A power consumption safety hidden danger grading early warning method, device and medium for different customers

Abstract

This invention relates to the field of electricity safety monitoring technology, and discloses a method, device, and medium for graded early warning of electricity safety hazards for different customers. The method includes acquiring user data and electricity hazard information from a multi-user environment, calculating problem priority ranking, and obtaining a preliminary scheduling plan; adjusting resource allocation according to the problem priority ranking to obtain an optimized resource scheduling scheme; allocating priority processing resources to key problems to obtain a processing sequence for key problems, acquiring data on unprocessed problems, determining problem escalation, and updating the scheduling plan; acquiring real-time feedback data from the multi-user environment, analyzing the timeliness of problem processing, obtaining a dynamic management scheme adapted to the multi-user environment, continuously monitoring changes in user importance and hazard severity, and updating and optimizing the scheduling plan accordingly. This invention can effectively improve resource utilization, optimize problem processing timeliness, and realize intelligent dynamic management of power system hazards.

Description

Technical Field

[0001] This invention relates to the field of electrical safety monitoring technology, and in particular to a method, equipment and medium for graded early warning of electrical safety hazards for different customers. Background Technology

[0002] In modern society, power safety management is a crucial field, directly impacting people's quality of life and the stable operation of the social economy. Especially in complex environments involving multiple users, ensuring power safety is not only a technical issue but also a vital link in safeguarding public safety. With the continuous growth of electricity demand and the diversification of user groups, effectively identifying and promptly addressing safety hazards has become a critical task that cannot be ignored in power management. However, current safety management methods often suffer from uneven resource allocation and low response efficiency when dealing with multi-user scenarios. Many solutions lack a comprehensive consideration of the importance of users and the urgency of problems when addressing safety issues for multiple users, leading to the neglect of the needs of important users or the failure to address serious issues in a timely manner. Furthermore, existing methods often cannot dynamically adjust strategies when dealing with multiple issues, easily resulting in wasted management resources or delayed responses.

[0003] Against this backdrop, the core challenges gradually emerge. First, in a multi-user environment, how to rationally prioritize and handle issues based on user importance and the severity of potential risks. This factor directly impacts whether resources can be used efficiently and whether significant losses can be avoided. Second, when multiple issues occur simultaneously, how to achieve intelligent scheduling under limited resources to ensure that critical issues are not delayed becomes a further challenge. These two aspects are closely related; the former determines the identification and classification of problems, while the latter directly affects the efficiency and effectiveness of actual processing.

[0004] Therefore, how to achieve intelligent priority management and resource scheduling in a multi-user environment, taking into account both user importance and the severity of potential risks, while ensuring that problems can be automatically escalated if they are not handled in a timely manner, has become a key issue that this research urgently needs to address. Summary of the Invention

[0005] In view of the aforementioned existing problems, this invention is proposed. Therefore, this invention provides a graded early warning method for electrical safety hazards tailored to different customers to solve the above problems.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a method for graded early warning of potential electrical safety hazards for different customers, including:

[0008] User data and electricity hazard information are obtained from a multi-user environment. By analyzing the importance of users and the severity of hazards, the problem priority is calculated and a preliminary scheduling plan is obtained.

[0009] Based on the scheduling plan, real-time resource allocation data is obtained. If resource allocation is insufficient, resource allocation is adjusted according to the problem priority ranking to obtain an optimized resource scheduling scheme.

[0010] Based on the optimized resource scheduling scheme, priority processing resources are allocated to key issues to determine whether the response time optimization requirements are met, the processing sequence of key issues is obtained, and data on issues that have not been processed in a timely manner is acquired. Issue escalation judgment is then performed to obtain escalated issue data, so as to dynamically adjust the scheduling mechanism and update the scheduling plan.

[0011] Based on the updated scheduling plan, real-time feedback data in a multi-user environment is obtained, the timeliness of problem handling is analyzed, a dynamic management solution adapted to the multi-user environment is obtained, and changes in user importance and the severity of potential risks are continuously monitored to update and optimize the scheduling plan.

[0012] As a preferred embodiment of the graded early warning method for electricity safety hazards for different customers described in this invention, the priority ranking of calculation problems includes:

[0013] User data and electricity hazard information are obtained from a multi-user environment to construct an initial dataset. User data is then classified into user groups using preset classification rules.

[0014] The severity of each user in the user group is analyzed, and combined with the information on potential electrical hazards, the hazard assessment result is determined. If the hazard assessment result exceeds a preset threshold, a severity score is obtained.

[0015] Based on the severity score and the user's level of importance, the priority of the issues is calculated, a solution for each issue is generated, and potential risk points are identified based on the distribution characteristics of high-priority issues.

[0016] As a preferred embodiment of the graded early warning method for electricity safety hazards for different customers described in this invention, the preliminary scheduling plan includes:

[0017] By extracting relevant information from user data sources and combining it with data on potential electricity hazards, a basic analysis dataset is constructed to obtain a preliminary set of problem categories. A weighted calculation method is then used to quantify the priority of the problems and determine the order of processing.

[0018] Based on the processing order number and the urgency level data, if the urgency level exceeds the preset threshold, the resource allocation ratio is adjusted first to obtain the adjusted scheduling framework. Combined with the time window, the processing sequence of each problem is optimized to obtain the optimized timing arrangement.

[0019] Based on the optimized timing arrangement and combined with the regional distribution map, the spatial clustering characteristics of potential hazards are analyzed to determine the range of high-risk areas. Through predictive processing, the distribution characteristics of potential hazards are obtained. Combined with data in the historical record database, the final scheduling execution plan is determined.

[0020] As a preferred embodiment of the graded early warning method for electricity safety hazards for different customers described in this invention, the optimized resource scheduling scheme includes:

[0021] Based on the data and real-time information of the scheduling plan, a dynamic view of resource allocation is constructed, and the comparison data between the total amount of resources and the current demand is obtained. If the total amount of resources is lower than the current demand, the range of resource gap is determined, and the resource allocation ratio of each task is sorted and adjusted in combination with the priority of the problem to obtain a preliminary resource reallocation plan.

[0022] Based on the resource reallocation scheme, the matching degree of demand coverage is analyzed. If the matching degree does not reach the preset standard, the resources of low-priority tasks are compressed to obtain the adjusted execution sequence. The stability of resource scheduling is evaluated, and the resource allocation details of each task are generated to obtain the optimized resource scheduling scheme.

[0023] As a preferred embodiment of the graded early warning method for electricity safety hazards for different customers described in this invention, the updated scheduling plan includes:

[0024] By optimizing the scheme, the key issues are initially divided into resource allocations, matching data of resource scheduling and processing order is obtained, and the key issues are classified. If the response time of the classified issues exceeds the optimization standard, the resource scheduling of the excess part is adjusted, and the adjusted allocation strategy is determined.

[0025] Based on the adjusted allocation strategy, the correspondence between the processing order of key issues and resource scheduling is obtained, it is determined whether there is an uneven distribution of resources, and further resource allocation optimization is carried out according to the time limit. The stability of the processing order is evaluated to obtain the processing sequence of key issues.

[0026] A dataset of unprocessed issues is generated from a sequence of key issues. The issue data is automatically escalated using a preset threshold. The degree to which the escalation conditions are met is evaluated to obtain a preliminary set of escalated issues. Issue features are then acquired for classification to determine a subset of issues after classification.

[0027] Based on the aforementioned subset of problems, determine whether the escalation conditions are met. If they are met, generate an escalation problem list; otherwise, mark it as a set of problems to be processed.

[0028] Extract problem priorities from the upgrade problem list, adjust the priority sorting according to preset thresholds, compare with real-time data to determine whether there are abnormal problems in the sequence, adjust the resource allocation ratio, and obtain a dynamically adjusted resource scheduling scheme.

[0029] The advantage of this preferred solution is that it can prioritize the allocation of processing resources for key issues and promptly identify and address issues that have not been addressed in a timely manner.

[0030] As a preferred embodiment of the graded early warning method for electricity safety hazards for different customers described in this invention, the method includes: acquiring real-time feedback data in a multi-user environment, determining whether resource utilization has reached a preset target, and obtaining the final resource allocation result, including:

[0031] Relevant data of the multi-user environment is obtained from the scheduling plan. The real-time feedback content is classified and stored to obtain the classified data set. Relevant information on resource utilization is obtained. If the resource utilization index is lower than the preset threshold, the abnormal data is marked and the abnormal data subset is determined.

[0032] Based on a subset of abnormal data, relevant information from environmental monitoring is obtained, and the resource allocation status in a multi-user environment is analyzed to determine whether there is uneven allocation and obtain the allocation status analysis results.

[0033] Based on the allocation status analysis results, a second check is performed using relevant data to determine the utilization rate. If the check results show that the resource utilization deviates from the preset threshold, the scheduling plan is partially adjusted to determine a temporary adjustment plan.

[0034] Based on the temporary adjustment plan, key information on the plan adjustment is obtained, and the resource allocation data before and after the adjustment is checked to obtain the difference detection results. The final state of resource allocation is verified. If the verification results show that the allocation plan deviates from the preset threshold, the feedback data is reorganized to determine the final allocation plan.

[0035] Based on the final allocation scheme, the resource utilization status in the multi-user environment is continuously tracked to obtain status records.

[0036] As a preferred embodiment of the graded early warning method for electricity safety hazards for different customers described in this invention, the method involves: based on the final resource allocation results, analyzing the timeliness of problem handling, and obtaining a dynamic management scheme adapted to multi-user environments, including:

[0037] Acquire real-time feedback data to monitor the resource allocation status in a multi-user environment, obtain a categorized dataset, extract key indicators of resource utilization efficiency, analyze indicator fluctuations, and determine efficiency fluctuation characteristics.

[0038] If the efficiency fluctuation characteristics indicate uneven resource allocation, abnormal data points are marked to obtain an abnormal data subset, user behavior correlation information is obtained, and resource demand changes are predicted through a time series analysis model to obtain demand prediction results.

[0039] Based on the demand forecast results, the scheduling plan parameters are adjusted to generate a dynamic adjustment plan. The resource allocation data before and after the adjustment are then checked for differences to obtain the difference detection results. If the difference detection results show that the resource allocation deviates from the preset threshold, the real-time feedback data is reorganized to determine the final dynamic management plan.

[0040] The beneficial effects of this preferred solution are that the present invention can continuously optimize resource allocation and use intelligent management capabilities to analyze the timeliness of problem handling.

[0041] As a preferred embodiment of the graded early warning method for electricity safety hazards for different customers described in this invention, the continuous monitoring of changes in the importance of users and the severity of hazards includes:

[0042] Real-time acquisition of data on user importance and risk severity, extraction of key feature values, determination of preliminary severity assessment results, classification of user importance and risk severity, and generation of a priority ranking list after classification.

[0043] Preliminary scheduling is performed based on the hierarchical priority list to generate an initial resource scheduling plan. The fluctuations in user importance and risk severity are continuously tracked to determine if there are any deviations. If the deviation exceeds a preset threshold, a dynamic plan adjustment mechanism is triggered to obtain the adjusted scheduling parameters.

[0044] The resource scheduling scheme is iteratively updated multiple times based on the adjusted scheduling parameters to determine the optimized dynamic scheme content and make real-time corrections. If new deviations are found, the loop execution process is restarted to obtain the final scheme update result.

[0045] In a second aspect, the present invention provides a computer device, comprising:

[0046] Memory and processor;

[0047] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method for graded early warning of electricity safety hazards for different customers are implemented.

[0048] Thirdly, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for graded early warning of electrical safety hazards for different customers.

[0049] Compared with existing technologies, the beneficial effects of this invention are as follows: By analyzing the importance of users and the severity of potential power hazards, this invention calculates problem priorities and formulates preliminary scheduling plans. Combined with real-time resource allocation data, this invention optimizes resource scheduling schemes, prioritizes the allocation of processing resources for critical issues, and uses preset thresholds to determine whether unresolved issues need to be automatically escalated, re-prioritizing the escalated issues. By acquiring real-time feedback data, this invention continuously optimizes resource allocation and uses intelligent management capabilities to analyze the timeliness of problem handling. This invention can also continuously monitor changes in user importance and the severity of potential hazards, cyclically execute priority ranking and resource scheduling, and achieve dynamic management in a multi-user environment. This method can effectively improve resource utilization, optimize problem handling timeliness, and realize intelligent dynamic management of potential power system hazards. Attached Figure Description

[0050] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a schematic diagram of the overall process of a graded early warning method for electricity safety hazards for different customers, as described in one embodiment of the present invention. Detailed Implementation

[0052] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0053] Reference Figure 1 As an embodiment of the present invention, a method for graded early warning of electricity safety hazards for different customers is provided, including:

[0054] S101: Obtain user data and electricity hazard information from a multi-user environment; calculate the priority of problems by analyzing the importance of users and the severity of hazards, and obtain a preliminary scheduling plan.

[0055] S102: Based on the scheduling plan, obtain real-time resource allocation data. If the resource allocation is insufficient, adjust the resource allocation according to the priority of the problem to obtain an optimized resource scheduling scheme.

[0056] S103, based on the optimized resource scheduling scheme, prioritize resources for critical issues, determine whether the response time optimization requirements are met, obtain the processing sequence of critical issues, acquire data on issues that have not been processed in a timely manner, and perform issue escalation judgment to obtain escalation issue data, so as to dynamically adjust the scheduling mechanism and update the scheduling plan;

[0057] S104, based on the updated scheduling plan, obtains real-time feedback data in a multi-user environment, analyzes the timeliness of problem handling, obtains a dynamic management solution adapted to the multi-user environment, continuously monitors changes in user importance and the severity of potential risks, and updates and optimizes the scheduling plan.

[0058] In a preferred embodiment, prioritizing computational problems includes:

[0059] User data and electricity hazard information are obtained from a multi-user environment to construct an initial dataset. User data is then classified into user groups using preset classification rules.

[0060] Analyze the criticality of each user in the user group, combine it with information on potential electrical hazards, determine the hazard assessment result, and if the hazard assessment result exceeds the preset threshold, obtain the severity score;

[0061] Based on the severity score and the user's level of importance, the priority of the issues is calculated, a solution for each issue is generated, and potential risk points are identified based on the distribution characteristics of high-priority issues.

