A power system load balancing control method

Abstract

The application discloses a power system load balancing control method, and belongs to the field of power system operation control and dispatching, which comprises the following steps: obtaining power grid global operation parameters to generate a load power target adjustment value; collecting load capacity data in each power supply subarea and aggregating to generate a virtual resource set; performing optimization calculation on the virtual resource set based on the load power target adjustment value to generate a load control instruction set; and issuing the instruction for execution and performing closed-loop compensation adjustment. The layered control architecture combining global optimization and regional autonomy can realize the collaborative optimization of macro load balancing and micro cost benefit, and perform emergency autonomous control when communication is interrupted, thereby significantly improving the reliability, robustness and precision of power grid operation and regulation.

Description

Technical Field

[0001] This invention relates to the field of power system operation control and dispatch, and in particular to a power system load balancing control method. Background Technology

[0002] In modern power systems, load balancing control is a key technology for ensuring the safe, stable, and economical operation of the power grid. With socio-economic development and the continuous growth of electricity load, especially the integration of large-scale renewable energy and distributed loads, the unidirectional power flow characteristics of traditional power systems have been disrupted, significantly increasing the complexity and uncertainty of grid operation. Particularly during peak electricity consumption periods, distribution transformers or lines in some areas are prone to heavy or even overloaded conditions, while other areas may experience light loads. This uneven load distribution severely impacts the reliability and quality of power supply.

[0003] Currently, existing technologies still have many limitations in solving the load balancing problem of power systems. Traditional load control methods mostly rely on human experience for observation and judgment, which is inefficient and makes it difficult to guarantee optimal decisions. Although some automated control systems can achieve load transfer, they are often limited to rigid transfer at the single grid level, failing to fully consider the flexible control capabilities of the user side and the actual operational constraints of the local grid. This can easily lead to poor control effects, expanded impact range, or the creation of new overload risks. In addition, existing load control systems generally lack emergency mechanisms to deal with communication interruptions, resulting in limited control capabilities during network anomalies and an inability to form effective closed-loop optimization. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a power system load balancing control method. It employs a hierarchical control architecture, acquiring global operating parameters and ultra-short-term load forecast data to generate target adjustment values ​​for load power in each power supply zone. By receiving and aggregating load capacity profiles from each power supply zone, a virtual resource set is generated. Based on the target adjustment values ​​and local data, the virtual resource set is optimized to generate a load control instruction set. This invention achieves synergistic optimization of macro-level load balancing and micro-level cost-effectiveness, and executes emergency autonomous control during communication interruptions, significantly improving the reliability, robustness, and precision of power grid operation.

[0005] The above objectives can be achieved through the following approach:

[0006] A power system load balancing control method includes: acquiring global operating parameters and ultra-short-term load forecast data from the power grid dispatch center; obtaining target adjustment values ​​for load power in each power supply zone through power flow calculation and with overall load balancing and line loss minimization as optimization objectives; collecting standardized load capacity profile data packages containing real-time operating power, maximum adjustable power, and sustainable adjustment time within each power supply zone, and aggregating them to generate a virtual resource set representing the adjustable load resources of the region; based on the target adjustment values ​​for load power, combined with voltage data collected in real-time from key nodes of the distribution network in the global operating parameters, and with the minimum adjustment cost as the optimization objective, sorting and combining the resources in the virtual resource set to generate a load control instruction set; issuing the load control instruction set to the corresponding load control terminal to execute load adjustment, monitoring the actual load changes and voltage changes, comparing the actual adjustment results with the target adjustment values ​​for load power, calculating the adjustment deviation, and executing closed-loop compensation adjustment.

[0007] Optionally, the method further includes: when communication between the power supply zone and the power grid dispatch center is interrupted, performing emergency autonomous optimization control based on local real-time monitoring data, and generating an emergency autonomous operation status record. Optionally, performing emergency autonomous optimization control includes: in emergency autonomous mode, the power supply zone continuously monitors the real-time voltage, current, and power changes of key nodes in the distribution network within its jurisdiction, forming a local real-time operation status snapshot; based on the local real-time operation status snapshot, performing strategy matching and filtering through a pre-stored emergency control strategy library to identify emergency control strategies that can solve the local load imbalance problem; generating an emergency load control instruction set according to the emergency control strategy, and issuing it to the corresponding load control terminal to perform load adjustment, and generating an emergency autonomous operation status record based on the operation process and results.

[0008] Optionally, the step of aggregating the collected load regulation capacity data to generate a virtual resource set representing the adjustable load resources of the region includes: collecting real-time operating data and load characteristic parameters of the load control terminal, and extracting the power characteristics, time characteristics, and response characteristics of the load control terminal; using a data standardization processing method, converting the extracted parameters into a standardized load capacity profile data package containing real-time operating power, maximum adjustable power, sustainable adjustment time, and adjustment response rate. The standardized load capacity profile data package is then parsed and aggregated, and a virtual resource set representing the adjustable load resources of the region is generated through a multi-dimensional resource aggregation algorithm.

[0009] Optionally, parsing and aggregating the standardized load capacity profile data package includes: performing a weighted summation calculation on the maximum adjustable power to generate the total regional adjustment capacity; calculating the average sustainable adjustment time to obtain the average sustainable duration; and filtering the adjustment response rate by minimum value to generate the fastest response rate.

[0010] Optionally, the step of sorting and optimizing the resources in the virtual resource set to generate a load control instruction set includes: establishing a multi-dimensional dynamic ranking model for load regulation priority based on the power characteristics, time characteristics, and response characteristics of each resource in the virtual resource set; establishing a regulation cost model that considers load response delay and user incentives to quantify the regulation cost of each resource in the virtual resource set; under the security constraints determined by the real-time voltage data of the key nodes of the distribution network, using the regulation cost model as the optimization objective and combining the results of the dynamic ranking model for load regulation priority, optimizing the combination of resources in the virtual resource set to obtain an optimized combination of regulation resources; and converting the optimized combination of regulation resources into a load control instruction set containing specific control objects and control parameters.

