Multi-scenario-oriented coordinated voltage regulation method for distribution network multi-element reactive power resources

Abstract

The present application relates to a kind of multi-element reactive resource coordinated voltage regulation method for distribution network multi-scene, belong to distribution network operation and control technical field.The method includes: obtaining the real-time operating data of multiple nodes in distribution network, calculate the global static voltage sensitivity matrix and dynamic voltage stability margin index of distribution network;Establish and identify the regulation scene of distribution network operating state, call corresponding coordinated optimization strategy model;Real-time operating data, static voltage sensitivity matrix and dynamic voltage stability margin index are input into the called coordinated optimization strategy model to solve, generate reactive power output instruction set for multi-element reactive resource;The time sequence coordination and equipment action feasibility check of reactive power output instruction set are carried out, and the executable coordinated control instruction sequence that meets time sequence and equipment constraint is generated.Compared with prior art, the present application can accurately distinguish the difference of different distribution network operating scene, realize the pertinence coordination of multi-element reactive resource.

Description

Technical Field

[0001] This invention relates to the field of power distribution network operation and control technology, and in particular to a method for coordinated voltage regulation of multiple reactive resources for various scenarios in power distribution networks. Background Technology

[0002] As the final link in the power system, the stable operation of the distribution network is directly related to the quality and security of power supply. Currently, distribution network voltage regulation primarily employs a single reactive power resource regulation mode or a fixed regulation strategy. This involves acquiring basic operational data such as voltage and power at some nodes in the distribution network, judging the network's operating status based on a single voltage amplitude index, and then calling a fixed optimization model to generate reactive power output commands to regulate the distribution network voltage.

[0003] Chinese patent application CN119362484A provides a method for coordinated control of static centralized and dynamic distributed reactive power / voltage in a distribution network. It utilizes static centralized control strategies and dynamic distributed control strategies to process event warning signal data for inter-regional coordinated control, obtaining coordinated control command data. However, it classifies scenarios (e.g., static centralized, dynamic distributed) based on multi-scale voltage stability domain data and event warning signals, primarily relying on historical fault data and multi-scale constraint analysis. It focuses on voltage stability domain division and event triggering, and its selection of dynamic distributed control strategies is based on a pre-set power grid dynamic control rule base to obtain event-resource-control rule data. Essentially, it still involves calling a fixed optimization model and does not achieve real-time dynamic control. Furthermore, it lacks detailed design for command timing conflicts and does not perform equipment feasibility verification, potentially leading to command conflicts during actual execution. Existing technologies mostly rely on single voltage amplitudes or fixed rule bases to identify scenarios.

[0004] It is evident that existing technologies, relying on a single voltage index, cannot comprehensively reflect the static voltage correlation characteristics and dynamic voltage stability level of the entire distribution network. This results in insufficient accuracy in identifying distribution network operation scenarios and difficulty in distinguishing control requirements under different operating conditions. The fixed optimization strategies invoked lack specificity and cannot adapt to the complex needs of multiple distribution network operation scenarios. Furthermore, after generating reactive power output commands, existing technologies do not systematically verify the commands, which can easily lead to problems such as timing conflicts between different reactive power resource actions and commands exceeding equipment operating limits. This results in commands being unable to be executed directly, affecting control effectiveness and equipment safety. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of the existing technology by proposing a multi-reactive-resource coordinated voltage regulation method for multiple scenarios in distribution networks.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A method for coordinated voltage regulation of multiple reactive resources in distribution networks across various scenarios includes the following steps: Acquire real-time operating data from multiple nodes in the power distribution network; Based on the acquired real-time operating data, the static voltage sensitivity matrix and dynamic voltage stability margin index of the entire distribution network are calculated. Based on the static voltage sensitivity matrix, the control scenarios of the current distribution network operation are identified, and based on the identified control scenarios, a collaborative optimization strategy model matching the control scenarios is invoked. The real-time operating data, static voltage sensitivity matrix, and dynamic voltage stability margin index are input into the called collaborative optimization strategy model for solution, generating a reactive power output instruction set for multiple reactive resources; The reactive power output instruction set is subjected to timing coordination and equipment action feasibility verification to generate an executable collaborative control instruction sequence that meets timing and equipment constraints.

[0007] As a further aspect of the present invention, the specific steps for calculating the static voltage sensitivity matrix and the dynamic voltage stability margin index include: Based on the real-time operating data, an equivalent admittance network model of the current distribution network is established; Based on the equivalent admittance network model, the Jacobian matrix is ​​obtained by using the Newton-Raphson method for power flow calculation. The Jacobian matrix is ​​inverted at a set operating point to extract the partial derivatives of voltage with respect to active and reactive power injections, thus obtaining the static voltage sensitivity matrix. Based on the equivalent admittance network model and the dynamic response model of multiple reactive resources, a small-disturbance stability analysis model for the distribution network is constructed. Solve for the eigenvalues ​​of the small disturbance stability analysis model, and obtain the eigenvectors corresponding to the minimum damping ratio and the key oscillation mode based on these eigenvalues; The dynamic voltage stability margin index is obtained based on the eigenvectors corresponding to the minimum damping ratio and the key oscillation modes.

[0008] As a further aspect of the present invention, identifying the control scenario to which the current distribution network operation status belongs is achieved through the following steps: Set thresholds for classifying control scenarios, including static voltage over-limit threshold, static voltage sensitivity threshold, and minimum damping ratio threshold; The deviation between the node voltage amplitude and the static voltage over-limit threshold is calculated to obtain the node voltage safety margin; The elements in the static voltage sensitivity matrix are compared with the static voltage sensitivity threshold to evaluate the sensitivity of node voltage to power changes. The calculated minimum damping ratio is compared with the minimum damping ratio threshold to evaluate the small disturbance stability level of the system. By combining the node voltage safety margin, the node voltage sensitivity to power changes, and the stability level under small disturbances, and based on the preset multi-level decision tree logic, the control scenario for the current operating state is determined.

[0009] As a further aspect of the present invention, the collaborative optimization strategy model includes a preventive optimization strategy model corresponding to the steady-state prevention scenario, a corrective optimization strategy model corresponding to the post-disturbance recovery scenario, and an emergency control strategy model corresponding to the emergency support scenario. The preventive optimization strategy model takes minimizing the total active power loss and the sum of squared voltage deviations as multiple objectives, and is constrained by the multi-dimensional reactive power resource regulation capability and the node voltage safety boundary. The corrective optimization strategy model prioritizes restoring all node voltages to a safe range as quickly as possible, and uses the coordination rule of adjusting the order of reactive resources with different speeds for collaboration. The emergency control strategy model takes suppressing voltage collapse as its core objective and the dynamic voltage support strength of key nodes as its optimization objective, while ignoring some economic constraints. Among them, the ability to regulate multiple reactive resources refers to the comprehensive ability to control the voltage of each node in the distribution network by adjusting the reactive power output of multiple reactive resources in the distribution network.

