Power distribution area voltage collaborative treatment method based on high-proportion new energy access scene

Abstract

The invention relates to the technical field of power system operation control, and discloses an area voltage autonomous and cooperative governance method based on a power distribution internet of things, comprising the following steps: monitoring output data of distributed photovoltaic in real time through an intelligent fusion terminal deployed in a power distribution area; when it is identified that a preset sudden change event occurs in the photovoltaic output, the intelligent fusion terminal generates a first voltage regulation instruction and executes the first voltage regulation instruction; after the first voltage regulation instruction is generated and before the first voltage regulation instruction is executed, the intelligent fusion terminal performs cooperative information interaction with the intelligent fusion terminal of at least one adjacent transformer area, and corrects the first voltage regulation instruction based on an interaction result; and converging the operation data of the intelligent fusion terminals, carrying out collaborative optimization on decision models preset in the intelligent fusion terminals, and issuing optimized model parameters to the corresponding intelligent fusion terminals to update the decision models in the intelligent fusion terminals. The core pain point of long response delay of centralized control is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system operation control, and particularly relates to a distribution area voltage autonomous and collaborative management method based on a power distribution Internet of Things. BACKGROUND

[0002] With the advancement of the double carbon strategy, the penetration rate of distributed photovoltaic and other new energy in distribution areas is continuously increasing. However, photovoltaic output is intermittent and volatile, and its large-scale access can easily lead to problems such as voltage out-of-limit (voltage being too high or too low) and severe fluctuations in distribution areas, seriously affecting power quality and the safe and stable operation of distribution networks.

[0003] Currently, distribution area voltage management mainly relies on traditional reactive power compensation devices (such as SVG, capacitor reactor switching) or on-load voltage regulating transformers. These methods are mostly local self-reliant adjustment modes, with slow response speed and lack of coordination between distribution areas. When the voltage in a distribution area rises due to sudden increase in photovoltaic output, its adjacent distribution area may be in a different operating state. Traditional localized management cannot utilize the adjustment potential of adjacent distribution areas, which may lead to problems such as poor adjustment effect, frequent device action, and even "over-regulation".

[0004] In addition, existing centralized optimization control methods rely on cloud master stations for calculation and instruction issuance, which has a long response delay and is difficult to cope with the rapid changes in photovoltaic output at the minute or even second level.

[0005] Therefore, there is an urgent need in the art for a collaborative management method that can balance fast response and global optimization in a smart grid to address the voltage challenges brought about by high-proportion new energy access. SUMMARY

[0006] The present application provides a distribution area voltage collaborative management method based on high-proportion new energy access scenarios to solve the above problems.

[0007] The present application is implemented by the following technical solutions:

[0008] A distribution area voltage collaborative management method based on high-proportion new energy access scenarios, comprising the following steps:

[0009] By deploying intelligent fusion terminals in distribution areas, the output data of distributed photovoltaic is monitored in real time; when a preset sudden change event of photovoltaic output is identified, the intelligent fusion terminal generates a first voltage adjustment instruction and executes it within a first time threshold;

[0010] Before generating the first voltage adjustment instruction, the intelligent fusion terminal performs cooperative information interaction with at least one intelligent fusion terminal of a neighboring cell through a local communication network, and modifies the first voltage adjustment instruction based on the interaction result;

[0011] The cloud service platform periodically collects operation data of the intelligent fusion terminals, cooperatively optimizes a decision model pre-stored in each intelligent fusion terminal, and distributes the optimized model parameters to the corresponding intelligent fusion terminal to update the decision model in the intelligent fusion terminal.

[0012] As an optimization, the data monitored by the intelligent fusion terminal in real time further includes a grid-connected point voltage change rate dV / dt and an environmental irradiance change rate G / dt, wherein V is the grid-connected point voltage, G is the environmental irradiance, and t is time.

[0013] As an optimization, based on the output data and the grid-connected point voltage change rate and the specific process of identifying that a preset sudden change event of photovoltaic output occurs, the specific process is as follows:

[0014] A1, real-time calculation of a change amount ΔP of photovoltaic output in a neighboring sampling period, and corresponding grid-connected point voltage change rate dV / dt and environmental irradiance change rate dG / dt;

[0015] A2, when a set of conditions are met at the same time, it is determined that a positive sudden change event of photovoltaic output occurs:

[0016] , is a positive output mutation threshold value;

[0017] , is a positive voltage mutation threshold value;

[0018] , is a positive irradiance mutation threshold value;

[0019] When another set of conditions are met at the same time, it is determined that a negative sudden change event of photovoltaic output occurs:

[0020] , is a negative output mutation threshold value;

[0021] , is a negative voltage mutation threshold value;

[0022] , is a negative irradiance mutation threshold value.

