A reactive power compensation fast response control method and system for distribution transformer area

Abstract

The application belongs to the technical field of power distribution network operation control, and discloses a reactive power compensation fast response control method and system for distribution transformer area, which comprises the following steps: generating change sequences and reliable sequences by aligning and evaluating the electrical quantity data of the distribution transformer area at all times, and performing fuzzy trend discrimination to obtain a trend state; inputting scene clusters and a plurality of candidate sequences into a response model in a rolling prediction domain to obtain predicted trajectories corresponding to each candidate sequence, so as to determine a target sequence; determining the first control quantity of the target sequence as an execution quantity, and sending a control instruction corresponding to the execution quantity to the reactive power compensation device of the distribution transformer area; obtaining a trajectory deviation according to the predicted trajectory corresponding to the execution quantity and the response window after the control instruction is executed, and updating the response model and the discrimination parameter according to the trajectory deviation, so as to enter the next control cycle. The application can improve the adaptability of the distribution transformer area to different change processes and improve the reactive power compensation accuracy.

Description

Technical Field

[0001] This invention relates to the field of power distribution network operation control technology, and in particular to a fast response control method and system for reactive power compensation in distribution transformer areas. Background Technology

[0002] With the increasing number of distributed power sources, power electronic equipment, and fluctuating loads in power distribution networks, the operating status of distribution transformer substations exhibits greater randomness and time-varying characteristics. Specifically, the start-up and shutdown of industrial equipment can cause sudden changes in reactive power within a short period, while residential electricity consumption exhibits significant peak-valley differences, easily leading to voltage fluctuations, decreased power factor, and increased line losses in the substations. Reactive power compensation control, which adjusts reactive power according to the operating status of the substation, is an important means to improve voltage quality and increase the utilization efficiency of power distribution equipment. Faced with rapidly changing loads in distribution transformer substations, the compensation device not only needs to provide appropriate compensation amounts but also needs to keep pace with changes in reactive power demand to avoid compensation actions lagging behind actual operating conditions.

[0003] Current reactive power compensation control in distribution transformer areas primarily relies on collected voltage or reactive power data for judgment. When the detected value exceeds a set range, it activates capacitor banks, static var generators, or other compensation equipment. Some control methods employ tiered switching or closed-loop continuous regulation, adjusting the compensation amount by continuously comparing the detected value with the target value. These methods can improve the power factor and voltage level when load changes are relatively gradual, but the control results largely depend on preset thresholds, empirical parameters, or current-moment detection data, failing to adequately consider the load change process itself.

[0004] However, the load types within a distribution transformer area are diverse, and the reactive power changes caused by different loads vary significantly in amplitude, rate, and duration. Furthermore, the short-term impact of industrial equipment start-up and shutdown can produce instantaneous changes similar to the continuous increase in residential load. Compensation based solely on current measured values ​​and preset thresholds is insufficient to promptly determine whether reactive power demand will rapidly decline or continue to increase. On the one hand, for short-term changes, excessive compensation can easily lead to a reverse deviation after load recovery; on the other hand, for continuous changes, successive corrections may cause compensation lag, and the transformer area measurement data may also be affected by sampling fluctuations and communication delays. If the controller sensitivity is too high, it may frequently change compensation commands; if the sensitivity is too low, it may fail to keep up with load changes in a timely manner.

[0005] Therefore, how to form a compensation control adapted to the current operating conditions based on the reactive power change process when there is a lack of internal load operating data and the transformer parameters change with the operating conditions is an urgent problem to be solved. Summary of the Invention

[0006] This invention provides a fast response control method and system for reactive power compensation in distribution transformer areas to solve the aforementioned problems in the prior art.

[0007] According to a first aspect of the present invention, a fast response control method for reactive power compensation for distribution transformer areas is provided.

[0008] In one embodiment, the reactive power compensation fast response control method for distribution transformer areas includes: generating change sequences and reliable sequences by performing time alignment and quality assessment on electrical quantity data of the distribution transformer area within a state window; performing fuzzy trend discrimination based on the change sequences and reliable sequences to obtain a trend state; the trend state is used to describe the probability of the current reactive power change continuing, falling back, or being uncertain within the prediction period; constructing a scenario cluster based on the trend state and the end state of the state window; inputting the scenario cluster and several candidate sequences into the response model within the rolling prediction domain to obtain the prediction trajectory corresponding to each candidate sequence; using the prediction trajectory to determine the trajectory cost, and determining the candidate sequence with the minimum trajectory cost as the target sequence; determining the first control quantity of the target sequence as the execution quantity, and sending a control command corresponding to the execution quantity to the reactive power compensation device of the distribution transformer area; obtaining the trajectory deviation based on the response window after the control command is executed and the prediction trajectory corresponding to the execution quantity, and updating the discrimination parameters of the response model and fuzzy trend discrimination based on the trajectory deviation to enter the next control cycle.

