A fault identification and processing method based on big data analysis

Abstract

The application discloses a fault identification and processing method based on big data analysis, and relates to the technical field of fault identification.The application solves the problem that the fault identification in the prior art only stays in the stage of obtaining an identification result, lacks further analysis and processing of the identification result, and causes managers to lack data reference based on the fault when making decisions.The application collects power data and meteorological data of each collection point; then, a fault identification model is used for fault identification, the geographical position of the collection point where the fault is identified is recorded, and an alarm signal is sent; further, the collection points are sorted based on the fault occurrence frequency in a period, and a frequently-occurring fault position is obtained; finally, a target function is established by taking the station construction cost and distance difference as influencing factors, and a station construction strategy can be obtained by solving the target function with the minimum target function as the goal.The application can provide a scientific basis for the maintenance and optimization of a power grid, help managers make more reasonable decisions, and improve the efficiency and effect of power grid management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault identification, in particular to a fault identification and processing method based on big data analysis. BACKGROUND

[0002] Power grid fault identification is crucial for ensuring the safe and stable operation of power systems. As the scale and complexity of power systems continue to expand, the reliability and stability of the power grid face greater challenges. Fault identification technology can quickly and accurately detect and locate abnormal conditions in the power grid, thereby reducing power outage time, improving power supply reliability, and ensuring the normal operation of the social economy. It plays a crucial role in preventing large-scale power outages, reducing economic losses, and maintaining the long-term healthy operation of the power grid. In addition, the progress of power grid fault identification technology also helps to improve the self-healing ability of the power system, by automatically isolating fault areas and reconfiguring the network to reduce the impact of power outages. The progress of this technology not only can timely warn of possible faults, but also can effectively reduce the potential threat of faults to the safety and reliability of the power grid operation, providing technical support and protection for the stable power supply of the power system.

[0003] Although power grid fault identification technology has made significant progress, there are still some shortcomings. Currently, many fault identification systems only stop at the stage of obtaining identification results, lacking further analysis and processing of the identification results. This means that although the system can identify faults, it does not further process the identification information to facilitate daily management and decision-making by management personnel, making it difficult for management personnel to make decisions based on fault-based data references.

[0004] In view of this, a fault identification and processing method based on big data analysis is needed. SUMMARY

[0005] In view of the problem in the prior art that fault identification only stops at the stage of obtaining identification results, lacking further analysis and processing of the identification results, making it difficult for management personnel to make decisions based on fault-based data references, the present application provides a fault identification and processing method based on big data analysis, which can sort the fault occurrence frequency of each collection point within a period after the fault identification model identifies the fault, and then select the frequently occurring fault location. Finally, a target function is established with the station construction cost and distance difference as influencing factors, and the target function is minimized to solve the target function and output the station construction strategy. The specific technical scheme is as follows:

[0006] A fault identification and processing method based on big data analysis, comprising the following steps:

[0007] Collecting power data and meteorological data of each collection point, and recording the location information of each collection point;

[0008] The fault identification model is used to identify faults, output results of whether faults exist at each collection point, record geographical positions of the collection points where faults are identified, and send an alarm signal;

[0009] The collection points are sorted based on the number of faults occurring at each collection point in a period, and a sorting threshold is set, and geographical positions of the collection points whose sorting positions are in front of the sorting threshold are taken as frequently-occurring fault positions;

[0010] A target function is established by taking the station construction cost and distance difference as influencing factors, and a station construction strategy is output by minimizing the target function;

[0011] The target function is solved to obtain the station construction strategy.

[0012] Preferably, the target function is specifically as follows:

[0013]

[0014] In the formula, C 建站 is a station construction unit cost, m is a number of stations, d 故障常发位置_i is a longitude and latitude coordinate of an i-th frequently-occurring fault position, n is a number of collection points whose sorting positions are in front of the sorting threshold, a number of frequently-occurring fault positions, d 站点_j is a j-th station position, F(d 故障常发位置_i , d 站点_j ) is a distance difference between the i-th frequently-occurring fault position and the j-th station position, and μ1 and μ2 are weight coefficients.

