Generator set abnormal data identification method and device, and electronic equipment

Abstract

The application discloses an abnormal data identification method and device of a generator set and electronic equipment. The method comprises the following steps: obtaining operation data of the generator set, wherein the operation data comprises wind speed data and power data; inputting the power data into a trained inverse model, and outputting predicted wind speed corresponding to the power data by the inverse model, wherein the inverse model is a neural network model, and is trained by a plurality of sets of training data, each set of training data comprising input power data and output wind speed data; and determining whether the operation data is abnormal according to whether a difference between the wind speed data and the predicted wind speed reaches a preset threshold. The problems that a large amount of calculation and low accuracy exist in the way of abnormal detection based on data statistical characteristics of the generator set in the related art are solved.

Description

Technical Field

[0001] This application relates to the field of new energy power generation, and more specifically, to a method and device for identifying abnormal data of generator sets, and electronic equipment. Background Technology

[0002] In the daily operation of wind turbine units, data acquisition systems typically collect a large amount of operational data continuously. Many of these parameters are crucial for the safe operation and efficiency improvement of the unit. The accuracy of these parameters is essential for optimized operation and online monitoring; however, due to various reasons, observed parameter values ​​often deviate from actual values. Therefore, before using this data for analysis, preprocessing is necessary to eliminate as many outliers as possible.

[0003] Existing research mainly focuses on cleaning abnormal data from wind turbines based on the statistical characteristics of the data. Chinese patent CN202210652602 primarily constructs a comprehensive abnormal data identification result matrix, optimizes weight coefficients and data cleaning thresholds through an optimization algorithm, and proposes a comprehensive scoring index to achieve a comprehensive evaluation of the cleaning effect of complex data. Chinese patent CN20 1910892258 uses an improved fuzzy clustering algorithm to cluster wind turbine operation data and employs statistical methods to calculate the feasible region matrix from the clustering results, subsequently identifying abnormal data. Chinese patent CN201510484365 uses random forest and gradient iterative decision tree models to predict samples and obtain prediction results, thereby achieving abnormal data identification.

[0004] As can be seen from the above, the existing correction methods based on data statistics are relatively simple in principle, but they have a certain amount of computation in actual operation and require a certain amount of data. Machine learning models have good fitting of nonlinear relationships, but the generality of the models is limited, and models built for a single object are difficult to apply to other objects.

[0005] The anomaly detection methods based on the statistical characteristics of generator set data in related technologies suffer from high computational complexity and low accuracy, and no effective solution has yet been proposed. Summary of the Invention

[0006] The main objective of this application is to provide a method, device, and electronic equipment for identifying abnormal data of generator sets, in order to solve the problems of high computational load and low accuracy in related technologies that rely on the statistical characteristics of generator set data for anomaly detection.

[0007] To achieve the above objectives, according to one aspect of this application, a method for identifying abnormal data of a generator set is provided, comprising: acquiring operating data of the generator set, wherein the operating data includes wind speed data and power data; inputting the power data into a trained inverse model, wherein the inverse model outputs a predicted wind speed corresponding to the power data, wherein the inverse model is a neural network model trained from multiple sets of training data, each set of training data including input power data and output wind speed data; and determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold.

[0008] Optionally, obtaining the generator set's operating data includes: collecting the generator set's operating data at a target time; matching wind speed data and power data using the time stamp of the operating data to form data pairs; and deleting invalid data from the operating data, wherein the invalid data includes wind speed data without corresponding power data, power data without corresponding wind speed data, and data without a timestamp.

[0009] Optionally, determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold includes: determining that the operating data is abnormal if the difference between the wind speed data and the predicted wind speed reaches the preset threshold; and determining that the operating data is normal if the difference between the wind speed data and the predicted wind speed does not reach the preset threshold.

