Hydropower station electrical equipment intelligent diagnosis method and system

Abstract

The invention relates to an intelligent diagnosis method and system for electrical equipment of a hydropower station, and the method comprises the following steps: obtaining a historical data set of a target hydropower station, including current measurement data, environment sensing data and equipment fault records of each electrical equipment; carrying out harmonic feature extraction on the current measurement data of each electrical device, carrying out environment feature extraction on the environment sensing data, and carrying out feature fusion on the extracted corresponding harmonic features and environment features to generate fusion features; extracting an equipment fault label through the equipment fault record, and constructing a sample set based on the fusion feature and the equipment fault label; constructing an electrical equipment fault diagnosis model based on a neural network, and training the electrical equipment fault diagnosis model through the sample set; and obtaining real-time environment data of the target hydropower station and real-time current measurement data of each electrical device, and obtaining a fault diagnosis result of each electrical device through the trained electrical device fault diagnosis model.

Description

Technical Field

[0001] The present invention relates to an intelligent diagnosis method and system for electrical equipment in a hydropower station, belonging to the technical field of equipment monitoring in a hydropower station. Background Art

[0002] As hydropower stations continue to expand in size and the number of electrical equipment increases, the safe and stable operation of these equipment has a significant impact on the reliability of the power system. Over long-term operation, these equipment is prone to various faults, such as insulation aging, bearing wear, and harmonic interference, due to environmental factors, equipment aging, and load variations. If these faults are not discovered and addressed promptly, they can lead to equipment damage and even system failures. Therefore, effective fault diagnosis of hydropower station electrical equipment is crucial.

[0003] Traditional fault diagnosis of hydropower station electrical equipment relies primarily on manual experience and judgment, assessing equipment status through regular inspections and equipment parameter measurements. With the development of sensor technology and data analysis methods, data-driven intelligent fault diagnosis technology has gradually been applied to condition monitoring and fault warning of hydropower station electrical equipment. For example, Chinese invention patent publication number CN119089109A discloses a hydropower station transformer fault diagnosis algorithm based on multi-source heterogeneous data feature fusion using a multi-channel neural network. This method collects multi-source heterogeneous data such as one-dimensional data and two-dimensional data related to the transformer, filters the data using variance elimination and Pearson correlation coefficient, establishes adaptive models for feature extraction and fusion for different types of data, and finally uses a deep residual network to extract the fused data features and output the diagnostic results.

[0004] However, the above scheme still has the following shortcomings: First, the diagnostic system does not fully consider the coupling effect of special environmental factors of hydropower stations (such as temperature and humidity, air pressure fluctuations, mechanical vibrations, etc.) on electrical equipment, resulting in insufficient diagnostic accuracy; Second, the existing diagnostic method based on current waveform analysis fails to effectively solve the problem of dynamic influence of environmental parameters, especially the lack of adaptability under complex working conditions; this affects the accuracy and reliability of equipment fault diagnosis results. Summary of the Invention

[0005] In order to solve the above problems in the prior art, the present invention proposes an intelligent diagnosis method and system for electrical equipment in a hydropower station.

[0006] The technical solutions of the present invention are as follows:

[0007] In one aspect, the present invention provides a method for intelligent diagnosis of electrical equipment in a hydropower station, comprising the following steps:

[0008] Obtain historical data sets for the target hydropower station, including current measurement data of each electrical device, environmental sensor data, and equipment failure records;

[0009] Harmonic features are extracted from the current measurement data of each electrical device, and environmental features are extracted from the environmental sensor data. The extracted corresponding harmonic features and environmental features are then fused to generate fused features.

[0010] Extract device fault labels from device fault records and construct a sample set based on fusion features and device fault labels;

[0011] Construct an electrical equipment fault diagnosis model based on a neural network, train the electrical equipment fault diagnosis model through a sample set, and obtain a trained electrical equipment fault diagnosis model;

[0012] The real-time environmental data of the target hydropower station and the real-time current measurement data of each electrical equipment are obtained. After feature extraction and feature fusion, they are input into the trained electrical equipment fault diagnosis model to obtain the fault diagnosis results of each electrical equipment.

[0013] As a preferred embodiment, the step of extracting harmonic features from the current measurement data of each electrical device includes:

[0014] The non-stationary current waveform is extracted from the current measurement data, and the improved CEEMDAN algorithm is used to adaptively decompose the non-stationary current waveform into several effective intrinsic mode components.

