Accident disposal decision-making system based on field risk factors

Abstract

The invention discloses an accident handling decision-making system based on field risk factors, and relates to the technical field of power grid fault handling, and the system comprises a field risk handling decision-making auxiliary platform which comprises an accident integration module, a field analysis module and a decision-making support module. According to the accident disposal decision-making system based on the field risk factors, a historical accident library is constructed according to power grid data collected when an accident occurs in history and power grid data within a certain period of time before the accident occurs; the method comprises the following steps: firstly, screening unstable data in power grid data of a power grid field accident district to screen a plurality of accident types from a historical accident library, then carrying out secondary matching on the plurality of accident types according to the change trend of the unstable data to accurately position the accident types, and combining with a reference scheme of a corresponding accident to accurately position the accident types. And reliable support is provided for the disposal decision of power grid field accidents.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid fault processing, and in particular to an accident handling decision-making system based on on-site risk factors. Background Art

[0002] At present, the system's power grid accident disposal plan generation method mainly relies on manual analysis and experience judgment. Although this method can cope with power grid accidents to a certain extent, it has many shortcomings. First, the manual analysis process is cumbersome and time-consuming, and it is difficult to quickly generate an effective disposal plan after a power grid accident occurs. Secondly, experience judgment is limited by personal knowledge and experience, and is prone to deviations and omissions, which affects the accuracy and practicality of the plan.

[0003] As described in application number 202410621760.8, a method, system, device and storage medium for generating a power grid accident handling plan realizes intelligent plan generation based on machine learning by conducting in-depth analysis and mining through artificial intelligence algorithms based on a large amount of power grid operation data, historical accident cases and expert experience.

[0004] Based on the retrieval of the above information, it can be seen that when making decisions on the existing power grid accident handling, the accident handling plan generated by relying on historical accident handling experience can effectively improve the efficiency of accident handling. However, when an accident occurs at the power grid site, when matching with the accident handling plan, there is a situation where the accident cannot be accurately located, which seriously restricts the handling efficiency of the power grid site accident. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an accident handling decision-making system based on on-site risk factors, which solves the problem that when an on-site accident occurs in the existing power grid, there is a situation where the accident cannot be accurately matched with the accident handling plan, affecting the efficiency of on-site accident handling in the power grid.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an accident handling decision-making system based on on-site risk factors, including an on-site risk handling decision-making auxiliary platform, the on-site risk handling decision-making auxiliary platform including: an accident integration module, an on-site analysis module and a decision support module, the accident integration module is respectively connected with the on-site analysis module and the decision support module, the on-site analysis module is connected with the decision support module; the accident integration module includes a historical data storage unit, a correlation analysis unit, an influencing factor introduction unit and a processing record unit, the historical data storage unit and the influencing factor introduction unit are both connected with the correlation analysis unit, and the correlation analysis unit is connected with the processing record unit; the on-site analysis module includes a data acquisition unit, a data analysis unit and a mutation factor acquisition unit, the data acquisition unit is connected with the data analysis unit, and the data analysis unit is connected with the mutation factor acquisition unit; the decision support module includes a data screening unit, a factor screening unit and an accident locating unit, the data screening unit is connected with the factor screening unit, and the factor screening unit is connected with the accident locating unit.

[0007] Furthermore, the historical data storage unit obtains the power grid data collected when historical accidents have occurred, as well as the power grid data within a certain period of time before the accident occurs, as accident-related power grid data; the influencing factor introduction unit determines whether the historical accident is triggered by human factors, and marks the relevant data; the correlation analysis unit automatically eliminates historical accidents that are triggered by human factors, and manually identifies the accident-related power grid data that have an impact on the remaining historical accidents, records them as valid correlation data, obtains the mapping relationship between accidents and valid correlation data, and constructs a historical accident library; the processing record unit records the processing plan for the corresponding accident as a reference plan for the corresponding accident.