[0009] According to a second aspect of the present invention, a fast response control system for reactive power compensation for distribution transformer areas is provided.

[0010] In one embodiment, the reactive power compensation fast response control system for distribution transformer areas includes: a sequence generation module, used to generate change sequences and reliable sequences by performing time alignment and quality assessment on electrical quantity data of the distribution transformer area within a state window; a trend discrimination module, used to perform fuzzy trend discrimination based on the change sequences and reliable sequences to obtain a trend state; the trend state is used to describe the probability of the current reactive power change continuing, falling back, or being uncertain within the prediction period; a trajectory decision module, used to construct a scenario cluster based on the trend state and the end state of the state window, and input the scenario cluster and several candidate sequences into the response model within the rolling prediction domain to obtain the predicted trajectory corresponding to each candidate sequence; the predicted trajectory is used to determine the trajectory cost, and the candidate sequence with the minimum trajectory cost is determined as the target sequence; and an execution update module, used to determine the first control quantity of the target sequence as the execution quantity, send the control command corresponding to the execution quantity to the reactive power compensation device of the distribution transformer area; obtain the trajectory deviation based on the response window after the control command is executed and the predicted trajectory corresponding to the execution quantity, and update the discrimination parameters of the response model and fuzzy trend discrimination based on the trajectory deviation to enter the next control cycle.

[0011] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: (1) This invention obtains electrical quantity data within the status window of the distribution transformer area, and determines a reliable sequence by combining data delay, missing degree and local residual. Based on the reliable data, a change sequence relative to the operating baseline is formed, which can reduce the interference of communication delay, sampling fluctuation and data missing on the judgment of reactive power change. This method does not require reading the internal operating data of the electrical equipment, and can complete the control based on the operating data collected from the low voltage side and feeder side of the distribution transformer. It is suitable for application environments where the load source is dispersed, the equipment type is complex and the parameters of the distribution transformer area are difficult to obtain completely.

[0012] (2) The present invention forms a trend state based on the intensity, duration, decline and judgment reliability of reactive power change, and uses the trend state to construct a continuous scenario, a decline scenario and a boundary scenario, so that the controller can determine whether the current reactive power change is more likely to continue to develop or gradually recover in the subsequent period; for short-term impacts that tend to decline, the compensation range can be limited to reduce the reverse deviation generated after the load recovers; for continuously developing reactive power changes, the adjustment range can be expanded in advance to reduce the compensation lag caused by successive catch-up, thereby improving the adaptability of the distribution transformer area to different change processes.

[0013] (3) The present invention inputs the scene cluster and candidate sequence into the response model, calculates the prediction trajectory corresponding to each candidate sequence in the rolling prediction domain, and determines the target sequence by combining reactive power deviation, voltage deviation, control quantity change and reverse compensation risk. The selected target sequence can not only follow the reactive power demand change quickly, but also suppress the problems of compensation quantity exceeding actual demand, repeated changes in control command and voltage deviation of transformer area, which is conducive to improving reactive power compensation accuracy, reducing unnecessary actions of compensation device, and maintaining the smoothness of compensation process.

[0014] (4) This invention uses the response window after the control command is executed to check the predicted trajectory. When the actual response deviates from the predicted result or the reactive power change trend changes, the control command is recalculated and the response model and discrimination parameters are updated according to the trajectory deviation. This allows the control method to be gradually corrected as the load composition and operating status of the distribution transformer area change. In the face of the alternating industrial impact load and residential peak and valley load, the influence of model parameter mismatch on the control effect can be reduced, maintaining a good dynamic following ability and compensation effect, thereby improving the voltage stability and operating efficiency of the distribution transformer area.

[0015] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0017] Figure 1 This is a flowchart illustrating a fast response control method for reactive power compensation in a distribution transformer area, according to an exemplary embodiment. Figure 2 This is a schematic block diagram of a reactive power compensation fast response control system for distribution transformer areas, according to an exemplary embodiment. Figure 3 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment. Detailed Implementation

[0018] The following description and accompanying drawings fully illustrate specific embodiments described herein to enable those skilled in the art to practice them. Some portions and features of certain embodiments may be included in or replace portions and features of other embodiments. The scope of the embodiments herein includes the entire scope of the claims and all available equivalents thereof. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments; similar or identical parts between embodiments can be referred to interchangeably.