[0015] For each effective eigenmode component, the instantaneous amplitude and phase parameters of each harmonic are calculated by Hilbert transform;

[0016] The instantaneous amplitude and phase parameters of each harmonic are constructed as harmonic features.

[0017] As a preferred embodiment, in the step of extracting environmental features from the environmental sensor data, the environmental features include short-term fluctuation indexes of temperature and humidity data and long-term features of air pressure data;

[0018] The short-term fluctuation index of the temperature and humidity data is obtained by performing a sliding window calculation on the temperature and humidity data in the environmental sensor data;

[0019] The long-term characteristics of the air pressure data are obtained by analyzing the air pressure data in the environmental sensor data using a pre-established ARIMA model.

[0020] As a preferred embodiment, the step of fusing the extracted corresponding harmonic features and environmental features to generate fused features includes:

[0021] Establish a spatiotemporal correlation matrix, where the rows of the spatiotemporal correlation matrix represent the time dimension and the columns represent the spatial dimension of the electrical equipment;

[0022] Preprocess the harmonic features and environmental features to make their time scales consistent, and establish the corresponding relationship between the harmonic features and environmental features in the spatiotemporal correlation matrix;

[0023] Based on the correspondence between harmonic features and environmental features, a fusion feature containing harmonic features and environmental features is constructed.

[0024] On the other hand, the present invention also provides an intelligent diagnostic system for electrical equipment in a hydropower station, comprising:

[0025] The data acquisition module is used to obtain historical data sets of the target hydropower station, including current measurement data of each electrical device, environmental sensor data, and equipment fault records;

[0026] The feature extraction module is used to extract harmonic features from the current measurement data of each electrical device, and extract environmental features from the environmental sensor data, and fuse the extracted corresponding harmonic features and environmental features to generate fused features;

[0027] The sample construction module is used to extract device fault labels from device fault records and construct a sample set based on fusion features and device fault labels;

[0028] The model training module is used to build an electrical equipment fault diagnosis model based on a neural network, and train the electrical equipment fault diagnosis model through a sample set to obtain a trained electrical equipment fault diagnosis model;

[0029] The fault diagnosis module is used to obtain the real-time environmental data of the target hydropower station and the real-time current measurement data of each electrical equipment. After feature extraction and feature fusion, the data is input into the trained electrical equipment fault diagnosis model to obtain the fault diagnosis results of each electrical equipment.

[0030] As a preferred embodiment, the step of extracting harmonic features from the current measurement data of each electrical device includes:

[0031] The non-stationary current waveform is extracted from the current measurement data, and the improved CEEMDAN algorithm is used to adaptively decompose the non-stationary current waveform into several effective intrinsic mode components.

[0032] For each effective eigenmode component, the instantaneous amplitude and phase parameters of each harmonic are calculated by Hilbert transform;

[0033] The instantaneous amplitude and phase parameters of each harmonic are constructed as harmonic features.

[0034] As a preferred embodiment, in the step of extracting environmental features from the environmental sensor data, the environmental features include short-term fluctuation indexes of temperature and humidity data and long-term features of air pressure data;

[0035] The short-term fluctuation index of the temperature and humidity data is obtained by performing a sliding window calculation on the temperature and humidity data in the environmental sensor data;

[0036] The long-term characteristics of the air pressure data are obtained by analyzing the air pressure data in the environmental sensor data using a pre-established ARIMA model.

[0037] As a preferred embodiment, the step of fusing the extracted corresponding harmonic features and environmental features to generate fused features includes:

[0038] Establish a spatiotemporal correlation matrix, where the rows of the spatiotemporal correlation matrix represent the time dimension and the columns represent the spatial dimension of the electrical equipment;

[0039] Preprocess the harmonic features and environmental features to make their time scales consistent, and establish the corresponding relationship between the harmonic features and environmental features in the spatiotemporal correlation matrix;

[0040] Based on the correspondence between harmonic features and environmental features, a fusion feature containing harmonic features and environmental features is constructed.

[0041] On the other hand, the present invention also proposes an electronic device, comprising a memory, a processor, and a computer program stored in the memory and runnable on the processor, wherein when the processor executes the program, the intelligent diagnosis method for electrical equipment of a hydropower station as described in any embodiment of the present invention is implemented.

[0042] On the other hand, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent diagnosis method for electrical equipment of a hydropower station as described in any embodiment of the present invention.