[0008] Furthermore, the data acquisition unit obtains the power grid data when the power grid accident occurs on site and within a certain period of time before the accident occurs, as on-site accident related data; the data analysis unit obtains the power grid topology map of the accident area, obtains the current flow direction, and then sorts the power grid data within a certain period of time before the accident occurs according to the current flow direction, takes the power grid data collection point located at the upstream of the current direction as the analysis point, obtains the change rate of the remaining power grid data under the same value collected by the analysis point, and uses it as the accident-related power grid analysis data; the mutation factor collection unit filters out the unstable power grid data from the accident-related power grid analysis data as fault data.

[0009] Furthermore, the data screening unit uses the fault data as an index to search for several accident types containing fault data from the historical accident database; the factor screening unit obtains the change trend of several power grid data corresponding to the fault data in several accident types within a certain period of time as a reference trend, and after performing a similarity comparison with the change trend of the fault data, obtains the accident type with the highest similarity as the suspected fault type; the accident locating unit is used to obtain a reference plan for the suspected fault type to provide support for decision-making and processing after the occurrence of a power grid on-site accident.

[0010] Furthermore, the method of acquiring the change rate of the remaining power grid data under the same value collected by the analysis point includes:

[0011] The grid data with the same value collected from the remaining grid data at the screening analysis point and without value change are removed, and the changed grid data are obtained and recorded as risk data;

[0012] Get the time interval between two adjacent analysis points for collecting the same data;

[0013] Calculate the change rate of risk data corresponding to the same power grid data collection point:

[0014]

[0015] In the formula, R n The risk data collected by a risk data collection point at time t n-1 to n The rate of change of time period, V n-1 For a risk data collection point at time t n-1 The collected power grid data, V n For a risk data collection point at time t n The collected power grid data;

[0016] Obtain the change rate set of risk data corresponding to the same power grid data collection point, and analyze the change trend of adjacent change rates:

[0017]

[0018] In the formula, R n-1 The risk data collected by a risk data collection point at time t n-2 to n-1 The rate of change in a time period, and the type of change trend that is increasing or decreasing is recorded as unstable.

[0019] Furthermore, the factor screening unit obtains reference trends and performs similarity comparison to determine the suspected faults in the following specific steps:

[0020] Step 1: Data preparation: extract some power grid data corresponding to the fault data from the historical accident database, covering a certain period before and after the accident;

[0021] Step 2: Data preprocessing, including: denoising, normalization and time alignment, wherein the time alignment is to ensure the alignment of the reference data and the fault data in the time dimension;

[0022] Step 3: Extract time series features, including: value change trend, change rate and fluctuation characteristics. The value change trend includes the increase and decrease trend of current and voltage. The calculation formula of the change rate is:

[0023]

[0024] In the formula, R n The risk data collected by a risk data collection point at time t n-1 to n The rate of change of time period, V n-1 For a risk data collection point at time t n-1 The collected power grid data, V n For a risk data collection point at time t n The collected power grid data;

[0025] The fluctuation characteristics are the mean and standard deviation of data extracted through statistical features;

[0026] Step 4: Reference trend generation: Generate a reference trend corresponding to the fault data based on the historical accident type data;

[0027] Step 5: Similarity calculation: Calculate the Euclidean distance between the fault data and the reference trend:

[0028]

[0029] Where D represents the Euclidean distance between the fault data and the reference trend. The smaller the value, the higher the similarity between the two. n represents the number of data points, that is, the length of the time series. i represents the value of the fault data at the i-th time point, y i Represents the value of the reference trend at the i-th time point;

[0030] Step 6, comparing fault data with reference trends: calculating the similarity between the changing trend of the fault data and all reference trends one by one and recording the similarity score of each group of comparison results;

[0031] Step 7: Similarity sorting: sort the historical accident types corresponding to the reference trend according to the similarity scores;

[0032] Step 8: Determine the accident type: select the accident type with the highest similarity score and mark it as the suspected fault type;

[0033] Step 9. Output results: Return the suspected fault type and its corresponding reference solution to provide support for accident site handling and record the calculation process and results for subsequent analysis and optimization.

