Method and system for dynamically planning flat and urgent states of power grid based on Hemma algorithm

Abstract

The invention discloses a method and a system for dynamically planning the level and urgent state of a power grid based on a Hemma algorithm, and relates to the technical field of power grid emergency regulation and control. The method comprises the following steps: according to historical emergency state data, constructing an initial river horse status influence relationship based on disaster stability and disaster types, and constructing an initial river position influence relationship based on disaster sources and terrain types; constructing a power grid emergency optimization objective function according to the highest participation complexity and the lowest damage probability based on a He-horse optimization algorithm architecture; constructing a power grid flat valley optimization objective function based on the state transition relation and the minimum flat valley fluctuation; and obtaining an initial river horse status and an initial river position, and obtaining a power grid dynamic planning strategy by using the power grid emergency optimization objective function and the power grid flat valley optimization objective function. According to the invention, emergency power supply regulation and control can be realized in time in an emergency state, and the timeliness and stability of regulation and control are improved.

Description

Technical Field

[0001] This application relates to the technical field of power grid emergency regulation, and particularly to a dynamic programming method and system for power grid normal and emergency states based on the hippopotamus algorithm. Background Art

[0002] In recent years, the access of distributed energy has relieved the carbon emission pressure for power grid power supply. However, since distributed energy usually needs to be built in open areas, it has poor resistance ability in disaster weather and a high probability of being damaged, resulting in unstable supply of distributed energy in abnormal weather. In natural disasters, large-scale damage to the power system may also lead to insufficient power supply in the affected areas. Medical facilities rely on small energy storage devices, affecting the rescue progress and unable to ensure the safety of personnel.

[0003] In related technologies, elastic quantitative assessment of the power grid is adopted, and the relationship between extreme disaster evolution and equipment state transition is established to clarify the component withdrawal situation in the time-varying scenario under disasters. However, the affected range of some disasters cannot be predicted, and each time the component damage caused by the disaster is simulated, the power grid cannot be regulated in time, resulting in low regulation efficiency. Moreover, emergency regulation is accompanied by rapid switching of power grid equipment, which is likely to cause damage to power grid equipment, and rapid switching will also bring power grid imbalance problems, affecting the stability of the power grid.

[0004] The patent application "Distribution Network Fault Reconfiguration Method Based on Improved Hippopotamus Optimization Algorithm", publication number: CN118868026A, publication date: October 29, 2024, specifically discloses using connected components to evaluate fault recovery, determining in advance the loads that can be restored through reconfiguration, and reducing the complexity of problem solving. Then, encoding is combined with the minimum cycle set to map the solution space of the distribution network fault reconfiguration problem to the search space of the algorithm. This encoding method can greatly increase the proportion of effective solutions. Then, the traditional hippopotamus optimization algorithm is improved in combination with the proposed encoding strategy to make it adaptable to discrete optimization problems. Finally, the improved hippopotamus optimization algorithm is used to quickly solve the distribution network fault reconfiguration strategy, thereby achieving rapid power supply restoration. This solution generates the initial population of the hippopotamus optimization algorithm through the proposed encoding strategy, improving the solution efficiency of the hippopotamus optimization algorithm, so as to improve the solution efficiency of the hippopotamus optimization algorithm for distribution network fault reconfiguration. However, it is a solution to the optimal way of restoration and reconstruction and cannot solve the problem of power supply interruption caused by power station damage in the emergency state.

[0005] Patent Application "A Method and System for Energy Storage Capacity Planning Considering Power Supply in Emergency Scenarios", Publication Number: CN119849881A, Publication Date: April 18, 2025, specifically discloses the following: Obtain annual operation data, and construct an energy dispatch optimization model for energy storage in normal scenarios and emergency power supply scenarios based on the annual operation data; according to the energy dispatch optimization model, considering the configuration parameters of the energy storage and the uncertainty caused by fluctuations in grid operation parameters, establish an energy storage capacity optimization model considering emergency power supply scenarios; input the relevant parameters of the energy storage into the energy storage capacity optimization model, and the energy storage capacity optimization model outputs the energy storage capacity planning result. This solution only considers the situation of energy storage devices for emergency power supply. However, in some disasters, such as in strong wind weather, although the power generation of wind power plants is affected, the power generation of hydropower plants can still provide emergency power supply compensation. Moreover, this solution is still based on the emergency switching of energy storage devices in the case of grid emergency power supply, and there will still be grid imbalance problems. Summary of the Invention

[0006] In view of the problems in the prior art that the supply stability is poor in the emergency state and the emergency switching is likely to damage the power station, the present application provides a method and system for dynamic planning of grid normal and emergency states based on the hippopotamus algorithm. By constructing the initial hippopotamus position, the iteration of hippopotamus optimization calculation is reduced, and the emergency planning efficiency is improved. And through the influence relationship of the initial river position, the position of the disaster source is continuously updated during the optimization process. At the same time, the adjustment of the normal valley state is carried out by using the time sequence of the conversion from the normal valley state to the emergency state, reducing the number of emergency switchings in the emergency state and ensuring the smoothness of power station regulation. Using the resistance behavior in the hippopotamus optimization algorithm to simulate the disaster avoidance (such as shutdown) of the power station, etc., improves the accuracy of power station emergency planning.