[0019] The modules in the apparatus or system of this application can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0020] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0021] Figure 1 An embodiment of the reactive power compensation fast response control method for distribution transformer areas according to the present invention is shown.

[0022] In this optional embodiment, the reactive power compensation fast response control method for distribution transformer areas includes: Step S101, generating a change sequence and a reliable sequence by performing time alignment and quality assessment on the electrical quantity data of the distribution transformer area within the state window; Step S102, performing fuzzy trend discrimination based on the change sequence and the reliable sequence to obtain a trend state; the trend state is used to describe the probability of the current reactive power change continuing, falling back, or being uncertain within the prediction period; Step S103, constructing a scenario cluster based on the trend state and the end state of the state window, and inputting the scenario cluster and several candidate sequences into the response model within the rolling prediction domain to obtain the prediction trajectory corresponding to each candidate sequence; the prediction trajectory is used to determine the trajectory cost, and the candidate sequence with the minimum trajectory cost is determined as the target sequence; Step S104, determining the first control quantity of the target sequence as the execution quantity, and sending a control command corresponding to the execution quantity to the reactive power compensation device of the distribution transformer area; obtaining the trajectory deviation based on the response window after the control command is executed and the prediction trajectory corresponding to the execution quantity, and updating the discrimination parameters of the response model and fuzzy trend discrimination based on the trajectory deviation to enter the next control cycle.

[0023] In this optional embodiment, electrical quantity data is obtained through measurement points on the low-voltage side of the distribution transformer and feeder measurement points. The electrical quantity data includes reactive power and voltage. A change sequence describes the change process of reactive power demand in the distribution area relative to the operating baseline, and a reliable sequence describes the availability of each change value in the change sequence. The reliable sequence is generated by: establishing a unified time axis according to the sampling time carried by each electrical quantity data point; mapping electrical quantity data located between adjacent times on the unified time axis to the unified time axis time closest to its sampling time; and determining the time delay based on the time difference before and after mapping; determining a reliable value based on the time delay, the degree of data loss within the status window, and the residual of the electrical quantity data relative to the local change trend; and arranging the reliable values ​​according to the sampling time to obtain the reliable sequence. The expression for the reliable value is: In the formula, For the first The reliable value at each sampling time; For the first The normalized time delay at each sampling moment; For the first The degree of data missing in the segment where each sampling time is located; For the first The normalized residuals at each sampling time; , and The evaluation coefficient is greater than zero.

[0024] Specifically, evaluation coefficient , and The impact of latency, data loss, and residual on reactive power change assessment can be determined, and calibration should be performed using historical sampling data from the distribution transformer area before control is implemented. The deviation of the change sequence of the three indicators as they increase from their smaller values ​​to the allowable upper limit is statistically analyzed, ensuring that the confidence values ​​corresponding to the three indicators reaching the allowable upper limit all decrease to a preset lower level. The proportions of the three are then adjusted according to the actual impact of communication latency, data loss, and sampling fluctuations in the distribution transformer area; the proportion is increased when communication is unstable. or Increase when measurement fluctuations are large And ensure that the confidence value corresponding to normal data is close to 1.

[0025] In this optional embodiment, the generation of the change sequence includes: determining a short time window close to the current moment in the state window, and determining the portion of the state window before the short time window as a reference window; generating an operating baseline based on the direction and rate of change of reactive power within the reference window, and extending the operating baseline to the short time window; calculating the difference between the reactive power at each sampling moment within the short time window and the operating baseline, and attenuating the difference using the confidence value of the corresponding sampling moment to obtain the change value; arranging the change values ​​according to the sampling moment to obtain the change sequence, so that the change sequence retains the direction, rate of change, and fallback process after the reactive power change.

[0026] Specifically, the operating baseline is obtained based on reactive power data with high confidence values ​​within a reference window. In this invention, the difference between adjacent reactive power values ​​is first calculated according to the sampling time. These differences are then weighted according to their confidence values ​​to determine the main direction and average rate of change of reactive power within the reference window. Next, starting from the weighted reactive power at the end of the reference window, the baseline is extended to a short-time window at each sampling time according to the stated average rate of change, forming the operating baseline. When the direction of change within the reference window is inconsistent or the average rate of change is small, the reliable weighted average value of the reactive power within the reference window can be used as the horizontal operating baseline to avoid local sampling fluctuations being mistaken for continuous changes.