[0010] Optionally, before inputting the power data into the trained inverse model and having the inverse model output the predicted wind speed corresponding to the power data, the method further includes: obtaining a training set based on the theoretical operating data of the generator set; creating the inverse model based on the neural network; and training the inverse model based on the training set to obtain the trained inverse model.

[0011] Optionally, obtaining a training set based on the theoretical operating data of the generator set includes: obtaining the theoretical power curve of the generator set, wherein the theoretical power curve is a theoretical variation curve of wind speed and power; determining curve data between the cut-in wind speed and the rated wind speed based on the theoretical power curve, wherein the cut-in wind speed is the wind speed at which the generator set begins to generate electricity, and the rated wind speed is the wind speed at which the generator set is fully generating electricity; and determining multiple data pairs of theoretical wind speed and corresponding theoretical power based on the curve data, as a training set.

[0012] Optionally, training the inverse model based on the training set to obtain the trained inverse model includes: using the theoretical power in the training set as input and the theoretical wind speed as output, training the inverse model through an adaptive momentum stochastic optimization method; and determining that the inverse model training is complete when the loss function of the inverse model meets a preset requirement.

[0013] Optionally, before determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold, the method further includes: dividing multiple wind speed ranges according to the theoretical power curve of the generator set, and determining a preset threshold for each wind speed range, wherein the preset threshold is the allowable wind speed deviation value of normal wind speed data; and determining the corresponding preset threshold according to the wind speed range in which the predicted wind speed is located.

[0014] To achieve the above objectives, according to another aspect of this application, an abnormal data identification device for a generator set is provided, comprising: an acquisition module for acquiring operating data of the generator set, wherein the operating data includes wind speed data and power data; a reverse calculation module for inputting the power data into a trained reverse calculation model, wherein the reverse calculation model outputs a predicted wind speed corresponding to the power data, wherein the reverse calculation model is a neural network model trained from multiple sets of training data, each set of training data including input power data and output wind speed data; and a determination module for determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold.

[0015] According to another aspect of this application, a computer-readable storage medium is also provided, the storage medium being used to store a program, wherein the program executes the abnormal data identification method for generator sets described in any of the preceding claims.

[0016] According to another aspect of this application, an electronic device is also provided, including one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the abnormal data identification method for generator sets described in any of the above.

[0017] This application employs the following steps: acquiring generator set operating data, including wind speed and power data; inputting the power data into a trained inverse model, which outputs a predicted wind speed corresponding to the power data. The inverse model is a neural network model trained on multiple sets of training data, each set including input power data and output wind speed data; determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold. By using a neural network model to inversely deduce the predicted wind speed from the power data, and then using the predicted wind speed and the actual wind speed data for anomaly identification, the application achieves the goal of identifying data anomalies based on directly and easily measurable wind speed data. This improves the efficiency of anomaly identification and reduces computational load. Using a neural network model as the inverse model, and employing this method to inversely deduce wind speed based on efficiency, offers high speed and accuracy, thus solving the problems of high computational load and low accuracy associated with related technologies that rely on the statistical characteristics of generator set data for anomaly detection. Attached Figure Description

[0018] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 This is a flowchart of an abnormal data identification method for a generator set according to an embodiment of this application;

[0020] Figure 2 This is a schematic diagram of the theoretical power curve of a wind turbine provided according to the embodiments of this application;

[0021] Figure 3 This is a schematic diagram of a single hidden layer neural network provided according to an embodiment of this application;

[0022] Figure 4-1 This is a schematic diagram of the actual power distribution of a certain type of wind turbine provided according to the embodiments of this application;

[0023] Figure 4-2 This is a schematic diagram of the actual power distribution of a certain type of wind turbine after removing abnormal data according to the embodiments of this application;

[0024] Figure 5-1 This is a schematic diagram of the actual power output distribution of a similar wind turbine provided according to the embodiments of this application;

[0025] Figure 5-2 This is a schematic diagram showing the actual power distribution of similar wind turbine units after removing abnormal data according to the embodiments of this application;

