Intelligent power grid monitoring alarm method and system

Abstract

The invention relates to the technical field of intelligent power grids, in particular to an intelligent power grid monitoring alarm method and system. According to the method, the traditional image analysis technology is not adopted, but the fault prediction of the power grid equipment is carried out based on the real-time operation data of the power grid, including current, voltage, temperature and humidity data. A prediction mode of combining an ANN prediction model and an LSTM prediction model is adopted, static features of collected data can be extracted, time sequence data can be associated, features of a time dimension can be reserved, and prediction precision is improved. And the health degree of the power grid in a period of time in the future can be evaluated based on the predicted data and the equipment state, and alarm information is sent to related personnel, so that the related personnel can be pre-warned about the equipment fault at the next moment and the state of the power grid in the period of time in the future, and the effect of pre-judgment in advance is achieved.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and in particular to a smart grid monitoring alarm method and system. Background Art

[0002] With the advancement of technology and the accumulation of big data, the fault detection and prediction method of smart grid based on artificial intelligence has become a key technology for the safe and stable operation of the power grid. Artificial intelligence has the advantages of real-time, high accuracy, big data processing capabilities and comprehensive analysis and prediction, providing a timely, efficient and reliable fault detection and prediction solution for smart grids. With the rapid development of digitalization, automation and Internet technology, the power system is also undergoing a profound change. The traditional power grid system is gradually transforming into a smart grid, which provides a more efficient, flexible and reliable power supply. But at the same time, the complexity of the power grid has also increased, resulting in more faults and unpredictable challenges. Therefore, real-time monitoring and prediction of the power grid is crucial to ensure its stable and safe operation.

[0003] Although existing technologies have made some progress in power grid fault detection and prediction, there are still some obvious shortcomings. Traditional inspection image analysis technology usually uses classical image processing algorithms. The manual features extracted by these algorithms are basically low-level visual features. Compared with the features of pure learning methods, the features are more interpretable but less adaptable to data. In addition, the operating environment of the power grid is complex and changeable. Fault detection and prediction often encounter fuzzy and uncertain situations, and existing technologies have limited ability to handle such fuzzy and uncertain information. Summary of the invention

[0004] In view of the problems in the prior art, the present invention provides a smart grid monitoring alarm method and system, and the specific technical solution is as follows:

[0005] A smart grid monitoring and alarm method comprises the following steps:

[0006] Step S1, collecting real-time operation data of the power grid, including current, voltage, temperature and humidity data;

[0007] Step S2, processing and analyzing the collected data to obtain processed current, voltage, temperature and humidity data, as well as analyzed change trend data of the current, voltage, temperature and humidity;

[0008] Step S3, inputting the collected data into the trained ANN prediction model to obtain a first prediction result; the first prediction result includes predicting the current, voltage, temperature and humidity data at the next moment;

[0009] Step S4, inputting the change trend data of current, voltage, temperature and humidity into the trained LSTM prediction model to obtain a second prediction result; the second prediction result includes the predicted current, voltage, temperature and humidity data at the next moment, and the change trend data of current, voltage, temperature and humidity starting from the next moment;

[0010] Step S5, splicing the first prediction result and the second prediction result to obtain the final current, voltage, temperature and humidity data at the next moment, whether the equipment is faulty at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment;

[0011] Step S6, sending an alarm message to relevant personnel when predicting equipment failure at the next moment;

[0012] Step S7, accumulate the cumulative number of failures of the equipment according to the prediction information, and evaluate the health of the power grid in the future period and predict the cumulative number of failures according to the change trend data of current, voltage, temperature and humidity starting from the next moment. When the health is less than the set threshold or the predicted cumulative number of failures reaches the set threshold, an alarm message is sent to relevant personnel.

[0013] Preferably, the processing of the collected data in step S2 includes cleaning, outlier screening, missing value filling, and standardization.

