Wind generating set health state assessment method and system based on SCADA (Supervisory Control And Data Acquisition)

Abstract

The invention discloses an SCADA-based wind turbine generator health state assessment method and system, and the method comprises the steps: testing a to-be-predicted wind turbine generator based on a prediction model obtained through the pre-training of a historical data set of the wind turbine generator and SCADA data of the to-be-predicted wind turbine generator in a healthy operation state, and obtaining a prediction error as a standard residual set; the actual power of the to-be-predicted wind turbine generator is obtained in real time, the residual error between the actual power and the predicted power of the to-be-predicted wind turbine generator is calculated, and the mahalanobis distance between the obtained residual error and the standard residual error set serves as the health state evaluation criterion of the to-be-predicted wind turbine generator to evaluate the health state of the to-be-predicted wind turbine generator. According to the method, the evaluation precision of the health state of the wind generating set can be greatly improved, and the evaluation workload of the health state of the wind generating set is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of wind turbine status assessment, and in particular relates to a wind turbine health status assessment method and system based on SCADA. Background Art

[0002] Timely and accurate understanding of the health status of wind turbines is of great significance for optimizing wind turbine control and operation and maintenance strategies, reducing wind turbine maintenance costs, and ensuring safe and efficient operation of wind turbines. There are two main health status assessment methods for wind turbines: model-based and data-driven. The model-based health status assessment method mainly establishes a model related to the health status of the subsystem of the wind turbine, and evaluates the health status of the subsystem based on the model. This method is difficult to implement, and the evaluation result depends on the accuracy of the established model. In addition, since the wind turbine subsystem model is established, the health status assessment of the entire machine cannot be achieved. The data-based health status assessment method is easy to implement and the evaluation results are accurate. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for evaluating the health status of a wind turbine generator set based on SCADA, so as to overcome the problems of poor accuracy and great difficulty in evaluating the health status of a wind turbine generator set in the prior art.

[0004] To achieve the above purpose, the technical solution adopted by the present invention is as follows: A method for evaluating the health status of a wind turbine generator set based on SCADA comprises the following steps: S1, the prediction model obtained by pre-training the historical data set of the wind turbine and the SCADA data of the wind turbine to be predicted in a healthy operating state are used to test the wind turbine to be predicted, and the prediction error is obtained as the standard residual set; S2, obtaining the actual power of the wind turbine to be predicted in real time, calculating the residual between the actual power of the wind turbine to be predicted and the predicted power, and evaluating the health status of the wind turbine to be predicted based on the Mahalanobis distance between the obtained residual and the standard residual set as the health status evaluation benchmark of the wind turbine to be predicted.

[0005] Preferably, the prediction model is a prediction model that takes the output power of the wind turbine as a prediction target and takes characteristic variables related to power as input.

[0006] Preferably, the quartile method is used to eliminate abnormal data of each characteristic variable in the wind turbine data acquisition and monitoring control system to obtain a historical data set of the wind turbine, and the maximum relevance minimum redundancy algorithm is used to select power-related characteristic variables.

[0007] Preferably, first, the data below the cut-in wind speed and with negative power values are excluded, and then the quartile method is used to preprocess the historical SCADA data of the wind turbine to exclude outliers.

[0008] Preferably, the maximum relevance and minimum redundancy algorithm is used to select the power-related features, and the mutual information between each feature variable and power is calculated according to Equation (2): (2) In Equation (2), 、 、 are the marginal probability density and joint probability density of the random variables 、 respectively; Calculate the maximum correlation index between the feature subset and power: (3) In Equation (3), is the feature subset, is the target variable, is the th feature, is the number of features; Calculate the minimum redundancy index to minimize the correlation between feature variables in the selected feature set: (4) Taking into account the maximum correlation index and minimum redundancy criterion, define the operator and calculate the mRMR value as: (5).

[0009] Preferably, the residual between the predicted power and the actual power of the training set is used as the standard residual set, and the Mahalanobis distance between the residual of the test set power prediction obtained by using the test set and the standard residual set is used to divide the health index of the wind turbine.