[0062] The specific plan for hazard assessment is as follows:

[0063] 1. Define user criticality

[0064] User types are assigned different weights. For example, commercial users have higher requirements for the continuity and stability of power supply, so a weight of 0.8 can be assigned; industrial users also have high requirements for the stability of power supply, so a weight of 0.7 can be assigned; and residential users have relatively lower requirements for power supply, so a weight of 0.5 can be assigned. The weighting is primarily based on the intensity of the business demand for power supply reliability from different user types, the impact of historical power outages on electricity consumption, and is determined through the experience of industry experts. In practical applications, this weighting system can be adaptively adjusted according to the structural characteristics of different regions (such as the proportion of industrial areas, commercial centers, and residential areas).

[0065] The larger the user base, the greater their dependence on electricity supply and the higher their criticality. For example, large commercial complexes or large factories may be more critical than small businesses and ordinary residents. Criticality can be quantified based on data such as user transformer capacity and historical maximum load in each region, through normalization and weighted according to regional importance determined by expert experience.

[0066] User importance is determined by the specific characteristics of certain users, even if they are not large in scale. For example, critical facilities such as hospitals and data centers can be assigned the highest level of importance. The identification and importance assignment of such users can be determined based on the real-time power grid usage of each user in each region, and set according to expert experience.

[0067] 2. Define the severity of the hazard.

[0068] The types of hazards have varying degrees of impact on the power system. For example, voltage fluctuations may cause equipment instability, with a severity level of 6; current overloads may cause equipment damage, with a severity level of 8; and equipment aging may lead to frequent failures, with a severity level of 5. The weight of each hazard type can be determined based on historical probability, average repair time, and severity data of equipment failures or power outages caused by that type of hazard in each region, and with reference to equipment technical parameters. The weights can be adjusted accordingly, as the health status and operating environment of equipment vary across different power grids.

[0069] The scope of a hazard's impact is determined by the number of users affected; the greater the number of users affected, the higher the severity. For example, a hazard affecting only a single user has relatively low severity, while a hazard affecting multiple users has higher severity. The scope of impact can be quantified by combining it with the criticality of users, calculating the weighted number of affected users, and setting the threshold based on the average user density and network structure of each region.

[0070] The frequency of hazard occurrence is a key factor; frequently occurring hazards require more attention than those that occur only occasionally. For example, a hazard that occurs multiple times a week is more serious than one that occurs only once a month. Frequency levels can be determined by analyzing historical alarm counts over time to establish thresholds.

[0071] The longer a hazard persists, the greater its impact on the power system. For example, a hazard lasting for several hours is more severe than one lasting only a few minutes. Duration levels can be categorized based on the permissible abnormal operating time standards for different equipment.

[0072] 3. Hazard Assessment Model

[0073] By combining the user's criticality and the severity of the hazard, a hazard assessment model is constructed. The specific formula is: Hazard assessment result = User criticality weight × Hazard severity score.

[0074] For example, suppose user A is a residential user with a criticality weight of 0.5. There is a potential voltage fluctuation hazard, and the severity score of this hazard is 6. Then the hazard assessment result is 0.5 × 6 = 3.

[0075] 4. Threshold determination

[0076] A threshold is set to determine whether a potential hazard requires further action. For example, the threshold could be set to 3. If the hazard assessment result exceeds this threshold, the hazard is considered to require further action; if it is below or equal to this threshold, it can be temporarily marked as a low-priority hazard, awaiting further observation or action. The threshold can be determined based on historical data, by analyzing the proportion of hazards that actually develop into failures (i.e., risk probability) within different historical hazard assessment value ranges, and selecting a critical point that balances early warning sensitivity and workload through expert experience. The threshold can be initially set as an empirical value (e.g., 3), and then optimized and adjusted based on local operational data.

[0077] Among them, obtaining the distribution characteristics of high-priority problems includes:

[0078] First, relevant data on high-priority issues need to be extracted from the ranking results, including information such as the location, type, and severity of the issues. Then, these issues are visualized in geospatial space using a Geographic Information System (GIS) or other data analysis tools to observe their distribution.

[0079] Specifically, data extraction is performed to filter out issues with a priority higher than a preset threshold from the sorted results. This threshold can be dynamically set based on the actual situation in each region and through expert experience. Detailed information about these issues is extracted, including location (e.g., latitude and longitude), type, and severity. Geographic visualization is then performed using GIS tools to mark these issues on a map and observe their spatial distribution. Heatmaps, dot maps, and other methods can be used to visually display the distribution density and concentrated areas of the issues.

[0080] Furthermore, determining the distribution characteristics includes:

[0081] Using geographic visualization tools, observe whether high-priority issues exhibit a clustering trend. This clustering trend can be identified in the following ways:

[0082] Hotspot analysis: Using hotspot analysis tools in GIS, areas with high problem density are identified.

[0083] Cluster analysis uses clustering algorithms (such as K-means) to cluster the problem locations and determine whether there are obvious clustering regions.

[0084] Spatial autocorrelation analysis uses spatial autocorrelation statistical methods to determine whether there is spatial clustering in the distribution of problems.

[0085] An improved logistic regression model was used to predict and analyze hazard data in concentrated areas to identify potential risk points, including:

[0086] Historical hazard data from concentrated areas are extracted as a training set, including known high-priority and low-priority issues. A logistic regression model is used to train the training set, and the model parameters are adjusted to improve prediction accuracy. Current hazard data from concentrated areas are input into the model to predict the potential risk level of each hazard point. The predictive performance of the model is evaluated using methods such as confusion matrix and receiver operating characteristic (ROC) curve to ensure the reliability and accuracy of the model.

[0087] Based on the model's predictions, potential risk points within the concentrated area are identified. These potential risk points can serve as the focus for subsequent resource allocation and handling, as detailed below:

[0088] Risk point screening involves selecting potential risk points with probabilities higher than a certain threshold (e.g., 0.8) based on the probability values ​​predicted by the model. This probability threshold can be set during the model validation phase by using a weighted average of the false positive rate and the false negative rate.

[0089] Risk assessment involves comprehensively evaluating potential risk points based on factors such as the type and severity of the hazard to determine their risk level. The specific plan is as follows:

[0090] 1. Construct an evaluation indicator system

[0091] Calculate the weights for different hazard types, as they have varying degrees of impact on the power system. For example, equipment aging hazards may primarily affect the reliability of local equipment, while voltage fluctuation hazards may affect all electrical equipment in the entire area. Assign a weight value to each hazard type. For instance, equipment aging hazards might have a weight of 0.6, voltage fluctuation hazards 0.8, and current overload hazards 0.9. These weight values ​​can be determined based on historical data and expert experience, or learned from large datasets using machine learning algorithms. The weighting can be based on local power grid equipment failure records, maintenance costs, and outage impact analysis. The weights for the same type of hazard may differ across regions due to variations in equipment models, service life, and environmental conditions.

[0092] The severity score of a potential hazard is calculated, and severity can be measured based on multiple dimensions. These include the hazard's impact range (the more users affected, the higher the severity), occurrence frequency (frequently occurring hazards require more attention than occasional ones), and duration (the longer a hazard lasts, the greater its impact on the power system). A scoring standard can be set for each dimension, and the scores from these dimensions are weighted and summed to obtain the total severity score. For example, if the impact range score is 1-5, the occurrence frequency score is 1-5, and the duration score is 1-5, with a total weight of 1, the total severity score = impact range score × 0.4 + occurrence frequency score × 0.3 + duration score × 0.3. The weights for each dimension (0.4, 0.3, 0.3) are for illustrative purposes; the actual weight allocation needs to be determined by analyzing the correlation between each dimension and the final failure loss in historical data, or by expert scoring using methods such as the Analytic Hierarchy Process (AHP). The emphasis varies in different business scenarios, and the weights are adjustable.

[0093] Obtain a user criticality coefficient, considering the importance of the user at the location of the potential hazard. For critical users (such as hospitals and data centers), even if the severity of the hazard itself is not high, the risk level may be increased due to the user's criticality. A criticality coefficient is assigned to each user based on factors such as user type, user size, and user importance. For example, the criticality coefficient is 1.2 for commercial users, 1.1 for industrial users, 1.0 for residential users, and 1.5 for critical facility users (such as hospitals). The coefficient setting must be logically consistent with the technical evaluation system for user criticality weighting. Its specific value is first determined based on the mandatory power supply reliability level classification in national and industry standards; then, it is calibrated by combining the user's load location in the power grid topology, the nature of the supplied load, and its quantitative impact on system operation safety, and finally determined through expert experience.

[0094] A comprehensive assessment model is constructed, which can be achieved by weighted summation, combining the hazard type weight, hazard severity score, and user criticality coefficient. The formula is: Comprehensive Risk Assessment Value = Hazard Type Weight × Hazard Severity Score × User Criticality Coefficient. For example, a hazard is current overload (hazard type weight is 0.9), its hazard severity score is 7 (calculated from the total severity score above), and its location is in a commercial user area (user criticality coefficient is 1.2). Then, the comprehensive risk assessment value = 0.9 × 7 × 1.2 = 7.56.

[0095] 2. Risk level classification: Risk levels can be divided into four levels:

[0096] Low risk indicates that the potential hazard has a minor impact on the power system and, even if it occurs, will not cause serious consequences. For example, minor equipment aging hazards, and where users in the area have low power supply requirements, result in a low overall risk assessment value.

[0097] Medium risk means that the potential hazard has a certain possibility of adversely affecting the power system, but the scope and extent of the impact are limited. For example, a voltage fluctuation hazard occurring in an ordinary residential area with a small fluctuation amplitude would be considered a medium-level risk in the overall risk assessment.

[0098] High risk: High-risk hazards may cause significant damage to the power system, affecting a wide range of areas and potentially disrupting normal electricity use for users. For example, a current overload hazard occurring in an industrial user area, with a high degree of overload, results in a high overall risk assessment value.

[0099] Extremely high risk is the highest level of risk, indicating that the potential hazard has a very serious impact on the power system and may lead to widespread power outages or severe damage to critical facilities. For example, the presence of multiple potential hazards (such as aging equipment and current overload) in an area with critical facilities (such as a hospital) results in an extremely high overall risk assessment value.

[0100] The risk level classification criteria are as follows:

[0101] Risk levels are classified based on the range of comprehensive risk assessment values. For example:

[0102] Low risk, overall risk assessment value <3

[0103] Medium risk, 3 ≤ comprehensive risk assessment value < 6

[0104] High risk, 6 ≤ comprehensive risk assessment value < 9

[0105] Extremely high risk, with a comprehensive risk assessment value ≥9.

[0106] The threshold values ​​(3, 6, 9) mentioned above are example values ​​for this embodiment. In practical applications, these thresholds can be calibrated by analyzing the correspondence between historical risk assessment values ​​and actual accidents. For example, the optimal classification threshold can be determined using ROC curves, or it can be derived backwards from management's tolerance for the quantity of risks at each level (such as the allowed number of high-risk hazards). Different regions can set different classification standards based on their risk tolerance and resource conditions.

[0107] When allocating resources, resources should be prioritized for handling based on risk level to ensure that high-risk hazards are resolved in a timely manner.

[0108] For example, the entire process of acquiring user data and electricity hazard information in a multi-user environment, and calculating the priority of problems by analyzing the importance of users and the severity of hazards, can be automated through information technology.

[0109] First, real-time electricity consumption data is collected from users through smart meters and IoT devices to construct an initial dataset. For example, this dataset might contain the electricity consumption, voltage fluctuations, and equipment operating status of 100 households in a residential community. Assume user A's average monthly electricity consumption is 500 kWh, and their voltage fluctuation rate is 5%, exceeding the normal range of 2%. Next, user information is integrated using a database, and users are categorized by importance with assigned weights. For example, commercial users have a weight of 0.8, ordinary residents 0.5, and user A is an ordinary resident with a weight of 0.5. Simultaneously, the severity of potential electricity hazards is assessed. Based on historical data and a rule engine, hazard levels are defined. For example, a voltage fluctuation rate exceeding 3% is considered a medium-level hazard, with a score of 6 (out of 10). 10) User A's hazard score is 6. Then, using the priority calculation formula: Priority = User Weight × Hazard Score × 10, User A's priority is calculated to be 0.5 × 6 × 10 = 30. Another user, B, is a business user with a weight of 0.8 and a hazard score of 8, so their priority is 0.8 × 8 × 10 = 64. Using a sorting algorithm (such as quicksort), all users' priorities are arranged from high to low to generate a priority processing list, showing that User B is ranked before User A. Furthermore, it can be linked to business scenarios, such as automatically pushing users with a priority higher than 50 to the maintenance scheduling system to ensure timely handling of hazards, forming a complete logical chain from data collection to problem sorting to business scheduling.

[0110] It should be noted that the above steps not only enable data-driven decision-making but also improve the efficiency and accuracy of handling potential electricity hazards. The coefficient "10" in the formula is used to amplify the priority value to a range that is easy to process and compare. Its size does not affect the relative relationship of the sorting and can be adjusted according to the actual system value range.

[0111] In a preferred embodiment, step S101, obtaining a preliminary scheduling plan includes:

[0112] By extracting relevant information from user data sources (user electricity consumption data, including electricity consumption, voltage fluctuations, and equipment operating status), and combining it with data on potential electricity hazards, a basic analysis dataset is constructed to obtain a preliminary set of problem categories. A weighted calculation method is then used to quantify the priority of the problems and determine the order of processing.

[0113] Based on the number of processing orders and the data on urgency levels, if the urgency level exceeds a preset threshold, the resource allocation ratio is adjusted first to obtain the adjusted scheduling framework. Combined with the time window, the processing sequence of each problem is optimized to obtain the optimized timing arrangement.

[0114] Based on the optimized time sequence and combined with the regional distribution map, the spatial clustering characteristics of potential hazards are analyzed to determine the range of high-risk areas. Through predictive processing, the distribution characteristics of potential hazards are obtained. Combined with data in the historical record database, the final scheduling and execution plan is determined.