[0010] As a further aspect of the present invention, generating a reactive power output instruction set for multiple reactive power resources is achieved through the following steps: Based on the invoked collaborative optimization strategy model, a mathematical optimization problem is constructed that includes the optimization objective, constraints, and coordination rules. The static voltage sensitivity matrix is ​​used as a simplified expression of the correlation between variables and the objective function in the optimization problem; The dynamic voltage stability margin index is used as a quantification parameter for stability constraints in the optimization problem; The real-time output and status information of the various reactive resources are used as the initial values ​​of the optimization variables and the boundary of the feasible region. The mathematical optimization problem is solved online using interior-point method, mixed-integer programming, or heuristic algorithm to obtain the reactive power output setpoints of various multi-dimensional reactive resources at each time point within a future control cycle, thus forming the reactive power output instruction set.

[0011] As a further aspect of the present invention, the timing coordination and equipment operation feasibility verification of the reactive power output instruction set is performed through the following steps: Check whether the change pace of different types of reactive power resource output instructions on the time axis in the reactive power output instruction set meets the preset coordination timing requirements. The coordination timing requirements specify the sequence of actions and coordination intervals between resources with fast adjustment capabilities and resources with slow adjustment capabilities. Verify whether each output command value in the reactive power output command set exceeds the physical adjustable range and action rate limit of the corresponding reactive power compensation device at the next moment. For instructions that do not meet the coordination timing requirements or exceed the feasibility limits of device actions, smooth correction and re-coordination shall be performed on the premise of meeting the core objectives of the coordination optimization strategy model. Generate an executable sequence of cooperative control instructions that fully satisfies timing and device constraints.

[0012] As a further aspect of the present invention, the multiple reactive power resources include a static var generator, a static synchronous compensator, a distributed photovoltaic inverter, an energy storage converter, and a grouped switching capacitor. The reactive power output instruction set includes continuous reactive current reference value instructions for static var generators, reactive power setpoint instructions for static synchronous compensators, reactive power factor or reactive power setpoint instructions for distributed photovoltaic inverters, reactive power setpoint instructions for energy storage converters, and switching status instructions for grouped switching capacitors.

[0013] As a further aspect of the present invention, when acquiring the real-time operating data, the load power data of key feeders in the distribution network, the actual and predicted output power data of distributed power sources, and the voltage support capability information of the upper-level grid interconnection points are acquired simultaneously as background data for confirming the network topology and boundary conditions of the mathematical optimization problem.

[0014] As a further aspect of the present invention, it also includes an online parameter update step for the collaborative optimization strategy model: After the executable coordinated control command sequence is executed, real-time operating data of the distribution network is acquired again to calculate new node voltages and system states; The control deviation is calculated by comparing the new node voltage and system state with the target state expected by the optimization model. Based on the magnitude and direction of the control deviation, the key parameters in the currently used collaborative optimization strategy model, including the objective function weights and constraint boundary relaxation factors, are adaptively fine-tuned. The fine-tuned parameters are updated in the collaborative optimization strategy model for use in the next round of optimization calculations.

[0015] As a further aspect of the present invention, after generating the executable coordinated control command sequence, the executable coordinated control command sequence is sent to the multi-reactive power resource control terminal deployed in the distribution network, and the multi-reactive power resource control terminal drives the corresponding reactive power compensation equipment to perform reactive power output adjustment, specifically including: The executable collaborative control instruction sequence is decomposed according to resource type and geographical location, and encapsulated into standardized instruction messages that can be recognized by different control terminals; Add a timestamp, instruction sequence number, and checksum to each standardized instruction message, and distribute it through the power distribution network communication network; After receiving the standardized instruction message, each multi-reactive power resource control terminal parses and executes the control commands in it, driving the static var generator, static synchronous compensator, distributed photovoltaic inverter, energy storage converter and group switching capacitor to perform corresponding reactive power output adjustment, and replies with an execution confirmation signal to the main station system.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention identifies control scenarios based on a dual-indicator approach: a global static voltage sensitivity matrix and a dynamic voltage stability margin index. The static voltage sensitivity matrix comprehensively captures the static correlation between voltage and reactive power output at each node of the distribution network, while the dynamic voltage stability margin index accurately reflects the dynamic voltage safety level of the distribution network. The synergistic effect of these two indicators enables a more accurate and comprehensive capture of the differences between the static and dynamic characteristics of the entire distribution network, solving the problems of limited scenario identification and difficulty in distinguishing complex operating conditions caused by single-indicator approaches in existing technologies.

[0017] 2. This invention establishes a collaborative optimization strategy model based on the identified scenarios and different reactive power resource optimization objectives, constraints, and coordination rules; it then invokes the collaborative optimization strategy model matched to the scenario, enabling the optimization direction, operational constraints, and coordination logic of different types of reactive power resources to adapt to the current distribution network operation status. This avoids the insufficient adaptability of fixed strategies in multiple scenarios, allowing various reactive power resources to play an optimal regulatory role according to scenario requirements, and achieving targeted collaboration of diverse reactive power resources.

[0018] 3. This invention performs timing coordination verification on the set of reactive power output commands for multiple reactive resources obtained from the model solution. Timing coordination verification can identify action conflicts in the time dimension of different reactive power output commands, avoid the cancellation of control effects caused by the disorder of action timing of various reactive resources, and ensure that different reactive resources are synchronized and coordinated in the control process.

[0019] 4. This invention performs equipment action feasibility verification on the set of reactive power output instructions obtained from the model solution for multiple reactive resources. It can screen out instructions that meet the equipment operating limits and action capability requirements, avoid equipment damage or control failure caused by instructions exceeding the equipment operating range, and enable the generated control instructions to be directly implemented, ensuring the safety and stability of the distribution network voltage regulation process and realizing efficient collaborative regulation of multiple reactive resources.

[0020] 5. This invention also includes online parameter updates for the collaborative optimization strategy model, which can adaptively fine-tune and iteratively optimize the control deviation, thereby achieving a continuous decrease in the control deviation until it stabilizes. It has good convergence and robustness, reflecting the adaptability and intelligence of the control method, and achieving the optimal balance between voltage safety and economic operation.

[0021] 6. Compared with existing technologies (such as CN119362484A), this invention not only coordinates control based on static / dynamic partitioning, but also accurately identifies three types of scenarios—steady-state prevention, post-disturbance recovery, and emergency support—through the static voltage sensitivity matrix and dynamic voltage stability margin index, and constructs optimization models with specific multi-objectives, constraints, and cooperative rules for each scenario. In particular, this invention adds a timing coordination and equipment action feasibility verification step after generating instructions, and performs adaptive online updates of the optimization model parameters based on control deviations, thereby solving the problems of easy command conflict, difficult execution, and insufficient adaptability in existing technologies. Attached Figure Description

[0022] Figure 1 This is a flowchart of the multi-reactive-resource coordinated voltage regulation method for multiple scenarios in distribution networks according to the present invention. Figure 2 A flowchart for establishing and identifying control scenarios of distribution network operation status using the method of the present invention; Figure 3 The original instruction timing conflict detection diagram for timing coordination verification of reactive power output instruction set of multiple reactive resources; Figure 4 An adaptive update diagram of parameters for the reactive power control model of the distribution network, reflecting the effect of online updating of parameters for the collaborative optimization strategy model; Figure 5 A visualization of the decomposition stages of the executable coordinated control command sequence for the distribution network. Detailed Implementation

