Dynamic cluster planning and control method and system for distribution network based on electrical indicators

Abstract

The present invention discloses a distribution network dynamic cluster planning and control method and system based on electrical indicators, which relates to the field of cluster collaborative control technology, and includes the following steps: obtaining device response data, and constructing a device dynamic response model, extracting the characteristics of the device under different adjustment signals through the model and generating a device flexibility index set; constructing a cluster boundary flexible layer state mapping matrix according to the device flexibility index set, and coupling it with the cluster external control interface to generate an inter-layer flexibility response interface model; based on the response interface model, constructing a hierarchical decoupling control strategy with upper and lower layer strategy linkage, and driving controllers at all levels to execute consistent control actions according to the control strategy; the present invention solves the problems of inconsistent control granularity, strategy conflict and control failure inside and outside the cluster by constructing a flexible state mapping matrix and using the response interface model as an intermediary to achieve flexible capability matching and control strategy linkage between the upper and lower layers, thereby improving the dynamic adjustment capability of the cluster.

Description

Technical Field

[0001] The present invention relates to the technical field of cluster collaborative control, and more specifically, to a method and system for dynamic cluster planning and control of a distribution network based on electrical indicators. Background Art

[0002] With the increasing interaction between sources and loads in new power systems, distribution networks are gradually showing operational characteristics of multi-source access, a high proportion of distributed energy resources connected to the grid, and increased flexible load participation. In this context, clustered collaborative control and dynamic response scheduling for flexible load resources have gradually become important research directions for improving the regulation capability and operational efficiency of distribution networks.

[0003] For example, the invention patent with announcement number CN106950840A discloses a hierarchical distributed coordinated control method for an integrated energy system for power grid peak shaving, which includes the following steps: 1) the lower-level control system performs self-optimization control on users; 2) the upper-level control system collects power information at the gateway to determine whether the overall peak value of the energy system meets actual demand; 3) the upper-level park increases the direct-regulated energy storage output and generator output; 4) the upper-level control system again determines whether the power at the gateway exceeds the limit; 5) the upper-level control system issues instructions to the lower-level adjustable users, and the interactive users reasonably adjust their own loads; 6) the upper-level control system continues to determine whether the power at the gateway exceeds the limit; 7) the upper-level park control system issues instructions to the lower-level interruptible users, and the interactive users reasonably interrupt their own loads. The present invention solves the problem of mutual coupling and difficulty in coordination and complementarity among multiple energy sources, shaving peaks and filling valleys, and achieving friendly interaction with the power grid.

[0004] For example, the invention patent with announcement number CN116700086A discloses an energy station control system, method, device, and medium, including a hierarchical control module for obtaining environmental data and user instructions of each energy station, and generating an intelligent control solution based on the environmental data and user instructions; a communication module for obtaining data and intelligent control solutions of each energy station, and optimizing the intelligent control solution based on the data of each energy station, and distributing the optimized intelligent control solution to each energy station; and a voltage stabilization module for obtaining voltage stability indicators and intelligent control solutions, thereby controlling the voltage of each bidirectional converter. The present invention generates an intelligent control solution by optimizing user instructions through a hierarchical control module, and further optimizes the intelligent control solution through a communication module, thereby adapting to a multi-energy flow coupled environment, efficiently processing complex data from different sources, effectively improving the operating efficiency of the energy system, and effectively solving the voltage fluctuation caused by the hierarchical control module by setting a voltage stabilization module.

[0005] The above disclosed technical solutions have at least the following technical problems:

[0006] While the control granularity of devices within a cluster is fine, the large-scale inter-cluster control lacks an effective hierarchical decoupling control strategy. This makes it difficult for lower layers to quickly follow the upper-layer strategy. Lower-layer control behavior in turn affects upper-layer target constraints, generating feedback conflicts. Furthermore, existing upper and lower-layer control systems are tightly coupled in architecture but lack flexible interface support, making it impossible to effectively map and decouple target parameters from responsiveness. This makes linkage strategies difficult to implement and can even lead to policy conflicts and control failures. This present invention proposes a solution to these problems. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a distribution network dynamic cluster planning and control method and system based on electrical indicators. By constructing a flexible state mapping matrix and using a response interface model as an intermediary to achieve flexible capability matching and control strategy linkage between the upper and lower layers, the problems of inconsistent control granularity inside and outside the cluster, strategy conflicts and control failures are solved, and the dynamic adjustment capability of the cluster is improved.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A dynamic cluster planning and control method for distribution networks based on electrical indicators includes the following steps: obtaining device response data, constructing a dynamic response model for the equipment, extracting multidimensional device characteristic parameters through the response model, and generating a set of device flexibility indicators; constructing a mapping matrix that characterizes the state of the flexible layer at the cluster boundary based on the device flexibility indicator set; dynamically coupling the layered mapping matrix with the cluster's external control interface to generate an inter-layer flexibility response interface model; and constructing a hierarchical decoupling control strategy for the upper-layer scheduling strategy and the lower-layer device control based on the response interface model, and driving the cluster's multi-level controllers to perform collaborative consistency control actions through the control strategy.

