A method for energy efficiency optimization control of power distribution cabinet clusters based on digital twins

By identifying the offsetting pseudo-steady state of bidirectional regulation in the distribution cabinet cluster, the energy efficiency extrapolation of the digital twin model is optimized, solving the problem of redundant regulation in the existing technology and improving the accuracy and stability of energy efficiency optimization control of the distribution cabinet cluster.

CN123073731APending Publication Date: 2026-11-10TIANJIN PUCHUAN CONTROL EQUIP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611562037.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-10
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing digital twin-based power distribution cabinet cluster energy efficiency optimization control technology cannot identify the temporary stable state formed by load branches under opposite-direction regulation, resulting in repeated regulation and repeated load migration, increasing line losses and the number of times the equipment operates, and reducing control accuracy and stability.

Method used

By acquiring the total active power of the distribution cabinet, the active power of the load branch, the target control power, and the control direction, the remaining power of the previous round of control and the expected end time of the reverse control are determined, forming a two-way control action sequence. The control offset false stability stage is identified, and cluster energy efficiency is extrapolated in the digital twin model to determine the target control power and control the load branch.

Benefits of technology

It effectively identifies bidirectional control offsetting, reduces redundant control, lowers the deviation between the predicted power and the actual power of the digital twin model, and improves the accuracy and stability of the energy efficiency optimization control of the power distribution cabinet cluster.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123073731A_ABST
    Figure CN123073731A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of power distribution system operation and control technology, and relates to a method for optimizing the energy efficiency of distribution cabinet clusters based on digital twins. It collects active power, target control power, and execution feedback from the distribution cabinets and load branches to determine the remaining control power corresponding to the previous round of control applied to different load branches within the same distribution cabinet, as well as the expected end times for both control directions. This is combined with the active power changes already generated by reverse control to form a bidirectional control action sequence. When the total active power change is within a preset power fluctuation range but the two control directions still produce opposite power changes, a false stability phase of control cancellation is determined. The cluster simulations of continued and stopped reverse control are compared in the digital twin model to determine and execute the target control power for the corresponding branch, reducing redundant control and repeated adjustments to the target control power.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of power distribution system operation and control technology, and relates to a method for energy efficiency optimization control of power distribution cabinet clusters based on digital twins. Background Technology

[0002] In scenarios involving collaborative power supply across multiple distribution cabinets, such as industrial parks, large commercial complexes, and manufacturing workshops, digital twin-based energy efficiency optimization control technology for distribution cabinet clusters is increasingly being adopted to reduce operating losses and improve load utilization efficiency. This technology typically collects voltage, current, active power, and load operation information from each distribution cabinet, establishes a mapping of cabinet operating states in a digital twin model, and generates load adjustment strategies based on the cluster load distribution. By controlling the corresponding load branches to change their operating power, it achieves optimized load allocation among the distribution cabinets.

[0003] However, in actual multi-round control processes, the execution response time of different load branches varies. Some load branches can quickly complete power adjustment after receiving control, while others require a certain amount of time to complete the target control power change due to equipment operation and actuator response processes. Therefore, when the previous round of control has not yet fully ended, the digital twin model may re-optimize based on the currently collected active power and generate a load adjustment strategy in the opposite direction to the previous round of control. At this time, some load branches corresponding to the previous round of control still generate power changes in the original control direction, while other load branches corresponding to the next round of control have already generated power changes in the opposite direction, resulting in two directions of control effects existing simultaneously within the distribution cabinet.

[0004] When the changes in active power in two directions are close, the change in the total active power of the distribution cabinet may be temporarily small. Existing technologies typically judge the control effect based on the change in total active power, easily mistaking this state for a stable load. In reality, this stable state is a temporary stability formed by the mutual cancellation of control effects in both directions and does not reflect the true load changes within the distribution cabinet. When the control effect in one direction ends prematurely, the remaining control effect in the other direction loses its cancellation effect, causing a significant shift in the total active power of the distribution cabinet in a short period. This leads to a discrepancy between the predicted and actual operating states of the digital twin model. These problems result in repeated control and repeated load migration within the distribution cabinet cluster, causing some distribution cabinets to frequently change their operating states, increasing line losses and the number of equipment actions, and reducing the accuracy and stability of energy efficiency optimization control of the distribution cabinet cluster. Therefore, existing digital twin-based energy efficiency optimization control technologies for distribution cabinet clusters still need to address the problem of failing to identify the internal control cancellation process behind the stable total power when control effects in opposite directions are present, and failing to predict the actual power change after the cancellation is lifted. Summary of the Invention

[0005] In view of this, in order to solve the problems mentioned in the background technology, a method for energy efficiency optimization control of power distribution cabinet clusters based on digital twins is proposed.

