Source-grid-load collaborative voltage regulation system and method based on interval robust optimization

Abstract

The application discloses a source-grid-load collaborative voltage regulation system and method based on interval robust optimization, which comprises the following steps: obtaining relevant data of a target power grid, establishing distributed new energy output interval and load demand interval, mapping the new energy output interval and the load demand interval into node voltage offset interval based on the voltage sensitivity relationship of the source, the grid and the load, and forming corresponding voltage safe operation boundaries, setting a joint optimization target, establishing an interval robust optimization model, identifying the most unfavorable scenario causing the node voltage to approach or exceed the safe boundary based on the interval robust optimization model, and collaboratively optimizing according to the regulation capacity and response characteristics of the source side, the grid side and the load side to obtain a voltage regulation control strategy meeting the voltage constraint. The application reduces the probability of voltage out-of-limit, simultaneously considers the coordination of voltage regulation resource use and the continuity of the regulation process, and improves the stability level of power grid voltage operation and the economy of the voltage regulation process.

Description

Technical Field

[0001] This invention belongs to the field of source-grid-load voltage regulation technology, specifically relating to a source-grid-load coordinated voltage regulation system and method based on interval robust optimization. Background Technology

[0002] With the high proportion of distributed renewable energy sources, such as photovoltaics and wind power, being integrated into the distribution network, the power grid operation mode is gradually evolving from traditional unidirectional power supply to multi-source coexistence and bidirectional power flow. The output of renewable energy is significantly affected by meteorological conditions, exhibiting obvious randomness and volatility. At the same time, the load side shows characteristics of diversification and increased uncertainty, causing frequent fluctuations in node voltage levels within a short period of time, which can easily lead to voltage over-limit problems.

[0003] Existing voltage regulation methods mostly employ deterministic optimization or optimization strategies based on a single prediction scenario, typically solving for reactive power distribution or voltage regulation equipment operation under given operating conditions. However, these methods do not adequately consider the uncertainties of renewable energy output and load fluctuations, often relying on single-point prediction or limited scenario analysis, making it difficult to cover extreme disturbances that may occur in actual operation. This results in voltage exceeding limits still occurring during actual operation. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention proposes a source-grid-load coordinated voltage regulation system and method based on interval robust optimization, which takes into account both uncertainty modeling and multi-resource coordinated regulation, thereby improving the voltage regulation capability of the power grid under complex operating environments.

[0005] The technical solution of the present invention is as follows:

[0006] A source-grid-load coordinated voltage regulation method based on interval robust optimization includes:

[0007] Acquire relevant data of the target power grid, establish the output range and load demand range of distributed new energy sources, and map the output range and load demand range of new energy sources and load demand ranges to node voltage offset ranges based on the source-grid-load voltage sensitivity relationship, and form corresponding voltage safety operation boundaries. The relevant data includes network topology parameters, line parameters, node type information, distributed new energy related data and load related data.

[0008] A robust optimization model for the interval is established with the joint optimization objectives of minimizing the risk of voltage exceeding the limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation.

[0009] Based on the aforementioned interval robust optimization model, the most unfavorable scenario that causes the node voltage to exceed the safety boundary is identified. The regulation capability and response characteristics of the source side, grid side and load side are coordinated for optimization to obtain a voltage regulation control strategy that satisfies the voltage constraint.

[0010] The voltage regulation control strategy is used to control various voltage regulation resources, and the output range of new energy sources and the load demand range are updated in combination with real-time operating data. When the operating state deviates, the range robust optimization model is re-executed to perform continuous voltage regulation control.

[0011] Furthermore, the acquisition of relevant data from the target power grid, the establishment of distributed renewable energy output ranges and load demand ranges, and the mapping of these ranges to node voltage offset ranges based on the source-grid-load voltage sensitivity relationship, and the formation of corresponding voltage safety operation boundaries, include:

[0012] The network topology parameters, line parameters, node type information, distributed renewable energy related data, and load related data of the target power grid are obtained. Each node is divided into renewable energy-dominant nodes, load-dominant nodes, and hybrid coupling nodes, forming source-side disturbance units, load-side disturbance units, and corresponding node association sets. The nodes include bus nodes, feeder nodes, and end-connection nodes.

[0013] Based on distributed new energy related data and load related data, fluctuation segments reflecting the rising process, falling process, stagnation process and sudden change process are extracted respectively. The fluctuation segments are reorganized according to the time adjacency relationship and node association to establish the output range basic element of each source-side disturbance unit and the demand range basic element of each load-side disturbance unit.

[0014] Based on the topological hierarchy and voltage transmission direction among the new energy dominant node, load dominant node and hybrid coupling node, the output interval basic element and the demand interval basic element are combined and paired to construct the source-load coupling disturbance sequence.

[0015] According to the order of influence of source-side active power disturbance, source-side reactive power disturbance, load-side active power disturbance and load-side reactive power disturbance on node voltage, a hierarchical voltage sensitivity relationship is established, and the source-load coupling disturbance sequence is screened according to the hierarchical voltage sensitivity relationship to determine the dominant disturbance sequence and the accompanying disturbance sequence participating in the node voltage offset calculation, and the voltage offset interval corresponding to each node is generated.

[0016] Based on the voltage offset interval, the upper and lower voltage offset segments of each node are segmented and transformed. The transformed node voltage offset results are then superimposed and verified with the node's allowable operating range to form the voltage safety operating boundary corresponding to different nodes.

[0017] Further, the step of screening the source-load coupling disturbance sequence according to the hierarchical voltage sensitivity relationship to determine the dominant disturbance sequence and accompanying disturbance sequence participating in the node voltage offset calculation, and generating the voltage offset interval corresponding to each node, includes:

[0018] The source-load coupled disturbance sequence is decomposed into source-side active disturbance, source-side reactive disturbance, load-side active disturbance, and load-side reactive disturbance. Based on the electrical connection relationship between nodes, the transmission path of each disturbance in the network is calibrated to form a set of disturbance effect paths corresponding to different nodes.

[0019] Based on the hierarchical voltage sensitivity relationship, each disturbance sequence in the disturbance action path set is matched node by node, and the disturbance sequences are sorted according to the transmission level and action order of the disturbance's influence on the target node voltage to form an ordered disturbance sequence queue for each node.

[0020] In the ordered perturbation sequence queue, according to the preset sequence truncation rules, the perturbation sequences located in the preceding order are aggregated to determine the dominant perturbation sequence set, and the remaining perturbation sequences are determined as the accompanying perturbation sequence set;

[0021] The dominant disturbance sequence set and the accompanying disturbance sequence set are combined and superimposed according to the order of their action paths, and node voltage offset process segments are constructed according to the entry order of the disturbance sequences at different action stages, and converged to form the voltage offset interval of the corresponding node.

[0022] Furthermore, based on the voltage offset interval, the upper and lower voltage offset segments of each node are segmented and transformed, and the transformed node voltage offset results are superimposed and verified with the node's allowable operating range to form a voltage safety operating boundary corresponding to different nodes, including:

[0023] The voltage offset interval is divided into an upper voltage offset segment and a lower voltage offset segment according to the direction of voltage change. Based on the order of appearance of each offset segment in the disturbance sequence, the upper voltage offset segment and the lower voltage offset segment are sorted into an offset segment sequence.

[0024] Based on the topology level of each node, the transmission order of adjacent nodes, and the reference voltage state corresponding to the current node, the turning point of the offset segment sequence is determined segment by segment.

[0025] Starting from the aforementioned turning point, the voltage upward offset segment and voltage downward offset segment of each node are segmented and turned to form a continuous voltage offset trajectory band corresponding to each node.

[0026] The continuous voltage offset trajectory of each node is superimposed and verified with the allowable operating range of that node segment by segment. Turning segments within the allowable operating range, turning segments overlapping with the allowable operating range, and turning segments exceeding the allowable operating range are identified. The boundary turning point is taken as the boundary turning point to form the voltage safety operating boundary corresponding to different nodes.

[0027] Furthermore, a robust interval optimization model is established with the joint optimization objectives of minimizing the risk of voltage exceeding the limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation operations. This model includes:

[0028] The joint optimization objective of minimizing the risk of voltage exceeding the limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation actions is broken down into risk control sub-objectives, cost constraint sub-objectives, and action sequence sub-objectives. All sub-objectives are then mapped to voltage regulation tasks on the source side, grid side, and load side, forming a set of voltage regulation tasks.

[0029] The new energy output range, load demand range, and voltage safety operation boundary are embedded into the voltage regulation task set to generate a regulation state sequence and construct a corresponding collaborative voltage regulation sequence.

[0030] Based on the aforementioned coordinated voltage regulation sequence, each sub-target is sorted, and the voltage constraints and voltage regulation action constraints under interval disturbances are merged to form a robust constraint set.

[0031] Based on the voltage regulation task set, the coordinated voltage regulation sequence, and the robust constraint set, an interval robust optimization model is constructed.

[0032] Furthermore, based on the aforementioned voltage regulation task set, coordinated voltage regulation sequence, and robust constraint set, an interval robust optimization model is constructed, including:

[0033] Based on the voltage regulation task set and the coordinated voltage regulation sequence, the regulation actions corresponding to various voltage regulation resources are serialized and expressed according to the triggering order, forming a voltage regulation variable sequence that includes source-side regulation variables, grid-side regulation variables and load-side regulation variables;

[0034] The new energy output range and the load demand range are expanded according to the disturbance entry sequence to form multiple consecutive interval disturbance segments. The interval disturbance segments are then matched with the voltage regulation variable sequence segment by segment to construct the connection relationship between the disturbance segments and the voltage regulation action.

[0035] The robust constraint set is embedded into the connection relationship between the disturbance segment and the voltage regulation action. The continuity constraint of voltage regulation action, node voltage constraint and regulation conflict constraint between different disturbance segments are encapsulated across segments to form a unified constraint system covering multiple disturbance segments.

[0036] Based on the voltage regulation variable sequence, interval disturbance segments, and unified constraint system, the voltage regulation process corresponding to each disturbance segment is connected as a whole to form an interval robust optimization model.

[0037] Furthermore, based on the aforementioned interval robust optimization model, the most unfavorable scenario causing the node voltage to exceed the safety boundary is identified. Coordinated optimization is then performed according to the regulation capabilities and response characteristics of the source side, grid side, and load side to obtain a voltage regulation control strategy that satisfies the voltage constraint, including:

[0038] Based on the aforementioned interval robust optimization model, the new energy output interval, load demand interval, and voltage regulation variable sequence are expanded accordingly to form multiple candidate disturbance scenarios corresponding to the node voltage change process. The candidate disturbance scenarios are then aggregated to form a scenario set corresponding to different nodes.

[0039] Each candidate disturbance scenario in the scenario set is compared with the voltage safe operation boundary of the corresponding node one by one, and the disturbance scenarios that cause the node voltage to enter the boundary adjacent segment, cross the boundary turning position, or continuously wander along the boundary are selected as boundary approach scenarios.

[0040] The voltage offset propagation process in the boundary approach scenario is rearranged to determine the most unfavorable scenario as the disturbance scenario that first triggers voltage constraint conflict or continuously occupies voltage regulation margin under the current coordinated voltage regulation sequence.

[0041] Based on the worst-case scenario, according to the adjustment initiation conditions, action succession relationships and response sequences of source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources, various voltage regulation resources are hierarchically ordered to form a coordinated voltage regulation process chain.

[0042] The collaborative voltage regulation process chain is embedded into the interval robust optimization model corresponding to the most unfavorable scenario. The entry order, holding interval and exit order of each voltage regulation resource in different adjustment segments are jointly obtained to generate a voltage regulation control strategy that satisfies the node voltage constraint.

[0043] Furthermore, based on the most unfavorable scenario, and according to the adjustment initiation conditions, action succession relationships, and response sequences of source-side voltage regulation resources, grid-side voltage regulation resources, and load-side voltage regulation resources, various voltage regulation resources are hierarchically ordered to form a coordinated voltage regulation process chain, including:

[0044] Based on the worst-case scenario, scenario matching is performed on the source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources respectively, and the voltage triggering state, power offset state and adjustable margin state corresponding to each type of voltage regulation resource are identified, and the set of adjustment start conditions for each type of voltage regulation resource is determined accordingly.

[0045] Based on the aforementioned set of adjustment initiation conditions, the adjustment actions of various voltage regulation resources are arranged and sequenced to form an action sequence.

[0046] The adjustment actions in the action sequence are folded, the adjustment actions with overlapping responses are grouped into parallel segments, the adjustment actions with staggered responses are divided into front and back segments, and the action sequence is divided into a preceding adjustment layer, a following adjustment layer, and a subsequent adjustment layer accordingly.

[0047] The preceding adjustment layer, the subsequent adjustment layer, and the subsequent adjustment layer are connected in series according to the disturbance progression order in the most unfavorable scenario. The entry order, maintenance order, and exit order of the adjustment actions within each layer are uniformly organized to form a coordinated voltage regulation process chain corresponding to the most unfavorable scenario.

[0048] Furthermore, the coordinated voltage regulation process chain is embedded into the interval robust optimization model corresponding to the most unfavorable scenario. The entry order, holding interval, and exit order of each voltage regulation resource in different adjustment segments are jointly calculated to generate a voltage regulation control strategy that satisfies the node voltage constraint, including:

[0049] The coordinated voltage regulation process chain is embedded into the corresponding interval robust optimization model according to the disturbance advancement order in the most unfavorable scenario, and each adjustment layer in the coordinated voltage regulation process chain is aligned to the different voltage offset stages corresponding to the most unfavorable scenario, forming a correspondence between the adjustment layer and the scenario stage.

[0050] Based on the correspondence between the regulation layer and the scenario stage, the voltage regulation process in the most unfavorable scenario is divided into regulation segments, and the source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources are respectively connected to the corresponding regulation segments to form a set of regulation segment resources.

[0051] Within the aforementioned robust optimization model, for each set of resources in each regulation segment, the entry order, holding interval, and exit order of each voltage regulation resource in the corresponding regulation segment are jointly calculated, and the continuity of the resource switching relationship between adjacent regulation segments is checked to form the resource action sequence corresponding to each regulation segment.

[0052] The resource action sequences corresponding to each adjustment segment are merged and organized according to the scenario stages in the most unfavorable scenario to form a voltage regulation control strategy. The voltage regulation control strategy includes the action entry arrangement, segment maintenance arrangement, and stage exit arrangement for each voltage regulation resource.