[0007] To achieve the above technical objectives, a technical solution provided by the present application is a method for dynamic planning of grid normal and emergency states based on the hippopotamus algorithm, including the following steps: Construct an initial hippopotamus position influence relationship for each power station based on the historical emergency state data according to the disaster stability degree and disaster type; construct an initial river position influence relationship for each disaster source based on the historical emergency state data according to the disaster source and terrain type; construct a grid emergency optimization objective function based on the hippopotamus optimization algorithm framework according to the highest participation complexity and the lowest damage probability; obtain the state conversion relationship based on the historical emergency state data and historical normal valley state data, and construct a grid normal valley optimization objective function based on the state conversion relationship and the minimum normal valley fluctuation; call the initial hippopotamus position influence relationship and the initial river position influence relationship according to the current regional state data to obtain the initial hippopotamus position and the initial river position, and use the grid emergency optimization objective function and the grid normal valley optimization objective function to obtain the grid dynamic planning strategy.

[0008] Further, the construction of the initial hippopotamus status influence relationship for each power station based on geographical location and disaster type according to historical emergency status data includes: extracting the disaster type and the degree of disaster-affected stability of each power station from the historical emergency status data; constructing the temporal hierarchical relationship between each power station based on the disaster-affected temporal hierarchy; using the degree of disaster-affected stability as the initial hippopotamus status evaluation criterion, establishing the correlation mapping between the initial hippopotamus status and the disaster type and the temporal hierarchical relationship, and obtaining the initial hippopotamus status influence relationship for each power station.

[0009] Further, the construction of the initial river position influence relationship for each disaster source based on the disaster source and the terrain type according to historical emergency status data includes: obtaining the movement path and the influence range of the disaster source from the historical emergency status data; using the movement path of the disaster source as the river path, establishing the correlation mapping between the initial river position and the influence range and the terrain type, and obtaining the initial river position influence relationship for each disaster source.

[0010] Further, the construction of the power grid emergency optimization objective function based on the hippopotamus optimization algorithm architecture according to the highest participation complexity and the lowest damage probability includes: calculating the participation complexity according to the number of power station types, and calculating the damage probability according to the entropy increase of the power station affected by the disaster.

[0011] Further, obtaining the state transition relationship based on the historical emergency status data and the historical flat valley status data, and constructing the power grid flat valley optimization objective function based on the state transition relationship and the minimum flat valley fluctuation includes: obtaining the historical flat valley status data of the disaster latent time period according to the historical emergency status data, extracting the conversion dimension according to the historical flat valley status data of the disaster latent time period, and constructing the state transition relationship based on the conversion dimension and the disaster latent time period.

[0012] Further, the construction of the power grid flat valley optimization objective function based on the state transition relationship and the minimum flat valley fluctuation includes: calculating the flat valley fluctuation based on the state transition relationship, the power grid load fluctuation, and the power supply range fluctuation.

[0013] Further, calling the initial hippopotamus status influence relationship and the initial river position influence relationship according to the current regional status data to obtain the initial hippopotamus status and the initial river position, and using the power grid emergency optimization objective function and the power grid flat valley optimization objective function to obtain the power grid dynamic programming strategy includes: obtaining the predicted disaster latent time period and the predicted disaster source according to the current regional status data and the state transition relationship; obtaining the initial hippopotamus status and the initial river position according to the predicted disaster source, the initial hippopotamus status influence relationship, and the initial river position influence relationship; obtaining the power grid emergency planning strategy according to the initial hippopotamus status and the initial river position based on the power grid emergency optimization objective function; obtaining the power grid flat valley planning strategy according to the predicted disaster latent time period, the power grid emergency planning strategy, and the power grid flat valley optimization objective function.

[0014] Further, using the disaster stability degree as the initial judgment criterion for the hippo status, establishing the correlation mapping between the initial hippo status and the disaster type and the time series level relationship, and obtaining the influence relationship of the initial hippo status of each power station includes: obtaining the disaster stability degree of all power stations under the same disaster type based on the time series level relationship, and performing clustering analysis based on the disaster stability degree to obtain the outliers and the clustering center. Taking the power station corresponding to the outlier as the little hippo, taking the power station corresponding to the highest disaster stability degree as the hippo leader, taking the power stations corresponding to the clustering centers except the hippo leader as the male hippos, and taking the power stations corresponding to the clustering points around the clustering center as the female hippos.

[0015] Further, the performing clustering analysis based on the disaster stability degree to obtain the outliers and the clustering center further includes: performing a first clustering on all power stations based on the environmental similarity and the power station type, and performing a second clustering on the results of the first clustering based on the disaster stability degree to obtain the outliers and the clustering center of each clustering result.

[0016] Further, the performing a first clustering on all power stations based on the environmental similarity and the power station type includes: calculating the longitude distance and the latitude distance of the power stations respectively based on the distance algorithm; obtaining the longitude distance weight, the latitude distance weight of the power stations and the latitude weight of the interval area type of the power stations according to the disaster type and the power station type; obtaining the environmental similarity according to the longitude distance weight, the latitude distance weight, the latitude weight of the interval area type of the power stations, the longitude distance, the latitude distance and the interval area type of the power stations.

[0017] Another technical solution provided by this application is a power grid normal and emergency state dynamic programming system based on the hippo algorithm, which is used to implement the above method.

[0018] Another technical solution provided by this application is that a computer program or instruction is stored in the storage medium, and when the computer program or instruction is executed by a processing device, the above method is implemented.