[0115] If the urgency level exceeds the preset threshold (e.g., 7 for residential users, 6 for commercial users, 6 for industrial users, and 5 for critical facilities (such as hospitals and data centers), the thresholds for urgency levels (5, 6, 7) can be set according to business regulations regarding the maximum allowed duration of potential hazards for different types of users. For example, critical facilities require the shortest response time, so their thresholds are the lowest (most likely to trigger priority adjustment). These thresholds are consistent with subsequent response time optimization standards (such as 6 hours in step S103) and can be adjusted according to the service commitment level of different regions), then the resource allocation ratio is adjusted first to obtain the adjusted scheduling framework. Combined with the time window, the processing sequence of each problem is optimized to obtain the optimized timing arrangement, as follows:

[0116] 1. Determine resource allocation priorities

[0117] Prioritization: Sort all issues according to their priority (combining user importance, severity of potential risks, and urgency level), and prioritize issues whose urgency level exceeds the threshold.

[0118] 2. Adjustment of resource allocation ratio

[0119] The resource allocation ratio is calculated initially based on the priority of the problem and the total amount of resources. For example, if the total resources are 100 units and there are 10 problems to be processed, the initial resource allocation ratio for each problem is 10 units. The resource allocation ratio is then adjusted: for problems whose urgency level exceeds a threshold, their resource allocation ratio is increased, while the resource allocation ratio for other problems is decreased to ensure a balance in the total resources.

[0120] Adjustment strategies include increasing the resource allocation ratio, the magnitude of which can be determined based on the urgency of the problem and the total available resources. For example, for a problem with an urgency level of 8, the resource allocation ratio could be increased by 20%. Alternatively, the resource allocation ratio can be decreased for non-urgent problems. The magnitude of the decrease can be determined based on resource gaps and the problem's priority. For example, for lower-priority problems, the resource allocation ratio could be decreased by 10%. The increase / decrease ratios (e.g., 20%, 10%) are example parameters, and these ratios can be designed as functions positively correlated with the extent to which the urgency level exceeds a threshold, for example: Increase Ratio = k × (Urgency Level - Threshold). The coefficient k needs to be determined through simulation or historical data analysis to ensure the efficiency and fairness of resource adjustments.

[0121] 3. Resource allocation ratio adjustment formula

[0122] Increase the resource allocation ratio. The adjusted resource allocation ratio = initial resource allocation ratio × (1 + increase ratio). For example, if the initial resource allocation ratio is 10 units and the increase ratio is 20%, then the adjusted resource allocation ratio is 10 × (1 + 0.2) = 12 units.

[0123] Reduce the resource allocation ratio. The adjusted resource allocation ratio = initial resource allocation ratio × (1 - reduction ratio). For example, if the initial resource allocation ratio is 10 units and the reduction ratio is 10%, then the adjusted resource allocation ratio is 10 × (1 - 0.1) = 9 units.

[0124] 4. Constraints on adjusting resource allocation ratios

[0125] To balance the total resources, the sum of the adjusted resource allocation ratios must not exceed the total resources. If the total resources are insufficient after adjustment, the allocation ratios need to be further adjusted, or additional resources need to be requested. For example, if the total resources are 100 units, and the adjusted allocation ratio is 110 units, the allocation ratios need to be readjusted, or an additional 10 units of resources need to be requested.

[0126] A minimum resource allocation ratio is required for each problem; this ratio must not fall below a minimum to ensure that the problem can be adequately addressed. For example, the minimum resource allocation ratio for each problem could be set to 5 units. This minimum resource allocation ratio can be determined based on the minimum human and material costs required to handle that type of problem.

[0127] For example, suppose there is a multi-user environment with a total resource of 100 units, and 5 issues that need to be addressed. The urgency levels of the issues and the initial resource allocation are as follows:

[0128] Question 1 has an urgency level of 7 and an initial resource allocation ratio of 20 units.

[0129] Question 2 has an urgency level of 6 and an initial resource allocation ratio of 20 units.

[0130] Question 3 has an urgency level of 8 and an initial resource allocation ratio of 20 units.

[0131] Question 4 has an urgency level of 5 and an initial resource allocation ratio of 20 units.

[0132] Question 5 has an urgency level of 9 and an initial resource allocation ratio of 20 units.

[0133] For example, let's determine an urgency level threshold. Assume the urgency level threshold is 7. Problems 3 and 5 exceed this threshold, requiring priority adjustment of resource allocation ratios. Further increasing the resource allocation ratio: Problem 3's urgency level is 8, increasing by 20% (20 × (1 + 0.2) = 24 units); Problem 5's urgency level is 9, increasing by 30% (20 × (1 + 0.3) = 26 units). Decreasing the resource allocation ratio: Problem 1's urgency level is 7, decreasing by 10% (20 × (1 - 0.1) = 18 units); Problem 2's urgency level is 6, decreasing by 10% (20 × (1 - 0.1) = 18 units); Problem 4's urgency level is 5, decreasing by 10% (20 × (1 - 0.1) = 18 units).

[0134] The adjusted resource allocation ratios are as follows: Problem 1: 18 units; Problem 2: 18 units; Problem 3: 24 units; Problem 4: 18 units; Problem 5: 26 units.

[0135] A resource balance check revealed that the adjusted total resource allocation ratio is 18 + 18 + 24 + 18 + 26 = 104 units. This exceeds the total resource limit of 100 units, requiring further adjustment. The resource allocation ratios for problems 3 and 5 can be appropriately reduced, or an additional 4 units of resources can be requested. Assuming an additional 4 units of resources are requested, the final resource allocation ratios would be: Problem 1 18 units, Problem 2 18 units, Problem 3 24 units, Problem 4 18 units, and Problem 5 26 units.

[0136] In the graded early warning system for electrical safety hazards, optimizing the processing sequence of each issue is a key step to ensure efficient resource utilization and timely problem resolution. The following are specific optimization schemes and examples:

[0137] 1. Processing timing optimization

[0138] Define a time window to define the start and end times for handling a problem. For example, a day can be divided into multiple time windows, such as the morning shift (8:00-14:00), the afternoon shift (14:00-20:00), and the evening shift (20:00-02:00). The division of time windows should be based on the urgency of the problem and the availability of resources. Urgent problems should be prioritized for handling within time windows with sufficient resources.

[0139] 2. Optimize objectives

[0140] Minimize total processing time by rationally scheduling processing times to reduce the overall processing time for all issues. Maximize resource utilization by ensuring resources are fully utilized within each time window, avoiding resource idleness or overload. Meet urgency requirements by prioritizing high-urgency issues and ensuring their completion within the stipulated time.

[0141] 3. Optimization Methods

[0142] Prioritization is performed by sorting issues according to their priority (combining user importance, hazard severity, and urgency level). Dynamic scheduling then occurs, adjusting the processing sequence based on real-time resource allocation data and issue priorities. If resources are insufficient for a particular issue, its processing time can be postponed to the next time window, or additional resources can be requested. Time windows are optimized by determining the optimal processing time window for each issue. For example, urgent issues are prioritized for processing within resource-sufficient time windows. If resources are insufficient within a particular time window, some issues can be moved to other time windows. The processing sequence is adjusted based on the issue's processing time and resource requirements. For example, if an issue has a long processing time, it can be scheduled for processing within resource-sufficient time windows. Urgent issues are processed within the current time window whenever possible. If resources are insufficient, the processing sequence of other issues can be adjusted appropriately to ensure that urgent issues are prioritized.

[0143] 4. Algorithm Optimization

[0144] A greedy algorithm can be used to prioritize high-urgency issues and process them sequentially downwards. Within each time window, based on resource allocation, the highest-priority issue is selected for processing. The problem-solving sequence optimization problem is modeled as a dynamic programming problem, and the optimal processing sequence is solved recursively. Dynamic programming can consider multiple time windows and resource allocation scenarios to find the globally optimal processing sequence.

[0145] Heuristic algorithms (such as genetic algorithms and simulated annealing algorithms) can also be used to find near-optimal processing timing. Heuristic algorithms can find better solutions in a shorter time and are suitable for optimizing the processing timing of complex problems.

[0146] For example, suppose there is a multi-user environment with a total resource of 100 units, and 5 issues that need to be addressed. The urgency levels of the issues and the initial resource allocation are as follows:

[0147] Problem 1 has an urgency level of 7, an initial resource allocation ratio of 20 units, and an estimated processing time of 2 hours.

[0148] Problem 2 has an urgency level of 6, an initial resource allocation ratio of 20 units, and an estimated processing time of 3 hours.

[0149] Problem 3 has an urgency level of 8, an initial resource allocation ratio of 20 units, and an estimated processing time of 1 hour.

[0150] Problem 4 has an urgency level of 5, an initial resource allocation ratio of 20 units, and an estimated processing time of 4 hours.

[0151] Problem 5 has an urgency level of 9, an initial resource allocation ratio of 20 units, and an estimated processing time of 2 hours.

[0152] For example, if the urgency level threshold is 7, then issues 3 and 5 exceed the urgency level and need to be prioritized. Assume a day is divided into three time windows: morning shift (08:00-14:00), afternoon shift (14:00-20:00), and evening shift (20:00-02:00).

[0153] The priorities are ranked as follows: Problem 5 (urgency level 9), Problem 3 (urgency level 8), Problem 1 (urgency level 7), Problem 2 (urgency level 6), and Problem 4 (urgency level 5). Dynamic scheduling is then implemented: For the morning shift (08:00-14:00), Problems 5 and 3 are prioritized. Problem 5, with an estimated processing time of 2 hours, is scheduled for 08:00-10:00, and Problem 3, with an estimated processing time of 1 hour, is scheduled for 10:00-11:00. For the afternoon shift (14:00-20:00), Problems 1 and 2 are processed. Problem 1, with an estimated processing time of 2 hours, is scheduled for 14:00-16:00, and Problem 2, with an estimated processing time of 3 hours, is scheduled for 16:00-19:00. For the evening shift (20:00-02:00), Problem 4 is processed. Problem 4, with an estimated processing time of 4 hours, is scheduled for 20:00-24:00.

[0154] Then, resource allocation is adjusted. The resource allocation ratio is increased: the urgency level of problem 3 is 8, so the resource allocation ratio is increased by 20% (20 × (1 + 0.2) = 24 units); the urgency level of problem 5 is 9, so the resource allocation ratio is increased by 30% (20 × (1 + 0.3) = 26 units). The resource allocation ratio is decreased: the urgency level of problem 1 is 7, so the resource allocation ratio is decreased by 10% (20 × (1 - 0.1) = 18 units); the urgency level of problem 2 is 6, so the resource allocation ratio is decreased by 10% (20 × (1 - 0.1) = 18 units); the urgency level of problem 4 is 5, so the resource allocation ratio is decreased by 10% (20 × (1 - 0.1) = 18 units).

[0155] The final processing schedule is as follows: Morning shift (08:00-14:00): Problem 5 will be processed from 08:00-10:00 with a resource allocation ratio of 26 units; Problem 3 will be processed from 10:00-11:00 with a resource allocation ratio of 24 units. Afternoon shift (14:00-20:00): Problem 1 will be processed from 14:00-16:00 with a resource allocation ratio of 18 units; Problem 2 will be processed from 16:00-19:00 with a resource allocation ratio of 18 units. Evening shift (20:00-02:00): Problem 4 will be processed from 20:00-24:00 with a resource allocation ratio of 18 units.

[0156] A resource balance check revealed that the adjusted total resource allocation ratio is 26 + 24 + 18 + 18 + 18 = 104 units. This exceeds the total resource limit of 100 units, requiring further adjustment. The resource allocation ratios for problems 3 and 5 could be appropriately reduced, or an additional 4 units of resources could be requested. Assuming an additional 4 units of resources are requested, the final resource allocation ratio and processing sequence are as follows: Early shift (08:00-14:00): Problem 5 is processed from 08:00-10:00 with a resource allocation ratio of 26 units; Problem 3 is processed from 10:00-11:00 with a resource allocation ratio of 24 units. Middle shift (14:00-20:00): Problem 1 is processed from 14:00-16:00 with a resource allocation ratio of 18 units; Problem 2 is processed from 16:00-19:00 with a resource allocation ratio of 18 units. Late shift (20:00-02:00): Problem 4 is processed from 20:00-24:00 with a resource allocation ratio of 18 units.

[0157] Furthermore, by analyzing the spatial clustering characteristics of potential hazard points, the high-risk area was determined to include:

[0158] Extract the location information (such as latitude and longitude) and related attributes (such as hazard type, severity, and treatment priority) of all potential hazards from the optimized time sequence; use Geographic Information System (GIS) tools to mark the potential hazards on the map and generate a potential hazard distribution map, which can intuitively display the distribution density and clustering of potential hazards through point maps, heat maps, and other methods;

[0159] During spatial analysis, hotspot analysis tools in GIS (such as kernel density estimation) can be used to identify areas with high density of potential hazard points and determine the scope of high-risk areas. Alternatively, clustering algorithms (such as density-based clustering or K-means) can be used to cluster the locations of potential hazard points, identify obvious clustering areas, and use spatial autocorrelation statistical methods to determine whether there is spatial clustering in the distribution of potential hazard points, further confirming high-risk areas.

[0160] Based on the results of cluster analysis or hotspot analysis, the boundaries of high-risk areas are delineated. Buffer analysis tools in GIS can be used to generate buffer zones of a certain radius for each cluster center, serving as high-risk areas. The rationality of these high-risk areas is verified using historical data and expert knowledge. If the density of potential hazards is found to be low in certain areas, the boundaries can be adjusted appropriately. Parameters such as the number of clusters K in clustering algorithms (e.g., K-means) or the neighborhood radius and minimum number of points in density-based algorithms (e.g., DBSCAN) can be determined through adjustments using historical or simulated data from each region, using indicators such as the silhouette coefficient, to find parameters that best reflect the spatial distribution characteristics of local potential hazards.