[0053] A source-grid-load coordinated voltage regulation system based on interval robust optimization is used to implement the source-grid-load coordinated voltage regulation method based on interval robust optimization as described above, including: a data processing module, a model building module, a strategy generation module, and a coordinated voltage regulation module;

[0054] The data processing module is used to acquire relevant data of the target power grid, establish distributed renewable energy output range and load demand range, and map the renewable energy output range and load demand range to node voltage offset range based on the source-grid-load voltage sensitivity relationship, and form corresponding voltage safety operation boundaries. The relevant data includes network topology parameters, line parameters, node type information, distributed renewable energy related data and load related data.

[0055] The model building module is used to establish an interval robust optimization model with the joint optimization objectives of minimizing the risk of voltage over-limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation.

[0056] The strategy generation module is used to identify the most unfavorable scenario that causes the node voltage to exceed the safety boundary based on the interval robust optimization model, and to perform collaborative optimization according to the adjustment capabilities and response characteristics of the source side, grid side and load side to obtain a voltage regulation control strategy that meets the voltage constraints.

[0057] The coordinated voltage regulation module is used to control various voltage regulation resources according to the voltage regulation control strategy, and to update the new energy output range and load demand range in combination with real-time operating data. When the operating state deviates, the range robust optimization model is re-executed to perform continuous voltage regulation control.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] This invention proposes a source-grid-load coordinated voltage regulation system and method based on interval robust optimization. An interval robust optimization model is established, and voltage regulation resources are hierarchically ordered and coordinated according to the source side, grid side, and load side. Simultaneously, a worst-case scenario screening and voltage regulation process chain embedding mechanism are introduced to jointly determine the entry order, maintenance process, and exit order of voltage regulation resources, thereby forming a coordinated voltage regulation strategy oriented towards the entire process of uncertain disturbances. Through this approach, the voltage regulation process is transformed from a single static optimization to a dynamic coordinated decision-making process oriented towards the evolution of interval disturbances. This allows for more comprehensive coverage of possible operating states when dealing with new energy output and load fluctuations, reducing the probability of voltage exceeding limits, while also considering the coordination of voltage regulation resource use and the continuity of the regulation process, thereby improving the stability of grid voltage operation and the economy of the voltage regulation process. Attached Figure Description

[0060] Figure 1 A flowchart of the source-grid-load coordinated voltage regulation method based on interval robust optimization provided by the present invention;

[0061] Figure 2 This is a schematic diagram of voltage offset provided by the present invention;

[0062] Figure 3 A schematic diagram of the interval robust optimization model provided by this invention;

[0063] Figure 4 The architecture diagram of the source-grid-load coordinated voltage regulation system based on interval robust optimization provided by the present invention is shown. Detailed Implementation

[0064] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. It should be understood that the following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Equivalent modifications or substitutions made by those skilled in the art without departing from the concept of the present invention should fall within the scope of protection defined by the appended claims.

[0065] Example 1:

[0066] Please see Figures 1-3 The present invention provides an embodiment of a source-grid-load coordinated voltage regulation method based on interval robust optimization, comprising the following specific steps:

[0067] Step S1: Obtain relevant data of the target power grid, establish the output range and load demand range of distributed renewable energy, and map the renewable energy output range and load demand range to node voltage offset range based on the source-grid-load voltage sensitivity relationship, and form the corresponding voltage safety operation boundary. The relevant data includes network topology parameters, line parameters, node type information, distributed renewable energy related data and load related data.

[0068] The specific steps of step S1 are as follows:

[0069] Step S101: Obtain the network topology parameters, line parameters, node type information, distributed renewable energy related data and load related data of the target power grid, divide each node into renewable energy-dominant nodes, load-dominant nodes and hybrid coupling nodes, and form source-side disturbance units, load-side disturbance units and corresponding node association sets. The nodes include bus nodes, feeder nodes and end access nodes.

[0070] In this embodiment, based on the network topology parameters and line connection relationships of the target power grid, a node-branch abstract model of the power grid is performed. Electrical connection locations are uniformly mapped to bus nodes, feeder nodes, and end-connection nodes, with nodes serving as the basic analysis units for power injection, transmission, and aggregation. Subsequently, combining the access location and capacity data of distributed renewable energy sources with the access location and electricity consumption characteristic data of loads, the power injection direction and changing trends of each node are analyzed. By classifying the temporal changes in power flow direction in historical operating data, the source-side dominant characteristics or load-side dominant characteristics exhibited by nodes in different operating periods are identified. Based on this, the relative relationship between the proportion of renewable energy output and the proportion of load demand at a node is used as the criterion to classify nodes into renewable energy-dominant nodes, load-dominant nodes, and hybrid coupled nodes where both have a significant impact. Furthermore, based on the above node classification results, nodes with the same dominant attributes and associated paths in the topology are merged to form source-side disturbance units and load-side disturbance units, respectively. Node association sets are constructed based on the electrical connection relationships and power transfer paths between nodes, so that each disturbance unit corresponds to a set of nodes with a clear scope of action in subsequent analysis.

[0071] For example, in a 10kV radial distribution network of a commercial park, one bus node B0, one renewable energy-dominant node N1, one load-dominant node N2, and one hybrid coupling node N3 are set up. N1 is connected to distributed photovoltaic (PV) power, N2 is connected to centralized loads, and N3 is connected to both small-capacity PV and flexible loads. Based on the type of resources connected to the nodes and the power flow direction, N1 is classified as a renewable energy-dominant node, N2 as a load-dominant node, and N3 as a hybrid coupling node. Nodes N1, N2, and N3 are connected to bus node B0 via feeders, forming a node association set with B0 as the upstream power supply node and N1 to N3 as downstream access nodes. In the initial operating state, the reference voltages of N1, N2, and N3 are 1.020 pu, 1.006 pu, and 1.018 pu, respectively, and the allowable operating range of each node voltage is set to 0.95 pu to 1.05 pu.

[0072] Step S102: Based on distributed new energy related data and load related data, extract fluctuation segments reflecting the rising process, falling process, stagnant process and sudden change process respectively, and reorganize the fluctuation segments according to the time adjacency relationship and node association to establish the output range basic element of each source-side disturbance unit and the demand range basic element of each load-side disturbance unit.

[0073] In this embodiment, distributed renewable energy-related data and load-related data undergo unified time-scale alignment and sequence cleaning. Missing points, abnormal jump points, and discontinuous sampling segments are removed or filled in to ensure the continuous comparability of data sequences corresponding to nodes within the same disturbance unit. Subsequently, the change trajectory of each data sequence is segmented and identified along the time axis. Segments that continuously increase and maintain a consistent direction of change are extracted as upward process fluctuation segments; segments that continuously decrease and maintain a consistent direction of change are extracted as downward process fluctuation segments; segments that maintain narrow fluctuations over a certain period of time are extracted as stagnant process fluctuation segments; and segments that undergo significant jumps between adjacent sampling times and are accompanied by local trend abrupt changes are extracted as abrupt process fluctuation segments. Among these, the fluctuation segments on the distributed renewable energy side focus on characterizing the power output evolution caused by irradiance changes, wind speed disturbances, and inverter operation switching; while those on the load side... The fluctuation segments focus on characterizing the demand evolution caused by load start-up and shutdown, aggregated response, and time period switching. After extracting single-node fluctuation segments, the segments that are connected end-to-end, have compatible trends, or are of different types but occur consecutively in time are spliced ​​together based on temporal adjacency. At the same time, the node fluctuation segments that are in the same disturbance unit and have transmission correlation in the electrical path are merged in parallel based on the node association set, so that the originally discrete local fluctuation segments are reorganized into a continuous fluctuation chain that reflects the unit-level disturbance behavior. Furthermore, the reorganized continuous fluctuation chain is enveloped and sorted according to its upper bound state, lower bound state, and intermediate transition state in time to form an output interval primitive that can characterize the range of output change of the source-side disturbance unit in a certain operating period, and a demand interval primitive that can characterize the range of demand change of the load-side disturbance unit in the corresponding operating period.

[0074] A voltage regulation cycle is divided into four consecutive time periods: T1, T2, T3, and T4, each lasting 15 minutes. Based on the forecasts of renewable energy output and load demand, fluctuation segments such as rapid rise in photovoltaic power, high photovoltaic output, load decline, and output drop are extracted and formed into interval primitives as shown in Table 1.

[0075] Table 1. Source load disturbance intervals and characteristics for each time period

[0076] T1 [0.02,0.08] [0.00,0.02] [0.05,0.12] [0.02,0.05] Normal fluctuations T2 [0.40,0.55] [0.05,0.10] [-0.05,0.05] [-0.02,0.02] Photovoltaics are rising rapidly. T3 [0.55,0.72] [0.08,0.12] [-0.18,-0.05] [-0.07,-0.02] High output of photovoltaic power and reduced load T4 [0.10,0.25] [0.02,0.04] [0.00,0.04] [0.00,0.02] Output decline

[0077] Among them, T2 corresponds to the rapid increase in renewable energy output, T3 corresponds to the period when renewable energy maintains high output and load demand decreases, and T4 corresponds to the period when renewable energy output declines.

[0078] Step S103: Based on the topological hierarchy between the new energy dominant node, the load dominant node and the hybrid coupling node and the voltage transmission direction, the output interval basic element and the demand interval basic element are combined and paired to construct a source-load coupling disturbance sequence containing the same-direction disturbance sequence, the opposite-direction disturbance sequence and the alternating disturbance sequence.

[0079] In this embodiment, based on the hierarchical expansion order from bus nodes to end access nodes in the power grid topology, and combined with the power transmission path from the source side to the load side, the topological hierarchical relationship between each new energy dominant node, load dominant node, and hybrid coupling node is determined. Based on this, the influence direction of output interval basic elements and demand interval basic elements in the network is calibrated. Calibration is performed through the power grid topology hierarchy and power transmission direction; that is, based on the hierarchical order from bus to end access nodes and the power flow direction from the source side to the load side, the positive or negative influence of each basic element along the electrical path on the node voltage is determined. Subsequently, using nodes within the same node association set or reachable through an electrical path as the pairing range, output interval basic elements that are temporally adjacent or appear in the same operating period are combined and paired with demand interval basic elements. During the pairing process, the changes of both types of interval basic elements are recorded simultaneously. The disturbance sequence is determined by the direction, order of occurrence, and duration. Based on this, according to the consistency or oppositeity of the change direction of the output interval element and the demand interval element within the same time period, the pairing results with the same change direction and superimposed effect on voltage are classified into the same-direction disturbance sequence, and the pairing results with opposite change direction and offsetting effect on voltage are classified into the opposite-direction disturbance sequence. For pairing results that alternate in time and switch change direction between the preceding and following stages, an alternating disturbance sequence is constructed according to their switching order. Furthermore, the above three types of disturbance sequences are rearranged according to the propagation path from upstream to downstream or from key nodes to associated nodes according to the topological hierarchy, so that each disturbance sequence reflects the transmission order and action path of the disturbance in the network, thereby forming a source-load coupled disturbance sequence that can simultaneously reflect the changes in the source and load intervals and their coupling propagation characteristics in the power grid.

[0080] During period T2, the photovoltaic output on the N1 side increases, while the load changes on the N2 and N3 sides are relatively small. At this time, the increase in source-side output has a lifting effect on the node voltage, forming a co-directional disturbance sequence dominated by source-side disturbance. During period T3, the photovoltaic output continues to remain high, while the load demand decreases. Both the increase in source-side output and the decrease in load on the load side will cause the node voltage to rise. The two are superimposed in the direction of voltage influence, thus forming a co-directional enhanced source-load coupled disturbance sequence. During period T4, the photovoltaic output falls back, the load demand tends to stabilize, and the voltage rise effect weakens, forming a falling disturbance sequence. Therefore, T1 to T4 can be summarized as a source-load coupled disturbance sequence of normal fluctuation - photovoltaic rise - co-directional enhancement of source and load - output fall.

[0081] Step S104: Establish a hierarchical voltage sensitivity relationship according to the order of influence of source-side active power disturbance, source-side reactive power disturbance, load-side active power disturbance and load-side reactive power disturbance on node voltage, and screen the source-load coupling disturbance sequence according to the hierarchical voltage sensitivity relationship to determine the dominant disturbance sequence and accompanying disturbance sequence participating in the node voltage offset calculation, and generate the voltage offset interval corresponding to each node.

[0082] It should be noted that the hierarchical voltage sensitivity relationship is determined jointly by the power grid's operating status and topology. Specifically, firstly, based on the target power grid's network topology and operating mode, the electrical connections between nodes are modeled, and the hierarchical position of each node in the voltage transmission path is identified. Subsequently, by analyzing the response characteristics of node voltage to changes in active and reactive power, the direction and degree of influence of different types of disturbances on node voltage are determined. For example, when the output of photovoltaic nodes increases (active power rises), the bus voltage will rise slightly; when the power consumption of load nodes increases (active power rises), the bus voltage drops, and the impact on voltage is negative, with the magnitude varying with the load size. Based on this, the impact of disturbances is hierarchically classified according to the topological distance between the disturbance source node and the target node, the connection path, and whether it passes through critical nodes. Disturbances directly affecting the target node are classified as first-level influences, disturbances transmitted through adjacent nodes are classified as second-level influences, and disturbances transmitted through multiple levels of nodes are classified as higher-level influences. Furthermore, within the same level, the order of influence of each disturbance is sorted according to the order of their arrival time and duration, thus forming a hierarchical voltage sensitivity relationship that includes both spatial hierarchy information and temporal sequence information, which is used for subsequent disturbance screening and voltage offset calculation.

[0083] The specific steps of step S104 are as follows:

[0084] Step S1041: Decompose the source-load coupled disturbance sequence into source-side active disturbance, source-side reactive disturbance, load-side active disturbance, and load-side reactive disturbance, and based on the electrical connection relationship between nodes, calibrate the transmission path of each disturbance in the network to form a set of disturbance action paths corresponding to different nodes.