[0027] In this optional embodiment, fuzzy trend discrimination is performed based on the change sequence and the reliable sequence to obtain the trend state, including: determining the change intensity based on the magnitude of the change values ​​in the change sequence, determining the continuity based on the degree to which adjacent change values ​​maintain the same direction; determining the fallback degree based on the degree to which the change values ​​recover from the peak to the operating baseline, and determining the discrimination reliability based on the reliable sequence; inputting the change intensity, continuity, fallback, and discrimination reliability into the fuzzy rule base to obtain the continuity weight, fallback weight, and uncertainty weight; normalizing the continuity weight, fallback weight, and uncertainty weight to obtain the trend state; wherein, when the discrimination reliability decreases, the proportion of uncertainty weight in the trend state is increased.

[0028] Specifically, the intensity of change is determined by the weighted average of the absolute value of the change within the short-term window, the difference between adjacent change values, and their confidence values; a larger value indicates a more drastic change in reactive power demand. The continuity is determined by the proportion of adjacent change values ​​with the same sign and the number of sampling moments where the change continues in the same direction; the longer the direction is maintained, the higher the continuity. The decline is determined by the reduction ratio of the absolute value of the change after the peak relative to the absolute value of the peak; the closer to the operating baseline, the higher the decline. In this invention, the confidence level is the weighted average of the confidence values ​​within the short-term window. The fuzzy rule base predefines the correspondence between continuity weight, decline weight, and uncertainty weight when the above four items are in low, medium, and high ranges. For example, when the intensity of change and the continuity are both high, the continuity weight is increased; when the decline is high, the decline weight is increased; when the confidence level is low or both the continuity and decline are insignificant, the uncertainty weight is increased.

[0029] In this optional embodiment, constructing a scenario cluster based on the end state of the trend state and the state window includes: extending the change direction of the end of the change sequence to the rolling prediction domain according to the continuation weight to generate a continuation scenario; generating a fallback scenario based on the fallback weight and the peak position of the change sequence to recover from the current reactive power demand to the operating baseline; determining the allowable change range of reactive power demand according to the uncertainty weight and the reliable sequence to generate a boundary scenario; and combining the reactive power demand trajectory and scenario weights corresponding to each scenario into a scenario cluster; wherein, the weight of each scenario is determined according to the corresponding continuation weight, fallback weight or uncertainty weight, and is corrected using the discrimination reliability.

[0030] Specifically, the rolling prediction domain starts from the current control moment and includes continuous... The future control interval for each prediction time can have a single prediction interval that is consistent with the control cycle or sampling interval. The continuation scenario continues to extrapolate reactive power demand within the prediction domain according to the direction of change at the end of the change sequence and the recent rate of change, and the extrapolation magnitude is limited by the continuation weight. The decline scenario starts from the current reactive power demand and gradually approaches the operating baseline according to the actual decline rate after the peak or the uniform recovery rate from the current value to the operating baseline. The boundary scenario gives the upper and lower change trajectory around the current reactive power demand, and its range expands as the uncertainty weight increases and the confidence sequence decreases. It is used to cover the reactive power demand changes that may occur when it is impossible to clearly determine whether it is continuation or decline.

[0031] In this optional embodiment, the response model is a local incremental model that is updated online based on the executed control commands and the corresponding response windows; the method for obtaining the predicted trajectory includes: based on the response model, determining the reactive power deviation and voltage deviation at the end of the state window as the initial state, determining each control quantity in the candidate sequence as the control input, and determining the change value of each reactive power demand trajectory in the scenario cluster as the scenario input; calculating the predicted value of reactive power deviation and the predicted value of voltage deviation time-by-time based on the initial state, control input, and scenario input, and forming the predicted trajectory from the predicted value of reactive power deviation and the predicted value of voltage deviation; the expression of the response model is: ; In the formula, For the first The scene in The state variables at each prediction time include the predicted values ​​of reactive power deviation and voltage deviation. For candidate sequences in the th Control quantity at each predicted time; For the first The scene in The change in reactive power demand at each predicted time point; For the first State parameter matrix for each control cycle; For the first Control parameter matrix for each control cycle; For the first The scenario parameter matrix for each control cycle.