[0034] Furthermore, the denoising process is to use a moving average method to smooth the noise, and the expression of the moving average method is:

[0035]

[0036] In the formula, represents the i-th data point after smoothing, x j represents the jth data point in the original data, k represents half of the window size, the window length is 2k+1, and i represents the index of the current data point.

[0037] Furthermore, the normalization process includes: minimum-maximum normalization and standardization, and the expression of the minimum-maximum normalization is:

[0038]

[0039] In the formula, x′ i represents the normalized ith data point, x i represents the original i-th data point, x min Represents the minimum value in the data, x max Indicates the maximum value in the data;

[0040] The standardized expression is:

[0041]

[0042] In the formula, x′ i represents the normalized ith data point, x i represents the original i-th data point, μ represents the mean of the data, and σ represents the standard deviation of the data.

[0043] The present invention provides an accident handling decision-making system based on on-site risk factors. It has the following beneficial effects:

[0044] The present invention constructs a historical accident database by using power grid data collected when accidents have occurred in the past and power grid data within a certain period of time before the accidents have occurred. When a power grid on-site accident occurs, the present invention first screens the unstable data in the power grid data of the power grid on-site accident area to screen several accident types from the historical accident database, and then performs secondary matching on several accident types through the change trend of the unstable data to accurately locate the accident type, and combines the reference plan for the corresponding accident to provide reliable support for the handling decision of the power grid on-site accident. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a structural schematic diagram of an accident handling decision-making system based on on-site risk factors of the present invention;

[0046] Figure 2 It is a structural schematic diagram of an accident integration module of an accident handling decision system based on on-site risk factors of the present invention;

[0047] Figure 3 It is a structural schematic diagram of a field analysis module of a field risk factor-based accident handling decision-making system of the present invention;

[0048] Figure 4 It is a structural schematic diagram of a decision support module of an accident handling decision system based on on-site risk factors of the present invention.

[0049] In the figure:

[0050] 1. On-site risk management decision-making support platform;

[0051] 2. Accident integration module; 201. Historical data storage unit; 202. Correlation analysis unit; 203. Influencing factor introduction unit; 204. Processing record unit;

[0052] 3. On-site analysis module; 301. Data collection unit; 302. Data analysis unit; 303. Mutation factor collection unit;

[0053] 4. Decision support module; 401. Data screening unit; 402. Factor screening unit; 403. Accident location unit. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0055] See also Figure 1-4 The embodiment of the present invention provides the following technical solutions: an accident handling decision system based on on-site risk factors, as shown in the attached Figure 1As shown, the on-site risk processing decision-making auxiliary platform 1 includes an accident integration module 2, an on-site analysis module 3 and a decision support module 4.

[0056] As a preferred solution, in order to generate a power grid accident handling plan, the accident integration module 2 includes a historical data storage unit 201, a correlation analysis unit 202, an influencing factor introduction unit 203 and a processing record unit 204. The historical data storage unit 201 obtains the power grid data collected when an accident has occurred in the past, as well as the power grid data within a certain period of time before the accident occurs, as accident-related power grid data. The influencing factor introduction unit 203 determines whether the historical accident is triggered by human factors and marks the relevant data.

[0057] The historical data storage unit 201 and the influencing factor introduction unit 203 are both connected to the correlation analysis unit 202. The correlation analysis unit 202 is used to automatically eliminate historical accidents that have been manually triggered, and manually identify accident-related power grid data that have an impact on the remaining historical accidents, record them as valid correlation data, obtain the mapping relationship between accidents and valid correlation data, and build a historical accident database.

[0058] The correlation analysis unit 202 is connected to the processing recording unit 204, and the processing recording unit 204 records the processing plan of the corresponding accident as a reference plan of the corresponding accident.