[0026] Figure 6 This is a schematic diagram of an abnormal data identification device for a generator set according to an embodiment of this application;

[0027] Figure 7 This is a schematic diagram of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0028] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0029] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0031] The present invention will now be described in conjunction with preferred implementation steps. Figure 1 This is a flowchart of an abnormal data identification method for a generator set according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0032] Step S101: Obtain the operating data of the generator set, including wind speed data and power data;

[0033] Step S102: Input the power data into the trained inverse model, and output the predicted wind speed corresponding to the power data from the inverse model. The inverse model is a neural network model, which is trained from multiple sets of training data. Each set of training data includes the input power data and the output wind speed data.

[0034] Step S103: Determine whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold.

[0035] Through the above steps, the operating data of the generator set is obtained, including wind speed data and power data. The power data is input into the trained inverse model, which outputs the predicted wind speed corresponding to the power data. The inverse model is a neural network model, which is trained from multiple sets of training data. Each set of training data includes the input power data and the output wind speed data. The abnormality of the operating data is determined based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold.

[0036] By using a neural network model to predict wind speed based on power data, and then using the predicted wind speed and actual wind speed data to identify data anomalies, the goal of identifying data anomalies based on directly and easily measurable wind speed data is achieved. This improves the efficiency of data anomaly identification and reduces the amount of data computation. By using a neural network model as a reverse inference model and employing a reverse inference model to deduce wind speed based on efficiency, the technology achieves high speed and accuracy. This solves the problems of high computational load and low accuracy in related technologies that rely on the statistical characteristics of generator data for anomaly detection.

[0037] The entity performing the above steps can be a processor, calculator, server, or other device with data processing and analysis capabilities. Alternatively, it can be a device with the aforementioned data processing and analysis capabilities, such as a computer with a processor, a smartphone, a wearable device, a data system with a server, or a computing system.

[0038] The aforementioned generator sets can be new energy generator sets, such as wind turbines. In other scenarios, they can also be solar generators; however, the wind speed data may be adjusted to reflect sunlight data, such as sunlight intensity and duration. The training data for the inverse model is also adjusted accordingly. In other words, the aforementioned inverse model can be adapted and applied to new energy power generation fields such as photovoltaics and hydropower.

[0039] The aforementioned inverse model is used to infer the corresponding predicted wind speed based on the actual measured power data. The predicted wind speed can be considered as the standard wind speed corresponding to that actual power. By comparing it with the measured wind speed, the difference between the two is used to determine whether there are any anomalies in the actual measured operating data.

[0040] Optionally, determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold includes: determining that the operating data is abnormal if the difference between the wind speed data and the predicted wind speed reaches the preset threshold; and determining that the operating data is normal if the difference between the wind speed data and the predicted wind speed does not reach the preset threshold.

[0041] Specifically, if the difference between the predicted wind speed and the measured wind speed exceeds a preset threshold, the operational data can be considered abnormal, and an anomaly warning and notification can be issued. If the difference between the predicted wind speed and the measured wind speed does not exceed the preset threshold, the operational data can be considered normal.

[0042] Optionally, acquiring the generator set's operating data includes: collecting the generator set's operating data at a target time; matching wind speed data and power data using the time stamp of the operating data to form data pairs; and deleting invalid data from the operating data, including wind speed data without corresponding power data, power data without corresponding wind speed data, and data without a timestamp.

[0043] After acquiring the operational data, wind speed and power data are mapped into data pairs using time tags. Invalid data is then removed, including wind speed data without corresponding power data, power data without corresponding wind speed data, and data without timestamps, which cannot form valid data pairs. Removing invalid data reduces the workload of subsequent data processing and improves data accuracy, thereby increasing the training speed and accuracy of the model.

[0044] Optionally, before inputting the power data into the trained inverse model and having the inverse model output the predicted wind speed corresponding to the power data, the method further includes: obtaining a training set based on the theoretical operating data of the generator set; creating an inverse model based on a neural network; and training the inverse model based on the training set to obtain the trained inverse model.