[0019] Advantages of this application: By constructing the initial influence relationship of hippo status and the initial influence relationship of river position based on historical emergency state data, the differences in the disaster-affected stability of power stations under different disaster types and the movement or influence of disaster sources under different terrain types can be reflected. Furthermore, by considering the participation complexity of the overall power station during disasters and the impact of the damage probability on emergency power supply stability, a power grid emergency optimization objective function is constructed. The position update, resistance behavior, and escape behavior of hippos relative to rivers in the hippo optimization algorithm are used to simulate the relative position update between power stations and disaster sources, the defense behavior of power stations, and the shutdown behavior of power stations in the event of disasters, improving the simulation accuracy of the impact on actual power station behavior. At the same time, the adjustment of the time sequence of the conversion from the flat valley state to the emergency state is considered. According to the state conversion time sequence and the regulated state of the demand in the emergency state, the adjustment of the flat valley stage is carried out with the minimum flat valley fluctuation of the power grid as the goal, avoiding the problem of grid imbalance caused by too fast connection, ensuring that emergency power supply regulation can be achieved in a timely manner in the emergency state, and improving the timeliness and stability of regulation. Description of the Drawings

[0020] Figure 1 It is a flowchart of the power grid flat and emergency state dynamic programming method based on the hippo algorithm of this application.

[0021] Figure 2 It is a schematic diagram of the initial hippo status in an embodiment of this application. Detailed Embodiment

[0022] To make the purpose, technical solution, and advantages of this application clearer, the following further elaborates on this application in combination with the drawings and embodiments. It should be understood that the specific embodiment described here is only the best embodiment of this application, only used to explain this application, and does not limit the protection scope of this application. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of this application.

[0023] As Figure 1 shown, as Embodiment 1 of this application, the power grid flat and emergency state dynamic programming method based on the hippo algorithm includes the following steps: Construct the initial influence relationship of hippo status for each power station based on historical emergency state data according to the disaster-affected stability and disaster type; Construct the initial influence relationship of river position for each disaster source based on historical emergency state data according to the disaster source and terrain type; Construct the power grid emergency optimization objective function based on the highest participation complexity and the lowest damage probability according to the hippo optimization algorithm framework; Obtain the state transition relationship based on historical emergency state data and historical valley - flat state data, and construct the grid valley - flat optimization objective function based on the state transition relationship and the minimum valley - flat fluctuation; Call the initial hippopotamus status influence relationship and the initial river position influence relationship according to the current regional state data to obtain the initial hippopotamus status and the initial river position, and use the grid emergency optimization objective function and the grid valley - flat optimization objective function to obtain the grid dynamic programming strategy.

[0024] In this embodiment, the initial hippopotamus status influence relationship and the initial river position influence relationship are constructed through historical emergency state data, so as to reflect the differences in the disaster - affected stability of power stations under different disaster types and the movement or influence of disaster sources under different terrain types. Furthermore, the grid emergency optimization objective function is constructed by considering the influence of the participation complexity and damage probability of the overall power station during disasters on emergency power supply stability. The relative position update between the power station and the disaster source, the defense behavior of the power station, and the shutdown behavior of the power station in the disaster - affected situation are simulated by using the position update, resistance behavior, and escape behavior of the hippopotamus relative to the river in the hippopotamus optimization algorithm, improving the simulation accuracy of the influence on the actual power station behavior. At the same time, the adjustment of the valley - flat state to the emergency state transition timing is considered. According to the state transition timing and the regulated state required in the emergency state, the adjustment of the valley - flat stage is carried out with the minimum grid valley - flat fluctuation as the goal, avoiding the problem of grid imbalance caused by too fast connection, ensuring that emergency power supply regulation can be realized in time in the emergency state, and improving the timeliness and stability of the regulation.

[0025] Specifically, the initial hippopotamus status influence relationship for each power station is constructed based on the disaster - affected stability and disaster type according to historical emergency state data, including: Extract the disaster type and disaster - affected stability of each power station from the historical emergency state data; Construct the time - series hierarchical relationship between each power station based on the disaster - affected time - series level; Taking the disaster - affected stability as the initial hippopotamus status evaluation criterion, establish the associated mapping between the initial hippopotamus status, the disaster type, and the time - series hierarchical relationship, and obtain the initial hippopotamus status influence relationship for each power station.

[0026] By obtaining the disaster - affected stability of each power station when an emergency state occurs in its own area, the resistance ability of each power station to different disaster types is obtained, and the difference in the disaster - affected degree between different power stations is constructed based on the difference in the disaster - affected time - series, and visualized through the hierarchical model, so as to allocate the initial hippopotamus status according to the behavior of the power station at different times and different disaster types.

[0027] Specifically, in the Hippopotamus Optimization Algorithm, the hippopotamus population consists of female hippopotamuses, baby hippopotamuses, male hippopotamuses, and the hippopotamus leader. Baby hippopotamuses often show a tendency to deviate from the group, while female hippopotamuses are distributed around the male. The position update of female hippopotamuses explores new areas while maintaining a certain distance from the dominant solution. After being expelled by the hippopotamus leader, male hippopotamuses need to establish a new group by competing or attracting females, so as to balance global search and local search. The hippopotamus leader is the individual with the best value in the current iteration. As Figure 2 shown, taking the disaster-affected stability degree as the initial evaluation criterion for hippopotamus status, establishing the correlation mapping between the initial hippopotamus status and the disaster type and the time-series hierarchical relationship, and obtaining the influence relationship of the initial hippopotamus status of each power station includes: Based on the time-series hierarchical relationship, obtain the disaster-affected stability degree of all power stations under the same disaster type, and perform clustering analysis based on the disaster-affected stability degree to obtain outliers and cluster centers. Take the power station corresponding to the outlier as the baby hippopotamus, the power station corresponding to the highest disaster-affected stability degree as the hippopotamus leader, the power stations corresponding to the cluster centers except the hippopotamus leader as male hippopotamuses, and the power stations corresponding to the cluster points around the cluster centers as female hippopotamuses.