[0161] Furthermore, for data within high-risk areas, a logistic regression model is used for predictive processing to obtain the distribution characteristics of potential hazards. Historical hazard data is extracted from high-risk areas as a training set, including known hazard points and non-hazard points. Data features can include location information, hazard type, severity, equipment aging, user electricity consumption patterns, etc. Adding more hazard-related features, such as equipment operating time, historical maintenance records, and weather conditions, can improve the model's predictive ability. The logistic regression model is used to train the training set, and the model's hyperparameters can be tuned using methods such as cross-validation to improve the model's generalization ability.

[0162] It should be noted that, due to the relatively small number of potential hazard points, data imbalance may occur. Oversampling (such as SMOTE) or undersampling methods can be used to balance the dataset. Feature importance analysis is used to select the features most helpful for hazard prediction, reducing model complexity. Current hazard data within high-risk areas are input into the model to predict the probability of potential hazards at each location. The model's predictive performance is evaluated using confusion matrices, ROC curves, etc., to ensure the model's reliability and accuracy.

[0163] Furthermore, based on the model's prediction results, the distribution characteristics of potential hazards within high-risk areas are determined. Locations with probabilities exceeding a certain threshold (e.g., 0.8) are selected as potential hazard points based on the model's predicted probability values. Combining hazard type and severity factors, a comprehensive assessment of these potential hazard points is conducted to determine their risk level. The spatial distribution characteristics of these potential hazard points, such as distribution density and clustering degree, are analyzed. A distribution map of these potential hazard points can be generated using GIS tools to visually demonstrate their distribution.

[0164] The determination of the probability threshold (e.g., 0.8) can be based on the model's performance on historical validation data, selecting the optimal classification boundary through quantitative analysis. After model training, it is tested using historical data not used in training (validation set) to obtain the predicted probability and true label for each sample. Precision and recall at different probability thresholds are calculated by plotting ROC curves and applying expert experience. For example, if business requirements dictate minimizing false positives (i.e., the selected points must be highly reliable), a probability threshold (perhaps 0.82 or 0.78) corresponding to achieving high precision (e.g., 95%) will be chosen as the application threshold. This process is entirely driven by historical data, ensuring the objectivity and reproducibility of the threshold. For models in different regions or with different hazard types, the threshold must be determined independently using local data.

[0165] Furthermore, considering factors such as the type and severity of the hazard, a comprehensive assessment of potential hazard points is conducted to determine their risk level, specifically including:

[0166] 1. Construction of Evaluation Indicator System

[0167] The weighting of hazard types is crucial, as different types of hazards have varying degrees of impact on the power system. A weight value is assigned to each hazard type. For example, equipment aging hazards have a weight of 0.6, voltage fluctuation hazards have a weight of 0.8, and current overload hazards have a weight of 0.9. Weights can be set based on equipment operation data and impact consequence analysis, namely: historical fault / alarm data, statistically analyzing the probability that various hazards that have occurred in the local power grid in the past will eventually evolve into actual faults (such as tripping or equipment damage); equipment impact range data, analyzing the location of equipment associated with different hazard types in the local power grid topology, and assessing the potential impact range (such as the number of feeders and users affected); and handling time and complexity data, based on historical work orders, analyzing the average time and required resources (such as personnel skill levels, spare parts, and power outage requirements) for handling different types of hazards. Hazards that are more complex and take longer to handle typically have higher weights. The above analysis uses the local power grid's historical operation database, equipment ledger, and topology data. Preliminary weighting relationships are derived through data analysis (such as calculating the actual failure rate caused by different types of hidden dangers). These relationships are then reviewed and confirmed by domain experts in conjunction with the characteristics of the local power grid, ultimately forming a set of weight values ​​adapted to the local power grid characteristics and operation and maintenance strategies. The weight values ​​should vary depending on the equipment condition, grid structure, and operation and maintenance capabilities of different regions.

[0168] The severity score for a hazard can be measured from multiple dimensions, such as the hazard's impact range, frequency of occurrence, and duration. A scoring criterion is set for each dimension, and the scores from these dimensions are weighted and summed to obtain the total severity score. Each scoring criterion is set based on expert experience and actual data from each region.

[0169] The model predicts probabilities by using logistic regression or other predictive models to forecast potential hazard points, obtaining a predicted probability value for each hazard point. A higher predicted probability value indicates a greater likelihood that the hazard point will become an actual hazard.

[0170] The user criticality coefficient takes into account the importance of users in the location of potential hazards. For critical users (such as hospitals and data centers), even if the severity of the hazard itself is not high, the risk level may be increased due to the user's criticality. Based on factors such as user type, user size, and user importance in each region, a criticality coefficient is set for each user through expert experience.

[0171] 2. Construct a comprehensive evaluation model

[0172] The comprehensive assessment model can use a weighted summation method to combine the hazard type weight, hazard severity score, model prediction probability, and user criticality coefficient. The formula is: Comprehensive Risk Assessment Value = Hazard Type Weight × Hazard Severity Score × Model Prediction Probability × User Criticality Coefficient.

[0173] 3. Risk Level Classification

[0174] Risk levels are classified based on the range of comprehensive risk assessment values. For example:

[0175] Low risk, overall risk assessment value <3

[0176] Medium risk, 3 ≤ comprehensive risk assessment value < 6

[0177] High risk, 6 ≤ comprehensive risk assessment value < 9

[0178] Extremely high risk, with a comprehensive risk assessment value ≥9.

[0179] Taking the threshold values ​​(3, 6, 9) as an example, which introduce predicted probabilities, the range of the comprehensive risk assessment value may differ from before, therefore the threshold needs to be recalibrated. The calibration method involves collecting a batch of historical assessment values ​​of "potential hazard points" and their subsequent actual development, and determining the optimal cutoff point through classification performance evaluation (such as ROC curves).

[0180] 4. Model prediction probability

[0181] Potential hazard points are screened by selecting locations with probabilities higher than a certain threshold (e.g., 0.8) based on the probability values ​​predicted by the model. For example, if the model predicts a probability of 0.85 for a certain hazard point, then that hazard point is selected as a potential hazard point.

[0182] In a comprehensive evaluation model, the probability value's weight can be considered an important weighting factor. For example, the weight of the model's predicted probability can be set to 0.5 to ensure the prediction results have an impact on the comprehensive evaluation. The probability value weight (e.g., 0.5) is an adjustment coefficient used in the comprehensive risk assessment model to balance the relative importance of "model prediction probability" and "traditional assessment elements (hazard type weight × severity score)". Its value is determined through a parameter optimization process based on historical data. Specifically, a batch of historical hazard data is collected, including its traditional assessment element values, model prediction probability values, and whether the hazard subsequently caused a failure (yes / no). Optimization objectives and indicators are defined to find the weights that make the "comprehensive risk value" most accurately distinguish between historical "failure points" and "non-failure points". Optimization indicators can be the area under the AUC-ROC curve or the F1 score. Parameters are searched and determined using methods such as grid search to try a series of candidate weight values ​​(e.g., 0, 0.1, 0.2, ..., 1.0). For each candidate weight, the comprehensive risk value of all historical samples is calculated, and its effectiveness in predicting real failures is evaluated (e.g., calculating AUC). Finally, the candidate weight that makes the effectiveness indicator reach the highest value is selected as the final value.

[0183] 5. Spatial Distribution Characteristics Analysis

[0184] Distribution density: GIS tools are used to calculate the density of potential hazard points within each area. For example, a heatmap can be used to visually display the distribution density of hazard points. Clustering degree: Cluster analysis methods (such as K-means or DBSCAN) are used to analyze the degree of clustering of potential hazard points. For example, cluster analysis can identify clustered areas of hazard points. Spatial autocorrelation analysis: Spatial autocorrelation statistical methods (such as Moran's I) are used to determine whether there is spatial clustering in the distribution of hazard points.

[0185] For example, suppose there are three potential hazard points in a high-risk area. Their relevant data includes: Hazard Point A, Hazard Type: Current Overload (weight 0.9), Severity Score: 7, Model Prediction Probability: 0.85, User Criticality Coefficient: 1.2; Hazard Point B, Hazard Type: Voltage Fluctuation (weight 0.8), Severity Score: 6, Model Prediction Probability: 0.9, User Criticality Coefficient: 1.0; Hazard Point C, Hazard Type: Equipment Aging (weight 0.6), Severity Score: 5, Model Prediction Probability: 0.8, User Criticality Coefficient: 1.1. Then, the comprehensive risk assessment values ​​are calculated: Hazard Point A's comprehensive risk assessment value = 0.9 × 7 × 0.85 × 1.2 = 6.426, high risk level; Hazard Point B's comprehensive risk assessment value = 0.8 × 6 × 0.9 × 1.0 = 4.32, medium risk level; and Hazard Point C's comprehensive risk assessment value = 0.6 × 5 × 0.8 × 1.1 = 2.64, low risk level. Spatial distribution characteristic analysis was performed, including the generation of a heat map using GIS tools to display the distribution density of potential hazard points A, B, and C. It was found that potential hazard points A and B are concentrated in the southeastern part of the city, with a higher distribution density. K-means clustering analysis was used to determine the degree of clustering, revealing that potential hazard points A and B cluster together, forming a clear high-risk cluster area. Spatial autocorrelation analysis using Moran's I statistical method confirmed the significant spatial clustering of the potential hazard points. Finally, a distribution map was generated using GIS tools, visually displaying the distribution of potential hazard points A, B, and C. This distribution map clearly shows the distribution density and clustering areas of the potential hazard points, providing a clear basis for resource allocation and treatment prioritization.

[0186] For example, suppose that in a certain urban area, the location information of all potential hazard points was extracted through an optimized time-series arrangement, and a hazard point distribution map was generated using GIS tools. Hotspot analysis revealed that the hazard points are mainly concentrated in the southeastern part of the city, forming a clear high-risk area. Then, historical hazard data was extracted from this high-risk southeastern area as a training set, including features such as hazard type, severity, and equipment aging. A logistic regression model was used to train the training set, and the model parameters were fine-tuned through cross-validation. After the model training was completed, current hazard data from the southeastern area was input into the model to predict the probability of potential hazards at each location. Based on the model prediction results, locations with a probability higher than 0.8 were selected as potential hazard points, and their risk levels were determined. Finally, a distribution map of potential hazard points was generated using GIS tools, visually displaying their distribution. It was found that potential hazard points are more densely distributed in certain specific areas of the southeastern region (such as near industrial parks).

[0187] For example, in a multi-user electricity management scenario, a weighted algorithm is used to prioritize issues and formulate a preliminary scheduling plan. First, abnormal electricity usage data for 200 users in a certain area is extracted from the database. For instance, user C's current load rate reaches 85%, exceeding the safety threshold by 15%, while user D's load rate is 90%, exceeding the safety threshold by 20%. Then, weight coefficients are assigned based on user type and electricity usage characteristics. Industrial users are assigned a weight of 0.9, and residential users 0.4. User C is a residential user with a weight of 0.4, and user D is an industrial user with a weight of 0.9. Next, a built-in evaluation model scores the severity of the anomalies, with a maximum score of 10 points. A load rate exceeding 10% to 20% scores 5, and exceeding 20% ​​scores 7. User C scores 5, and user D scores... The priority is set at 7. Further, a weighted calculation formula is used: Priority = Weight × Severity Score × 5. User C's priority is calculated as 0.4 × 5 × 5 = 10, and user D's priority is 0.9 × 7 × 5 = 31.5. Then, a heap sort algorithm is used to sort all users' priorities from highest to lowest, generating a ranking result where user D is ranked before user C. Finally, the priority results are linked to resource allocation to automatically generate a scheduling plan. For example, users with a priority higher than 20 will be associated with the emergency response queue. Based on current maintenance resource data, it is calculated that user D needs to be processed within 24 hours, while user C is scheduled for processing within 72 hours. Simultaneously, resource occupancy status is automatically updated, forming a complete closed-loop logic from anomaly assessment to scheduling arrangement, ensuring a reasonable processing order and balanced resource utilization. The coefficient "5" and the threshold "20" for triggering an emergency response in the formula are both example parameters. The coefficient "5" is used to adjust the unit and range of the priority value. The threshold "20" should be determined by expert experience based on the historical processing capacity of each region, the distribution of the number of problems, and the expected proportion of response levels. For example, if it is desired to classify about 20% of the problems into the emergency queue through analysis, the threshold can be deduced accordingly.

[0188] In a preferred embodiment, step S102, obtaining the optimized resource scheduling scheme includes:

[0189] Based on scheduling plan data and real-time information, a dynamic view of resource allocation is constructed to obtain comparison data between the total amount of resources and the current demand. If the total amount of resources is lower than the current demand, the range of resource gap is determined, and the resource allocation ratio of each task is sorted and adjusted in combination with the priority of the problem to obtain a preliminary resource reallocation plan.

[0190] Based on the resource reallocation scheme, the matching degree of demand coverage is analyzed. If the matching degree does not reach the preset standard (85 units), resources for low-priority tasks are compressed, the adjusted execution sequence is obtained, and the stability of resource scheduling is evaluated. Detailed resource allocation for each task is generated, resulting in an optimized resource scheduling scheme. The preset matching degree standard (e.g., 0.85) is a key performance indicator (KPI) reflecting the adequacy of resource allocation. Setting this standard requires a trade-off between resource costs and security risks and can be determined through expert experience. For example, if historical data shows that a matching degree of 0.85 guarantees a completion rate of over 95% for high-priority tasks, then this standard can be set. This standard can be adjusted appropriately for different regions based on resource scarcity.

[0191] Specifically, by extracting basic data from the scheduling plan and combining it with real-time information, a dynamic view of resource allocation is constructed to obtain the current resource distribution status. Simultaneously, a comparison of the total resource amount with the current demand is obtained. If the total resource amount is lower than the current demand, a resource shortage judgment is triggered, determining the range of resource gaps. Based on priority, the resource allocation ratios of each task are sorted and adjusted to obtain a preliminary resource reallocation plan. The degree of matching in demand coverage is analyzed. If the matching degree does not meet the preset standard, resources for low-priority tasks are further compressed to determine the optimized allocation structure. Combined with real-time information, the timing of resource scheduling is dynamically adjusted to obtain the adjusted execution sequence. For the adjusted execution sequence, a logistic regression model is used to evaluate the stability of resource scheduling, resulting in the final resource scheduling framework. Resource allocation details for each task are generated, and specific implementation paths are determined.

