Intelligent power grid load balance control method and system

Abstract

The invention belongs to the technical field of intelligent power grids, and discloses an intelligent power grid load balance control method and system. The method comprises the steps of dividing a power grid into a plurality of regions based on a predetermined rule; based on the regional power supply quantity and the regional power consumption quantity, identifying an overload region with insufficient power supply quantity; formulating and executing a load balancing strategy for the overload area by adopting a natural heuristic optimization algorithm, wherein the load balancing strategy is a load scheduling strategy taking scheduling time minimization as a target; calculating the priority coefficient of the electric equipment in the overload area, and supplying power to each electric equipment according to the priority coefficient; effective balance and optimal allocation of intelligent power grid loads are achieved, emergency power supply is guaranteed to the maximum extent in natural disaster time, and therefore the power grid operation level and livelihood guarantee capacity are improved, and the power supply reliability and the power utilization efficiency of the power grid are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grids, and more particularly, to a smart grid load balancing control method and system. Background Art

[0002] A smart microgrid is a miniaturized, self-sufficient power network that integrates distributed power sources, energy storage systems, controllable loads, and energy management systems. Distributed power sources include renewable energy sources such as solar photovoltaics, wind power, and small hydropower. Energy storage systems store energy during periods of surplus power and release it during peak demand. Controllable loads can flexibly adjust power consumption based on grid conditions. Leveraging advanced big data analytics, artificial intelligence, and machine learning technologies, smart microgrids can accurately and efficiently allocate and utilize power resources, achieving a dynamic balance between supply and demand on the grid and significantly improving the overall efficiency and stability of the grid system.

[0003] The document with publication number CN118263843A discloses a method for energy-saving and carbon reduction control of regional distribution networks, including: establishing a regional distribution network model based on the operating data of the regional distribution network, and performing flow calculation on the regional distribution network model to obtain voltage and power distribution; performing sensitivity analysis based on the change data of voltage and power distribution, and selecting sensitive nodes as reactive compensation nodes; monitoring the voltage of the reactive compensation node in real time, and issuing reactive compensation instructions according to a preset reactive control strategy to suppress voltage fluctuations; determining a knife switch reconstruction plan based on the line loss minimization target, photovoltaic absorption maximization target, and load balancing target using distribution network reconstruction software, and formulating and executing a knife switch operation plan based on the knife switch reconstruction plan to complete line reconstruction; improving the distribution network's adaptability to the intermittent characteristics of photovoltaic power generation, enhancing its absorption capacity for distributed photovoltaics, and realizing voltage control coordination within the region to ensure efficient, stable and safe operation of the distribution network.

[0004] However, existing smart microgrid load balancing control technologies mostly focus on minimizing line losses and maximizing photovoltaic absorption, and often ignore the key factor of dispatch time. In situations where minimizing dispatch time is necessary, such as in areas where natural disasters (such as earthquakes, mudslides, typhoons, etc.) or man-made disasters (such as fires, etc.) occur suddenly, the demand for the operation of various emergency facilities, such as hospitals, fire stations, and rescue centers, increases, leading to a significant increase in electricity demand. In other words, additional electricity is needed to maintain the normal operation of medical equipment, communication systems, lighting, and other equipment so that rescue work can proceed smoothly. In order to achieve rapid rescue and protect life safety, the focus of load balancing should be to ensure that rescue operations and the operation of key facilities can be supported quickly and effectively. In other words, load balancing should fully consider the issue of minimizing dispatch time.

[0005] In view of this, the present invention proposes a smart grid load balancing control method and system to solve the above problems. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned objectives, the present invention provides the following technical solution: a smart grid load balancing control method, comprising:

[0007] Dividing the power grid into multiple areas based on predetermined rules;

[0008] Identify overloaded areas with insufficient power supply based on regional power supply and regional power consumption;

[0009] A load balancing strategy is formulated and executed for overloaded areas using a nature-inspired optimization algorithm. The load balancing strategy is a load scheduling strategy that aims to minimize scheduling time.

[0010] Calculate the priority coefficient for electrical equipment in the overload area and supply power to each electrical equipment according to the priority coefficient.

[0011] Furthermore, the steps of formulating a load balancing strategy include:

[0012] Step 1: Randomly generate a set of solutions, which include multiple scheduling sets, set set labels for the scheduling sets, and preset iteration thresholds;

[0013] Step 2: Define the fitness function and use the scheduling time as the evaluation indicator;

[0014] Step 3: Iterate and optimize the solution through cloning, mutation, and recombination operations. When the preset number of iterations is reached, filter the set label with the largest fitness and obtain the corresponding scheduling set based on the set label.