[0028] In this embodiment, clustering analysis is performed through the disaster-affected stability degree of power stations at different times and under different disaster types, and different cluster centers are obtained through the clustering results of different disaster-affected stability degrees. The form of cluster points surrounding the cluster centers is used to simulate the form of female hippopotamuses surrounding male hippopotamuses. The power station with the highest disaster-affected stability degree is defined as the initial optimal solution and defined as the initial hippopotamus leader. The power stations that deviate from the cluster center, that is, the power stations whose disaster-affected stability degree does not conform to the other power stations, are used as baby hippopotamuses, thus showing the state of baby hippopotamuses being out of the group. By defining the initial hippopotamus status of power stations according to the disaster-affected stability degree, while increasing the search dimension in the iterative optimization process, the exploration time of the algorithm iteration can be reduced according to the initial hippopotamus status as the initial solution, and the accuracy and efficiency of power station emergency planning are improved by dynamically simulating the power station state changes using hippopotamus behavior.

[0029] In some other embodiments, the power stations can also be clustered once according to the environmental similarity and / or power station type, and then the different states of each power station can be further refined according to the results of the first clustering and the disaster-affected stability degree.

[0030] In some cases, performing clustering analysis based on the disaster-affected stability degree to obtain outliers and cluster centers further includes: Performing a first clustering on all power stations based on the environmental similarity, and performing a second clustering on the results of the first clustering based on the disaster-affected stability degree to obtain the outliers and cluster centers of each clustering result.

[0031] In this embodiment, through the environmental similarity between power stations, each power station is first clustered once, and then the clustering analysis results after the first clustering are subjected to secondary clustering to obtain the clustering situation corresponding to the disaster-affected stability degree of power stations in similar environments, so as to incorporate the environmental dimension into the consideration of the initial definition of the Hippo status. When the disaster-affected stability degree in a similar environment deviates from the large group, it is considered that the power station is far from the group.

[0032] In some other cases, performing clustering analysis based on the disaster-affected stability degree to obtain outliers and cluster centers further includes: Performing a first clustering on all power stations based on the power station type, and performing a second clustering on the results of the first clustering based on the disaster-affected stability degree to obtain the outliers and cluster centers of each clustering result.

[0033] In this case, a first clustering is performed on all power stations by power station type, and then a second clustering is performed on each set of power station types, so as to show the different response states of different power stations to different disaster types, avoid the deviation of the selection of outliers caused by the different behaviors of different types of power stations, further improve the accuracy of the initial definition of the Hippo status, reduce the number of subsequent optimization iterations, and improve the accuracy of the output of the emergency planning strategy.

[0034] In still some other cases, performing clustering analysis based on the disaster-affected stability degree to obtain outliers and cluster centers further includes: Performing a first clustering on all power stations based on the environmental similarity and power station type, and performing a second clustering on the results of the first clustering based on the disaster-affected stability degree to obtain the outliers and cluster centers of each clustering result.

[0035] Among them, the environmental similarity is calculated according to the power station longitude and latitude and the type of the area between power stations: ; ; ; ; ; Among them, represents the environmental similarity; represents the longitude and latitude similarity weight; represents the weight of the type of the area between power stations; represents the distance between the longitudes and latitudes of power stations; represents the distance of the type of the area between power stations; represents the relative position relationship between power station i and power station j; represents the difference in latitude radians between power station i and power station j, ; represents the latitude radian of power station i; Represents the latitude radian of power station j; Represents the difference in longitude radian between power station i and power station j, , Represents the longitude radian of power station j; Represents the longitude radian of power station i; Represents the radius of the earth; Represents the terrain height of the g-th interval area; Represents the terrain height of power station i; G represents the total number of interval areas.

[0036] The disaster-affected stability degree is calculated based on the disaster-affected proportion, average recovery time, and power supply: ; ; ; ; Among them, C represents the disaster-affected stability degree; Represents the weight of the disaster-affected proportion, Represents the weight of the average recovery time, Represents the weight of the power supply; Represents the disaster-affected proportion; Represents the average recovery time, Represents the n-th disaster-affected recovery time, and N represents the total number of disaster-affected recoveries; Represents the power supply.

[0037] , , , , Can be obtained through expert experience or calculated by hierarchical analysis based on historical data. In this embodiment, = 0.6, = 0.4, = 0.5, = 0.2, = 0.3.

[0038] Perform a clustering on all power stations based on environmental similarity and power station type, initially divide the power stations according to their locations and types, improve the sensitivity of outlier detection, and simulate the behavior of female hippos surrounding male hippos to reflect the synergistic effect of power stations of the same type, improving the exploration efficiency and accuracy in the optimization calculation process.

[0039] In some other embodiments, calculating the environmental similarity based on the power station longitude and latitude and the power station interval area type further includes: Calculating the power station longitude distance and the power station latitude distance respectively based on the distance algorithm; Obtain the power station longitude distance weight, power station latitude distance weight, and power station interval area type latitude weight according to the disaster type and power station type; Obtain the environmental similarity according to the power station longitude distance weight, power station latitude distance weight, power station interval area type latitude weight, power station longitude distance, power station latitude distance, and power station interval area type.