[0023] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0024] Example 1 This embodiment provides a method for coordinated voltage regulation of multiple reactive resources in various scenarios of distribution networks. (See reference...) Figure 1 First, real-time operating data of multiple nodes in the distribution network is obtained. The real-time operating data includes node voltage amplitude and phase, line active power and reactive power, and real-time output and status information of various reactive resources. Based on the acquired real-time operational data, a state assessment is performed, and the static voltage sensitivity matrix and dynamic voltage stability margin of the distribution network are calculated. Specifically, this includes: Using the obtained node voltage amplitude and phase, line active power and reactive power data, an equivalent admittance network model of the current distribution network is established. Based on the established equivalent admittance network model, the Newton-Raphson method is used to calculate the power flow, solve the system power balance equation, obtain the convergent solution of node voltage and phase angle, and obtain the Jacobian matrix at the operating point. By inverting the Jacobian matrix at a set operating point, the partial derivatives of voltage with respect to active and reactive power injection at the nodes are extracted. These partial derivatives constitute the static voltage sensitivity matrix, which reflects the linear response of node voltage changes to power injection.

[0025] Based on the equivalent admittance network model and the dynamic response model of multiple reactive resources, a small-disturbance stability analysis model for distribution networks is constructed to analyze small-range disturbances in the system. Preferably, the dynamic response model of multiple reactive resources includes differential algebraic equations for static var generators, static synchronous compensators, distributed photovoltaic inverters, energy storage converters, and grouped switching capacitors. Solve for the eigenvalues ​​of the small disturbance stability analysis model, and obtain the eigenvectors corresponding to the minimum damping ratio and the key oscillation mode based on these eigenvalues; obtain the dynamic voltage stability margin index based on the eigenvectors corresponding to the minimum damping ratio and the key oscillation mode. Preferably, the small-disturbance stability analysis model obtains the state matrix by linearizing the system equations, and then solves for the eigenvalues ​​of the state matrix. The eigenvalues ​​satisfy the formula: det(A-λI)=0, where A represents the linearized system state matrix, λ represents the eigenvalues, and I represents the identity matrix. All eigenvalues ​​are calculated, and the minimum damping ratio and the eigenvector corresponding to the key oscillation mode are used as components of the dynamic voltage stability margin index. The minimum damping ratio is calculated from the real and imaginary parts of the eigenvalues, and the key oscillation mode corresponds to the dominant eigenvalue and its eigenvector.

[0026] Specifically, when identifying the control scenario to which the current distribution network operation status belongs based on the static voltage sensitivity matrix and dynamic voltage stability margin index, the control scenario is divided into three categories: steady-state prevention scenario, post-disturbance recovery scenario, and emergency support scenario. To set thresholds for different control scenarios, please refer to [reference needed]. Figure 2 The thresholds include the static voltage over-limit threshold, the static voltage sensitivity threshold, and the minimum damping ratio threshold. The static voltage over-limit threshold is defined as the upper and lower limits of the safe operation of the node voltage; the static voltage sensitivity threshold is defined as the limit value of the sensitivity of the node voltage to power changes; and the minimum damping ratio threshold is defined as the minimum required value for the stability of the system under small disturbances. The elements in the calculated static voltage sensitivity matrix are compared with the static voltage sensitivity threshold to determine the sensitivity of the node voltage to power changes. For example, if an element in the static voltage sensitivity matrix exceeds the static voltage sensitivity threshold, the corresponding node voltage is considered to be highly sensitive. The calculated minimum damping ratio is compared with the minimum damping ratio threshold to evaluate the small disturbance stability level of the system. A minimum damping ratio lower than the minimum damping ratio threshold indicates that the system is not damped enough. Based on the deviation between the current node voltage amplitude and the static voltage over-limit threshold, as well as the evaluation results of sensitivity and stability level, the current operating state is determined to belong to one of the following scenarios according to the preset multi-level decision tree logic: steady-state prevention scenario, disturbance recovery scenario, or emergency support scenario. The multi-level decision tree logic achieves scenario classification through hierarchical judgment conditions. Understandably, in the example scenarios, the data comparison is reflected in the threshold comparison process. For instance, in one distribution network operation state, the static voltage sensitivity matrix shows that the voltage sensitivity of multiple nodes to reactive power injection is higher than the static voltage sensitivity threshold, while the calculated minimum damping ratio is close to but higher than the minimum damping ratio threshold, and the node voltage amplitudes are all within the static voltage over-limit threshold range. Through multi-level decision tree logic, this scenario might be classified as a steady-state prevention scenario. In another example scenario, the static voltage sensitivity matrix shows that key nodes have high sensitivity, the calculated minimum damping ratio is lower than the minimum damping ratio threshold, and the voltage amplitudes of some nodes exceed the static voltage over-limit threshold. Through multi-level decision tree logic, this scenario might be classified as an emergency support scenario. These comparisons are based on a direct comparison of calculated data and preset thresholds, without relying on empirical assumptions. Optionally, the implementation of multi-level decision tree logic can be based on a rule engine; The first level determines whether the node voltage exceeds the static voltage limit threshold. If it does, it enters the emergency or recovery scenario determination sub-logic. The second level of judgment determines whether the minimum damping ratio is lower than the minimum damping ratio threshold. If it is lower, it tends to be an emergency support scenario. The third level combines the comparison results of the static voltage sensitivity matrix elements and the static voltage sensitivity threshold to comprehensively determine the scene classification.

[0027] All the above judgments are based on real-time calculated data to ensure that the scene recognition matches the current operating status.

[0028] Preferably, both the calculation process and scene recognition are executed online, utilizing real-time data acquisition to update the equivalent admittance network model and sensitivity matrix to achieve dynamic tracking. The static voltage sensitivity matrix and dynamic voltage stability margin index are calculated using numerical methods to ensure efficient solutions even when the distribution network scale changes. The output results of the scene recognition module are directly used for subsequent strategy model calls, forming the basis for closed-loop control.

[0029] In practical implementation, when invoking the collaborative optimization strategy model matching the identified control scenario, a pre-defined collaborative optimization strategy model matching the three control scenarios is used, including: A preventive optimization strategy model corresponding to the steady-state prevention scenario. This preventive optimization strategy model has multiple objectives, namely minimizing the active power loss of the entire network and minimizing the sum of squares of voltage deviation, and uses the multi-dimensional reactive power resource regulation capability and node voltage safety boundary as constraints. A corrective optimization strategy model is preset to correspond to the recovery scenario after the disturbance. The primary goal of this corrective optimization strategy model is to restore the voltage of all nodes to the safe range as quickly as possible, and the coordination rule is the order of coordination of reactive resources with different adjustment speeds. An emergency control strategy model corresponding to the emergency support scenario is preset. The core objective of this emergency control strategy model is to suppress voltage collapse, and the optimization objective is the dynamic voltage support strength of key nodes. In this scenario, some economic constraints can be ignored. Based on the identified specific scenario, select one of the preset preventive optimization strategy models, corrective optimization strategy models, and emergency control strategy models as the current collaborative optimization strategy model.