[0011] Optionally, the step of establishing a multi-dimensional dynamic ranking model for load regulation priorities includes: calculating a load regulation margin index characterizing the current adjustable potential of resources based on the real-time operating power and the total regional regulation capacity; calculating a load regulation durability index and a load regulation sensitivity index characterizing the stability and response speed of resource regulation based on the average sustainability time and the fastest response rate; using a multi-index weighted fusion algorithm to comprehensively score the load regulation margin index, the load regulation durability index, and the load regulation sensitivity index to obtain the final regulation priority of each resource, and ranking the resources in the virtual resource set according to the regulation priority.

[0012] Optionally, establishing a regulation cost model that considers load response delay and user incentives includes: for each resource in the virtual resource set, obtaining a basic regulation cost parameter characterizing the inherent regulation cost per unit power of that resource; based on the regulation response rate in the standardized load capacity profile data packet, calculating a response delay cost parameter that quantifies the potential impact of resource response delay on stability through a nonlinear mapping function; and, according to a preset segmented user incentive policy, calculating a user incentive cost parameter that incentivizes users to participate in regulation through a piecewise linear function.

[0013] Optionally, comparing the actual adjustment result with the target adjustment value of the load power, calculating the adjustment deviation, and performing closed-loop compensation adjustment includes: monitoring the actual load change, actual voltage change, and actual adjustment response time after the execution of the load control command set to obtain the actual adjustment result; comparing the actual adjustment result with the target adjustment value of the load power and the expected response time in multiple dimensions to obtain a comprehensive adjustment deviation including power deviation, voltage deviation, and time deviation; based on the magnitude and nature of the comprehensive adjustment deviation, using a hierarchical compensation decision algorithm, selecting low-power loads with high precision and fast response characteristics from the virtual resource set, and generating compensation adjustment commands; and issuing the compensation adjustment commands to the corresponding load control terminal to perform compensation adjustment and eliminate the comprehensive adjustment deviation.

[0014] Based on the same inventive concept, this invention also provides a power system load balancing control system, the system comprising: a data acquisition and fusion module, used to acquire global operating parameters and ultra-short-term load forecast data from the power grid dispatch center, collect standardized load capacity profile data packets within each power supply zone, and aggregate the standardized load capacity profile data packets to generate a virtual resource set representing the adjustable load resources of the region; an optimization decision module, used to sort and combine resources in the virtual resource set based on the load power target adjustment value, combined with voltage data collected in real time from key nodes of the distribution network in the global operating parameters, with the minimum adjustment cost as the optimization objective, to generate a load control instruction set; an execution and monitoring module, used to send the load control instruction set to the corresponding load control terminal to execute load adjustment, and monitor the actual load change and voltage change to obtain the actual adjustment result; and a closed-loop compensation module, used to compare the actual adjustment result with the load power target adjustment value, calculate the adjustment deviation, and execute closed-loop compensation adjustment.

[0015] Compared with the prior art, the present invention has the following advantages:

[0016] 1. This invention adopts a hierarchical control architecture that combines global optimization with regional autonomy. By performing power flow calculations at the global level and using overall load balancing and line loss minimization as optimization objectives, it provides precise load power target adjustment values ​​for each power supply zone, while allowing zones to perform optimization calculations based on local real-time data. This hierarchical decision-making mechanism effectively solves the computational complexity problem in large-scale power grid dispatching, ensuring the effectiveness of macro-control and the flexibility of micro-execution, and significantly improving the overall efficiency and reliability of power grid operation.

[0017] 2. This invention constructs a multi-dimensional load capacity profile and a virtual resource set, enabling refined management and abstract utilization of load regulation resources. By establishing a regulation cost model that considers load response delay and user incentives, as well as a multi-dimensional dynamic ranking model, it can intelligently select the best load resources, avoiding the drawbacks of traditional rigid regulation. Simultaneously, in the event of a communication interruption, the regional agent can enter an emergency autonomous mode, independently executing optimized control based on local data, greatly enhancing its resilience and robustness, and ensuring continuous power supply under non-ideal conditions.

[0018] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating a power system load balancing control method according to an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the multi-index weighted fusion algorithm according to an embodiment of the present invention.

[0022] Figure 3 This is a nonlinear relationship diagram between response delay cost and adjustment response rate in an embodiment of the present invention.

[0023] Figure 4 This is a schematic diagram of the structure of a power system load balancing control system according to an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Reference Figure 1One embodiment of this invention proposes a power system load balancing control method. It employs a hierarchical control architecture combining global optimization and regional autonomy. The global optimization module acquires global operating parameters and ultra-short-term load forecast data, generating target adjustment values ​​for load power in each power supply zone. By receiving and aggregating load capacity profiles from each power supply zone, a virtual resource set is generated. Based on the target adjustment values ​​and local data, the virtual resource set is optimized to generate a load control instruction set. This invention achieves synergistic optimization of macro-level load balancing and micro-level cost-effectiveness, and executes emergency autonomous control during communication interruptions, significantly improving the reliability, robustness, and precision of power grid operation.

[0026] The method described in this embodiment specifically includes:

[0027] The global operating parameters and ultra-short-term load forecast data of the power grid dispatch center are obtained. Through power flow calculation and with the optimization objectives of overall load balance and line loss minimization, the load power target adjustment values ​​of each power supply zone are obtained.

[0028] Specifically, the power grid dispatch center collects the topology and real-time operating status parameters of the main power grid lines to form a snapshot of the current power grid status. The dispatch center obtains load forecast curves for a future period and combines them with the current power grid status snapshot to perform power flow calculations. With overall load balancing and line loss minimization as optimization objectives, the power flow calculation model is solved to obtain the target adjustment values ​​for load power in each power supply zone.