[0010] A health status assessment system for SCADA-based wind turbines includes a preprocessing module and an evaluation module; The preprocessing module tests the wind turbine to be predicted based on the prediction model pre-trained from the historical data set of the wind turbine and the SCADA data under the healthy operating state of the wind turbine to be predicted, and obtains the prediction error as the standard residual set; The evaluation module calculates the residual between the actual power and the predicted power of the wind turbine to be predicted according to the actual power of the wind turbine to be predicted obtained in real time, and uses the Mahalanobis distance between the obtained residual and the standard residual set as the health status evaluation benchmark of the wind turbine to be predicted to evaluate the health status of the wind turbine to be predicted.

[0011] Preferably, the prediction model is a prediction model with the output power of the wind turbine as the prediction target and power-related characteristic variables as the input.

[0012] An electronic device includes a processor and a memory. The processor is configured to execute a computer program stored in the memory to implement the above-mentioned SCADA-based wind turbine health status assessment method.

[0013] A computer-readable storage medium stores at least one instruction, and when the at least one instruction is executed by a processor, the above-mentioned SCADA-based wind turbine health status assessment method is implemented.

[0014] Compared with the prior art, the present invention has the following beneficial technical effects: A method for assessing the health status of a SCADA-based wind turbine according to the present invention includes pre-training a prediction model based on a historical data set of the wind turbine, and testing the wind turbine to be predicted with the SCADA data under the healthy operating state of the wind turbine to be predicted to obtain a prediction error as a standard residual set. The actual power of the wind turbine to be predicted is obtained in real time, the residual between the actual power and the predicted power of the wind turbine to be predicted is calculated, and the Mahalanobis distance between the obtained residual and the standard residual set is used as the health status assessment benchmark for the wind turbine to be predicted to evaluate the health status of the wind turbine to be predicted. Guided by the prediction result for evaluation, the present invention can greatly improve the accuracy of the health status assessment of the wind turbine and reduce the workload of the health status assessment of the wind turbine.

[0015] The present invention takes the output power of the wind turbine as the prediction target, optimizes the parameters of the CNN-BiGRU network using the sparrow search algorithm, constructs a health assessment index based on the Mahalanobis distance between the residual between the real-time power prediction value and the actual value and the standard residual set, and the evaluation result is more objective and accurate. Compared with the SCADA system, the deterioration trend of the wind turbine can be monitored in advance and early warnings can be given. Description of the Drawings

[0016] Figure 1 is a flowchart of the SCADA-based wind turbine health status assessment method in an embodiment of the present invention.

[0017] Figure 2 is a wind speed-power scatter plot before data preprocessing in an embodiment of the present invention.

[0018] Figure 3 is a wind speed-power scatter plot after data preprocessing using the quartile method in an embodiment of the present invention.

[0019] Figure 4It is the CNN-BiGRU network structure diagram in the embodiment of the present invention.

[0020] Figure 5 It is the curve of the CNN-BiGRU network prediction result and the actual monitoring value in the embodiment of the present invention.

[0021] Figure 6 It is the Mahalanobis distance between the real-time residual and the standard residual set in the embodiment of the present invention.

[0022] Figure 7 It is the unit health index in the embodiment of the present invention. Detailed implementation manners

[0023] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0025] As Figure 1 shown, the present invention provides a method for evaluating the health status of a wind turbine generator based on a Supervisory Control and Data Acquisition (SCADA) system, to achieve the evaluation of the health status of the wind turbine generator; specifically including the following steps: S1, based on the prediction model pre-trained with the historical data set of the wind turbine generator and the SCADA data under the healthy operating state of the wind turbine generator to be predicted, test the wind turbine generator to be predicted, and obtain the prediction error as the standard residual set; S2, obtaining the actual power of the wind turbine to be predicted in real time, calculating the residual between the actual power of the wind turbine to be predicted and the predicted power, and evaluating the health status of the wind turbine to be predicted based on the Mahalanobis distance between the obtained residual and the standard residual set as the health status evaluation benchmark of the wind turbine to be predicted.