[0059] As a preferred solution, in order to achieve effective organization of data related to power grid on-site accidents, to better match accident handling plans, and to achieve accurate positioning of accident types, the accident integration module 2 is connected to the on-site analysis module 3. The on-site analysis module 3 includes a data acquisition unit 301, a data analysis unit 302, and a mutation factor acquisition unit 303. The data acquisition unit 301 obtains power grid data when a power grid on-site accident occurs and within a certain period of time before the accident occurs, as on-site accident related data.

[0060] The data acquisition unit 301 is connected to the data analysis unit 302. The data analysis unit 302 obtains the power grid topology map of the accident area, and after obtaining the current flow direction, sorts the power grid data within a certain period before the accident according to the current flow direction, takes the power grid data collection point located at the upstream of the current direction as the analysis point, and obtains the change rate of the remaining power grid data under the same value collected by the analysis point as the accident-related power grid analysis data, wherein the method of obtaining the change rate of the remaining power grid data under the same value collected by the analysis point includes:

[0061] The grid data with the same value collected from the remaining grid data at the screening analysis point and without value change are removed, and the changed grid data are obtained and recorded as risk data;

[0062] Get the time interval between two adjacent analysis points for collecting the same data;

[0063] Calculate the change rate of risk data corresponding to the same power grid data collection point:

[0064]

[0065] In the formula, R n The risk data collected by a risk data collection point at time t n-1 to n The rate of change of time period, V n-1 For a risk data collection point at time t n-1 The collected power grid data, V n For a risk data collection point at time t n The collected power grid data;

[0066] Obtain the change rate set of risk data corresponding to the same power grid data collection point, and analyze the change trend of adjacent change rates:

[0067]

[0068] In the formula, R n-1 The risk data collected by a risk data collection point at time t n-2 to n-1 The rate of change in a time period, and the type of change trend that is increasing or decreasing is recorded as unstable.

[0069] The data analysis unit 302 is connected to the mutation factor collection unit 303, and the mutation factor collection unit 303 selects unstable power grid data from the accident-related power grid analysis data as fault data.

[0070] As a preferred solution, in order to achieve effective matching of power grid on-site accidents and accident handling plans, the accident integration module 2 is connected to the decision support module 4, and the on-site analysis module 3 is connected to the decision support module 4. The decision support module 4 includes a data screening unit 401, a factor screening unit 402 and an accident location unit 403. The data screening unit 401 uses fault data as an index to search for several accident types containing fault data from the historical accident library.

[0071] The data screening unit 401 is connected to the factor screening unit 402. The factor screening unit 402 obtains the change trend of several power grid data corresponding to the fault data in several accident types within a certain period of time as a reference trend. After comparing the similarity with the change trend of the fault data, the accident type with the highest similarity is obtained as the suspected fault type.

[0072] The factor screening unit 402 is connected to the accident location unit 403. The accident location unit 403 is used to obtain a reference plan for the suspected fault type to provide support for decision-making after the power grid on-site accident occurs.

[0073] As a more preferred solution, the factor screening unit obtains reference trends and performs similarity comparison to determine the suspected fault in the following specific steps:

[0074] Step 1: Data preparation: extract some power grid data corresponding to the fault data from the historical accident database, covering a certain period before and after the accident;

[0075] Step 2: Data preprocessing, including: denoising, normalization and time alignment, wherein the time alignment is to ensure the alignment of the reference data and the fault data in the time dimension;

[0076] Step 3: Extract time series features, including: value change trend, change rate and fluctuation characteristics. The value change trend includes the increase and decrease trend of current and voltage. The calculation formula of the change rate is:

[0077]

[0078] In the formula, R n The risk data collected by a risk data collection point at time t n-1 to n The rate of change of time period, V n-1 For a risk data collection point at time t n-1 The collected power grid data, V n For a risk data collection point at time t n The collected power grid data;

[0079] The fluctuation characteristics are the mean and standard deviation of data extracted through statistical features;