[0192] Specifically, the aforementioned real-time information includes:

[0193] 1. Resource Status Information

[0194] Resource availability: Real-time monitoring of the quantity and status of currently available resources. For example, the number of maintenance teams, equipment availability, and spare parts inventory.

[0195] Resource location information, including the current location of the resource, such as the real-time location of the maintenance team and the deployment location of the equipment. This helps to quickly dispatch resources to where they are needed.

[0196] Resource utilization rate refers to the current usage of resources, such as the workload of the maintenance team and the operating status of equipment. This helps determine whether resources are being fully utilized and whether there are idle or overloaded situations.

[0197] 2. User requirements information

[0198] The system monitors user electricity usage in real time, including parameters such as current power consumption, voltage, and current. This helps determine whether a user's electricity demand is normal and whether there are any abnormal power usage situations.

[0199] The system receives real-time reports of electrical hazards from users or automatically, including hazard type, severity, time of occurrence, and scope of impact. This helps to adjust resource allocation plans promptly and prioritize the handling of high-priority hazards.

[0200] User feedback, including real-time feedback such as evaluations of service quality and expectations regarding processing time, helps optimize resource allocation strategies and improve user satisfaction.

[0201] 3. Environmental Information

[0202] Weather conditions, real-time weather information, including temperature, humidity, wind speed, precipitation, etc. Severe weather may affect resource allocation and the efficiency of handling potential hazards, so preparations need to be made in advance;

[0203] Traffic conditions, real-time traffic information, including road congestion and traffic control measures. This helps optimize the maintenance team's route planning and reduce arrival time;

[0204] Geographic information, or Geographic Information System (GIS) data, includes topography, landforms, and regional divisions. This helps to more accurately locate potential hazards and resource locations, and optimize resource allocation paths.

[0205] 4. System Status Information

[0206] System load is monitored in real time to assess the overall load of the power system, including the operating status of the power grid and the load of substations. This helps in determining the system's stability and potential risks.

[0207] Equipment status: Real-time monitoring of equipment operation, including health status and fault alarms. This helps to identify potential equipment problems in a timely manner and perform maintenance proactively.

[0208] System logs provide real-time system information, including operation records and alarm messages. This helps to trace the root cause of problems and optimize the handling process.

[0209] 5. Forecast Information

[0210] Demand forecasting, based on historical and real-time data, uses time series analysis or machine learning models to predict future resource demand, which helps to adjust resource allocation plans in advance and avoid resource shortages or surpluses.

[0211] Hazard prediction, based on historical hazard data and real-time hazard reports, uses predictive models to predict the time and location of potential hazards, which helps to take preventive measures in advance and reduce the impact of hazards.

[0212] For example, suppose that in a certain urban area, the power system needs to allocate resources to address potential power outages. By extracting basic data from the dispatch plan and combining it with the following real-time information, a dynamic view of resource allocation is constructed, including: resource status information: currently, 10 maintenance teams are available, of which 3 are handling other tasks, 7 are on standby, and there is sufficient spare parts inventory in the warehouse, but some equipment is operating at 80% capacity, close to full load. User demand information: real-time monitoring shows that electricity consumption in a commercial area has suddenly increased, with voltage fluctuations reaching 5%, exceeding the normal range; user reports indicate that a low voltage problem has occurred in a residential area, affecting multiple users. Environmental information: current weather conditions are good, but traffic information shows that some roads are congested due to construction, which may affect the arrival time of maintenance teams; Geographic Information System (GIS) data shows that potential problems are concentrated in the southeastern part of the city, requiring priority resource allocation. Status information: the overall grid load is within the normal range, but the load of a certain substation is close to its limit and requires attention; real-time monitoring shows that a certain piece of equipment has a fault alarm and needs timely repair. Based on historical and real-time data, forecasts indicate that electricity demand in the commercial area will continue to increase in the next two hours, potentially requiring additional resource support. The forecast model also suggests that new potential hazards may emerge in the southeastern region within the next 24 hours, requiring advance preparation.

[0213] In multi-user power management scenarios, based on existing scheduling plans, real-time resource allocation data is obtained through automated processes to optimize resource scheduling schemes. First, the real-time status of maintenance resources in the current area is extracted from the resource management database. For example, if there are currently 10 maintenance units available, but the expected demand is 15, there is a shortage of 5 units. Next, through dynamic monitoring and analysis of resource occupancy, it is found that 3 maintenance units will complete their current tasks and release resources within 12 hours, while the remaining 2 shortages need to be addressed through allocation adjustments. Subsequently, the priority ranking database is accessed to extract the urgency data of the current 100 users. For example, user A's priority is 28.5, and user B's is 18.2. Considering the resource shortage, an optimization algorithm is initiated to allocate resources... Priority is given to users with a priority level higher than 25. User A is found to meet the criteria, while user B's processing will be delayed. Simultaneously, based on a resource release time prediction model, it is estimated that three additional repair units can be added after 12 hours. User B's processing time is automatically adjusted to 16 hours later, and the resource pool status is updated. To ensure a closed-loop logic, a backup resource mobilization mechanism is also implemented. If resources are still insufficient within the predicted time, support is automatically requested from neighboring areas, for example, requesting two backup units, with an estimated arrival time of within 24 hours. This forms a complete process from resource monitoring to dynamic adjustment, ensuring reasonable resource allocation and timely response. The threshold "25" in the example is a priority filter line under resource scarcity conditions. Its setting ensures that the number of filtered issues roughly matches the current and recent available resource quantity. This threshold can be determined by solving the equation: "Number of issues with priority ≥ threshold ≈ Number of issues that available resources can handle".

[0214] In a preferred embodiment, step S103, updating the scheduling plan includes:

[0215] By optimizing the scheme, the key issues are initially divided into resource allocations, matching data of resource scheduling and processing order is obtained, and the key issues are classified. If the response time of the classified issues exceeds the optimization standard, the resource scheduling of the excess part is adjusted, and the adjusted allocation strategy is determined.

[0216] Based on the adjusted allocation strategy, the correspondence between the processing order of key issues and resource scheduling is obtained, it is determined whether there is an uneven distribution of resources, and further resource allocation optimization is carried out according to the time limit. The stability of the processing order is evaluated to obtain the processing sequence of key issues.

[0217] A dataset of unprocessed issues is generated from a sequence of key issues. The issue data is automatically escalated using a preset threshold. The degree to which the escalation conditions are met is evaluated to obtain a preliminary set of escalated issues. Issue features are then acquired for classification to determine a subset of issues after classification.

[0218] Based on a subset of issues, determine whether the escalation conditions are met. If they are met, generate a list of escalation issues; otherwise, mark them as a set of issues to be processed.

[0219] The problem priority is extracted from the upgrade problem list, the priority sorting is adjusted according to the preset threshold, and the sequence is compared with real-time data to determine whether there are abnormal problems. The resource allocation ratio is then adjusted to obtain a dynamically adjusted resource scheduling scheme.

[0220] Specifically, through the optimized resource scheduling scheme, priority resources are allocated to key issues to determine whether the response time optimization requirements are met, thus obtaining the processing sequence for key issues, including:

[0221] By optimizing the scheme, a preliminary allocation of resources is made for key issues (potential problems that have a significant impact on the reliability, security and stability of power supply, determined by factors such as the severity of the problem, urgency and user feedback), and matching data of resource scheduling and processing order is obtained to determine the preliminary allocation structure.

[0222] Based on the initial allocation structure and the principle of prioritization, key issues are categorized to determine whether they meet the response time optimization criteria (e.g., a maximum of 6 hours), thus determining the categorized processing priorities. The response time optimization criteria (e.g., 6 hours) are core business metrics derived from Service Level Agreements (SLAs), internal operations and maintenance performance standards, or customer expectations, and are determined through expert experience. Different user categories or different levels of potential hazards may have different response time criteria. These criteria form the basis for setting relevant thresholds (such as the urgency level thresholds for key issues below).

[0223] If there are parts of the processing priority that exceed the optimization standard in response time after classification, the scheduling will be adjusted for the part that exceeds the standard. A preset threshold (e.g., the urgency level of critical issues is ≥7, and the urgency level of ordinary issues is <7) will be used to limit resource allocation and determine the adjusted allocation strategy.

[0224] By adjusting the allocation strategy, the correspondence between the processing order of key issues and resource scheduling is obtained, it is determined whether there is an uneven distribution of resources, and a balanced scheduling framework is obtained.

[0225] Uneven resource allocation refers to a significant difference in the amount of resources allocated to different tasks or problems during resource scheduling. This results in some tasks receiving too many resources, while others receive too few, failing to meet their normal processing needs. Such uneven resource allocation can affect task processing efficiency and the stability of the overall system. Specific manifestations include: differences in task processing time, where some tasks take too long to process due to insufficient resources, while others take less time; for example, task A is allocated fewer resources and takes 10 hours, while task B is allocated more resources and takes 2 hours. Differences in resource utilization, where some tasks have excessively high resource utilization, while others have excessively low utilization; for example, task A has a resource utilization rate of only 40%, while task B has a utilization rate as high as 90%. Mismatch between task priority and resource allocation, where high-priority tasks are allocated insufficient resources, while low-priority tasks are allocated excessive resources; for example, high-priority task A is allocated 10 units of resources, while low-priority task B is allocated 20 units. Differences in task completion rate, where some tasks have low completion rates due to insufficient resources, while others have high completion rates; for example, task A has a completion rate of 60%, while task B has a completion rate of 90%. Methods for determining uneven resource allocation include calculating the standard deviation of resource allocation (resource allocation for each task, average resource allocation); calculating a fairness index for resource allocation (such as the Gini coefficient), where a Gini coefficient closer to 0 indicates fairer resource allocation, and closer to 1 indicates more uneven allocation; calculating the difference in resource utilization, where the difference = max(utilization) - min(utilization), and calculating the difference between the highest and lowest resource utilization rates across all tasks. Further, resource allocation can be adjusted to achieve balance. This can be done by proportionally adjusting resource allocation, where the adjusted allocation = initial resource allocation × (1 - adjustment ratio); reallocating resources, allocating resources saved from low-priority tasks to high-priority tasks; and dynamically adjusting resource allocation based on real-time information to ensure fairness and efficiency. Finally, a balanced scheduling framework is generated, producing detailed resource allocation for each task to ensure that the resource allocation for each task meets its processing needs. The stability of the adjusted resource allocation scheme is evaluated using logistic regression models or other assessment methods.

[0226] Based on the balanced scheduling framework, further resource allocation optimization is performed to address time constraints. A logistic regression model is used to evaluate the stability of the processing order and determine the final execution sequence.

[0227] The time limit refers to the maximum processing time set for each task or problem during resource scheduling. This time limit is set based on the urgency of the task, the importance of the user, and the feasibility of actual operation. The specific value of the time limit needs to be determined according to the actual situation. Time limits can be set according to the urgency level; the higher the urgency level, the shorter the time limit. For example, urgency levels 1-3 have a time limit of 24 hours, urgency levels 4-6 have a time limit of 12 hours, and urgency levels 7-10 have a time limit of 6 hours. Time limits can also be set according to user importance; critical users (such as hospitals and data centers) have shorter time limits. For example, critical facilities (such as hospitals and data centers) have a time limit of 4 hours, commercial users have a time limit of 8 hours, and residential users have a time limit of 12 hours. Time limits can also be set according to the type and complexity of the task. Time limits for complex tasks can be appropriately relaxed; for example, voltage fluctuations have a time limit of 6 hours, current overloads have a time limit of 8 hours, and equipment aging have a time limit of 12 hours. The aforementioned time limits are a concretization of the aforementioned response time optimization standards, consistent with SLAs or internal procedures, and take into account practical factors such as traffic and staffing in different regions, and are ultimately determined through expert experience.

[0228] The stability of the processing order refers to the rationality and reliability of the task processing order during resource scheduling. It reflects whether the task processing order can maintain consistency and efficiency in a dynamically changing environment, avoiding frequent changes and task delays. The stability of the processing order is one of the key factors ensuring the effective execution of resource scheduling schemes. Key data and calculation methods for processing order stability include: Task delay rate, which is the ratio of the number of tasks that fail to be completed on time to the total number of tasks: Task delay rate = (Number of tasks not completed on time / Total number of tasks) × 100%. The lower the task delay rate, the higher the stability of the processing order. It also includes the volatility of task processing time, which refers to the degree of difference between the actual processing time and the expected processing time: Task processing time volatility = Standard deviation (Actual processing time - Expected processing time). The smaller the volatility, the higher the stability of the processing order. Finally, it includes the matching degree between task priority and actual processing order, which refers to the degree to which high-priority tasks are processed first: Matching degree = (Sum of actual processing orders of high-priority tasks / Total number of high-priority tasks) / (Total number of tasks / 2). The closer the matching degree is to 1, the higher the stability of the processing order. This also includes the fairness of resource allocation, which refers to whether resource allocation reasonably considers task priorities and needs. The fairness of resource allocation is calculated as the Gini coefficient (the resource allocation ratio for each task). The closer the Gini coefficient is to 0, the fairer the resource allocation and the higher the stability of the processing order. It also includes the task completion rate, which is the ratio of the number of tasks completed within a specified time to the total number of tasks. The task completion rate is calculated as (number of completed tasks / total number of tasks) × 100%. The higher the task completion rate, the higher the stability of the processing order. The stability of the processing order is evaluated using a logistic regression model.

[0229] If the response time of some key issues in the final execution sequence still does not meet the optimization standard, a second resource scheduling is performed in combination with the allocation strategy, real-time data is obtained for comparison, and an optimized execution arrangement is obtained.

[0230] By optimizing the execution schedule, the proportion of resource allocation is dynamically adjusted according to the requirements of problem classification and priority handling, and it is determined whether all key issues meet the response time standards to obtain the final solution.