[0014] Preferably, the specific steps of ANN prediction model training in step S3 are as follows:

[0015] Data collection and processing: Collect and process the historical operation data of the power grid, and divide the historical operation data into a training set, a validation set, and a test set; the historical operation data includes current, voltage, temperature, humidity data, and whether the equipment is faulty; Initialization: Establish an ANN prediction model, which includes an input layer, two hidden layers, and an output layer, and use random numbers to initialize the weights and biases of each layer of the ANN prediction model;

[0016] Forward propagation: input the training set and validation set into the initialized ANN prediction model to calculate the output of each neuron;

[0017] Loss calculation: Calculate the value of the loss function to evaluate the gap between the model prediction result and the true value;

[0018] Backpropagation: Calculate the gradient of the loss function with respect to each weight; Update weights and biases using gradient descent;

[0019] Iterative training: Repeat forward propagation, loss calculation, and back propagation until the prediction accuracy of the ANN prediction model converges or reaches the preset number of training times to obtain a trained ANN prediction model;

[0020] Model evaluation: Use the test set to evaluate the trained ANN prediction model, calculate the prediction accuracy and evaluation index, and output the trained ANN prediction model when the prediction accuracy is greater than or equal to the set value and the evaluation index meets the set conditions, otherwise retrain.

[0021] Preferably, the specific steps of LSTM prediction model training in step S4 are as follows:

[0022] Data collection and processing: Collect historical operation data of the power grid and process and analyze it, and divide the historical operation data into training set, verification set and test set; the historical operation data includes current, voltage, temperature and humidity data and whether the equipment is faulty; analyze and obtain the change trend data of historical current, voltage, temperature and humidity;

[0023] Initialization: Establish an LSTM prediction model, which includes an input layer, two LSTM layers and an output layer, and use random numbers to initialize the weights and biases of each layer of the LSTM prediction model;

[0024] Forward propagation: Input the training set and validation set into the initialized LSTM prediction model and calculate the output of each LSTM element;

[0025] Loss calculation: Calculate the value of the loss function to evaluate the gap between the model prediction result and the true value;

[0026] Backpropagation: Calculate the gradient of the loss function with respect to each weight; Update weights and biases using gradient descent;

[0027] Iterative training: Repeat forward propagation, loss calculation, and back propagation until the prediction accuracy of the LSTM prediction model converges or reaches the preset number of training times to obtain a trained LSTM prediction model;

[0028] Model evaluation: Use the test set to evaluate the trained LSTM prediction model, calculate the prediction accuracy and evaluation index, and output the trained LSTM prediction model when the prediction accuracy is greater than or equal to the set value and the evaluation index meets the set conditions, otherwise retrain.

[0029] Preferably, the step S5 of combining the first prediction result and the second prediction result specifically includes the following steps:

[0030] Assume that the prediction result at time t+1 is X(t+1), then:

[0031] X(t+1)=ω1X1(t+1)+ω2X2(t+1);

[0032]

[0033] Among them, X1(t+1) is the predicted value of current / voltage / temperature / humidity in the first prediction result of the ANN prediction model; ω1 is the allocation weight of X1(t+1); X2(t+1) is the predicted value of current / voltage / temperature / humidity in the second prediction result of the LSTM prediction model; ω2 is the allocation weight of X2(t+1); D1 is the prediction accuracy of the ANN prediction model; D2 is the prediction accuracy of the LSTM prediction model.

[0034] Preferably, the health of the power grid in step S7 is calculated as follows:

[0035]

[0036] Among them, J is the health of the power grid; Z i is the predicted state variable of device i, which is a 0-1 variable. It is 0 when the device fails and 1 otherwise. n is the number of power grid devices.

[0037] Preferably, step S8 is also included, in which a large-screen visualization is performed to display whether the equipment is faulty at the next moment, the changing trend data of the current, voltage, temperature and humidity starting from the next moment, the health of the power grid in the future period and the predicted cumulative number of faults.

[0038] A smart grid monitoring and alarm system, applied to the method described, comprising:

[0039] Data acquisition module, used to collect real-time operation data of the power grid, including current, voltage, temperature and humidity data;

[0040] The data processing and analysis module is used to process and analyze the collected data to obtain the processed current, voltage, temperature and humidity data, as well as the change trend data of the current, voltage, temperature and humidity obtained by analysis;

[0041] A first prediction module, used to input the collected data into a trained ANN prediction model to obtain a first prediction result; the first prediction result includes predicting the current, voltage, temperature and humidity data at the next moment;

[0042] The second prediction module is used to input the change trend data of current, voltage, temperature and humidity into the trained LSTM prediction model to obtain a second prediction result; the second prediction result includes the predicted current, voltage, temperature and humidity data at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment;

[0043] A result splicing module is used to splice the first prediction result and the second prediction result to obtain the final current, voltage, temperature and humidity data at the next moment, whether the equipment is faulty at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment;

[0044] Prediction alarm module, used to send alarm information to relevant personnel when predicting equipment failure at the next moment;

[0045] The evaluation and alarm module is used to accumulate the cumulative number of equipment failures according to the prediction information, and evaluate the health of the power grid in the future period and predict the cumulative number of failures based on the change trend data of current, voltage, temperature and humidity starting from the next moment. When the health is less than the set threshold or the predicted cumulative number of failures reaches the set threshold, an alarm message is sent to relevant personnel.