[0015] Furthermore, the method for identifying an overloaded area with insufficient power supply includes:

[0016] An overloaded area refers to a regional state when the regional power supply is less than the regional power consumption;

[0017] The sum of electricity consumption in each region is the regional electricity consumption; the sum of power generation in each region is the regional power generation; the sum of energy storage in each region is the regional energy storage; the sum of regional power generation and regional energy storage in each region is the regional power supply; and energy storage is the amount of electricity stored in each energy storage unit in the region.

[0018] Furthermore, in step 1, except for the overloaded area, the energy storage units in all areas are marked as dispatching units; all dispatching units are randomly combined, and the result of the combination is taken as a unit set; the energy storage of the dispatching units in each unit set is sequentially added to obtain the collective energy storage; the regional power consumption in the overloaded area is marked as overloaded power consumption, the collective energy storage of each unit set is compared with the overloaded power consumption, and the unit sets whose collective energy storage is greater than or equal to the overloaded power consumption are retained and marked as deployment sets;

[0019] Subtract the overload power consumption from the collective energy storage of each deployment set to obtain the energy storage difference; compare the energy storage of the dispatching units in each deployment set with the energy storage difference in turn, retain the deployment sets in which the energy storage of each dispatching unit is greater than or equal to the energy storage difference, and mark them as dispatching sets; set a different digital label for each dispatching set and mark it as a set label.

[0020] Furthermore, in step 2, the method for obtaining the scheduling time includes:

[0021] Get all scheduling units in the scheduling set and mark them as analysis units;

[0022] Obtain the topological structure of the power grid and draw a topological map of the power grid, which includes nodes and edges; obtain the route between each analysis unit and each node in the overload area according to the topological map of the power grid; obtain the length of each transmission line in each route according to the topological map of the power grid; add the transmission line lengths corresponding to each route in sequence to obtain the total length of the line corresponding to each route; compare the total lengths of the lines corresponding to multiple routes of each analysis unit, and take the total length of the line with the smallest value as the shortest length of the corresponding analysis unit; divide the shortest length of each analysis unit by the speed of light to obtain the transmission time of each analysis unit; preset a start-up time set, which includes the start-up time of each energy storage unit in the power grid, and the start-up time is the time consumed by the energy storage unit from startup to the start of power transmission; obtain the start-up time corresponding to each analysis unit according to the start-up time set;

[0023] The transmission time of each analysis unit is added to the corresponding start time to obtain the scheduling time of each analysis unit; the scheduling time of each analysis unit is compared, and the scheduling time with the largest value is used as the scheduling time of the scheduling set.

[0024] Furthermore, step 3 also includes verifying whether the energy storage capacity of the new scheduling set meets the constraint conditions after mutation and reorganization. The constraint conditions are that the total energy storage capacity of the new scheduling set covers the overload power consumption and the energy storage capacity of the scheduling unit is greater than the energy storage difference.

[0025] Furthermore, the method for calculating the priority coefficient of the electrical equipment in the overload area includes:

[0026] Collect the power value of each electrical device in the overload area; compare the power value of each electrical device with 0, mark the electrical devices with power values greater than 0 as running devices, and do not mark the electrical devices with power values equal to 0; assign different digital labels to the electrical devices in the overload area and mark them as device labels; obtain the device label corresponding to each running device and mark it as the running label;

[0027] Collect regional images of overloaded areas, use the trained disaster recognition model to identify the regional images, and output the recognition results, which are digital labels corresponding to the natural disaster types; mark the digital labels corresponding to the natural disaster types as disaster labels; use a device label and a disaster label as a set of analysis data, input each set of analysis data into the trained priority analysis model, and predict the corresponding priority coefficient.

[0028] Furthermore, the method further comprises: collecting and analyzing the operating parameters of each electrical device in the overload area, screening out the electrical devices in an online state, and marking them as operating devices;

[0029] Operating parameters include power, temperature, and vibration. The device tag of each electrical device is obtained, and the device tag and operating parameters corresponding to each electrical device are used as a set of test data. Each set of test data is sequentially input into the trained state analysis model to predict the state tag corresponding to each electrical device. The state tag is the digital label corresponding to the device state. The device state includes running state, standby state, shutdown state, and fault state. Different device states have different corresponding digital labels.

[0030] The online status includes the running status and the standby status; the status tag corresponding to the running status is marked as the running tag, and the status tag corresponding to the standby status is marked as the standby tag; the running tag and the standby tag are used as the status tags corresponding to the online status; the status tag of each electrical device is compared with the status tag of the online status; the electrical devices with the same status tags as the running tag or the standby tag are filtered out and marked as running devices.