[0023] As optimization, when a preset sudden change event of photovoltaic output is identified, a first voltage adjustment instruction is generated by the intelligent fusion terminal on site within a first time threshold and a specific process is executed, which comprises:

[0024] B1, determining a corresponding voltage adjustment direction according to the identified sudden change event type; if it is a positive sudden change event, the adjustment direction is to reduce the grid connection point voltage; if it is a negative sudden change event, the adjustment direction is to increase the grid connection point voltage;

[0025] B2, based on the related data of the sudden change event, an initial adjustment amount is dynamically calculated by a first calculation model in the decision model pre-installed in the terminal; the first calculation model takes the absolute value of the change amount of the photovoltaic output and the change rate of the grid connection point voltage as input, and outputs an adjustment amount value positively correlated with the input;

[0026] B3, combining the adjustment direction and the initial adjustment amount to generate the first voltage adjustment instruction.

[0027] As optimization, the information delivered by the collaborative information interaction includes: the reactive power support capability curve in the future preset time period predicted based on the state of the intelligent fusion terminal of the adjacent station area.

[0028] As optimization, the specific process of correcting the first voltage adjustment instruction based on the interaction result is:

[0029] An optimization problem is constructed to minimize the grid connection point voltage deviation of the whole network and minimize the total cost of collaborative adjustment;

[0030] Based on the received reactive power support capability curves of each adjacent intelligent fusion terminal, the optimization problem is solved to obtain an optimal reactive power distribution scheme;

[0031] According to the optimal reactive power distribution scheme, the reactive power value requested to each of the adjacent intelligent fusion terminals is determined, and the reactive power adjustment amount borne by the local intelligent fusion terminal in the first voltage adjustment instruction is corrected accordingly.

[0032] As optimization, the specific process of solving the optimization problem is:

[0033] C1, each intelligent fusion terminal dynamically adjusts the weight coefficients of the voltage deviation term and the adjustment cost term in the optimization target according to the real-time collected local grid connection point voltage change rate;

[0034] C2, when the absolute value of the local grid connection point voltage change rate exceeds a preset threshold, a distributed iterative solving process is activated;

[0035] C3. During the iteration process, each intelligent fusion terminal calculates the correction direction of the reactive power adjustment based on the current weight coefficient, and exchanges the correction information with adjacent intelligent fusion terminals.

[0036] C4. Each intelligent converged terminal updates its own reactive power regulation based on the correction information of itself and adjacent intelligent converged terminals, and constrains the updated reactive power regulation within the safe operating range issued by the cloud platform.

[0037] C5. When the change in reactive power adjustment of all the intelligent fusion terminals is less than the convergence threshold in two consecutive updates, the iteration stops and the current result is taken as the optimal reactive power allocation scheme.

[0038] As an optimization, the specific process of dynamically adjusting the weight coefficients of the voltage deviation term and the adjustment cost term in the optimization objective is as follows: Set the weight coefficients of the voltage deviation term. for:

[0039] ;

[0040] in, Let be the absolute value of the rate of change of the grid connection point voltage monitored by the i-th intelligent fusion terminal; It is the absolute value of the rate of change of grid connection point voltage monitored by the nearest adjacent smart fusion terminal that is electrically closest to the i-th smart fusion terminal, obtained through the communication network; The neighboring area influence factor has a value range of (0, 1]. This is the sensitivity coefficient; This is the minimum weighting coefficient for the voltage deviation term.

[0041] As an optimization, the information transmitted in the collaborative information interaction includes: the predicted value of the impact on the grid connection point voltage monitored by the adjacent intelligent fusion terminal after receiving the collaborative request and simulating the operation based on its own decision model.

[0042] As an optimization, the specific process of correcting the first voltage adjustment command based on the interaction results is as follows:

[0043] If the predicted value of the grid connection point voltage impact received from one or more adjacent smart converged terminals exceeds the safety threshold set by the local smart converged terminal, the adjustment amount in the first voltage adjustment command is reduced proportionally, or the request for support from the transformer area where the specific adjacent smart converged terminal is located is canceled.

[0044] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0045] In this invention, by rapidly generating the first command locally and performing pre-execution neighbor cell collaborative correction, when the intelligent fusion terminal locally identifies a photovoltaic sudden change event, it can generate and prepare to execute the first voltage regulation command within a first time threshold (e.g., 100 milliseconds) without waiting for any external commands. This quickly suppresses the initial voltage deterioration trend and solves the core pain point of long response delay in centralized control. The neighbor cell collaboration introduced before execution, through distributed optimization calculation, yields a better solution that comprehensively considers the voltage deviation and regulation cost of the entire network. This makes the final regulation command no longer a reckless local decision, fundamentally solving the over-regulation or regulation conflict problems that may be caused by purely local control.

[0046] This invention sets dynamic weighting coefficients so that when the voltage change rate is large, The voltage automatically approaches 1, with the system objective approximating minimizing voltage deviation. At this point, the system will mobilize all resources (including high-cost equipment) to quickly stabilize the voltage, regardless of cost, as if entering a wartime state, greatly improving the grid's robustness against sudden disturbances. When the voltage stabilizes, Reduce to The weight of economic objectives has increased. The system will prioritize the use of the lowest-cost reactive resources, avoiding unnecessary equipment operation, significantly reducing grid operation losses and equipment wear, and extending equipment life.