[0045] The aforementioned reverse inference model is a neural network model, built from a neural network. Preferably, in this embodiment, the reverse inference model is built from a single hidden layer neural network, with the specific structure as follows: Figure 3 As shown, it includes an input layer, a hidden layer (i.e., a single hidden layer), and an output layer.

[0046] The inverse model is trained using the aforementioned training set until it converges and meets the requirements of the loss function. This indicates that the inverse model training is complete. The trained inverse model can then be used to derive the corresponding predicted wind speed from measured power data.

[0047] Optionally, based on the theoretical operating data of the generator set, obtaining the training set includes: obtaining the theoretical power curve of the generator set, wherein the theoretical power curve is the theoretical variation curve of wind speed and power; determining the curve data between the cut-in wind speed and the rated wind speed based on the theoretical power curve, wherein the cut-in wind speed is the wind speed at which the generator set starts generating electricity, and the rated wind speed is the wind speed at which the generator set is generating electricity at full capacity; and determining multiple data pairs of theoretical wind speed and corresponding theoretical power based on the curve data, as the training set.

[0048] When acquiring a training set based on operational data, this can be done by obtaining the theoretical power curve of the generator set, which is the theoretical curve of wind speed versus power. Firstly, the accuracy and authority of the theoretical power curve are relatively high, with a very low probability of error. Secondly, the curve corresponds to countless values, ensuring a large data volume. Thirdly, it avoids the complex processes, low efficiency, and high costs associated with actual measurement data.

[0049] Specifically, the curve data between the cut-off wind speed and the rated wind speed is used. The power in this curve segment increases with the wind speed, avoiding the situation where the power does not change with the wind speed in the curve segment at full load power generation, resulting in data without differentiation and making it difficult to effectively and accurately train the model.

[0050] Optionally, training the inverse model based on the training set to obtain the trained inverse model includes: using the theoretical power in the training set as input and the theoretical wind speed as output, training the inverse model through an adaptive momentum stochastic optimization method; and determining that the inverse model training is complete when the loss function of the inverse model meets the preset requirements.

[0051] During training, the aforementioned inverse model employs an adaptive momentum stochastic optimization method, such as the Adam optimization algorithm, which yields better training results and can improve training speed and accuracy.

[0052] Optionally, before determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold, the method further includes: dividing multiple wind speed ranges according to the theoretical power curve of the generator set, and determining a preset threshold for each wind speed range, wherein the preset threshold is the allowable wind speed deviation value of normal wind speed data; and determining the corresponding preset threshold according to the wind speed range in which the predicted wind speed is located.

[0053] Considering the different power variations across different wind speed ranges, this embodiment divides the wind speed into multiple ranges based on specific wind speed values. Each wind speed range corresponds to a preset threshold for determining whether the data is abnormal. This preset threshold can be an empirical value. A flexible quantification index is used for the preset threshold to fully consider the distribution of measured data, thereby improving the accuracy of data anomaly detection.

[0054] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0055] It should be noted that this application also provides an optional implementation method, which will be described in detail below.

[0056] This embodiment provides a method for identifying abnormal data of wind turbine generators using flexible quantitative indicators. The specific implementation involves the following steps:

[0057] 1. For a specific model or a specific wind turbine in the site, obtain its theoretical power curve and select the area between the cut-in wind speed and the rated wind speed (i.e., the area from when the wind turbine starts generating power normally to when it reaches full power) for backup. Figure 2 This is a schematic diagram of the theoretical power curve of a wind turbine provided according to the embodiments of this application. The theoretical power curve is as follows: Figure 2 The figure shown is the theoretical curve of power generation as a function of wind speed. Figure 2 In the middle, v A To cut off the wind speed, v B This is the rated wind speed.