[0031] A smart grid load balancing control system implements a smart grid load balancing control method, including:

[0032] A microgrid partitioning module that divides the grid into multiple areas based on predetermined rules;

[0033] The overload determination module identifies overloaded areas with insufficient power supply based on regional power supply and regional power consumption;

[0034] A load balancing module is used to formulate and execute a load balancing strategy for overloaded areas using a nature-inspired optimization algorithm, wherein the load balancing strategy is a load scheduling strategy with the goal of minimizing scheduling time;

[0035] The sequential power supply module is used to calculate the priority coefficient of the electrical equipment in the overload area and supply power to each electrical equipment according to the priority coefficient.

[0036] The status analysis module is used to collect and analyze the operating parameters of each electrical device in the overload area, filter out the electrical devices in the online state, and mark them as operating devices.

[0037] The technical effects and advantages of the smart grid load balancing control method and system of the present invention are as follows:

[0038] 1. Through regional division and energy data collection, each area of the power grid is monitored in real time to accurately identify overloaded areas; taking full account of scheduling time factors, a clonal selection algorithm is used to formulate and implement load balancing strategies for overloaded areas; and based on the specific electrical equipment and natural disaster types in the overloaded area, the priority power supply sequence for each electrical equipment is calculated to effectively ensure priority power supply to life safety facilities; the effective balance and optimal allocation of smart grid loads are achieved, and emergency power supply is guaranteed to the greatest extent possible during natural disasters, thereby improving the power grid's operation level and people's livelihood protection capabilities, and improving the power supply reliability and power efficiency of the power grid.

[0039] 2. Adopt multi-source information fusion technology to fuse the multi-dimensional operating parameters collected from electrical equipment, such as power, temperature, and vibration values; through deep learning technology, complete the subdivision of the operating status, standby status, shutdown status, and fault status of electrical equipment, realize end-to-end identification of equipment status, and more accurately identify the actual equipment status of electrical equipment; avoid misjudgment of equipment status caused by a single judgment, thereby providing a more reliable basis for grid load scheduling, ensuring priority power supply for key emergency equipment, and improving emergency rescue efficiency and the reliability of grid load balancing control. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a schematic diagram of a smart grid load balancing control system according to embodiment 1 of the present invention;

[0041] Figure 2 This is a flow chart of the steps for formulating a load balancing strategy according to embodiment 1 of the present invention;

[0042] Figure 3 This is a schematic diagram of a smart grid load balancing control system according to embodiment 2 of the present invention;

[0043] Figure 4 This is a flow chart of the smart grid load balancing control method according to embodiment 3 of the present invention. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0045] Example 1:

[0046] See also Figure 1 and Figure 2 As shown, this embodiment provides a smart grid load balancing control system, including a microgrid partitioning module, an overload determination module, a load balancing module and a sequential power supply module; each module is connected by wired and / or wireless means to realize data transmission between modules.

[0047] The microgrid partitioning module divides the grid into multiple zones based on predefined rules. For example, based on geographic information and energy topology, it combines the main grid and microgrid into multiple autonomous units. This regionalization facilitates the collection of energy data from each zone and monitors load changes in each zone, enhancing load management capabilities. It also enables independent load analysis and control for each zone, achieving optimal load balance within the region.

[0048] The overload determination module identifies overloaded areas with insufficient power supply based on regional power supply and regional power consumption.

[0049] Methods for identifying overloaded areas with insufficient power supply include:

[0050] Energy data is collected in each area; energy data includes electricity consumption, power generation and energy storage; electricity consumption is the amount of electricity used by each electrical device in the area, and the electricity consumption is obtained by the smart meters installed on the regional electrical devices; power generation is the amount of electricity generated by each power generation unit in the area, and the power generation is obtained by the power sensor installed at the output end of the regional power generation unit; energy storage is the amount of energy stored in each energy storage unit in the area, and the energy storage is obtained by the built-in energy storage management system of each energy storage unit. The energy storage management system has a built-in battery management module for monitoring the power storage capacity of the energy storage unit.

[0051] The sum of electricity consumption in each region is the regional electricity consumption; the sum of power generation in each region is the regional power generation; the sum of energy storage in each region is the regional energy storage; the sum of regional power generation and regional energy storage in each region is the regional power supply; the regional power supply of each region is compared with the corresponding regional power consumption to determine whether each region is in an overloaded state. An overloaded area refers to a regional state when the regional power supply is less than the regional power consumption. Areas that are not in an overloaded state are not marked.