[0043] The beneficial effects of the present invention include:

[0044] The present invention breaks through the limitations of traditional single-dimensional diagnosis by constructing a multimodal data fusion diagnostic framework based on the "electrical-environmental" feature coupling model. It can more comprehensively consider the impact of environmental factors on electrical equipment and provide more accurate fault diagnosis results.

[0045] Additional aspects and advantages of the present invention will be set forth in the following description, and some of them will be obvious from the description, or may be learned by practicing the present invention. In addition, the various aspects and advantages of the present invention may be realized and obtained by the method steps and combinations particularly pointed out in the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a flow chart of the method according to the first embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0048] It should be understood that the step numbers used herein are only for convenience of description and are not intended to limit the order in which the steps are to be executed.

[0049] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0050] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0051] The term "and / or" refers to and includes any and all possible combinations of one or more of the associated listed items.

[0052] Example 1:

[0053] See also Figure 1 This embodiment provides an intelligent diagnosis method for electrical equipment in a hydropower station, which specifically includes the following steps:

[0054] S100, obtaining a historical data set of the target hydropower station, including current and voltage measurement data of each electrical device, environmental sensor data, and equipment fault records;

[0055] In this embodiment, a historical data set is first obtained through the data acquisition system of the hydropower station. This historical data set contains three types of data: current and voltage measurement data of electrical equipment, environmental sensor data, and equipment fault records. The current and voltage measurement data of electrical equipment are collected through current transformers and voltage transformers installed on various key electrical equipment in the hydropower station. The sampling frequency is 10kHz, the data accuracy is 16 bits, and the collection time is 12 consecutive months. Environmental sensor data includes temperature, humidity, and air pressure data in the hydropower station machine room. The sampling frequency is 1 time / minute, and the data accuracy is 0.1°C (temperature), 0.5%RH (humidity), and 0.1hPa (air pressure). Equipment fault records include information such as the time of fault occurrence, fault type, fault equipment number, and fault handling method. They are recorded by the hydropower station operation and maintenance personnel and stored in the operation and maintenance management system.

[0056] S200, extracting harmonic features from the current and voltage measurement data of each electrical device, and extracting environmental features from the environmental sensor data, and fusing the extracted corresponding harmonic features and environmental features to generate fused features;

[0057] S300, extracting device fault labels from device fault records, and constructing a sample set based on fusion features and device fault labels;

[0058] In this example, fault labels are first extracted from equipment fault records. Fault labels include two dimensions: fault type and fault severity. Fault types are categorized based on common faults of hydropower station electrical equipment: insulation aging, bearing failure, poor contact, overheating, and other. Fault severity is divided into four levels: normal, minor, moderate, and severe.

[0059] Then, the extracted fusion features are time-aligned with the fault labels to construct a sample set. The format of the sample set is {(X1, Y1), (X2, Y2), ..., (X□, Y□)}, where X i represents the fusion feature of the i-th sample, Y i represents the fault label of the i-th sample. To balance the number of fault samples of each class, the SMOTE algorithm is used to oversample the minority class samples, so that the ratio of the number of samples of each class is close to 1:1. The final sample set constructed includes a training set, a validation set, and a test set with a ratio of 7:1:2.

[0060] S400, constructing an electrical equipment fault diagnosis model based on a neural network, and training the electrical equipment fault diagnosis model using a sample set to obtain a trained electrical equipment fault diagnosis model;

[0061] S500: Acquire real-time environmental data of the target hydropower station and real-time current and voltage measurement data of each electrical device, and input them into a trained electrical equipment fault diagnosis model after feature extraction and feature fusion to obtain fault diagnosis results of each electrical device.

[0062] In this embodiment, the hydropower station's real-time data acquisition system acquires real-time environmental data and real-time current and voltage measurement data from each electrical device. Real-time environmental data includes temperature, humidity, and air pressure, and is sampled once per minute; real-time current and voltage measurement data is sampled at a frequency of 10 kHz. Simultaneously, feature extraction and feature fusion are performed on the acquired data according to the method in step S200 to generate fused features. These fused features are then input into a trained electrical equipment fault diagnosis model to obtain fault diagnosis results for each electrical device. The diagnostic results include fault type, fault severity, and fault probability. The system sets different alarm levels based on fault probability: no alarm is triggered when the fault probability is below 30%; a pre-alarm is triggered when the fault probability is between 30% and 70%; and an alarm is triggered when the fault probability is above 70%. Alarm information is displayed to operation and maintenance personnel through the hydropower station monitoring system, providing fault cause analysis and recommended solutions.