[0030] Then, based on the identified control scenario, the matching collaborative optimization strategy model is invoked. The real-time operating data, static voltage sensitivity matrix and dynamic voltage stability margin index are substituted into the invoked collaborative optimization strategy model for solution, generating a set of reactive power output instructions for multiple reactive resources. Based on the definition of the selected collaborative optimization strategy model, construct a complete mathematical optimization problem that includes the optimization objective, constraints, and coordination rules; The static voltage sensitivity matrix can be used as a simplified expression of the relationship between variables and the objective function in optimization problems. For example, in the preventive optimization strategy model, the static voltage sensitivity matrix can be used to quickly estimate the approximate impact of multi-dimensional reactive power resource output adjustment on node voltage and network loss. The dynamic voltage stability margin index is used as a quantitative parameter of stability constraints in optimization problems. For example, in the emergency control strategy model, the minimum damping ratio is used as part of the inequality constraint, requiring the minimum damping ratio of the optimized system to be higher than a set threshold. The real-time output and status information of multiple reactive resources are used as the initial values ​​of optimization variables and the boundary of the feasible region. The real-time reactive current value of the static var generator is used as the initial value of its optimization variables. The current number of switching groups of the grouped capacitors determines the starting point of its discrete action.

[0031] It is understandable that the solution process employs interior-point methods, mixed-integer programming, or heuristic algorithms to solve the constructed mathematical optimization problem online. When the collaborative optimization strategy model is a preventive optimization strategy model and contains only continuous variables, the interior point method is used for solving it; When the collaborative optimization strategy model includes discrete switching action variables for grouped capacitor switching, a mixed integer programming algorithm is used to solve it; In emergency support scenarios where time constraints are extremely stringent, specific heuristic algorithms can be used to quickly solve emergency control strategy models. The solution yields the reactive power output setpoints of various multi-source reactive power resources at discrete moments within a future control cycle. These setpoints are organized according to time sequence and equipment type to form a complete reactive power output instruction set.

[0032] Preferably, the mathematical optimization problem in constructing the preventive optimization strategy model involves multi-objective optimization, with the objective function in the form of: in: This indicates that the system has a total active power loss. This represents the sum of squares of voltage deviations at all nodes. and These are weighting coefficients; The constraints of the preventive optimization strategy model include the upper and lower limits of active and reactive power output of multiple reactive resources, the ramp rate limit of multiple reactive resources, and the safe operating boundary of node voltage amplitude. The mathematical optimization problem constructed by the corrective optimization strategy model takes minimizing the voltage recovery time of all over-limit nodes as its primary objective. Its coordination rule is reflected in setting time-sharing action sequences for fast resources such as static var generators and static synchronous compensators and slow resources such as group switching capacitors in the constraints. The mathematical optimization problem for constructing the emergency control strategy model aims to minimize the transient voltage drop at the critical weak node. The critical weak node is determined with the help of eigenvector information in the dynamic voltage stability margin index. The constraints of the emergency control strategy model focus on the instantaneous reactive power output limit and response rate of multiple reactive resources.

[0033] Optionally, in the example scenario, the data comparison is reflected in the selection and construction of different collaborative optimization strategy models. If the scenario identification module determines that the current scenario is a steady-state prevention scenario, it calls the preventive optimization strategy model. The static voltage sensitivity matrix in its input data will be used to efficiently construct an approximate relationship between network loss and voltage deviation and control variables. The resulting reactive power output instruction set focuses on the balance between economy and safety. If the scenario identification module determines that the current scenario is an emergency support scenario, it calls the emergency control strategy model. The dynamic voltage stability margin index in its input data is directly converted into stability constraints. The resulting reactive power output instruction set prioritizes ensuring dynamic voltage support strength, without considering economic indicators. This difference in data flow and model construction based on the scenario reflects the core of multi-strategy scheduling.

[0034] In practical implementation, the solution module of the collaborative optimization strategy model receives a trigger signal and model type identifier from the scene recognition module. Based on the model type identifier, the solution module loads the complete mathematical description of the corresponding preventative optimization strategy model, corrective optimization strategy model, or emergency control strategy model from the strategy library. The solution module reads the static voltage sensitivity matrix, dynamic voltage stability margin index, and real-time output and status information of multiple reactive power resources from the real-time database, and fills them into the corresponding input parameter positions of the loaded model. The solver performs numerical calculations on the parameterized model, and after iterative convergence, outputs a sequence of setpoints for multiple reactive power resources over several future time segments, i.e., a reactive power output instruction set, completing the computational closed loop from strategy model invocation to instruction generation.

[0035] Finally, the timing coordination and equipment action feasibility of the reactive power output instruction set are verified. The change pace of different types of reactive power resource output instructions on the time axis in the reactive power output instruction set is checked to see if they meet the preset coordination timing requirements. The coordination timing requirements clearly define the order of actions and coordination time intervals between resources with fast adjustment capabilities and resources with slow adjustment capabilities. Verify each output command value in the reactive power output command set to determine whether it exceeds the physical adjustable range and action rate limit of the corresponding reactive power compensation equipment at the next moment. For instructions found during the inspection that do not meet the coordination timing requirements or exceed the feasibility limits of equipment actions, they should be smoothly corrected and re-coordinated on the premise of meeting the core objectives of the coordination optimization strategy model. After the above verification and correction process, a final executable collaborative control instruction sequence that fully meets the timing requirements and physical constraints of the equipment is generated and output.

[0036] Example 2 This embodiment provides a method for coordinated voltage regulation of multiple reactive resources in various scenarios of distribution networks. (See reference...) Figure 1First, real-time operating data of multiple nodes in the distribution network is obtained. The real-time operating data includes node voltage amplitude and phase, line active power and reactive power, and real-time output and status information of various reactive resources. Based on the acquired real-time operating data, the static voltage sensitivity matrix and dynamic voltage stability margin index of the entire distribution network are calculated. Based on the static voltage sensitivity matrix and dynamic voltage stability margin index, the control scenarios of the current distribution network operation are identified, and based on the identified control scenarios, a collaborative optimization strategy model matching the control scenarios is invoked. The real-time operating data, static voltage sensitivity matrix, and dynamic voltage stability margin index are input into the called collaborative optimization strategy model for solution, generating a set of reactive power output instructions for multiple reactive resources. The reactive power output instruction set is subjected to timing coordination and equipment action feasibility verification to generate an executable collaborative control instruction sequence that meets timing and equipment constraints.

[0037] Furthermore, the specific steps for calculating the static voltage sensitivity matrix and the dynamic voltage stability margin index include: Based on the node voltage amplitude and phase, and the active and reactive power data of the lines, an equivalent admittance network model of the current distribution network is established. Based on the equivalent admittance network model, the Jacobian matrix is ​​obtained by using the Newton-Raphson method for power flow calculation. The Jacobian matrix is ​​inverted at a set operating point to extract the partial derivatives of voltage with respect to active and reactive power injections, thus obtaining the static voltage sensitivity matrix. Based on the equivalent admittance network model and the dynamic response model of multiple reactive resources, a small-disturbance stability analysis model for the distribution network is constructed. Solve for the eigenvalues ​​of the small disturbance stability analysis model, and obtain the eigenvectors corresponding to the minimum damping ratio and the key oscillation mode based on these eigenvalues; The dynamic voltage stability margin index is obtained based on the eigenvectors corresponding to the minimum damping ratio and the key oscillation modes.