[0052] The load balancing module is used to formulate and execute a load balancing strategy for the overloaded area using a nature-inspired optimization algorithm to achieve load balance in the overloaded area. The load balancing strategy is a load scheduling strategy with the goal of minimizing scheduling time.

[0053] Naturally inspired optimization algorithms such as genetic algorithm, particle swarm optimization algorithm, simulated annealing algorithm, clonal selection algorithm, etc.; this embodiment takes the clonal selection algorithm as an example to introduce the formulation of load balancing strategy; Figure 2 As shown, the steps to develop a load balancing strategy include:

[0054] Step 1: Randomly generate a set of solutions, which include multiple scheduling sets, set set labels for the scheduling sets, and preset iteration thresholds;

[0055] Step 2: Define the fitness function and use the scheduling time as the evaluation indicator;

[0056] Step 3: Iterate and optimize the solution through cloning, mutation, and recombination operations. When the preset number of iterations is reached, filter the set label with the largest fitness and obtain the corresponding scheduling set based on the set label.

[0057] The energy storage units in all areas except the overloaded area are marked as dispatching units; all dispatching units are randomly combined, and the result of one combination is regarded as a unit set; the energy storage of the dispatching units in each unit set is added up in sequence to obtain the collective energy storage; the regional electricity consumption in the overloaded area is marked as overloaded electricity consumption, and the collective energy storage of each unit set is compared with the overloaded electricity consumption. The unit sets with collective energy storage greater than or equal to the overloaded electricity consumption are retained and marked as deployment sets; to ensure that the amount of electricity that can be dispatched by each deployment set during energy dispatch can meet the electricity demand of the overloaded area.

[0058] Subtract the overload power consumption from the collective energy storage of each deployment set to obtain the energy storage difference; compare the energy storage of the dispatching units in each deployment set with the energy storage difference in turn, retain the deployment sets in which the energy storage of each dispatching unit is greater than or equal to the energy storage difference, and mark them as dispatching sets; set a different digital label for each dispatching set and mark it as a set label.

[0059] For example, the overload power consumption is 15, and the collective energy storage of the first deployment set is 22. Then the energy storage difference of the first deployment set is 7, where the energy storage of each scheduling unit is 2, 4, 7, and 9, respectively. Since 2 and 4 are both less than 7, only energy storage units with energy storage of 7 and 9 are needed in the first deployment set to meet the power demand of the overload area, and energy storage units with energy storage of 2 and 4 are not needed. Therefore, the first deployment set is not marked as a scheduling set; the collective energy storage of the second deployment set is 20, then the energy storage difference of the second deployment set is 5, where the energy storage of each scheduling unit is 6, 7, and 7, respectively. Since 6, 7, and 7 are all greater than 5, all scheduling units in the second deployment set are needed to meet the power demand of the overload area, so the second deployment set is marked as a scheduling set.

[0060] It should be understood that the purpose of filtering out the scheduling set from the deployment set is to avoid the appearance of redundant scheduling units in the scheduling set, making the scheduling process more streamlined and efficient, reducing the complexity of the scheduling process, and improving the decision-making speed; and being able to maximize the use of existing resources to avoid resource waste and redundancy; at the same time, redundant scheduling units will cause unnecessary energy loss. By removing redundant scheduling units, energy loss can be reduced and the overall efficiency of the system can be improved.

[0061] Methods for obtaining scheduling time include:

[0062] Get all scheduling units in the scheduling set and mark them as analysis units;

[0063] Obtain the topology of the power grid and draw a grid topology map. The topology of the power grid is obtained by technical personnel in this field from the power grid company. The topology of the power grid includes nodes (such as substations, distribution cabinets, power generation units, etc.) and edges (such as transmission lines, distribution lines, etc.); obtain the route between each analysis unit and each node in the overload area according to the power grid topology map; obtain the length of each transmission line in each route according to the power grid topology map; add the transmission line lengths corresponding to each route in sequence to obtain the total line length corresponding to each route; compare the total line lengths of multiple routes corresponding to each analysis unit, and take the total line length with the smallest value as the shortest length of the corresponding analysis unit; divide the shortest length of each analysis unit by the speed of light to obtain the transmission time of each analysis unit.

[0064] A preset start-up time set includes the start-up time of each energy storage unit in the power grid. The start-up time is the time elapsed from the start-up of the energy storage unit to the start of power transmission. The start-up time set is obtained by technicians in this field by measuring the time elapsed from the start-up of each energy storage unit in the power grid to the start of power transmission during historical power grid load balancing. Based on the start-up time set, the start-up time corresponding to each analysis unit is obtained.