[0047] This invention combines voltage change rate-activated iteration with a distributed optimization algorithm, ensuring that time-consuming distributed iteration is only activated when the voltage change rate exceeds a threshold. During most of the grid's stable operation, the terminals remain silent, significantly saving computing resources and communication bandwidth. Each terminal autonomously reaches the optimal solution by exchanging correction information (gradients), independent of real-time instructions from the cloud master station. This achieves decentralized collaborative control, reducing cloud pressure and enabling the system to maintain regional autonomy even in abnormal situations such as communication interruptions, thus improving overall reliability. Attached Figure Description

[0048] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0049] Figure 1 This is a flowchart of a distribution substation voltage collaborative management method based on a high proportion of new energy access scenario, as described in this invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0051] This embodiment 1 provides a method for coordinated voltage management of distribution transformer areas based on a high proportion of new energy access scenarios, such as... Figure 1 As shown, it includes the following steps:

[0052] S1. Through intelligent fusion terminals deployed in the distribution transformer area, the grid connection point voltage and output data of distributed photovoltaic power are monitored in real time. When a preset sudden change event in photovoltaic output is detected, the intelligent fusion terminal generates and executes a first voltage regulation command locally within a first time threshold. The first time threshold is in the range of milliseconds to seconds, for example, 100 milliseconds, to achieve rapid local control.

[0053] In some embodiments, the data monitored in real time by the intelligent fusion terminal also includes the grid connection point voltage change rate dV / dt and the ambient irradiance change rate G / dt of the distributed photovoltaic system, where V is the grid connection point voltage, G is the ambient irradiance, and t is time.

[0054] That is, through intelligent integrated terminals deployed in the distribution area, the output data of distributed photovoltaic power, the rate of change of grid connection voltage dV / dt, and the rate of change of ambient irradiance dG / dt are monitored in real time. When a preset sudden change event in photovoltaic output is detected, the intelligent integrated terminal generates a first voltage regulation command locally within the first time threshold.

[0055] The specific process for identifying sudden change events is as follows:

[0056] Real-time calculation of changes in photovoltaic power output between adjacent sampling periods , and the corresponding dV / dt and dG / dt.

[0057] A positive sudden change event in photovoltaic power output is determined when the following set of conditions are met simultaneously:

[0058] , This is the threshold for positive output mutation.

[0059] , This is the threshold for a positive voltage surge.

[0060] , This is the threshold for positive irradiance mutation.

[0061] A sudden negative change in photovoltaic output is determined when the following other set of conditions are met simultaneously:

[0062] , This is the threshold for negative output mutation.

[0063] , The threshold for negative voltage change;

[0064] , This is the threshold for negative irradiance mutation.

[0065] In some embodiments, generating and executing a first voltage regulation command specifically involves:

[0066] B1. Determine the corresponding voltage adjustment direction based on the identified sudden change event type; where, if it is a positive sudden change event, the adjustment direction is to decrease the grid connection point voltage; if it is a negative sudden change event, the adjustment direction is to increase the grid connection point voltage.

[0067] B2. Based on the relevant data of the sudden change event, the initial adjustment amount is dynamically calculated using the first calculation model in the decision model pre-installed in the terminal; the first calculation model uses the change in photovoltaic output... and the voltage change rate at the grid connection point The absolute value is used as input, and the output is a value of adjustment that is positively correlated with the input; the first calculation model is part of the decision model pre-installed in the intelligent fusion terminal.

[0068] B3. Combine the adjustment direction with the initial adjustment amount to generate the first voltage adjustment command.

[0069] S2. After generating the first voltage adjustment command and before executing the first voltage adjustment command, the intelligent fusion terminal interacts with the intelligent fusion terminal of at least one adjacent transformer station through the local communication network, and modifies the first voltage adjustment command based on the interaction results.

[0070] The information transmitted includes: the reactive power support capability curve of the adjacent transformer area intelligent fusion terminal based on its own state prediction within a preset time period in the future, and / or the predicted value of the impact on its own voltage generated by the adjacent transformer area intelligent fusion terminal after receiving the coordination request and simulating the operation based on its own decision model.

[0071] a. If the transmitted information is a reactive power support capability curve, the specific process for correcting the first voltage regulation command based on the interaction result is as follows:

[0072] An optimization problem is constructed with the objectives of minimizing the voltage deviation at the grid connection points and minimizing the total cost of coordinated regulation. Based on the received reactive power support capability curves of each adjacent smart fusion terminal, the optimization problem is solved to obtain the optimal reactive power allocation scheme. According to the optimal reactive power allocation scheme, the reactive power value requested from each of the adjacent smart fusion terminals is determined, and the reactive power regulation amount undertaken by the local smart fusion terminal in the first voltage regulation command is adjusted accordingly.

[0073] The logic modified in this invention is based on a simple principle: the reactive power regulation amount undertaken by the local terminal is equal to the total regulation demand minus the reactive power supported by the neighboring terminal.

[0074] Assume the total reactive power regulation required by the first voltage regulation command initially generated by the local transformer area (terminal A) is... (Positive values ​​represent reactive power generated, and negative values ​​represent reactive power absorbed.)

[0075] 1. Determine total requirements and support plans:

[0076] By solving the optimization problem, an optimal reactive power allocation scheme was obtained. This scheme specifies the reactive power that each of the multiple terminals needs to bear.