[0006] This invention proposes a method for optimizing the energy efficiency control of power distribution cabinet clusters based on digital twins, comprising the following steps: The system obtains the total active power of the distribution cabinet, the active power of the load branch, the target control power, the control issuance time, the execution feedback and the control direction, determines the remaining control power corresponding to the previous round of control for different load branches in the same distribution cabinet and the expected end time of the two control directions, and forms a two-way control action sequence by combining the active power change generated by the reverse control. When the change in the total active power of the distribution cabinet is within the preset power fluctuation range, and the active power change directions corresponding to the two control directions are opposite, the control cancellation false stability stage is determined. Compare the expected end times of the two control directions, take the earlier expected end time as the first end time, and obtain the predicted active power of the distribution cabinet after the cancellation based on the change in the remaining active power of the other control direction after the first end time. In the digital twin model, cluster energy efficiency simulations are performed for both the continuation and cessation of reverse regulation. Based on the simulation results, the target regulation power is determined, and the corresponding load branches are controlled.

[0007] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This invention forms a bidirectional control action sequence by determining the remaining control power of the previous round of control and the change in active power generated by the reverse control. When the change in the total active power of the distribution cabinet is within the preset power fluctuation range and the change in active power corresponding to the two control directions still exists, the control cancellation false stability stage is determined. This can identify the bidirectional power cancellation behind the temporary stability of the total active power and reduce the situation of mistaking the temporary stability as the completion of control.

[0008] (2) By comparing the expected end times of two control directions, this invention determines the change in the remaining active power corresponding to the end time of the first control direction and the end time of the other control direction, and obtains the predicted active power of the distribution cabinet after the cancellation is removed. It can reflect the power offset caused by the continued execution of the other control direction before the end of one control direction, and reduce the deviation between the predicted power and the actual power of the digital twin model.

[0009] (3) The present invention uses the predicted active power of the distribution cabinet after the cancellation is removed as the input for the energy efficiency simulation of the digital twin cluster, compares the continued execution and the termination of the reverse regulation, and determines the load allocation and control of the corresponding load branch accordingly, thereby reducing repeated regulation and repeated adjustment of the target regulation power, and reducing unnecessary repeated actions of the load controller. Attached Figure Description

[0010] Figure 1 This is an overall flowchart of the energy efficiency optimization control method for a power distribution cabinet cluster based on digital twins according to the present invention. Figure 2 This is a flowchart illustrating the process of determining the remaining control power and the expected end time in this invention. Figure 3 This is a flowchart illustrating the process of determining the predicted active power of the distribution cabinet after the cancellation is performed in this invention. Figure 4 This is a schematic diagram of power change when the previous round of regulation and the reverse regulation exist simultaneously in this invention. Detailed Implementation

[0011] The following will describe, with reference to the accompanying drawings of the embodiments of the present invention, a method for optimizing the energy efficiency control of a distribution cabinet cluster based on digital twins. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0012] The following description, in conjunction with the accompanying drawings and embodiments, further illustrates the present invention's method for optimizing the energy efficiency of a distribution cabinet cluster based on digital twins.

[0013] Example 1: See Figures 1 to 4 A method for optimizing the energy efficiency control of a distribution cabinet cluster based on digital twins, comprising: This embodiment applies to a scenario in an industrial park where multiple distribution cabinets jointly supply power to production equipment, ventilation equipment, water pumps, and auxiliary electrical equipment. Each distribution cabinet is equipped with an energy metering device on its incoming line side, and each adjustable load branch is equipped with a branch energy metering device connected to the corresponding load controller. A digital twin model obtains the total active power of each distribution cabinet, the active power of each load branch, and execution feedback, and determines the control process based on the load distribution requirements of the distribution cabinet cluster. Because different load branches have different start-up, shutdown, and power adjustment speeds, the next round of control may have already acted in the opposite direction on other load branches before the previous round of control is completed, temporarily stabilizing the total active power of the distribution cabinets while internal power changes in the opposite direction still exist.

[0014] In response to the above situation, this embodiment determines the remaining control power corresponding to the previous round of control, the change in active power corresponding to the reverse control, and the expected end time of the two control directions; when the change in the total active power of the distribution cabinet is within the preset power fluctuation range and the active power corresponding to the two control directions is still changing in opposite directions, the control cancellation false stability stage is determined, and subsequent cluster energy efficiency is extrapolated based on the corrected predicted active power of the distribution cabinet.

