A method for preventing misoperation in power grid dispatching operations with cross-level collaboration

Abstract

The present invention discloses a method for preventing misoperation in power grid dispatching operations with cross-level collaboration, which relates to the technical field of preventing misoperation in dispatching operations. The method includes determining the dispatching operation characteristics of the power grid dispatching operation to be analyzed; judging whether the dispatching operation characteristics meet a preset first condition to obtain a first judgment result; if the first judgment result meets the first condition, judging whether the dispatching operation characteristics meet a preset second condition to obtain a second judgment result; if the second judgment result meets the second condition, outputting a first scanning result; if the first judgment result does not meet the first condition, outputting a second scanning result; the second scanning result also updates the operation execution delay according to historical data, and through the automation of the power grid dispatching process, it solves the problem that the support capabilities of existing automation and intelligent systems are limited, unable to effectively process dispatching tasks, and increasing the risk of misoperation.

Description

Technical Field

[0001] The present invention relates to the technical field of anti-error dispatching operations. More specifically, the present invention relates to an anti-error method for power grid dispatching operations with cross-level collaboration. Background Art

[0002] Existing power grid dispatching operations cover all links from power generation, transmission to distribution, and mainly conduct real-time monitoring, command dispatching, and equipment management through an automated system. Dispatching operations include starting and stopping of generating units, load distribution, line control, voltage regulation, etc., to ensure the safe and stable operation of the power grid under different load conditions. At the same time, dispatchers also need to adjust the power grid configuration based on real-time data to avoid equipment overload or power grid failures. To improve the operation efficiency and safety of the power grid, intelligent technologies have also been incorporated into power grid dispatching, such as optimizing dispatching plans through machine learning and big data analysis to improve the accuracy and scientificity of decision-making.

[0003] Anti-error means that through a series of technical means and measures, it is ensured that during the power grid dispatching process, any potential errors can be detected and corrected in a timely manner, preventing operation errors from causing abnormal power grid operation or failures. The core goal of anti-error operation is to ensure the compliance of dispatching instructions, the accuracy of equipment status, and the reliability of the dispatching process.

[0004] Power grid dispatching operations are the key to ensuring the stable, safe, and efficient operation of the power grid system. With the growth of power demand and the expansion of the power grid scale, the complexity of power grid dispatching operations has been increasing continuously. Dispatchers need to handle various emergency situations and complex decisions, and are prone to misoperations during the operation process.

[0005] For example, a method, system, device, and medium for controlling the voltage of an AC-DC hybrid distribution network disclosed in the invention patent announcement with the publication number of CN117239768B includes the following steps:

[0006] It includes constructing the objective function and constraint conditions of the upper-level, middle-level, and lower-level voltage coordination control models for the AC-DC hybrid distribution network, solving the optimal solution through the upper-level, middle-level, and lower-level voltage coordination control models, and controlling the voltage of the AC-DC hybrid distribution network according to the optimal solution; the exchange power between the upper-level decision area and the AC-DC hybrid distribution network, as well as between areas; the middle-level decides the exchange power between each feeder within the area; the lower-level decides the output of distributed power sources, energy storage, and loads on each section of the feeder; giving full play to the regulation autonomy of distributed resources between different voltage levels, improving the robustness of voltage control, and finally achieving the maximization of in-situ consumption of new energy and promoting the cross-level collaborative interaction of adjustable resources.

[0007] For example, a power grid dispatching management system and a management control method announced in the invention patent announcement with the announcement number of CN113572172B include a power grid energy system, a substation operation monitoring module, a main station self-check module, and an energy dispatching module. The power grid energy system supplies power to a local area. The substation operation monitoring module monitors the power supply substations in real time. The main station self-check module is used to detect the remaining power after the main energy station transmits electricity. The energy dispatching module is used to transmit energy to the power supply substations. Through the setting of the substation operation monitoring module, the present invention monitors the power consumption of the power supply substations in real time, immediately judges the situation of the power supply substations in case of energy shortage, and when the power supply substations are in the section of excessive power consumption, the energy dispatching module screens out the power supply substations for energy dispatching, ensuring the continuity of energy use, controlling the dispatching and use of the power supply substations, and facilitating the management and coordination of each energy station.

[0008] In the above-mentioned disclosed technical solution, there are at least the following technical problems:

[0009] The existing power grid dispatching anti-error technology mainly relies on manual intervention and traditional single anti-error mechanisms, and cannot effectively cope with complex dispatching scenarios of multiple levels, cross-targets and multiple devices. When dealing with complex and sudden situations, dispatchers are prone to make wrong decisions due to insufficient information or excessive pressure. Especially in emergency moments, there is a lack of efficient decision support and clear coping strategies. In addition, the support capabilities of existing automated and intelligent systems are limited and cannot effectively handle all dispatching tasks, resulting in a large amount of manual intervention, which increases the risk of misoperation.

[0010] Moreover, during the power grid dispatching anti-error process, the length of the operation execution delay has an important impact on the system stability. Although a shorter delay can quickly recover, it may ignore potential problems due to hasty operations, increasing risks; while a longer delay can ensure more accurate repair, but it may lead to too long a duration of the fault, affecting the system stability.