[0063] In one embodiment, in step S200, the step of extracting harmonic features from the current and voltage measurement data of each electrical device specifically includes:

[0064] S210. Adaptively decompose the non-stationary current waveform using an improved CEEMDAN algorithm. This improved CEEMDAN algorithm introduces an adaptive adjustment mechanism for the ambient noise level during the decomposition process, dynamically selecting valid intrinsic mode components based on the kurtosis coefficient, and adding constraints on the number of decomposition layers based on the device load rate. Specifically, the adaptive adjustment mechanism dynamically adjusts the white noise amplitude based on the current ambient noise intensity, with the noise amplitude range being 0.05-0.2 times the signal standard deviation. The kurtosis coefficient threshold is set to 3.5, and when the kurtosis coefficient of an intrinsic mode component exceeds this threshold, it is determined to be a valid component. The constraint relationship between the device load rate and the number of decomposition layers is as follows: when the load rate is below 30%, the number of decomposition layers does not exceed 8; when the load rate is between 30% and 70%, the number of decomposition layers is 8-12; and when the load rate is above 70%, the number of decomposition layers is 12-15.

[0065] S211. Calculate the instantaneous amplitude and phase parameters of each harmonic using the Hilbert transform. Apply the Hilbert transform to each effective eigenmode component to obtain an analytical signal, and then calculate the instantaneous amplitude and phase. The instantaneous amplitude reflects changes in harmonic intensity, while the instantaneous phase reflects changes in harmonic phase. For the fundamental component, record its amplitude change rate and phase stability. For higher-order harmonic components, calculate their amplitude ratio and phase difference with the fundamental.

[0066] S222. Construct a three-dimensional harmonic feature tensor containing the fundamental distortion rate and the odd-even harmonic energy ratio. The fundamental distortion rate is defined as the ratio of the higher harmonic amplitude to the fundamental amplitude; the odd-even harmonic energy ratio is defined as the ratio of the odd harmonic energy to the even harmonic energy. The dimension of the three-dimensional harmonic feature tensor is [number of devices × harmonic feature dimension × time window length]. The harmonic feature dimension includes features such as the fundamental distortion rate, odd-even harmonic energy ratio, harmonic amplitude ratio, and phase difference. The time window length is 10 minutes.

[0067] In one embodiment, in step S200, the step of extracting environmental features from environmental sensing data includes:

[0068] S221. Use a sliding window algorithm to calculate the short-term fluctuation index of temperature and humidity parameters. Set a 30-minute sliding time window, calculate the standard deviation and coefficient of variation of the temperature and humidity data within the window, and perform feature normalization based on the device heat dissipation coefficient. Specifically, the formula for calculating the short-term temperature fluctuation index is: Temperature Fluctuation Index = Temperature Standard Deviation × Device Temperature Sensitivity Coefficient / Temperature Average Value, where the device temperature sensitivity coefficient ranges from 0.8 to 1.5 depending on the device type. The formula for calculating the short-term humidity fluctuation index is: Humidity Fluctuation Index = Humidity Coefficient of Variation × Device Humidity Sensitivity Coefficient, where the device humidity sensitivity coefficient ranges from 0.6 to 1.2 depending on the device type.

[0069] S222. Establish the long-term degradation trend characteristics of the air pressure parameters based on the ARIMA model. First, perform differential processing on the air pressure data to eliminate non-stationarity, then determine the order of the ARIMA model through the autocorrelation function and partial autocorrelation function, and finally estimate the model parameters through the maximum likelihood estimation method. In this embodiment, the parameters of the ARIMA model are (2,1,1), that is, the autoregressive order p = 2, the difference order d = 1, and the moving average order q = 1. The long-term change trend of the air pressure is extracted through this model, including characteristics such as trend slope, periodicity, and abnormal fluctuations.

[0070] In one embodiment, in step S200, the step of fusing the extracted corresponding harmonic features and environmental features to generate fused features includes:

[0071] S230. Establish a spatiotemporal correlation matrix to perform multi-scale alignment of minute-level environmental data and millisecond-level electrical data. The rows of the spatiotemporal correlation matrix represent the time dimension, and the columns represent the spatial dimension (the location of different devices and sensors). For the environmental data, linear interpolation is used to interpolate the minute-level data into millisecond-level data. For the electrical data, a sliding average method is used to aggregate the millisecond-level data into minute-level data. Then, a corresponding relationship between the two types of data is established in the spatiotemporal correlation matrix.