[0231] For example, in the field of multi-user power management, based on an optimized resource scheduling scheme, priority resources are allocated to critical issues, and it is determined whether the response time optimization requirements are met, ultimately generating a processing sequence for critical issues. First, power fault data to be processed in the current area is extracted from the problem database. Assuming there are 50 fault points, 10 of which are marked as critical issues, involving users with power outages exceeding 6 hours; using a fault impact assessment algorithm, the severity score of each critical issue is calculated. For example, fault point C has a score as high as 35.7, and fault point D has a score of 22.3. The score is calculated based on a weighted average of the number of users and the duration of the outage. Next, combined with the resource scheduling scheme, the distribution of currently available maintenance resources is analyzed. Assuming there are 8 maintenance teams that can be deployed immediately, these are prioritized for critical issues with scores higher than 30. The calculation results show that fault point C is prioritized for processing, while fault point D enters the processing queue; simultaneously, through response... The time prediction model estimates the processing time for fault point C to be 4.5 hours, which is lower than the optimization requirement of 6 hours and meets the standard. For fault point D, it predicts that processing will only be possible after two teams are available, with an estimated delay of 8 hours. To form a closed-loop logic, a fault impact diffusion analysis was also performed. If delayed processing may lead to an amplified impact, for example, if three secondary fault points around fault point D might be affected, resource scheduling will be automatically adjusted, calling in one backup team, expected to arrive within 6 hours, ensuring a reasonable processing sequence. Finally, a processing sequence list is generated, with fault point C at the top and fault point D moved to second place, and the database status is updated, forming a complete process from problem identification to resource allocation to time optimization. The scoring threshold "30" in the example is used to filter out a subset of critical issues that require immediate processing. This threshold should be based on the current available resources (8 teams) and the average processing time of the issues to ensure that the number of filtered issues is within the range where resources can process them in a timely manner. This threshold can be dynamically calculated through simulation or optimization algorithms.

[0232] Specifically, from the critical issue processing sequence, data on unprocessed issues is obtained, and a preset threshold is used to determine whether an issue needs to be automatically escalated, resulting in a list of escalated issues, including:

[0233] Data on unprocessed issues were extracted from the sequence of key issues. Data filtering methods were used to determine the unprocessed dataset based on processing status (unprocessed, processed, etc.) and time constraints (e.g., an 8-hour limit).

[0234] If the non-empty dataset is not processed in time, the problem data will be automatically upgraded based on a preset threshold (upgrade threshold 25). A logistic regression model will be used to evaluate the degree of satisfaction of the upgrade conditions (for example, if the upgrade condition is that the risk score of the problem exceeds 25, then the degree of satisfaction can be the difference between the actual risk score and 25. If the degree of satisfaction is positive, it means that the problem meets the upgrade conditions; if it is negative, it does not meet the conditions). A preliminary set of upgraded problems will be obtained. The setting of the upgrade threshold (e.g., 25) determines the sensitivity of problem upgrade. This threshold can be set by experts based on the analysis of the consequences of delayed processing.

[0235] The problem characteristics (including but not limited to the risk level of the problem, the number of users or regions affected by the problem, the urgency level of the problem, the current processing progress of the problem, and the amount of resources required to process the problem) are obtained from the initial escalation problem set. Cluster analysis is used to classify the problem data (high-risk problems, medium-risk problems, and low-risk problems). Based on time constraints and processing status, a subset of problems after classification is determined.

[0236] For the categorized subset of issues, a judgment logic is used to check whether the escalation conditions are met (e.g., a risk score threshold of 25). If they are met, an escalation issue list is generated; otherwise, it is marked as a set of issues to be processed.

[0237] Extract problem priorities from the upgrade problem list, adjust the priority ranking according to a preset threshold (a certain indicator of a problem (such as a risk score threshold of 30) exceeding this value will increase the priority of the problem, and the threshold can be determined by expert experience), and use a sorting algorithm to rearrange the problem list to obtain an optimized upgrade sequence;

[0238] For the optimized upgrade sequence, real-time data is acquired for comparison. Data matching methods are used to determine whether there are any abnormal issues in the sequence (such as a high-risk issue being ranked after a low-risk issue, or an urgent issue taking longer than expected to be processed), and the final list of upgrade issues is determined.

[0239] Extract relevant data (other information about the issues, including the specific details of the issues, a list of affected users, and the resources already allocated to the issues) from the final upgrade issue list, and adjust the resource allocation ratio according to the judgment logic mentioned above to obtain a dynamically adjusted resource scheduling scheme.

[0240] For example, in the field of multi-user power management, data on unprocessed issues is extracted from the critical issue processing sequence to automatically determine whether an upgrade is needed, ultimately generating an upgrade issue list. First, fault points that have not been repaired within the specified time are filtered from the processing sequence database. Assuming there are currently 30 fault points in the sequence, and 5 of them have exceeded the preset 8-hour limit in processing time (e.g., fault point E has a processing delay of 10.2 hours, and fault point F has 9.5 hours), then a delay impact assessment algorithm is called to comprehensively calculate the degree of impact of the delay on users. Based on a weighted formula of delay duration and the number of affected users, the algorithm gives an impact score of 28.4 for fault point E and 25.6 for fault point F. Since the preset upgrade threshold is 25.0, both are marked as requiring upgrade processing. Simultaneously, the geographical location correlation of the fault points is further analyzed through regional impact assessment. The impact propagation model identified two low-priority fault points near fault point E. If not addressed promptly, these could trigger a chain reaction. Therefore, the model automatically elevated their priority to the highest level and generated an escalation report. Next, based on historical data analysis, it predicted the additional resources needed after the escalation. It estimated that fault point E would require an additional emergency response team, and fault point F would require additional equipment support. The predicted processing time could be reduced to within 4.8 hours. Finally, fault points meeting the escalation criteria were compiled into a list, including fault points E and F. Their status was automatically updated in the escalation issue database, and related notifications synchronized the information, ensuring smooth subsequent resource allocation and forming a complete closed-loop logic from delay identification to impact assessment to escalation processing.

[0241] Specifically, based on the list of upgrade issues, the scheduling mechanism is dynamically adjusted, and the upgrade issues are re-prioritized to obtain an updated scheduling plan, including:

[0242] The core data of the upgrade issues are obtained from the issue list. Each data item is classified and processed. The data is initially filtered using preset rules to determine the data set after the initial filtering.

[0243] The core data consists of key information describing the escalation issue, used for subsequent processing and decision-making. Core data typically includes: Issue ID (a unique identifier for each issue), Risk Score (the degree of risk, usually a numerical value, with higher scores indicating greater risk), Impact Scope (the number of users or regions affected by the issue), Urgency Level (the urgency level of the issue, usually a numerical value or category, such as high, medium, or low), Processing Status, Current Processing Progress (not started, in progress, completed), Resource Requirements (the amount of resources required to process the issue), and Estimated Processing Time (the time required to complete the process). Predefined rules are pre-defined filtering conditions used to select data from a large dataset that meets specific criteria. These rules may include: Risk Score Threshold (only issues with risk scores higher than a certain value are filtered); Impact Scope Value (only issues with an impact scope exceeding a certain value are filtered); Urgency Level (only issues with high or medium urgency are filtered); and Processing Status (only issues not started or in progress are filtered).

[0244] For the initially screened dataset, relevant information on priority levels is obtained, such as issue ID, risk score, and urgency. If the priority level is higher than a preset threshold (e.g., risk score threshold of 30 and urgency threshold of 7, set by expert experience based on historical data for each region), it is marked as a high-priority issue, thus obtaining a subset of high-priority issues.

[0245] Based on the high-priority problem subset, data extraction methods are used to obtain the correlation information of problem ranking (problem priority, a comprehensive priority calculated based on risk score, urgency, etc.; problem dependencies, some problems may depend on the completion of other problems; resource allocation, the amount of resources currently allocated to each problem; estimated processing time, the time required to complete the problem processing). The problem ranking is then optimized and adjusted using a logistic regression model (evaluating the rationality of the problem ranking and adjusting it according to the prediction results, adjusting the processing order of the problems to ensure that high-priority problems are processed first), and the optimized ranking result is determined.

[0246] Based on the optimized sorting results, obtain the necessary information for sequence adjustment (including the current problem processing order, the problem priority predicted by the logistic regression model, the amount of resources currently allocated to each problem, and the processing time limit for each problem). If the sequence adjustment conforms to preset rules (used to determine whether the problem processing order needs to be adjusted, including priority differences (if the difference between the model-predicted priority and the current processing order is too high, adjustment is needed); resource allocation rationality (if the resource allocation for a certain problem is insufficient while the resources for other problems are excessive, adjustment is needed); and time limit (if the estimated processing time for a certain problem exceeds the time limit, adjustment is needed), then the scheduling plan is dynamically updated to obtain a temporary scheduling scheme.

[0247] Key points for plan updates are extracted from the temporary scheduling scheme (used to evaluate the rationality and effectiveness of the scheduling scheme, including the processing order of high-priority issues to ensure that high-priority issues are processed first; the rationality of resource allocation to ensure that resource allocation can meet the processing needs of the issues; and the adherence to time limits to ensure that the processing time of each issue does not exceed the time limit). Comparative analysis tools are used to detect the differences before and after the plan update to determine whether there are any deviation issues, and the deviation detection results are obtained (priority deviation, the difference between the priority predicted by the model and the current processing order; resource allocation deviation, the difference between the actual resource allocation and the ideal resource allocation; time deviation, the difference between the actual processing time and the expected processing time).

[0248] Based on the deviation detection results, the deviation issues are reassessed. If the deviation exceeds the preset range (e.g., priority deviation ±5%, resource allocation deviation ±10%, time deviation ±2 hours), the scheduling plan is adjusted a second time to determine the final scheduling scheme. The preset deviation ranges (±5, ±10%, ±2 hours) define acceptable scheduling errors. These ranges should be set based on the flexibility and accuracy requirements of business operations through expert experience. For example, a time deviation of ±2 hours might mean allowing reasonable delays due to uncertainties such as traffic. Setting the range too strictly may lead to frequent and unnecessary adjustments, while setting it too loosely may mask problems. Gradual optimization through operational experience is necessary.

[0249] The latest status of the problem list is obtained from the final scheduling plan, and the processing progress of the problem list is tracked using data logging tools to obtain a complete processing record.

[0250] For example, in the field of multi-user power management, the system automatically adjusts the scheduling mechanism based on the generated list of upgrade issues, re-prioritizing the upgrade issues and ultimately forming an updated scheduling plan. First, it extracts data on fault points currently marked as high priority from the upgrade issue database. Assuming there are three upgrade fault points, with fault point A having an urgency score of 32.5, fault point B 29.8, and fault point C 27.3, based on the priority ranking algorithm, the scores are combined with the scope of impact to calculate the comprehensive weight. The formula is: Weight = Urgency Score × 0.6 + Number of Affected Users × 0.4. Therefore, fault point A has a weight of 38.2 (affecting 140,000 users), fault point B 34.6 (affecting 120,000 users), and fault point C 31.9 (affecting 110,000 users). Subsequently, through a resource allocation optimization model, these three fault points are compared and analyzed with other tasks in the existing scheduling plan, and the task queue is automatically adjusted, prioritizing fault point A. The task is prioritized and processed first, with fault points B and C following in sequence. Simultaneously, a geolocation optimization algorithm is used to analyze the distance between fault point A and another ordinary task point D, finding it to be only 2.3 kilometers. Task D is automatically prioritized to reduce resource allocation costs. Next, dynamic scheduling is updated. Based on current resource pool data, processing fault point A requires two maintenance units, with an estimated completion time of 3.5 hours. Fault points B and C are each allocated one unit, with estimated completion times of 4.2 hours and 4.7 hours respectively, ensuring resource utilization exceeds 85%. Finally, an updated scheduling plan is generated, and the adjusted task sequence and resource allocation scheme are stored in the scheduling database and synchronized to the execution system via an interface, forming a complete logical closed loop from priority reordering to resource optimization.

[0251] Resource allocation optimization models are used to optimize resource allocation under given resource constraints to achieve specific objectives, such as minimizing processing time, maximizing resource utilization, or ensuring high-priority tasks are completed first. These models typically combine multiple optimization techniques, such as linear programming, integer programming, and heuristic algorithms, to solve complex resource allocation problems. Their objective functions can be minimizing total processing time, maximizing resource utilization, minimizing the latency of high-priority tasks, etc., and the constraints can be total resource limits, task processing time limits, and task priority limits. Assume we have the following fault points and task data: Fault point A has an urgency score of 32.5, affecting 140,000 users, with an estimated processing time of 3.5 hours; Fault point B has an urgency score of 29.8, affecting 120,000 users, with an estimated processing time of 4.2 hours; Fault point C has an urgency score of 27.3, affecting 110,000 users, with an estimated processing time of 4.7 hours; and ordinary task D has an estimated processing time of 2 hours and is 2.3 kilometers away from fault point A. The comprehensive weight is calculated using a resource allocation optimization model. The weight value is calculated as: urgency score × 0.6 + number of affected users × 0.4. The comprehensive weight for each fault point is calculated as follows: Fault point A has a weight of 32.5 × 0.6 + 14 × 0.4 = 38.2; Fault point B has a weight of 29.8 × 0.6 + 12 × 0.4 = 34.6; and Fault point C has a weight of 27.3 × 0.6 + 11 × 0.4 = 31.9. Comparing fault points A, B, and C with other tasks in the existing scheduling plan, fault point A has the highest weight and is prioritized, followed by fault point B, and then fault point C has the lowest weight and is processed last. The task queue is automatically adjusted based on the comprehensive weight and geographical location optimization algorithm. Fault point A is placed at the top of the priority processing sequence, followed by fault point B. Fault point C is processed last. Ordinary task D, being closer to fault point A, is processed earlier to reduce resource allocation costs. Simultaneously, dynamic scheduling updates are performed. Based on the current resource pool data, resource allocation and estimated completion times are calculated. Fault point A is allocated 2 maintenance units with an estimated completion time of 3.5 hours; fault point B is allocated 1 maintenance unit with an estimated completion time of 4.2 hours; and fault point C is allocated 1 maintenance unit with an estimated completion time of 4.7 hours. An updated scheduling plan table is generated, and the adjusted task sequence and resource allocation scheme are stored in the scheduling database. The coefficients (0.6 and 0.4) in the weight calculation formula determine the relative importance of urgency and impact scope. These coefficients need to be determined through business analysis or optimization algorithms (such as reinforcement learning based on historical scheduling feedback).