[0085] In this embodiment, taking the source-load coupled disturbance sequence formed in step S103 as input, each coupled disturbance sequence is first decoupled and identified according to the disturbance source and power attribute. That is, the power injection change from the distributed renewable energy side is split into source-side active power disturbance and source-side reactive power disturbance, and the power absorption change from the load side is split into load-side active power disturbance and load-side reactive power disturbance, thereby restoring the originally mixed-expression coupled disturbance sequence into several basic disturbance components with single-effect attributes. After completing the disturbance attribute splitting, combined with the node-branch topology relationship obtained in step S101, for each basic disturbance component, starting from its occurrence node, it is traced step by step along the electrical connection path to its potentially affected adjacent nodes, upstream nodes, and downstream nodes, and according to the line connection sequence and branch bifurcation position. The propagation path of disturbances in the network is marked segment by segment according to the node level and power transfer direction, so that each disturbance component corresponds to one or more traceable transmission paths. For disturbance components that affect different areas simultaneously through multiple branches, they are expanded in a way that prioritizes the main path and records the branch paths in parallel, so as to avoid simplifying the propagation of multiple branches into a single path and losing the coupling relationship between nodes. Furthermore, the entry nodes, passing nodes and arrival nodes of each basic disturbance component on different paths are sequentially associated to form a disturbance effect path record oriented to specific nodes. These path records are then collected according to the target node, so that each node obtains a set of disturbance effect paths composed of source-side active disturbance, source-side reactive disturbance, load-side active disturbance and load-side reactive disturbance respectively.

[0086] Step S1042: Based on the hierarchical voltage sensitivity relationship, perform node-by-node matching on each disturbance sequence in the disturbance action path set, and sort the disturbance sequences according to the transmission level and action order of the disturbance's influence on the target node voltage to form an ordered disturbance sequence queue for each node.

[0087] In this embodiment, a corresponding path reception view is established for each target node. Source-side active power disturbance, source-side reactive power disturbance, load-side active power disturbance, and load-side reactive power disturbance that can propagate to the target node along the electrical connection path are extracted and matched node-by-node with a pre-established hierarchical voltage sensitivity relationship. This hierarchical voltage sensitivity relationship does not only reflect the static correspondence between the disturbance quantity and the node voltage change, but also distinguishes the disturbance's level, the target node's level, and the progressive influence order of the disturbance's cross-level propagation. This ensures that each disturbance sequence can obtain an influence level identifier corresponding to the target node during matching. Subsequently, each disturbance sequence entering the target node is compared one by one according to the following order: first determining whether it directly affects the target node, then determining whether it is transmitted across levels, and finally determining whether it is transferred via a hybrid coupling node. Those directly affecting the target node or those locally associated with the target node are identified. Disturbances within the target node's range are prioritized as preceding sequences, while those requiring transmission through upstream, downstream, or intermediate coupling nodes are identified as subsequent sequences. After hierarchical division, the disturbance sequences within the same level are further refined and sorted based on their temporal entry order, path arrival order, and overlap with other disturbances within the node's local area. Disturbances that enter and act first and form a continuous chain of influence on the target node are placed first, while those that intervene later or only have a secondary effect are placed later. For disturbances that simultaneously satisfy the same level and approximate entry order, they are reordered based on whether their path passes through key nodes, whether they form repeated effects within the node association set, and whether they have a connection with existing sequences. Finally, an ordered queue of disturbance sequences is formed for each target node, arranged sequentially according to transmission level and action order.

[0088] Step S1043: In the ordered perturbation sequence queue, according to the preset sequence truncation rules, the perturbation sequences located in the preceding order are aggregated to determine the dominant perturbation sequence set, and the remaining perturbation sequences are determined to be the accompanying perturbation sequence set. The sequence truncation rules are constructed based on the superposition and correlation relationship of perturbations on different propagation paths.

[0089] In this embodiment, each disturbance sequence is examined segment by segment along the order of the sequence queue. The disturbance sequence located at the previous position is taken as the candidate aggregation object, and combined with the transmission path information marked in step S1041, the path nodes, branch nodes, and convergence nodes that each disturbance passes through before entering the target node are identified. On this basis, the correlation analysis of disturbance sequences in adjacent or close positions is performed based on the criteria of "whether they form continuous transmission on the same electrical path, whether they have superposition effects within the node association set, and whether they form a sequential relationship in time". For disturbance sequences that show continuous entry on the same path or adjacent paths and form superposition effects in the local area of ​​the node, they are combined and merged according to their order of appearance, and gradually expanded to form a set of disturbance subsets that continuously cover the voltage change process of the target node. During the combination process, when When subsequent disturbance sequences are detected to separate along the path, exhibit significant temporal discontinuities, or no longer form a direct connection with the current aggregated sequence within the node range, the current aggregation process is terminated, and the aggregated subset of disturbances is determined as the dominant disturbance sequence set. Subsequently, the remaining disturbance sequences in the original ordered disturbance sequence queue that were not included in the above aggregation process are uniformly classified into the accompanying disturbance sequence set, and their original hierarchical position and path identification information are retained for use as supplementary influence items in subsequent offset interval combinations. Through the above sequence truncation and aggregation method based on path superposition correlation, the dominant disturbance sequence set can cover the key disturbance processes that have a continuous dominant effect on node voltage changes, while maintaining the accompanying disturbance sequences as a supplementary source of secondary influences, thereby obtaining a disturbance sequence division result with a clear structure and distinct hierarchy.

[0090] It should be noted that the preset sequence truncation rules are constructed based on the correlation of disturbance paths and temporal continuity. Specifically, when processing disturbance sequences, the sequences are first arranged according to the time order in which the disturbances enter the target node. Then, starting from the beginning of the sequence, the correlation between each disturbance along the electrical path is analyzed one by one. For disturbance sequences that appear consecutively on the same or adjacent paths and do not have obvious temporal discontinuities, they are grouped into the same dominant disturbance set. When subsequent disturbances show obvious separation along the path, or are interrupted in time, or no longer have a continuous influence relationship with the preceding disturbances, the expansion of the current dominant sequence is terminated, and the remaining disturbances are classified as accompanying disturbance sequences.

[0091] Step S1044: Combine and superimpose the dominant disturbance sequence set and the accompanying disturbance sequence set according to the order of their action paths, and construct node voltage offset process segments according to the entry order of the disturbance sequences at different action stages, and converge to form the voltage offset interval of the corresponding node.

[0092] In this embodiment, according to the transmission paths marked in step S1041, the order in which each disturbance sequence enters the target node is uniformly sorted. The disturbance sequence that arrives at the target node first along the main path is taken as the preceding sequence, and the disturbance sequence that arrives via branch paths or after being transferred through intermediate nodes is taken as the subsequent sequence. Based on this, a path sequence framework oriented towards the target node is constructed. Subsequently, with the path sequence framework as a reference, each disturbance in the dominant disturbance sequence set is unfolded segment by segment according to its entry order, serving as the backbone of the dominant process of node voltage change. The disturbances in the accompanying disturbance sequence set are embedded into the corresponding stages according to the intersection of their corresponding paths and the dominant sequence, so that disturbances from different sources and different paths can be integrated into the target node. The perturbations along the path form corresponding superimposed relationships in time and space. After the sequence embedding is completed, the node voltage change process is segmented according to the order of entry of each perturbation sequence at different stages. The perturbations that act together in the same stage are merged into a voltage offset process segment, and the segments are connected in series according to the stage progression order to form a segment sequence that reflects the entire voltage change process. Furthermore, the upper and lower bound change trajectories of each voltage offset process segment are extracted respectively, and they are aggregated as a whole while maintaining the segment order. The voltage offset ranges that may occur in each stage are continuously spliced ​​to form the voltage offset interval of the corresponding target node under the action of this type of perturbation.

[0093] Step S105: Based on the voltage offset interval, according to the node type, topology location and the offset transmission order of adjacent nodes, the upper and lower voltage offset segments of each node are segmented and transformed. The transformed node voltage offset results are superimposed and verified with the node's allowable operating range to form the voltage safety operating boundary corresponding to different nodes.

[0094] like Figure 2 As shown, the specific steps of step S105 are as follows:

[0095] Step S1051: Divide the voltage offset interval into voltage upper offset segment and voltage lower offset segment according to the direction of voltage change, and sort the voltage upper offset segment and voltage lower offset segment according to the order of appearance of each offset segment in the disturbance sequence to form the offset segment sequence corresponding to each node.

[0096] In this embodiment, the voltage offset interval is analyzed segment by segment along its change process. Segments showing a continuous upward trend in voltage due to disturbance are identified as voltage upward offset segments, and segments showing a continuous downward trend are identified as voltage downward offset segments. The segments are further divided at the points where adjacent offset directions switch, thus decomposing the original continuous voltage offset interval into several offset segments with a single direction of change. Subsequently, each offset segment is back-linked to its source disturbance sequence to determine its corresponding position in the dominant disturbance sequence and accompanying disturbance sequences. Furthermore, considering the order in which the disturbances enter the target node, the generation order of each offset segment and its phase during the voltage change process are recorded. For stage location; after completing direction splitting and source calibration, the voltage upper offset segment and voltage lower offset segment are independently organized. Offset segments originating from adjacent disturbance stages and connected in time are sequentially spliced ​​together. Offset segments that are discontinuous in time or introduced by different paths are kept segmented records, thus forming two sets of offset segment sequences with clear stage boundaries. Furthermore, according to the overall progression order of the disturbance sequence, the above upper offset segment sequence and lower offset segment sequence are uniformly sorted so that the arrangement of each offset segment in the sequence reflects both its voltage change direction and its actual occurrence order in the disturbance propagation process, finally obtaining a structured offset segment sequence oriented towards each node.

[0097] Step S1052: Based on the topology level of each node, the transmission order of adjacent nodes, and the reference voltage state corresponding to the current node, determine the turning point for each segment of the offset segment sequence. For offset segments formed by transmission from adjacent nodes, the state at the beginning of the segment when it is transmitted into the current node is used as the turning point. For offset segments formed by disturbances of the current node itself, the reference voltage state corresponding to the current node is used as the turning point.

[0098] In this embodiment, the hierarchical position of each node is marked according to the power grid topology, clarifying its upstream, downstream, and adjacent nodes in the electrical path. Based on the transmission order of disturbances in the network, the source of each offset segment is traced back to determine whether it is an external offset segment transmitted from adjacent nodes along the electrical path or a local offset segment directly caused by changes in the current node's internal output or load. On this basis, for external offset segments, the corresponding disturbance effect path is traced to the boundary position where the offset segment enters the current node, and the initial state of the offset segment when entering the node is used as the turning point. The starting point ensures that the transition process maintains a connection with the voltage change process of the upstream node. For the local offset segment, the reference voltage state of the current node before the disturbance of this segment is superimposed is used as the starting reference, so that its change process can revolve around the node's own operating state. Subsequently, the above two types of transition base points are written segment by segment in the order of the offset segment sequence, so that each offset segment has a clear starting reference position and maintains a connection between adjacent offset segments. That is, the transition base point of the subsequent offset segment can take over the termination state of the previous offset segment when necessary, thus forming a continuous and traceable transition starting point sequence.

[0099] Step S1053: Starting from the turning base point, the voltage upward offset segment and voltage downward offset segment of each node are segmented and turned separately, so that the offset segments with different sources and different orders in the same node are converted into turning segments that are continuously unfolded along the unified voltage coordinate. The turning segments are spliced ​​together according to the segment order to form a continuous voltage offset trajectory band corresponding to each node.

[0100] In this embodiment, for each offset segment, its corresponding turning point is used as the starting reference. The voltage-up offset segment is expanded in the direction of gradual voltage increase, and the voltage-down offset segment is expanded in the direction of gradual voltage decrease. This transforms the offset segments originally dispersed from different disturbance sources and different stages into continuous and comparable turning segments under the same voltage coordinate system. During the expansion process, for offset segments with different sources but adjacent in the sequence, the termination state of the previous turning segment is used as the starting reference for the next turning segment, so that the voltage changes caused by different disturbances can be continuously connected on the same trajectory. For offset segments that are discontinuous in the original sequence, their independence is maintained. While recording the interval relationship between the current segment and the preceding segment, path distortion caused by artificial splicing is avoided. Subsequently, according to the segment order sorting results formed in step S1051, the beginning and end of each transition segment are spliced ​​together, and the voltage upward offset segment and voltage downward offset segment are arranged alternately according to their actual occurrence order, so that the rise and fall of the node voltage in each disturbance stage can be represented in the form of a continuous trajectory. Furthermore, during the splicing process, multiple transition segments formed by parallel disturbances in the same stage are merged in parallel, so that they are represented as multiple change branches in the same stage in the trajectory band. Finally, all transition segments are integrated into a continuously unfolded voltage change trajectory band under a unified voltage coordinate.

[0101] Step S1054: The continuous voltage offset trajectory of each node is superimposed and verified with the allowable operating range of that node segment by segment. Turning segments within the allowable operating range, turning segments overlapping with the allowable operating range, and turning segments exceeding the allowable operating range are identified. The boundary turning point is taken as the boundary turning point to form the voltage safety operating boundary corresponding to different nodes.

[0102] In this embodiment, the allowable operating range of voltage for the corresponding node under the current operating mode is obtained, and this allowable operating range is mapped to a voltage coordinate system consistent with the voltage offset trajectory band. Subsequently, each turning segment is scanned segment by segment along the unfolding order of the trajectory band. The start and end positions of each turning segment are compared with the upper and lower boundaries of the allowable operating range to determine whether the turning segment is entirely within the range, partially crosses the range boundary, or is completely outside the range. Turning segments in different states are classified and labeled. During the identification process, for turning segments that overlap with the allowable operating range, their entry point is located by tracing the continuous change path of the turning segment in the trajectory band. The boundaries of the permitted operating range and the specific locations outside these boundaries are designated as key inflection points during voltage changes, representing the transition from a safe zone to a boundary state or from a boundary state to a cross-boundary state. Furthermore, these key inflection points are connected in series according to the arrangement of the transition segments within the trajectory band. Combined with the stable sections formed by the transition segments located within the permitted operating range, the voltage change process is divided into intervals, allowing for segmented representation of the node voltage's operating state at different stages. Finally, using the key inflection points as boundary markers, the trajectory intervals corresponding to each transition segment are reorganized to form a segmented structure that characterizes the node voltage's approach to, contact with, and crossing of the boundary under disturbance.