[0072] S231. Use a dynamic time warping algorithm to eliminate clock skew between different environmental sensors. This algorithm calculates the optimal alignment path between two time series to eliminate time skew caused by sensor clock asynchrony. The algorithm uses the Euclidean distance metric and the path constraint is the Sakoe-Chiba bandwidth, which is set to 10% of the sequence length. This algorithm accurately aligns data collected by different sensors in the time dimension.

[0073] S232. Construct a three-dimensional fusion tensor containing harmonic time-frequency features and environmental trend features. The fusion tensor has a dimension of [number of devices × fusion feature dimension × time window length], where the fusion feature dimension includes a combination of harmonic and environmental features, and the time window length is 10 minutes. Harmonic time-frequency features include fundamental distortion rate and odd-even harmonic energy ratio; environmental trend features include short-term temperature and humidity fluctuation index and long-term air pressure degradation trend. The fusion method uses a combination of feature concatenation and weighted fusion, with weight coefficients obtained through historical data training.

[0074] In one embodiment, in step S400, the constructed electrical equipment fault diagnosis model includes a dual-channel feature encoder, a dynamic attention fusion module and a diagnosis decision layer.

[0075] The dual-channel feature encoder includes an electrical feature channel and an environmental feature channel. The electrical feature channel uses a temporal processing structure that combines 1D CNN and BiGRU, where 1D CNN is used to extract local time-frequency features and BiGRU is used to capture long-term temporal dependencies. The 1D CNN contains three convolutional layers with kernel sizes of 5, 3, and 3, and the number of convolution kernels is 32, 64, and 128, respectively. Each convolutional layer is followed by a BatchNormalization layer and a ReLU activation function. The BiGRU contains two layers, 256 hidden units, and a dropout rate of 0.3. The environmental feature channel uses a Transformer architecture to extract long-term dependencies. It contains four self-attention heads, a hidden layer dimension of 512, a feedforward network dimension of 2048, three layers, and a dropout rate of 0.1.

[0076] The dynamic attention fusion module generates a weight matrix for coupling environmental and electrical features. This module first calculates the correlation matrix between electrical and environmental features. It then converts this correlation matrix into an attention weight matrix using a softmax function. Finally, it performs a weighted fusion of the two feature types based on the attention weight matrix. The temperature parameter of the attention mechanism is set to 0.5, and the number of attention heads is 8.

[0077] The diagnostic decision layer combines the device's physical model to construct a cross-entropy loss function constrained by prior knowledge. The diagnostic decision layer consists of three fully connected layers, with 512 neurons, 256 neurons, and the number of output categories (number of fault types × number of fault severities). The activation function uses Reinforced Luminance (ReLU), and the final layer uses a softmax function to output the fault probability distribution. Prior knowledge constraints are implemented by adding a regularization term to the loss function. The regularization term is based on the relationship between fault modes and features defined by the device's physical model. The formula for the cross-entropy loss function is: L = -∑(y_true*log(y_pred))+λ*R(θ,M), where y_true is the true label, y_pred is the predicted label, λ is the regularization coefficient (set to 0.1), and R(θ,M) is the regularization term for the model parameter θ based on the physical model M.

[0078] The model was trained using the Adam optimizer with an initial learning rate of 0.001, a batch size of 64, and 100 training epochs. An early stopping strategy was used to prevent overfitting, with a patience value of 10. The learning rate adjustment strategy was to reduce the learning rate by 0.5 if the validation set loss did not decrease for five consecutive epochs. During training, model performance was evaluated on the validation set every 10 epochs, with metrics such as accuracy, precision, recall, and F1 score recorded. The model with the best performance on the validation set was selected as the trained electrical equipment fault diagnosis model.

[0079] Example 2:

[0080] This embodiment provides an intelligent diagnostic system for electrical equipment in a hydropower station, including:

[0081] A data acquisition module, configured to acquire a historical data set of the target hydropower station, including current measurement data of each electrical device, environmental sensor data, and equipment fault records; this module is configured to implement the function of step S100 in the first embodiment and will not be described in detail here;

[0082] A feature extraction module is used to extract harmonic features from the current measurement data of each electrical device, and to extract environmental features from the environmental sensor data, and to fuse the extracted corresponding harmonic features and environmental features to generate a fused feature. This module is used to implement the function of step S200 in Example 1 and is not further described here.