[0252] In a preferred embodiment, step S104 involves acquiring real-time feedback data in a multi-user environment, determining whether resource utilization has reached a preset target, and obtaining the final resource allocation result, including:

[0253] Relevant data of the multi-user environment is obtained from the scheduling plan. The real-time feedback content is classified and stored to obtain the classified data set. Relevant information on resource utilization is obtained. If the resource utilization index is lower than the preset threshold, the abnormal data is marked and the abnormal data subset is determined.

[0254] Based on a subset of abnormal data, relevant information from environmental monitoring is obtained, and the resource allocation status in a multi-user environment is analyzed to determine whether there is uneven allocation and obtain the allocation status analysis results.

[0255] Based on the allocation status analysis results, a secondary check is performed using relevant data on utilization rate. If the check results show that resource utilization deviates from the preset threshold (preset resource utilization target value of 82%), the scheduling plan is partially adjusted to determine a temporary adjustment plan. The preset resource utilization target value (e.g., 82%) is a key efficiency indicator. This target value can be set based on statistical analysis of historical resource utilization data, while also allowing for a certain margin to cope with sudden demand. For example, if the analysis shows that the utilization rate fluctuates between 75% and 80% for a long period, the target can be set at 82% to drive efficiency improvement while retaining 18% flexibility. Different regions can set different target values ​​based on their resource redundancy and load fluctuations.

[0256] Based on the temporary adjustment plan, key information regarding the plan adjustment is obtained (parts of the temporary adjustment plan that require special attention, used to evaluate the rationality and effectiveness of the scheduling plan; key information typically includes task processing order (i.e., the adjusted task processing order), resource allocation (i.e., the amount of resources allocated to each task), estimated processing time (i.e., the estimated processing time for each task), task priority (i.e., the priority of each task), and task dependencies (i.e., some tasks may depend on the completion of other tasks). Difference detection is performed on the resource allocation data before and after the adjustment, and the difference detection results are obtained. Finally, the final state of resource allocation (referring to the final result of the adjusted resource allocation plan in actual execution, including actual...) is determined. The resource allocation (the actual amount of resources allocated to each task), actual processing time (the actual time each task takes to complete), task completion status (whether each task is completed on time), and resource utilization rate (the overall utilization rate of resources) are verified. If the verification results show that the allocation plan deviates from the preset thresholds (resource utilization rate threshold 85%, task completion rate threshold 90%, processing time deviation threshold ±2 hours), the feedback data is reorganized to determine the final allocation plan. The preset thresholds (85%, 90%, ±2 hours) are standards used to evaluate whether the adjusted plan is "qualified" and can be determined by expert experience based on the historical task scheduling situation of each region.

[0257] Based on the final allocation scheme, the resource utilization status in the multi-user environment is continuously tracked to obtain status records.

[0258] Specifically, this embodiment can use data acquisition tools to classify and store feedback data (which can be classified by task, time, or resource type) to obtain a classified dataset. For the classified dataset, relevant information on resource utilization is obtained (data describing resource usage, used to evaluate the rationality and efficiency of resource allocation, including resource utilization rate (the ratio of actual resources used to total resources), resource allocation volatility (the degree of change in resource allocation ratios between different tasks), task completion rate (the ratio of the number of tasks completed within a specified time to the total number of tasks), and the fairness of resource allocation (whether resource allocation reasonably considers task priorities and needs). If the resource utilization indicators are lower than preset thresholds (resource utilization rate threshold 85%, task completion rate threshold 90%), abnormal data is labeled using data filtering tools to determine an abnormal data subset. Furthermore, related environmental monitoring information is obtained (environmental data related to resource allocation and task processing, used to evaluate the rationality and efficiency of resource allocation, including user electricity usage status). The system includes real-time electricity consumption data, hazard reports (real-time reports of potential electricity hazards), user feedback (user evaluations of service quality), and environmental information such as weather conditions and traffic conditions that may affect resource allocation and task processing. A logistic regression model can be used to analyze resource allocation in a multi-user environment to determine if there is uneven distribution. The analysis results include: fairness of resource allocation (whether resource allocation reasonably considers task priorities and needs), resource utilization rate (the ratio of actual resources used to total resources), task completion rate (the ratio of tasks completed within a specified time to the total number of tasks), volatility of resource allocation (the degree of change in resource allocation ratios across different tasks), and abnormal data (data points that do not conform to normal patterns or expectations, which may require further investigation and processing). Furthermore, a comparison tool can be used to detect differences in resource allocation data before and after adjustments, and a data recording tool can be used to continuously track the resource utilization status in the multi-user environment to obtain a complete status record.

[0259] For example, real-time feedback data is obtained through updated scheduling plans to determine whether resource utilization has reached the preset target, ultimately forming the resource allocation result. First, real-time feedback data is extracted from the scheduling execution database. Assuming there is current power load data for three key areas: area X has a peak load of 180.5 MW, area Y has 162.3 MW, and area Z has 145.7 MW, the resource demand index for each area is calculated based on the load balancing algorithm. The formula is: demand index = peak load × 0.7 + historical volatility × 0.3, where the historical volatility of area X is 12.4%, area Y is 10.8%, and area Z is 9.5%. The calculated demand indexes are 130.1 for area X, 116.9 for area Y, and 104.8 for area Z. Next, these indices are compared with the preset resource utilization target value of 82%. Analysis revealed that the actual utilization rate of region X was 78.6%, lower than the target value, while that of region Y was 83.2% and region Z was 81.5%. This automatically triggered a resource reallocation model. Analysis of the current available capacity of the resource pool, at 320 MW, calculated that 15.3 MW of reserve capacity needed to be added to region X to improve utilization, while 4.7 MW was reduced from region Y to avoid resource surplus. Subsequently, considering the grid connection status between regions, analysis revealed that the transmission loss rate between region X and region Y was only 1.2%. Resources were prioritized for allocation from Y to X to ensure the efficiency of the adjustment process. Finally, the adjusted resource allocation data was updated to the central control database, generating the final allocation result. The utilization rate of region X increased to 82.3%, region Y to 81.9% after adjustment, and region Z remained unchanged, forming a complete process from feedback data collection to resource optimization. The weights (0.7, 0.3) in the demand index calculation formula reflect the trade-off between current load and stability. These weights can be learned through regression analysis of historical optimal dispatch decisions, or set by dispatch experts based on experience. Different power grids have different stability characteristics, and the weights should be adjusted accordingly.

[0260] The resource reallocation model dynamically adjusts resource allocation based on real-time feedback data and preset targets to achieve optimal resource utilization efficiency. It combines various optimization techniques, such as linear programming, integer programming, and heuristic algorithms, to solve complex resource allocation problems. The objective function can be set to maximize resource utilization, minimize resource waste, or minimize the imbalance in resource allocation. Constraints can include: total resource allocation limit (total allocation ≤ total resource quantity); regional resource demand limit (resource allocation ≥ demand index); and transmission loss limit (transmission loss rate ≤ preset transmission loss rate). Assume we have the following regional electricity load data and resource demand indices: Region X, peak load 180.5 MW, historical volatility 12.4%, demand index 130.1, actual utilization 78.6%; Region Y, peak load 162.3 MW, historical volatility 10.8%, demand index 116.9, actual utilization 83.2%; Region Z, peak load 145.7 MW, historical volatility 9.5%, demand index 104.8, actual utilization 81.5%. Calculate the demand index: Demand Index = Peak Load × 0.7 + Historical Volatility × 0.3. Calculate the demand index for each region: Region X demand index = 180.5 × 0.7 + 12.4 × 0.3 = 130.1, Region Y demand index = 162.3 × 0.7 + 10.8 × 0.3 = 116.9, Region Z demand index = 145.7 × 0.7 + 9.5 × 0.3 = 104.8. Compare the demand indices of each region with the preset resource utilization target of 82%. Region X's actual utilization rate is 78.6% < the target value of 82%, Region Y's actual utilization rate is 83.2% > the target value of 82%, and Region Z's actual utilization rate is 81.5% < the target value of 82%, triggering the resource reallocation model. The current available capacity of the resource pool is 320 MW. Calculate resource allocation: Region X needs to add 15.3 MW of reserve capacity to improve utilization, and Region Y needs to reduce 4.7 MW to avoid resource surplus. The transmission loss rate is 1.2% between regions X and Y, so resources are preferentially allocated from Y to X. Dynamic scheduling updates are performed, and the adjusted resource allocation is as follows: Region X is allocated 130.1 + 15.3 = 145.4 MW, Region Y is allocated 116.9 - 4.7 = 112.2 MW, and Region Z is allocated 104.8 MW (unchanged). The updated utilization rates are 82.3% for Region X, 81.9% for Region Y, and 81.5% for Region Z. The final allocation result is generated, and the adjusted resource allocation data is updated to the central control database. The final allocation result is: Region X 145.4 MW, utilization rate 82.3%; Region Y 112.2 MW, utilization rate 81.9%; Region Z 104.8 MW, utilization rate 81.5%.The "preset transmission loss rate" limit (e.g., ≤1.2%) in the resource reallocation model is a technical constraint. Excessive loss means uneconomical scheduling. This value is determined by experts based on the technical parameters of the power grid lines and the cost of electricity. Different lines have different upper limits for loss rates.

[0261] In a preferred embodiment, based on the final resource allocation result, the timeliness of problem handling is analyzed to obtain a dynamic management scheme adapted to a multi-user environment, including:

[0262] Acquire real-time feedback data to monitor the resource allocation status in a multi-user environment, obtain a categorized dataset, extract key indicators of resource utilization efficiency (resource utilization rate, task completion rate, and resource allocation volatility), analyze indicator fluctuations, and determine efficiency fluctuation characteristics.

[0263] If the efficiency fluctuation characteristics indicate uneven resource allocation, abnormal data points are marked to obtain an abnormal data subset. User behavior correlation information is obtained (user electricity consumption pattern, i.e., peak and off-peak electricity consumption periods, user feedback, i.e., user evaluation of service quality, user type, i.e., whether the user is a residential / commercial / industrial user, and user equipment status, i.e. whether the user's equipment is operating normally). The resource demand changes are predicted through a time series analysis model to obtain the demand prediction results.

[0264] Based on the demand forecast results, the scheduling plan parameters are adjusted to generate a dynamic adjustment plan. The resource allocation data before and after the adjustment are then compared to obtain the difference detection results. If the difference detection results show that the resource allocation deviates from the preset threshold, the real-time feedback data is reorganized to determine the final dynamic management plan.

[0265] The efficiency index threshold (e.g., 0.5) for judging uneven resource allocation needs to be determined based on the distribution of efficiency indices under the "balanced allocation" state in historical data. For example, the median or a certain quantile of the efficiency index for each historical period can be calculated as a benchmark threshold. A resource allocation difference threshold of ±10% means that a change of up to 10% in the allocation amount before and after adjustment is allowed; exceeding this may mean that the adjustment is excessive or insufficient, and the plan needs to be reorganized.

[0266] Specifically, real-time feedback data is acquired, and data acquisition tools are used to monitor the resource allocation status in a multi-user environment. Simultaneously, a logistic regression model is used to analyze indicator fluctuations and determine efficiency fluctuation characteristics (resource utilization efficiency index and fluctuation amplitude). If the efficiency fluctuation characteristics indicate uneven resource allocation (e.g., efficiency index less than 0.5), outlier data points are marked using data filtering tools to obtain an outlier subset. Furthermore, time series analysis models (e.g., autoregressive integral moving average model, seasonal decomposition model, etc.) are used to predict changes in resource demand. Parameter optimization tools are used to generate dynamic adjustment schemes and determine the optimized allocation parameters (resource allocation ratio, i.e., the proportion of each task or region). The allocation of resources, processing priority (i.e., task processing order), time window (i.e., task processing time range), and total available resources are all considered. A comparative analysis tool is used to detect differences in resource allocation data before and after the adjustment, yielding the results (resource allocation difference, i.e., the difference in resource allocation before and after the adjustment; efficiency change, i.e., the change in resource utilization efficiency before and after the adjustment; task completion rate change, i.e., the change in task completion rate before and after the adjustment). If the difference detection results show that resource allocation deviates from preset thresholds (resource utilization rate threshold 85%, task completion rate threshold 90%, resource allocation difference threshold ±10%), then the real-time feedback data is reorganized using a data integration tool to determine the final dynamic management plan.

[0267] For example, based on the final resource allocation results, intelligent management capabilities are used to analyze the timeliness of problem handling and formulate an adaptive dynamic management solution. First, response time data for each region is extracted from the historical operation database. Assuming there are three regions with average response times of 3.2 hours for region A, 4.1 hours for region B, and 2.8 hours for region C, a timeliness optimization algorithm is used to calculate the processing efficiency index for each region. The formula is: Efficiency Index = Reciprocal of Response Time × 0.6 + Historical Stability × 0.4. The historical stability of region A is 85.3%, region B is 79.6%, and region C is 88.2%. Therefore, the calculated efficiency indexes are 0.55 for region A, 0.42 for region B, and 0.61 for region C. Next, the efficiency indices are compared with a preset timeliness target value of 0.5. It is found that region B is below the target value and requires optimization of the processing flow, thus automatically initiating... The dynamic timeliness improvement model analyzes the current task queue backlog in region B, which is 12.5 task units. It calculates that 2.3 virtual processing nodes need to be added to share the load. Subsequently, considering the task coordination status between regions, it analyzes that the node load rate in region C is only 62.7%, allowing for the temporary allocation of 1.8 node resources to region B with an estimated transmission delay of 0.3 hours, ensuring that the optimization process does not affect overall operation. Finally, the adjusted node allocation data (resource allocation ratio, processing priority, time window, total resources, etc.) is synchronized to the task scheduling center. The response time in region B is shortened to 3.4 hours, and the efficiency index is improved to 0.51. The operating status of regions A and C remains stable, forming a complete process from timeliness analysis to dynamic management. The timeliness target value (0.5) and the weights (0.6, 0.4) in the efficiency index calculation formula are parameters that need to be calibrated. These values ​​can be set with reference to industry benchmarks or internal historical best levels. The weights reflect the relative importance of response time and stability, and can be determined by analyzing their contribution to user satisfaction or system reliability.