[0010] In a preferred embodiment, device response data is obtained and a device dynamic response model is constructed, specifically: the device response data of each flexible load device in the cluster under different control instructions is obtained, and the device response data includes control input, output response quantity and disturbance scenario label; the device response data is standardized and preprocessed, and the dynamic response model of each device is established for the preprocessed data through the mapping relationship between control input and output response quantity.

[0011] In a preferred embodiment, a dynamic response model is established for each device based on the mapping relationship between the control input and output response quantity for the preprocessed data, specifically: according to the parallel convolutional neural network and recurrent neural network structure, the control input and output response quantity feature sequences are extracted through a multi-window sliding sampling method, and the multi-window sliding sampling method includes setting a short-time window to extract local features and setting a long-time window to extract trend features. The convolutional neural network processes the short-time window, and the recurrent neural network processes the long-time window; the disturbance scene label is embedded as a conditional vector and spliced ​​and trained with the feature sequence to construct a dynamic response model.

[0012] In a preferred embodiment, the characteristics of the device under different adjustment signals are extracted through a model and a set of device flexibility indicators is generated, specifically: the dynamic response model is disturbed to extract the dynamic characteristics of each device; multiple response curves are generated according to the dynamic characteristics, and a response curve cluster for each device is constructed; based on the response curve cluster, the flexibility indicator set of each device is calculated, and the flexibility indicator set includes dynamic adjustment margin, time response capability, continuous adjustment capability and elastic recovery capability.

[0013] In a preferred embodiment, a cluster boundary flexible layer state mapping matrix is ​​constructed based on the equipment flexibility index set, specifically: based on the degree of discreteness of the flexibility index set, the initial weight of the flexibility index is calculated by the entropy weight method; the system frequency deviation characteristic quantity is obtained, and the characteristic quantity is mapped into a level correction factor based on a preset fuzzy rule library, and the initial weight is dynamically corrected; the equipment flexibility index set is standardized to obtain an initial matrix, and is fused with the corrected weight to construct a flexible layer mapping vector; the flexible layer mapping vector is aggregated based on the topological hierarchy and physical connection method of the equipment to generate a flexible layer state mapping matrix.

[0014] In a preferred embodiment, coupling is performed with the cluster external control interface to generate an inter-layer flexibility response interface model, specifically: obtaining the cluster external control interface characteristics, the external control interface characteristics including response delay tolerance, load adjustment granularity, minimum callable flexibility threshold, and adjustment duration period requirements; constructing a flexible response adaptation function through the flexible layer state mapping matrix and the external control interface characteristics; coupling the layer with the response adaptation function to construct a flexibility response interface model.

[0015] In a preferred embodiment, based on the response interface model, a hierarchical decoupling control strategy with linkage between upper and lower layer strategies is constructed, specifically: the control target vector of the upper control system is obtained and input into the response interface model, the control target vector includes a power regulation target, a response time constraint and an adjustment continuity constraint; based on the flexible response adaptation function, the adjustable capability vector group is extracted through the matching relationship between the flexible layer state mapping matrix and the control target vector, the adjustable capability vector group includes a dynamic power adjustment range, a time response capability commitment value, a sustainable adjustment time commitment value and an elastic recovery time commitment value; according to the adjustable capability vector group, the decoupling mapping parameters between the upper and lower layers are identified, the decoupling mapping parameters include an adjustment redundancy interval, a response time lag matching interval and an adjustment continuity overlap interval; the decoupling mapping parameters are passed to the lower controller, and several local control strategy sets are generated within the range of its local controllable devices; each local control strategy set is evaluated and fed back to the upper controller for optimization and execution of consistency control actions.

[0016] In a preferred embodiment, the adjustable capability vector group is extracted through the matching relationship between the flexible layer state mapping matrix and the control target vector, specifically: based on the dynamic adjustment margin, the power adjustment target in the control target vector is matched to obtain the dynamic power adjustment range; based on the time response capability, the response time constraint in the control target vector is compared with the boundary to obtain the time response capability commitment value; based on the continuous adjustment capability, the window is matched with the adjustment continuity constraint in the control target vector to obtain the sustainable adjustment time commitment value; based on the elastic recovery capability, the disturbance recovery time requirement in the external control interface characteristics is verified to obtain the elastic recovery time commitment value.

[0017] In a preferred embodiment, each local control strategy set is evaluated and fed back to the upper-level controller for optimization and execution of consistency control actions, specifically: key performance indicators corresponding to the local control strategy set are obtained and an adjustment strategy evaluation model is constructed; the output results of the adjustment strategy evaluation model are fed back to the upper-level controller; the upper-level controller receives several model output results and screens out the optimal control strategy set based on the hierarchical analysis method of the main target; the target control vector of the upper-level controller is corrected according to the optimal control strategy set to generate a linkage control strategy; and each level of controller executes its own control action according to the linkage control strategy.