[0011] In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0012] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a cross-level collaborative power grid dispatching operation anti-error method, which automates the power grid dispatching process to solve the problem that the support capabilities of existing automated and intelligent systems are limited and cannot effectively handle dispatching tasks, resulting in a large amount of manual intervention and increasing the risk of misoperation.

[0013] To achieve the above object, the present invention provides the following technical solutions:

[0014] A method for preventing misoperation in power grid dispatching operations with cross - level collaboration includes the following steps: determining the dispatching operation characteristics of the power grid dispatching operation to be analyzed; judging whether the dispatching operation characteristics meet a preset first condition to obtain a first judgment result, where the first condition includes an incorrect operation condition; if the first judgment result meets the first condition, judging whether the dispatching operation characteristics meet a preset second condition to obtain a second judgment result, where the second condition includes a cross - level collaboration operation condition; if the second judgment result meets the second condition, outputting a first scanning result, where the first scanning result is that the power grid dispatching operation is error - free; if the first judgment result does not meet the first condition, outputting a second scanning result, where the second scanning result is the termination of the operation; the second scanning result also updates the operation execution delay according to historical data.

[0015] In a preferred embodiment, the determining the dispatching operation characteristics of the power grid dispatching operation to be analyzed is specifically feature extraction of dispatching based on event - driven; the dispatching operation characteristics include event characteristics, trigger characteristics, and risk characteristics; based on intelligent sensors in the power grid, sudden events during the operation of the power grid are monitored in real - time, and the event characteristics of each type of sudden event are recorded to form a real - time event data set; based on a convolutional neural network, feature extraction is performed on the real - time event data set to capture non - linear and temporal correlations, obtaining dynamic event trigger characteristics, where the trigger characteristics reflect the changes and impacts of the power grid state when an event is triggered; based on the immediate state of the power grid when a real - time event occurs, combined with historical event data, the risk of dispatching operations caused after the event is predicted based on the time series of the neural network to form risk characteristics.

[0016] In a preferred embodiment, the first scanning result further includes cross - level collaborative regulation of dispatching errors and resolving dispatching errors at other levels.

[0017] In a preferred embodiment, the incorrect operation condition is specifically as follows: dispatching rules are defined, and each rule is judged according to different operation characteristics. The dispatching rules include single rules and overall rules; the single rule judges whether it meets the incorrect operation condition based on a decision tree; the overall rule judges whether it meets the incorrect operation condition based on an improved DTW.

[0018] In a preferred embodiment, the cross - level collaboration operation condition includes a collaborative operation condition and a hierarchical repair condition. The collaborative operation condition is specifically as follows: obtaining the total delay of the dispatching operation, and judging the dispatching stage of the dispatching operation by accumulating the total delay; obtaining the maximum preset time for each stage, and then calculating the remaining adjustable time according to the executed delay; obtaining the time required for cross - level collaborative operation, and judging whether repair can be performed by comparing the remaining adjustable time and the time required for cross - level collaborative operation. If the remaining adjustable time is greater than or equal to the time required for cross - level collaborative operation, it is considered that repair can be performed.

[0019] In a preferred embodiment, the hierarchical repair condition is specifically as follows: If the time required for cross - hierarchical collaborative operation exceeds the remaining adjustable time and does not meet the feasibility condition of cross - hierarchical collaborative operation, it is determined whether to perform error repair through the hierarchical control mechanism; if the total repair delay is less than or equal to the remaining adjustable time, it indicates that the hierarchical control can solve the scheduling error; when the remaining time is not sufficient to support repair or cross - hierarchical collaborative operation, the operation must be terminated.

[0020] In a preferred embodiment, if the first judgment result does not meet the first condition, the first scan result is also output; the first scan result also includes cross - hierarchical collaborative regulation of the scheduling error to solve the scheduling error at other levels.

[0021] In a preferred embodiment, the operation termination also includes: when it is determined that the first condition is not met, the operation termination avoids the spread of system errors by dynamically adjusting the scope of the current operation instead of terminating immediately; a multi - path recovery mechanism is included during the operation termination process, specifically: when it is determined that there are potential risks in the system, a preset recovery strategy is first triggered, the backup system and standby equipment are preferentially enabled, and the preset power grid parameters are adjusted for risk isolation until the operation returns to normal completely and then a complete termination is performed.

[0022] In a preferred embodiment, the update of the operation execution delay is specifically as follows: The update of the operation execution delay is achieved based on an improved PID control mechanism; the judgment results of each scheduling operation are obtained, including whether the operation is terminated and whether the error condition is met, and the feedback parameters of the operation are analyzed to establish a feedback model to capture the relationship with the operation execution delay to obtain the target delay. The feedback parameters include the termination frequency, operation risk, and operation efficiency; the difference between the current operation delay and the target delay is defined as the error of the PID control mechanism; the PID control mechanism is used to dynamically adjust the operation execution delay.

[0023] In a preferred embodiment, the single rule is to check a single scheduling condition; the overall rule is based on improved DTW to determine whether it meets the wrong operation condition, specifically: each scheduling operation feature is transformed into a scheduling operation feature time series; the corresponding overall rule feature time series is extracted for the overall rule; the scheduling operation feature time series and the overall rule feature time series are respectively processed hierarchically to extract the feature time series and overall rule features at different time scales; the DTW algorithm is applied for matching at different scales to minimize the DTW difference between the scheduling operation feature time series and the overall rule feature time series to obtain the most similar overall rule, and the DTW distance corresponding to the most similar overall rule is compared with the preset dynamic similarity threshold to determine whether it meets the wrong operation condition.