[0268] In a preferred embodiment, continuous monitoring of changes in user importance and risk severity includes:

[0269] Real-time acquisition of data on user importance and risk severity, extraction of key feature values, determination of preliminary severity assessment results, classification of user importance and risk severity, and generation of a priority ranking list after classification.

[0270] Preliminary scheduling is performed based on the tiered priority list to generate an initial resource scheduling plan. Fluctuations in user importance and hazard severity are continuously tracked to determine if there are any deviations (importance deviation, hazard severity deviation). If the deviation exceeds preset thresholds (importance threshold 10%, hazard severity threshold 15%), a dynamic plan adjustment mechanism is triggered to obtain the adjusted scheduling parameters. The preset deviation thresholds (10%, 15%) define the monitoring sensitivity and are set based on historical data analysis of the actual importance and hazard severity of each area, and through expert experience.

[0271] The dynamic scheme adjustment mechanism refers to an automatic adjustment mechanism that is activated when the deviation exceeds the preset threshold. This mechanism optimizes resource allocation schemes, including: reassessment (re-evaluating user importance and the severity of potential risks); priority adjustment (adjusting task processing priorities based on the reassessment results); resource reallocation (re-allocating resources according to the new priorities to ensure reasonable and efficient resource allocation); and updating the scheduling scheme (updating the adjusted resource allocation and priorities to the scheduling scheme and synchronizing it with the task scheduling center to ensure real-time execution).

[0272] The resource scheduling scheme is iteratively updated multiple times based on the adjusted scheduling parameters to determine the optimized dynamic scheme content and make real-time corrections. If new deviations are found, the loop execution process is restarted to obtain the final scheme update result.

[0273] Specifically, based on the dynamic management plan, changes in user importance and risk severity are continuously monitored, and priority ranking and resource scheduling are executed cyclically to obtain a continuously optimized management process, as follows:

[0274] Data on user importance and hazard severity are collected in real time using information collection tools. Key feature values ​​(user type, user scale, hazard type, hazard impact range, hazard occurrence frequency, hazard duration, etc.) are extracted from these data to determine preliminary severity assessment results (a comprehensive assessment result of user importance and hazard severity). Data classification tools are then used to classify user importance and hazard severity to obtain a priority ranking list.

[0275] For the priority list after classification, a resource allocation tool is used for initial scheduling to generate an initial resource scheduling plan. Change analysis is used to continuously track the fluctuations in user importance and the severity of potential risks.

[0276] For the adjusted scheduling parameters, a loop execution tool is used to iterate and update the resource scheduling scheme multiple times to determine the optimized dynamic scheme content. The management process is corrected in real time through a data integration tool to generate an updated management process framework. The results of subsequent change analysis are monitored by continuous tracking. If new deviations are found, the loop execution process is re-entered to obtain the final updated scheme result.

[0277] For example, by continuously monitoring changes in user importance and hazard severity, a continuously optimized management process is formed by iteratively executing priority ranking and resource scheduling. First, user electricity demand data is obtained from a real-time database. Assume there are three user groups: user group X has a peak electricity demand of 1200 kW, user group Y has 900 kW, and user group Z has 1500 kW. The scheduling priority of each group is calculated based on a priority ranking algorithm, using the formula: Priority Index = Peak Electricity Demand × 0.7 + Hazard Severity × 0.3. The hazard severity is determined through historical fault data analysis; user group X has a hazard severity of 75.4%, user group Y has 82.1%, and user group Z has 68.7%. The calculated priority index is 860.2 for user group X, 654.3 for user group Y, and 1055.1 for user group Z. The priority index is then compared with a threshold of 800... Comparison revealed that user group Y was below the threshold, necessitating priority adjustment of resource allocation. Subsequent analysis of the power supply line load rates for each group showed that user group X had a load rate of 85.6%, user group Y 92.3%, and user group Z 78.4%. Using a load balancing algorithm, it was calculated that 300 kW of power capacity needed to be borrowed from user group Z to user group Y, with an estimated transmission loss of 15 kW. After adjustment, the scheduling strategy was updated in real-time, reducing the load rate of user group Y to 87.5% and improving power supply stability to 90.2%. To ensure continuous optimization, the power demand fluctuations and potential risks of each group were monitored every 30 minutes, dynamically adjusting the priority index and resource allocation ratio, ultimately forming a closed-loop management process from user demand monitoring to resource scheduling. The priority index threshold (800) was used to identify user groups requiring immediate attention; this threshold could be set based on total resources and the number of groups requiring service. The cyclical monitoring period (30 minutes) balanced response speed with power safety overhead; the period length should be determined based on the speed of changes in user demand and potential risks, and the optimal value could be selected through experimental data.

[0278] It should be noted that this invention analyzes the importance of users and the severity of potential power hazards, calculates problem priorities, and formulates preliminary scheduling plans. Combined with real-time resource allocation data, this invention optimizes resource scheduling schemes, prioritizing resource allocation for critical issues. For issues not addressed in a timely manner, it determines whether automatic escalation is necessary and re-prioritizes the escalated issues. By acquiring real-time feedback data, this invention continuously optimizes resource allocation and employs intelligent management capabilities to analyze the timeliness of problem handling. Furthermore, this invention continuously monitors changes in user importance and the severity of potential hazards, cyclically executing priority ranking and resource scheduling to achieve dynamic management in a multi-user environment. This method can effectively improve resource utilization, optimize problem handling timeliness, and realize intelligent dynamic management of potential power system hazards.

[0279] The above is an illustrative scheme of a graded early warning method for electricity safety hazards targeting different customers, as described in this embodiment. It should be noted that the technical solution of this graded early warning system for electricity safety hazards targeting different customers belongs to the same concept as the technical solution of the aforementioned graded early warning method for electricity safety hazards targeting different customers. Details not described in detail in this embodiment can be found in the description of the aforementioned graded early warning method for electricity safety hazards targeting different customers.

[0280] This embodiment also provides a computer device suitable for graded early warning of electrical safety hazards for different customers, including:

[0281] The system includes a memory and a processor. The memory stores computer-executable instructions, and the processor executes these instructions to implement the graded early warning method for electricity safety hazards for different customers, as proposed in the above embodiments.

[0282] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements a graded early warning method for electricity safety hazards for different customers as proposed in the above embodiment.

[0283] The storage medium proposed in this embodiment and the method for implementing graded early warning of electrical safety hazards for different customers proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0284] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0285] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for graded early warning of electricity safety hazards for different customers, characterized in that, include: User data and electricity hazard information are obtained from a multi-user environment. By analyzing the importance of users and the severity of hazards, the problem priority is calculated and a preliminary scheduling plan is obtained. Based on the scheduling plan, real-time resource allocation data is obtained. If resource allocation is insufficient, resource allocation is adjusted according to the problem priority ranking to obtain an optimized resource scheduling scheme. Based on the optimized resource scheduling scheme, priority processing resources are allocated to key issues to determine whether the response time optimization requirements are met, the processing sequence of key issues is obtained, and data on issues that have not been processed in a timely manner is acquired. Issue escalation judgment is then performed to obtain escalated issue data, so as to dynamically adjust the scheduling mechanism and update the scheduling plan. Based on the updated scheduling plan, real-time feedback data in a multi-user environment is obtained, the timeliness of problem handling is analyzed, a dynamic management solution adapted to the multi-user environment is obtained, and changes in user importance and the severity of potential risks are continuously monitored to update and optimize the scheduling plan. Prioritization of computational problems includes: User data and electricity hazard information are obtained from a multi-user environment to construct an initial dataset. User data is then classified into user groups using preset classification rules. The severity of each user in the user group is analyzed, and combined with the information on potential electrical hazards, the hazard assessment result is determined. If the hazard assessment result exceeds a preset threshold, a severity score is obtained. Based on the severity score and the user's level of importance, the priority of the issues is calculated, the handling plan for the issues is generated, and potential risk points are identified based on the distribution characteristics of high-priority issues. The preliminary scheduling plan includes: By extracting relevant information from user data sources and combining it with data on potential electricity hazards, a basic analysis dataset is constructed to obtain a preliminary set of problem categories. A weighted calculation method is then used to quantify the priority of the problems and determine the order of processing. Based on the processing order number and the urgency level data, if the urgency level exceeds the preset threshold, the resource allocation ratio is adjusted first to obtain the adjusted scheduling framework. Combined with the time window, the processing sequence of each problem is optimized to obtain the optimized timing arrangement. Based on the optimized timing arrangement and combined with the regional distribution map, the spatial clustering characteristics of potential hazards are analyzed to determine the range of high-risk areas. Through predictive processing, the distribution characteristics of potential hazards are obtained. Combined with data in the historical record database, the final scheduling execution plan is determined. The optimized resource scheduling scheme includes: Based on the data and real-time information of the scheduling plan, a dynamic view of resource allocation is constructed, and the comparison data between the total amount of resources and the current demand is obtained. If the total amount of resources is lower than the current demand, the range of resource gap is determined, and the resource allocation ratio of each task is sorted and adjusted in combination with the priority of the problem to obtain a preliminary resource reallocation plan. Based on the resource reallocation scheme, the matching degree of demand coverage is analyzed. If the matching degree does not reach the preset standard, the resources of low problem priority tasks are compressed to obtain the adjusted execution sequence. The stability of resource scheduling is evaluated, and the resource allocation details of each task are generated to obtain the optimized resource scheduling scheme. The updated scheduling plan includes: By optimizing the scheme, the key issues are initially divided into resource allocations, matching data of resource scheduling and processing order is obtained, and the key issues are classified. If the response time of the classified issues exceeds the optimization standard, the resource scheduling of the excess part is adjusted, and the adjusted allocation strategy is determined. Based on the adjusted allocation strategy, the correspondence between the processing order of key issues and resource scheduling is obtained, it is determined whether there is an uneven distribution of resources, and further resource allocation optimization is carried out according to the time limit. The stability of the processing order is evaluated to obtain the processing sequence of key issues. A dataset of unprocessed issues is generated from a sequence of key issues. The data is automatically escalated based on a preset threshold. The degree to which the escalation conditions are met is evaluated to obtain a preliminary set of escalated issues. Issue features are then acquired for classification to determine a subset of issues after classification. Based on the aforementioned subset of problems, determine whether the escalation conditions are met. If they are met, generate an escalation problem list; otherwise, mark it as a set of problems to be processed. Extract problem priorities from the upgrade problem list, adjust the priority ranking according to preset thresholds, compare with real-time data to determine if there are any abnormal problems in the sequence, adjust the resource allocation ratio, and obtain a dynamically adjusted resource scheduling scheme.

2. The method for graded early warning of electricity safety hazards for different customers as described in claim 1, characterized in that, Acquire real-time feedback data in a multi-user environment to determine whether resource utilization has reached the preset target, and obtain the final resource allocation result, including: Relevant data of the multi-user environment is obtained from the scheduling plan. The real-time feedback content is classified and stored to obtain the classified data set. Relevant information on resource utilization is obtained. If the resource utilization index is lower than the preset threshold, the abnormal data is marked and the abnormal data subset is determined. Based on a subset of abnormal data, relevant information from environmental monitoring is obtained, and the resource allocation status in a multi-user environment is analyzed to determine whether there is uneven allocation and obtain the allocation status analysis results. Based on the allocation status analysis results, a second check is performed using relevant data to determine the utilization rate. If the check results show that the resource utilization deviates from the preset threshold, the scheduling plan is partially adjusted to determine a temporary adjustment plan. Based on the temporary adjustment plan, key information on the plan adjustment is obtained, and the resource allocation data before and after the adjustment is checked to obtain the difference detection results. The final state of resource allocation is verified. If the verification results show that the allocation plan deviates from the preset threshold, the feedback data is reorganized to determine the final allocation plan. Based on the final allocation scheme, the resource utilization status in the multi-user environment is continuously tracked to obtain status records.

3. The method for graded early warning of electricity safety hazards for different customers as described in claim 2, characterized in that, Based on the final resource allocation results, the timeliness of problem handling is analyzed, and a dynamic management solution adapted to multi-user environments is obtained, including: Acquire real-time feedback data to monitor the resource allocation status in a multi-user environment, obtain a categorized dataset, extract key indicators of resource utilization efficiency, analyze indicator fluctuations, and determine efficiency fluctuation characteristics. If the efficiency fluctuation characteristics indicate uneven resource allocation, abnormal data points are marked to obtain an abnormal data subset, user behavior correlation information is obtained, and resource demand changes are predicted through a time series analysis model to obtain demand prediction results. Based on the demand forecast results, the scheduling plan parameters are adjusted to generate a dynamic adjustment plan. The resource allocation data before and after the adjustment are then compared to obtain the difference detection results. If the difference detection results show that the resource allocation deviates from the preset threshold, the real-time feedback data is reorganized to determine the final dynamic management plan.

4. The method for graded early warning of electricity safety hazards for different customers as described in claim 3, characterized in that, Continuous monitoring of changes in user importance and risk severity includes: The system acquires data on user importance and risk severity in real time, extracts key feature values, determines preliminary severity assessment results, classifies user importance and risk severity into levels, and obtains a priority ranking list after classification. Preliminary scheduling is performed based on the hierarchical priority list to generate an initial resource scheduling plan. Fluctuations in user importance and risk severity are continuously tracked to determine if there are any deviations. If the deviation exceeds a preset threshold, a dynamic plan adjustment mechanism is triggered to obtain the adjusted scheduling parameters. The resource scheduling scheme is iteratively updated multiple times based on the adjusted scheduling parameters to determine the optimized dynamic scheme content and make real-time corrections. If new deviations are found, the loop execution process is restarted to obtain the final scheme update result.

5. A computer device, characterized in that, include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the method for graded early warning of electrical safety hazards for different customers as described in any one of claims 1 to 4.

6. A computer-readable storage medium, characterized in that, It stores computer-executable instructions, which, when executed by a processor, implement the steps of the graded early warning method for electrical safety hazards for different customers as described in any one of claims 1 to 4.

Images