[0058] 2. Figure 3 This is a schematic diagram of a single hidden layer neural network provided according to an embodiment of this application, such as... Figure 3 As shown, a single hidden layer neural network is used to reverse-engineer the backup region of the theoretical power curve. The input of the model is the theoretical power value and the output is the corresponding wind speed. The network training adopts an adaptive momentum stochastic optimization algorithm, and the training terminates after the training requirements are met. The network expression is as follows.

[0059] y=f(x|(W,b)) (1)

[0060] In equation (1): x is the model input, y is the model output, and (W,b) are the model-related parameters;

[0061] 3. Obtain nacelle wind speed and wind power data from actual operating wind turbine units, match them one by one using time tags, and delete invalid data;

[0062] 4. Input the measured wind power data into the trained single hidden layer neural network, and the model output is the wind speed value that theoretically corresponds to the power value.

[0063] 5. Compare the calculated theoretical wind speed value with the measured wind speed value, and determine whether the data is abnormal according to the following formula;

[0064] |v real -v model|≤α (2)

[0065] In equation (2): v real The measured wind speed value is v. model For the calculated theoretical wind speed value, α is a preset threshold. α can be set to a single constant; or it can be based on v. model The data values ​​are segmented to set flexible quantitative indicators;

[0066] Data that meets the requirements of the above formula is considered reasonable; data that does not meet the requirements is considered abnormal.

[0067] This implementation uses a single hidden-layer neural network to model the theoretical power-wind speed relationship of wind turbines, and derives the theoretical incoming wind speed based on this. Deriving theoretical wind speed from power values ​​makes it easier to identify outliers compared to the conventional method of deriving theoretical power from wind speed values. Instead of rigid quantification indicators for outlier judgment, flexible quantification indicators are set based on the magnitude of data values, fully considering the actual distribution of measured data. For areas where measured data is concentrated, the reasonable range is expanded accordingly.

[0068] By combining the theoretical power curve of the wind turbine, the corresponding theoretical incoming wind speed is derived from the power value, and anomaly detection is achieved. First, the power curve derivation uses a single hidden layer neural network, which has higher fitting accuracy compared to commonly used polynomial functions. Second, anomaly detection is achieved by comparing theoretical and measured wind speeds, which yields better results than comparing power values, and the implementation steps are simple. Third, the anomaly detection adopts a flexible quantification standard, further considering the actual distribution of measured data.

[0069] The following describes this embodiment in further detail with reference to the accompanying drawings and specific implementation details.

[0070] For a specific type of wind turbine in a selected wind farm, its theoretical power curve is obtained by consulting relevant materials, and raw data such as nacelle wind speed and actual power generated by the turbine are collected for later use.

[0071] Select as Figure 2 The region shown is between the cut-in wind speed and the rated wind speed. (Using...) Figure 3 The single hidden layer neural network in the model reverse-engineers the region. The input of the model is the theoretical power value, and the output value corresponds to the incoming wind speed. The training of the network adopts the backpropagation algorithm. After reaching a certain number of iterations or a certain accuracy, the training is completed, and the reverse-engineering model f(x) in formula (1) is obtained.

[0072] The nacelle wind speed and actual generator power data are matched one by one using time tags, and invalid data is deleted.

[0073] The filtered actual power generation data is input into the trained neural network to obtain the corresponding theoretical incoming wind speed value;

[0074] The preset threshold α is set to 2, and abnormal data is identified using formula (2). The value of α is continuously adjusted according to the identification performance until the identification performance reaches its optimal level. The results before and after identification are as follows: Figure 4-1 and Figure 4-2 As shown, Figure 4-1 This is a schematic diagram of the actual power distribution of a certain type of wind turbine provided according to the embodiments of this application. Figure 4-2 This is a schematic diagram of the actual power distribution of a certain type of wind turbine after removing abnormal data according to the embodiments of this application;

[0075] The inverse model f(x) and the final value α' of the preset threshold are retained as the model and related parameters used for identifying abnormal data of this type of wind turbine. Data from another wind turbine of the same type is used for testing, and the test results are as follows: Figure 5-1 and Figure 5-2 As shown, Figure 5-1 This is a schematic diagram of the actual power distribution of similar wind turbine units provided according to the embodiments of this application. Figure 5-2 This is a schematic diagram showing the actual power distribution of similar wind turbine units after removing abnormal data according to the embodiments provided in this application. The test results verify the effectiveness of the proposed method for identifying abnormal data of wind turbine units using flexible quantitative indicators.