[0026] In the embodiment of the present application, the prediction model of the present application is a prediction model that takes the output power of the wind turbine as a prediction target and takes power-related characteristic variables as input.

[0027] In the embodiment of the present application, the quartile method is used to eliminate abnormal data of each characteristic variable in the wind turbine data acquisition and supervisory control system (SCADA), to obtain a historical data set of the wind turbine, and the maximum relevance and minimum redundancy (Max-Relevance and Min-Redundancy, mRMR) algorithm is used to select power-related characteristic variables.

[0028] Specifically, the steps for preprocessing the data stored in the SCADA system include: first, eliminating the data below the cut-in wind speed and with negative power values, and then using the quartile method to preprocess the historical SCADA data of the wind turbine to eliminate abnormal values.

[0029] The characteristic historical data in SCADA are preprocessed, and the wind speed-power (VP) and power-wind speed (PV) data sequences are arranged in order of size and divided into 4 equal parts using the quartile method. Each equal part accounts for 25% of the total data. The data at the 25%, 50%, and 75% cutoff points are called quartiles, which are recorded as the first quartile ( ), the second quartile ( ), the third quartile ( ), the third quartile ( ) and the first quartile ( ) is also called the interquartile range (IQR): (1) In statistics, it is considered , Data outside the range are outliers, so the data outside the above range are removed. Figure 2 , Figure 3 Wind speed-power scatter plots before and after data preprocessing, respectively.

[0030] For the prediction error, the maximum relevance and minimum redundancy (mRMR) algorithm is used to select the power-related features. The mutual information between each feature variable and power is calculated according to Equation (2): (2) In Equation (2), 、 、 are the marginal probability density and joint probability density of the random variables 、 respectively; Calculate the maximum correlation index between the feature subset and power: (3) In the formula (3), is the feature subset, is the target variable, is the th feature, is the number of features.

[0031] Calculate the minimum redundancy index to minimize the correlation between the feature variables in the selected feature set: (4) Taking into account the maximum correlation index and the minimum redundancy index criteria, define the operator , and calculate the mRMR value as: (5) Table 1 shows the mRMR values of different feature variables.

[0032] Table 1

[0033] Select the feature parameters with mRMR values greater than 1, including: torque, wind direction, wind turbine speed, wind speed, generator torque, and gearbox bearing temperature, as the input parameters of the power prediction model.

[0034] Such as Figure 4 Construct a CNN-BiGRU power prediction model, including an input layer, a CNN layer, a BiGRU layer, and an output layer. Set the convolution kernel size to 3, the sliding step size to 1, use the ReLU function as the activation function, use average pooling for the pooling layer, set the maximum number of training times to 100, use the Adam optimization algorithm, and use the sparrow search algorithm to optimize the number of hidden nodes, the initial learning rate, and the L2 regularization coefficient of the BiGRU network. The optimization results are shown in Table 2.

[0035] Table 2

[0036] The preprocessed data is divided into a training set and a test set. The residual between the predicted power and the actual power of the training set is used as the standard residual set as follows: (6) In formula (6), is the predicted power value, is the actual power value.

[0037] The Mahalanobis distance between the residual of the test set power prediction obtained by using the test set and the standard residual set is calculated: (7) In formula (7), is the residual of the power prediction at time is the standard residual set, and are respectively the mean and covariance of the distribution .

[0038] To eliminate the influence of random errors, the size of the sliding window is set to m, and the Mahalanobis distance of the moving average at time t is obtained as: According to this Mahalanobis distance, the health index of the wind turbine is defined as: The operating state of the wind turbine is divided into three levels: healthy, general, and abnormal. The health index HI > 0.9 is the healthy level; the α value is determined according to the distribution of the Mahalanobis distance between the residual of the power prediction in the healthy operating state of the wind turbine and the standard residual set. According to the 3σ criterion, the Mahalanobis distance at μ + 3σ of this distribution is used as the demarcation point between the healthy and general levels, and α = 12.74 is obtained. And the index for the abnormal state alarm of the wind turbine is determined as HI = 0.86 based on the maximum value of this distribution.