[0046] A computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the smart grid monitoring and alarm method.

[0047] A processor is used to run a program, wherein the smart grid monitoring and alarm method is executed when the program is running.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The present invention provides a smart grid monitoring and alarm method and system, which does not use traditional image analysis technology, but predicts the faults of power grid equipment based on the real-time operation data of the power grid, including current, voltage, temperature and humidity data. The prediction method using the ANN prediction model and the LSTM prediction model combined can not only extract the static features of the collected data, but also be associated with the time series data, retain the features of the time dimension, and improve the prediction accuracy. Moreover, based on the predicted data and equipment status, the health of the power grid in the future period can be evaluated, and alarm information can be sent to relevant personnel, which can warn relevant personnel of the equipment failure at the next moment and the status of the power grid in the future period, achieving the effect of early prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0051] Figure 1 The present invention is a flow chart of the method.

[0052] Figure 2 It is a system principle diagram of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0055] It should also 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, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0056] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0057] Embodiment 1:

[0058] like Figure 1 As shown, this embodiment provides a smart grid monitoring alarm method, comprising the following steps:

[0059] Step S1, collect real-time operation data of the power grid, including current, voltage, temperature and humidity data. Among them, voltage (V): collect the voltage value of each node of the power grid to monitor voltage fluctuations and anomalies; current (A): collect current data to monitor current anomalies and load changes; temperature (℃): collect the temperature of key equipment (such as transformers, distribution cabinets) to monitor overheating; humidity (%): collect environmental humidity data to monitor the impact of humidity on equipment operation. The method of data collection is to install sensors on the equipment to continuously collect data. Different data requires the use of different sensors, temperature and humidity sensors, current transformers, and voltage transformers. Data acquisition equipment should have high precision, high reliability and real-time transmission capabilities.

[0060] Step S2, processing and analyzing the collected data to obtain processed current, voltage, temperature and humidity data, as well as analyzed change trend data of the current, voltage, temperature and humidity.

[0061] The processing of the collected data includes cleaning, outlier screening, missing value filling, and standardization.

[0062] Among them, cleaning includes removing noisy data.

[0063] Outlier removal is the removal of outliers (e.g., abnormal readings due to sensor failures) using statistical methods (e.g.,

[0064] 3σ principle) or machine learning methods to detect and handle outliers.

[0065] Missing value filling uses interpolation or mean filling methods to fill in missing values.

[0066] Normalization is to scale the data to the range of [0,1] or [-1,1] to speed up model training and improve convergence speed.

[0067] Min-Max normalization or Z-score normalization method.

[0068] Step S3, inputting the collected data into the trained ANN prediction model to obtain a first prediction result; the first prediction result includes predicting the current, voltage, temperature and humidity data at the next moment.

[0069] The specific steps of ANN prediction model training are as follows:

[0070] (1) Data collection and processing: The historical operation data of the power grid is collected and processed, and the historical operation data is divided into a training set, a validation set, and a test set; the historical operation data includes current, voltage, temperature, humidity data, and whether the equipment is faulty.

[0071] (2) Initialization: An ANN prediction model is established, wherein the ANN prediction model includes an input layer, two hidden layers, and an output layer, and the weights and biases of each layer of the ANN prediction model are initialized using random numbers; the hidden layer uses a tanh activation function, and the output layer uses a linear activation function.

[0072] (3) Forward propagation: The training set and validation set are input into the initialized ANN prediction model to calculate the output of each neuron.

[0073] (4) Loss calculation: Calculate the value of the loss function to evaluate the gap between the model prediction result and the true value.

[0074] (5) Backpropagation: Calculate the gradient of the loss function with respect to each weight; use gradient descent to update weights and biases.

[0075] (6) Iterative training: Repeat forward propagation, loss calculation, and back propagation until the prediction accuracy of the ANN prediction model converges or reaches the preset number of training times to obtain a trained ANN prediction model.