[0083] A sample construction module is used to extract device fault labels from device fault records and construct a sample set based on the fusion features and the device fault labels. This module is used to implement the function of step S300 in the first embodiment and will not be repeated here.

[0084] A model training module is used to construct a neural network-based electrical equipment fault diagnosis model, train the electrical equipment fault diagnosis model using a sample set, and obtain a trained electrical equipment fault diagnosis model; this module is used to implement the function of step S400 in Example 1 and will not be described in detail here;

[0085] The fault diagnosis module is used to obtain the real-time environmental data of the target hydropower station and the real-time current measurement data of each electrical device, and input the data into the trained electrical equipment fault diagnosis model after feature extraction and feature fusion to obtain the fault diagnosis results of each electrical device; this module is used to implement the function of step S500 in Example 1 and will not be repeated here.

[0086] As a preferred implementation of this embodiment, the step of extracting harmonic features from the current measurement data of each electrical device includes:

[0087] The non-stationary current waveform is extracted from the current measurement data, and the improved CEEMDAN algorithm is used to adaptively decompose the non-stationary current waveform into several effective intrinsic mode components.

[0088] For each effective eigenmode component, the instantaneous amplitude and phase parameters of each harmonic are calculated by Hilbert transform;

[0089] The instantaneous amplitude and phase parameters of each harmonic are constructed as harmonic features.

[0090] As a preferred implementation of this embodiment, in the step of extracting environmental features from the environmental sensor data, the environmental features include short-term fluctuation indexes of temperature and humidity data and long-term features of air pressure data;

[0091] The short-term fluctuation index of the temperature and humidity data is obtained by performing a sliding window calculation on the temperature and humidity data in the environmental sensor data;

[0092] The long-term characteristics of the air pressure data are obtained by analyzing the air pressure data in the environmental sensor data using a pre-established ARIMA model.

[0093] As a preferred implementation of this embodiment, the step of fusing the extracted corresponding harmonic features and environmental features to generate fused features includes:

[0094] Establish a spatiotemporal correlation matrix, where the rows of the spatiotemporal correlation matrix represent the time dimension and the columns represent the spatial dimension of the electrical equipment;

[0095] Preprocess the harmonic features and environmental features to make their time scales consistent, and establish the corresponding relationship between the harmonic features and environmental features in the spatiotemporal correlation matrix;

[0096] Based on the correspondence between harmonic features and environmental features, a fusion feature containing harmonic features and environmental features is constructed.

[0097] Example 3:

[0098] This embodiment proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the intelligent diagnosis method for electrical equipment of a hydropower station as described in any embodiment of the present invention is implemented.

[0099] Example 4:

[0100] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the intelligent diagnosis method for electrical equipment of a hydropower station as described in any embodiment of the present invention is implemented.

[0101] In the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can represent: a, b, c, a and b, a and c, b and c or a and b and c, where a, b, c can be single or multiple.

[0102] Those skilled in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented using a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0103] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0104] In the several embodiments provided in this application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of this application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), magnetic disk or optical disk, and other media that can store program code.

[0105] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention's description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An intelligent diagnostic method for electrical equipment in a hydropower station, characterized in that: The following steps are involved: Obtain historical data sets for the target hydropower station, including current measurement data of each electrical device, environmental sensor data, and equipment failure records; Harmonic features are extracted from the current measurement data of each electrical device, and environmental features are extracted from the environmental sensor data. The extracted corresponding harmonic features and environmental features are then fused to generate fused features. Extract device fault labels from device fault records and construct a sample set based on fusion features and device fault labels; Construct an electrical equipment fault diagnosis model based on a neural network, train the electrical equipment fault diagnosis model through a sample set, and obtain a trained electrical equipment fault diagnosis model; The real-time environmental data of the target hydropower station and the real-time current measurement data of each electrical equipment are obtained. After feature extraction and feature fusion, they are input into the trained electrical equipment fault diagnosis model to obtain the fault diagnosis results of each electrical equipment.