[0029] Collect standardized load capacity profile data packages containing real-time operating power, maximum adjustable power, and sustainable adjustment time within each power supply zone, and aggregate them to generate a virtual resource set representing the adjustable load resources of the region;

[0030] Specifically, within each power supply zone, the load control terminal continuously monitors the operating parameters of its connected loads, including current power, maximum adjustable power, and sustainable adjustment time. These operating parameters are packaged into standardized load capacity profile data packets and uploaded to the power supply zone via the communication network. The power supply zone receives and parses all profile data packets within its jurisdiction, weights and aggregates the adjustable capacities within them, and generates a virtual resource set representing the adjustable load resources of its region.

[0031] Based on the load power target adjustment value, and combined with the voltage data collected in real time from key nodes of the distribution network in the global operating parameters, with the goal of minimizing adjustment cost, the resources in the virtual resource set are sorted and combined to generate a load control instruction set.

[0032] Specifically, the power supply zone reads the load power target adjustment value from the power grid dispatch center as the adjustment target for local optimization calculations. Simultaneously, it obtains real-time voltage measurements from voltage measurement equipment at key nodes of the distribution network as safety constraints in the optimization calculations. The local optimization algorithm aims to minimize adjustment costs and, under safety constraints, sorts and combines resources in the virtual resource set to obtain an optimized set of adjustment resource combinations. These optimized adjustment resource combinations are then transformed into a load control instruction set containing specific control objects and control parameters.

[0033] The load control command set is sent to the corresponding load control terminal to perform load adjustment, monitor the actual load change and voltage change, compare the actual adjustment result with the target adjustment value of load power, calculate the adjustment deviation and perform closed-loop compensation adjustment.

[0034] Specifically, after receiving the load control command set, the load control terminal executes the corresponding load adjustment operation. During this operation, the load control terminal continuously monitors the actual load changes and voltage changes at key nodes of the distribution network to obtain the actual adjustment results. The actual adjustment results are compared with the target load power adjustment value to calculate the adjustment deviation. If the adjustment deviation exceeds a preset threshold, a fine-tuning mechanism is automatically activated, generating a compensation adjustment command, which is then sent to the load control terminal to execute closed-loop compensation adjustment.

[0035] A hierarchical decision-making mechanism is adopted, which sets macro-control targets at the global level while allowing regional agents to perform micro-optimization based on local real-time data and conduct compensation adjustments in a closed-loop manner. This method can effectively solve the computational complexity problem in large-scale power grid dispatching, ensuring the effectiveness of macro-control and the flexibility of micro-execution, and significantly improving the overall efficiency, reliability, and accuracy of power grid operation.

[0036] Optionally, the method further includes:

[0037] When communication between the power supply sub-district and the power grid dispatch center is interrupted, emergency autonomous optimization control is executed based on local real-time monitoring data, and an emergency autonomous operation status record is generated. Specifically, the power supply sub-district continuously monitors the communication link status between its jurisdiction and the power grid dispatch center. When the communication link experiences continuous interruption, and the interruption time exceeds a preset emergency activation threshold, the power supply sub-district will automatically enter emergency autonomous mode. During emergency autonomous operation, the power supply sub-district will record all operation processes and results into an emergency autonomous operation status record, which will be uploaded to the power grid dispatch center after communication is restored.

[0038] Optionally, the execution of emergency autonomous optimization control includes:

[0039] In emergency autonomous mode, the power supply zone continuously monitors the real-time voltage, current and power changes of key nodes in the distribution network within its jurisdiction, forming a snapshot of the local real-time operating status.

[0040] Specifically, in emergency autonomous mode, the power supply zone continuously receives data from voltage and current sensors at key nodes of the distribution network, as well as load power data from various load control terminals. This data is sampled at high frequency and timestamped to form a local real-time operational snapshot. This snapshot comprehensively reflects the current physical operational status of the local distribution network.

[0041] Based on the local real-time operating status snapshot, the emergency control strategy library is used to perform strategy matching and filtering to identify emergency control strategies that can solve the local load imbalance problem.

[0042] Specifically, each power supply zone pre-stores an emergency control strategy library, containing various contingency plans for different load imbalance scenarios. Each strategy in the library is associated with one or more power grid operating states. For example, one emergency control strategy is to immediately cut off non-critical loads and send a command to activate distributed reactive power compensation devices when the local grid voltage drops by more than 5%. Another strategy is to initiate a temperature increase procedure for all air conditioning loads in the area when the critical line load rate exceeds 110%. Based on a snapshot of the local real-time operating status, the power supply zone uses a feature matching algorithm to match and score the current power grid feature vector with the strategies in the strategy library. This matching algorithm uses the decision tree principle, taking features such as voltage deviation and line load rate as decision nodes, and through layer-by-layer matching, selects the emergency control strategy that best matches the current power grid state. The selection process prioritizes minimizing local voltage deviation and maximizing the continuity of power supply to critical loads. When the strategy with the highest score is matched, it is adopted as the emergency control strategy.

[0043] Based on the emergency control strategy, an emergency load control instruction set is generated and sent to the corresponding load control terminal to perform load adjustment. The operation process and results are then used to generate an emergency autonomous operation status record.

[0044] Specifically, the power supply zone transforms the selected emergency control strategies into a series of specific emergency load control commands. The command set includes the controlled object, control parameters, and control timing. The command set is then distributed to the corresponding load control terminal via the local communication network to execute emergency load adjustments. Simultaneously, the power supply zone records the operation log, grid status changes, and command execution results throughout the entire emergency autonomous operation period, forming a complete emergency autonomous operation status record for uploading and analysis after communication is restored.

[0045] Optionally, the step of aggregating the collected load regulation capacity data to generate a virtual resource set representing the adjustable load resources of the region includes:

[0046] Collect real-time operating data and load characteristic parameters from the load control terminal, and extract the power characteristics, time characteristics, and response characteristics of the load control terminal;

[0047] Specifically, the Load Control Terminal (LCT) continuously monitors the real-time operating data of the loads it is connected to, including real-time operating power. Maximum adjustable power With sustainable adjustment time In addition, LCT also extracts the load's response characteristics during regulation, such as adjusting the response rate. These parameters together constitute a complete description of load regulation capability.