[0077] Assume the following is specified in the plan:

[0078] The reactive power required by neighboring terminal B is

[0079] The reactive power required by the neighboring terminal C is

[0080] ...(and so on, multiple neighboring areas may participate)

[0081] 2. Calculate the adjusted amount for the local terminal:

[0082] The reactive power regulation that local terminal A needs to handle (That is, the corrected amount), is obtained by subtracting the support amount of all neighboring regions from the total adjustment amount:

[0083] ;

[0084] That is, the adjustment task that originally needed to be completed independently by local equipment. Currently, the workload is shared by local equipment and multiple neighboring equipment.

[0085] 3. Correct the first voltage regulation command:

[0086] The original first voltage regulation command contains the regulation amount as follows: .

[0087] In the revised instructions, the adjustment amount that local terminal A needs to perform has been updated to... .

[0088] Simultaneously, local terminal A will send coordination request commands to terminals B, C, etc., via the communication network, requesting them to output respectively. and .

[0089] It should be noted that since voltage deviation and coordinated regulation are parameters with different dimensions, this optimization problem can be defined as:

[0090] ;

[0091] J is the overall objective function. This is a normalized grid connection point voltage deviation index. For example, it can be defined as the ratio of the sum of squares of all grid connection point voltage deviations to a reference value, or the per-unit value of the maximum voltage deviation. , This is the voltage reference value. Let be the voltage at the i-th grid connection point. Where is the reference voltage, and N is the number of grid connection points.

[0092] This is a normalized indicator of total cost for coordinated adjustment.

[0093] For example, it can be defined as the ratio of the total cost of all participating regulating equipment to the maximum possible cost: ; The reactive power provided or absorbed by the j-th reactive power compensation device (such as a smart capacitor, static var generator, etc.). This indicates the generation of reactive power (to compensate for inductive loads and boost voltage). This indicates the absorption of reactive power (to offset excess capacitive reactive power and reduce voltage). is the cost weighting coefficient of the j-th device; M is the total number of devices participating in the adjustment, that is, the number of all available reactive power compensation devices in this collaborative governance. The maximum possible cost is used for normalization. It is a preset benchmark value that represents the weighted total cost that the system may generate under the most extreme and costly adjustment conditions.

[0094] This is the weighting coefficient for the voltage deviation term. This coefficient determines the system's emphasis on voltage quality.

[0095] In some embodiments, the specific process of solving the optimization problem is as follows:

[0096] C1. Each intelligent fusion terminal dynamically adjusts the weighting coefficients of the voltage deviation term and the adjustment cost term in the optimization objective based on the real-time collected local grid connection point voltage change rate.

[0097] In some embodiments, the dynamic adjustment formula for the weighting coefficient of the voltage deviation term is:

[0098] ;

[0099] in, Let be the absolute value of the rate of change of the grid connection point voltage monitored by the i-th intelligent fusion terminal; It is the absolute value of the rate of change of grid connection point voltage monitored by the nearest adjacent smart fusion terminal that is electrically closest to the i-th smart fusion terminal, obtained through the communication network; The neighboring area influence factor has a value range of [value range missing]. ; This is the sensitivity coefficient; This is the minimum weighting coefficient for the voltage deviation term. It is usually set based on engineering experience, system operation requirements and expert knowledge.

[0100] By dynamically adjusting the weight of the voltage deviation term, the system can automatically and smoothly adjust its focus according to the real-time urgency of the power grid.

[0101] In the event of severe voltage fluctuations (crisis management mode), the system focuses almost entirely on minimizing voltage deviations. In this mode, the system mobilizes all resources (including costly regulation methods) to quickly stabilize the voltage and prevent accidents such as voltage collapse, reflecting the paramount principle of safety first.

[0102] When the voltage is stable (economic operating mode), the system ensures voltage quality (but is still affected by voltage fluctuations). Under the premise of constraints, the equipment with the lowest operating cost will be selected first to avoid unnecessary equipment operation, thereby saving energy consumption, reducing equipment wear and tear, and extending equipment life.

[0103] Sudden changes in weights can cause the optimization scheme to jump back and forth between different strategies, leading to frequent device switching and system oscillations. The dynamic adjustment formula of this invention provides a smooth transition, ensuring the consistency of the control strategy and the overall stability of the system.

[0104] Furthermore, even if the local voltage change is not drastic, the system will increase the weight in advance if the voltage of the nearest neighboring cell begins to deteriorate rapidly. They entered a more vigilant state, prepared to provide coordinated support, nipped the voltage problem in the bud, and achieved regional joint defense.

[0105] C2. When the absolute value of the voltage change rate at the local grid connection point exceeds a preset threshold, a round of distributed iterative solution process is activated.

[0106] The design of iteratively solving the problem only when the rate of change exceeds a threshold has the greatest advantage of avoiding unnecessary computational overhead, thereby ensuring the system's rapid response capability at critical moments.

[0107] If the distributed iterative process runs continuously, it will constantly consume the computing resources and communication bandwidth of the intelligent fusion terminal. However, the power grid is in a relatively stable state most of the time, making this continuous computing wasteful. On-demand activation greatly saves system resources.

