state of the wind turbine generator, is favorable for objectively and comprehensively knowing the current running state of the wind turbine generator, timely adjusts the running state and reduces the probability of the wind turbine generator failing.

In order to implement the foregoing embodiments, the present application further proposes a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for evaluating the energy efficiency state of the wind turbine generator set is implemented as described in any one of the foregoing embodiments.

In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.

Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or to implicitly indicate the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present application, "plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.

Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present application in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present application.

The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.

It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.

It will be understood by those skilled in the art that all or part of the steps carried out in the method of implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and the program, when executed, includes one or a combination of the steps of the method embodiments.

In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.

The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc. Although embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application, and that variations, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.

Claims

1. The method for evaluating the energy efficiency state of the wind turbine generator is characterized by comprising the following steps of:

comprehensively analyzing an energy efficiency mechanism of the wind turbine generator, and establishing an energy efficiency state evaluation index system containing multi-stage economic indexes;

2. The method of claim 1, wherein the calculating the sensitivity coefficient of each feature parameter corresponding to the comprehensive index by a support vector machine regression technique comprises:

establishing a regression function representing the mapping relation between the comprehensive index and the characteristic parameters based on a regression model of a support vector machine;

introducing an insensitive loss function to convert the regression function into a first function, and introducing a structure risk function to convert the first function into a second function;

converting the second function into a third function having a first set of constraints by solving an optimal solution for the convex optimization dual of the second function;

processing the first set of constraints by a Lagrangian multiplier to convert the third function to a fourth function;

converting the fourth function into a fifth function through a support vector machine regression model algorithm, wherein the fifth function has a second constraint condition set, and a decision function of the support vector machine regression model is obtained by calculating an optimal solution of the fifth function;

and solving a first-order partial derivative of any characteristic parameter by the decision function to obtain a sensitivity coefficient of any characteristic parameter.

3. The method according to claim 1 or 2, wherein the determining of the reference value and the threshold value of each characteristic parameter under each working condition according to the historical data comprises:

dividing the historical data into training sample sets with corresponding numbers according to working condition division results;

setting the energy efficiency index value of each characteristic parameter as a corresponding reference value when the energy efficiency level of the wind turbine generator to be evaluated is highest in each training sample set;

calculating the average value and standard deviation of each characteristic parameter under each working condition, and calculating the threshold value of each characteristic parameter by the following formula:

wherein, X_iIs a threshold value for the characteristic parameter(s),

4. The method according to claim 3, wherein the calculating the deviation degree of each characteristic parameter according to the actual value, the reference value and the threshold value of each characteristic parameter under the current working condition comprises:

calculating the deviation value of each characteristic parameter according to the actual value and the reference value of each characteristic parameter;

calculating a degree of deviation of each of the characteristic parameters by the following formula:

is a reference value of the characteristic parameter.

5. The method of claim 1, wherein determining the weight of the characteristic parameter according to the sensitivity coefficient comprises:

and performing regularization processing on the sensitivity coefficient, and labeling the regularized sensitivity coefficient.

6. The method according to claim 1, characterized in that the energy efficiency state evaluation result of the wind turbine to be evaluated is calculated by the following formula:

7. The energy efficiency state evaluation system of the wind turbine generator is characterized by comprising the following components:

8. The system of claim 7, wherein the first determining module is specifically configured to:

converting the second function to a third function having a first set of constraints based on a convex optimization and a dual algorithm;

solving the fourth function to convert the fourth function into a fifth function representing the optimization problem of the support vector machine regression model, wherein the fifth function has a second constraint condition set, and a decision function of the support vector machine regression model is obtained by calculating an optimal solution of the fifth function;

9. The system according to claim 7 or 8, wherein the second determining module is specifically configured to:

setting the energy efficiency index value of each characteristic parameter as a corresponding reference value when the energy efficiency level of the wind turbine generator to be evaluated is the highest in each training sample set;

wherein, X_iIs a threshold value for the characteristic parameter(s),

10. A non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for evaluating the energy efficiency status of a wind turbine according to any one of claims 1 to 6.