[0039] The health state level index of the wind turbine is shown in Table 3.

[0040] Table 3

[0041] The health index of the wind turbine is calculated according to the Mahalanobis distance between the residual of the real-time power prediction value and the actual value of the wind turbine and the standard residual set, and the current health level is determined according to the classification result of the above steps. Specific embodiment: Taking the actual SCADA monitoring data of a 2MW wind turbine, 5000 groups of data in the healthy operating state of the unit are selected as training samples for power prediction network training, Figure 5The curve of the power prediction result of the CNN-BiGRU network proposed by the present invention and the actual monitoring value is shown. It can be seen that the CNN-BiGRU network constructed by the present invention can accurately predict the output power of the wind turbine generator set.

[0043] Using the historical data of the three days before the abnormal shutdown of the unit at 7:30 on June 8, 2017 for verification, torque, wind direction, wind turbine speed, wind speed, generator torque, and the temperature of the gearbox bearing 2 are selected as the input variables of the power prediction model. The Mahalanobis distance between the real-time residual of the prediction result and the standard residual set and the corresponding health index value are respectively as Figure 6 , Figure 7 shown. It can be seen from Figure 6 that the Mahalanobis distance began to increase on June 7. From the health index of the wind turbine generator set characterized by the Mahalanobis distance shown in Figure 7 , it can be found that the health index dropped to the interval of [0.86, 0.9] many times starting from 00:50 on June 7, and then dropped below 0.86 at 16:20 on June 7, indicating that the wind turbine generator set was in an abnormal state until the unit issued a shutdown alarm. The data experiment results show that the method proposed by the present invention can detect the abnormal state of the wind turbine generator set 15 hours earlier than the SCADA system.

[0044] In another embodiment of the present invention, a SCADA-based wind turbine generator set health status evaluation system is provided, including a preprocessing module and an evaluation module; The preprocessing module tests the wind turbine generator set to be predicted based on the prediction model pre-trained from the historical data set of the wind turbine generator set and the SCADA data under the healthy operation state of the wind turbine generator set to be predicted, and obtains the prediction error as the standard residual set; The evaluation module calculates the residual between the actual power of the wind turbine generator set to be predicted and the predicted power according to the actual power of the wind turbine generator set to be predicted obtained in real time, and uses the Mahalanobis distance between the obtained residual and the standard residual set as the health status evaluation benchmark of the wind turbine generator set to be predicted to evaluate the health status of the wind turbine generator set to be predicted.

[0045] Preferably, the prediction model is a prediction model with the output power of the wind turbine generator set as the prediction target and the power-related characteristic variables as the input.

[0046] In another embodiment of the present invention, a terminal device is provided. The terminal device includes a processor and a memory. The memory is used to store a computer program, and the computer program includes program instructions. The processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding functions. The processor described in the embodiment of the present invention can be used for the operation of the SCADA wind turbine health status assessment method.

[0047] In another embodiment of the present invention, a storage medium is also provided, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is a memory device in the terminal device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and, of course, the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, and the operating system of the terminal is stored in this storage space. And in this storage space, one or more instructions suitable for being loaded and executed by the processor are also stored. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the SCADA wind turbine health status assessment method in the above embodiments.

[0048] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0049] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks.

[0050] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one or more of the flows Figure 1 or blocks.

[0051] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks.

[0052] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the specific implementation manners of the present invention can still be modified or equivalently replaced, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

[0053] This invention does not require establishing mathematical models for each subsystem of the wind turbine. By using easily obtainable SCADA data, with the output power of the wind turbine as the prediction target, the sparrow search algorithm is adopted to optimize the parameters of the CNN-BiGRU network. A health assessment index is constructed based on the Mahalanobis distance between the residual between the real-time power prediction value and the actual value and the standard residual set, and the evaluation result is more objective and accurate. Compared with the SCADA system, it can monitor the deterioration trend of the wind turbine in advance and give early warnings.