[0108] Meanwhile, iterative calculations take time. This somewhat expensive collaborative calculation process is only initiated when the system faces a genuine emergency involving a sudden voltage change. This ensures that, in times of crisis, the system's computing power can be focused on solving the most pressing problems, thus enabling the fastest collaborative response.

[0109] This reduces the continuous operating time of the terminal processor and communication module, helping to extend their service life.

[0110] It should be noted that the preset threshold for activating distributed iteration can be the same as the voltage surge threshold for identifying sudden events, or it can be set to a different threshold used to trigger collaborative computing.

[0111] C3. During the iteration process, each intelligent fusion terminal calculates the correction direction of the reactive power adjustment based on the current weight coefficient and exchanges the correction information with adjacent intelligent fusion terminals.

[0112] The specific process is as follows:

[0113] C3.1 Local Gradient Calculation: Each intelligent fusion terminal calculates the current reactive power adjustment... Calculate the partial derivative of the overall objective function J with respect to the adjustment amount to obtain the local correction direction. This direction comprehensively reflects the direction and magnitude of local reactive power regulation adjustments that should be made under the current system conditions to simultaneously optimize voltage quality and regulation costs. This is the reactive power regulation calculated by the j-th intelligent fusion terminal in the k-th iteration; This is the correction direction calculated by the j-th intelligent fusion terminal in the k-th iteration.

[0114] C3.2, Neighborhood Information Exchange: Each intelligent fusion terminal will exchange its local correction direction. Broadcast to all adjacent smart converged terminals, and simultaneously receive corrected directions from adjacent terminals. .

[0115] C3.3 Collaborative Consensus Update (This is a preparation for step C4): After acquiring neighborhood information, each intelligent fusion terminal will combine its own correction direction with the correction direction of the neighborhood, and calculate the reactive power adjustment amount for the next iteration according to the rules of a pre-set distributed optimization algorithm (such as a consensus algorithm or the alternating direction multiplier method). This process ensures that each terminal's decision-making, while pursuing its local objective, also coordinates with the decisions of its neighbors, jointly approaching the globally optimal solution.

[0116] C4. Each intelligent fusion terminal updates its reactive power regulation based on its own and adjacent intelligent fusion terminal correction information, and constrains the updated reactive power regulation within the safe operating range issued by the cloud platform (the safe operating range is a set of feasible intervals of reactive power regulation periodically calculated and issued by the cloud platform, and the determination of this set is subject to the constraint that the voltage of all grid connection points in the distribution substation cluster does not exceed the limit).

[0117] The specific execution process is as follows:

[0118] C4.1 Perform collaborative update calculation:

[0119] Each intelligent fusion terminal corrects the direction based on the neighborhood set obtained in step C3.2. and its own direction of correction According to the update rules of a pre-defined distributed optimization algorithm (such as Alternating Direction Multiplier Method (ADMM) or consensus algorithm), the unconstrained provisional value of reactive power adjustment for the next iteration is calculated. The calculation aims to achieve consensus between local and neighborhood decisions.

[0120] C4.2, Apply safe operation constraints:

[0121] The above provisional values Within the safe operating range specified by the cloud platform, a feasible reactive power regulation amount is obtained for the next iteration. :

[0122] ;

[0123] Where Proj is the projection operation, defined as:

[0124] ;

[0125] This refers to the safe operating range for reactive power regulation of the j-th intelligent fusion terminal, issued by the cloud platform.

[0126] C4.3, Preparing for the next iteration:

[0127] Updated reactive power regulation As the current value, it prepares for the convergence check in step C5 or the next iteration (C3 to C4).

[0128] C5. When the change in reactive power adjustment of all the intelligent fusion terminals is less than the convergence threshold in two consecutive updates, the iteration stops and the current result is taken as the optimal reactive power allocation scheme.

[0129] In other words, the collaborative governance method in this invention includes two working modes:

[0130] Collaborative optimization mode: When the absolute value of the voltage change rate at the local grid connection point exceeds a preset threshold, the distributed iterative solution process is activated to collaboratively correct the first voltage regulation command and execute the corrected command.

[0131] Fast Local Mode: When the absolute value of the voltage change rate at the local grid connection point does not exceed the preset threshold, the distributed iterative solution process is not activated. The intelligent fusion terminal will directly execute the locally generated first voltage regulation command to achieve the fastest single response.

[0132] By automatically switching between the two modes, the method ensures a rapid response in most cases, while initiating more resource-intensive collaborative optimization only in emergency scenarios where voltage fluctuates drastically, thus achieving an optimal balance between response speed and governance effectiveness.

[0133] b. If, after receiving a coordination request, the intelligent converged terminal of an adjacent transformer area generates a predicted value based on its own decision-making model simulation operation regarding the impact on its own grid connection point voltage... .

[0134] If the voltage impact prediction value received from one or more adjacent smart converged terminals exceeds the safety threshold set by the local smart converged terminal, the adjustment amount in the first voltage adjustment command is reduced proportionally, or the request for support from the specific neighboring substation is canceled.