[0076] This application also provides an abnormal data identification device for generator sets. It should be noted that the abnormal data identification device for generator sets in this application can be used to execute the abnormal data identification method for generator sets provided in this application. The abnormal data identification device for generator sets provided in this application will be described below.

[0077] Figure 6 This is a schematic diagram of an abnormal data identification device for a generator set according to an embodiment of this application, such as... Figure 6 As shown, the device includes: an acquisition module 61, a reverse calculation module 62, and a determination module 63. The device will be described in detail below.

[0078] The acquisition module 61 is used to acquire the operating data of the generator set, wherein the operating data includes wind speed data and power data; the reverse calculation module 62 is connected to the acquisition module 61, and is used to input the power data into a trained reverse calculation model, and the reverse calculation model outputs the predicted wind speed corresponding to the power data, wherein the reverse calculation model is a neural network model, which is trained by multiple sets of training data, each set of training data including input power data and output wind speed data; the determination module 63 is connected to the reverse calculation module 62, and is used to determine whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold.

[0079] The abnormal data identification device for generator sets provided in this application embodiment acquires the operating data of the generator set, including wind speed data and power data; inputs the power data into a trained inverse model, and the inverse model outputs the predicted wind speed corresponding to the power data. The inverse model is a neural network model, which is trained from multiple sets of training data, each set of training data including the input power data and the output wind speed data; and determines whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold.

[0080] By using a neural network model to predict wind speed based on power data, and then using the predicted wind speed and actual wind speed data to identify data anomalies, the goal of identifying data anomalies based on directly and easily measurable wind speed data is achieved. This improves the efficiency of data anomaly identification and reduces the amount of data computation. By using a neural network model as a reverse inference model and employing a reverse inference model to deduce wind speed based on efficiency, the technology achieves high speed and accuracy. This solves the problems of high computational load and low accuracy in related technologies that rely on the statistical characteristics of generator data for anomaly detection.

[0081] The abnormal data identification device for the generator set includes a processor and a memory. The aforementioned acquisition module 61, reverse calculation module 62, determination module 63, etc., are all stored in the memory as program units. The processor executes the aforementioned program units stored in the memory to realize the corresponding functions.

[0082] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured. By adjusting kernel parameters, the problem of high computational complexity and low accuracy in anomaly detection methods based on generator set data statistical characteristics can be addressed.

[0083] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0084] This invention provides a computer-readable storage medium storing a program that, when executed by a processor, implements a method for identifying abnormal data of a generator set.

[0085] This invention provides a processor for running a program, wherein the program executes an abnormal data identification method for the generator set.

[0086] Figure 7 This is a schematic diagram of an electronic device provided according to an embodiment of this application, such as... Figure 7 As shown, this application embodiment provides an electronic device 70, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of any of the above methods.

[0087] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0088] This application also provides a computer program product that, when executed on an abnormal data identification device for a generator set, is adapted to execute a program that initializes any of the above-described method steps.