[0018] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0019] 1. By introducing a multi-scale deep neural network to construct a dynamic response model at the granularity of a single device, it is possible to accurately model the transient and long-term response characteristics of various flexible loads under various disturbance scenarios. By extracting disturbance features, a set of flexibility indicators is constructed, forming a full-dimensional characterization of the dynamic adjustment capabilities of the equipment. Combining the entropy weight method with the fuzzy level correction mechanism to construct a flexible layer state mapping matrix, this achieves structured and weighted dynamic modeling of flexible resources in the cluster. Furthermore, by constructing a flexible response adaptation function and an inter-layer response interface model, an adaptive coupling mechanism is formed between the cluster's flexible characteristics and upper-level control requirements, significantly improving the scheduling and matching accuracy of flexible resources and the real-time and stability of system regulation. This provides a highly scalable algorithmic foundation and model framework for achieving refined, multi-level coordinated control of flexible resources.

[0020] 2. By constructing an upper and lower layer strategy linkage mechanism based on a flexible response interface model, precise decoupling and dynamic coordination between multi-level control systems is achieved. The adjustable capability vector group is parsed through the flexible layer state mapping matrix, and the decoupling mapping parameters are extracted in combination with the time window matching algorithm, so that the upper-level control objectives can be accurately transmitted to the lower-level controller based on the actual response capability, avoiding problems such as instruction rigidity and response conflicts under traditional strategies. The lower-level controller generates a strategy set based on local status and equipment capabilities, and cooperates with multi-dimensional performance indicator evaluation and upper-level global optimization based on hierarchical analysis method to realize a closed-loop control process from strategy setting, boundary identification to consistent execution. The overall solution significantly improves the system's regulation efficiency, strategy adaptability and controllability of flexible resource response, and provides a solution path with a clear structure, complete model and strong practicality for the coordinated control of multi-level resources in new power systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flow chart of a method for dynamic cluster planning and control of a distribution network based on electrical indicators provided in an embodiment of the present application.

[0022] Figure 2 A schematic diagram of the process structure of a dynamic cluster planning control system for a distribution network based on electrical indicators provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] The following will provide a clear and complete description of 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. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0024] Example 1, Figure 1The flow chart of the method for dynamic cluster planning and control of a distribution network based on electrical indicators provided in an embodiment of the present application includes the following steps:

[0025] S1, obtain the device response data and build a dynamic response model of the device. Through the model, extract the characteristics of the device under different adjustment signals and generate a set of device flexibility indicators.

[0026] In this embodiment, a dynamic device response model accurately characterizes its transient response characteristics under different regulation commands, enabling control strategies to be refined down to the granularity of individual devices, rather than simply targeting the average behavior of a load group. This enables device response prediction and dynamic estimation of regulation margins, providing a priori support for control strategies. Furthermore, this model enables personalized dynamic regulation of each flexible device in the cluster. This supports subsequent flexibility feature extraction and provides a data source for the flexibility index set.

[0027] Furthermore, the device response data is obtained and a dynamic response model of the device is constructed, specifically:

[0028] Obtain device response data for each flexible load device in the cluster under different control instructions (including switch switching, power disturbance, and voltage step). The device response data includes control input, output response amount, and disturbance scenario label.

[0029] Performing standardized preprocessing on the device response data to ensure data quality suitable for dynamic modeling, wherein the standardized preprocessing includes data denoising, time series alignment, feature regularization, and outlier removal;

[0030] The dynamic response model of each device is established based on the mapping relationship between the control input and output response quantity for the preprocessed data.

[0031] Among them, device response data refers to the response behavior exhibited by the device under the action of external control, and records the input and output variables of the device in the form of time series data. These data are used to reflect the dynamic characteristics of the device and serve as the basis for model training and evaluation. Control input refers to the external stimulus or operating instruction applied to the device by the dispatching or control system. It is a direct variable that affects the change of the device state, including changes in the power set value and the switching quantity of the control signal. The output response quantity refers to the observable response behavior exhibited by the device within a certain time window after receiving the control input, including active power, reactive power, and frequency response. The disturbance scenario label is used to identify the experimental or operating scenario corresponding to a certain segment of device response data, providing contextual environment information for model training. Disturbance scenario types include: normal operation, power disturbance scenario, voltage disturbance scenario, and frequency disturbance scenario.

[0032] Furthermore, the pre-processed data is used to establish a dynamic response model for each device based on the mapping relationship between the control input and the output response, specifically:

[0033] Based on the parallel convolutional neural network and recurrent neural network structure, a multi-window sliding sampling method is used to extract the characteristic sequences of control input and output response quantities. The multi-window sliding sampling method includes setting a short-term window to extract local features and a long-term window to extract trend features. The convolutional neural network processes the short-term window to quickly extract high-frequency response features within a short period after the disturbance, and the recurrent neural network processes the long-term window to capture the recovery trend and memory path of the device over a long time scale.

[0034] The disturbance scene label is embedded as a conditional vector and concatenated and trained with the feature sequence to build a dynamic response model.

[0035] The specific calculation formula of the dynamic response model is as follows:

[0036]

[0037] Where, For in time The device output response at the moment, Extract local features for short time windows, Extract trend features for long time windows, is the perturbation scene label, is the output lag order, is the input lag order, For the The device output response at the moment, For the Control input at any moment.