[0015] Preferably, a calculation formula of the station construction cost is as follows:

[0016] C 建站 =C 建筑 +C 工具 +C 人力

[0017] In the formula, C 建筑 is a construction cost of station construction, i.e., a construction cost of station infrastructure itself; C 工具 is a tool cost of station construction, i.e., a cost of configuring tools in the station, and C 人力 is a personnel cost of station construction, i.e., a cost of configuring personnel in the station.

[0018] Preferably, a calculation formula of F(d 故障常发位置_i , d 站点_j ) is as follows:

[0019]

[0020] In the formula, R is a radius of the earth, lat i is a latitude of the i-th frequently-occurring fault position, and lat jis the latitude of the jth station location, lon i is the longitude of the ith frequent failure location, lon j is the longitude of the jth station location.

[0021] Preferably, if the latitude and longitude coordinates are in degrees, they need to be converted to radians first, and the conversion formula is:

[0022]

[0023] Preferably, constraints are set, including station number constraints:

[0024] 0 < m ≤ m max

[0025] In the formula, m max is the maximum number of stations.

[0026] Preferably, the constraints also include station cost constraints:

[0027]

[0028] In the formula, C max is the maximum total station cost.

[0029] Preferably, the training process of the fault identification model is as follows:

[0030] S1: Data collection: First, the power data and weather data of each collection point need to be collected, including voltage, current, power, device temperature, device environment humidity, weather temperature, weather humidity, wind speed, light intensity, and season. Data collection is the basis for fault prediction, and its correctness and rationality directly affect the model construction and accuracy;

[0031] S2: Data preprocessing: clean and preprocess the collected data, including denoising and normalization processing;

[0032] S3: Feature extraction: extract features that are helpful for fault identification from the preprocessed data, including statistical features and frequency domain features. Feature extraction uses automatic extraction technology DFS to construct features and select features based on filtering;

[0033] S4: Model construction: a neural network model is constructed using deep neural network DNN, which combines low-level features to form more abstract high-level representation attributes or features, thereby discovering the internal feature representation of the data;

[0034] S5: Model training: use the extracted features and labeled data to train the neural network model. In the model training process, the data set is divided into training set, validation set and test set;

[0035] S6: Model evaluation and optimization: Evaluate the performance of the model, including accuracy, recall, and optimize the model according to the evaluation results.

[0036] Preferably, the gradient descent algorithm is used to solve the objective function, and the specific steps are as follows:

[0037] S01: Select initial point: Select an initial point x0, which can be random or selected according to prior knowledge of the problem;

[0038] S02: Set learning rate: Set learning rate α;

[0039] S03: Set convergence threshold: Set convergence threshold ∈, determine when the algorithm stops;

[0040] S04: Calculate gradient: Calculate the gradient of the objective function f(x) at the current point x t The gradient is a vector pointing in the direction of the fastest growth of the function;

[0041] S05: Update parameters: Update parameters using gradient and learning rate, as follows:

[0042]

[0043] S06: Check convergence: Calculate the magnitude of parameter update ||x t+1 -x t ||, if the magnitude is less than the convergence threshold ∈, the algorithm is considered to have converged to the minimum value, and the algorithm is stopped, if the algorithm has not converged, repeat steps S04 to S06 until the convergence condition is met.

[0044] A processor for running a program, wherein the program performs the fault identification processing method based on big data analysis as described above when running.

[0045] Compared with the prior art, the beneficial effects of the present application are:

[0046] ​The application collects power data and meteorological data of each collection point, and records the position information of each collection point; uses a fault identification model to identify faults, outputs the result of whether each collection point has a fault, records the geographic position of the collection point where the fault is identified, and sends an alarm signal; sorts the collection points based on the number of faults occurring in each collection point within a period, and sets a sorting threshold, taking the geographic position of the collection points whose sorting positions are in front of the sorting threshold as the frequently-occurring fault position; finally, a target function is established with the station construction cost and distance difference as influencing factors, and the station construction strategy is obtained by solving the target function with the goal of minimizing the target function. The application realizes fast response and accurate positioning of power grid faults, thereby improving the reliability and safety of the power grid. Moreover, by analyzing the statistical characteristics of the fault data, scientific basis can be provided for the maintenance and optimization of the power grid, helping managers make more reasonable decisions and improving the efficiency and effectiveness of power grid management. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed in the specific embodiments or prior art description will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale.