[0076] (7) Model evaluation: Use the test set to evaluate the trained ANN prediction model and calculate the prediction accuracy and evaluation index. When the prediction accuracy is greater than or equal to the set value and the evaluation index meets the set conditions, the trained ANN prediction model is output. Otherwise, the model's hyperparameters, such as learning rate, batch size, number of network layers, and number of units, are adjusted and retrained.

[0077] Evaluation indicators include mean square error (MSE), mean absolute error (MAE), etc. The accuracy and robustness of the model are evaluated to ensure its reliability in practical applications.

[0078] Step S4, input the changing trend data of current, voltage, temperature and humidity into the trained LSTM prediction model to obtain a second prediction result; the second prediction result includes the predicted current, voltage, temperature and humidity data at the next moment, and the changing trend data of current, voltage, temperature and humidity starting from the next moment.

[0079] The specific steps of LSTM prediction model training are as follows:

[0080] (1) Data collection and processing: The historical operation data of the power grid is collected and processed and analyzed, and the historical operation data is divided into a training set, a validation set, and a test set; the historical operation data includes current, voltage, temperature, and humidity data, as well as the status of whether the equipment is faulty; the change trend data of the historical current, voltage, temperature, and humidity are obtained by analysis. Specifically, the sliding window method can be used to generate time series data to facilitate the LSTM model to perform time series prediction.

[0081] (2) Initialization: Establish an LSTM prediction model, which includes an input layer, two LSTM layers and an output layer, and use random numbers to initialize the weights and biases of each layer of the LSTM prediction model; the LSTM layer uses a tanh activation function.

[0082] (3) Forward propagation: Input the training set and validation set into the initialized LSTM prediction model and calculate the output of each LSTM element.

[0083] (4) Loss calculation: Calculate the value of the loss function to evaluate the gap between the model prediction result and the true value;

[0084] Backpropagation: Calculate the gradient of the loss function with respect to each weight; Update weights and biases using gradient descent;

[0085] (5) Iterative training: Repeat forward propagation, loss calculation, and back propagation until the prediction accuracy of the LSTM prediction model converges or reaches the preset number of training times, and obtain a trained LSTM prediction model;

[0086] (6) Model evaluation: Use the test set to evaluate the trained LSTM prediction model and calculate the prediction accuracy and evaluation index. When the prediction accuracy is greater than or equal to the set value and the evaluation index meets the set conditions, the trained LSTM prediction model is output. Otherwise, adjust the model's hyperparameters, such as learning rate, batch size, number of network layers, and number of units, and retrain.

[0087] ANN is used to process static features (such as voltage, current, temperature, and humidity at the current moment). ANN can learn nonlinear relationships and extract high-level feature representations. LSTM is used to process time series data (such as the trend of voltage, current, temperature, and humidity at historical moments). LSTM can capture long-term dependencies in time series and is suitable for predicting future states. The outputs of ANN and LSTM are fused (such as weighted fusion or splicing) to comprehensively consider static features and time series features.

[0088] Step S5, concatenate the first prediction result and the second prediction result to obtain the final current, voltage, temperature and humidity data at the next moment, whether the equipment is faulty at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment.

[0089] The splicing of the first prediction result and the second prediction result specifically includes the following steps:

[0090] Assume that the prediction result at time t+1 is X(t+1), then:

[0091] X(t+1)=ω1X1(t+1)+ω2X2(t+1);

[0092]

[0093] Among them, X1(t+1) is the predicted value of current / voltage / temperature / humidity in the first prediction result of the ANN prediction model; ω1 is the allocation weight of X1(t+1); X2(t+1) is the predicted value of current / voltage / temperature / humidity in the second prediction result of the LSTM prediction model; ω2 is the allocation weight of X2(t+1); D1 is the prediction accuracy of the ANN prediction model; D2 is the prediction accuracy of the LSTM prediction model. In the present invention, the higher the prediction accuracy, the greater the proportion of its prediction results.

[0094] Step S6, sending an alarm message to relevant personnel when predicting equipment failure at the next moment.

[0095] Step S7, accumulate the cumulative number of failures of the equipment according to the prediction information, and evaluate the health of the power grid in the future period and predict the cumulative number of failures according to the change trend data of current, voltage, temperature and humidity starting from the next moment. When the health is less than the set threshold or the predicted cumulative number of failures reaches the set threshold, an alarm message is sent to relevant personnel.

[0096] The health of the power grid is calculated as follows:

[0097]

[0098] Among them, J is the health of the power grid; Z i is the predicted state variable of device i, which is a 0-1 variable. It is 0 when the device fails and 1 otherwise. n is the number of power grid devices.