[0024] In a preferred embodiment, the specific method for obtaining the feedback model is as follows: The feedback model is a specific regression model; a corresponding regression method is selected according to the relationship between the feedback parameters to establish a regression model. Specifically, when the relationship between the operation delay, termination frequency, operation risk, and operation efficiency is a linear relationship, a linear regression model is constructed as the feedback model; when the relationship between the operation delay, termination frequency, operation risk, and operation efficiency is non-linear, a polynomial regression model is constructed based on the support vector machine as the feedback model; the feedback model is trained using the feedback parameters and the corresponding operation delay as training data, with the feedback parameters as the input and the target delay as the output.

[0025] In a preferred embodiment, the method for obtaining the dynamic similarity threshold is as follows: State evaluation factors are obtained according to different states of the power grid and the complexity of dispatching operations; the state evaluation factors include load state evaluation factors, equipment health state evaluation factors, external environmental factor evaluation factors, and dispatching operation complexity evaluation factors; the dynamic similarity threshold is calculated by weighted summation of the state evaluation factors; in the overall rule matching process, the state evaluation factors are preset and have been determined before judgment.

[0026] The technical effects and advantages of a cross-level collaborative power grid dispatching operation error prevention method of the present invention:

[0027] 1. The present invention uses intelligent sensors to monitor emergencies in the operation of the power grid in real time, extracts event features using a convolutional neural network, and combines historical data to predict dispatching operation risks. By judging whether the dispatching operation meets the wrong operation conditions, including load overlimit, abnormal equipment temperature, etc., a multi-scale dynamic time warping (DTW) method is used to improve the matching accuracy, and the judgment criteria are flexibly adjusted to adapt to different power grid states and the complexity of dispatching operations. In addition, the method also introduces a cross-level collaborative operation and hierarchical repair mechanism, and judges whether to perform repair or termination operations according to the remaining adjustable time, providing an efficient error repair solution to ensure the stability and reliability of power grid dispatching.

[0028] 2. The present invention dynamically updates the operation execution delay based on an improved PID control mechanism, significantly improving the response speed and stability of the system. By real-time monitoring the judgment results of scheduling operations (such as whether the operation terminates, whether error conditions are met, etc.) and combining the feedback parameters of the operation (such as termination frequency, operation risk, and operation efficiency), a feedback model is established to accurately capture the relationship between the operation delay and these parameters. Using the proportional, integral, and derivative parts of the PID control mechanism, the system can adjust the operation delay according to the real-time error, avoiding system instability caused by excessive or too small delays. This dynamic adjustment mechanism ensures that the system can adapt to the changing operation environment, thus guaranteeing the smooth operation and efficient response of power grid scheduling. In addition, by selecting an appropriate regression method to construct the feedback model, the adaptability and accuracy of the system are further enhanced, and appropriate regulation methods can be selected according to different operation delay characteristics, achieving more precise and intelligent operation execution. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic flowchart of a method for preventing misoperation of power grid scheduling with cross-level collaboration according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0031] Embodiment 1 Figure 1 A method for preventing misoperation of power grid scheduling with cross-level collaboration according to the present invention is given, including the following steps:

[0032] S1, determining the scheduling operation characteristics of the power grid scheduling operation to be analyzed;

[0033] The determination of the scheduling operation characteristics of the power grid scheduling operation to be analyzed is specifically the extraction of scheduling characteristics based on event-driven;

[0034] The scheduling operation characteristics include event characteristics, trigger characteristics, and risk characteristics;

[0035] Based on intelligent sensors in the power grid, sudden events during power grid operation are real-time monitored, and the event characteristics of each type of sudden event are recorded to form a real-time event data set;

[0036] Based on a convolutional neural network, feature extraction is performed on the real-time event data set to capture non-linearity and temporal correlation, obtaining dynamic event trigger characteristics, and the trigger characteristics reflect the changes and impacts of the power grid state when the event is triggered;

[0037] Based on the immediate state of the power grid when a real-time event occurs, combined with historical event data, a time series based on a neural network is used to predict the scheduling operation risks caused after the event occurs, and risk characteristics are formed.

[0038] S2. Determine whether the scheduling operation characteristics meet a preset first condition to obtain a first judgment result, where the first condition includes an incorrect operation condition;

[0039] The specific incorrect operation condition is:

[0040] Define scheduling rules, which are derived from the operation specifications of the power grid, the operation manuals of equipment, and historical experience; each rule is determined according to different operation characteristics, and the scheduling rules include single rules and overall rules;

[0041] For example, load exceeding the rated capacity of the equipment: When a scheduling operation causes the load of a certain equipment (such as a transformer, line) to exceed its rated capacity, it is considered an incorrect operation.

[0042] Abnormal equipment temperature: If a scheduling operation causes the temperature of some equipment (such as a transformer, cable) to exceed the safety threshold, it is determined as an incorrect operation.

[0043] Violating the scheduling priority: If a scheduling operation violates the predetermined load priority and causes a load that should not be powered off to be cut off, it is also an incorrect operation.