[0032] Specifically, reactive power deviation is the difference between reactive power at the end of the state window and the reactive power target value, and voltage deviation is the difference between voltage at the end of the state window and the voltage target value. Both serve as the initial states of the response model. Each control variable in the candidate sequence is the amount of reactive power compensation that needs to be increased or decreased at each prediction time relative to the previous time, and is limited by the capacity and adjustment rate of the compensation device. The scenario cluster consists of the reactive power demand trajectories of the continuous scenario, the fallback scenario, and the boundary scenario, as well as the corresponding scenario weights. The response model calculates the predicted values ​​of reactive power deviation and voltage deviation time by time based on the reactive power demand change value and the candidate control variable for each scenario, and arranges them according to the prediction time to form the prediction trajectory of the candidate sequence in the corresponding scenario.

[0033] In this optional embodiment, the trajectory cost is determined as follows: for each candidate sequence, the reactive power deviation, voltage deviation, and changes between adjacent control quantities in the predicted trajectory are accumulated according to the weights of each scenario; when the predicted reactive power deviation value in the fallback scenario exceeds the target value and changes direction due to the compensation effect of the candidate sequence, the reverse compensation risk is determined based on the magnitude of exceeding the target value; the trajectory cost is determined based on the reactive power deviation, voltage deviation, changes, and reverse compensation risk, and the control quantities in the candidate sequence are constrained not to exceed the capacity range and adjustment rate range of the reactive power compensation device; the expression for the trajectory cost is: ; In the formula, Candidate sequences The trajectory cost; The number of scenes contained in a scene cluster; For the first Scene weights for each scenario; This represents the number of prediction times within the rolling prediction domain. For the first The scene in Predicted reactive power deviation at each prediction time; For the first The scene in Predicted voltage deviation at each prediction time; For the first Control quantity at each predicted time; For the first The scene in Risk of reverse compensation at each predicted time; , , and The cost coefficient is greater than zero.

[0034] Specifically, when the predicted reactive power deviation in a fallback scenario crosses from one side of the target value to the other, the absolute value of the reactive power deviation after the crossover is taken as the magnitude of the crossover of the target value, and the square of this magnitude can be determined as the reverse compensation risk. When the target value is not exceeded, Set to zero. In this invention, the cost coefficient... , , and Based on the principle that each factor has a similar impact on trajectory cost when it reaches its allowable upper limit, the following parameters are determined: Normalization is performed on allowable reactive power deviation, voltage deviation, single control quantity change, and allowable reverse deviation, and then corresponding coefficients are set; when prioritizing reactive power compensation accuracy, the coefficients are increased. Prioritize limiting voltage offset when increasing Increase when reducing frequent movements Suppressing overcompensation increases .

[0035] In this optional embodiment, the method for determining the target sequence includes: determining the allowable range of the first control quantity within the rolling prediction domain based on the trend state, and the allowable change of each subsequent control quantity relative to the previous control quantity; forming a sequence search domain from the allowable range and allowable change of the first control quantity; expanding the allowable range of the first control quantity along the compensation direction of the current reactive power change when the continuation weight increases; narrowing the allowable range of the first control quantity and retaining a reverse adjustment margin within the capacity range of the reactive power compensation device when the uncertainty weight increases; reducing the allowable change of each subsequent control quantity relative to the previous control quantity when the uncertainty weight increases; selecting control quantities at each prediction time within the sequence search domain to generate several coarse-selected sequences; selecting a preset number of retained sequences from the coarse-selected sequences based on the trajectory cost of each coarse-selected sequence; generating a refined sequence within the sequence search domain, centered on the control quantities in each retained sequence, and determining the trajectory cost of the retained sequence and the refined sequence respectively; and determining the retained sequence or refined sequence with the minimum trajectory cost as the target sequence.

[0036] Specifically, the trajectory cost is a comprehensive evaluation value of the control effect of the candidate sequence within the rolling prediction domain. It is obtained by accumulating the predicted reactive power deviation, predicted voltage deviation, changes in adjacent control quantities, and the risk of reverse compensation according to the weights of each scenario. The smaller the trajectory cost, the closer the candidate sequence is to the reactive power target and voltage target under different reactive power variation scenarios. At the same time, the control quantity changes are relatively stable and the possibility of reverse compensation is low. Therefore, the target sequence can be determined by comparing the trajectory costs of each coarsely selected sequence, retained sequence, and refined sequence.

[0037] In this optional embodiment, the trajectory deviation is obtained by: taking the sending time of the control command as the starting time of the response window, and obtaining the actual value of reactive power and the actual value of voltage within the response window; determining the actual reactive power deviation based on the difference between the actual reactive power value and the reactive power target value, and determining the actual voltage deviation based on the difference between the actual voltage value and the voltage target value; re-evaluating the quality of the electrical quantity data within the response window to obtain the response reliability value corresponding to each sampling time; matching the actual reactive power deviation and the actual voltage deviation with the predicted trajectory corresponding to the executed quantity, and determining the trajectory deviation using the response reliability value at the corresponding time.