2. The intelligent diagnosis method for electrical equipment of a hydropower station according to claim 1, characterized in that: The step of extracting harmonic features from the current measurement data of each electrical device includes: The non-stationary current waveform is extracted from the current measurement data, and the improved CEEMDAN algorithm is used to adaptively decompose the non-stationary current waveform into several effective intrinsic mode components. For each effective eigenmode component, the instantaneous amplitude and phase parameters of each harmonic are calculated by Hilbert transform; The instantaneous amplitude and phase parameters of each harmonic are constructed as harmonic features.

3. The intelligent diagnosis method for electrical equipment of a hydropower station according to claim 1, characterized in that: In the step of extracting environmental features from the environmental sensor data, the environmental features include short-term fluctuation indexes of temperature and humidity data and long-term features of air pressure data; The short-term fluctuation index of the temperature and humidity data is obtained by performing a sliding window calculation on the temperature and humidity data in the environmental sensor data; The long-term characteristics of the air pressure data are obtained by analyzing the air pressure data in the environmental sensor data using a pre-established ARIMA model.

4. The intelligent diagnosis method for electrical equipment of a hydropower station according to claim 1, characterized in that: The step of fusing the extracted corresponding harmonic features and environmental features to generate fused features includes: Establish a spatiotemporal correlation matrix, where the rows of the spatiotemporal correlation matrix represent the time dimension and the columns represent the spatial dimension of the electrical equipment; Preprocess the harmonic features and environmental features to make their time scales consistent, and establish the corresponding relationship between the harmonic features and environmental features in the spatiotemporal correlation matrix; Based on the correspondence between harmonic features and environmental features, a fusion feature containing harmonic features and environmental features is constructed.

5. An intelligent diagnostic system for electrical equipment in a hydropower station, characterized in that: include: The data acquisition module is used to obtain historical data sets of the target hydropower station, including current measurement data of each electrical device, environmental sensor data, and equipment fault records; The feature extraction module is used to extract harmonic features from the current measurement data of each electrical device, and extract environmental features from the environmental sensor data, and fuse the extracted corresponding harmonic features and environmental features to generate fused features; The sample construction module is used to extract device fault labels from device fault records and construct a sample set based on fusion features and device fault labels; The model training module is used to build an electrical equipment fault diagnosis model based on a neural network, and train the electrical equipment fault diagnosis model through a sample set to obtain a trained electrical equipment fault diagnosis model; The fault diagnosis module is used to obtain the real-time environmental data of the target hydropower station and the real-time current measurement data of each electrical equipment. After feature extraction and feature fusion, the data is input into the trained electrical equipment fault diagnosis model to obtain the fault diagnosis results of each electrical equipment.

6. The intelligent diagnostic system for electrical equipment of a hydropower station according to claim 5, characterized in that: The step of extracting harmonic features from the current measurement data of each electrical device includes: The non-stationary current waveform is extracted from the current measurement data, and the improved CEEMDAN algorithm is used to adaptively decompose the non-stationary current waveform into several effective intrinsic mode components. For each effective eigenmode component, the instantaneous amplitude and phase parameters of each harmonic are calculated by Hilbert transform; The instantaneous amplitude and phase parameters of each harmonic are constructed as harmonic features.

7. The intelligent diagnostic system for electrical equipment of a hydropower station according to claim 5, characterized in that: In the step of extracting environmental features from the environmental sensor data, the environmental features include short-term fluctuation indexes of temperature and humidity data and long-term features of air pressure data; The short-term fluctuation index of the temperature and humidity data is obtained by performing a sliding window calculation on the temperature and humidity data in the environmental sensor data; The long-term characteristics of the air pressure data are obtained by analyzing the air pressure data in the environmental sensor data using a pre-established ARIMA model.

8. The intelligent diagnostic system for electrical equipment of a hydropower station according to claim 5, characterized in that: The step of fusing the extracted corresponding harmonic features and environmental features to generate fused features includes: Establish a spatiotemporal correlation matrix, where the rows of the spatiotemporal correlation matrix represent the time dimension and the columns represent the spatial dimension of the electrical equipment; Preprocess the harmonic features and environmental features to make their time scales consistent, and establish the corresponding relationship between the harmonic features and environmental features in the spatiotemporal correlation matrix; Based on the correspondence between harmonic features and environmental features, a fusion feature containing harmonic features and environmental features is constructed.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the intelligent diagnosis method for electrical equipment of a hydropower station according to any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the intelligent diagnosis method for electrical equipment of a hydropower station as claimed in any one of claims 1 to 4 is implemented.

Images