[0135] The specific process is as follows:

[0136] The local intelligent converged terminal receives predicted values ​​of grid connection point voltage impact from each adjacent terminal. (i is the index of the adjacent terminal). Its physical meaning is: after receiving a reactive power support request from a local terminal, the adjacent intelligent fusion terminal i, based on its local network parameters, load status, and the requested reactive power value... It performs rapid forward calculations using its internally pre-built decision model to predict the expected change in voltage at its own grid connection point caused by the action. This value is used to determine whether providing support would pose a risk to the voltage safety of the adjacent terminal.

[0137] The local terminal has a security threshold. (is a positive number). The verification condition is: for any adjacent terminal i, if it satisfies

[0138] If so, then the adjacent terminal i is determined to be in a voltage safety sensitive state.

[0139] Based on the above verification results, the collaboration request should be corrected according to the following rules:

[0140] Scenario 1 (Minor Limit Exceedance / Single Neighbor Cell Limit Exceedance): If one or more neighboring terminals are determined to be in a voltage safety sensitive state, but the predicted voltage impact exceeds the threshold by a manageable amount, the local terminal will request reactive power values ​​from these specific neighboring cells. The reduction is proportional, and the reduction formula is:

[0141] ;

[0142] in, This represents the actual reactive power requested after reduction. This ensures voltage safety in neighboring areas while providing partial support.

[0143] Scenario 2 (Severe Limit Exceedance / Critical Neighbor Area Limit Exceedance): If the voltage of a certain adjacent terminal affects the predicted value. Much larger (For example, exceeding twice the threshold), or if the neighboring cell is the closest and most important support provider in terms of electrical distance, the local terminal cancels the support request to that specific neighboring cell (one or more), i.e., it... Set to zero.

[0144] b.3. Local instruction re-correction:

[0145] After revising the coordination request, the local intelligent fusion terminal recalculates its own reactive power regulation load based on the revised neighbor cell support scheme. (The calculation formula is the same as before), and the first voltage adjustment command is finally updated.

[0146] It should be noted that the safety threshold It is configured by the local intelligent fusion terminal and used to verify the predicted voltage impact values ​​fed back by all adjacent terminals. This design is based on the following considerations:

[0147] Ensuring consistency in coordination strategies: As the initiator of a coordination request, the local terminal must ensure that its request will not cause regional impacts on the distribution network. Using a unified local bit threshold for evaluation can avoid decision-making fragmentation caused by differences in security standards among different adjacent terminals, ensuring that coordinated actions are carried out under unified security standards.

[0148] Improved system robustness: This scheme reduces reliance on neighboring terminal parameter configurations and communication reliability. Even if some neighboring terminals fail to correctly set or synchronize their own security parameters, the local terminal can still make security decisions based on its own reliable policies, enhancing the overall system's fault tolerance.

[0149] Simplified collaborative interaction protocol: This solution eliminates the need to synchronize dynamically changing security threshold parameters between terminals, simplifies the message structure and human-machine interface configuration of collaborative communication, and makes the system easier to implement and maintain.

[0150] However, with this mechanism, there may be situations where the predicted value of adjacent terminals does not exceed the safety threshold of the local terminal, but this predicted value may cause voltage safety risks to the adjacent terminals themselves. To address this situation, the present invention can also set the following protection mechanism:

[0151] The collaborative governance approach includes a two-tiered security protection mechanism to ensure the local security of each transformer station area:

[0152] First layer (requester verification): As mentioned above, the local intelligent converged terminal uses a unified security threshold. The received predicted values ​​of the impact of neighboring cell voltage are verified. This is a preliminary screening aimed at avoiding the initiation of obviously unsafe collaborative requests from the perspective of the requester.

[0153] The second layer (autonomous decision-making by the executor): After receiving a specific reactive power support instruction, the adjacent intelligent converged terminal must perform a local safety check before final execution. Even if its predicted value has passed the requester's verification, if the reactive power adjustment required by the instruction exceeds its local safe operating range (such as the equipment capacity limit or causing local voltage to exceed the limit), the adjacent terminal has the right to refuse execution or autonomously reduce the support amount to its local (adjacent intelligent converged terminal) safe range and report this situation to the local terminal.

[0154] The local security range refers to the set of pre-set constraints within each intelligent converged terminal to ensure the safe and stable operation of its local transformer area. This range is independently determined and maintained by the terminal based on the physical capacity of the reactive power compensation equipment it monitors, the qualified voltage limit of the grid connection point, and relevant protection settings. It is the ultimate security boundary that the terminal cannot infringe when executing any control command (including responding to coordination requests).

[0155] Through the above two-layer mechanism, the efficiency and consistency of collaborative requests are guaranteed, and the absolute security of each substation as an autonomous unit is ensured, forming a decentralized and reliable security defense system.

[0156] S3. The cloud service platform periodically collects the operating data of each of the intelligent fusion terminals, performs collaborative optimization on the parameters of the decision models pre-installed in each of the intelligent fusion terminals, and sends the optimized model parameters to the corresponding intelligent fusion terminal to update the decision models in the intelligent fusion terminal.