[0065] The transmission time of each analysis unit is added to the corresponding start time to obtain the scheduling time of each analysis unit; the scheduling time of each analysis unit is compared, and the scheduling time with the largest value is used as the scheduling time of the scheduling set.

[0066] It should be noted that the propagation speed of electric energy in high-voltage transmission lines is close to the speed of light, with very little difference, and there is no need to consider the subtle differences in propagation speed between different lines. Therefore, when calculating the scheduling time, the scheduling speed of electric energy is replaced by the speed of light.

[0067] In step 3 above, the cloning operation describes how, based on the fitness function (minimizing scheduling time), schedule sets with high fitness are selected for replication. Schedule sets with higher fitness (shorter scheduling time) are replicated more often. The mutation operation randomly adjusts some of the scheduling units in the cloned schedule set to increase population diversity. The recombination operation swaps some of the scheduling units of the two parent schedule sets to generate a new offspring. Through crossover recombination, the offspring explores new areas of the solution space while inheriting the excellent characteristics of the parent.

[0068] Constraint processing: After mutation and reorganization, it is necessary to verify whether the energy storage of the new scheduling set meets the constraints. The constraints are that the total energy storage of the new scheduling set must cover the overload power consumption, and the energy storage of the scheduling unit must be greater than the energy storage difference, ensuring that the algorithm efficiently optimizes the scheduling time under the physical limitations of the power grid.

[0069] The sequential power supply module is used to calculate the priority coefficient of the electrical equipment in the overload area and supply power to each electrical equipment according to the priority coefficient to achieve load balancing of each electrical equipment.

[0070] Methods for calculating the priority coefficient of electrical equipment in overload areas include:

[0071] The power value of each electrical device in the overload area is collected by the power sensor built into each device. The power value of each device is compared with 0. Devices with power values greater than 0 are marked as operating devices, while devices with power values equal to 0 are not marked. Since the power value must be greater than or equal to 0, the power value cannot be less than 0.

[0072] Different digital labels are set for electrical equipment in the overload area and marked as equipment labels; the equipment label corresponding to each running equipment is obtained and marked as the running label.

[0073] Collect regional images of overloaded areas. Regional images are obtained through satellite remote sensing technology. Satellite remote sensing technology uses satellite sensors (such as optical satellites, radar satellites, etc.) to detect and photograph various areas of the power grid to obtain large-scale high-resolution image data; use a trained disaster recognition model to identify the regional images and output the recognition results. The recognition results are digital labels corresponding to the natural disaster types, such as earthquakes, floods, forest fires, etc.

[0074] The specific training process of the disaster identification model includes:

[0075] Collect e regional images in advance, where e is an integer greater than 1, and annotate the natural disaster phenomena in each regional image as natural disaster types; convert different natural disaster types into digital labels, and they are inconsistent with the device labels. For example, earthquakes are converted to 1, floods are converted to 2, and forest fires are converted to 3; the annotated regional images are divided into training sets and test sets, with 70% of the regional images used as training sets and 30% of the regional images used as test sets; use the training set to train the disaster recognition model, and use the test set to test the disaster recognition model; preset an error threshold, and when the mean of the prediction errors of all regional images in the test set is less than the error threshold, output the disaster recognition model; wherein the calculation formula for the mean prediction error is: ,in is the mean prediction error, is the number of the region image, For the The predicted annotation corresponding to the region image, For the The actual annotation corresponding to the regional image is U, and U is the number of regional images in the test set; the error threshold is pre-set according to the accuracy required by the disaster identification model;

[0076] The above-mentioned disaster identification model is specifically a convolutional neural network model.

[0077] The digital labels corresponding to the natural disaster types are marked as disaster labels;

[0078] A device label and a disaster label are taken as a set of analysis data. Each set of analysis data is input into the trained priority analysis model to predict the corresponding priority coefficient.

[0079] The training process of the priority analysis model includes:

[0080] G groups of analysis data are collected in advance, and corresponding priority coefficients are set for each of the g groups of analysis data, where g is an integer greater than 1. The analysis data and the corresponding priority coefficients are converted into a corresponding set of feature vectors. The priority coefficients corresponding to the analysis data are collected by a person skilled in the art during the historical power grid load balancing process. Under the natural disaster conditions corresponding to the disaster labels corresponding to each group of analysis data, the corresponding priority coefficients are set according to the actual priority operation degree of the electrical equipment corresponding to the corresponding equipment labels. The corresponding priority coefficients are set for the g groups of analysis data in sequence.