[0048] The extracted parameters are converted into a standardized load capacity profile data package containing real-time operating power, maximum adjustable power, sustainable adjustment time, and adjustment response rate by adopting a data standardization processing method.

[0049] Specifically, to facilitate data transmission and unified processing, a data standardization method is adopted to encapsulate the extracted parameters in a unified format, forming a standardized load capacity profile data package. For example, the unified format of the sustainable adjustment time value can be represented as:

[0050] ,

[0051] in, It is the preset minimum time unit, and It is an integer multiple to ensure that the duration parameters of different loads can be processed by the algorithm on the same scale.

[0052] The standardized load capacity profile data packets are parsed and aggregated, and a virtual resource set representing the adjustable load resources in this region is generated through a multi-dimensional resource aggregation algorithm.

[0053] Specifically, after receiving the standardized load capacity profile data packet, the power supply zone parses its parameters. Then, the power supply zone employs a multi-dimensional resource aggregation algorithm to aggregate the standardized load capacity profile data packets from different load cells (LCTs). The aggregation algorithm comprehensively considers the maximum adjustable power, sustainable adjustment time, and adjustment response rate of each load. Through weighted summation, average calculation, and minimum value filtering of the original parameters, it generates a virtual resource pool (VRP) that characterizes the overall adjustment capacity and characteristics of the region. The VRP contains the total regional adjustment capacity. Average Durable Time and fastest response rate These key metrics provide a unified and abstract view for local optimization computation.

[0054] Optionally, parsing and aggregating the standardized load capacity profile data packet includes:

[0055] The total regional regulation capacity is generated by weighted summation of the maximum adjustable power.

[0056] Specifically, after receiving standardized load capacity profile data packets from various load control terminals, the power supply zone determines the maximum adjustable power of each load. Perform a weighted summation. Weighting coefficients. The total regional regulation capacity can be dynamically set based on factors such as the importance of the load and the reliability of the response. The calculation formula can be expressed as:

[0057] ,

[0058] in, It is the total number of adjustable loads within the region. Representing the One load, The weighting factor can be determined using the following method: quantify the importance of the load, for example, divide the load into three levels: critical load, general load, and non-critical load, and assign importance scores to each level. For example, 3, 2, 1. Simultaneously, this module calculates the load's response reliability score based on historical load control response data, such as response success rate and average response deviation. Then, the importance score and reliability score are weighted and fused to calculate the final weighting coefficient for the load. The calculation formula can be expressed as:

[0059] ,

[0060] in, It is the first The weighting factor for each load, and It is a preset weighting factor used to balance the impact of importance and reliability.

[0061] The average sustainable duration is obtained by averaging the sustainable adjustment time.

[0062] Specifically, the power supply zone considers the sustainable adjustment time in all received standardized load capacity profile data packets. Calculate the arithmetic mean. Average duration. The calculation formula can be expressed as:

[0063] .

[0064] The minimum value is selected from the adjustment response rate to generate the fastest response rate.

[0065] Specifically, the power supply zone adjusts its response rate based on all received standardized load capacity profile data packets. Compare and filter. Fastest response rate. The calculation formula can be expressed as:

[0066] .

[0067] Optionally, the step of sorting and optimizing the resources in the virtual resource set to generate a load control instruction set includes:

[0068] Based on the power characteristics, time characteristics, and response characteristics of each resource in the virtual resource set, a multi-dimensional dynamic ranking model for load adjustment priority is established.

[0069] Specifically, a multi-dimensional dynamic ranking model for load regulation priorities is established based on the power characteristics, time characteristics, and response characteristics of each resource in the virtual resource set. This model calculates a comprehensive score for each adjustable resource by weighting and fusing multiple key indicators, which serves as its priority in regulation.

[0070] Establish a regulation cost model that considers load response delay and user incentives to quantify the regulation cost of each resource in the virtual resource set;

[0071] Specifically, an Adjustment Cost Model (ACM) is established to quantify the adjustment costs of various resources in a refined manner. The ACM comprehensively considers multiple cost components, including basic adjustment costs, response latency costs, and user incentive costs. (Resource adjustment costs) It can be calculated using the following formula:

[0072] ,

[0073] in, Representing the The inherent adjustment cost per unit of power of a resource is typically a fixed parameter. It is the cost caused by response latency, which increases non-linearly with the increase of response time; This is the economic compensation that must be paid to encourage users to participate in regulation, and its size depends on the specific incentive policy.

[0074] Under the security constraints determined by the real-time voltage data of the key nodes of the distribution network, the resources in the virtual resource set are combined and optimized by taking the adjustment cost model as the optimization objective and combining the results of the load adjustment priority dynamic ranking model to obtain the optimized adjustment resource combination.

[0075] Specifically, under the security constraints determined by the real-time voltage data of key nodes in the distribution network, the optimization objective is to minimize the total regulation cost. This optimization is achieved by combining the results of a multi-dimensional load regulation priority dynamic ranking model with the results of the combined optimization of resources in the virtual resource set. This combined optimization problem can be formalized as a complex mathematical programming problem, whose objective function and constraints can be expressed as follows:

[0076]

[0077]

[0078]

[0079]

[0080] in, It is its regulating power, This represents the selected set of resources. It is a node voltage, This represents the total target adjustment value for load power. By solving this optimization problem, we can obtain the optimized combination of regulation resources that balances cost and safety / reliability.

[0081] The optimized combination of regulation resources is transformed into a load control instruction set containing specific control objects and control parameters.