[0038] Specifically, the actual reactive power and voltage values ​​are the reactive power and voltage measurements re-acquired and time-aligned by the low-voltage side measurement points of the distribution transformer or the corresponding feeder measurement points within the response window after the control command is sent, according to a unified time axis. In this invention, for data from multiple measurement points at the same time, the data can be weighted according to the reliability values ​​of each measurement point to obtain the actual reactive power and voltage values ​​used for verification at that time. When there is time delay, missing data, or data that deviates significantly from the local trend, the response reliability value is used to reduce its impact on the trajectory deviation.

[0039] In this optional embodiment, updating the response model and the discrimination parameters of the fuzzy trend discrimination based on the trajectory deviation to enter the next control cycle includes: when the trajectory deviation exceeds the preset deviation boundary, or when the trend state obtained again according to the response window is inconsistent with the trend state before sending the control command, terminating the execution of the unexecuted control quantity in the target sequence and re-executing the steps of generating the change sequence and the reliable sequence (step S101); when the trajectory deviation does not exceed the preset deviation boundary, updating the response model according to the trajectory deviation and the response reliability value at the corresponding time, updating the discrimination parameters according to the trajectory deviation, and entering the next control cycle.

[0040] Specifically, the preset deviation boundary is determined based on the deviation between the predicted trajectory and the actual response under normal operating conditions. Before control is implemented, multiple response windows with no significant load fluctuations and high data reliability are selected. The difference between the actual reactive power deviation and the predicted reactive power deviation, as well as the difference between the actual voltage deviation and the predicted voltage deviation, are calculated respectively. The larger value of the normal deviation or its average absolute value plus a preset margin is determined as the corresponding deviation boundary. At the same time, this boundary is ensured to not exceed the allowable reactive power deviation and voltage deviation of the transformer area. When sampling fluctuations or communication delays increase, the boundary can be appropriately widened; when the compensation accuracy requirements increase, the boundary can be appropriately narrowed.

[0041] In this optional embodiment, the discrimination parameters include the membership intervals corresponding to the intensity of change, duration, and fallback. The updating method of the discrimination parameters includes: determining the actual turning point when the actual change direction of reactive power changes based on the response window, and determining the predicted turning point when the change direction of the predicted value of reactive power deviation changes based on the prediction trajectory corresponding to the execution quantity; when there is no change in direction within the response window or prediction trajectory, the end time of the corresponding time period is taken as the turning point; determining the adjustment direction according to the sequential relationship between the actual turning point and the predicted turning point; using the intensity of change, duration, and fallback as the sample features of the current sample, determining the update weight of the current sample using the trajectory deviation, and adjusting the membership interval that matches the sample features according to the adjustment direction and update weight.

[0042] Figure 2 An embodiment of a fast response control system for reactive power compensation in distribution transformer areas according to the present invention is shown.

[0043] In this optional embodiment, the reactive power compensation fast response control system for distribution transformer areas includes: a sequence generation module 201, used to generate change sequences and reliable sequences by performing time alignment and quality assessment on electrical quantity data of distribution transformer areas within a state window; a trend discrimination module 202, used to perform fuzzy trend discrimination based on the change sequences and reliable sequences to obtain a trend state; the trend state is used to describe the probability of the current reactive power change continuing, falling back, or being uncertain within the prediction period; a trajectory decision module 203, used to construct a scenario cluster based on the trend state and the end state of the state window, and input the scenario cluster and several candidate sequences into the response model within the rolling prediction domain to obtain the prediction trajectory corresponding to each candidate sequence; the prediction trajectory is used to determine the trajectory cost, and the candidate sequence with the minimum trajectory cost is determined as the target sequence; and an execution update module 204, used to determine the first control quantity of the target sequence as the execution quantity, send the control command corresponding to the execution quantity to the reactive power compensation device of the distribution transformer area; obtain the trajectory deviation based on the response window after the control command is executed and the prediction trajectory corresponding to the execution quantity, and update the discrimination parameters of the response model and fuzzy trend discrimination based on the trajectory deviation to enter the next control cycle.

[0044] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0045] In one embodiment, the computer device may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores static and dynamic data. The network interface communicates with external terminals via a network. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0046] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0047] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0048] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by computer program instructions controlling related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0049] This invention is not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this invention is limited only by the appended claims.