[0081] For example, in areas where natural disasters occur, the priority coefficient of medical equipment should be greater than the priority coefficient of communication equipment, and greater than the priority coefficient of lighting equipment; the reason is that areas where natural disasters occur usually result in a large number of casualties and require emergency treatment; the second is communication equipment, and restoring communications plays a key role in intelligence transmission and search and rescue command; finally, lighting equipment; and for medical equipment, if an earthquake occurs, intensive care equipment must take priority over conventional medical equipment because the number of seriously ill patients may be large; if a flood occurs, equipment for diagnosing and treating infectious diseases has the highest priority coefficient to prevent the spread of infectious diseases; if a forest fire occurs, respiratory treatment equipment must take precedence over other medical equipment because the smoke caused by the fire has a great impact on the respiratory system.

[0082] Each set of feature vectors is used as the input of the priority analysis model. The priority analysis model outputs a set of predicted priority coefficients corresponding to each set of analysis data, and uses the actual priority coefficients corresponding to each set of analysis data as the prediction target. The actual priority coefficients are the priority coefficients corresponding to the analysis data collected in advance. The training goal is to minimize the sum of the prediction errors of all analysis data. The calculation formula for the prediction error is: ,in is the prediction error, K is the group number of the eigenvector corresponding to the analysis data, is the prediction priority coefficient corresponding to the Kth group of analysis data, is the actual priority coefficient corresponding to the Kth group of analysis data; the priority analysis model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

[0083] The above-mentioned priority analysis model is specifically a deep neural network model; it includes an input layer, a hidden layer and an output layer; each hidden layer includes multiple neurons, each neuron is connected to the neurons in the next layer, and the connection contains weights, which determine the importance and influence of data transmission in the neural network; an activation function is applied to each neuron between the hidden layer and the output layer, and the activation function reflects nonlinearity, allowing the network to learn more complex patterns and features.

[0084] It should be noted that the reason for supplying power to each electrical equipment according to the priority coefficient is that since the scheduling set includes multiple scheduling units, each scheduling unit will dispatch energy to the overload area at a different time. The electric energy that is preferentially dispatched to the overload area will be supplied to the electrical equipment with a large priority coefficient, which can effectively utilize limited power resources, maximize the protection of life safety, maintain public safety and improve rescue efficiency; it not only helps to respond to disasters in a timely manner, but also reduces secondary disasters and losses, and provides reliable power support for disaster areas.

[0085] This embodiment conducts real-time monitoring of various power grid areas through regional division and energy data collection, accurately identifying overloaded areas; fully considers scheduling time factors and adopts a clonal selection algorithm to formulate and implement load balancing strategies for overloaded areas; and, based on the specific electrical equipment and natural disaster types in the overloaded areas, calculates the priority power supply sequence for each electrical equipment, effectively ensuring priority power supply to life safety facilities; achieves effective balancing and optimal allocation of smart grid loads, maximizes emergency power supply during natural disasters, thereby improving the power grid's operation level and people's livelihood protection capabilities, and improves the power supply reliability and power efficiency of the power grid.

[0086] Example 2:

[0087] See also Figure 3 As shown, this embodiment further improves the design on the basis of embodiment 1. In embodiment 1, the operation of the equipment is judged by the size relationship between the power value and 0; however, in actual conditions, when the equipment fails or is in an abnormal working state, the power value may not be 0, but there is no need to supply power to the equipment to avoid energy waste; when the equipment is in standby or sleep state, the power value may be 0, but it is still necessary to supply power to the equipment to avoid automatic shutdown after a complete power outage, so as to ensure that it can be quickly started and put into use when needed; if the method in embodiment 1 is still used to judge the operation of the equipment, it will lead to misjudgment of the equipment operation status, thereby resulting in the inability to give priority to power supply to life safety facilities, thereby affecting emergency rescue work; therefore, the smart grid load balancing control system provided by this embodiment also includes a state analysis module.

[0088] The status analysis module is used to collect and analyze the operating parameters of each electrical device in the overload area, filter out the electrical devices in the online state, and mark them as operating devices.

[0089] The operating parameters include power value, temperature value and vibration value; the power value is obtained by the power sensor installed in each electrical device; the temperature value is obtained by the temperature sensor installed in each electrical device; and the vibration value is obtained by the vibration sensor installed in each electrical device.

[0090] Obtain the device label of each electrical device, and use the device label and operating parameters corresponding to an electrical device as a set of test data; input each set of test data into the trained state analysis model in turn to predict the state label corresponding to each electrical device. The state label is the digital label corresponding to the device state. The device state includes running state, standby state, shutdown state and fault state. Different device states have different digital labels. The training process of the state analysis model is consistent with the training process of the priority analysis model, and both are deep neural network models.