[0038] Furthermore, identifying the control scenario to which the current distribution network operation status belongs is achieved through the following steps: Set thresholds for classifying control scenarios, including static voltage over-limit threshold, static voltage sensitivity threshold, and minimum damping ratio threshold; The deviation between the node voltage amplitude and the static voltage over-limit threshold is calculated to obtain the node voltage safety margin; The elements in the static voltage sensitivity matrix are compared with the static voltage sensitivity threshold to evaluate the sensitivity of node voltage to power changes. The calculated minimum damping ratio is compared with the minimum damping ratio threshold to evaluate the small disturbance stability level of the system. By combining the node voltage safety margin, the node voltage sensitivity to power changes, and the stability level under small disturbances, and based on the preset multi-level decision tree logic, the control scenario for the current operating state is determined.

[0039] Furthermore, the collaborative optimization strategy model includes: The preventive optimization strategy model corresponding to the steady-state prevention scenario takes the minimization of the total network active power loss and the minimization of the sum of squares of voltage deviation as multiple objectives, and is constrained by the multi-dimensional reactive power resource regulation capability and the node voltage safety boundary. The corrective optimization strategy model corresponding to the post-disturbance recovery scenario takes the fastest possible recovery of all node voltages to the safe range as its primary objective and uses the coordination rule of the order of reactive resources with different adjustment speeds to coordinate. The emergency control strategy model corresponding to the emergency support scenario takes suppressing voltage collapse as its core objective and the dynamic voltage support strength of key nodes as its optimization objective, while ignoring some economic constraints.

[0040] Furthermore, generating a set of reactive power output instructions for multiple reactive power resources is achieved through the following steps: Based on the invoked collaborative optimization strategy model, a mathematical optimization problem is constructed that includes the optimization objective, constraints, and coordination rules. The static voltage sensitivity matrix is ​​used as a simplified expression of the correlation between variables and the objective function in the optimization problem; The dynamic voltage stability margin index is used as a quantification parameter for stability constraints in the optimization problem; The real-time output and status information of the various reactive resources are used as the initial values ​​of the optimization variables and the boundary of the feasible region. The mathematical optimization problem is solved online using interior-point method, mixed-integer programming, or heuristic algorithm to obtain the reactive power output setpoints of various multi-dimensional reactive resources at each time point within a future control cycle, thus forming the reactive power output instruction set.

[0041] Furthermore, the timing coordination and equipment operation feasibility verification of the reactive power output instruction set are performed through the following steps: Check whether the change pace of different types of reactive power resource output instructions on the time axis in the reactive power output instruction set meets the preset coordination timing requirements. The coordination timing requirements specify the sequence of actions and coordination intervals between resources with fast adjustment capabilities and resources with slow adjustment capabilities. Verify whether each output command value in the reactive power output command set exceeds the physical adjustable range and action rate limit of the corresponding reactive power compensation device at the next moment. For instructions that do not meet the coordination timing requirements or exceed the feasibility limits of device actions, smooth correction and re-coordination shall be performed on the premise of meeting the core objectives of the coordination optimization strategy model. Generate an executable sequence of cooperative control instructions that fully satisfies timing and device constraints.

[0042] In practice, timing coordination verification is performed based on preset coordination timing requirements, which specify the action sequence of resources with different response speeds. An exemplary coordination rule requires that resources with millisecond-level fast response, such as static var generators and static synchronous compensators, act first, followed by resources with second-level response, such as distributed photovoltaic inverters and energy storage converters, and finally resources with minute-level slow response, such as grouped switching capacitors. The timing verification module checks whether the timestamp sequence of various instructions in the reactive power output instruction set conforms to this order, and checks whether the time interval between fast resource and slow resource instructions is greater than the preset minimum coordination interval. The equipment operation feasibility verification check examines whether each command value is within the equipment's capability range, such as the reactive current reference value command for the static var generator. Must meet: in: and These are the lower and upper limits of the reactive current output of the static var generator at the current operating point, respectively. Simultaneously verify the rate of change of the command value. For example, the difference in the number of switching groups between adjacent control cycles of the grouped capacitor switching must be less than or equal to the maximum allowable number of switching groups in a single operation. See Table 1.

[0043] Table 1: Timing Coordination Verification Rules Table Understandably, in the example scenario, the data comparison is reflected in the difference in the instruction sequence before and after verification. Assume the original reactive power output instruction set requires the distributed photovoltaic inverter to perform reactive power regulation at time T, while the group-switching capacitors operate at time T+0.05. According to the coordination timing requirements in Table 1, the operation of the group-switching capacitors (Level 3) must be at least 2 seconds later than the operation of the distributed photovoltaic inverter (Level 2). This time interval of 0.05 seconds is less than the minimum operation interval of 2 seconds, which does not meet the coordination timing requirements. The verification and correction module delays the operation instruction of the group-switching capacitors to after T+2 seconds, thereby generating an executable coordination control instruction sequence that meets the timing requirements. Another data comparison involves equipment feasibility. For example, the original command requires the static var generator to output reactive current changes exceeding its maximum ramp rate limit within 1 second. After verification, this change is decomposed into multiple step commands to ensure that the change rate of each step is within the limit.

[0044] Optionally, the smoothing correction and re-coordination process is carried out within the framework of the optimization problem; For instruction sequences that do not meet timing or device constraints, they are used as initial solutions. The goal of correction is to minimize the deviation from the core objective of the original optimization problem. With strict timing requirements and device physical limitations as constraints, a local quadratic programming or linear programming problem is constructed for fast solution to obtain the corrected feasible instruction sequence. The final executable collaborative control instruction sequence is a data structure containing timestamps, device identifiers, and control instruction values.

[0045] In practical implementation, the verification and correction module receives the reactive power output instruction set from the optimization solution module and accesses the equipment model library containing various static and dynamic parameters of reactive power resources. The specific steps are as follows: Analyze the reactive power output instruction set and classify it according to resource type; Based on the parameters in the equipment model library, the numerical feasibility of each instruction is verified item by item; The timing logic of the instructions is verified according to the preset collaborative timing rule table (as shown in Table 1). For any conflicts found, the module calls the internal small-scale optimization and correction device to handle them. Output a final executable collaborative control instruction sequence that passes all checks, has strictly incrementing timestamps, and corresponds one-to-one with the device status. This sequence can be directly sent to the control system for execution.

[0046] See Figure 3 This is a timing conflict detection diagram of the original instruction set, visually illustrating the timing violations. In the original instructions, the interval between the action time of the grouped capacitors (minute-level) and the action time of the photovoltaic inverter (second-level) is only 0.05 seconds, far less than the minimum compliance interval of 2 seconds. The premature action of minute-level resources violates the coordination rule of second-level resources acting first and minute-level resources following, potentially causing chaotic system regulation rhythm and affecting voltage control stability. This diagram is core evidence in the timing coordination verification phase, clarifying the compliance issues of the original instruction set and serving as the basis for subsequent instruction correction. The action instructions for the grouped capacitors need to be delayed until the photovoltaic inverter completes its action and meets the 2-second interval requirement; that is, the corrected capacitor action time must be ≥2 seconds to meet the coordinated control requirements.