[0044] Current / voltage exceeding the safe range: If a scheduling instruction causes abnormal current and voltage in the system (such as overvoltage, overcurrent, etc.), it will be determined as an incorrect operation.

[0045] The single rule is to check a single scheduling condition;

[0046] The overall rule is to conduct a multi-dimensional analysis and make a comprehensive judgment by combining multiple incorrect operation conditions; the overall rule is used to judge the similarity between different scheduling operations, and can discover potential risks and incorrect operations in the operation process; the overall rule can consider the interactivity, timing, and combinability of constraints in scheduling, and judge whether the scheduling operation is safe based on these comprehensive factors;

[0047] The single rule is based on a decision tree to judge whether it meets the incorrect operation condition;

[0048] For example, input parameters:

[0049] The scheduling information of the current power grid scheduling operation (such as load demand, current voltage, temperature, equipment status, etc.);

[0050] The real-time data of each device in the power grid (such as device temperature, load, frequency, etc.);

[0051] Operation specifications and constraints of the equipment (such as equipment rated capacity, temperature range, operation priority, etc.);

[0052] Rule matching: Rule 1: If the load > the equipment rated capacity, output "wrong operation";

[0053] Rule 2: If the equipment temperature > the maximum operating temperature of the equipment, output "wrong operation";

[0054] Rule 3: If the scheduling operation violates the operation priority rule, output "wrong operation".

[0055] The overall rule is based on the improved DTW to judge whether it meets the wrong operation condition. Specifically:

[0056] Convert each scheduling operation feature into a scheduling operation feature time series;

[0057] Extract the corresponding overall rule feature time series for the overall rule;

[0058] Perform hierarchical processing on the scheduling operation feature time series and the overall rule feature time series respectively to extract the feature time series and overall rule features at different time scales;

[0059] Apply the DTW algorithm for matching at different scales to minimize the DTW difference between the scheduling operation feature time series and the overall rule feature time series, obtain the most similar overall rule, and judge whether it meets the wrong operation condition by comparing the DTW distance corresponding to the most similar overall rule with the preset dynamic similarity threshold.

[0060] The method for obtaining the dynamic similarity threshold is as follows:

[0061] Obtain the state evaluation factors according to different states of the power grid and the complexity of the scheduling operation, including:

[0062] Load state: The load level of the power grid (such as high load, low load) will directly affect the flexibility of the scheduling operation. In a high-load state, the scheduling of the power grid may be more tense, requiring a looser matching threshold to avoid overly strict wrong judgments; while in a low-load state, the power grid is more stable and may require a stricter judgment criterion.

[0063] Equipment health state: The health state of key equipment in the power grid (such as transformers, lines) has an important impact on the stability of the scheduling operation. In the case of equipment failure or maintenance, the similarity threshold should be stricter to reduce the occurrence of wrong operations.

[0064] External environmental factors, such as weather and seasonal changes, can also affect the load and dispatching operations of the power grid. For example, abnormal weather may cause equipment to operate overloaded, or the load change may be more drastic. In this case, the threshold can be appropriately relaxed to tolerate a certain degree of uncertainty.

[0065] Complexity of dispatching operations: Dynamically adjust the similarity threshold according to the complexity of dispatching operations (such as multiple operations being carried out simultaneously). Complex dispatching operations may require a higher error tolerance rate, so the similarity threshold can be increased accordingly.

[0066] Perform a weighted sum of the state evaluation factors to calculate a dynamically adjusted similarity threshold; the dynamically adjusted similarity threshold will determine the matching degree of DTW and avoid misjudging incorrect operations under extreme conditions.

[0067] For example: In the calculation of the similarity threshold, assume that at a specific moment, the operating state of the power grid is as follows:

[0068] The power grid is in a high-load state, and the load impact factor T1 = 1.2 (the threshold is increased by 20%).

[0069] The equipment is in good health, and the equipment impact factor T2 = 1.0 (no change).

[0070] The current weather conditions are normal, and the environmental impact factor T3 = 1.0.

[0071] The dispatching operation is relatively complex, and the operation complexity factor T4 = 1.2 (the threshold is increased by 20%).

[0072] According to the preset state evaluation factors, the dynamic similarity threshold can be calculated:

[0073]

[0074] It should be noted that the data points of each feature represent the values at a certain time point. For example, the time series of the load may include the changes in the load during an operation cycle. The time series of the equipment state may include the changes of the equipment from the normal state to the faulty state.

[0075] It should be noted that in the overall rule matching process, the state evaluation factors are preset and have been determined before the judgment. They are set according to the real-time state of the power grid (such as load, equipment health, external environment, etc.) to reflect the risk tolerance of dispatching operations. These factors will play a decisive role in the dynamic adjustment of the similarity threshold under different circumstances, so as to ensure that the system can flexibly respond to incorrect operations and maintain dispatching safety under specific conditions.

[0076] It should be noted that for the single rule: each rule corresponds to a specific condition, and as long as one condition is not met, the scheduling operation is determined to be incorrect. For example, when the load exceeds the rated capacity, the scheduling operation is considered incorrect.

[0077] Overall rule: Considering the combination of multiple conditions, the operation must meet specific constraint conditions in multiple dimensions. Only when the comprehensive conditions are met is the scheduling operation considered correct. The overall rule usually includes a set of multiple rules and the mutual constraints between conditions.