[0099] Step S8: Visualize the large screen display of whether the equipment will fail at the next moment, the trend data of current, voltage, temperature and humidity starting from the next moment, the health of the power grid in the future and the predicted cumulative number of failures. Specifically, data visualization technology can be used, such as charts and heat maps.

[0100] The present invention specifically provides a graphical interface to display the power grid operation status and prediction results, and supports the visual display of real-time data and historical data.

[0101] The present invention has the following characteristics:

[0102] High-precision prediction: Combining the advantages of ANN and LSTM, it can more accurately predict changes in power grid status.

[0103] Real-time: Real-time data collection and analysis can detect and handle anomalies and faults in a timely manner.

[0104] The present invention realizes comprehensive monitoring of the operating status of the power grid and accurate prediction of future trends by combining real-time data acquisition and prediction models. First, by collecting key parameters such as current, voltage, temperature and humidity in real time, it is possible to ensure accurate grasp of the current state of the power grid. The real-time nature of these data is crucial for rapid response to emergencies in the power grid. Secondly, the fault identification model trained using these real-time data can timely identify abnormal conditions in the power grid, providing the possibility for fault prevention and rapid response. Furthermore, the data prediction model predicts future power grid parameters and equipment status based on historical and real-time data, allowing operation and maintenance personnel to prepare in advance and take preventive measures, thereby reducing potential failures and losses. This combination of real-time and predictive properties not only improves the reliability of the power grid, but also provides strong support for the intelligent management of the power grid.

[0105] The present invention combines the ANN model with the LSTM model, and can learn complex patterns and associations from a large amount of historical data, which are difficult to capture with traditional methods. Through the continuous optimization of the training set and the adjustment of the model parameters, the accuracy of the model can be continuously improved. In addition, the real-time alarm mechanism and the record of the cumulative number of faults provide instant feedback for fault diagnosis, allowing operation and maintenance personnel to respond quickly and shorten the time for troubleshooting. This automated and intelligent fault diagnosis process not only improves efficiency, but also reduces the possibility of human error, providing a guarantee for the stable operation of the power grid.

[0106] By predicting the possible failures that may occur in the future, the operation and maintenance team can dispatch technicians and materials in advance to achieve more efficient maintenance operations. For example, for those equipments that are predicted to be high-risk, inspection and maintenance can be arranged in priority to avoid or reduce the occurrence of failures. At the same time, by comparing the statistics of the cumulative number of failures with the actual number of failures, the accuracy of the fault identification model can be evaluated, and then the model can be adjusted and optimized to improve the accuracy of the prediction. This data-driven decision support system makes resource allocation more scientific and reasonable, and maintenance strategies more accurate and effective, thereby reducing maintenance costs and improving the operating efficiency of the power grid.

[0107] The present invention enhances the adaptability and resilience of the power grid by real-time monitoring and prediction of the power grid status. When faced with a complex and changeable power grid operating environment, it can quickly adapt to various emergencies, such as weather changes, load fluctuations, etc., and adjust the power grid operation strategy in time to ensure the stable power supply of the power grid. The introduction of the prediction model makes power grid management more proactive rather than passively responding to failures. By predicting the future state of the power grid, measures can be taken in advance to avoid or mitigate the impact of failures, thereby improving the power grid's ability to resist various risks. In addition, the real-time alarm of the present invention provides data support for the health management of the power grid, which helps to promptly discover and solve potential problems, thereby extending the service life of equipment, reducing the occurrence of large-scale failures, and enhancing the overall resilience of the power grid.

[0108] Embodiment 2:

[0109] like Figure 2 As shown, based on the same inventive concept as that of Embodiment 1, this embodiment provides a smart grid monitoring and alarm system, which is applied to the method described, and includes:

[0110] Data acquisition module, used to collect real-time operation data of the power grid, including current, voltage, temperature and humidity data;

[0111] The data processing and analysis module is used to process and analyze the collected data to obtain the processed current, voltage, temperature and humidity data, as well as the change trend data of the current, voltage, temperature and humidity obtained by analysis;

[0112] A first prediction module, used to input the collected data into a trained ANN prediction model to obtain a first prediction result; the first prediction result includes predicting the current, voltage, temperature and humidity data at the next moment;