[0091] The online status includes the running status and the standby status; the status tag corresponding to the running status is marked as the running tag, and the status tag corresponding to the standby status is marked as the standby tag; the running tag and the standby tag are used as the status tags corresponding to the online status; the status tag of each electrical device is compared with the status tag of the online status; the electrical devices with the same status tags as the running tag or the standby tag are filtered out and marked as running devices.

[0092] It should be noted that the reason for judging the equipment status of the electrical equipment by the operating parameters is that when the electrical equipment is in the operating state, the operating parameters are all within the corresponding normal range; when the electrical equipment is in the standby state, the operating parameters will all decrease, and the power value may be equal to 0, but because the equipment is still in the online state, there is still vibration and heat dissipation, so the vibration value should be greater than 0 but less than the vibration value in the operating state, and the temperature value should be higher than the ambient temperature but less than the temperature value in the operating state; when the electrical equipment is in the shutdown state, the power value and vibration value in the operating parameters are both 0, and the temperature value should be close to the ambient temperature; when the electrical equipment is in a fault state, it will affect the normal operation of the equipment, so there is at least one parameter in the operating parameters that exceeds the corresponding normal range; in summary, when the electrical equipment is in different equipment states, the corresponding operating parameters are inconsistent, so the equipment status of the electrical equipment can be accurately judged by the operating parameters.

[0093] This embodiment adopts multi-source information fusion technology to fuse the multi-dimensional operating parameters collected from electrical equipment, such as power, temperature, and vibration values; through deep learning technology, it completes the subdivision of the operating status, standby status, shutdown status, and fault status of electrical equipment, realizes end-to-end identification of equipment status, and can more accurately identify the actual equipment status of electrical equipment; avoids misjudgment of equipment status caused by a single judgment, thereby providing a more reliable basis for power grid load scheduling, ensuring priority power supply to key emergency equipment, and improving emergency rescue efficiency and the reliability of power grid load balancing control.

[0094] Example 3:

[0095] See also Figure 4As shown, for the parts not described in detail in this embodiment, please refer to the description of Embodiment 1 and Embodiment 2. A smart grid load balancing control method is provided, and the method includes:

[0096] Dividing the power grid into multiple areas based on predetermined rules;

[0097] Identify overloaded areas with insufficient power supply based on regional power supply and regional power consumption;

[0098] A load balancing strategy is formulated and executed for overloaded areas using a nature-inspired optimization algorithm. The load balancing strategy is a load scheduling strategy that aims to minimize scheduling time.

[0099] Calculate the priority coefficient for electrical equipment in the overload area and supply power to each electrical equipment according to the priority coefficient.

[0100] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0101] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A smart grid load balancing control method, characterized in that: include: Dividing the power grid into multiple areas based on predetermined rules; Identify overloaded areas with insufficient power supply based on regional power supply and regional power consumption; A load balancing strategy is formulated and executed for the overloaded area using a nature-inspired optimization algorithm, wherein the load balancing strategy is a load scheduling strategy with the goal of minimizing scheduling time, and the nature-inspired optimization algorithm is a clonal selection algorithm; Calculate the priority coefficient for electrical equipment in the overload area and supply power to each electrical equipment according to the priority coefficient.

2. The smart grid load balancing control method according to claim 1, characterized in that: The steps to developing a load balancing strategy include: Step 1: Randomly generate a set of solutions, which include multiple scheduling sets, set set labels for the scheduling sets, and preset iteration thresholds; Step 2: Define the fitness function and use the scheduling time as the evaluation indicator; Step 3: Iterate and optimize the solution through cloning, mutation, and recombination operations. When the preset number of iterations is reached, filter the set label with the largest fitness and obtain the corresponding scheduling set based on the set label.

3. The smart grid load balancing control method according to claim 2, characterized in that: The method for identifying an overloaded area with insufficient power supply comprises: An overloaded area refers to a regional state when the regional power supply is less than the regional power consumption; The sum of electricity consumption in each region is the regional electricity consumption; the sum of power generation in each region is the regional power generation; the sum of energy storage in each region is the regional energy storage; the sum of regional power generation and regional energy storage in each region is the regional power supply; and energy storage is the amount of electricity stored in each energy storage unit in the region.