[0078] The improved dynamic time warping (DTW) method applicable to multiple scales can more effectively capture the changes and correlations of power grid scheduling operation characteristics at different time scales. By comparing time series at multiple scales, it can identify potential complex patterns and long-term and short-term dependencies, improving the matching accuracy and fault tolerance. Compared with the traditional single-scale DTW, the improved multi-scale method can provide more sensitive and accurate error judgments when dealing with scheduling operations with strong nonlinearity and time series correlation, reducing the possibility of false alarms and missed alarms, thereby enhancing the reliability and security of power grid scheduling.

[0079] The introduction of a dynamic similarity threshold can flexibly adjust the judgment criteria according to the real-time state and characteristic changes of power grid scheduling operations, making the overall rule matching more accurate and adaptable. Compared with the static threshold, the dynamic threshold can automatically adapt to fluctuations and abnormal situations in different operation scenarios, avoiding misjudgments or missed judgments caused by overly strict or loose criteria. It can adjust the tolerance of similarity in real time based on historical data, current operation characteristics, and system status, thereby effectively improving the error detection accuracy and stability in the power grid scheduling process.

[0080] S3. If the first judgment result meets the first condition, determine whether the scheduling operation characteristics meet the preset second condition to obtain a second judgment result, where the second condition includes a cross-level collaborative operation condition.

[0081] The cross-level collaborative operation condition is specifically:

[0082] The cross-level collaborative operation condition includes a collaborative operation condition and a hierarchical repair condition.

[0083] Obtain the total delay of the scheduling operation, and judge the scheduling stage at which the scheduling operation is located by accumulating the total delay. The specific formula for stage identification is as follows:

[0084]

[0085] is the total delay of the scheduling operation. Let \(t_i\) be the time delay of the \(i\)-th stage, where \(i = 1, 2,\cdots, n\) and \(n\) is the total number of stages of the scheduling operation. If the current executed time delay is the sum of the time delays before the \(k\)-th time, then the current operation is in the \(k\)-th stage;

[0086] Obtain the maximum preset time for each stage, and then calculate the remaining adjustable time according to the executed time delay. Specifically:

[0087]

[0088] Let \(T\) be the remaining adjustable time, Let \(T_{max}\) be the maximum preset time delay for each stage, Let \(t\) be the executed time delay;

[0089] Obtain the time required for cross-level collaborative operation. By comparing the remaining adjustable time and the time required for cross-level collaborative operation, determine whether repair can be performed. If the remaining adjustable time is greater than or equal to the time required for cross-level collaborative operation, the operation is considered feasible. Specifically:

[0090]

[0091] In the formula, Let \(T_{co}\) be the time required for cross-level collaborative operation;

[0092] If the time required for cross-level collaborative operation exceeds the remaining adjustable time and does not meet the feasibility condition of cross-level collaborative operation, then determine whether to perform error repair through the hierarchical control mechanism. The total repair time for multiple levels is:

[0093]

[0094] In the formula, Let \(T_{total}\) be the total repair time delay, Let \(L\) be the number of levels, Let \(T_{level}\) be the repair time required for each level. If the total repair time delay is less than or equal to the remaining adjustable time, it means that the hierarchical control can solve the scheduling error;

[0095] When the remaining time is not enough to support repair or cross-level collaborative operation, the operation must be terminated; the specific method for determining whether to terminate the operation is: if the remaining time is less than the minimum repair time delay threshold or the total repair time delay exceeds the remaining time, the operation needs to be terminated;

[0096] The decision model for determining whether the scheduling operation characteristics meet the preset second condition is as follows:

[0097] .

[0098] The cross - level collaborative operation condition provides a flexible and efficient solution for the repair of scheduling errors by comprehensively judging whether the collaborative operation is feasible and whether the hierarchical repair is feasible. First, the conditional judgment of the collaborative operation ensures that when the remaining scheduling time is sufficient, the problem can be resolved through the coordination of the scheduling forces at the other levels, thus avoiding the termination of the operation. If the collaborative operation is not feasible, then it turns to judge whether it is possible to rely on the collaboration between levels to alleviate or repair the error through hierarchical repair. The advantage of this method is that it makes full use of multi - level scheduling resources and repair strategies, making the error repair process more intelligent and dynamic, and avoiding premature termination of scheduling, thereby improving the stability and reliability of the system. In addition, by dynamically judging the feasibility of different repair paths, it can better adapt to the changing time delays and resource constraints in actual scheduling operations.

[0099] It should be noted that the process of cross - level collaborative operation judgment first judges whether there is enough time for cross - level regulation by evaluating the time delay of the current scheduling operation and the stage at which the operation is located. In this process, when judging the feasibility of cross - level collaborative operation, factors such as the feasibility of the collaborative operation, the availability of resource allocation, and whether specific time constraints are met are considered. Hierarchical repair, as an existing technology, has been widely used. Its core concept is to rely on the coordination and repair between levels to ensure that even when an operation at a certain level fails, the other levels can intervene in a timely manner and take corresponding repair measures, so it will not be elaborated here. In this solution, the judgment of the cross - level collaborative operation condition mainly depends on the real - time state and dynamic changes of the scheduling execution, so as to flexibly decide whether it is possible to repair or alleviate errors through cross - level operations.