[0038] It should be noted that by introducing parallel convolutional neural network and recurrent neural network structures, and combining them with a multi-window sliding sampling method, the multi-scale modeling of the dynamic response characteristics of flexible load equipment is effectively achieved. Convolutional neural networks have the ability to quickly capture high-frequency disturbance responses within a short-term window, and can accurately characterize the immediate reaction of the equipment to sudden control instructions; recurrent neural networks are good at modeling temporal dependencies in long-term windows, which helps to restore the recovery process and memory behavior of the equipment response. In addition, by embedding the disturbance scenario label as a conditional vector to participate in training, the model's ability to distinguish and generalize dynamic characteristics under different control scenarios is enhanced, significantly improving the model's adaptability and prediction accuracy under multi-scenario conditions, and providing a solid data foundation and model support for subsequent response evaluation and flexibility analysis.

[0039] Furthermore, by modeling the device response and measuring its flexibility, the device flexibility status can be dynamically fed back to the scheduling control layer to generate an adaptive target adjustment mechanism, thus realizing the feedback closed-loop control logic of “response → adjustment → re-response”.

[0040] The model extracts the characteristics of the equipment under different adjustment signals and generates a set of equipment flexibility indicators, specifically:

[0041] The dynamic response model is perturbed by injecting multiple adjustment signal scenarios, and the dynamic characteristics of each device are extracted. The dynamic characteristics include response speed, device power output, adjustment amplitude and steady-state recovery time, and input control variable. The multiple adjustment signal scenarios include step disturbance (for testing system fast response), ramp change (for testing output following capability), random disturbance (for testing robustness and adaptability), and sudden stop and start (for testing dynamic recovery capability under nonlinear switching);

[0042] generating a plurality of response curves according to the dynamic characteristics and constructing a response curve cluster for each device;

[0043] Based on the response curve cluster, a flexibility index set of each device is calculated, wherein the flexibility index set includes dynamic adjustment margin, time response capability, continuous adjustment capability and elastic recovery capability.

[0044] The dynamic adjustment margin is specifically calculated as follows:

[0045]

[0046] The specific calculation formula of the time response capability is as follows:

[0047]

[0048] The specific calculation formula of the continuous adjustment capability is as follows:

[0049]

[0050] The specific calculation formula of the elastic recovery capacity is as follows:

[0051]

[0052] Where, To dynamically adjust the margin, is the time response capability, For continuous adjustment capabilities, For elastic recovery, For the current signal adjustment scenario, To adjust the signal scene set, For the The power output of the device in each scenario, is the rated capacity when the disturbance occurs, For the current moment, is the disturbance start time, is the rated active power of the equipment, For devices in disturbance scenarios The time when its power output reaches 90% for the first time is For the scene The change range of the input control quantity applied to the device under For disturbance scenarios The target output value of the device is The continuous period of time during which the device output remains within the tolerance band of the target value. is the tolerance band coefficient, is the time it takes for the device to first recover to the steady state before the disturbance, For disturbance scenarios The time when the control instruction ends, is the penalty factor for recovering the error variance, is the steady-state error variance term.

[0053] It should be noted that the dynamic adjustment margin refers to the ratio of the maximum power adjustment amplitude achieved by the device in the response curve cluster per unit time to the rated capacity, which is used to measure the upper limit of the device's rapid output capability; the time response capability refers to the shortest time required for the device to reach and maintain a certain proportion (90%) of the target disturbance value from the occurrence of the disturbance to the first output, indicating the response speed limit; the continuous adjustment capability refers to the longest time period that the device can maintain the target output within the set tolerance band (±δ%) after the adjustment action is completed, reflecting the adjustment stability and continuous energy supply capability; the elastic recovery capability refers to the ability of the device to return to the original steady state after the adjustment is completed, taking into account the two dimensions of recovery time and error stability.

[0054] S2, based on the equipment flexibility index set, constructs a state mapping matrix representing the cluster boundary flexibility layer, and couples it with the cluster external control interface to generate an inter-layer flexibility response interface model.

[0055] In this embodiment, by constructing a cluster boundary flexible layer state mapping matrix, it is possible to effectively connect the multi-dimensional response relationship between the flexible characteristics of the equipment within the cluster and the external control requirements, realize dynamic adaptation and optimization of the inter-layer flexible control interface, and improve the accuracy of the scheduling instructions and the efficiency of cluster flexibility utilization.

[0056] Furthermore, a cluster boundary flexibility layer state mapping matrix is ​​constructed based on the equipment flexibility index set, specifically:

[0057] Get flexibility index set The initial weight of the flexibility index is calculated by the entropy weight method based on the degree of dispersion in the equipment group;

[0058] Obtain the characteristic quantity of the current operation state of the cluster, map the characteristic quantity into a level correction factor based on a preset fuzzy rule library, and dynamically correct the initial weight using the fuzzy level correction factor. The characteristic quantity of the operation state is the frequency deviation of the real-time acquisition system.