[0040] In this embodiment, the targeted calculation of environmental similarity is carried out by using the environmental impact differences between longitude and latitude. For example, in terms of climate, different latitudes receive different amounts of solar radiation, and different longitudes are affected by factors such as atmospheric circulation, resulting in humidity differences. Since distributed energy is affected differently, such as photovoltaic power stations being more susceptible to longitude differences and wind power stations being more susceptible to atmospheric circulation factors, and different disaster types are also affected by longitude and latitude differences. And the degree of weakening or strengthening of different disaster types affected by the interval area type is different. For example, the weakening ability of a typhoon is worse when the interval area type is a plain compared to when the interval area type is a hilly area. Therefore, the power station longitude distance weight, power station latitude distance weight, and power station interval area type latitude weight are calculated according to the power station type and disaster type. For example, the weakening or strengthening amplitude of different disasters in different interval area types can be obtained according to the disaster weakening degree in historical disaster data, and the power supply difference of different power station types under different longitudes and latitudes can be used to calculate the power station longitude distance weight, power station latitude distance weight, and power station interval area type latitude weight corresponding to the disaster type and power station type. Through targeted calculation, the environmental similarity is more in line with the actual situation, reducing the impact of the overall environment on the local environment and improving the result adaptability.

[0041] Construct the initial river position influence relationship for each disaster source based on the disaster source and terrain type according to historical emergency state data, including: Obtain the movement path and influence range of the disaster source from the historical emergency state data; Take the movement path of the disaster source as the river path, establish the association mapping between the initial river position, influence range, and terrain type, and obtain the initial river position influence relationship for each disaster source.

[0042] In this embodiment, the movement path of the disaster source is simulated as a river path. Based on the river path, an associated mapping with the influence range and terrain type is established. During this process, the terrain type can be thermally encoded for digital processing. In the hippopotamus optimization algorithm, the river attributes are mainly manifested as attraction attributes and resistance attributes. In this embodiment, the movement of the hippopotamus towards the river position is converted into an update of the relative position. The situation where the disaster source is blocked by the terrain is mapped to the resistance attribute of the river, and the situation where the power station is affected by the disaster source is mapped to the attraction attribute of the river, so as to realize the position update of the disaster source according to the pre-acquired movement path. At each update time sequence, according to the direction, distance of the movement path and the "resistance" influence brought by the terrain type, the movement distance and direction of the river are adjusted to realize the prediction of the movement of the disaster source. After the river position is updated, the relative positions of each hippopotamus and the river are immediately calculated, and the initial hippopotamus status and the initial river position are obtained through the initial hippopotamus status influence relationship and the initial river position influence relationship, reducing the number of iterative calculations and improving the calculation efficiency.

[0043] Based on the hippopotamus optimization algorithm framework, a power grid emergency optimization objective function is constructed according to the highest participation complexity and the lowest damage probability, including: Calculate the participation complexity according to the number of power station types, and calculate the damage probability according to the entropy increase situation of the power station affected by the disaster.

[0044] Due to the differences in the operating principles, equipment compositions and disaster resistance characteristics of different types of power stations, when there are more types of power stations in the power grid, due to the differences in aspects such as the disaster response mechanism and equipment tolerance, they will not be forced to shut down due to the same disaster factor at the same time, that is, the probability of synchronous shutdown of power stations decreases. When there are more different types of power stations, the probability of synchronous shutdown of power stations is lower, and thus the overall resistance ability is stronger. The entropy increase situation of the power station affected by the disaster is used to calculate the damage probability, that is, the damage probability of the current power station is calculated through the entropy increase influence of the power station's resistance action on the power station components and the entropy increase influence of the disaster source on the power station components.

[0045] Calculating the damage probability according to the entropy increase situation of the power station affected by the disaster includes: Obtain the historical action data of the power station, and calculate the component entropy increase influence relationship corresponding to the power station conversion action according to the historical action data of the power station; Obtain the historical environmental data of the power station, and calculate the environmental entropy increase influence relationship according to the historical environmental data of the power station; Calculate the damage probability based on the component entropy increase influence relationship and the environmental entropy increase influence relationship.

[0046] During the iterative calculation process of the hippopotamus optimization algorithm, calculate the environmental impact caused by the relative positions of each power station and the disaster source, calculate the environmental entropy increase influence relationship and the component entropy increase influence relationship according to the environmental impact, and calculate the damage probability of the current power station accordingly.

[0047] The objective function for power grid emergency optimization is as follows: ; ; Among them, represents the objective function for power grid emergency optimization, represents the number of power station types, represents the damage probability, represents the weight of the number of power station types, represents the weight of the damage probability.

[0048] The weight of the number of power station types and the weight of the damage probability can be calculated through expert experience or historical data. In some cases, the weight of the number of power station types and the weight of the damage probability are divided into two cases. When in a conservative situation, take = 0.3, = 0.7. When in an aggressive strategy, that is, when trying to ensure power supply as much as possible = 0.6, = 0.4.

[0049] Obtain the state transition relationship based on historical emergency state data and historical valley - flat state data. The power grid valley - flat optimization objective function constructed based on the state transition relationship and the minimum valley - flat fluctuation includes: Obtain the historical valley - flat state data of the disaster latent time segment according to the historical emergency state data. Extract the conversion dimension according to the historical valley - flat state data of the disaster latent time segment, and construct the state transition relationship based on the conversion dimension and the disaster latent time segment.