4. The smart grid load balancing control method according to claim 3, characterized in that: In step 1, except for the overloaded area, the energy storage units in all areas are marked as dispatching units; all dispatching units are randomly combined, and the result of the combination is taken as a unit set; the storage energy of the dispatching units in each unit set is sequentially added to obtain the collective storage energy; the regional power consumption in the overloaded area is marked as overloaded power consumption, the collective storage energy of each unit set is compared with the overloaded power consumption, and the unit sets whose collective storage energy is greater than or equal to the overloaded power consumption are retained and marked as deployment sets; Subtract the overload power consumption from the collective energy storage of each deployment set to obtain the energy storage difference; compare the energy storage of the dispatching units in each deployment set with the energy storage difference in turn, retain the deployment sets in which the energy storage of each dispatching unit is greater than or equal to the energy storage difference, and mark them as dispatching sets; set a different digital label for each dispatching set and mark it as a set label.

5. The smart grid load balancing control method according to claim 4, characterized in that: In step 2, the method for obtaining the scheduling time includes: Get all scheduling units in the scheduling set and mark them as analysis units; Obtain the topological structure of the power grid and draw a topological map of the power grid, which includes nodes and edges; obtain the route between each analysis unit and each node in the overload area according to the topological map of the power grid; obtain the length of each transmission line in each route according to the topological map of the power grid; add the transmission line lengths corresponding to each route in sequence to obtain the total length of the line corresponding to each route; compare the total lengths of the lines corresponding to multiple routes of each analysis unit, and take the total length of the line with the smallest value as the shortest length of the corresponding analysis unit; divide the shortest length of each analysis unit by the speed of light to obtain the transmission time of each analysis unit; preset a start-up time set, which includes the start-up time of each energy storage unit in the power grid, and the start-up time is the time consumed by the energy storage unit from startup to the start of power transmission; obtain the start-up time corresponding to each analysis unit according to the start-up time set; The transmission time of each analysis unit is added to the corresponding start time to obtain the scheduling time of each analysis unit; the scheduling time of each analysis unit is compared, and the scheduling time with the largest value is used as the scheduling time of the scheduling set.

6. The smart grid load balancing control method according to claim 3, characterized in that: Step 3 also includes verifying whether the energy storage capacity of the new scheduling set meets the constraints after mutation and reorganization. The constraints are that the total energy storage capacity of the new scheduling set covers the overload power consumption and the energy storage capacity of the scheduling unit is greater than the energy storage difference.

7. The smart grid load balancing control method according to claim 6, characterized in that: The method for calculating the priority coefficient of electrical equipment in the overload area includes: Collect the power value of each electrical device in the overload area; compare the power value of each electrical device with 0, mark the electrical devices with power values greater than 0 as running devices, and do not mark the electrical devices with power values equal to 0; assign different digital labels to the electrical devices in the overload area and mark them as device labels; obtain the device label corresponding to each running device and mark it as the running label; Collect regional images of overloaded areas, use the trained disaster recognition model to identify the regional images, and output the recognition results, which are digital labels corresponding to the natural disaster types; mark the digital labels corresponding to the natural disaster types as disaster labels; use a device label and a disaster label as a set of analysis data, input each set of analysis data into the trained priority analysis model, and predict the corresponding priority coefficient.

8. The smart grid load balancing control method according to claim 7, characterized in that: The method further includes: collecting and analyzing operating parameters of each electrical device in the overload area, screening out electrical devices in an online state, and marking them as operating devices; Operating parameters include power, temperature, and vibration. The device tag of each electrical device is obtained, and the device tag and operating parameters corresponding to each electrical device are used as a set of test data. Each set of test data is sequentially input into the trained state analysis model to predict the state tag corresponding to each electrical device. The state tag is the digital label corresponding to the device state. The device state includes running state, standby state, shutdown state, and fault state. Different device states have different corresponding digital labels. The online status includes the running status and the standby status; the status tag corresponding to the running status is marked as the running tag, and the status tag corresponding to the standby status is marked as the standby tag; the running tag and the standby tag are used as the status tags corresponding to the online status; the status tag of each electrical device is compared with the status tag of the online status; the electrical devices with the same status tags as the running tag or the standby tag are filtered out and marked as running devices.

9. A smart grid load balancing control system, implementing the smart grid load balancing control method according to any one of claims 1 to 8, characterized in that: include: A microgrid partitioning module that divides the grid into multiple areas based on predetermined rules; The overload determination module identifies overloaded areas with insufficient power supply based on regional power supply and regional power consumption; A load balancing module, configured to formulate and execute a load balancing strategy for overloaded areas using a nature-inspired optimization algorithm, wherein the load balancing strategy is a load scheduling strategy with the goal of minimizing scheduling time; The sequential power supply module is used to calculate the priority coefficient of the electrical equipment in the overload area and supply power to each electrical equipment according to the priority coefficient.

Images