[0047] Example 3 This embodiment provides a method for coordinated voltage regulation of multiple reactive resources in various scenarios of distribution networks. (See reference...) Figure 1 First, real-time operating data of multiple nodes in the distribution network is obtained. The real-time operating data includes node voltage amplitude and phase, line active power and reactive power, and real-time output and status information of various reactive resources. Based on the acquired real-time operating data, the static voltage sensitivity matrix and dynamic voltage stability margin index of the entire distribution network are calculated. Based on the static voltage sensitivity matrix and dynamic voltage stability margin index, the control scenarios of the current distribution network operation are identified, and based on the identified control scenarios, a collaborative optimization strategy model matching the control scenarios is invoked. The real-time operating data, static voltage sensitivity matrix, and dynamic voltage stability margin index are input into the called collaborative optimization strategy model for solution, generating a set of reactive power output instructions for multiple reactive resources. The reactive power output instruction set is subjected to timing coordination and equipment action feasibility verification to generate an executable collaborative control instruction sequence that meets timing and equipment constraints.

[0048] Preferably, when acquiring real-time operating data of multiple nodes in the distribution network, load power data of key feeders in the distribution network, actual and predicted output data of distributed power sources, and voltage support capability information of upstream grid interconnection points are acquired simultaneously.

[0049] In practice, the load power data of the critical feeder includes the active and reactive power measurements at the beginning of the critical feeder and at important sectional switches along the line. The actual and predicted output data of distributed power sources include the actual output power of photovoltaic inverters, wind turbine converters, etc. at the current moment and their predicted power curves in a future scheduling cycle. The voltage support capacity information of the upstream power grid interconnection point includes the bus voltage amplitude, short-circuit capacity, and available reactive power support capacity range of the upstream substation. This data is acquired synchronously through a data acquisition and monitoring system and corresponding measurement devices.

[0050] Optionally, load power data of critical feeders, actual and predicted output data of distributed generation sources, and voltage support capability information of upstream grid interconnection points will be used as background data for network topology and boundary condition confirmation when constructing mathematical optimization problems in the collaborative optimization strategy model. Network topology confirmation refers to using load power data of critical feeders and actual output data of distributed generation sources, combined with the network model, to perform state estimation in order to verify or correct the connection state of the network topology used in the current calculation; boundary condition confirmation refers to treating the upstream grid interconnection point as an equivalent virtual node with a specific voltage amplitude range and reactive power injection range in the optimization model, with its parameters set according to the voltage support capability information of the upstream grid interconnection point. Meanwhile, the predicted output data of distributed power sources are used as known, time-varying node injection power boundary conditions within the optimization cycle to participate in the construction of the optimization problem.

[0051] Preferably, the method in this embodiment further includes an online parameter update step for the collaborative optimization strategy model: After the executable coordinated control command sequence is issued and executed, the real-time operation data of the distribution network is acquired again to calculate the new node voltage and system operation status. The control deviation is calculated by comparing the new node voltage and system state with the target state expected to be achieved when solving the optimization model. Based on the magnitude and direction of the control deviation, the key parameters in the currently used collaborative optimization strategy model are adaptively fine-tuned. These key parameters include the weights of each item in the objective function and the boundary relaxation factors of the constraints. The fine-tuned parameters are updated in the collaborative optimization strategy model for use in the next round of optimization calculations.

[0052] The new node voltages and system status include the voltage amplitude, phase angle, and power of critical lines at each measurement node after the command is executed.

[0053] Control deviations include voltage deviation, reactive power output deviation, and stability index deviation, which can be calculated as follows: in: Indicates the overall control deviation. and Let these represent the actual voltage and the target voltage at node i, respectively. and These represent the actual reactive power output and the command value of resource j, respectively. and These are the weighting coefficients. and These represent the number of monitoring nodes and the number of controllable resources, respectively. See Table 2.

[0054] Table 2 Online Update Record of Collaborative Optimization Strategy Model Parameters It is understandable that, based on the magnitude and direction of the control deviation, the key parameters in the currently used collaborative optimization strategy model are adaptively fine-tuned. These key parameters include the weighting coefficients between the objective of minimizing the total active power loss and the objective of minimizing the sum of squared voltage deviations in the preventative optimization strategy model's objective function. and This includes boundary relaxation factors for constraints in various collaborative optimization strategy models, such as the upper and lower bound relaxation factors for node voltage safety constraints. Adaptive fine-tuning follows preset rules; for example, when the voltage control deviation remains positive, the weight of the voltage deviation term is appropriately increased. Alternatively, a relaxation factor can be used to loosen the upper limit of the voltage constraint. The fine-tuned parameters are then updated in the collaborative optimization strategy model for the next round of optimization calculations, thereby achieving incremental tracking and adaptation of the model parameters to the actual response characteristics of the system.

[0055] Optionally, the online parameter update step is completed in a separate post-evaluation and parameter tuning module. This module starts after each control command is executed and after a set observation delay, which is used to wait for the system's dynamic processes to subside. The module reads the latest round of real-time running data, calculates the current system state, and compares it with the target state expected to be achieved in the previous round of optimization calculation. The comparison result triggers the internal parameter tuning logic, which outputs the adjustment amount for specific model parameters according to the mapping relationship and tuning strategy shown in Table 2. The updated parameters are written to the parameter configuration file of the collaborative optimization strategy model, ensuring that if the scene recognition module calls the same type of collaborative optimization strategy model again in the next moment, the model will use the updated parameters to construct and solve the optimization problem.

[0056] In this embodiment, the online parameter update process demonstrates the adaptability of data-driven approaches. In the example scenario, if after multiple solutions and executions of the preventative optimization strategy model, the calculation reveals that the overall deviation (control deviation E) between the actual node voltage and the target voltage consistently exceeds the threshold, while the network loss control effect is good, the voltage deviation weight can be determined by referring to the rules shown in Table 2. The value is relatively low. Therefore, the parameter adjustment logic automatically increases the voltage deviation weight. The value, and correspondingly reduce the network loss weight. The value can be adjusted, for example, from (0.3, 0.7) to (0.2, 0.8). In the next round of preventative optimization, the optimization problem will focus more on reducing voltage deviation. This comparison and parameter adjustment based on historical control performance data allows the collaborative optimization strategy model to adapt to slow changes in network topology or load levels.

[0057] See Figure 4 This is an adaptive update diagram of parameters in a distribution network reactive power control model, visually presenting the dynamic adjustment process of control deviation and key model parameters. The control deviation continuously decreased from an initial 0.080 to a final 0.030, a reduction of 62.5%, proving that the online parameter update strategy effectively improved control accuracy. The voltage deviation weight and network loss weight adjusted inversely, dynamically optimizing between voltage safety and economic operation, which aligns with the core objectives of distribution network control. The voltage upper limit relaxation factor remained at 0.00 throughout, indicating that the system strictly adhered to voltage safety constraints and did not experience any risk of exceeding limits. After five iterations, the control deviation continued to decrease and stabilized, demonstrating the model's good convergence and robustness. The dynamic adjustment logic of multi-objective optimization weights is clearly demonstrated, reflecting the method's adaptability and intelligence. This provides data reference for tuning multi-objective optimization weights, achieving the optimal balance between voltage safety and economic operation.