[0082] Specifically, the optimized combination of regulation resources obtained through combinatorial optimization is transformed into a load control instruction set containing specific control objects and control parameters. This instruction set explicitly specifies key information such as the load control terminal ID to be regulated, the specific regulation power amount, the regulation duration, and the instruction validity period, and is sent to the corresponding load control terminal through an encrypted communication network to ensure the accurate and secure execution of the instructions.

[0083] Optionally, the step of establishing a multi-dimensional dynamic ranking model for load adjustment priorities includes:

[0084] Based on the real-time operating power and the total regional regulation capacity, a load regulation margin index characterizing the current adjustable potential of resources is calculated.

[0085] Specifically, based on the real-time operating power of each resource in the virtual resource set and the total regional regulation capacity, a load regulation margin index characterizing the current adjustable potential of the resources is calculated. This indicator reflects the contribution of a single adjustable load to alleviating regional load gaps, and its calculation formula can be expressed as:

[0086] ,

[0087] in, It is the first The maximum adjustable power of each resource. It is its real-time operating power. This refers to the total regional regulation capacity. This indicator comprehensively considers the inherent regulatory potential of resources and the current load status of the power grid; the larger the value, the greater the regulatory potential.

[0088] Based on the average sustainability time and the fastest response rate, load regulation durability index and load regulation sensitivity index, which characterize the stability and response speed of resource regulation, are calculated.

[0089] Specifically, the load regulation durability index is calculated based on the average sustainability time and fastest response rate in the virtual resource set. and load regulation sensitivity index The load regulation durability index aims to measure the ability of resources to contribute stably to the system load during regulation. Its calculation formula can be expressed as:

[0090] ,

[0091] The load regulation sensitivity index is used to quantify the response speed of resources to ensure a rapid response during emergency adjustments. Its calculation formula can be expressed as:

[0092] ,

[0093] in, It is the first The sustainable adjustment time of a resource. This refers to its regulation response rate. The higher the values ​​of the load regulation durability index and the load regulation sensitivity index, the better the resource regulation stability and response speed.

[0094] A multi-index weighted fusion algorithm is used to comprehensively score the load adjustment margin index, load adjustment durability index, and load adjustment sensitivity index to obtain the final adjustment priority of each resource, and then sort the resources in the virtual resource set according to the adjustment priority.

[0095] Specifically, a multi-index weighted fusion algorithm is used to comprehensively score the load adjustment margin index, load adjustment durability index, and load adjustment sensitivity index to obtain the final adjustment priority for each resource. The formula for calculating the overall score is as follows:

[0096] ,

[0097] in, , , These are weighting coefficients that are dynamically adjusted based on the current power grid operating status and dispatching strategies. When the overall power grid load factor exceeds 90%, the weighting coefficients are adjusted to prioritize power grid security. Increase the weighting factor to prioritize loads with higher margins. When the voltage compliance rate decreases, the weighting factor... The algorithm increases its weighting coefficients to prioritize loads with the fastest response for regulation. This dynamic adjustment of weighting coefficients allows the algorithm to adapt to different power grid operating conditions. Figure 2 As shown, the principle of the multi-index weighted fusion algorithm is described. By displaying the scores of different resources on each index and their weighted total score, the dynamic ranking process of load adjustment priority is intuitively reflected.

[0098] Optionally, establishing a regulation cost model that considers load response delay and user incentives includes:

[0099] For each resource in the virtual resource set, obtain the basic adjustment cost parameter that characterizes the inherent adjustment cost per unit power of the resource;

[0100] Specifically, for each resource in the virtual resource set, a Base Adjustment Cost Parameter (BACP) is obtained. The BACP characterizes the inherent adjustment cost per unit power of that resource, and is set based on factors such as the load's equipment type, energy efficiency, and maintenance costs. For industrial production line loads, the BACP may be higher; while for non-critical loads such as lighting equipment, the BACP is relatively lower, providing differentiated cost considerations for subsequent optimization calculations.

[0101] Based on the adjustment response rate in the standardized load capacity profile data packet, the response delay cost parameter that quantifies the potential impact of resource response delay on stability is calculated through a nonlinear mapping function.

[0102] Specifically, based on the adjustment response rate in the standardized load capacity profile data package, a response delay cost parameter is calculated through a nonlinear mapping function. This function aims to quantify the potential impact of resource response delay on grid stability. For example, a slower response rate results in a greater risk of load fluctuations to the grid, and its cost parameters also increase non-linearly. An exemplary non-linear mapping function can be expressed as:

[0103] ,

[0104] in, It is a positive cost coefficient used to adjust for cost sensitivity. This is a preset minimum response rate threshold. This function ensures that costs increase sharply when the response rate approaches the threshold, thus prioritizing faster-responding load resources during optimization. For example... Figure 3 As shown, the nonlinear mapping relationship between response delay cost and response rate adjustment is illustrated, clearly depicting how the cost increases sharply as the response rate decreases, providing an intuitive cost consideration for optimization algorithms when making decisions.

[0105] Based on the pre-defined segmented user incentive policy, the user incentive cost parameters required to incentivize users to participate in regulation are calculated using a piecewise linear function.

[0106] Specifically, based on the preset segmented user incentive policy, a user incentive cost parameter is calculated using a piecewise linear function. This piecewise linear function provides users with tiered economic compensation based on parameters such as the load's regulating power and the regulating time period. For example, when the regulating power is at the first level, the compensation coefficient is... When in the second tier, the compensation coefficient is The calculation of user incentive cost parameters can be expressed as:

[0107] ,

[0108] in, , These are preset incentive coefficients. , These are segmented power thresholds. In this way, the cost of incentivizing user participation in regulation can be precisely quantified, providing a basis for optimization decisions.

[0109] Optionally, the step of comparing the actual adjustment result with the target adjustment value of the load power, calculating the adjustment deviation, and performing closed-loop compensation adjustment includes:

[0110] The actual load change, actual voltage change, and actual regulation response time are monitored after the execution of the load control command set to obtain the actual regulation results.