[0100] S4, if the second judgment result meets the second condition, output the first scan result, where the first scan result is that the power grid scheduling operation is error - free;

[0101] If the first judgment result does not meet the first condition, the first scan result is also output.

[0102] The first scan result also includes cross - level collaborative regulation of the scheduling error to resolve the scheduling error at other levels.

[0103] It should be noted that cross - level collaborative regulation not only involves judging whether the error can be resolved at other levels, but also needs to consider real - time data feedback, information flow between levels, and decision - making support. During the judgment, real - time monitoring mechanisms and dynamic adjustments are crucial because they can correct scheduling operations in a timely manner according to changes in the current state of the power grid. In addition, the effectiveness of the cross - level collaborative operation scheme also needs to be ensured through timeliness and efficiency evaluation to ensure its rapid response and effective reduction of system errors, and to avoid greater - scope impacts caused by delays.

[0104] S5. If the first judgment result does not meet the first condition, output a second scan result, where the second scan result is the termination of the operation.

[0105] The termination of the operation also includes that when it is judged that the first condition is not met, the termination of the operation avoids the spread of system errors by dynamically adjusting the scope of the current operation, rather than terminating immediately.

[0106] For example, by gradually reducing the amplitude or scope of the scheduling operation, first perform local operation termination, then observe whether there are further risks, and then decide whether to completely terminate the operation.

[0107] The termination of the operation also includes a multi-path recovery mechanism during the operation termination process;

[0108] Specifically, when it is judged that there are potential risks in the system, first trigger a preset recovery strategy, give priority to enabling backup systems and standby devices, and adjust preset grid parameters for risk isolation until the operation returns to normal completely and then perform complete termination.

[0109] By introducing dynamic adjustment and multi-path recovery mechanisms, a more flexible response strategy is provided during the operation termination process, avoiding the "one-size-fits-all" method of traditional operation termination. Specifically, when it is judged that the first condition is not met, first dynamically adjust the scope of the scheduling operation, gradually reduce the operation amplitude or scope, so as to effectively avoid the spread of system errors and reduce unnecessary risks. By performing local operation termination and observing further risks, it is possible to more accurately evaluate whether the operation needs to be completely terminated, improving the accuracy of the operation and the stability of the system. At the same time, the added multi-path recovery mechanism can, when there are potential risks in the system, through preset recovery strategies such as enabling backup systems, standby devices or adjusting grid parameters, give priority to risk isolation, ensuring that the system is further terminated after returning to the normal state. Such innovative measures improve the flexibility, fault tolerance and stability of the system, effectively prevent large-scale power grid failures caused by emergencies, and improve the safety and reliability of power grid scheduling operations.

[0110] This implementation monitors emergencies in the power grid operation in real time through intelligent sensors, extracts event features using convolutional neural networks, and combines historical data to predict scheduling operation risks. This method determines whether the scheduling operation meets the wrong operation conditions, including load overlimit, abnormal equipment temperature, etc., and uses a multi-scale dynamic time warping (DTW) method to improve the matching accuracy and flexibly adjust the judgment criteria to adapt to the complexity of different power grid states and scheduling operations. In addition, the method also introduces cross-level collaborative operations and hierarchical repair mechanisms, and judges whether to perform repair or termination operations according to the remaining adjustable time, providing an efficient error repair solution to ensure the stability and reliability of power grid scheduling.

[0111] Example 2, S6. The second scan result also updates the operation execution time delay according to historical data.

[0112] The update of the operation execution time delay is specifically as follows:

[0113] The update of the operation execution time delay is achieved based on an improved PID control mechanism;

[0114] Obtain the judgment results of each scheduling operation, including whether the operation is terminated and whether the error condition is met, analyze to obtain the feedback parameters of the operation, establish a feedback model to capture the relationship with the operation execution time delay, and obtain the target time delay. The feedback parameters include termination frequency, operation risk, and operation efficiency;

[0115] Define the difference between the current operation time delay and the target time delay as the error of the PID control mechanism;

[0116] Use the PID control mechanism to dynamically adjust the operation execution time delay;

[0117] The controller adjustment formula is as follows:

[0118] In the formula, is the adjustment amount of the operation execution time delay, is the proportional coefficient to the error, is the error at the previous moment, is the integral coefficient for accumulating past errors, is the error function, is the differential coefficient of the PID controller.

[0119] It should be noted that the PID controller usually consists of three parts:

[0120] Proportional part (P): Directly proportional to the error, representing the influence of the operation error on the execution time delay, such as in this application .

[0121] Integral part (I): The effect of accumulating past errors, ensuring that the error deviation remains effective, such as in this application .

[0122] Differential part (D): Predict the trend of future errors and reduce the system overshoot phenomenon, such as in this application .

[0123] For example, if a scheduling operation frequently triggers errors or terminations, the system will consider that this operation requires more time to evaluate and handle potential risks, so the execution time delay is appropriately increased; conversely, if the scheduling operation is less terminated, the system will appropriately reduce the time delay to improve the execution efficiency.

[0124] It should be noted that the termination frequency refers to the frequency at which operations were terminated over a past period (e.g., the number of terminations per hour or per day).