[0050] In this embodiment, historical flat valley state data of the disaster latent time series segment is obtained based on historical emergency state data. For example, when gradually occurring disasters such as floods and droughts occur, there must be a time series that is not in the emergency state, but its environmental characteristics are coherent with those of the emergency state. For example, before the flood emergency state occurs, there must be a flat valley time series of rainfall. Thus, this time series is used as the disaster latent time series segment to obtain the corresponding flat valley state data. Then, conversion dimensions are extracted based on the historical flat valley state data. For example, in this embodiment, the conversion dimensions are rainfall, river water level depth, humidity, wind force, ice coating thickness, temperature, and the corresponding maintenance time, and a mapping relationship, that is, a state conversion relationship, among the conversion dimensions, the disaster latent time series segment, and the disaster source is constructed. At this time, according to the state conversion relationship, the disaster latent time series segment and the disaster source matching the current dimension data are retrieved, and used as the target time series for the current flat valley state adjustment and the disaster source for emergency adjustment. For example, when there is an average rainfall of 30 mm and a maintenance time of 6 h in a certain area, and a certain water level depth exceeds the warning depth, it is considered that a flood disaster will occur. The possible occurrence time of the flood disaster is calculated based on the current data, and the position where the water level depth exceeds the warning depth is used as the disaster source for pre-planning of emergency state regulation and pre-regulation warning of the flat valley state.

[0051] Based on the state conversion relationship and the minimum flat valley fluctuation, the grid flat valley optimization objective function includes: Calculate the flat valley fluctuation based on the state conversion relationship, grid load fluctuation, and power supply range fluctuation.

[0052] Obtain the target time series for the current flat valley state adjustment according to the state conversion relationship, calculate the grid load fluctuation and power supply range fluctuation based on the target time series and the emergency planning strategy, and obtain the optimal adjustment strategy under the flat valley state.

[0053] The grid flat valley optimization objective function is: ; Wherein, represents the grid flat valley optimization objective function, Q represents the total number of regulated power stations output by the grid emergency planning strategy, represents the load fluctuation caused by the adjustment of power station q when the target time series is , represents the range fluctuation caused by the adjustment of power station q when the target time series is .

[0054] In this embodiment, the range fluctuation refers to the power supply range fluctuation during the power station adjustment process. When the power station adjusts towards the state output by the emergency planning strategy in the flat valley state, both the load fluctuation situation during the power station adjustment is considered, and the power supply range problem during the power station adjustment is minimized as much as possible. Under normal conditions, the power supply range of each power station usually has a certain stability. When there is a power station adjustment in the emergency state, there are not only adjustments such as power station power and shutdown, but also the situation of emergency power supply with a stable power station. At this time, the power supply range also has corresponding adjustments. During the conversion process from the flat valley state to the emergency state, it cannot be ensured that the emergency state will definitely occur. Therefore, a large-scale power supply range adjustment will instead cause a decline in the user experience. Especially in abnormal weather such as extremely cold and extremely hot, users' demand for power supply stability is stronger. Therefore, the optimal adjustment strategy in the flat valley state is output according to the two-way adjustment of the load fluctuation and the power supply range, so that the flat valley adjustment can not only reduce the load fluctuation caused by the emergency switching of the emergency adjustment, but also ensure normal power consumption in the flat valley state as much as possible. In some cases, the importance of the load fluctuation and the power supply range can also be adjusted by using the weight coefficient. For example, when the power station equipment is vulnerable to emergency switching and damage and there are a large number of flexible loads in the current power supply area that can be used as the adjustment range, the weight coefficient of the load fluctuation is greater than the weight coefficient of the power supply range. Of course, different weight coefficients can also be constructed for each power station equipment and then comprehensively calculated to improve the accuracy.

[0055] Call the initial hippopotamus position influence relationship and the initial river position influence relationship according to the current regional state data to obtain the initial hippopotamus position and the initial river position, and use the power grid emergency optimization objective function and the power grid flat valley optimization objective function to obtain the power grid dynamic programming strategy, including: Obtain the predicted disaster latent time sequence and the predicted disaster source according to the current regional state data and the state conversion relationship; Obtain the initial hippopotamus position and the initial river position according to the predicted disaster source, the initial hippopotamus position influence relationship and the initial river position influence relationship; Based on the initial hippopotamus position and the initial river position, obtain the power grid emergency planning strategy according to the power grid emergency optimization objective function; Obtain the power grid flat valley planning strategy according to the predicted disaster latent time sequence, the power grid emergency planning strategy and the power grid flat valley optimization objective function.

[0056] In this embodiment, the state data at least includes environmental data, terrain data and power station operation data. The power station operation data at least includes the power supply amount, component damage data and power supply topology structure. The predicted disaster latent time sequence and the predicted disaster source are retrieved according to the matching result of the current regional state data and the state conversion relationship, and the initial hippopotamus position and the initial river position are obtained according to the matching result of the predicted disaster source and the initial hippopotamus position influence relationship and the initial river position influence relationship.

[0057] Based on the initial hippopotamus status and the initial river position, perform hippopotamus optimization iterative calculation according to the power grid emergency optimization objective function, including the following steps: Initialize the hippopotamus population, and assign an initial position to each hippopotamus individual according to the initial river position and the power station position; Calculate the fitness value of each hippopotamus according to the power grid emergency optimization objective function and the initial hippopotamus status, and update the hippopotamus status; When a hippopotamus individual enters the periphery of the disaster source, trigger an escape behavior (power station power outage avoidance); When there is no significant change in the Pareto front for multiple consecutive generations, or the maximum number of iterations Tmax is reached, terminate the iteration and output the Pareto optimal solution set.

[0058] Among them, the Pareto optimal solution set is the power grid emergency planning strategy. In this embodiment, a power supply quantity constraint and a power supply range constraint are constructed for the power grid emergency optimization objective function. The power supply quantity constraint is obtained according to the minimum demand power quantity of the current region, and the power supply range constraint is obtained according to the current disaster prevention area, that is, the power grid emergency planning needs to meet the minimum demand power quantity of the current region and at least meet the power supply demand of the disaster prevention area.