[0015] The digital twin model in this embodiment runs on a computer connected to the power metering device and load controller, and employs an active power calculation model updated according to a unified sampling period. The model is organized hierarchically according to load branches, distribution cabinets, and distribution cabinet clusters: the load branch layer records the branch number, the distribution cabinet number, the communication address of the metering device and load controller, the current active power, the target control power, the upper and lower limits of active power, the control direction, the control issuance time, and the execution feedback; the distribution cabinet layer records the power supply connection, the load branch, the total active power, and the upper limit of allowable active power, and saves the bidirectional control action sequence and the predicted active power of the distribution cabinet; the distribution cabinet cluster layer aggregates the above information from each distribution cabinet and executes the load allocation calculation and the deduction of the two control methods in step S4. The branch number and distribution cabinet number are used for hierarchical aggregation calculation and for locating the corresponding load controller.

[0016] During model construction, the affiliation of each distribution cabinet and load branch is first entered according to the on-site power supply connection, and the communication addresses of metering devices and load controllers are bound to their corresponding numbers. Then, the upper and lower power limits are entered according to the equipment operating limits, allowable adjustment range, and current production requirements. The total active power of the distribution cabinets and the active power of the branches sampled in the first group are used as the initial values ​​of the model, and the target control power, direction, and issuance time of any incomplete control are read. The unified sampling period is not less than the longest update period of each metering device. The incoming power and branch power at the same sampling time are used in the same judgment window. If the sample is missing, the sampling time cannot be matched, or the execution feedback cannot be associated with the control record, the sample group will not be used for bidirectional power comparison, and the continuous sampling period will be re-accumulated after the effective sampling is restored.

[0017] In historical normal control records, the shortest control duration that maintains the target control power and direction, ensures continuous and effective feedback, and allows for continuous adjacent comparisons is selected. This duration is then rounded down by the ratio of the selected duration to a uniform sampling period to determine the number of continuous sampling periods, ensuring that these periods can form adjacent sampling differences. Each period corresponds to an adjacent sampling interval, and sampling values ​​covering all intervals are used. If historical records are insufficient, records are first obtained through normal on-site control. If the actual bidirectional control overlap time does not cover the required window, confirmation of the control cancellation false stability phase is not performed. When the target control power or control direction changes, the judgment window is re-accumulated, and samples before and after the adjustment are not mixed for judgment.

[0018] The data flow of the model operation consists of measured input, branch calculation, distribution cabinet summary, and cluster inference in sequence. In each sampling period, the active power with number and sampling time is written into the corresponding load branch and distribution cabinet, and the execution feedback is matched with the control record according to the branch number and control issuance time. Step S1 reads the measured power and incomplete control records of this period, outputs the remaining control power, power change rate and expected end time of each branch, and forms a two-way control action sequence for each distribution cabinet according to the branch affiliation. Step S2 reads the sequence and the total active power change of the distribution cabinet, and outputs the determination of the control to offset the false stability stage. Step S3 calculates the power change correction time and the predicted active power of the distribution cabinet after the offset is removed for the distribution cabinet determined to be in this stage, and then passes it to step S4 along with the branch affiliation, the upper and lower limits of the active power of the branch and the incomplete control records.

[0019] Each cluster simulation uses the measured power and control records at the same sampling time to simulate the continuation and cessation of reverse control. Power changes in the two simulations are used for their respective simulations, without overwriting the measured power or directly triggering the load controller. The branch layer calculates active power periodically according to the current power change rate, and remains unchanged after reaching the target control power in the corresponding simulation. The distribution cabinet layer uses the current measured total active power as a starting point, and adds the active power changes of each branch relative to the current measured value to obtain the total active power of the distribution cabinet at each simulation sampling time. The cluster layer calculates the load rate, load distribution, and power fluctuation accordingly. This preserves the initial difference between the incoming line metering value and the branch metering value, and avoids double-counting of completed control power.

[0020] like Figure 1 As shown, the overall control process of this embodiment includes the following steps: S1, obtaining the total active power of the distribution cabinet, the active power of the load branch, the target control power, the control issuance time, the execution feedback and the control direction, determining the remaining control power and the expected end time corresponding to the previous round of control and the reverse control for different load branches in the same distribution cabinet, and combining the change in active power already generated by the load branch corresponding to the reverse control to form a bidirectional control action sequence.