[0103] like Figure 2 As shown, the distribution network uses the busbar as the upstream power supply node, and multiple feeders are connected to the busbar, such as feeder 1, feeder 2, and feeder 3. Each feeder has a new energy node and a load node. New energy nodes can be distributed photovoltaic, wind power, or energy storage grid-connected nodes, while load nodes can be residential loads, industrial loads, or charging loads. Figure 2 The right side shows the process of node voltage offset changing over time. Based on the allowable offset range of node voltage, an upper boundary and a lower boundary of voltage offset are preset. The node voltage offset interval is formed between the two boundaries. When the node voltage curve fluctuates within this interval, it indicates that the voltage offset is still within an acceptable range. When the voltage curve approaches the upper boundary, it indicates that the node has an overvoltage risk. When the voltage curve approaches the lower boundary, it indicates that the node has an undervoltage risk.

[0104] The voltage shift process can also be divided into multiple shift segments, such as shift segment 1, shift segment 2 and shift segment 3. Each shift segment corresponds to a relatively stable voltage change stage, such as a voltage continuous rise stage, a voltage slow fall stage or a voltage oscillation stage. By identifying the start and end time, shift direction, shift amplitude and duration of different shift segments, it is possible to determine whether the current voltage disturbance is a short-term fluctuation, a continuous shift or a complex disturbance formed by multiple sources alternating.

[0105] Using hybrid coupled node N3 as the main verification node, a linearized voltage sensitivity relationship for node N3 is established based on the target power grid topology and line parameters: ΔVN3=0.050ΔPG+0.025ΔQG-0.040ΔPL-0.025ΔQL; where an increase in active and reactive power on the source side will raise the voltage of node N3, while an increase in active and reactive power on the load side will lower the voltage of node N3; when the load on the load side decreases, ΔPL and ΔQL are negative, thus further raising the node voltage. According to this sensitivity relationship, the source-load disturbance intervals of each time period are mapped to the voltage offset intervals of node N3, and the results are shown in Table 2.

[0106] Table 2 Voltage Deviation Range and Safety Status of Node N3 at Different Time Periods

[0107] T1 1.0130 1.0200 Safety T2 1.0368 1.0505 Entering the upper boundary adjacent segment and slightly exceeding the upper limit T3 1.0500 1.0660 Clearly exceeding the upper limit T4 1.0214 1.0294 Safety

[0108] Therefore, it can be concluded that there is a slight overvoltage risk at node N3 during period T2, and a significant overvoltage risk at node N3 during period T3.

[0109] Assume the allowable operating range of node voltage is 0.95 pu to 1.05 pu, and divide the range from 1.045 pu to 1.05 pu into the upper boundary adjacent segment. Synchronous verification is performed on nodes N1, N2, and N3, and the upper bounds of the node voltages under no voltage regulation state are shown in Table 3:

[0110] Table 3. Upper voltage limits and major risk nodes of each node under no voltage regulation condition.

[0111] T1 1.0228 1.0088 1.0200 none T2 1.0480 1.0268 1.0505 N3 T3 1.0591 1.0393 1.0660 N1, N3 T4 1.0311 1.0133 1.0294 none

[0112] It can be seen that during the T2 period, the voltage of node N3 enters the upper boundary adjacent segment and slightly exceeds the limit. During the T3 period, both nodes N1 and N3 exceed the voltage upper limit. Therefore, the voltage trajectory segments corresponding to T2 and T3 are marked as boundary approach segments, with T3 being the key limit-exceeding segment.

[0113] Step S2: Establish an interval robust optimization model with the joint optimization objectives of minimizing the risk of voltage exceeding the limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation actions.

[0114] The specific steps of step S2 are as follows:

[0115] Step S201: The joint optimization objective of minimizing the risk of voltage exceeding the limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation actions is broken down into risk control sub-objectives, cost constraint sub-objectives, and action sequence sub-objectives. All sub-objectives are then mapped to voltage regulation tasks on the source side, grid side, and load side, forming a voltage regulation task set.

[0116] In this embodiment, the joint optimization objective is functionally decomposed based on the potential voltage out-of-bounds states of nodes, the costs associated with the investment of voltage regulation resources, and the continuity requirements of voltage regulation actions in the time dimension. Situations involving voltage approaching or exceeding safe operating boundaries are categorized as risk control sub-objectives; resource consumption and operational constraints related to the activation, switching, and maintenance of different voltage regulation resources are categorized as cost constraint sub-objectives; and situations involving abrupt changes, overlaps, or disordered switching during the sequential advancement of voltage regulation actions are categorized as action sequencing sub-objectives. After completing the objective decomposition, the above sub-objectives are further spatially mapped by combining the topological hierarchical position of each node in the power grid and the type of voltage regulation resources it is connected to. This mapping categorizes the objectives related to distributed renewable energy output regulation. The target content is mapped to the source-side voltage regulation task, the target content related to the regulation of network equipment such as on-load tap changers and reactive power compensation devices is mapped to the grid-side voltage regulation task, and the target content related to load response and demand regulation is mapped to the load-side voltage regulation task. During the mapping process, each sub-target is further subdivided according to the range of influence of different voltage regulation resources on node voltage, response path, and correspondence with the node association set, so that each sub-target corresponds to a set of voltage regulation task units with clear action nodes and action paths. Furthermore, the above-mentioned source-side, grid-side, and load-side voltage regulation task units are integrated according to the node association relationship and voltage transmission path, and voltage regulation tasks that act on the same node or are transmitted along the same electrical path are merged to form a structured voltage regulation task set.

[0117] It should be noted that the specific formula for the joint optimization objective is as follows: ,in, Indicates the joint optimization objective. , and These represent the weighting coefficients of the risk control sub-objective, cost constraint sub-objective, and action sequence sub-objective, respectively, for example, 0.4, 0.2, and 0.4, with a sum of 1. These coefficients are adjusted based on the target power grid's operating mode, voltage regulation control objectives, and historical operating results, and remain unchanged throughout the dispatching cycle. , and These represent the normalized baseline values ​​for the risk control sub-objective, the cost constraint sub-objective, and the action sequence sub-objective, respectively. Take the maximum reference value or rated reference value of the voltage over-limit risk indicator from historical operation, or the calculation result under the initial scenario. Take the maximum comprehensive voltage regulation cost, the upper limit of the rated cost, or the initial solution result of the voltage regulation resource within the baseline scheduling period. Take the upper limit of the sum of changes in actions in adjacent time periods, the upper limit of the historical statistical average, or the smoothing index value under the benchmark scenario. All of the above benchmark values ​​are positive and are used to eliminate the differences in the dimensions of different sub-objectives.

[0118] , and These represent the risk control sub-objective value, the cost constraint sub-objective value, and the action sequence sub-objective value, respectively. Where t represents the time period index, T represents the voltage regulation time period set, v represents the node index, and N represents the set of nodes participating in the optimization. This represents the voltage value of node v during time period t, obtained from real-time measurements and state estimates. This represents the upper limit of the voltage at node v. This represents the lower voltage limit of node v, which is preset based on the allowable operating range of the node voltage and can be set separately according to different node types. This indicates the deviation above the upper limit when the node voltage exceeds the upper limit. This indicates the deviation from the lower limit when the node voltage is below the lower limit; , where k represents the voltage regulation resource index, and K represents the voltage regulation resource set. This represents the adjustment amount of the voltage regulating resource k during time period t. It is the actual response amount of the voltage deviation, i.e., the value taken is the amount adjusted. The coefficient representing the unit regulation cost of voltage regulation resource k can include power loss, equipment loss, cost of power curtailment, and user-side response compensation. This coefficient can be determined by equipment operating parameters, maintenance experience values, dispatch strategy requirements, or a preset cost table. This indicates the adjustment flag of the voltage regulating resource k during time period t. It can be 0 or 1. When the voltage regulating resource is engaged, disengaged, switches gears, or its adjustment status changes during time period t relative to the previous time period, it indicates the adjustment status of the voltage regulating resource k. A value of 1 indicates that the corresponding voltage regulation resource was activated during that time period. If the resource maintains the same regulation state in adjacent time periods, A value of 0 indicates that the event was not triggered. This represents the single-action cost coefficient of the voltage regulation resource k. The cost may include mechanical wear, control switching losses, communication execution costs, etc. This coefficient can be determined by equipment operating parameters, maintenance experience values, scheduling strategy requirements, or preset cost tables. This indicates the adjustment range of the voltage regulating resource k during the corresponding time period. , This represents the amount of adjustment change of voltage regulation resource k between two adjacent time periods. The smaller this term is, the smoother the voltage regulation action, the fewer the abrupt changes, and the better the continuity. This represents the total number of time periods included in the voltage regulation time period set T.

[0119] Before entering the smoothing sub-target, the adjustment amount of each voltage regulation resource is first converted into a unified equivalent adjustment amount or dimensionless adjustment amplitude.

[0120] Step S202: Embed the new energy output range, load demand range and voltage safety operation boundary into the voltage regulation task set, generate the regulation state sequence of various voltage regulation resources under different disturbance ranges, and construct the corresponding coordinated voltage regulation sequence.

[0121] In this embodiment, the new energy output range and load demand range obtained in step S102 are mapped according to the node association set, so that each voltage regulation task unit corresponds to the power change range of its working node under different disturbance ranges. At the same time, the voltage safety operation boundary formed in step S1054 is superimposed on the task unit corresponding to each node as the basis for determining the triggering and constraint of voltage regulation tasks. Subsequently, based on the temporal progression of the disturbance ranges, each voltage regulation task unit is expanded interval by interval, identifying the types of voltage regulation resources that need to be regulated in different intervals. According to the regulation capability range, adjustable state, and correspondence with node voltage offset of the voltage regulation resources, each voltage regulation resource is assigned a corresponding regulation state identifier in each disturbance range, for example: 0. 1: No participation in regulation; 1: Increase output (positive regulation); -1: Decrease output (reverse regulation), thus forming a state change sequence oriented towards a single voltage regulation resource; After completing the construction of the state sequence of each resource, the regulation states of different resources in the same disturbance interval are horizontally aligned according to the electrical connection relationship between nodes and the transmission path of the voltage regulation task in the network, and vertically connected according to the order of the disturbance intervals, so that the regulation states of different resources at different stages can form a continuous expression on a unified time axis; On this basis, voltage regulation resources with synergistic relationships on the same node or the same path are combined, and their corresponding state sequences are aligned and fused to construct a collaborative voltage regulation sequence that can reflect the collaborative participation of multiple resources in the voltage regulation process in each disturbance interval.

[0122] Step S203: Based on the coordinated voltage regulation sequence, sort each sub-target and merge the voltage constraints and voltage regulation action constraints under the interval disturbance to form a robust constraint set.

[0123] In this embodiment, along the advancement direction of the coordinated voltage regulation sequence, the risk control sub-objective, cost constraint sub-objective, and action sequence sub-objective are sequentially aligned, so that the position of each sub-objective on different disturbance intervals, different nodes, and different voltage regulation resources is correspondingly marked. Among them, the risk control attribute is given priority to the task segment related to the node voltage approaching or touching the safe operation boundary; the cost constraint attribute is given to the task segment involving the continuous input, maintenance, or switching of multiple types of voltage regulation resources; and the action sequence attribute is given to the task segment involving the order of action entry, maintenance continuity, and exit connection. After the target alignment is completed, the above sub-objectives are further sequenced according to the sequential relationship of each voltage regulation action in the coordinated voltage regulation sequence. That is, the preceding target segment that has a direct impact on the node voltage boundary state is first determined, and then the cost verification segment and action connection segment that cooperate with the preceding target segment are sequentially attached to it. This organizes the target requirements of different natures within the same disturbance interval into a target sequence with sequential logic, and makes adjacent disturbance intervals... The transition relationship between objectives is maintained continuously. Subsequently, the node voltage constraints, resource adjustment range constraints, action switching connection conditions, and compatibility conditions for multiple concurrent resources corresponding to each objective sequence are extracted and merged segment by segment according to the disturbance interval. Among them, the node voltage constraints are used to define the allowable offset boundary of each node under different disturbance stages, the resource adjustment range constraints are used to define the accessible and maintainable range of various voltage regulation resources in the corresponding task segment, the action switching connection conditions are used to define the succession order between the previous and subsequent voltage regulation actions, and the compatibility conditions for multiple concurrent resources are used to define the combination boundary when different resources participate at the same time. Furthermore, the above-mentioned constraints are encapsulated in the manner of "prioritizing the merging of the same node, continuously merging the same path, and merging across intervals", so that each type of constraint does not only correspond to a single moment or a single interval, but can cover the continuous process of node voltage change and voltage regulation action change during the advancement of the disturbance interval, and finally form a robust constraint set for the entire process of interval disturbance.

[0124] Step S204: Based on the voltage regulation task set, the collaborative voltage regulation sequence, and the robust constraint set, construct an interval robust optimization model to generate a collaborative voltage regulation strategy that satisfies the voltage regulation constraints under different disturbance intervals.

[0125] The specific steps of step S204 are as follows:

[0126] Step S2041: Based on the voltage regulation task set and the coordinated voltage regulation sequence, the regulation actions corresponding to various voltage regulation resources are serialized according to the triggering order to form a voltage regulation variable sequence containing source-side regulation variables, grid-side regulation variables and load-side regulation variables.

[0127] In this embodiment, according to the correspondence between the node association set and the voltage regulation task unit, various resources participating in voltage regulation are assigned tasks. Distributed new energy inverter regulation and energy storage power regulation are classified as source-side regulation actions; on-load tap-changing transformer tap-changing, capacitor bank switching, and reactive power compensation device switching are classified as grid-side regulation actions; and interruptible loads, transferable loads, and flexible load responses are classified as load-side regulation actions. The position of each type of regulation action in the coordinated voltage regulation sequence is marked as the entry, maintenance, and exit stages. Subsequently, based on the triggering order of each regulation action in the coordinated voltage regulation sequence, the regulation actions within the same node, the same path, and the same disturbance range are sequentially sorted out. That is, the preceding actions that directly respond to the voltage offset of the node are first identified, and then the actions that inherit the preceding actions are identified. The subsequent actions are identified to supplement boundary margins or complete phase switching, thus organizing the originally scattered adjustment behaviors across different resources into action chains that unfold sequentially along the disturbance progression. After completing the action chain organization, each adjustment action is further abstracted into variable expression units that can participate in subsequent optimization solutions. Source-side adjustment actions correspond to source-side adjustment variables, grid-side adjustment actions correspond to grid-side adjustment variables, and load-side adjustment actions correspond to load-side adjustment variables. Each variable retains information about its node, corresponding disturbance range, triggering stage, and preceding and following relationships. Finally, according to the overall progression order of the coordinated voltage regulation sequence, the above three types of adjustment variables are connected and organized in a unified order to form a voltage regulation variable sequence containing source-side adjustment variables, grid-side adjustment variables, and load-side adjustment variables.