[0059] After outputting the power grid emergency planning strategy according to the predicted disaster source, obtain the predicted disaster latent time sequence according to the predicted disaster source, and use the power grid valley optimization objective function to output a valley planning strategy that can ensure the conversion to the power grid emergency planning strategy, minimize the switching load fluctuation, and minimize the impact on the current valley power supply under the predicted disaster latent time sequence. Adjust the power station in advance before entering the emergency state to avoid power grid fluctuations caused by emergency switching. At the same time, continuously update the predicted river position through the influence relationship of the initial river position, and output the corresponding power grid emergency planning strategy to adapt to the uncertainty of disasters. It can be understood that when entering the current emergency state, the power grid valley optimization objective function is no longer called for dynamic optimization of the valley state, and all computing resources are concentrated on the emergency optimization calculation, and the influence relationship of the initial hippopotamus status and the influence relationship of the initial river position are used to reduce the calculation difficulty or random inaccuracy of the initial solution, and further improve the planning efficiency.

[0060] As the second embodiment of the present application, a power grid normal and emergency state dynamic planning system based on the hippopotamus algorithm includes: A data acquisition unit for acquiring status data, where the status data at least includes historical emergency state data, historical valley state data, and current regional state data; An initial solution calculation unit for constructing an initial hippopotamus status influence relationship and an initial river position influence relationship based on the historical emergency state data according to the disaster-affected stability degree, disaster type, disaster source, and terrain type; Optimization construction unit, used to construct the power grid emergency optimization objective function and the power grid flat valley optimization function based on the hippopotamus optimization algorithm architecture; Dynamic programming unit, used to output the power grid dynamic programming strategy according to the initial hippopotamus status influence relationship, the initial river position influence relationship, the power grid emergency optimization objective function, and the power grid flat valley optimization function.

[0061] In this embodiment, the data acquisition unit is connected to the initial solution calculation unit, the optimization construction unit, and the dynamic programming unit. The initial solution calculation unit and the optimization construction unit are connected to the dynamic programming unit. The dynamic programming unit receives the current regional status data from the data acquisition unit, calls the hippopotamus status influence relationship and the initial river status influence relationship to obtain the initial hippopotamus status and the initial river status, calculates the power grid emergency planning strategy according to the power grid emergency optimization objective function, and outputs the flat valley planning strategy under the predicted disaster latent time sequence of the flat valley state according to the power grid emergency planning strategy and the power grid flat valley optimization function, realizing pre-adjustment and calculation pressure differentiation, and improving the accuracy and efficiency of emergency planning.

[0062] Data acquisition unit, used to obtain status data, where the status data at least includes historical emergency status data, historical flat valley status data, and current regional status data; Initial solution calculation unit, used to construct the initial hippopotamus status influence relationship and the initial river position influence relationship based on the historical emergency status data according to the disaster stability degree, disaster type, disaster source, and terrain type; Optimization construction unit, used to construct the power grid emergency optimization objective function and the power grid flat valley optimization function based on the hippopotamus optimization algorithm architecture; Dynamic programming unit, used to output the power grid dynamic programming strategy according to the initial hippopotamus status influence relationship, the initial river position influence relationship, the power grid emergency optimization objective function, and the power grid flat valley optimization function.

[0063] As the third embodiment of the present application, a computer-readable storage medium is used to store computer programs or instructions. When the computer programs or instructions are executed by a processing device, the above-mentioned power grid flat and emergency state dynamic programming method based on the hippopotamus algorithm is realized. The computer-readable storage medium can be any available medium that a computing device can store or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid-state drive), etc.

[0064] The above-described specific implementation manners are the preferred implementation manners of the power grid normal and emergency state dynamic programming method and system based on the hippopotamus algorithm of the present application, and do not limit the specific implementation scope of the present application. The scope of the present application includes but is not limited to this specific implementation manner. Any equivalent changes made according to the shape and structure of the present application are within the protection scope of the present application.

Claims

1. A dynamic programming method for power grid normal and emergency states based on the hippopotamus algorithm, characterized in that: It includes the following steps: Based on historical emergency state data, construct the initial Hippo status influence relationship for each power station according to the disaster stability degree and disaster type; Based on historical emergency state data, construct the initial river position influence relationship for each disaster source according to the disaster source and terrain type; Based on the Hippo optimization algorithm framework, construct the power grid emergency optimization objective function according to the highest participation complexity and the lowest damage probability; Based on historical emergency state data and historical flat valley state data, obtain the state transition relationship, and construct the power grid flat valley optimization objective function based on the state transition relationship and the minimum flat valley fluctuation; According to the current regional state data, call the initial Hippo status influence relationship and the initial river position influence relationship to obtain the initial Hippo status and the initial river position, and use the power grid emergency optimization objective function and the power grid flat valley optimization objective function to obtain the power grid dynamic programming strategy.

2. The power grid flat and emergency state dynamic programming method based on the Hippo algorithm according to claim 1, characterized in that: The step of constructing the initial Hippo status influence relationship for each power station based on historical emergency state data according to the geographical location and disaster type includes: Extract the disaster type and disaster stability degree of each power station from the historical emergency state data; Construct the time sequence level relationship between each power station based on the disaster time sequence level; Taking the disaster stability degree as the initial Hippo status evaluation criterion, establish the correlation mapping between the initial Hippo status and the disaster type and the time sequence level relationship, and obtain the initial Hippo status influence relationship for each power station.