[0058] Example 4 This embodiment provides a method for coordinated voltage regulation of multiple reactive resources in various scenarios of distribution networks. (See reference...) Figure 1 First, real-time operating data of multiple nodes in the distribution network is obtained; Based on the acquired real-time operating data, the static voltage sensitivity matrix and dynamic voltage stability margin index of the entire distribution network are calculated. Based on the static voltage sensitivity matrix and dynamic voltage stability margin index, the control scenarios of the current distribution network operation are identified, and based on the identified control scenarios, a collaborative optimization strategy model matching the control scenarios is invoked. The real-time operating data, static voltage sensitivity matrix, and dynamic voltage stability margin index are input into the called collaborative optimization strategy model for solution, generating a set of reactive power output instructions for multiple reactive resources. The reactive power output instruction set is subjected to timing coordination and equipment action feasibility verification to generate an executable collaborative control instruction sequence that meets timing and equipment constraints.

[0059] Optionally, after generating the final executable coordinated control command sequence, it is sent to the multi-reactive power resource control terminal deployed in the distribution network and drives the equipment to execute it. The executable coordinated control command sequence is sent to the multi-reactive power resource control terminal deployed in the distribution network, and the multi-reactive power resource control terminal drives the corresponding reactive power compensation equipment to perform reactive power output adjustment. The specific steps include: The executable collaborative control instruction sequence is decomposed according to the various reactive resource types and geographical locations. The decomposition process is based on the list of reactive compensation equipment under the control terminal and the network topology relationship, and is encapsulated into standardized instruction messages that can be recognized by different control terminals. Add a timestamp, instruction sequence number, and checksum to each standardized instruction message. The timestamp is used to indicate the absolute or relative time of instruction execution, the instruction sequence number is used to ensure the order and traceability of instructions, and the checksum is used to verify the integrity of the instruction message during transmission. Standardized instruction messages with added safety information are sent to the corresponding multi-source reactive power resource control terminals through the power distribution network communication network. After receiving the standardized instruction message, each multi-reactive-resource control terminal parses and executes the control commands within it, driving the static var generator, static synchronous compensator, distributed photovoltaic inverter, energy storage converter, and grouped switching capacitors to perform corresponding reactive power output adjustments. After the adjustment is completed, the multi-reactive-resource control terminal replies with an execution confirmation signal to the main station system.

[0060] Preferably, the encapsulation of the standardized command message follows a specific format. The message consists of a header, a command body, and a check field. The header contains the target control terminal address, message type, and length information. The command body contains the specific control command and its parameters. The check field is a checksum calculated based on the contents of the header and command body.

[0061] Preferably, the checksum calculation method uses cyclic redundancy check, and the formula is: in: This represents the calculated 16-bit checksum. This represents the 16-bit cyclic redundancy check (CRC) calculation function. Indicates the header byte sequence, Represents the instruction body byte sequence, symbol This indicates the concatenation of byte sequences. The timestamp is recorded in Coordinated Universal Time (UTC) format, and the instruction sequence number is a monotonically increasing sequence number within the same control terminal. The timestamp, instruction sequence number, and checksum are appended together to the end of the message.

[0062] In practical implementation, the process of parsing and executing control commands by the multi-source reactive power resource control terminal is localized and automated. The multi-source reactive power resource control terminal continuously listens to the communication network. When it receives a standardized command message sent to its own address, the terminal first verifies the correctness of the checksum. After the checksum verification is successful, the terminal parses the control command in the command body. Specifically: For a static var generator, the control command is a continuous reactive current reference value instruction, which is converted into a pulse width modulation signal by the multi-source reactive resource control terminal to drive the converter power module.

[0063] For static synchronous compensators, the control command is a reactive power setpoint instruction, and the multi-reactive power resource control terminal tracks this setpoint through its internal controller.

[0064] For distributed photovoltaic inverters, the control command is the reactive power factor setpoint or reactive power setpoint instruction. The multi-reactive resource control terminal calculates and executes the corresponding reactive power output based on the current active power output.

[0065] For energy storage converters, the control command is a reactive power setpoint instruction, which the multi-reactive power resource control terminal executes under the premise of satisfying the active power dispatch plan of the energy storage unit. For grouped switching capacitors, the control command is a switching switch status instruction, which the multi-reactive power resource control terminal drives relays or circuit breakers to perform the switching operation.

[0066] It is understandable that the issuance and execution processes involve timing logic and confirmation mechanisms. In the example scenario, the master station system generates an executable collaborative control command sequence containing five device instructions. The master station system decomposes this sequence into five independent sub-instructions based on resource type and geographical location, and encapsulates each sub-instruction into a standardized command message, adding a timestamp accurate to milliseconds, a unique command sequence number, and a checksum. The master station system then issues these five messages almost synchronously through the distribution network communication network.

[0067] The data comparison reveals that, with synchronous delivery, all multi-functional reactive power resource control terminals receive instructions at approximately the same time, but the execution time is determined by the timestamp in the instruction, which may be the same or sequential. With asynchronous delivery or network latency, different terminals receive instructions at different times, but each terminal executes the action based on the timestamp in the message, ensuring time synchronization of actions across devices.

[0068] After execution, the five multi-source reactive power resource control terminals respectively reply with execution confirmation signals to the master station system. The signals carry the corresponding instruction sequence number and execution result status code. The master station system completes the closed loop of this control cycle by comparing the confirmation signals with the issued instruction sequence number.

[0069] See Figure 5This is a visualization of the decomposition phase of executable collaborative control command sequences in a distribution network, intuitively presenting the decomposition process from the master station's overall command to the terminal's sub-commands. The deviation between the sub-commands of all devices and the original overall command is ≤2%, indicating almost no information loss during the command decomposition process and ensuring the consistency of the global optimization objective. Millisecond-level fast resources (SVG / STATCOM) exhibit the highest command decomposition accuracy, consistent with their high-priority positioning as core voltage support equipment; minute-level slow resources (capacitor banks) show slight attenuation due to communication and execution characteristics, consistent with engineering realities. The red line visually quantifies the execution effectiveness of the decomposed commands, providing a quantitative basis for subsequent command verification and closed-loop feedback. The command decomposition accuracy of different types of reactive power resources is quantified, providing data support for subsequent communication protocol optimization and terminal control strategy adjustment. The reliability of the command decomposition stage is quantified as a system performance evaluation indicator.