[0128] To address the aforementioned voltage over-limit risks, voltage regulation resources are configured on the source side, grid side, and load side as shown in Table 4:

[0129] Table 4 Voltage Regulation Resource Parameters and Voltage Regulation Characteristics

[0130] K1 Source-side resources Photovoltaic inverter reactive power absorption 0 to 0.20 Mvar Absorbing 0.10 Mvar reduces the N3 voltage by approximately 0.005 pu. Fast response, suitable for advance adjustment K2 Network-side resources On-load tap changer 0 gear, -1 gear, -2 gear Each reduction of one level lowers the downstream voltage by approximately 0.010 pu. The adjustment range is significant, but the movements should not be too frequent. K3 Network-side resources SVG reactive power compensation device 0 to 0.15 Mvar Absorbing 0.10 Mvar reduces the N3 voltage by approximately 0.004 pu. It has a fast response time and is suitable for compensation adjustment. K4 Load side resources Flexible load increase 0 to 0.15MW An increase of 0.10MW in load causes the N3 voltage to decrease by approximately 0.004 pu. It can be used as a backup adjustment.

[0131] Therefore, K1 is represented as the source-side regulation variable, K2 and K3 as the grid-side regulation variables, and K4 as the load-side regulation variable. A voltage regulation variable sequence of K1 priority, K2 take over, K3 compensate, and K4 standby is formed according to the response speed and regulation characteristics.

[0132] Step S2042: Expand the new energy output range and the load demand range according to the disturbance entry sequence to form multiple consecutive range disturbance segments, and match the range disturbance segments with the voltage regulation variable sequence segment by segment to construct the connection relationship between the disturbance segments and the voltage regulation action.

[0133] In this embodiment, the new energy output range and load demand range are simultaneously expanded along the time axis. Based on the order in which disturbances enter each node and their propagation order within the node association set, the original range is divided into several interval disturbance segments with continuous succession relationships. Each disturbance segment corresponds to a relatively stable disturbance combination state, and the transition relationship between adjacent segments is preserved. Subsequently, for each interval disturbance segment, combined with the source-load coupling disturbance sequence and its path calibration results formed in steps S103 and S104, the range of influence of the disturbance segment on each node and its dominant influence stage on voltage offset are determined, thereby matching the corresponding voltage regulation requirement for the disturbance segment. After completing the disturbance segment division, voltage regulation is implemented. The variable sequence is segmented and mapped to its respective node according to its triggering order. For each disturbance segment, voltage regulation variables that meet the entry conditions at this stage are selected, and according to the sequential relationship of the variables in the sequence, they are aligned to the beginning, continuation, or transition position of the corresponding disturbance segment. Furthermore, the voltage regulation variables between adjacent disturbance segments are connected, so that the voltage regulation variables that are in a maintenance state in the previous disturbance segment can continue or exit in the next disturbance segment, while the newly entering voltage regulation variables complete the connection at the segment boundary, thereby forming a continuous and traceable correspondence between disturbance segments and voltage regulation actions. Finally, a set of structured connection results is obtained, arranged in the order of disturbance segments, with each segment associated with a corresponding voltage regulation variable subsequence.

[0134] Step S2043: Embed the robust constraint set into the connection relationship between the disturbance segment and the voltage regulation action, and encapsulate the continuity constraint of the voltage regulation action, node voltage constraint and regulation conflict constraint between different disturbance segments across segments to form a unified constraint system covering multiple disturbance segments.

[0135] In this embodiment, according to the sequence of disturbance segments, the matched voltage regulation variables and their corresponding nodes within each segment are constrained and aligned. The node voltage boundary constraints are mapped to the voltage offset stages of the corresponding nodes in each segment. Simultaneously, the entry range, sustainable state, and exit conditions of the voltage regulation variables are embedded into the voltage regulation action positions within the corresponding segments, thus clearly expressing the constraint relationships within a single disturbance segment. Subsequently, regarding the connection relationship between adjacent disturbance segments, voltage regulation variables already in an active or sustainable state in the previous segment are marked with cross-segment continuation identifiers, and connection conditions are established for newly entering or exiting voltage regulation variables in the subsequent segment. This ensures that the voltage regulation action meets the constraint requirements for continuous transition at the segment boundaries, thereby avoiding action breaks or repeated triggering between segments. Based on this, ... The overlapping effects that different types of voltage regulation resources may produce within the same disturbance segment and across segments are uniformly analyzed. The simultaneous action relationships between voltage regulation variables on the source side, grid side, and load side at the same node or along the same path are conflict identified, and the conflict situations are classified and marked according to their action paths and succession relationships, thereby forming corresponding regulation conflict constraints. Furthermore, the above-mentioned node voltage constraints, voltage regulation action continuity constraints, and regulation conflict constraints are merged and encapsulated in a manner that prioritizes the integration of constraints within segments, sequentially inherits constraints from adjacent segments, and uniformly covers constraints across multiple segments, so that various constraints form a unified expression structure in multiple continuous disturbance segments. Finally, a unified constraint system is obtained that can cover the entire evolution process of the disturbance interval and simultaneously reflect the evolution relationship between node voltage limits and voltage regulation actions.

[0136] Step S2044: Based on the voltage regulation variable sequence, interval disturbance segments, and unified constraint system, the voltage regulation process corresponding to each disturbance segment is connected as a whole to form an interval robust optimization model for continuous changes in interval disturbances.

[0137] In this embodiment, following the temporal progression of the disturbance segments, the established voltage regulation variables and their corresponding constraints within each segment are locally combined, so that each disturbance segment corresponds to a segment-level voltage regulation process expression containing variable states, node constraints, and action succession conditions. Subsequently, based on the established voltage regulation action succession relationships between adjacent disturbance segments, the variable termination state in the previous segment is aligned with the variable start state in the next segment, ensuring that each segment-level voltage regulation process forms a continuous transmission relationship at the variable level and guaranteeing that the voltage regulation actions across segments are not broken or redefined. After completing the variable connection between segments, the cross-segment constraints in the unified constraint system are further embedded one by one into the above continuous structure, enabling node voltage limits, The requirement for continuous voltage regulation and the conflict constraints among multiple resources can be applied simultaneously across multiple disturbance segments, thus transforming the originally scattered local constraints within each segment into a holistic constraint expression covering the entire process. Based on this, the voltage regulation processes of all disturbance segments are interconnected according to their order of arrangement within the disturbance interval, forming a voltage regulation decision structure that spans the entire evolution of the interval disturbance. In this structure, the state changes, entry and exit relationships, and constraint ranges of each voltage regulation variable in different segments are consistently expressed. Finally, through the aforementioned processes of variable continuity, segment concatenation, and unified constraint embedding, the correspondence between discrete disturbance intervals and voltage regulation actions is integrated into a unified whole, resulting in an interval robust optimization model oriented towards continuous changes in interval disturbances.

[0138] It should be noted that the interval robust optimization model is constructed using a structured modeling approach, comprising a data input module, a feature construction module, an interval modeling module, a scenario generation module, a robust decision-making module, and a strategy output module. These modules are sequentially connected via data flow to form a unified decision-making chain. Specifically, the data input module acquires the power grid topology, node parameters, renewable energy output data, and load demand data; the feature construction module extracts node power change characteristics and topological correlation characteristics; the interval modeling module transforms renewable energy output and load demand into interval representations; the scenario generation module generates multiple candidate disturbance scenarios based on the interval data; the robust decision-making module identifies the most unfavorable scenario among the candidate scenarios and performs voltage regulation optimization; and the strategy output module generates control strategies for various voltage regulation resources.

[0139] In terms of input-output relationship, the model takes power grid structure data, output and load data and voltage regulation resource parameters as input. After interval construction and scenario expansion, it forms intermediate data such as node disturbance interval, disturbance propagation path and candidate scenario set, and finally outputs the most unfavorable scenario identification result and the corresponding voltage regulation control strategy, including the entry order, holding interval and exit arrangement of voltage regulation resources, thus forming a complete data processing and decision-making closed loop.

[0140] In the process of model construction and application, statistical analysis of the output of new energy sources and load fluctuation characteristics is carried out based on historical operating data to establish an initial range. The range and feature weights are continuously updated through actual operating data. At the same time, the node topology, voltage regulation resource constraints and voltage operation boundaries in the model are all derived from actual power grid operating conditions, so that the model output results can be directly applied to the source-grid-load coordinated voltage regulation scenario.

[0141] The interval robust optimization model established in this invention is as follows: Figure 3 As shown, with the objective function, constraints, and uncertainty interval as core elements, it can generate the optimal solution across the entire range of new energy output and load fluctuations, forming a robust optimization scheduling scheme.

[0142] Step S3: Based on the interval robust optimization model, identify the most unfavorable scenario that causes the node voltage to approach or exceed the safety boundary, and perform collaborative optimization according to the regulation capabilities and response characteristics of the source side, grid side and load side to obtain a voltage regulation control strategy that meets the voltage constraints.

[0143] The specific steps of step S3 are as follows:

[0144] Step S301: Based on the interval robust optimization model, expand the new energy output interval, load demand interval and voltage regulation variable sequence to form multiple candidate disturbance scenarios corresponding to the node voltage change process, and collect the candidate disturbance scenarios to form a scenario set corresponding to different nodes.

[0145] In this embodiment, the renewable energy output range and load demand range are combined and expanded according to their distribution position and time progression in the node association set. This allows the power change states corresponding to different nodes at different disturbance stages to be listed one by one, and matched with the trigger positions of each variable in the voltage regulation variable sequence. This forms a multi-dimensional correspondence including "disturbance state - voltage regulation response - node action" under a unified modeling framework. Subsequently, using nodes as analysis units, the above combined expansion results are projected node by node. That is, the voltage change path of each node under different disturbance combinations and voltage regulation variable participation conditions is extracted. The voltage evolution process that the same node may experience under different range combinations is expanded one by one, forming several corresponding to different disturbance stages. Candidate disturbance scenarios are generated based on their input sequence and voltage regulation action participation method. After generating candidate scenarios, scenarios originating from different nodes but having coupling relationships on the electrical path are further associated and merged based on the topological relationships between nodes and the disturbance propagation path, so that scenarios that form continuous effects on the same disturbance propagation chain are integrated into a unified expression. At the same time, scenarios that only have an impact within the local node range are independently retained to maintain the integrity of the scenario set. Finally, the above candidate disturbance scenarios are classified and summarized according to the node identifier, and all candidate disturbance scenarios related to the voltage change of the same node are grouped into the scenario set corresponding to that node, so that each node corresponds to a set of voltage evolution scenarios covering different disturbance combinations and voltage regulation response conditions.

[0146] Step S302: Compare each candidate disturbance scenario in the scenario set with the voltage safety operation boundary of the corresponding node one by one, and select the disturbance scenarios that cause the node voltage to enter the boundary adjacent segment, cross the boundary turning position, or continuously wander along the boundary as boundary approach scenarios.

[0147] In this embodiment, for each node, its corresponding candidate disturbance scenarios are unfolded into voltage evolution trajectories one by one according to the disturbance progression order. These trajectories are then aligned with the node's voltage safety operating boundary in the same voltage coordinate system, ensuring that the voltage state of each scenario at each stage corresponds segment-by-segment with the safety boundary. Subsequently, a path scan is performed on each candidate disturbance scenario to identify its relative positional changes with the safety operating boundary during voltage evolution. Stages where the voltage trajectory gradually approaches but has not yet entered the boundary interval are marked as boundary adjacency processes. Stages where the trajectory moves from the safety interval into the boundary interval or from the boundary interval to the boundary are also identified. The outer scenario is marked as a boundary turning process. For cases where the trajectory fluctuates near the boundary interval in multiple consecutive stages without significantly retreating to the safe interval, it is marked as a boundary wandering process. After completing the above marking, the entry order of the disturbance sequence in the scenario and the participation status of the voltage regulation variable are further combined to uniformly screen different types of boundary approach behaviors. Only scenarios that can continuously maintain the above three types of boundary-related states during the disturbance propagation process are retained, while scenarios that only briefly touch the boundary in a single stage and do not form a continuous impact are eliminated. Finally, the screened scenarios are uniformly classified as boundary approach scenarios to cover the entire process of node voltage approaching the safe boundary.

[0148] It should be noted that the voltage safety operating boundary refers to the upper and lower limits of the allowable change of node voltage under disturbance and its segmented transition structure.

[0149] Step S303: Rearrange the voltage offset propagation process in the boundary approach scenario, and determine the disturbance scenario that first triggers voltage constraint conflict or continuously occupies voltage regulation margin under the current cooperative voltage regulation sequence as the most unfavorable scenario.

[0150] In this embodiment, the voltage offset propagation process in each boundary approach scenario is path reconstructed, and the corresponding disturbance entry sequence, offset segment overlap relationship, and voltage regulation variable participation stage are uniformly unfolded. The effect of each disturbance on the node voltage is rearranged according to the time progression, so that the superposition relationship of disturbances from different sources on the same node is presented in the form of a continuous process. Subsequently, in the rearranged propagation process, the voltage regulation variable participation state of each stage is calibrated, the voltage regulation resources occupied in that stage and their remaining adjustable range are identified, and combined with the relative position of the node voltage and the safe operating boundary, it is determined segment by segment whether the voltage approaches or touches the boundary and the voltage regulation resources can no longer bear the load, thereby identifying the state. The key stage of triggering voltage constraint conflict is identified. Based on this, a horizontal comparison is made of each boundary approach scenario. Scenarios that show the above-mentioned constraint conflict in an early stage or that continuously occupy the main voltage regulation resources in multiple consecutive stages without forming an effective release are prioritized for screening. At the same time, scenarios that only show local constraint tension in later stages but do not affect the overall voltage regulation capacity are eliminated. Furthermore, the selected candidate scenarios are checked for consistency by combining the coverage of their disturbance propagation path and the breadth of voltage regulation variable participation to ensure that the selected scenarios are representative at both the node level and the path level. Finally, the disturbance scenario that meets the characteristics of prioritizing the triggering of constraint conflict or continuously occupying voltage regulation margin under the current coordinated voltage regulation sequence is identified as the most unfavorable scenario.