[0111] Specifically, after the load control command set is issued, the actual operating power data of each load control terminal is continuously collected at high frequency, and its changes are recorded. Simultaneously, this module also monitors voltage changes at key nodes in the distribution network to obtain actual voltage variation information. It also accurately times the actual adjustment response time from the issuance of the command to the start of load response. These data together constitute the actual adjustment results.

[0112] The actual adjustment results are compared with the target adjustment value of load power and the expected response time in multiple dimensions to obtain a comprehensive adjustment deviation that includes power deviation, voltage deviation and time deviation.

[0113] Specifically, the actual adjustment results will be compared with the target adjustment value for load power. and expected response time A multi-dimensional comparison is performed. The comparison process will yield deviations in three dimensions: power deviation. Voltage deviation and time deviation The formula for calculating the deviation can be expressed as:

[0114]

[0115]

[0116]

[0117] in, This is the actual load change. It is the actual voltage change. This is the preset optimal voltage value. These three deviations are weighted and fused to obtain a comprehensive adjustment deviation. This is used to comprehensively evaluate the effectiveness of regulation.

[0118] Based on the magnitude and nature of the comprehensive adjustment deviation, a hierarchical compensation decision algorithm is used to select low-power loads with high precision and fast response characteristics from the virtual resource set and generate compensation adjustment instructions.

[0119] Specifically, a compensation strategy is formulated based on the magnitude and nature of the overall regulation deviation. The algorithm explicitly defines "high precision" as a minimum adjustment step size of less than 1kW for the load control terminal and "fast response" as a regulation response rate greater than 5MW per second. When the overall regulation deviation is small, the algorithm will prioritize compensating low-power loads that meet the above definitions from the virtual resource set. The selection logic is as follows: first, sorting in descending order based on the regulation response rate, and then sorting in descending order based on the regulation margin index. This algorithm, through multi-objective decision-making, ensures the accuracy and smoothness of the compensation command, avoiding secondary fluctuations that may be caused by using high-power loads for coarse adjustments.

[0120] The compensation adjustment command is sent to the corresponding load control terminal to perform compensation adjustment and eliminate the overall adjustment deviation.

[0121] Specifically, the compensation decision algorithm transforms the selected compensation resource combinations into compensation adjustment instructions, explicitly specifying the compensation load control terminal ID, compensation power amount, and compensation duration. These instructions are then sent to the corresponding load control terminals via the local communication network to execute compensation adjustments, thereby precisely eliminating overall adjustment deviations and ensuring the ultimate achievement of the load balancing target.

[0122] Based on the same inventive concept, such as Figure 4 As shown, the present invention also provides a power system load balancing control system, the system comprising:

[0123] The data acquisition and fusion module is used to acquire global operating parameters and ultra-short-term load forecast data from the power grid dispatch center, collect standardized load capacity profile data packets in each power supply zone, and aggregate the standardized load capacity profile data packets to generate a virtual resource set representing the adjustable load resources of the region.

[0124] The optimization decision module is used to sort and combine resources in the virtual resource set based on the load power target adjustment value and the voltage data collected in real time from key nodes of the distribution network in the global operating parameters, with the goal of minimizing adjustment cost, and generate a load control instruction set.

[0125] The execution and monitoring module is used to send the load control command set to the corresponding load control terminal to perform load adjustment, and monitor the actual load change and voltage change to obtain the actual adjustment result;

[0126] The closed-loop compensation module is used to compare the actual adjustment result with the target adjustment value of the load power, calculate the adjustment deviation, and perform closed-loop compensation adjustment.

[0127] Example 1:

[0128] To verify the feasibility of this invention in practice, it was applied to a large city power grid company. Within the power grid area managed by this company, the traditional dispatching mode based on fixed operating mode tables is facing challenges due to the large-scale integration of new loads and power sources such as electric vehicle charging stations and distributed photovoltaic systems. Uneven power flow distribution on main lines and frequent local voltage exceedances are common problems. The company hopes to use the method of this invention to achieve global optimization and regional autonomous collaborative control of the entire network load, thereby improving the operational balance and security of the power grid.

[0129] In this embodiment, the power grid company deployed a global optimization decision-making module in its energy management system (EMS) at its dispatch center, and deployed regional autonomous agents in two typical power supply zones, A and B. Zone A is a commercial center, with loads mainly consisting of large central air conditioning units and shopping mall lighting; Zone B is a residential area, with loads mainly consisting of residential air conditioning units and electric vehicle charging stations. Each adjustable load is equipped with a smart control terminal. This embodiment selects operational data from a typical peak day in summer and compares it with control data from a day using traditional dispatching methods to verify the beneficial effects of the present invention.

[0130] During implementation, at 9:00 AM, the EMS (Electronic Management System) of the power grid dispatch center first collected the topology and real-time operating parameters of the main lines. Combined with the prediction results from the ultra-short-term load forecasting module, it was determined that a severe evening peak load would occur between 5:00 PM and 6:00 PM, with the load rate of some main lines expected to exceed 90%. To address this situation, the global optimization decision module was activated, performing optimization calculations with the objectives of overall load balancing and minimizing line losses. The weighting coefficient for overall load balancing... Set to 0.4, the weighting factor for minimizing line loss. The value was set to 0.6 to prioritize ensuring operational balance. The calculation results generated target load power adjustments: Zone A needs to reduce its load by 20MW, and Zone B needs to reduce its load by 15MW.

[0131] After the target value was issued, the regional autonomous agents in Zones A and B began executing local optimizations. The regional autonomous agent in Zone A aggregated load capacity profiles from loads such as central air conditioning in office buildings and lighting systems in shopping malls within its jurisdiction, forming a virtual resource set with a total adjustable capacity of 35MW. Upon receiving the 20MW reduction target, the agent aimed for the lowest possible adjustment cost and, based on real-time voltage data obtained from key nodes in the distribution network, set voltage safety constraints for optimization calculations. The optimization results prioritized the central air conditioning systems of several large office buildings with lower compensation costs for control, and generated a load control instruction set containing specific adjustment power and duration for execution.