[0125] Operation risk: Based on historical error types and risk judgments during the scheduling process (such as equipment failures, improper operations).

[0126] Operation efficiency: The efficiency of the system in completing scheduling operations (e.g., the time required, the degree of completion).

[0127] The specific method for obtaining the feedback model is as follows:

[0128] The feedback model is a specific regression model;

[0129] Select a corresponding regression method to establish a regression model according to the relationship between feedback parameters. Specifically:

[0130] When the relationship between operation delay and termination frequency, operation risk, and operation efficiency is a linear relationship, construct a linear regression model as the feedback model;

[0131] When the relationship between operation delay and termination frequency, operation risk, and operation efficiency is non-linear, construct a polynomial regression model based on support vector machines as the feedback model;

[0132] Use the feedback parameters and the corresponding operation delay as training data to train the feedback model, with the feedback parameters as the input and the target delay as the output.

[0133] Updating the operation delay through the PID control mechanism can effectively improve the response speed and stability of the system. PID control can precisely control the system's real-time changes by adjusting the three parameters of proportional (P), integral (I), and derivative (D) in real-time. This mechanism can automatically adjust the operation delay, avoiding system instability caused by excessive or too small delays, thus ensuring the smooth operation and response accuracy of the system.

[0134] The improved PID control mechanism further optimizes the performance of the traditional PID control. By introducing an adaptive adjustment strategy or an improved parameter update algorithm, the control system can more flexibly respond to environmental changes or fluctuations in operating conditions. Compared with the traditional PID control mechanism, the improved PID control mechanism can provide higher accuracy, faster response speed, and reduce the risk of over-regulation or system oscillation in a wider range of application scenarios. This mechanism is especially suitable for complex systems with high requirements for real-time performance and stability, and can effectively improve the overall performance and reliability.

[0135] By selecting an appropriate regression method to construct a feedback model based on the relationship between the operation delay and other key parameters (such as termination frequency, operation risk, operation efficiency), the accuracy and adaptability of the model can be effectively improved. When these relationships are linear, using a linear regression model can simplify the calculation and improve the calculation efficiency; while when the relationships are non-linear, a polynomial regression model based on support vector machines can better capture complex non-linear relationships, thereby improving the prediction accuracy. In addition, training with feedback parameters as input and target delay as output enables the feedback model to be dynamically adjusted during actual operations, thus optimizing the operation delay and efficiency of the system and ensuring more accurate scheduling operations.

[0136] In this embodiment, by dynamically updating the operation execution delay based on an improved PID control mechanism, the response speed and stability of the system are significantly improved. This method establishes a feedback model by real-time monitoring the judgment results of scheduling operations (such as whether the operation terminates, whether error conditions are met, etc.) and combining the feedback parameters of the operation (such as termination frequency, operation risk, and operation efficiency), and accurately captures the relationship between the operation delay and these parameters. Using the proportional, integral, and derivative parts of the PID control mechanism, the system can adjust the operation delay according to the real-time error, avoiding system instability caused by excessive or too small delays. This dynamic adjustment mechanism ensures that the system can adapt to the changing operation environment, thus ensuring the smooth operation and efficient response of the power grid scheduling. In addition, by selecting an appropriate regression method to construct the feedback model, the adaptability and accuracy of the system are further enhanced, and it can select a suitable control method according to different operation delay characteristics, realizing more accurate and intelligent operation execution.

[0137] The above formulas are all dimensionless and take their numerical values for calculation. The formula is obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0138] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0139] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0140] In addition, in each embodiment of the present application, each functional module may be integrated into one processing module, may exist separately physically for each module, or two or more modules may be integrated into one module.

[0141] As described above, it is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0142] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A cross-level collaborative power grid dispatching operation error prevention method, characterized in that: The steps include: extracting dispatching operation characteristics of the power grid dispatching operation to be analyzed; Preset judgment conditions, compare the scheduling operation characteristics with the judgment conditions, obtain a judgment result, and output a scanning result according to the judgment result, wherein the scanning result includes a judgment on the delay update of the operation execution; The preset judgment condition includes a first condition and a second condition, the first condition includes an erroneous operation condition, and the second condition includes a cross-level collaborative operation condition; Comparing the scheduling operation characteristic with the first condition to obtain a first judgment result; When the first judgment result satisfies the first condition, the scheduling operation characteristic is compared with the second condition to obtain a second judgment result; The output scanning result includes a first scanning result and a second scanning result, If the second judgment result satisfies the second condition, the first scanning result is output, and the first scanning result indicates that the power grid dispatching operation is correct; If the first judgment result is that the first condition is not met, a second scanning result is output, and the second scanning result is that the operation is terminated; The first condition includes judging the overall rule based on the dynamic similarity threshold and the DTW improved by the hierarchical features.

2. The cross-level collaborative power grid dispatching operation error prevention method according to claim 1 is characterized in that: The first scanning result also includes cross-level coordinated regulation of scheduling errors, and the second scanning result also updates the operation execution delay based on historical data.

3. The cross-level collaborative power grid dispatching operation error prevention method according to claim 2 is characterized in that: The specific judgment process of the erroneous operation condition is as follows: The scheduling rules are defined, each rule is judged according to different operation characteristics, and the scheduling rules include single rules and overall rules; the single rules are judged based on the decision tree, and the overall rules are judged based on the improved DTW.