[0151] It should be noted that the boundary approach scenario refers to the disturbance evolution process where the node voltage approaches or touches the safe operating boundary; the determination of the most unfavorable scenario is based on the quantitative evaluation and ranking of candidate disturbance scenarios. Specifically, firstly, the renewable energy output range and load demand range are combined and expanded to form multiple candidate disturbance scenarios, each scenario corresponding to a specific disturbance entry sequence and power change path; then, each candidate disturbance scenario is expanded into a voltage change process at the node level and compared segment by segment with the node's safe operating voltage boundary; during the evaluation process, each candidate scenario is comprehensively evaluated from the following three aspects: firstly, the degree of voltage exceeding the limit, i.e., the magnitude of the node voltage deviation from the allowable operating range; secondly, the voltage regulation resource occupancy, i.e., the number of voltage regulation resources to be called up in this scenario and their usage intensity; and thirdly, the continuity of the voltage regulation process, i.e., whether the voltage regulation resources are continuously in a high load state in multiple stages; based on the above evaluation indicators, all candidate disturbance scenarios are ranked, and the scenario that causes the voltage to approach or exceed the boundary in an earlier stage and simultaneously occupies more voltage regulation resources is selected as the most unfavorable scenario.

[0152] After ranking the candidate disturbance scenarios from T1 to T4, it was found that during the T3 period, when the source-side output is taken as the upper limit and the load-side load as the lower limit, the node voltage exceedance is the largest, and at the same time, it occupies more voltage regulation resources. Therefore, this scenario is determined as the most unfavorable scenario, which can be expressed as: ΔPG = 0.72MW, ΔQG = 0.12Mvar, ΔPL = -0.18MW, ΔQL = -0.07Mvar. At this time, the unregulated voltage of node N3 is:

[0153] VN3 = 1.018 + 0.050 × 0.72 + 0.025 × 0.12 - 0.040 × (-0.18) - 0.025 × (-0.07) = 1.06595 pu, which is approximately 1.0660 pu. Since 1.0660 pu exceeds the voltage limit of 1.05 pu, this scenario belongs to the most unfavorable scenario that prioritizes the triggering of voltage constraint conflict.

[0154] Step S304: Based on the most unfavorable scenario, according to the adjustment start conditions, action succession relationship and response sequence of source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources, the various voltage regulation resources are hierarchically sorted to form a coordinated voltage regulation process chain including the first adjustment layer, the subsequent adjustment layer and the subsequent adjustment layer.

[0155] The specific steps of step S304 are as follows:

[0156] Step S3041: Based on the most unfavorable scenario, perform scenario matching for source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources respectively, identify the voltage trigger state, power offset state and adjustable margin state corresponding to each type of voltage regulation resource, and determine the set of adjustment start conditions for each type of voltage regulation resource accordingly.

[0157] In this embodiment, the voltage offset propagation process in the most unfavorable scenario is aligned with the node association set. The voltage state, power injection state, and disturbance source of each node at different stages are unfolded segment by segment. Source-side voltage regulation resources, grid-side voltage regulation resources, and load-side voltage regulation resources are mapped to the corresponding scenario stages according to their access nodes and action paths. Subsequently, the participation conditions of each type of voltage regulation resource in the most unfavorable scenario are analyzed segment by segment. For source-side voltage regulation resources, the response segments caused by output changes during the node voltage offset process are identified, and the active or reactive power regulation directions and available adjustment ranges corresponding to them at different stages are marked. For grid-side voltage regulation resources, their position in the voltage transmission path is identified, and the intervention point and adjustable tap or compensation state that affect the node voltage in different voltage offset stages are determined. For load-side voltage regulation resources, their role in each stage is determined based on the impact process of load changes on the node voltage. The system includes triggering sections and response ranges for regulation. After completing the above-mentioned categorized matching, the system further combines the order of participation of each resource in the most unfavorable scenario and the order of its effect on voltage offset to summarize the state of each resource at each stage. The system extracts the triggering conditions, power offset performance, and remaining adjustable space when the voltage is close to the boundary, in the boundary adjacency state, and crossing the boundary. For example, if the upper limit of the voltage of the bus node is 1.05 and the lower limit is 0.95, in the most unfavorable scenario, for the voltage regulation resource of the bus node, the grid-side transformer intervenes when the voltage is continuously close to the upper limit of 1.05. Its triggering condition is that the voltage is higher than 1.049, the tap position is adjusted from 0 to -1, the power offset performance is a reduction in voltage amplitude of 0.01, and the remaining adjustable space is 1. Finally, the above information is summarized by resource category to form a set of regulation starting conditions that includes voltage triggering state, power offset state, and adjustable margin state.

[0158] Step S3042: Based on the set of adjustment initiation conditions, and following the rules of priority for those that arrive at the adjustment initiation conditions first, adjacent for those that form continuous effects on the same node, and subsequent for those that replace or continue the previous adjustment action, the adjustment actions of various voltage regulation resources are arranged to form an action succession sequence across the source side, grid side and load side.

[0159] In this embodiment, following the disturbance progression sequence in the most unfavorable scenario, the source-side voltage regulation resources, grid-side voltage regulation resources, and load-side voltage regulation resources are sequentially aligned to identify the order in which each resource reaches its regulation initiation condition. The regulation action that first meets the initiation condition and can directly affect the voltage offset process of the current target node is identified as the preliminary candidate action. Subsequently, based on the identified preliminary candidate action, other regulation actions that continue to affect the same node, the same path, or the same voltage offset segment in their subsequent stages are further searched. Regulation actions that can maintain voltage regulation continuity after the preliminary candidate action and form a direct connection relationship in the node's range of action are arranged as adjacent actions, so that the regulation chain on the same node remains continuously unfolding. On this basis, regulation actions not included in the aforementioned continuous action chain are further... Functional discrimination is performed. For adjustment actions that can take over the function when the adjustment margin of the preceding adjustment action weakens, exits, or is no longer suitable for the current offset stage, or for adjustment actions that can continue to expand the adjustment path and supplement the adjustment range based on the preceding action, they are uniformly arranged after the preceding action, thus forming a substitute or successor type of subsequent action. Furthermore, the above-mentioned preceding candidate actions, adjacent continuous actions, and subsequent substitute or successor actions are uniformly rearranged according to the cross-action relationship between the source side, grid side, and load side. For actions that have parallel entry conditions within the same stage, the adjustment actions that are more closely coupled with the current voltage offset propagation path are preferentially retained, and the remaining actions are embedded into the subsequent positions in sequence according to the succession relationship, finally forming an action succession sequence that unfolds segment by segment along the disturbance propagation process and spans the source side, grid side, and load side.

[0160] Step S3043: Fold each adjustment action in the action succession sequence according to the response time sequence, group the adjustment actions with overlapping responses into parallel segments, divide the adjustment actions with staggered responses into front and back segments, and divide the action succession sequence into the preceding adjustment layer, the following adjustment layer, and the subsequent adjustment layer accordingly.

[0161] In this embodiment, the triggering time, duration, and exit time of each adjustment action in the most unfavorable scenario are uniformly calibrated along the action sequence, mapping the response intervals of different voltage regulation resources in the time dimension to the same time coordinate. Subsequently, the above adjustment actions are compared one by one, and adjustment actions with overlapping time relationships are identified, grouped into the same response segment, and formed into parallel adjustment units, so that the parallel action of multiple voltage regulation resources in the same stage can be uniformly expressed. For adjustment actions that are staggered in time, they are segmented according to their triggering and ending order, so that the action sequence remains clear and continuous on the time axis. After completing the parallel segmentation and segmentation, the voltage offset propagation process of each adjustment action is further combined. The system divides each segment into hierarchical levels based on the position of the action and its relationship with the preceding and following actions. The action that enters first in the initial stage of voltage offset and guides subsequent adjustments is assigned to the preceding adjustment layer. The actions that follow the preceding adjustment layer and maintain the continuous change of voltage state are assigned to the continuation adjustment layer. Actions used in later stages to supplement the adjustment range or complete adjustment convergence are assigned to the subsequent adjustment layer. Finally, the adjustment actions within each layer are organized according to their relative positions in the original action sequence, ensuring that actions within the same layer maintain their original succession logic, while different layers form a hierarchical progression relationship according to the time progression. This transforms the original linear action sequence into a hierarchical expression of adjustment actions.

[0162] Step S3044: Connect the preliminary adjustment layer, the subsequent adjustment layer and the subsequent adjustment layer in series according to the disturbance advancement order in the most unfavorable scenario, and unify the entry order, maintenance order and exit order of the adjustment actions in each layer to form a coordinated voltage regulation process chain corresponding to the most unfavorable scenario.

[0163] In this embodiment, taking the time progression path of the disturbance in the most unfavorable scenario as the main line, the effective ranges of the three types of regulation layers in different stages are aligned and calibrated. The regulation actions corresponding to each layer are mapped to the specific stages in which they participate in voltage offset regulation, so that the regulation behavior of different levels and the disturbance evolution process are in one-to-one correspondence. Subsequently, for the regulation actions within each regulation layer, the entry order of each regulation action is uniformly organized according to its position in the original action succession sequence and the triggering order in the corresponding stage, so that the regulation actions that respond first to the current voltage offset trend are placed in the first position, and the regulation actions that have a continuous effect on the same node or the same path are kept adjacent. After completing the entry order organization, the maintenance process of each regulation action in its respective stage is further uniformly calibrated, and those that continue to act in the same stage and have not been replaced are included. The adjustment actions of the relationship are grouped into the same maintenance section, and the actions with successive relationships are set to be sequentially connected, so that different voltage regulation resources can form a continuous adjustment chain during the maintenance phase. Subsequently, combined with the transition position of the disturbance phase, the exit time of each adjustment action is uniformly organized. Resources that no longer meet the adjustment conditions at the end of the current phase or need to be replaced by subsequent adjustment actions are arranged as exit actions, and their exit order is connected with the entry order of the subsequent level. Finally, the three types of adjustment layers are connected in series according to the order of disturbance advancement, so that the first adjustment layer, the subsequent adjustment layer and the subsequent adjustment layer unfold sequentially on the time axis, and the layers are seamlessly connected through the entry and exit relationships, thus forming a collaborative voltage regulation process chain that runs through the entire process of the most unfavorable scenario and can reflect the order of participation and the relationship of action of multiple types of voltage regulation resources.

[0164] It should be noted that the coordinated voltage regulation process chain is represented and managed in a structured sequence manner. Specifically, the adjustment actions of each voltage regulation resource are arranged in chronological order, and each adjustment action is assigned a corresponding node identifier, adjustment type, entry time, duration, and exit condition, thereby forming an adjustment action recording unit containing multi-dimensional information. Subsequently, multiple adjustment actions are connected in series according to the disturbance advancement sequence to form a complete voltage regulation process chain.

[0165] Based on the response speed and connection relationship of each voltage regulation resource, a coordinated voltage regulation process chain is formed as shown in Table 5:

[0166] Table 5. Synergistic Voltage Regulation Process Chain and Voltage Regulation Effect

[0167] Pre-adjustment layer T2 N3 enters the upper boundary adjacent segment Source-side pre-regulation K1 absorbs 0.10 Mvar of reactive power. N3 decreased from 1.0505 to 1.0455 pu. Continuation adjustment layer T3 N1 and N3 clearly exceed the upper limit. Source-side + network-side coordination K1 is increased to 0.20 Mvar, and K2 is adjusted to -1 level. The N3's power consumption decreased from 1.0660 to 1.0460 PU, and the N1's power consumption decreased from 1.0591 to 1.0411 PU. Subsequent adjustment layer T3 Continuation Phase Voltage maintained within a safe range Maintain within the segment K1 remains at 0.20 Mvar, K2 remains at -1 gear. Node voltage stabilizes within safe boundaries

[0168] Therefore, the voltage regulation process chain unfolds in the order of rapid regulation at the source side, regulation at the grid side, and subsequent regulation.

[0169] Step S305: Embed the collaborative voltage regulation process chain into the interval robust optimization model corresponding to the most unfavorable scenario, and jointly calculate the entry order, holding interval and exit order of each voltage regulation resource in different regulation segments to generate a voltage regulation control strategy that satisfies the node voltage constraint.

[0170] The specific steps of step S305 are as follows:

[0171] Step S3051: Embed the coordinated voltage regulation process chain into the corresponding interval robust optimization model according to the disturbance advancement order in the most unfavorable scenario, and align each regulation layer in the coordinated voltage regulation process chain to the different voltage offset stages corresponding to the most unfavorable scenario, forming a correspondence between regulation layers and scenario stages.

[0172] In this embodiment, following the progression of the disturbance in the most unfavorable scenario, the corresponding interval disturbance segment is unfolded segment by segment in the model, and the entire coordinated voltage regulation process chain is embedded into this unfolded structure, so that the adjustment actions of each adjustment layer can form a unified expression with the corresponding disturbance segment in the model. Subsequently, for the preceding adjustment layer, the subsequent adjustment layer, and the subsequent adjustment layer in the coordinated voltage regulation process chain, their respective active segments in the process chain are extracted, and they are aligned segment by segment with different stages of node voltage offset in the most unfavorable scenario. This ensures that the preceding adjustment layer corresponds to the initial stage of voltage offset and the boundary approach stage, the subsequent adjustment layer corresponds to the continuous voltage offset and the boundary adjacency stage, and the subsequent adjustment layer corresponds to the voltage approach stage. The adjustment phase after extreme values ​​or boundary crossings establishes a stage mapping relationship between the regulation layer and the voltage offset stage. After completing the hierarchical alignment, the regulation actions within each regulation layer are further embedded into the corresponding stage's interval robust optimization model structure according to their entry, maintenance, and exit order in the process chain. This ensures that each disturbance segment in the model contains not only disturbance variables and constraint information but also the corresponding regulation layer action structure. Finally, through the above embedding and alignment processing, a one-to-one structural relationship is formed between the disturbance segments, voltage regulation variables, and regulation layers in the interval robust optimization model, thereby obtaining a unified modeling result that can simultaneously reflect the disturbance evolution process and the hierarchical regulation logic.

[0173] It should be noted that an interval disturbance segment refers to an interval unit that has a relatively stable combination of disturbances over a continuous time period.