[0059] The device flexibility index set is normalized to obtain the initial matrix, which is then fused with the modified weights to construct the flexibility layer mapping vector for each device.

[0060] The flexible layer mapping vectors are aggregated based on the topological level and physical connection mode of the devices to generate a flexible layer state mapping matrix.

[0061] The specific calculation formula for the discrete degree is as follows:

[0062]

[0063] The specific calculation formula of the initial weight is as follows:

[0064]

[0065] Where, is the information entropy, which measures the numerical dispersion of this indicator in all devices, where k=1,2,3,4. is the total number of devices involved in the flexibility evaluation, For equipment The proportion of each flexibility index in the normalized value of the flexibility index set, is the initial weight.

[0066] The specific calculation formula of the modified weight is as follows:

[0067]

[0068] Where, is the modified weight, is the initial weight, is the weight after fuzzy correction.

[0069] It should be noted that the fuzzy level correction factor is used to dynamically adjust the entropy weight. This factor is set based on the current operating status of the cluster and has the following characteristics: If the system is running in a normal state: , indicating that the weight is not adjusted; running in a tense or emergency state: the corresponding flexibility index set , to increase its weight; other indicators can be set , to weaken the impact.

[0070] The specific calculation formula of the flexible layer mapping vector is as follows:

[0071]

[0072] Where, is the flexible layer mapping vector, is the modified weight, is the flexibility index set.

[0073] The flexible layer state mapping matrix is ​​as follows:

[0074]

[0075] Where, is the flexible layer state mapping matrix, The number of rows represented by the number of logical levels of the cluster boundary, is the number of columns represented by the total number of flexible devices, is the flexible layer mapping vector, For the The device in The flexible layer mapping vector of each boundary layer.

[0076] It should be noted that the flexible layer mapping vectors are aggregated based on the topological level and physical connection mode of the device to generate a flexible layer state mapping matrix. This can be understood as the topological level of the device can be used to obtain the number of logical levels of the cluster boundary and the set of devices at each layer. Considering the impact of the physical connection relationship on the device flexibility index set, the flexible layer mapping vector of each device is constructed using the corrected weights. In this way, each device obtains a flexible layer mapping vector that has been corrected by the connection relationship. The flexible layer mapping vectors of each layer are combined to obtain the final flexible layer state mapping matrix. The flexible layer mapping vector represents its weighted comprehensive capability under all flexibility indicators.

[0077] Furthermore, it is coupled with the cluster external control interface to generate an inter-layer flexibility response interface model, specifically:

[0078] Obtain the external control interface characteristics of the upper control system outside the cluster to the cluster boundary, the external control interface characteristics including the target power adjustment amount initiated by the interface , Maximum acceptable response time , expected continuous adjustment time , Maximum tolerable recovery time ;

[0079] A flexible response adaptation function is constructed based on the flexible layer state mapping matrix and the external control interface characteristics to measure the response adaptability of the edge device to the control interface characteristics.

[0080] The response adaptation functions of all layers and all external interfaces are coupled to build an inter-layer flexibility response interface model. This model represents the degree of adaptation between the flexibility mapping and the control requirements in the form of a matching matrix. The layer boundary can be The total amount of flexible response capability provided by each interface task is used for hierarchical deployment of control strategies.

[0081] The specific calculation formula of the flexible response adaptation function is as follows:

[0082]

[0083] Where, is the flexible response adaptation function, is the flexible layer state mapping matrix, External control interface characteristic adaptation coefficient, indicating the matching degree between the device response capability and the interface requirements. is the number of columns represented by the total number of flexible devices, For external interface.

[0084] The specific calculation formula of the response interface model is as follows:

[0085]

[0086] Where, To respond to the interface model, For the The number of rows represented by the logical level of each cluster boundary, For external interface.

[0087] It should be noted that the inter-layer flexibility response interface model is a two-dimensional matrix composed of the response adaptation functions of all layers and all external interfaces, which represents the total amount of flexible response capability that the boundary can provide for interface tasks, and is used for the upper-level control system to perceive the adjustable capability of the cluster boundary in real time.

[0088] Among them, the external control interface characteristic adaptation coefficient is calculated as follows:

[0089]

[0090] Where, is the maximum adjustment capability of the device, is the actual minimum response time of the device, Continuously adjust the time for the device, For device recovery capabilities, 、 、 、 is the weight coefficient, It is a response compatibility function between external control interface characteristics and device capability values.

[0091] The response compatibility function indicates whether the device capability is sufficient to meet the demand: the device's maximum adjustment capability, the device's actual minimum response time, the device's sustainable adjustment time, and the device's recovery capability. It can be expressed in the following form: , The perturbation term to prevent division by zero is a very small positive number used to avoid division by zero. The actual minimum response time, sustainable adjustment time, and recovery capability of the remaining devices are similar.

[0092] S3, based on the response interface model, constructs a hierarchical decoupling control strategy of the upper-level scheduling strategy and the lower-level device control, and drives the cluster multi-level controller to perform collaborative consistency control actions through the control strategy.