[0080] Step 4: Reference trend generation: Generate a reference trend corresponding to the fault data based on the historical accident type data;

[0081] Step 5: Similarity calculation: Calculate the Euclidean distance between the fault data and the reference trend:

[0082]

[0083] Where D represents the Euclidean distance between the fault data and the reference trend. The smaller the value, the higher the similarity between the two. n represents the number of data points, that is, the length of the time series. i represents the value of the fault data at the i-th time point, y i Represents the value of the reference trend at the i-th time point;

[0084] Step 6, comparing fault data with reference trends: calculating the similarity between the changing trend of the fault data and all reference trends one by one and recording the similarity score of each group of comparison results;

[0085] Step 7: Similarity sorting: sort the historical accident types corresponding to the reference trend according to the similarity scores;

[0086] Step 8: Determine the accident type: select the accident type with the highest similarity score and mark it as the suspected fault type;

[0087] Step 9. Output results: Return the suspected fault type and its corresponding reference solution to provide support for accident site handling and record the calculation process and results for subsequent analysis and optimization.

[0088] More preferably, the denoising process is to use a moving average method to smooth the noise, and the expression of the moving average method is:

[0089]

[0090] In the formula, represents the i-th data point after smoothing, x j represents the jth data point in the original data, k represents half of the window size, the window length is 2k+1, and i represents the index of the current data point.

[0091] More preferably, the normalization process includes: minimum-maximum normalization and standardization, and the expression of the minimum-maximum normalization is:

[0092]

[0093] In the formula, x′ i represents the normalized ith data point, x i represents the original i-th data point, x min Represents the minimum value in the data, x max Indicates the maximum value in the data;

[0094] The standardized expression is:

[0095]

[0096] In the formula, x′ i represents the normalized ith data point, x i represents the original i-th data point, μ represents the mean of the data, and σ represents the standard deviation of the data.

[0097] The actual method of using the above system includes the following steps:

[0098] S1. Generation of accident handling plan: The power grid data collected when the historical accident has occurred and the power grid data within a certain period before the accident are obtained through the historical data storage unit 201 as the accident-related power grid data. The influencing factor introduction unit 203 determines whether the historical accident is triggered manually. After marking the historical accident triggered manually, the historical accident triggered manually is eliminated by the association analysis unit 202, and the accident-related power grid data that has an impact on the remaining historical accidents is manually identified and recorded as valid related data. The mapping relationship between the accident and the valid related data is obtained, and a historical accident database is constructed. At the same time, the processing plan of the corresponding accident is recorded by the processing record unit 204 as a reference plan for the corresponding accident.

[0099] S2. Data processing of power grid on-site accidents: The data acquisition unit 301 acquires the power grid data when the power grid on-site accident occurs and within a certain period of time before the accident occurs as on-site accident related data, and acquires the power grid topology map of the accident area. After acquiring the current flow direction, the data analysis unit 302 sorts the power grid data within a certain period of time before the accident occurs according to the current flow direction, and takes the power grid data acquisition point located at the upstream of the current direction as the analysis point, and acquires the change rate of the remaining power grid data under the same value collected by the analysis point. The change rate calculation method includes:

[0100] The grid data with the same value collected from the remaining grid data at the screening analysis point and without value change are removed, and the changed grid data are obtained and recorded as risk data;

[0101] Get the time interval between two adjacent analysis points for collecting the same data;

[0102] Calculate the change rate of risk data corresponding to the same power grid data collection point:

[0103]

[0104] In the formula, R n The risk data collected by a risk data collection point at time t n-1 to n The rate of change of time period, V n-1 For a risk data collection point at time t n-1 The collected power grid data, V n For a risk data collection point at time t n The collected power grid data;

[0105] Obtain the change rate set of risk data corresponding to the same power grid data collection point, and analyze the change trend of adjacent change rates:

[0106]