[0174] Step S3052: Based on the correspondence between the regulation layer and the scenario stage, the voltage regulation process in the most unfavorable scenario is divided into regulation segments, and the source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources are respectively connected to the corresponding regulation segments to form a set of regulation segment resources that includes the main regulation resources within the segment and the inter-segment support resources.

[0175] In this embodiment, the entire voltage regulation process is divided into segments according to the progression of the voltage offset stage in the most unfavorable scenario. The process from the initial voltage offset to boundary adjacency, from boundary adjacency to the critical state, and from the critical state to subsequent adjustment are defined as several consecutive adjustment segments, so that each adjustment segment corresponds to a relatively stable voltage change stage. Subsequently, for each adjustment segment, according to its corresponding adjustment layer type, the adjustment actions that are active or in the entry state in that stage are extracted from the collaborative voltage regulation process chain. According to the access node and action path of the voltage regulation resources, the source-side voltage regulation resources, grid-side voltage regulation resources, and load-side voltage regulation resources are respectively attached to the corresponding adjustment segments, so that the participation position of each type of resource in different stages is clearly marked. After completing the resource... After the source is connected, the voltage regulation resources in each regulation segment are further classified by role. The resource that intervenes first in this stage and plays a dominant role in the node voltage deviation is identified as the main regulation resource in the segment. The resource that continues the function of the previous regulation segment or provides transition support for the next regulation segment in this stage is identified as the inter-segment receiving resource. Combined with their entry and exit relationships in the process chain, it is ensured that the same resource has continuous connection capability between adjacent regulation segments. Finally, the main regulation resources and receiving resources in each regulation segment are uniformly organized to form a resource set structure divided by regulation segment. This ensures that each regulation segment corresponds to a clear set of resource configurations and their functions, thus obtaining a regulation segment resource set that reflects both the segmented characteristics of the voltage regulation process and the resource receiving logic.

[0176] It should be noted that the coordinated voltage regulation process chain refers to a sequence of coordinated effects of multiple types of voltage regulation resources organized in chronological order.

[0177] Step S3053: Within the interval robust optimization model, for each set of resources in each regulating segment, according to the connection rule of responding first and then taking over within the segment and exiting first and then entering between segments, the entry order, holding interval and exit order of each voltage regulating resource in the corresponding regulating segment are jointly obtained, and the resource switching relationship between adjacent regulating segments is continuously checked to form the resource action sequence corresponding to each regulating segment.

[0178] In this embodiment, the model extracts the corresponding source-side, grid-side, and load-side voltage regulation resources for each regulation segment. Based on the voltage offset stage of that regulation segment, the entry conditions and sustainable sections of each resource are constrained and aligned. Subsequently, within the same regulation segment, according to the rule of "respond first, then take over," the voltage regulation resources are sequentially sorted. Resources that can meet the triggering conditions at the beginning of the stage and directly affect the voltage offset are prioritized as preceding entry resources. Resources that rely on the voltage state formed by the preceding resources or continue to maintain the regulation effect based on it are arranged as subsequent receiving resources, thus forming the entry order of resources within the segment. After determining the entry order, the continuous action range of each resource within the regulation segment is further uniformly defined. Resources that continuously participate in regulation within the same stage and do not undergo replacement are grouped into the same sustaining range. For resources with alternating or successive relationships, segmented holding intervals are set to ensure continuous expression of the adjustment process in the time dimension. Subsequently, based on the boundary positions between adjustment segments, the exit order of each resource is organized, and resources that no longer meet the participation conditions or need to be replaced by resources in the next adjustment segment are arranged as exit resources, and their exit positions are linked to the entry positions of the corresponding resources in the next adjustment segment. On this basis, the continuity of resource switching relationships between adjacent adjustment segments is checked, that is, the connection between the exit resources of the previous adjustment segment and the entry resources of the next adjustment segment is checked one by one in terms of node position, action path and time sequence, to ensure that there are no resource gaps or duplicate actions. Finally, the entry order, holding interval and exit order within each adjustment segment are uniformly organized to form a resource action sequence divided by adjustment segment.

[0179] Step S3054: Merge and organize the resource action sequences corresponding to each regulation segment according to the scenario stages in the most unfavorable scenario to form a voltage regulation control strategy covering the voltage regulation resources on the source side, grid side and load side. The voltage regulation control strategy includes the action entry arrangement, segment maintenance arrangement and stage exit arrangement of each voltage regulation resource.

[0180] In this embodiment, the resource action sequences corresponding to each regulation segment are globally aligned according to the progression order of the voltage offset stages in the most unfavorable scenario. The stage position of each regulation segment in the scenario is uniformly identified, so that different regulation segments form a continuous correspondence in the time dimension. Subsequently, the resource action sequences within each regulation segment are merged and organized. The actions of source-side voltage regulation resources, grid-side voltage regulation resources, and load-side voltage regulation resources within the same stage are integrated according to their entry order in the sequence, so that different resources that work together in the same stage form a unified set of stage actions, and the original succession relationship is not disrupted. After completing the integration within the stage, the action sequences between each regulation segment are further spliced ​​across segments, combining the actions from the previous regulation segment. Resources in the exit state are aligned with their corresponding entry state in the next adjustment phase to ensure continuous switching between different stages and avoid adjustment gaps or redundant adjustments. Based on this, the integrated action sequence is structurally expressed, unifying the entry arrangements of each voltage regulation resource in different stages, their maintenance arrangements within the corresponding stages, and their exit arrangements at the end of each stage, and arranging them according to the scenario stage sequence. Finally, a comprehensive voltage regulation control strategy covering source-side, grid-side, and load-side voltage regulation resources is formed. This strategy, with stage-based advancement as the main line, fully describes the participation sequence and functional relationships of various voltage regulation resources under the most unfavorable scenario, thus obtaining control sequence results that can be directly used for voltage regulation execution.

[0181] The final voltage regulation control strategy is as follows: During period T2, the photovoltaic inverter K1 is first controlled to enter the reactive power absorption state, absorbing 0.10Mvar of reactive power; during period T3, the reactive power absorption of K1 is increased to 0.20Mvar, and the tap of the on-load tap changer K2 is adjusted from 0 to -1; the above adjustment state is maintained for the remaining time of T3; during period T4, when the photovoltaic output is detected to drop and the voltage of node N3 is lower than 1.045pu, K2 is first restored to 0, and then K1 is gradually withdrawn from the reactive power absorption state to avoid voltage reverse fluctuation caused by sudden changes in voltage regulation.

[0182] After implementing the above voltage regulation strategy, the upper bounds of the voltage at each node are shown in Table 6:

[0183] Table 6. Upper bounds of voltage at each node and constraint satisfaction after voltage regulation.

[0184] T1 1.0228 1.0088 1.0200 satisfy T2 1.0440 1.0268 1.0455 satisfy T3 1.0411 1.0293 1.0460 satisfy T4 1.0311 1.0133 1.0294 satisfy

[0185] Step S4: Control various voltage regulation resources according to the voltage regulation control strategy, and update the new energy output range and load demand range in combination with real-time operation data. When the operating status deviates, re-execute the range robust optimization model to carry out continuous voltage regulation control.

[0186] In this embodiment, according to the entry arrangement, intra-segment maintenance arrangement, and phase exit arrangement of each voltage regulation resource in the voltage regulation control strategy, the source-side voltage regulation resources, grid-side voltage regulation resources, and load-side voltage regulation resources are controlled in stages. This ensures that each type of resource participates in the voltage regulation process in a predetermined order at the corresponding node and time period. During execution, the voltage status, power flow direction, and resource response status of each node are continuously collected. Subsequently, the real-time collected operating data is compared with the original new energy output range and load demand range. The offset parts are re-divided, and new fluctuation segments that occur during actual operation are included in the range expression, thereby updating the original range primitives to reflect the actual disturbance range under the current operating state. After completing the range update, the updated data is further... The output and demand ranges are remapped to the node voltage offset ranges, and the relationship between the current voltage state and the safe operating boundary is determined. When the node voltage offset trend is detected to deviate from the disturbance range covered by the original voltage regulation strategy or a new boundary approach state appears, the interval robust optimization model is re-solved. During the re-solution process, the updated interval data, the current resource state, and the execution results of the existing voltage regulation actions are used as inputs to regenerate a voltage regulation control strategy that matches the current operating state, and smoothly connect it with the control process being executed. Finally, through the above-execution-monitoring-update-reconstruction-re-execution cycle, the voltage regulation control can be continuously adjusted according to the output of new energy sources and load changes, thereby forming a continuous voltage regulation control mechanism oriented towards the real-time operating environment.

[0187] For example, taking a rural 10kV distribution substation as an example, the substation includes one upstream bus node B0, one rooftop photovoltaic access node N1, and one end-user residential load node N2, where N2 is the end-user voltage-sensitive node. The allowable operating range of the node voltage is 0.95pu to 1.05pu, and the reference voltage of node N2 is 0.982pu.

[0188] In the evening, photovoltaic output gradually decreases while residential lighting, air conditioning, and kitchen appliance loads gradually increase, which can easily cause a drop in voltage at the end node N2. Therefore, three consecutive voltage regulation periods, T1, T2, and T3, each lasting 15 minutes, were selected, and the interval data shown in Table 7 were obtained:

[0189] Table 7. Description of Source-Load Disturbance Ranges and Scenarios in Rural Transmission Areas at Different Time Periods

[0190] T1 [-0.05,0.02] [0.05,0.10] Solar power declined slightly, while load increased slightly. T2 [-0.25,-0.15] [0.25,0.45] Solar power sales have declined significantly, while load has risen rapidly. T3 [-0.10,0.00] [0.10,0.20] The load stabilized and the voltage gradually recovered.

[0191] Here, ΔPPV represents the change in photovoltaic output relative to the baseline state, with a negative value indicating a decrease in photovoltaic output; ΔPL represents the change in load demand relative to the baseline state, with a positive value indicating an increase in load. A decrease in photovoltaic output will lower the node voltage, and an increase in load will also lower the node voltage, thus potentially creating an undervoltage risk during the T2 period.

[0192] Under the operating mode of this distribution area, based on the line impedance and historical power flow calculation results, a simplified voltage sensitivity relationship for the terminal node N2 is established: ΔVN2=0.040ΔPPV-0.060ΔPL, where, for every 1MW increase in photovoltaic output, the voltage of node N2 increases by approximately 0.040pu; for every 1MW increase in load, the voltage of node N2 decreases by approximately 0.060pu.

[0193] The voltage ranges for each time period are shown in Table 8:

[0194] Table 8. Voltage Deviation Ranges and Risk Assessment for Node N2 in Rural Transformer Areas During Different Time Periods

[0195] T1 0.9740 0.9798 Safety T2 0.9450 0.9610 Risk of undervoltage T3 0.9660 0.9760 Safety

[0196] Therefore, the T2 period is the boundary approach scenario; when the photovoltaic output is taken as the lower limit of -0.25MW and the load demand is taken as the upper limit of 0.45MW, the voltage of the N2 node is the lowest, which is 0.9450pu. This scenario is the most unfavorable scenario in this example.

[0197] To address the undervoltage risk during period T2, voltage regulation resources are configured as shown in Table 9:

[0198] Table 9. Voltage Regulation Resource Parameters and Voltage Regulation Effects in Rural Transformer Areas

[0199] K1 Source side Photovoltaic inverter reactive power support 0 to 0.15 Mvar The 0.10 Mvar reactive power output caused the N2 voltage to rise by 0.004 pu. K2 Net side On-load tap changer 0 gear, +1 gear Adjusting to +1 increases the N2 voltage by 0.008 pu. K3 Load side Interruptible load reduction 0 to 0.10MW Reducing the load by 0.10MW causes the N2 voltage to rise by 0.006pu.

[0200] The voltage regulation principle is as follows: prioritize the use of source-side photovoltaic inverters for reactive power support; if the voltage is still below the lower limit, then call up the grid-side transformer to increase the voltage level; load-side interruptible loads can be used as backup resources to avoid frequent impacts on user-side power consumption.

[0201] The final voltage regulation control strategy is shown in Table 10:

[0202] Table 10. Voltage Regulation Control Strategies and Implementation Results in Rural Transmission Areas

[0203] T1 Voltage safety Do not start the voltage regulation resource N2 minimum 0.9740 pu T2 N2 has an undervoltage risk. K1 outputs 0.15 Mvar of reactive power. N2 increased from 0.9450 pu to 0.9510 pu T3 Voltage recovery K1 is gradually being phased out of reactive power support. N2 minimum 0.9660 pu

[0204] In specific execution, at the beginning of T2, if the voltage of node N2 is detected to enter the lower boundary adjacent section and there is a risk of it falling below 0.95 pu, the photovoltaic inverter K1 is prioritized to enter the reactive power support state and generate 0.15 Mvar reactive power. If the voltage of node N2 is still detected to be below 0.95 pu in real time, the on-load tap changer K2 is further controlled to rise to +1 level. If the voltage constraint cannot be met after adjusting K1 and K2, the load-side interruptible load K3 is started to reduce some non-critical loads. In this example, only K1 needs to be activated to meet the voltage constraint, so K2 and K3 remain in standby state.

[0205] Table 11 compares the lower bound of the N2 node voltage before and after voltage regulation:

[0206] Table 11 Comparison of the lower limit of the N2 node voltage before and after voltage regulation in rural transformer substations

[0207] T1 0.9740 0.9740 satisfy T2 0.9450 0.9510 satisfy T3 0.9660 0.9660 satisfy

[0208] As shown in the above example, in a scenario where photovoltaic output decreases in rural areas in the evening while residential load increases rapidly, the end node N2 may experience undervoltage risk. By identifying the T2 period as the most unfavorable scenario through an interval robust optimization model and prioritizing the use of the source-side photovoltaic inverter for reactive power support, the voltage at node N2 increased from 0.9450 pu to 0.9510 pu, returning to a safe operating range.

[0209] Example 2:

[0210] Please see Figure 4 Another embodiment of the present invention provides: a source-grid-load coordinated voltage regulation system based on interval robust optimization, comprising: a data processing module, a model building module, a strategy generation module and a coordinated voltage regulation module;

[0211] The data processing module is used to acquire relevant data of the target power grid, establish the output range and load demand range of distributed new energy sources, map the output range and load demand range of new energy sources to node voltage offset ranges based on the source-grid-load voltage sensitivity relationship, and form the corresponding voltage safety operation boundary. The relevant data includes network topology parameters, line parameters, node type information, distributed new energy related data and load related data.