[0132] At 18:05, after the command execution window ended, the closed-loop verification step was triggered. The regional autonomous agent in Area A calculated the actual load reduction to be 18.2MW by aggregating the power data from the load control terminal. This actual value deviated from the target value of 20MW by 1.8MW, with an adjustment deviation calculated as 9%. Since this deviation exceeded the preset 5% threshold, the fine-tuning mechanism was immediately activated. The regional autonomous agent selected the shopping mall's intelligent LED lighting system from the virtual resource set, which had a response speed within seconds and high adjustment accuracy, and issued a supplementary fine-tuning command to reduce the load by 1.8MW, ultimately bringing the adjustment result precisely to the target.

[0133] After the entire adjustment event is completed, the incentive management step is activated. Based on the recorded actual adjustment contributions of each user and according to the preset peak-hour compensation electricity price, the compensation due to each user is calculated to be 1200 yuan. Subsequently, the record containing the user's identifier, adjustment amount, duration, and compensation amount is packaged into a transaction and stored on the blockchain network, generating an immutable incentive certificate to ensure the fairness and transparency of incentive settlement.

[0134] In another test, a scenario was simulated where communication between the regional autonomous agent in Area A and the dispatch center was interrupted for 30 minutes. Upon detecting the communication interruption, the system immediately switched to emergency autonomous mode. Based on the last received instruction to "maintain the 10MW reduction," and combined with local real-time voltage monitoring data, it continued load control, ensuring the basic stability of the regional power grid. After communication was restored, the agent automatically uploaded all cached operational status records from the past 30 minutes to the dispatch center, achieving seamless data synchronization.

[0135] Table 1 Comparison of Load Balancing Control Effects

[0136]

[0137] Table 2A: Data on the Accuracy of Autonomous Control in Partitions

[0138]

[0139] Table 3 Example of User Incentive and Emergency Response Records

[0140]

[0141] As can be seen from the data in Tables 1 to 3 above, the method of the present invention has achieved significant technical effects. The data in Table 1 shows that, compared with the traditional scheduling method, the method of the present invention reduces the maximum load rate of the line during peak hours from the dangerous range of over 90% to a safe level of below 78%, while the standard deviation of the load rate of the trunk line is significantly reduced from 0.28 to 0.11, proving the superiority of the present method in achieving load balancing across the entire network.

[0142] Table 2 clearly demonstrates the effectiveness of the closed-loop verification and fine-tuning mechanism. The initial 9% deviation generated by open-loop control was corrected to 0.5% through closed-loop feedback and high-precision fine-tuning, which is far below the control requirements. This ensures the accurate achievement of the global optimization goal and reflects the robustness and precision of the system control.

[0143] The records in Table 3 verify the incentive mechanism and fault tolerance of this invention. The incentive ensures fairness and transparency in user participation, while the emergency handling mechanism for communication interruptions guarantees reliable operation under extreme failures, demonstrating the great potential of this invention in building an efficient, reliable, and economically feasible next-generation load control system.

[0144] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0145] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A power system load balancing control method, characterized in that, The method includes: The global operating parameters and ultra-short-term load forecast data of the power grid dispatch center are obtained. Through power flow calculation and with the optimization objectives of overall load balance and line loss minimization, the load power target adjustment values ​​of each power supply zone are obtained. Standardized load capacity profile data packets containing real-time operating power, maximum adjustable power, and sustainable adjustment time are collected from each power supply zone and aggregated to generate a virtual resource set representing the adjustable load resources of the region. Specifically, real-time operating data and load characteristic parameters of load control terminals are collected, and the power characteristics, time characteristics, and response characteristics of the load control terminals are extracted. Using a data standardization processing method, the extracted parameters are converted into standardized load capacity profile data packets containing real-time operating power, maximum adjustable power, sustainable adjustment time, and adjustment response rate. These standardized load capacity profile data packets are parsed and aggregated, and a virtual resource set representing the adjustable load resources of the region is generated through a multi-dimensional resource aggregation algorithm. Based on the target load power adjustment value, and combined with the voltage data collected in real time from key nodes of the distribution network in the global operating parameters, the resources in the virtual resource set are sorted and combined for optimization with the lowest adjustment cost as the optimization objective, generating a load control instruction set. Specifically, a multi-dimensional dynamic ranking model for load adjustment priority is established based on the power characteristics, time characteristics, and response characteristics of each resource in the virtual resource set; a adjustment cost model considering load response delay and user incentives is established to quantify the adjustment cost of each resource in the virtual resource set; under the safety constraints determined by the real-time voltage data of the key nodes of the distribution network, the resources in the virtual resource set are combined for optimization with the adjustment cost model as the optimization objective, combined with the results of the dynamic ranking model for load adjustment priority, resulting in an optimized adjustment resource combination; the optimized adjustment resource combination is then transformed into a load control instruction set containing specific control objects and control parameters. The load control command set is sent to the corresponding load control terminal to perform load adjustment, monitor the actual load change and voltage change, compare the actual adjustment result with the target adjustment value of load power, calculate the adjustment deviation and perform closed-loop compensation adjustment. The parsing and aggregation of the standardized load capacity profile data package includes: performing a weighted summation calculation on the maximum adjustable power to generate the total regional adjustment capacity; calculating the average value of the sustainable adjustment time to obtain the average sustainable duration; and filtering the adjustment response rate by minimum value to generate the fastest response rate. The steps of establishing a multi-dimensional dynamic ranking model for load regulation priorities include: calculating a load regulation margin index characterizing the current adjustable potential of resources based on the real-time operating power and the total regional regulation capacity; calculating a load regulation durability index and a load regulation sensitivity index characterizing the stability and response speed of resource regulation based on the average sustainability time and the fastest response rate; using a multi-index weighted fusion algorithm to comprehensively score the load regulation margin index, the load regulation durability index, and the load regulation sensitivity index to obtain the final regulation priority of each resource, and ranking the resources in the virtual resource set according to the regulation priority; The establishment of a regulation cost model considering load response delay and user incentives includes: for each resource in the virtual resource set, obtaining a basic regulation cost parameter representing the inherent regulation cost per unit power of the resource; based on the regulation response rate in the standardized load capacity profile data packet, calculating a response delay cost parameter that quantifies the potential impact of resource response delay on stability through a nonlinear mapping function; and, according to a preset segmented user incentive policy, calculating a user incentive cost parameter that incentivizes users to participate in regulation through a piecewise linear function.