Claims

1. A method for evaluating the health status of a wind turbine based on SCADA, characterized in that The following steps are involved: S1, the prediction model obtained by pre-training the historical data set of the wind turbine and the SCADA data of the wind turbine to be predicted in a healthy operating state are used to test the wind turbine to be predicted, and the prediction error is obtained as the standard residual set; S2, obtaining the actual power of the wind turbine to be predicted in real time, calculating the residual between the actual power of the wind turbine to be predicted and the predicted power, and evaluating the health status of the wind turbine to be predicted based on the Mahalanobis distance between the obtained residual and the standard residual set as the health status evaluation benchmark of the wind turbine to be predicted.

2. The method for evaluating the health status of a wind turbine based on SCADA according to claim 1, wherein The prediction model takes the output power of wind turbines as the prediction target and the power-related characteristic variables as the input.

3. The health state assessment method of a wind turbine based on SCADA according to claim 1, characterized in that The quartile method is used to eliminate abnormal data of each characteristic variable in the data acquisition and monitoring control system of the wind turbine set, and the historical data set of the wind turbine set is obtained. The maximum relevance minimum redundancy algorithm is used to select the power-related characteristic variables.

4. The method for evaluating the health status of a wind turbine based on SCADA according to claim 2, wherein, Firstly, the data with cut-in wind speed below the specified value and negative power value are eliminated. Then, the historical SCADA data of wind turbines are preprocessed by using the quartile method to eliminate abnormal values.

5. The method for evaluating the health status of a wind turbine based on SCADA according to claim 2, wherein, The maximum relevance minimum redundancy algorithm is used to select power-related features, and the mutual information between each feature variable and power is calculated according to formula (2): （2） In formula (2) , , are respectively the marginal probability density and the joint probability density of the random variables and ; Compute the maximum correlation metric for a feature subset with power: （3） In formula (3), is the feature subset, is the target variable, is the th feature, is the number of features; Calculate the minimum redundancy index to minimize the correlation between the feature variables in the selected feature set: （4） Considering the maximum relevance index and the minimum redundancy index comprehensively, an operator is defined , and the mRMR value is calculated as follows: （5）。 6. The method for evaluating the health status of a wind turbine based on SCADA according to claim 1, wherein, The residual between the predicted power and the actual power of the training set is taken as the standard residual set, and the Mahalanobis distance between the residual of the test set power prediction obtained by calculating the test set and the standard residual set is used to divide the health indicators of the wind turbines.

7. A health status assessment system for a wind turbine generator based on SCADA, characterized in that, Includes preprocessing module and evaluation module; The preprocessing module tests the wind turbine to be predicted based on the prediction model obtained by pre-training the historical data set of the wind turbine and the SCADA data of the wind turbine to be predicted in a healthy operating state, and obtains the prediction error as the standard residual set; The evaluation module calculates the residual between the actual power and the predicted power of the wind turbine to be predicted based on the actual power of the wind turbine to be predicted obtained in real time, and evaluates the health status of the wind turbine to be predicted based on the Mahalanobis distance between the obtained residual and the standard residual set as the health status evaluation benchmark of the wind turbine to be predicted.

8. The health status assessment system of a wind turbine based on SCADA according to claim 7, characterized in that, The prediction model takes the output power of wind turbines as the prediction target and the power-related characteristic variables as the input.

9. An electronic device, characterized in that, It comprises a processor and a memory, wherein the processor is used to execute a computer program stored in the memory to implement the SCADA-based wind turbine health status assessment method as claimed in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one instruction, and when the at least one instruction is executed by the processor, the SCADA-based wind turbine health status assessment method according to any one of claims 1 to 6 is implemented.

Images