[0048] Figure 1 The overall flowchart of the present application is shown in the figure.

[0049] Figure 2 The construction process of the fault identification model is shown in the figure.

[0050] Figure 3 The flowchart of solving the target function of the present application is shown in the figure. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0052] It should be understood that when used in the present specification and the appended claims, the terms "comprise" and "include" indicate the presence of the described features, integers, steps, operations, elements, and / or components, but do not exclude one or more other features, integers, steps, operations, elements, components, and / or sets thereof.

[0053] It should also be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification and the appended claims of the present application, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0054] It should be further understood that the term "and / or" used in the specification and the appended claims of the present application means one or more of the associated listed items and all possible combinations thereof, and includes these combinations.

[0055] In one embodiment of the present application, a fault identification and processing method based on big data analysis is provided, as shown in Figure 1 The method comprises the following steps:

[0056] Step 1: Collecting power data and weather data at each collection point, the power data including voltage, current, power, device temperature, and device environment humidity, and the weather data including weather temperature, weather humidity, wind speed, light intensity, and season. In addition, the location information of each collection point, i.e., the latitude and longitude coordinates, is recorded.

[0057] Step 2: Using the collected historical power data and weather data to train a fault identification model for subsequent fault identification.

[0058] As shown in Figure 2 The training process of the fault identification model is as follows:

[0059] S1: Data collection: First, the power data and weather data at each collection point need to be collected. The power data includes voltage, current, power, device temperature, and device environment humidity, and the weather data includes weather temperature, weather humidity, wind speed, light intensity, and season. Data collection is the basis of fault prediction, and its correctness and rationality directly affect the model construction and accuracy.

[0060] S2: Data preprocessing: cleaning and preprocessing the collected data, including denoising, normalization processing, etc., to improve data quality. For example, MinMaxScaler can be used for normalization processing.

[0061] S3: Feature extraction: extracting features that are helpful for fault identification from the preprocessed data, including statistical features, frequency domain features, etc. In this embodiment, the DFS automatic extraction technology is used for feature construction and filter-based feature selection based on the original collected data.

[0062] S4: Model construction: A neural network model is constructed using a deep neural network (DNN). The deep neural network can form more abstract high-level representation attributes or features by combining low-level features, thereby discovering the intrinsic feature representation of the data.

[0063] S5: Model training: The neural network model is trained using the extracted features and labeled data (i.e., whether a failure occurred). During model training, the data set can be divided into a training set, a validation set, and a test set, where the data in the test set is removed from the data within a certain period of time before the failure occurs.

[0064] S6: Model evaluation and optimization: The performance of the model is evaluated, including accuracy and recall, and the model is optimized based on the evaluation results.

[0065] Through the above steps, a neural network-based fault identification model can be constructed to realize real-time monitoring and prediction of power system failures. In addition, existing trained fault identification models can be used directly for subsequent fault identification. That is, the model can be trained by itself, or the trained model can be downloaded from various third-party channels for subsequent use.

[0066] Step three: Obtain real-time power data and weather data collected at each collection point, use the fault identification model to identify faults, output the results of whether each collection point has a fault, record the geographic location of the collection point where the fault is identified, and send an alarm signal.

[0067] Step four: Sort the number of faults occurring at each collection point within a period, and set a sorting threshold. The geographic location of the collection point whose sorting position is before the sorting threshold is considered as a frequently occurring fault location.

[0068] For the frequently occurring fault location, a number of small workstations are established to quickly repair the faults at the frequently occurring fault location. The small workstations are used to store the tools required for fault repair and are equipped with corresponding repair personnel.

[0069] Step five: Establish a target function with station construction cost and distance difference as influencing factors, and minimize the target function as the goal. Solve the target function and output the station construction strategy.