[0070] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for coordinated voltage regulation of multiple reactive resources in distribution networks across various scenarios, characterized in that, Includes the following steps: Acquire real-time operating data from multiple nodes in the power distribution network; Based on the acquired real-time operating data, the static voltage sensitivity matrix and dynamic voltage stability margin index of the entire distribution network are calculated. Based on the static voltage sensitivity matrix, the control scenarios of the current distribution network operation are identified, and based on the identified control scenarios, a collaborative optimization strategy model matching the control scenarios is invoked. The real-time operating data, static voltage sensitivity matrix, and dynamic voltage stability margin index are input into the called collaborative optimization strategy model for solution, generating a reactive power output instruction set for multiple reactive resources; The reactive power output instruction set is subjected to timing coordination and equipment action feasibility verification to generate an executable cooperative control instruction sequence that meets timing and equipment constraints; The specific steps for calculating the static voltage sensitivity matrix and the dynamic voltage stability margin index include: Based on the real-time operating data, an equivalent admittance network model of the current distribution network is established; Based on the equivalent admittance network model, the Jacobian matrix is ​​obtained by using the Newton-Raphson method for power flow calculation. The Jacobian matrix is ​​inverted at a set operating point to extract the partial derivatives of voltage with respect to active and reactive power injections, thus obtaining the static voltage sensitivity matrix. Based on the equivalent admittance network model and the dynamic response model of multiple reactive resources, a small-disturbance stability analysis model for the distribution network is constructed. Solve for the eigenvalues ​​of the small disturbance stability analysis model, and obtain the eigenvectors corresponding to the minimum damping ratio and the key oscillation mode based on these eigenvalues; The dynamic voltage stability margin index is obtained based on the eigenvectors corresponding to the minimum damping ratio and the key oscillation modes. Identifying the control scenario to which the current distribution network operation status belongs involves the following steps: Set thresholds for classifying control scenarios, including static voltage over-limit threshold, static voltage sensitivity threshold, and minimum damping ratio threshold; The deviation between the node voltage amplitude and the static voltage over-limit threshold is calculated to obtain the node voltage safety margin; The elements in the static voltage sensitivity matrix are compared with the static voltage sensitivity threshold to obtain the sensitivity of the node voltage to power changes. The calculated minimum damping ratio is compared with the minimum damping ratio threshold to obtain the small disturbance stability level of the system; Combining the node voltage safety margin, the node voltage sensitivity to power changes, and the stability level under small disturbances, the control scenario for the current operating state is determined based on the preset multi-level decision tree logic. The collaborative optimization strategy model includes a preventive optimization strategy model corresponding to the steady-state prevention scenario, a corrective optimization strategy model corresponding to the post-disturbance recovery scenario, and an emergency control strategy model corresponding to the emergency support scenario. The preventive optimization strategy model takes minimizing the total active power loss and the sum of squared voltage deviations as multiple objectives, and is constrained by the multi-dimensional reactive power resource regulation capability and the node voltage safety boundary. The corrective optimization strategy model prioritizes restoring all node voltages to a safe range as quickly as possible, and uses the coordination rule of adjusting the order of reactive resources with different speeds for collaboration. The emergency control strategy model takes suppressing voltage collapse as its core objective and the dynamic voltage support strength of key nodes as its optimization objective, while ignoring some economic constraints. Among them, the ability to regulate multiple reactive resources refers to the comprehensive ability to control the voltage of each node in the distribution network by adjusting the reactive power output of multiple reactive resources in the distribution network. The timing coordination and equipment operation feasibility verification of the aforementioned reactive power output instruction set is performed through the following steps: Check whether the change pace of different types of reactive power resource output instructions on the time axis in the reactive power output instruction set meets the preset coordination timing requirements. The coordination timing requirements specify the sequence of actions and coordination intervals between resources with fast adjustment capabilities and resources with slow adjustment capabilities. Verify whether each output command value in the reactive power output command set exceeds the physical adjustable range and action rate limit of the corresponding reactive power compensation device at the next moment. For instructions that do not meet the coordination timing requirements or exceed the feasibility limits of device actions, smooth correction and re-coordination shall be performed on the premise of meeting the core objectives of the coordination optimization strategy model. Generate an executable sequence of cooperative control instructions that fully satisfies timing and device constraints.

2. The method for coordinated voltage regulation of multiple reactive resources for distribution networks in multiple scenarios according to claim 1, characterized in that, Generating a reactive power output instruction set for multiple reactive power resources involves the following steps: Based on the invoked collaborative optimization strategy model, a mathematical optimization problem is constructed that includes the optimization objective, constraints, and coordination rules. The static voltage sensitivity matrix is ​​used as a simplified expression of the correlation between variables and the objective function in the optimization problem; The dynamic voltage stability margin index is used as a quantification parameter for stability constraints in the optimization problem; The real-time output and status information of the various reactive resources are used as the initial values ​​of the optimization variables and the boundary of the feasible region. The mathematical optimization problem is solved online using interior-point method, mixed-integer programming, or heuristic algorithm to obtain the reactive power output setpoints of various multi-dimensional reactive resources at each time point within a future control cycle, thus forming the reactive power output instruction set.

3. The method for coordinated voltage regulation of multiple reactive resources for distribution networks in multiple scenarios according to claim 2, characterized in that, When acquiring the real-time operating data, load power data of key feeders in the distribution network, actual and predicted output data of distributed power sources, and voltage support capability information of upstream grid interconnection points are acquired simultaneously as background data for confirming the network topology and boundary conditions of the mathematical optimization problem.

4. The method for coordinated voltage regulation of multiple reactive resources for distribution networks in multiple scenarios according to claim 1, characterized in that, The diverse reactive power resources include static var generators, static synchronizing compensators, distributed photovoltaic inverters, energy storage converters, and grouped switching capacitors. The reactive power output instruction set includes continuous reactive current reference value instructions for static var generators, reactive power setpoint instructions for static synchronous compensators, reactive power factor or reactive power setpoint instructions for distributed photovoltaic inverters, reactive power setpoint instructions for energy storage converters, and switching status instructions for grouped switching capacitors.

5. The method for coordinated voltage regulation of multiple reactive resources for distribution networks in multiple scenarios according to claim 1, characterized in that, It also includes the online parameter update steps for the collaborative optimization strategy model: After the executable coordinated control command sequence is executed, real-time operating data of the distribution network is acquired again to calculate new node voltages and system states; The control deviation is calculated by comparing the new node voltage and system state with the target state expected by the optimization model. Based on the magnitude and direction of the control deviation, the key parameters in the currently used collaborative optimization strategy model, including the objective function weights and constraint boundary relaxation factors, are adaptively fine-tuned. The fine-tuned parameters are updated in the collaborative optimization strategy model for use in the next round of optimization calculations.

6. The method for coordinated voltage regulation of multiple reactive resources for distribution networks in multiple scenarios according to claim 1, characterized in that, After generating the executable coordinated control command sequence, the executable coordinated control command sequence is sent to the multi-reactive power resource control terminal deployed in the distribution network. The multi-reactive power resource control terminal drives the corresponding reactive power compensation equipment to perform reactive power output adjustment, specifically including: The executable collaborative control instruction sequence is decomposed according to resource type and geographical location, and encapsulated into standardized instruction messages that can be recognized by different control terminals; Add a timestamp, instruction sequence number, and checksum to each standardized instruction message, and distribute it through the power distribution network communication network; After receiving the standardized instruction message, each multi-reactive power resource control terminal parses and executes the control commands in it, driving the static var generator, static synchronous compensator, distributed photovoltaic inverter, energy storage converter and group switching capacitor to perform corresponding reactive power output adjustment, and replies with an execution confirmation signal to the main station system.

Images