[0157] Decision models include, but are not limited to:

[0158] The first computational model used to quickly generate initial adjustment instructions;

[0159] Used for dynamically calculating the weighting coefficient of the voltage deviation term. The weight adjustment model.

[0160] The optimization parameters of the weight adjustment model include, but are not limited to, the sensitivity coefficient. Influence factors of neighboring regions By optimizing these parameters, the terminal's sensitivity to voltage change rate and its attention to neighboring cell states can be corrected, thereby enabling dynamic weighting. The real-time calculation results are more in line with the optimal strategy for the long-term operation of the entire network.

[0161] The specific process includes the following sub-steps:

[0162] S3.1 Data Aggregation and Preprocessing:

[0163] The cloud service platform aggregates encrypted historical operational data from intelligent converged terminals across the entire domain on a minute-by-minute or hourly basis. This data includes at least:

[0164] The time series V(t) of the grid connection point voltage monitored by each terminal.

[0165] Time series of output data of distributed photovoltaic power .

[0166] Action record sequence of each reactive power compensation device (adjustment time and adjustment amount Q(t)).

[0167] Neighbor interaction information recorded in collaborative events.

[0168] The cloud cleans, aligns, and formats the data to build a high-quality training dataset for model optimization.

[0169] S3.2 Centralized Model Training and Collaborative Optimization:

[0170] The cloud leverages its powerful computing capabilities and aggregates data from across the entire network to process pre-installed decision-making models (such as the first calculation model and the weight adjustment model that calculates weight coefficients) within the terminal. (etc.) to conduct intensive retraining.

[0171] Optimization objective: The overall goal is to minimize long-term voltage deviation at grid connection points, governance costs, and equipment operation frequency.

[0172] Optimization method: Employ advanced machine learning algorithms, such as deep reinforcement learning (DRL) or federated learning.

[0173] Through deep reinforcement learning, the cloud can simulate the entire power distribution network environment and train a decision model with better performance.

[0174] Federated learning allows for the collaborative training of a global model using the local computing power of each device without aggregating raw data (to protect privacy).

[0175] Output: Optimized global model parameters obtained after training. .

[0176] Global model parameters include:

[0177] Parameters of the first computational model The first computational model can be a simple linear function, a lookup table, or a small neural network;

[0178] Parameters of the weighted model Specifically, there is a sensitivity coefficient. Neighboring area influence factors It can also have a minimum weight. .

[0179] S3.3 Model parameter differencing and security verification:

[0180] The cloud will optimize the global model parameters The model parameters currently being used by each terminal Compare them.

[0181] Generate parameter differential package .

[0182] The differential package is simulated and verified for security to ensure that the new parameters will not cause system instability in typical and extreme scenarios.

[0183] S3.4 Parameter Distribution and Silent Update:

[0184] The optimized model parameters will be transmitted to the cloud via a secure, encrypted channel. Or a validated parameter differential package The data is distributed to the corresponding intelligent fusion terminals. When communication conditions are good and the system is stable, the terminals complete silent updates and hot switching of the model in the background without manual intervention.

[0185] In other words, the method of this invention constructs a closed-loop optimization system with two time scales: at the second-level real-time scale, each terminal uses its local fixed decision model and real-time data to quickly calculate dynamic weights and adjustment instructions; at the hour-level offline scale, the cloud platform performs global collaborative optimization and distribution of decision model parameters within the terminals based on long-term operational data. This mechanism enables the system to possess both the agility of real-time response and the intelligence of continuous evolution.

[0186] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for coordinated voltage management of distribution transformer areas based on a high proportion of renewable energy access scenarios, characterized in that, Includes the following steps: The intelligent integrated terminal deployed in the distribution area monitors the output data of distributed photovoltaic power in real time; when a preset sudden change event in photovoltaic output is detected, the intelligent integrated terminal generates and executes a first voltage regulation command locally within the first time threshold. After generating the first voltage adjustment command and before executing the first voltage adjustment command, the intelligent fusion terminal interacts with the intelligent fusion terminal of at least one adjacent transformer station through the local communication network, and modifies the first voltage adjustment command based on the interaction results. The cloud service platform periodically aggregates the operating data of each intelligent fusion terminal, collaboratively optimizes the decision models pre-installed in each intelligent fusion terminal, and sends the optimized model parameters to the corresponding intelligent fusion terminal to update the decision models in the intelligent fusion terminal.

2. The method for coordinated voltage management of distribution substations based on a high proportion of renewable energy access scenario, as described in claim 1, is characterized in that... The data monitored in real time by the intelligent fusion terminal also includes the grid connection point voltage change rate dV / dt and the environmental irradiance change rate G / dt of the distributed photovoltaic system, where V is the grid connection point voltage, G is the environmental irradiance, and t is time.

3. The method for coordinated voltage management of distribution substations based on a high proportion of renewable energy access scenario, as described in claim 2, is characterized in that... The specific process of identifying a preset sudden change event in photovoltaic output based on the aforementioned power output data, grid connection point voltage change rate, and detection is as follows: A1. Calculate in real time the change in photovoltaic output ΔP between adjacent sampling periods, as well as the corresponding grid connection point voltage change rate dV / dt and environmental irradiance change rate dG / dt; A2. A positive sudden change event in photovoltaic power output is determined when the following set of conditions are met simultaneously: , This is the threshold for positive output mutation. , This is the threshold for a positive voltage surge. , This is the threshold for positive irradiance abrupt change. A sudden negative change in photovoltaic output is determined when the following other set of conditions are met simultaneously: , This is the threshold for negative output mutation. , The threshold for negative voltage change; , This is the threshold for negative irradiance mutation.