[0070] The target function is as follows:

[0071]

[0072] In the formula, C 建站 is the unit cost of station construction, m is the number of stations, d 故障常发位置_i is the latitude and longitude coordinates of the i-th frequently occurring fault location, n is the number of collection points whose sorting position is before the sorting threshold, and the number of frequently occurring fault locations is d 站点_jF(d 故障常发位置_i , d 站点_j ) is the distance difference between the ith frequent failure location and the jth site location, and μ1, μ2 are weight coefficients.

[0073] F(d 故障常发位置_i , d 站点_j ) is calculated as follows:

[0074]

[0075] where R is the radius of the earth, lat i is the latitude of the ith frequent failure location, lat j is the latitude of the jth site location, lon i is the longitude of the ith frequent failure location, and lon j is the longitude of the jth site location.

[0076] In addition, if the latitude and longitude coordinates are in degrees, they need to be converted to radians first, and the conversion formula is as follows:

[0077]

[0078] The formula for calculating the station construction cost is as follows:

[0079] C 建站 = C 建筑 + C 工具 + C 人力

[0080] where C 建筑 is the construction cost of the station, i.e., the construction cost of the station infrastructure itself; C 工具 is the tool cost of the station, i.e., the cost of configuring tools in the station, and C 人力 is the cost of personnel configured in the station.

[0081] Set the constraint conditions, including the station number constraint:

[0082] 0 < m ≤ m max

[0083] where m max is the maximum number of stations;

[0084] Station cost constraint:

[0085]

[0086] where C max is the maximum total station cost.

[0087] Step six: Solve the objective function, and then obtain the station building strategy for the relevant processing personnel to refer to.

[0088] In this embodiment, the gradient descent algorithm is used to solve the objective function. The gradient descent algorithm is an iterative optimization algorithm used to find the local minimum of a function. It starts from an initial point and updates the parameters step by step in the opposite direction of the gradient (i.e. the steepest descent direction) of the objective function until the minimum of the function is found or the stopping condition is met. The algorithm is simple and intuitive, and is widely used in parameter optimization problems in machine learning, deep learning and other fields.

[0089] As shown in Figure 3 , the specific steps of the gradient descent algorithm are as follows:

[0090] S01: Select an initial point: Select an initial point x0, which can be random or selected according to prior knowledge of the problem.

[0091] S02: Set the learning rate: Set the learning rate α, which usually needs to be adjusted through experiments to find the best convergence speed and accuracy.

[0092] S03: Set the convergence threshold: Set the convergence threshold ∈, which determines when the algorithm stops.

[0093] S04: Calculate the gradient: Calculate the gradient of the objective function f(x) at the current point x t The gradient is a vector pointing in the direction of the fastest growth of the function.

[0094] S05: Update the parameters: Update the parameters using the gradient and the learning rate, as follows:

[0095]

[0096] This step is to move in the opposite direction of the gradient, because our goal is to find the minimum of the function.

[0097] S06: Check convergence: Calculate the magnitude of the parameter update ||x t+1 -x t ||. If this magnitude is less than the convergence threshold ∈, the algorithm is considered to have converged to the minimum, and the algorithm is stopped. If the algorithm has not converged, repeat steps S04 to S06 until the convergence condition is met.

[0098] ​To sum up, the application collects power data and meteorological data of each collection point, and records the position information of each collection point; the fault identification model is used for fault identification, and the result of whether each collection point has a fault is output, the geographic position of the collection point with the fault identified is recorded, and an alarm signal is sent; based on the number of faults of each collection point in a period, the collection points are sorted, and a sorting threshold is set, and the geographic position of the collection point before the sorting threshold is set as the frequently-occurring fault position; finally, a target function is established by taking the station construction cost and distance difference as influencing factors, and the station construction strategy is obtained by solving the target function.

[0099] Those skilled in the art can appreciate that the units of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both, and in order to clearly illustrate the interchangeability of hardware and software, the components of each example have been described in the above description in general terms. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0100] In the embodiments provided by the present application, it should be understood that the division of units is only a logical functional division, and actual implementation can have another division manner, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0101] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0102] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0103] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and 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 application, and they should be covered in the scope of the claims and the specification of the present application.