[0212] The model building module is used to establish a robust optimization model for the interval with the joint optimization objectives of minimizing the risk of voltage over-limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation.

[0213] The strategy generation module is used to identify the most unfavorable scenarios that cause the node voltage to approach or exceed the safety boundary based on the interval robust optimization model, and to perform collaborative optimization according to the regulation capabilities and response characteristics of the source side, grid side and load side to obtain a voltage regulation control strategy that meets the voltage constraints.

[0214] The coordinated voltage regulation module is used to control various voltage regulation resources according to the voltage regulation control strategy, and to update the output range of new energy sources and the load demand range in combination with real-time operating data. When the operating status deviates, the range robust optimization model is re-executed to perform continuous voltage regulation control.

[0215] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A source-grid-load coordinated voltage regulation method based on interval robust optimization, characterized in that, include: Acquire relevant data of the target power grid, establish the output range and load demand range of distributed new energy sources, and map the output range and load demand range of new energy sources and load demand ranges to node voltage offset ranges based on the source-grid-load voltage sensitivity relationship, and form corresponding voltage safety operation boundaries. The relevant data includes network topology parameters, line parameters, node type information, distributed new energy related data and load related data. A robust optimization model for the interval is established with the joint optimization objectives of minimizing the risk of voltage exceeding the limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation. Based on the aforementioned interval robust optimization model, the most unfavorable scenario that causes the node voltage to exceed the safety boundary is identified. The regulation capabilities and response characteristics of the source side, grid side, and load side are coordinated for optimization to obtain a voltage regulation control strategy that satisfies the voltage constraint. The voltage regulation control strategy is used to control various voltage regulation resources, and the output range of new energy sources and the load demand range are updated in combination with real-time operating data. When the operating state deviates, the range robust optimization model is re-executed to perform continuous voltage regulation control.

2. The source-grid-load coordinated voltage regulation method based on interval robust optimization according to claim 1, characterized in that, The process involves acquiring relevant data from the target power grid, establishing distributed renewable energy output ranges and load demand ranges, mapping these ranges to node voltage offset ranges based on the source-grid-load voltage sensitivity relationship, and forming corresponding voltage safety operation boundaries, including: The network topology parameters, line parameters, node type information, distributed renewable energy related data, and load related data of the target power grid are obtained. Each node is divided into renewable energy-dominant nodes, load-dominant nodes, and hybrid coupling nodes, forming source-side disturbance units, load-side disturbance units, and corresponding node association sets. The nodes include bus nodes, feeder nodes, and end-connection nodes. Based on distributed new energy related data and load related data, fluctuation segments reflecting the rising process, falling process, stagnation process and sudden change process are extracted respectively. The fluctuation segments are reorganized according to the time adjacency relationship and node association to establish the output range basic element of each source-side disturbance unit and the demand range basic element of each load-side disturbance unit. Based on the topological hierarchy and voltage transmission direction among the new energy dominant node, load dominant node and hybrid coupling node, the output interval basic element and the demand interval basic element are combined and paired to construct the source-load coupling disturbance sequence. According to the order of influence of source-side active power disturbance, source-side reactive power disturbance, load-side active power disturbance and load-side reactive power disturbance on node voltage, a hierarchical voltage sensitivity relationship is established, and the source-load coupling disturbance sequence is screened according to the hierarchical voltage sensitivity relationship to determine the dominant disturbance sequence and the accompanying disturbance sequence participating in the node voltage offset calculation, and the voltage offset interval corresponding to each node is generated. Based on the voltage offset interval, the upper and lower voltage offset segments of each node are segmented and transformed. The transformed node voltage offset results are then superimposed and verified with the node's allowable operating range to form the voltage safety operating boundary corresponding to different nodes.

3. The source-grid-load coordinated voltage regulation method based on interval robust optimization according to claim 2, characterized in that, The step of screening the source-load coupling disturbance sequence according to the hierarchical voltage sensitivity relationship to determine the dominant disturbance sequence and accompanying disturbance sequence participating in the node voltage offset calculation, and generating the voltage offset interval corresponding to each node, includes: The source-load coupled disturbance sequence is decomposed into source-side active disturbance, source-side reactive disturbance, load-side active disturbance, and load-side reactive disturbance. Based on the electrical connection relationship between nodes, the transmission path of each disturbance in the network is calibrated to form a set of disturbance effect paths corresponding to different nodes. Based on the hierarchical voltage sensitivity relationship, each disturbance sequence in the disturbance action path set is matched node by node, and the disturbance sequences are sorted according to the transmission level and action order of the disturbance's influence on the target node voltage to form an ordered disturbance sequence queue for each node. In the ordered perturbation sequence queue, according to the preset sequence truncation rules, the perturbation sequences located in the preceding order are aggregated to determine the dominant perturbation sequence set, and the remaining perturbation sequences are determined as the accompanying perturbation sequence set; The dominant disturbance sequence set and the accompanying disturbance sequence set are combined and superimposed according to the order of their action paths, and node voltage offset process segments are constructed according to the entry order of the disturbance sequences at different action stages, and converged to form the voltage offset interval of the corresponding node.

4. The source-grid-load coordinated voltage regulation method based on interval robust optimization according to claim 3, characterized in that, The voltage offset interval is used to segment and transform the upper and lower voltage offset segments of each node, and the transformed node voltage offset results are superimposed and verified with the node's allowable operating range to form the voltage safety operating boundary corresponding to different nodes, including: The voltage offset interval is divided into an upper voltage offset segment and a lower voltage offset segment according to the direction of voltage change. Based on the order of appearance of each offset segment in the disturbance sequence, the upper voltage offset segment and the lower voltage offset segment are sorted into an offset segment sequence. Based on the topology level of each node, the transmission order of adjacent nodes, and the reference voltage state corresponding to the current node, the turning point of the offset segment sequence is determined segment by segment. Starting from the aforementioned turning point, the voltage upward offset segment and voltage downward offset segment of each node are segmented and turned to form a continuous voltage offset trajectory band corresponding to each node. The continuous voltage offset trajectory of each node is superimposed and verified with the allowable operating range of that node segment by segment. Turning segments within the allowable operating range, turning segments overlapping with the allowable operating range, and turning segments exceeding the allowable operating range are identified. The boundary turning point is taken as the boundary turning point to form the voltage safety operating boundary corresponding to different nodes.

5. The source-grid-load coordinated voltage regulation method based on interval robust optimization according to claim 1, characterized in that, The proposed robust optimization model, which takes minimizing the risk of voltage exceeding the limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation operations as joint optimization objectives, includes: The joint optimization objective of minimizing the risk of voltage exceeding the limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation actions is broken down into risk control sub-objectives, cost constraint sub-objectives, and action sequence sub-objectives. All sub-objectives are then mapped to voltage regulation tasks on the source side, grid side, and load side, forming a set of voltage regulation tasks. The new energy output range, load demand range, and voltage safety operation boundary are embedded into the voltage regulation task set to generate a regulation state sequence and construct a corresponding coordinated voltage regulation sequence. Based on the aforementioned coordinated voltage regulation sequence, each sub-target is sorted, and the voltage constraints and voltage regulation action constraints under interval disturbances are merged to form a robust constraint set. Based on the voltage regulation task set, the coordinated voltage regulation sequence, and the robust constraint set, an interval robust optimization model is constructed.

6. The source-grid-load coordinated voltage regulation method based on interval robust optimization according to claim 5, characterized in that, Based on the aforementioned voltage regulation task set, coordinated voltage regulation sequence, and robust constraint set, an interval robust optimization model is constructed, including: Based on the voltage regulation task set and the coordinated voltage regulation sequence, the regulation actions corresponding to various voltage regulation resources are serialized and expressed according to the triggering order, forming a voltage regulation variable sequence that includes source-side regulation variables, grid-side regulation variables and load-side regulation variables; The new energy output range and the load demand range are expanded according to the disturbance entry sequence to form multiple consecutive interval disturbance segments. The interval disturbance segments are then matched with the voltage regulation variable sequence segment by segment to construct the connection relationship between the disturbance segments and the voltage regulation action. The robust constraint set is embedded into the connection relationship between the disturbance segment and the voltage regulation action. The continuity constraint of voltage regulation action, node voltage constraint and regulation conflict constraint between different disturbance segments are encapsulated across segments to form a unified constraint system covering multiple disturbance segments. Based on the voltage regulation variable sequence, interval disturbance segments, and unified constraint system, the voltage regulation process corresponding to each disturbance segment is connected as a whole to form an interval robust optimization model.

7. The source-grid-load coordinated voltage regulation method based on interval robust optimization according to claim 6, characterized in that, Based on the aforementioned interval robust optimization model, the most unfavorable scenario causing the node voltage to exceed the safety boundary is identified. Coordinated optimization is then performed according to the regulation capabilities and response characteristics of the source side, grid side, and load side to obtain a voltage regulation control strategy that satisfies the voltage constraint, including: Based on the aforementioned interval robust optimization model, the new energy output interval, load demand interval, and voltage regulation variable sequence are expanded accordingly to form multiple candidate disturbance scenarios corresponding to the node voltage change process. The candidate disturbance scenarios are then aggregated to form a scenario set corresponding to different nodes. Each candidate disturbance scenario in the scenario set is compared with the voltage safe operation boundary of the corresponding node one by one, and the disturbance scenarios that cause the node voltage to enter the boundary adjacent segment, cross the boundary turning position, or continuously wander along the boundary are selected as boundary approach scenarios. The voltage offset propagation process in the boundary approach scenario is rearranged to determine the most unfavorable scenario as the disturbance scenario that first triggers voltage constraint conflict or continuously occupies voltage regulation margin under the current coordinated voltage regulation sequence. Based on the worst-case scenario, according to the adjustment initiation conditions, action succession relationships and response sequence of source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources, various voltage regulation resources are hierarchically ordered to form a coordinated voltage regulation process chain. The collaborative voltage regulation process chain is embedded into the interval robust optimization model corresponding to the most unfavorable scenario. The entry order, holding interval and exit order of each voltage regulation resource in different adjustment segments are jointly obtained to generate a voltage regulation control strategy that satisfies the node voltage constraint.

8. The source-grid-load coordinated voltage regulation method based on interval robust optimization according to claim 7, characterized in that, Based on the most unfavorable scenario, and according to the adjustment initiation conditions, action succession relationships, and response sequences of source-side, grid-side, and load-side voltage regulation resources, various voltage regulation resources are hierarchically ordered to form a coordinated voltage regulation process chain, including: Based on the worst-case scenario, scenario matching is performed on the source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources respectively, and the voltage triggering state, power offset state and adjustable margin state corresponding to each type of voltage regulation resource are identified, and the set of adjustment start conditions for each type of voltage regulation resource is determined accordingly. Based on the aforementioned set of adjustment initiation conditions, the adjustment actions of various voltage regulation resources are arranged and sequenced to form an action sequence. The adjustment actions in the action sequence are folded, the adjustment actions with overlapping responses are grouped into parallel segments, the adjustment actions with staggered responses are divided into front and back segments, and the action sequence is divided into a preceding adjustment layer, a following adjustment layer, and a subsequent adjustment layer accordingly. The preceding adjustment layer, the subsequent adjustment layer, and the subsequent adjustment layer are connected in series according to the disturbance progression order in the most unfavorable scenario. The entry order, maintenance order, and exit order of the adjustment actions within each layer are uniformly organized to form a coordinated voltage regulation process chain corresponding to the most unfavorable scenario.

9. The source-grid-load coordinated voltage regulation method based on interval robust optimization according to claim 8, characterized in that, The coordinated voltage regulation process chain is embedded into the interval robust optimization model corresponding to the most unfavorable scenario. The entry order, holding interval, and exit order of each voltage regulation resource in different adjustment segments are jointly calculated to generate a voltage regulation control strategy that satisfies the node voltage constraint, including: The coordinated voltage regulation process chain is embedded into the corresponding interval robust optimization model according to the disturbance advancement order in the most unfavorable scenario, and each adjustment layer in the coordinated voltage regulation process chain is aligned to the different voltage offset stages corresponding to the most unfavorable scenario, forming a correspondence between the adjustment layer and the scenario stage. Based on the correspondence between the regulation layer and the scenario stage, the voltage regulation process in the most unfavorable scenario is divided into regulation segments, and the source-side voltage regulation resources, grid-side voltage regulation resources and load-side voltage regulation resources are respectively connected to the corresponding regulation segments to form a set of regulation segment resources. Within the aforementioned robust optimization model, for each set of resources in each regulation segment, the entry order, holding interval, and exit order of each voltage regulation resource in the corresponding regulation segment are jointly calculated, and the continuity of the resource switching relationship between adjacent regulation segments is checked to form the resource action sequence corresponding to each regulation segment. The resource action sequences corresponding to each adjustment segment are merged and organized according to the scenario stages in the most unfavorable scenario to form a voltage regulation control strategy. The voltage regulation control strategy includes the action entry arrangement, segment maintenance arrangement, and stage exit arrangement for each voltage regulation resource.

10. A source-grid-load coordinated voltage regulation system based on interval robust optimization, used to implement the source-grid-load coordinated voltage regulation method based on interval robust optimization as described in any one of claims 1 to 9, characterized in that, include: Data processing module, model building module, strategy generation module, and collaborative voltage regulation module; The data processing module is used to acquire relevant data of the target power grid, establish distributed renewable energy output range and load demand range, and map the renewable energy output range and load demand range to node voltage offset range based on the source-grid-load voltage sensitivity relationship, and form corresponding voltage safety operation boundaries. The relevant data includes network topology parameters, line parameters, node type information, distributed renewable energy related data and load related data. The model building module is used to establish an interval robust optimization model with the joint optimization objectives of minimizing the risk of voltage over-limit across the entire network, minimizing the overall voltage regulation cost, and optimizing the smoothness of voltage regulation action. The strategy generation module is used to identify the most unfavorable scenario that causes the node voltage to exceed the safety boundary based on the interval robust optimization model, and to perform collaborative optimization according to the adjustment capabilities and response characteristics of the source side, grid side and load side to obtain a voltage regulation control strategy that meets the voltage constraints. The coordinated voltage regulation module is used to control various voltage regulation resources according to the voltage regulation control strategy, and to update the new energy output range and load demand range in combination with real-time operating data. When the operating state deviates, the range robust optimization model is re-executed to perform continuous voltage regulation control.

Images