Claims

1. A fast-response control method for reactive power compensation in distribution transformer areas, characterized in that, include: By performing time alignment and quality assessment on the electrical quantity data of the distribution transformer area within the status window, change sequences and reliable sequences are generated. Fuzzy trend discrimination is performed based on the change sequence and the reliable sequence to obtain the trend state; The trend state is used to describe the degree to which the current reactive power change will continue, decline, or remain uncertain during the forecast period; Scenario clusters are constructed based on the end states of trend states and state windows. Within the rolling prediction domain, the scenario clusters and several candidate sequences are input into the response model to obtain the prediction trajectory corresponding to each candidate sequence. The predicted trajectory is used to determine the trajectory cost, and the candidate sequence with the minimum trajectory cost is determined as the target sequence. The first control variable of the target sequence is determined as the execution variable, and the control command corresponding to the execution variable is sent to the reactive power compensation device of the distribution transformer area. The trajectory deviation is obtained based on the response window after the execution of the control command and the predicted trajectory corresponding to the execution variable. The discrimination parameters of the response model and fuzzy trend discrimination are updated based on the trajectory deviation to enter the next control cycle.

2. The reactive power compensation fast response control method for distribution transformer areas according to claim 1, characterized in that, The electrical quantity data is obtained through measurement points on the low-voltage side of the distribution transformer and feeder measurement points, and the electrical quantity data includes reactive power and voltage. The change sequence is used to describe the change process of reactive power demand in the transformer area relative to the operating baseline, and the reliable sequence is used to describe the availability of each change value in the change sequence. The reliable sequence is generated in the following ways: A unified time axis is established based on the sampling time carried by each electrical quantity data. Electrical quantity data located between adjacent times on the unified time axis are mapped to the unified time axis time closest to their sampling time, and the time delay is determined based on the time difference before and after mapping. The confidence values ​​are determined based on the time delay, the degree of data loss within the state window, and the residual of electrical quantity data relative to the local change trend. The confidence values ​​are then arranged according to the sampling time to obtain a confidence sequence.

3. The reactive power compensation fast response control method for distribution transformer areas according to claim 2, characterized in that, The methods for generating the change sequence include: In the state window, a short time window close to the current time is determined, and the portion of the state window before the short time window is determined as a reference window; An operating baseline is generated based on the direction and rate of change of reactive power within the reference window, and the operating baseline is extended to this short time window. Calculate the difference between the reactive power and the operating baseline at each sampling time within the short time window, and use the confidence value of the corresponding sampling time to attenuate the difference to obtain the change value; arrange the change values ​​according to the sampling time to obtain the change sequence.

4. The fast response control method for reactive power compensation for distribution transformer areas according to claim 1, characterized in that, The step of performing fuzzy trend discrimination based on the change sequence and the reliable sequence to obtain the trend state includes: The intensity of change is determined based on the magnitude of the change values ​​in the change sequence, and the duration is determined based on the degree to which adjacent change values ​​maintain the same direction. The degree of decline is determined based on the extent to which the changed value recovers from the peak to the operating baseline, and the confidence level of the judgment is determined based on the confidence sequence. Input the intensity of change, the degree of continuation, the degree of fallback, and the confidence level of the judgment into the fuzzy rule base to obtain the continuation weight, the fallback weight, and the uncertainty weight; Normalize the continuation rights, decline rights, and uncertainty rights to obtain the trend state.

5. The reactive power compensation fast response control method for distribution transformer areas according to claim 4, characterized in that, The scenario cluster constructed based on trend state and state window includes: Based on the continuation right, the change direction at the end of the change sequence is extended to the rolling prediction domain to generate a continuation scene; Based on the fallback weight and the peak position of the change sequence, a fallback scenario is generated to restore the current reactive power demand to the operating baseline; The allowable range of reactive power demand is determined based on the uncertainty weight and the reliable sequence, and a boundary scenario is generated. The reactive power demand trajectory and scenario weights corresponding to each scenario are combined into scenario clusters.

6. The reactive power compensation fast response control method for distribution transformer areas according to claim 1, characterized in that, The response model is a local incremental model that is updated online based on the executed control commands and the corresponding response windows; The methods for obtaining the predicted trajectory include: Based on the response model, the reactive power deviation and voltage deviation at the end of the state window are determined as the initial state, each control quantity in the candidate sequence is determined as the control input, and the change value of each reactive power demand trajectory in the scenario cluster is determined as the scenario input. The reactive power deviation prediction value and voltage deviation prediction value are calculated time-by-time based on the initial state, control input, and scenario input, and the prediction trajectory is formed by the reactive power deviation prediction value and voltage deviation prediction value.