[0113] The second prediction module is used to input the change trend data of current, voltage, temperature and humidity into the trained LSTM prediction model to obtain a second prediction result; the second prediction result includes the predicted current, voltage, temperature and humidity data at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment;

[0114] A result splicing module is used to splice the first prediction result and the second prediction result to obtain the final current, voltage, temperature and humidity data at the next moment, whether the equipment is faulty at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment;

[0115] Prediction alarm module, used to send alarm information to relevant personnel when predicting equipment failure at the next moment;

[0116] The evaluation and alarm module is used to accumulate the cumulative number of equipment failures according to the prediction information, and evaluate the health of the power grid in the future period and predict the cumulative number of failures based on the change trend data of current, voltage, temperature and humidity starting from the next moment. When the health is less than the set threshold or the predicted cumulative number of failures reaches the set threshold, an alarm message is sent to relevant personnel.

[0117] Embodiment 3:

[0118] Based on the same inventive concept as Example 1, this embodiment provides a computer-readable storage medium, which includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the smart grid monitoring and alarm method.

[0119] Embodiment 4:

[0120] Based on the same inventive concept as that of Embodiment 1, this embodiment provides a processor, which is used to run a program, wherein the smart grid monitoring and alarm method is executed when the program is running.

[0121] Those of ordinary skill in the art will appreciate that the modules of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0122] In the embodiments provided by the present invention, it should be understood that the division of modules is only a logical function division, and there may be other division methods in actual implementation, for example, multiple modules can be combined into one module, one module can be split into multiple modules, or some features can be ignored, etc.

[0123] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0124] If the integrated module is implemented in the form of a software function module 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 the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-0nlyMemory), random access memory (RAM, RandomAccessMemory), mobile hard disk, magnetic disk or optical disk, etc., various media that can store program codes.

[0125] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A smart grid monitoring and alarm method, characterized in that: The following steps are involved: Step S1, collecting real-time operation data of the power grid, including current, voltage, temperature and humidity data; Step S2, processing and analyzing the collected data to obtain processed current, voltage, temperature and humidity data, as well as analyzed change trend data of the current, voltage, temperature and humidity; Step S3, inputting the collected data into the trained ANN prediction model to obtain a first prediction result; the first prediction result includes predicting the current, voltage, temperature and humidity data at the next moment; Step S4, inputting the change trend data of current, voltage, temperature and humidity into the trained LSTM prediction model to obtain a second prediction result; the second prediction result includes the predicted current, voltage, temperature and humidity data at the next moment, and the change trend data of current, voltage, temperature and humidity starting from the next moment; Step S5, splicing the first prediction result and the second prediction result to obtain the final current, voltage, temperature and humidity data at the next moment, whether the equipment is faulty at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment; Step S6, sending an alarm message to relevant personnel when predicting equipment failure at the next moment; Step S7, accumulate the cumulative number of failures of the equipment according to the prediction information, and evaluate the health of the power grid in the future period and predict the cumulative number of failures according to the change trend data of current, voltage, temperature and humidity starting from the next moment. When the health is less than the set threshold or the predicted cumulative number of failures reaches the set threshold, an alarm message is sent to relevant personnel.

2. A smart grid monitoring and alarm method according to claim 1, characterized in that: The processing of the collected data in step S2 includes cleaning, outlier screening, missing value filling, and standardization.

3. A smart grid monitoring and alarm method according to claim 1, characterized in that: The specific steps of ANN prediction model training in step S3 are as follows: Data collection and processing: Collect and process the historical operation data of the power grid, and divide the historical operation data into a training set, a validation set, and a test set; the historical operation data includes current, voltage, temperature, humidity data, and whether the equipment is faulty; Initialization: Establishing an ANN prediction model, the ANN prediction model includes an input layer, two hidden layers and an output layer, and using random numbers to initialize the weights and biases of each layer of the ANN prediction model; Forward propagation: input the training set and validation set into the initialized ANN prediction model to calculate the output of each neuron; Loss calculation: Calculate the value of the loss function to evaluate the gap between the model prediction result and the true value; Backpropagation: Calculate the gradient of the loss function with respect to each weight; Update weights and biases using gradient descent; Iterative training: Repeat forward propagation, loss calculation, and back propagation until the prediction accuracy of the ANN prediction model converges or reaches the preset number of training times to obtain a trained ANN prediction model; Model evaluation: Use the test set to evaluate the trained ANN prediction model, calculate the prediction accuracy and evaluation index, and output the trained ANN prediction model when the prediction accuracy is greater than or equal to the set value and the evaluation index meets the set conditions, otherwise retrain.