[0093] In this embodiment, the interface model connects the strategy mapping and feedback channels between the upper and lower layer controllers, opening up the linkage between the upper and lower layer strategies and facilitating control granularity matching. The upper-level control system is the control logic entity at the top of the control instruction chain in the hierarchical control architecture. It is responsible for generating control strategies or target parameters based on the grid operation objectives from a global system perspective and transmitting strategy information to the lower-level control system through the interface model. The lower-level control system refers to various specific control units at the station control level, device level, and equipment level (such as distributed energy storage controllers, flexible load regulators, substation controllers, etc.), which are responsible for strategy execution, action implementation, and data feedback.

[0094] Furthermore, based on the response interface model, a hierarchical decoupling control strategy is constructed for the upper-layer scheduling strategy and the lower-layer device control. The control strategy drives the cluster multi-level controllers to perform coordinated consistency control actions, specifically:

[0095] Obtaining a control target vector set by the upper-level control system in the current operation cycle, wherein the control target includes a power regulation target, a response time constraint, a regulation continuity constraint, and a disturbance recovery time;

[0096] Inputting the control target vector into the response interface model, calculating the matching relationship between the flexible layer state mapping and the target vector through the flexible response adaptation function, and generating an adjustable capability vector group, wherein the adjustable capability vector group includes a dynamic power adjustment range, a time response capability commitment value, a sustainable adjustment capability commitment value, and an elastic recovery time commitment value;

[0097] Identifying decoupling mapping parameters between the upper-layer control target and the lower-layer response capability using a time window matching algorithm based on the adjustable capability vector group, wherein the decoupling mapping parameters include an adjustment redundancy interval, a response time lag matching interval, and an adjustment persistence overlap interval;

[0098] The decoupling mapping parameters are sent to the lower-level controller to guide the strategy boundary definition and dynamic correction range determination when the lower-level controller generates the strategy. In combination with the local operating status and device adjustment capabilities, several local control strategy sets are generated within the range of its locally controllable devices.

[0099] Each local control strategy set is evaluated in multiple dimensions, and the evaluation results are fed back to the upper-level controller for optimization and execution of consistency control actions.

[0100] The matching relationship between the flexible layer state mapping and the target vector is calculated through the flexible response adaptation function to generate an adjustable capability vector group, specifically:

[0101] Dynamic adjustment margin based on equipment The power regulation target in the upper control target vector Perform margin matching to generate dynamic power adjustment range , the specific formula is as follows:

[0102] in, ;

[0103] Based on the time response capability of the device With the response time constraint in the target vector Perform boundary comparisons and generate time response capability commitments , the specific formula is as follows:

[0104] in, ;

[0105] Where, is the safety margin factor;

[0106] Continuous device-based adjustment capabilities and the regulation persistence constraint in the control target vector Perform window matching to generate a sustainable regulation capacity commitment value , the specific formula is as follows:

[0107] ;

[0108] If it is 0, the ratio meets the requirements, the same below;

[0109] Device-based resilience Disturbance recovery time requirements for external control interfaces Verify and obtain the elastic recovery time commitment value , the specific formula is as follows:

[0110] .

[0111] It should be noted that the adjustable capability vector group refers to a set of parameterized vectors derived from the upper-layer control objectives and the current flexible layer state in the response interface model. This vector group quantitatively represents the upper and lower limits and capability boundaries of the lower-layer system's adjustable capability in dimensions such as power, voltage, time, and frequency that can be achieved within the current time period. The boundary flexible layer, a state mapping and capability perception intermediary layer between the upper-layer and lower-layer control systems, represents the current cluster boundary adjustable capability and is a core component of the response interface model. The local operating state is used to dynamically characterize the operating conditions of each controllable device within the lower-layer region. The device adjustable capability refers to the capability boundaries, such as the upper and lower limits of power adjustment, response delay, adjustment duration, and adjustment mode, that can be achieved by each device or device group under the current operating state. The control strategy set refers to a set of feasible control strategy candidates constructed by the lower-layer controller based on the local operating state and adjustable capability, under the constraints of decoupled mapping parameters, for screening or negotiation by the upper layer. The decoupling mapping parameters refer to the mapping constraint parameter set for generating a local adjustment strategy that is transmitted by the upper-level control target vector to the lower-level controller through the flexibility response interface model, and are used to guide the lower layer to design an adjustment scheme within its adjustable capability, and ensure the consistency and executability of the strategy linkage. The adjustment redundancy interval refers to the potential adjustable capability range that the lower-level device has in addition to the current adjustment task, which can be used for strategy diversity expansion, adjustment path selection or emergency backup resource scheduling. The response lag matching interval refers to the acceptable deviation range between the response delay (start-up time) required by the upper-level control strategy and the response capability of each lower-level device, and the adjustment continuity overlap interval refers to the overlapping time range between the adjustment time period required by the upper layer and the maximum continuous adjustment time capability of the lower-level device.