7. The reactive power compensation fast response control method for distribution transformer areas according to claim 1, characterized in that, The methods for determining the trajectory cost include: For each candidate sequence, the reactive power deviation, voltage deviation, and changes between adjacent control variables in the predicted trajectory are accumulated according to the weights of each scenario. When the predicted reactive power deviation in the fallback scenario exceeds the target value and changes direction due to the compensation effect of the candidate sequence, the reverse compensation risk is determined based on the magnitude of exceeding the target value. The trajectory cost is determined based on reactive power deviation, voltage deviation, change, and reverse compensation risk, and the control quantities in the candidate sequence are constrained not to exceed the capacity range and regulation rate range of the reactive power compensation device. The expression for the trajectory cost is: ； In the formula, Candidate sequences The trajectory cost; The number of scenes contained in a scene cluster; For the first Scene weights for each scenario; This represents the number of prediction times within the rolling prediction domain. For the first The scene in Predicted reactive power deviation at each prediction time; For the first The scene in Predicted voltage deviation at each prediction time; For the first Control quantity at each predicted time; For the first The scene in Risk of reverse compensation at each predicted time; , , and The cost coefficient is greater than zero.

8. The fast response control method for reactive power compensation for distribution transformer areas according to claim 1, characterized in that, The methods for determining the target sequence include: Based on the trend state, determine the allowable range of the first control quantity within the rolling prediction domain, and the allowable change of each subsequent control quantity relative to the previous control quantity; form a sequence search domain from the allowable range and allowable change of the first control quantity; When the continuation right increases, the allowable range of the first control quantity is expanded along the compensation direction of the current reactive power change; when the fallback right increases, the allowable range of the first control quantity is reduced, and a reverse adjustment margin is retained within the capacity range of the reactive power compensation device; when the uncertainty right increases, the allowable change of each subsequent control quantity relative to the previous control quantity is reduced. Within the sequence search domain, control quantities at each prediction time are selected to generate several coarse-selected sequences; based on the trajectory cost of each coarse-selected sequence, a preset number of retained sequences are selected from the coarse-selected sequences. Centered on the control variables in each retained sequence, a refined sequence is generated within the sequence search domain, and the trajectory costs of the retained and refined sequences are determined respectively; the retained or refined sequence with the minimum trajectory cost is determined as the target sequence.

9. The fast response control method for reactive power compensation for distribution transformer areas according to claim 1, characterized in that, The methods for obtaining the trajectory deviation include: The start time of the response window is taken as the time when the control command is sent, and the actual values ​​of reactive power and voltage are obtained within the response window. The actual reactive power deviation is determined based on the difference between the actual reactive power value and the reactive power target value; the actual voltage deviation is determined based on the difference between the actual voltage value and the voltage target value. The electrical quantity data within the response window are reassessed to obtain a response confidence value; the actual reactive power deviation and actual voltage deviation are matched with the predicted trajectory corresponding to the executed quantity at different times, and the trajectory deviation is determined using the response confidence value at the corresponding time.

10. A fast-response control system for reactive power compensation in distribution transformer areas, characterized in that, include: The sequence generation module is used to generate change sequences and reliable sequences by performing time alignment and quality assessment on the electrical quantity data of the distribution transformer area within the status window; The trend discrimination module is used to perform fuzzy trend discrimination based on the change sequence and the reliable sequence to obtain the trend state; the trend state is used to describe the degree of possibility that the current reactive power change will continue, decline, or be uncertain during the prediction period; The trajectory decision module is used to construct scene clusters based on the trend state and the end state of the state window. Within the rolling prediction domain, the scene clusters and several candidate sequences are input into the response model to obtain the predicted trajectory corresponding to each candidate sequence. The predicted trajectory is used to determine the trajectory cost, and the candidate sequence with the minimum trajectory cost is determined as the target sequence. The execution update module is used to determine the first control quantity of the target sequence as the execution quantity and send the control command corresponding to the execution quantity to the reactive power compensation device of the distribution transformer area; The trajectory deviation is obtained by comparing the response window after the control command is executed with the predicted trajectory corresponding to the executed quantity. The response model and the discrimination parameters of the fuzzy trend discrimination are updated based on the trajectory deviation in order to enter the next control cycle.

Images