[0107] In the formula, R n-1 The risk data collected by a risk data collection point at time t n-2 to n-1 The rate of change in a time period, and the type of change trend that is increasing or decreasing is recorded as unstable;

[0108] Then, the mutation factor collection unit 303 is used to filter out unstable power grid data from the accident-related power grid analysis data as fault data;

[0109] S3. Matching of fault data and accident handling plans: With fault data as index, the data screening unit 401 searches for several accident types containing fault data from the historical accident database, and the factor screening unit 402 obtains the change trend of several power grid data corresponding to the fault data in several accident types within a certain period of time as a reference trend. After similarity comparison with the change trend of the fault data, the accident type with the highest similarity is obtained as the suspected fault type. The accident location unit 403 is used to obtain a reference plan for the suspected fault type, so as to provide support for decision-making after the occurrence of the power grid on-site accident.

[0110] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0111] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An accident handling decision system based on on-site risk factors, comprising an on-site risk handling decision support platform (1), characterized in that: The on-site risk processing decision support platform (1) comprises: an accident integration module (2), an on-site analysis module (3) and a decision support module (4); the accident integration module (2) is connected to the on-site analysis module (3) and the decision support module (4) respectively, and the on-site analysis module (3) is connected to the decision support module (4); the accident integration module (2) comprises a historical data storage unit (201), a correlation analysis unit (202), an influencing factor introduction unit (203) and a processing record unit (204); the historical data storage unit (201) and the influencing factor introduction unit (203) are both connected to the correlation analysis unit (202), and the correlation analysis unit (204) is connected to the on-site analysis module (3) and the decision support module (4) respectively; 2) is connected to the processing and recording unit (204); the on-site analysis module (3) includes a data acquisition unit (301), a data analysis unit (302) and a mutation factor acquisition unit (303), the data acquisition unit (301) is connected to the data analysis unit (302), and the data analysis unit (302) is connected to the mutation factor acquisition unit (303); the decision support module (4) includes a data screening unit (401), a factor screening unit (402) and an accident location unit (403), the data screening unit (401) is connected to the factor screening unit (402), and the factor screening unit (402) is connected to the accident location unit (403).

2. The accident handling decision-making system based on on-site risk factors according to claim 1 is characterized by: The historical data storage unit (201) obtains the power grid data collected when a historical accident has occurred, as well as the power grid data within a certain period before the accident occurs, as accident-related power grid data; the influencing factor introduction unit (203) determines whether the historical accident is triggered by human factors, and marks the relevant data; the association analysis unit (202) automatically removes the historical accidents that have been triggered by human factors, and manually identifies the accident-related power grid data that has an impact on the remaining historical accidents, records them as valid associated data, obtains the mapping relationship between the accident and the valid associated data, and constructs a historical accident database; the processing record unit (204) records the processing plan for the corresponding accident as a reference plan for the corresponding accident.

3. The accident handling decision-making system based on on-site risk factors according to claim 1 is characterized by: The data acquisition unit (301) acquires power grid data when a power grid on-site accident occurs and within a certain period of time before the accident occurs, as on-site accident-related data; the data analysis unit (302) acquires a power grid topology diagram of the accident-occurring area, acquires the current flow direction, and then sorts the power grid data within a certain period of time before the accident occurs according to the current flow direction, takes the power grid data acquisition point located at the upstream of the current direction as the analysis point, and acquires the change rate of the remaining power grid data under the same value acquired at the analysis point, as accident-related power grid analysis data; the mutation factor acquisition unit (303) selects unstable power grid data from the accident-related power grid analysis data as fault data.

4. The accident handling decision-making system based on on-site risk factors according to claim 1 is characterized by: The data screening unit (401) uses the fault data as an index to search for several accident types containing the fault data from the historical accident database; the factor screening unit (402) obtains the change trend of several power grid data corresponding to the fault data in a certain period of time from the several accident types as a reference trend, and after performing a similarity comparison with the change trend of the fault data, obtains the accident type with the highest similarity as the suspected fault type; The accident location unit (403) obtains a reference plan for the suspected fault type to provide support for decision-making after a power grid on-site accident occurs.