4. A smart grid monitoring and alarm method according to claim 1, characterized in that: The specific steps of LSTM prediction model training in step S4 are as follows: Data collection and processing: Collect historical operation data of the power grid and process and analyze it, and divide the historical operation data into training set, verification set and test set; the historical operation data includes current, voltage, temperature and humidity data and whether the equipment is faulty; Analyze and obtain historical current, voltage, temperature and humidity change trend data; Initialization: Establish an LSTM prediction model, which includes an input layer, two LSTM layers and an output layer, and use random numbers to initialize the weights and biases of each layer of the LSTM prediction model; Forward propagation: Input the training set and validation set into the initialized LSTM prediction model and calculate the output of each LSTM element; Loss calculation: Calculate the value of the loss function to evaluate the gap between the model prediction result and the true value; Backpropagation: Calculate the gradient of the loss function with respect to each weight; Update weights and biases using gradient descent; Iterative training: Repeat forward propagation, loss calculation, and back propagation until the prediction accuracy of the LSTM prediction model converges or reaches the preset number of training times to obtain a trained LSTM prediction model; Model evaluation: Use the test set to evaluate the trained LSTM prediction model, calculate the prediction accuracy and evaluation index, and output the trained LSTM prediction model when the prediction accuracy is greater than or equal to the set value and the evaluation index meets the set conditions, otherwise retrain.

5. The smart grid monitoring and alarm method according to claim 1, characterized in that: The step S5 of combining the first prediction result with the second prediction result specifically includes the following steps: Assume that the prediction result at time t+1 is X(t+1), then: X(t+1)=ω1X1(t+1)+ω2X2(t+1); Among them, X1(t+1) is the predicted value of current / voltage / temperature / humidity in the first prediction result of the ANN prediction model; ω1 is the allocation weight of X1(t+1); X2(t+1) is the predicted value of current / voltage / temperature / humidity in the second prediction result of the LSTM prediction model; ω2 is the allocation weight of X2(t+1); D1 is the prediction accuracy of the ANN prediction model; D2 is the prediction accuracy of the LSTM prediction model.

6. A smart grid monitoring and alarm method according to claim 1, characterized in that: The health of the power grid in step S7 is calculated as follows: Among them, J is the health of the power grid; Z i is the predicted state variable of device i, which is a 0-1 variable. It is 0 when the device fails and 1 otherwise. n is the number of power grid devices.

7. A smart grid monitoring and alarm method according to claim 1, characterized in that: It also includes step S8, which visualizes on a large screen whether the equipment will fail at the next moment, the changing trend data of current, voltage, temperature and humidity starting from the next moment, the health of the power grid in the future period and the predicted cumulative number of failures.

8. A smart grid monitoring and alarm system, characterized in that: The method applied to any one of claims 1 to 7 comprises: a data acquisition module for collecting real-time operation data of the power grid, including current, voltage, temperature and humidity data; The data processing and analysis module is used to process and analyze the collected data to obtain the processed current, voltage, temperature and humidity data, as well as the change trend data of the current, voltage, temperature and humidity obtained by analysis; A first prediction module, used to input the collected data into a trained ANN prediction model to obtain a first prediction result; the first prediction result includes predicting the current, voltage, temperature and humidity data at the next moment; The second prediction module is used to input the change trend data of current, voltage, temperature and humidity into the trained LSTM prediction model to obtain a second prediction result; the second prediction result includes the predicted current, voltage, temperature and humidity data at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment; A result splicing module is used to splice the first prediction result and the second prediction result to obtain the final current, voltage, temperature and humidity data at the next moment, whether the equipment is faulty at the next moment, and the change trend data of the current, voltage, temperature and humidity starting from the next moment; Prediction alarm module, used to send alarm information to relevant personnel when predicting equipment failure at the next moment; The evaluation and alarm module is used to accumulate the cumulative number of equipment failures according to the prediction information, and evaluate the health of the power grid in the future period and predict the cumulative number of failures based on the change trend data of current, voltage, temperature and humidity starting from the next moment. When the health is less than the set threshold or the predicted cumulative number of failures reaches the set threshold, an alarm message is sent to relevant personnel.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the smart grid monitoring and alarm method according to any one of claims 1 to 7.

10. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the smart grid monitoring and alarm method described in any one of claims 1 to 7.

Images