[0112] Each local control strategy set is evaluated and fed back to the upper controller for optimization and execution of consistency control actions, specifically:

[0113] Obtaining key performance indicators corresponding to the local control strategy set and building a regulation strategy evaluation model, wherein the key performance indicators include regulation effect, control dissipation, and load fluctuation smoothness;

[0114] Feedback the output of the adjustment strategy evaluation model to the upper controller;

[0115] The upper-level controller receives several model output results, constructs a three-layer structure of goal-criteria-scheme based on the hierarchical analysis method of the main goal, uses the pairwise comparison method to determine the criterion weights, and combines the output results of each strategy in each index adjustment strategy evaluation model to obtain a comprehensive score. Finally, the control strategy set with the highest score is selected as the optimal solution, thus selecting the optimal control strategy set;

[0116] According to the optimal control strategy set, the target control vector of the upper controller is modified to generate a linkage control strategy;

[0117] Each controller at each level performs its own control actions according to the linkage control strategy to achieve linkage consistency regulation. The control actions include adjusting power command issuance, local load switching, and energy storage start and stop control.

[0118] The specific calculation formula of the adjustment strategy evaluation model is as follows:

[0119]

[0120] Where, is the adjustment strategy evaluation value, To adjust the effect, To control dissipation, is the load fluctuation smoothness, 、 、 are weight coefficients respectively.

[0121] It should be noted that the aforementioned hierarchical decoupling and linkage control scheme achieves refined decoupling and linkage coordination among the power grid's multi-level control systems by constructing a flexible response interface model and a coordinated strategy generation and feedback mechanism between upper and lower layers. Driven by control objectives, the system accurately extracts adjustable capability vectors based on the state mapping matrix of the flexible layer at the cluster boundary, ensuring that the adjustment range, response time, sustainability, and resilience all meet the target constraints. A time window matching algorithm is used to map parameters between objectives and capabilities, effectively avoiding conflicts and resource waste between upper and lower layer control commands. Simultaneously, lower-level controllers autonomously generate local control strategies within the defined strategy boundaries and conduct quantitative evaluations based on multidimensional performance indicators, forming a closed feedback loop that prompts upper-level controllers to optimize the overall linkage strategy using the analytic hierarchy process. Ultimately, each control layer executes consistent regulation operations according to the optimal strategy, improving the coordination, robustness, and real-time performance of the system response while strengthening the distributed dynamic management capabilities of complex loads and energy storage units. This effectively supports the coordinated scheduling and proactive service capabilities of flexible resources in new power systems.

[0122] Example 2, Figure 2 This is a schematic diagram of the process structure of a distribution network dynamic cluster planning control system based on electrical indicators provided in an embodiment of the present application, including a flexibility indicator set generation module, a model construction module, and a hierarchical control strategy generation and execution module. There are connections between the modules:

[0123] The flexibility index set generation module is used to obtain device response data and build a dynamic response model of the device. The model extracts the characteristics of the device under different adjustment signals and generates a device flexibility index set.

[0124] The model construction module is used to construct the cluster boundary flexibility layer state mapping matrix according to the equipment flexibility index set, and couple it with the cluster external control interface to generate the inter-layer flexibility response interface model;

[0125] The hierarchical control strategy generation and execution module is used to construct a hierarchical decoupling control strategy of the upper-level scheduling strategy and the lower-level device control based on the response interface model, and drive the cluster multi-level controller to perform collaborative consistency control actions through the control strategy.

[0126] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0127] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0128] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0129] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0130] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0131] 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 dynamic cluster planning and control method for distribution network based on electrical indicators, characterized in that: include: Obtain equipment response data, build an equipment dynamic response model, extract multi-dimensional equipment characteristic parameters through the response model, and generate an equipment flexibility index set; A mapping matrix representing the state of the cluster boundary flexibility layer is constructed based on the equipment flexibility index set; The mapping matrix representing the state of the cluster boundary flexibility layer is constructed based on the device flexibility index set, specifically: Based on the discrete degree of the flexibility index set, the initial weight of the flexibility index is calculated by the entropy weight method; Obtain the system frequency deviation characteristic quantity, map the characteristic quantity into the level correction factor based on the preset fuzzy rule library, and dynamically correct the initial weight; The equipment flexibility index set is normalized to obtain the initial matrix, which is then fused with the modified weights to construct the flexibility layer mapping vector. Aggregate the flexible layer mapping vectors based on the topological level and physical connection mode of the devices to generate a flexible layer state mapping matrix; Dynamically couple the layered mapping matrix with the cluster's external control interface to generate an inter-layer flexibility response interface model. The dynamic coupling with the cluster external control interface generates an inter-layer flexibility response interface model, specifically: Acquire cluster external control interface characteristics, where the external control interface characteristics include target power adjustment amount, maximum response time limit, continuous adjustment time, and maximum tolerable recovery time; The flexible response adaptation function is constructed through the flexible layer state mapping matrix and the external control interface characteristics; Couple the hierarchy with the response adaptation function to build a flexible response interface model; Based on the response interface model, a hierarchical decoupling control strategy of the upper-level scheduling strategy and the lower-level device control is constructed, and the cluster multi-level controllers are driven by the control strategy to perform collaborative consistency control actions.