3. The power grid flat and emergency state dynamic programming method based on the Hippo algorithm according to claim 1, characterized in that: The step of constructing the initial river position influence relationship for each disaster source based on historical emergency state data according to the disaster source and terrain type includes: Obtain the moving path and influence range of the disaster source from the historical emergency state data; Taking the moving path of the disaster source as the river path, establish the correlation mapping between the initial river position and the influence range and the terrain type, and obtain the initial river position influence relationship for each disaster source.

4. The power grid flat and emergency state dynamic programming method based on the Hippo algorithm according to claim 1, characterized in that: The step of constructing the power grid emergency optimization objective function based on the Hippo optimization algorithm framework according to the highest participation complexity and the lowest damage probability includes: Calculate the participation complexity according to the number of power station types, and calculate the damage probability according to the entropy increase situation of power station disasters.

5. The power grid flat and emergency state dynamic programming method based on the Hippo algorithm according to claim 1, characterized in that: The step of obtaining the state transition relationship based on historical emergency state data and historical flat valley state data, and constructing the power grid flat valley optimization objective function based on the state transition relationship and the minimum flat valley fluctuation includes: Obtain the historical flat valley state data of the disaster latent time sequence segment according to the historical emergency state data, extract the conversion dimension according to the historical flat valley state data of the disaster latent time sequence segment, and construct the state transition relationship based on the conversion dimension and the disaster latent time sequence segment.

6. The power grid flat and emergency state dynamic programming method based on the Hippo algorithm according to claim 1, characterized in that: The grid valley optimization objective function constructed based on the state transition relationship and the minimum valley - flat fluctuation includes: Calculate the valley - flat fluctuation based on the state transition relationship, grid load fluctuation, and power supply range fluctuation.

7. The grid normal - emergency state dynamic programming method based on the hippopotamus algorithm according to claim 1, characterized in that: The method of obtaining the initial hippopotamus position and the initial river position by calling the initial hippopotamus position influence relationship and the initial river position influence relationship according to the current regional state data, and obtaining the grid dynamic programming strategy by using the grid emergency optimization objective function and the grid valley optimization objective function includes: Obtain the predicted disaster latent time sequence and the predicted disaster source according to the current regional state data and the state transition relationship; Obtain the initial hippopotamus position and the initial river position according to the predicted disaster source, the initial hippopotamus position influence relationship, and the initial river position influence relationship; Obtain the grid emergency planning strategy based on the initial hippopotamus position and the initial river position according to the grid emergency optimization objective function; Obtain the grid valley planning strategy according to the predicted disaster latent time sequence, the grid emergency planning strategy, and the grid valley optimization objective function.

8. The grid normal - emergency state dynamic programming method based on the hippopotamus algorithm according to claim 2, characterized in that: The method of establishing the correlation mapping between the initial hippopotamus position and the disaster type and time - sequence level relationship with the disaster - affected stability degree as the initial hippopotamus position evaluation criterion, and obtaining the initial hippopotamus position influence relationship of each power station includes: Obtain the disaster - affected stability degree of all power stations under the same disaster type based on the time - sequence level relationship, and perform clustering analysis based on the disaster - affected stability degree to obtain outliers and cluster centers. Take the power station corresponding to the outlier as the little hippopotamus, the power station corresponding to the highest disaster - affected stability degree as the hippopotamus leader, the power stations corresponding to the cluster centers except the hippopotamus leader as male hippopotamuses, and the power stations corresponding to the cluster points around the cluster centers as female hippopotamuses.

9. The grid normal - emergency state dynamic programming method based on the hippopotamus algorithm according to claim 8, characterized in that: The method of performing clustering analysis based on the disaster - affected stability degree to obtain outliers and cluster centers further includes: Perform a first - level clustering on all power stations based on environmental similarity and power station type, and perform a second - level clustering on the results of the first - level clustering based on the disaster - affected stability degree to obtain outliers and cluster centers of each clustering result.

10. The grid normal - emergency state dynamic programming method based on the hippopotamus algorithm according to claim 9, characterized in that: The method of performing a first - level clustering on all power stations based on environmental similarity and power station type includes: Calculate the longitude distance and latitude distance of power stations respectively based on the distance algorithm; Obtain the longitude distance weight of the power station, the latitude distance weight of the power station, and the latitude weight of the area type between power stations according to the disaster type and power station type; Obtain the environmental similarity according to the longitude distance weight of the power station, the latitude distance weight of the power station, the latitude weight of the area type between power stations, the longitude distance of the power station, the latitude distance of the power station, and the area type between power stations.

11. A grid normal - emergency state dynamic programming system based on the hippopotamus algorithm, used to implement the method according to any one of claims 1 to 10, characterized in that: A data acquisition unit for obtaining status data, where the status data at least includes historical emergency status data, historical flat valley status data, and current regional status data; An initial solution calculation unit for constructing an initial Hippopotamus position influence relationship and an initial river position influence relationship based on the historical emergency status data according to the disaster-affected stability degree, disaster type, disaster source, and terrain type; An optimization construction unit for constructing a power grid emergency optimization objective function and a power grid flat valley optimization function based on the Hippopotamus optimization algorithm architecture; A dynamic programming unit for outputting a power grid dynamic programming strategy according to the initial Hippopotamus position influence relationship, the initial river position influence relationship, the power grid emergency optimization objective function, and the power grid flat valley optimization function.

12. A computer-readable storage medium, characterized in that: The storage medium stores a computer program or instruction, and when the computer program or instruction is executed by a processing device, the method described in any one of claims 1 to 10 is implemented.

Images