2. The power system load balancing control method according to claim 1, characterized in that, The method further includes: When communication between the power supply zone and the power grid dispatch center is interrupted, emergency autonomous optimization control is executed based on local real-time monitoring data, and an emergency autonomous operation status record is generated.

3. The power system load balancing control method according to claim 2, characterized in that, The execution of emergency autonomous optimization control includes: In emergency autonomous mode, the power supply zone continuously monitors the real-time voltage, current and power changes of key nodes in the distribution network within its jurisdiction, forming a snapshot of the local real-time operating status. Based on the local real-time operating status snapshot, the emergency control strategy library is used to perform strategy matching and filtering to identify emergency control strategies that can solve the local load imbalance problem. Based on the emergency control strategy, an emergency load control instruction set is generated and sent to the corresponding load control terminal to perform load adjustment. The operation process and results are then used to generate an emergency autonomous operation status record.

4. The power system load balancing control method according to claim 1, characterized in that, The step of comparing the actual adjustment result with the target adjustment value of the load power, calculating the adjustment deviation, and performing closed-loop compensation adjustment includes: The actual load change, actual voltage change, and actual regulation response time are monitored after the execution of the load control command set to obtain the actual regulation results. The actual adjustment results are compared with the target adjustment value of load power and the expected response time in multiple dimensions to obtain a comprehensive adjustment deviation that includes power deviation, voltage deviation and time deviation. Based on the magnitude and nature of the comprehensive adjustment deviation, a hierarchical compensation decision algorithm is used to select low-power loads with high precision and fast response characteristics from the virtual resource set and generate compensation adjustment instructions. The compensation adjustment command is sent to the corresponding load control terminal to perform compensation adjustment and eliminate the overall adjustment deviation.

5. A power system load balancing control system, characterized in that, The system is applied to a power system load balancing control method as described in any one of claims 1-4, and the system comprises: The data acquisition and fusion module is used to obtain global operating parameters and ultra-short-term load forecast data from the power grid dispatch center, collect standardized load capacity profile data packages from each power supply zone, and aggregate these standardized load capacity profile data packages to generate a virtual resource set representing the adjustable load resources of the region. Specifically, it collects real-time operating data and load characteristic parameters from load control terminals, and extracts the power characteristics, time characteristics, and response characteristics of the load control terminals. Using a data standardization processing method, the extracted parameters are converted into standardized load capacity profile data packages containing real-time operating power, maximum adjustable power, sustainable adjustment time, and adjustment response rate. The standardized load capacity profile data packages are then parsed and aggregated, and a virtual resource set representing the adjustable load resources of the region is generated through a multi-dimensional resource aggregation algorithm. The optimization decision module is used to sort and combine resources in the virtual resource set based on the target load power adjustment value and in combination with voltage data collected in real time from key nodes of the distribution network in the global operating parameters, with the goal of minimizing adjustment cost. This generates a load control instruction set. Specifically, a multi-dimensional dynamic ranking model for load adjustment priority is established based on the power characteristics, time characteristics, and response characteristics of each resource in the virtual resource set; an adjustment cost model considering load response delay and user incentives is established to quantify the adjustment cost of each resource in the virtual resource set; under the safety constraints determined by the real-time voltage data of the key nodes of the distribution network, the resources in the virtual resource set are combined and optimized using the adjustment cost model as the optimization objective and in combination with the results of the dynamic ranking model for load adjustment priority, resulting in an optimized adjustment resource combination; and the optimized adjustment resource combination is then transformed into a load control instruction set containing specific control objects and control parameters. The execution and monitoring module is used to send the load control command set to the corresponding load control terminal to perform load adjustment, and monitor the actual load change and voltage change to obtain the actual adjustment result; The closed-loop compensation module is used to compare the actual adjustment result with the target adjustment value of the load power, calculate the adjustment deviation, and perform closed-loop compensation adjustment. The parsing and aggregation of the standardized load capacity profile data package includes: performing a weighted summation calculation on the maximum adjustable power to generate the total regional adjustment capacity; calculating the average value of the sustainable adjustment time to obtain the average sustainable duration; and filtering the adjustment response rate by minimum value to generate the fastest response rate. The steps of establishing a multi-dimensional dynamic ranking model for load regulation priorities include: calculating a load regulation margin index characterizing the current adjustable potential of resources based on the real-time operating power and the total regional regulation capacity; calculating a load regulation durability index and a load regulation sensitivity index characterizing the stability and response speed of resource regulation based on the average sustainability time and the fastest response rate; using a multi-index weighted fusion algorithm to comprehensively score the load regulation margin index, the load regulation durability index, and the load regulation sensitivity index to obtain the final regulation priority of each resource, and ranking the resources in the virtual resource set according to the regulation priority; The establishment of a regulation cost model considering load response delay and user incentives includes: for each resource in the virtual resource set, obtaining a basic regulation cost parameter representing the inherent regulation cost per unit power of the resource; based on the regulation response rate in the standardized load capacity profile data packet, calculating a response delay cost parameter that quantifies the potential impact of resource response delay on stability through a nonlinear mapping function; and, according to a preset segmented user incentive policy, calculating a user incentive cost parameter that incentivizes users to participate in regulation through a piecewise linear function.

Images