2. The method for dynamic cluster planning and control of distribution network based on electrical indicators according to claim 1, characterized in that: The acquisition of device response data and the construction of a device dynamic response model are specifically as follows: Obtain device response data of each flexible load device in the cluster under different control instructions, wherein the device response data includes control input, output response amount and disturbance scenario label; The device response data is subjected to standardized preprocessing, and a device dynamic response model is established for each device by controlling the nonlinear mapping relationship between input and output response quantities of the preprocessed data.

3. The method for dynamic cluster planning and control of distribution network based on electrical indicators according to claim 2, characterized in that: The pre-processed data is used to establish a device dynamic response model for each device by controlling the nonlinear mapping relationship between the input and output response quantities, specifically: According to the parallel convolutional neural network and recurrent neural network structure, the control input and output response quantity feature sequences are extracted through a multi-window sliding sampling method. The multi-window sliding sampling method includes setting a short-time window to extract local features and setting a long-time window to extract trend features. The convolutional neural network processes the short-time window, and the recurrent neural network processes the long-time window; The disturbance scene label is embedded as a conditional vector and concatenated and trained with the feature sequence to build a dynamic response model.

4. The method for dynamic cluster planning and control of distribution network based on electrical indicators according to claim 1, characterized in that: The multi-dimensional equipment characteristic parameters are extracted through the response model to generate the equipment flexibility index set, specifically: Perturb the dynamic response model and extract the dynamic characteristics of each device; Generate multiple response curves based on dynamic characteristics and construct a response curve cluster for each device; Based on the response curve cluster, a flexibility index set of each device is calculated, wherein the flexibility index set includes dynamic adjustment margin, time response capability, continuous adjustment capability and elastic recovery capability.

5. The method for dynamic cluster planning and control of distribution network based on electrical indicators according to claim 1, characterized in that: Based on the response interface model, a hierarchical decoupling control strategy is constructed for the upper-layer scheduling strategy and the lower-layer device control. The control strategy drives the cluster multi-level controller to perform coordinated consistency control actions, specifically: Obtain the control target vector of the upper control system and input it into the response interface model. Calculate the matching relationship between the flexible layer state mapping matrix and the target vector through the flexible response adaptation function to generate an adjustable capability vector group. identifying decoupling mapping parameters according to the adjustable capability vector group, the decoupling mapping parameters including an adjustment redundancy interval, a response time lag matching interval, and an adjustment persistence overlap interval; Send the decoupling mapping parameters to the lower-level controller to generate a local control strategy set within the controllable area of ​​the device; Perform multi-dimensional evaluation on each policy set and feed back the evaluation results to the upper-level controller; Execute cluster consistency control actions based on the evaluation results.

6. The method for dynamic cluster planning and control of distribution network based on electrical indicators according to claim 5, characterized in that: The matching relationship between the flexible layer state mapping matrix and the target vector is calculated by the flexible response adaptation function to generate an adjustable capability vector group, specifically: Generate a dynamic power adjustment range by matching the device's dynamic adjustment margin with the power adjustment target in the upper-layer control target vector. Based on the time response capability of the device and the response time constraint in the target vector, a bound comparison is performed to generate a time response capability commitment value; Based on the continuous adjustment capability of the equipment and the adjustment continuity constraint in the control target vector, a window is matched to generate a sustainable adjustment time commitment value; The elastic recovery capability of the equipment is verified against the disturbance recovery time requirements of the external control interface to obtain the elastic recovery time commitment value.

7. The method for dynamic cluster planning and control of distribution network based on electrical indicators according to claim 6, characterized in that: The hierarchical decoupling control strategy of constructing the upper-layer scheduling strategy and the lower-layer device control is used to drive the cluster multi-level controller to perform coordinated consistency control actions. Specifically, Obtain the local control strategy set of each lower-level controller, extract the key performance indicator vector of each strategy set, and build a regulation strategy evaluation model based on the performance indicator vector; Feedback the output of the adjustment strategy evaluation model to the upper controller; The upper controller selects the optimal control strategy set based on the hierarchical analysis method of the main target; Modify the upper target control vector according to the optimal control strategy set to generate a linkage control strategy; Drive controllers at all levels to execute distributed control actions consistent with the linkage control strategy.

8. A system using the distribution network dynamic cluster planning and control method based on electrical indicators according to any one of claims 1 to 7, characterized in that: It includes a flexibility index set generation module, a model construction module, and a hierarchical control strategy generation and execution module. There are connections between the modules: The flexibility index set generation module is used to obtain device response data and build a dynamic response model of the device. The model extracts the characteristics of the device under different adjustment signals and generates a device flexibility index set. The model construction module is used to construct the cluster boundary flexibility layer state mapping matrix according to the equipment flexibility index set, and couple it with the cluster external control interface to generate the inter-layer flexibility response interface model; The hierarchical control strategy generation and execution module is used to construct a hierarchical decoupling control strategy of the upper-level scheduling strategy and the lower-level device control based on the response interface model, and drive the cluster multi-level controller to perform collaborative consistency control actions through the control strategy.

Images