4. The method for coordinated voltage management of distribution substations based on a high proportion of renewable energy access scenario, as described in claim 1, is characterized in that... When a preset sudden change event in photovoltaic output is detected, the intelligent fusion terminal generates and executes a first voltage regulation command locally within a first time threshold. Specifically: B1. Determine the corresponding voltage adjustment direction based on the identified sudden change event type; where, if it is a positive sudden change event, the adjustment direction is to decrease the grid connection point voltage; if it is a negative sudden change event, the adjustment direction is to increase the grid connection point voltage. B2. Based on the relevant data of the sudden change event, the initial adjustment amount is dynamically calculated using the first calculation model in the decision model pre-installed in the terminal; the first calculation model uses the change in photovoltaic output... and the voltage change rate at the grid connection point The absolute value is used as input, and the output is a value of adjustment that is positively correlated with the input; B3. Combine the adjustment direction with the initial adjustment amount to generate the first voltage adjustment command.

5. The method for coordinated voltage management of distribution transformer areas based on a high proportion of new energy access scenario as described in claim 1, characterized in that, The information transmitted during collaborative information exchange includes: reactive power support capability curves for a future preset time period, provided by the intelligent converged terminals of adjacent transformer substations based on their own state predictions.

6. A method for coordinated voltage management of distribution transformer areas based on a high proportion of renewable energy access scenario, as described in claim 5, is characterized in that... The specific process of correcting the first voltage adjustment command based on the interaction results is as follows: Construct an optimization problem with the objectives of minimizing the voltage deviation at all grid connection points and minimizing the total cost of coordinated regulation; Based on the received reactive power support capability curves of each adjacent intelligent fusion terminal, the optimization problem is solved to obtain the optimal reactive power allocation scheme. Based on the optimal reactive power allocation scheme, the reactive power value requested from each of the adjacent intelligent fusion terminals is determined, and the reactive power adjustment amount undertaken by the local intelligent fusion terminal in the first voltage adjustment command is adjusted accordingly.

7. A method for coordinated voltage management of distribution substations based on a high proportion of renewable energy access scenario, as described in claim 6, is characterized in that... The specific process of solving the optimization problem is as follows: C1. Each intelligent fusion terminal dynamically adjusts the weighting coefficients of the voltage deviation term and the adjustment cost term in the optimization objective based on the real-time collected local grid connection point voltage change rate. C2. When the absolute value of the voltage change rate at the local grid connection point exceeds a preset threshold, the distributed iterative solution process is activated. C3. During the iteration process, each intelligent fusion terminal calculates the correction direction of the reactive power adjustment based on the current weight coefficient, and exchanges the correction information with adjacent intelligent fusion terminals. C4. Each intelligent converged terminal updates its own reactive power regulation based on the correction information of itself and adjacent intelligent converged terminals, and constrains the updated reactive power regulation within the safe operating range issued by the cloud platform. C5. When the change in reactive power adjustment of all the intelligent fusion terminals is less than the convergence threshold in two consecutive updates, the iteration stops and the current result is taken as the optimal reactive power allocation scheme.

8. A method for coordinated voltage management of distribution substations based on a high proportion of renewable energy access scenario, as described in claim 7, is characterized in that... The specific process of dynamically adjusting the weighting coefficients of the voltage deviation term and the adjustment cost term in the optimization objective is as follows: Set the weighting coefficient of the voltage deviation term. for: ； in, Let be the absolute value of the rate of change of the grid connection point voltage monitored by the i-th intelligent fusion terminal; It is the absolute value of the rate of change of grid connection point voltage monitored by the nearest adjacent smart fusion terminal that is electrically closest to the i-th smart fusion terminal, obtained through the communication network; The neighboring area influence factor has a value range of (0, 1]. This is the sensitivity coefficient; This is the minimum weighting coefficient for the voltage deviation term.

9. A method for coordinated voltage management of distribution transformer areas based on a high proportion of renewable energy access scenario, as described in claim 1, is characterized in that... The information transmitted in the collaborative information interaction includes: the predicted value of the impact on the grid connection point voltage monitored by the intelligent fusion terminal of the adjacent transformer area after receiving the collaborative request and simulating the operation based on its own decision model.

10. A method for coordinated voltage management of distribution transformer areas based on a high proportion of new energy access scenario, as described in claim 9, is characterized in that... The specific process of correcting the first voltage adjustment command based on the interaction results is as follows: If the predicted value of the grid connection point voltage impact received from one or more adjacent smart converged terminals exceeds the safety threshold set by the local smart converged terminal, the adjustment amount in the first voltage adjustment command is reduced proportionally, or the request for support from the transformer area where the specific adjacent smart converged terminal is located is canceled.

Images