[0089] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable generator set anomaly data identification device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable generator set anomaly data identification device, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0091] These computer program instructions may also be stored in a computer-readable storage medium that can direct anomaly data identification devices of a computer or other programmable generator set to function in a specific manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions can also be loaded onto a computer or other programmable generator set's abnormal data identification device, causing a series of operational steps to be executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0093] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0094] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0095] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0096] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0097] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0098] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for identifying abnormal data in a generator set, characterized in that, include: Acquire the operating data of the generator set, wherein the operating data includes wind speed data and power data; Based on the theoretical operating data of the generator set, a training set is obtained; a backpropagation model is created based on the neural network; the backpropagation model is trained based on the training set to obtain the trained backpropagation model; wherein, the method for obtaining the training set is as follows: obtaining the theoretical power curve of the generator set, wherein the theoretical power curve is the theoretical change curve of wind speed and power; determining the curve data between the cut-off wind speed and the rated wind speed based on the theoretical power curve, wherein the cut-off wind speed is the wind speed at which the generator set starts generating electricity, and the rated wind speed is the wind speed at which the generator set is fully generating electricity; determining multiple data pairs of theoretical wind speed and corresponding theoretical power based on the curve data, as the training set; The power data is input into the trained inverse model, and the inverse model outputs the predicted wind speed corresponding to the power data. The inverse model is a neural network model, which is trained from multiple sets of training data. Each set of training data includes the input power data and the output wind speed data. Whether the operating data is abnormal is determined by whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold.

2. The method according to claim 1, characterized in that, Obtaining generator set operating data includes: Collect the operating data of the generator set at the target time; By using time tags to map wind speed data and power data, data pairs are formed. Delete invalid data from the operational data, including wind speed data without corresponding power data, power data without corresponding wind speed data, and data without timestamps.

3. The method according to claim 1, characterized in that, Determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold includes: If the difference between the wind speed data and the predicted wind speed reaches a preset threshold, the operating data is determined to be abnormal. If the difference between the wind speed data and the predicted wind speed does not reach a preset threshold, the operating data is determined to be normal.

4. The method according to claim 1, characterized in that, The back-inference model is trained based on the training set to obtain the trained back-inference model, which includes: The theoretical power in the training set is used as input and the theoretical wind speed is used as output. The inverse model is trained by an adaptive momentum stochastic optimization method. If the loss function of the inverse model meets the preset requirements, the training of the inverse model is considered complete.

5. The method according to any one of claims 1 to 4, characterized in that, Before determining whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold, the method further includes: Based on the theoretical power curve of the generator set, multiple wind speed ranges are divided, and a preset threshold for each wind speed range is determined, wherein the preset threshold is the allowable wind speed deviation value of normal wind speed data. Based on the wind speed range in which the predicted wind speed is located, a corresponding preset threshold is determined.

6. An abnormal data identification device for a generator set, characterized in that, include: The acquisition module is used to acquire the operating data of the generator set, wherein the operating data includes wind speed data and power data; The training module is used to obtain a training set based on the theoretical operating data of the generator set; create a backpropagation model based on a neural network; and train the backpropagation model based on the training set to obtain a trained backpropagation model. The process of obtaining the training set includes: obtaining the theoretical power curve of the generator set, where the theoretical power curve is the theoretical change curve between wind speed and power; determining the curve data between the cut-off wind speed and the rated wind speed based on the theoretical power curve, where the cut-off wind speed is the wind speed at which the generator set begins generating electricity, and the rated wind speed is the wind speed at which the generator set is fully generating electricity; and determining multiple data pairs of theoretical wind speed and corresponding theoretical power based on the curve data as the training set. The reverse calculation module is used to input the power data into the trained reverse calculation model, and the reverse calculation model outputs the predicted wind speed corresponding to the power data. The reverse calculation model is a neural network model, which is trained by multiple sets of training data. Each set of training data includes the input power data and the output wind speed data. The determination module is used to determine whether the operating data is abnormal based on whether the difference between the wind speed data and the predicted wind speed reaches a preset threshold.

7. A computer-readable storage medium, characterized in that, The storage medium is used to store a program, wherein the program executes the abnormal data identification method for generator sets according to any one of claims 1 to 5.

8. An electronic device, characterized in that, It includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the abnormal data identification method for generator sets according to any one of claims 1 to 5.

Images