Claims

1. A fault identification and handling method based on big data analysis, characterized in that, Includes the following steps: Collect power and meteorological data from each collection point and record the location information of each collection point; The fault identification model is used to identify faults, output the results of whether there are faults at each collection point, record the geographical location of the collection point with the fault and issue an alarm signal. The data collection points are sorted based on the number of faults occurring at each collection point within the period, and a sorting threshold is set. The geographical locations of the collection points whose sorting positions are before the sorting threshold are considered as the locations where faults frequently occur. An objective function is established with website construction cost and distance difference as influencing factors. The goal is to minimize the objective function in order to solve for the website construction strategy. The objective function is solved to derive the website building strategy; The objective function is as follows: In the formula, For website building costs, m For the number of stations, For the first i Latitude and longitude coordinates of the locations where faults frequently occur. n This represents the number of data collection points whose sorting position is before the sorting threshold, i.e., the number of frequently occurring fault locations. For the first j Each site location For the i-th fault location and the i-th fault location j Distance difference between each station location These are the weighting coefficients; The formula for calculating website construction costs is as follows: In the formula, The construction cost of the station, that is, the construction cost of the station's infrastructure itself; The cost of website building tools, that is, the cost of configuring tools on the website. The cost of personnel configured at the site.

2. The fault identification and handling method based on big data analysis according to claim 1, characterized in that, The calculation formula is as follows: In the formula, R is the radius of the Earth. It is the latitude of the i-th frequently occurring fault location. It is the latitude of the j-th station location. It is the longitude of the i-th frequently occurring fault location. It is the longitude of the j-th station location.

3. The fault identification and processing method based on big data analysis according to claim 2, characterized in that, If latitude and longitude coordinates are in degrees, they need to be converted to radians first. The conversion formula is: 。 4. A fault identification and handling method based on big data analysis according to any one of claims 1-3, characterized in that, There are constraints, including a constraint on the number of websites: In the formula, This represents the maximum number of websites that can be created.

5. The fault identification and processing method based on big data analysis according to claim 4, characterized in that, The constraints also include website construction cost constraints: In the formula, This represents the maximum total cost of building a website.

6. The fault identification and processing method based on big data analysis according to claim 1, characterized in that, The training process of the fault identification model is as follows: S1: Data Acquisition: First, it is necessary to collect power data and meteorological data from each collection point. The power data includes voltage, current, power, equipment temperature, and equipment ambient humidity. The meteorological data includes weather temperature, weather humidity, wind speed, light intensity, and season. S2: Data preprocessing: Cleaning and preprocessing the collected data, including noise reduction and normalization; S3: Feature Extraction: Extract features that are helpful for fault identification from the preprocessed data, including statistical features and frequency domain features. The feature extraction adopts the automatic extraction technology DFS to construct features and select features based on filtering from the original collected data. S4: Model Building: A neural network model is built using a deep neural network (DNN). By combining low-level features, a more abstract high-level representation of attribute categories or features is formed, thereby discovering the inherent feature representation of the data. S5: Model Training: Use extracted features and labeled data to train a neural network model. During model training, the dataset is divided into a training set, a validation set, and a test set. S6: Model Evaluation and Optimization: Evaluate the model's performance, including accuracy and recall, and optimize the model based on the evaluation results.

7. The fault identification and processing method based on big data analysis according to claim 1, characterized in that, The objective function is solved using the gradient descent algorithm, and the specific steps are as follows: S01: Select Initial Point: Select an initial point x 0 initial point x 0 It is selected randomly or based on prior knowledge of the problem; S02: Set the learning rate: Set the learning rate α ; S03: Set convergence threshold: Set the convergence threshold to determine when the algorithm stops; S04: Calculate the gradient: at the current point x t Calculate the objective function at the location f ( x The gradient of ∇ f ( x t The gradient is a vector that points in the direction of the fastest growth of the function. S05: Update parameters: Update parameters using gradients and learning rate, specifically: ; S06: Check convergence: Calculate the magnitude of parameter updates. If this amplitude is less than the convergence threshold, the algorithm is considered to have converged to the minimum value and the algorithm is stopped. If the algorithm has not yet converged, repeat steps S04 to S06 until the convergence condition is met.

8. A processor, characterized in that, The processor is used to run a program, wherein the program executes the fault identification and processing method based on big data analysis as described in claim 1.

Images