5. The accident handling decision-making system based on on-site risk factors according to claim 3 is characterized by: The method of obtaining the change rate of the remaining power grid data under the same value collected by the analysis point includes: The grid data with the same value collected from the remaining grid data at the screening analysis point and without value change are removed, and the changed grid data are obtained and recorded as risk data; Get the time interval between two adjacent analysis points for collecting the same data; Calculate the change rate of risk data corresponding to the same power grid data collection point: In the formula, R n The risk data collected by a risk data collection point at time t n-1 to n The rate of change of time period, V n-1 For a risk data collection point at time t n-1 The collected power grid data, V n For a risk data collection point at time t n The collected power grid data; Obtain the change rate set of risk data corresponding to the same power grid data collection point, and analyze the change trend of adjacent change rates: In the formula, R n-1 The risk data collected by a risk data collection point at time t n-2 to n-1 The rate of change in a time period, and the type of change trend that is increasing or decreasing is recorded as unstable.

6. The accident handling decision-making system based on on-site risk factors according to claim 4 is characterized by: The factor screening unit (402) obtains the reference trend and performs similarity comparison to determine the suspected fault in the following specific steps: Step 1: Data preparation: extract some power grid data corresponding to the fault data from the historical accident database, covering a certain period before and after the accident; Step 2: Data preprocessing, including: denoising, normalization and time alignment, wherein the time alignment is to ensure the alignment of the reference data and the fault data in the time dimension; Step 3: Extract time series features, including: value change trend, change rate and fluctuation characteristics. The value change trend includes the increase and decrease trend of current and voltage. The calculation formula of the change rate is: In the formula, R n The risk data collected by a risk data collection point at time t n-1 to n The rate of change of time period, V n-1 For a risk data collection point at time t n-1 The collected power grid data, V n For a risk data collection point at time t n The collected power grid data; The fluctuation characteristics are the mean and standard deviation of data extracted through statistical features; Step 4: Reference trend generation: Generate a reference trend corresponding to the fault data based on the historical accident type data; Step 5: Similarity calculation: Calculate the Euclidean distance between the fault data and the reference trend: Where D represents the Euclidean distance between the fault data and the reference trend. The smaller the value, the higher the similarity between the two. n represents the number of data points, that is, the length of the time series. i represents the value of the fault data at the i-th time point, y i Represents the value of the reference trend at the i-th time point; Step 6, comparing fault data with reference trends: calculating the similarity between the changing trend of the fault data and all reference trends one by one and recording the similarity score of each group of comparison results; Step 7: Similarity sorting: sort the historical accident types corresponding to the reference trend according to the similarity scores; Step 8: Determine the accident type: select the accident type with the highest similarity score and mark it as the suspected fault type; Step 9. Output results: Return the suspected fault type and its corresponding reference solution to provide support for accident site handling and record the calculation process and results for subsequent analysis and optimization.

7. The accident handling decision-making system based on on-site risk factors according to claim 6 is characterized by: The denoising process is to use a moving average method to smooth the noise, and the expression of the moving average method is: In the formula, represents the i-th data point after smoothing, x j represents the jth data point in the original data, k represents half of the window size, the window length is 2k+1, and i represents the index of the current data point.

8. The accident handling decision-making system based on on-site risk factors according to claim 6 is characterized by: The normalization process includes: minimum-maximum normalization and standardization. The expression of the minimum-maximum normalization is: In the formula, x′ i represents the normalized ith data point, x i represents the original i-th data point, x min Represents the minimum value in the data, x max Indicates the maximum value in the data; The standardized expression is: In the formula, x′ i represents the normalized ith data point, x i represents the original i-th data point, μ represents the mean of the data, and σ represents the standard deviation of the data.

Images