4. The cross-level collaborative power grid dispatching operation error prevention method according to claim 3 is characterized in that: The cross-level collaborative operation conditions include collaborative operation conditions, specifically: Obtain the total delay of the scheduling operation, and determine the scheduling stage of the scheduling operation by accumulating the total delay; Get the maximum preset time for each stage, and calculate the remaining adjustable time based on the executed delay; Obtain the time required for cross-level collaborative operations, and determine whether the collaborative operation conditions are met by comparing the remaining adjustable time and the time required for cross-level collaborative operations.

5. The cross-level coordinated power grid dispatching operation error prevention method according to claim 4 is characterized in that: The cross-level collaborative operation conditions also include layered repair conditions, specifically: If the time required for cross-level collaborative operations exceeds the remaining adjustable time, it is determined whether to perform error repair through the hierarchical control mechanism; Obtain the total repair delay and compare it with the remaining adjustable time to determine whether the layered repair conditions are met; When the collaborative operation conditions and layered repair conditions are not met, the operation is terminated.

6. The cross-level coordinated power grid dispatching operation error prevention method according to claim 5 is characterized in that: The operation termination also includes a dynamic adjustment mechanism and a multipath recovery mechanism: The dynamic adjustment mechanism is specifically: When it is determined that the first condition is not satisfied, the operation is terminated by dynamically adjusting the scope of the current operation; The multipath recovery mechanism is specifically: When the second scan result is output, the preset recovery strategy is triggered first, the backup system and spare equipment are enabled first, and the preset power grid parameters are adjusted to isolate risks until the operation is fully restored to normal and then completely terminated.

7. The cross-level coordinated power grid dispatching operation error prevention method according to claim 6 is characterized in that: The updating of the operation execution delay is implemented based on an improved PID control mechanism, specifically: Obtain the judgment result of each scheduling operation, analyze and obtain the feedback parameters of the operation, establish a feedback model to capture the relationship between the operation execution delay, and obtain the target delay. The judgment result includes whether the operation is terminated and whether the error condition is met. The feedback parameters include termination frequency, operation risk, and operation efficiency. The difference between the current operation delay and the target delay is defined as the error of the PID control mechanism, and the operation execution delay is dynamically adjusted based on the PID control mechanism.

8. The cross-level coordinated power grid dispatching operation error prevention method according to claim 7 is characterized in that: The overall rule is based on the improved DTW, specifically: Convert each scheduling operation feature into a scheduling operation feature time series, and extract the overall rule feature time series; The characteristic time series of scheduling operation and the characteristic time series of overall rule are processed in layers respectively, and the characteristic time series and overall rule features of different time scales are extracted; Based on the DTW algorithm, the overall rule matching is performed to minimize the DTW difference between the scheduling operation feature time series and the overall rule feature time series, and the most similar overall rule is obtained by traversing; The DTW distance corresponding to the most similar overall rule is compared with a preset dynamic similarity threshold to determine whether the first condition is met.

9. The cross-level coordinated power grid dispatching operation error prevention method according to claim 8, characterized in that: The establishment of the feedback model captures the relationship between the operation execution delay and obtains the target delay, specifically including: Determine the linear relationship between operation delay and feedback parameters; If it is a linear relationship, a linear regression model is constructed as a feedback model; If it is a nonlinear relationship, a polynomial regression model is constructed based on the support vector machine as the feedback model; The feedback model is trained based on the feedback parameters and the corresponding operation delay as training data, with the feedback parameters as input and the target delay as output.

10. The cross-level coordinated power grid dispatching operation error prevention method according to claim 9, characterized in that: The method for obtaining the dynamic similarity threshold is: Acquire state assessment factors according to different states of the power grid and the complexity of the dispatching operation, wherein the state assessment factors include a load state assessment factor, an equipment health state assessment factor, an external environment factor assessment factor, and a dispatching operation complexity assessment factor; The dynamic similarity threshold is calculated by weighted summation of the state evaluation factors.

11. The cross-level coordinated power grid dispatching operation error prevention method according to claim 10, characterized in that: The specific judgment formula of the scheduling stage is as follows: In the formula, is the total delay of the scheduling operation, is the delay of the i-th stage, i=1, 2, …, n, n is the total number of stages of the scheduled operation. If the currently executed delay is the sum of the delay before the k-th time, the current operation is in the k-th stage.

12. The cross-level coordinated power grid dispatching operation error prevention method according to claim 11, characterized in that: The remaining adjustable time is specifically: For the remaining adjustable time, is the maximum preset delay for each stage, is the executed delay; The total repair delay is specifically: In the formula, To fix the total delay, is the number of layers, The time required for repair at each level. If the total repair delay is less than or equal to the remaining adjustable time, it means that hierarchical control can solve the scheduling error; The decision model for judging whether the scheduling operation characteristics meet the preset second condition is as follows: In the formula, The time required for cross-level collaborative operations.

13. The cross-level coordinated power grid dispatching operation error prevention method according to claim 12, characterized in that: The PID control mechanism is specifically: In the formula, is the adjustment amount for the operation execution delay, The proportionality coefficient with the error, is the error at the previous moment, is the integral coefficient for accumulating past errors, is the error function, is the differential coefficient of the PID controller.

Images