[0021] like Figure 2As shown, step S1 further includes steps S101 to S103. Step S101: Based on the execution feedback, determine the execution status of the previous round of control, and obtain the active power change process based on the sampling values ​​of the branch power metering device. When the current active power of the load branch has not exceeded the target control power and the direction of active power change is consistent with the control direction, determine the absolute value of the power difference between the current active power and the target control power that has not yet been completed as the remaining control power. Step S102: Based on the active power change within the continuous sampling period of the load branch, determine the current power change rate, and determine the expected end time based on the remaining control power and the current power change rate. Step S103: When the previous round of control involves multiple load branches, compare the expected end times of each load branch, and take the latest end time as the expected end time corresponding to the previous round of control.

[0022] like Figure 2 As shown, the process for determining the remaining control power and the expected end time is as follows: The power metering device on the incoming side of each distribution cabinet obtains the total active power of the distribution cabinet, and the power metering device on each load branch obtains the active power of the corresponding load branch. The load controller returns the execution feedback. The load branch, control direction, target control power, and control issuance time are recorded in the control record, and the active power of the branch, the total active power of the distribution cabinet, and the execution feedback are obtained according to a unified sampling period.

[0023] For the load branch corresponding to the previous round of regulation, the execution feedback is used to indicate when the load controller starts execution, regulation is completed, action is interrupted, and effect is stopped. The change in active power of the load branch is used to determine whether the regulation effect has actually occurred. When the execution feedback confirms that the corresponding load controller has started execution, and the active power of the load branch changes along the direction of the previous round of regulation at adjacent sampling times, it is determined that the load branch has generated an actual regulation effect. The target regulation power is the active power that the corresponding load branch should reach after the completion of this round of regulation. When the current active power of the load branch has not exceeded the target regulation power and the direction of active power change is consistent with the direction of the previous round of regulation, the current active power of the load branch and the target regulation power corresponding to the previous round of regulation are read. The absolute value of the power difference between the two is determined as the remaining regulation power of the load branch, and the regulation direction is recorded separately. For multiple load branches involved in the previous round of regulation within the same distribution cabinet, the remaining regulation power of each load branch in the same regulation direction is accumulated to obtain the remaining regulation power corresponding to the previous round of regulation.

[0024] The absolute value of the difference between the current active power and the target controlled power of the load branch does not exceed the upper limit of the active power error of the branch's energy metering device, and after the feedback confirmation is completed, the remaining controlled power is treated as completed. The upper limit of the active power error is determined according to the metering device's technical documents and actual measurement conditions. If the current active power exceeds the target controlled power, the direction of change is inconsistent with the recording direction, or the feedback indication action is interrupted, the calculation will not continue along the original direction; it will be recalculated after obtaining continuous sampling values ​​consistent with the effective controlled record.

[0025] Subsequently, for each load branch where the target controlled power and control direction remain unchanged, the most recent consecutive effective sampled value within the determined continuous sampling period is taken. The ratio of the absolute value of the difference between adjacent active power to the sampling time interval is calculated, and the arithmetic mean of these ratios is taken as the current power change rate. The direction of active power change in each window used should be consistent with the corresponding control direction. The expected end time of each load branch is determined according to the following formula: ; In the formula, For the first The expected end time of each load branch; This is the current sampling time; For the first The remaining control power of each load branch; For the first The absolute value of the rate of change of current power in each load branch; This refers to the load branch number involved in the previous round of regulation. The active power of the load branch is in the same power unit. It is expressed as the change in power per unit time, therefore This indicates the duration required for the corresponding load branch to complete the remaining regulation power. For multiple load branches, take each... The latest time in the calculation is taken as the expected end time corresponding to the previous round of regulation; when only one load branch is involved, the expected end time of that load branch is directly used. If the absolute value of the difference between the current active power and the target controlled power does not exceed the upper limit of the active power error of the branch power metering device, and the feedback confirmation of control completion is executed, then it is treated as a completed control; otherwise, the record of incomplete control is retained, the estimated end time of the current cycle is uncertain, and it is recalculated in the next effective sampling cycle. If the estimated end time of any branch participating in the comparison cannot be determined, the power prediction based on that time is not executed, and the missing time is not treated as completed.

[0026] For the next round of regulation, when its regulation direction is opposite to that of the previous round, it is defined as reverse regulation. After the execution feedback confirms the start of reverse regulation and the active power of the corresponding load branch changes along the reverse regulation direction, the active power sample value of the corresponding load branch at the time of regulation issuance is read and compared with the current active power to obtain the amount of active power change generated by the reverse regulation. The remaining regulation power corresponding to the previous round of regulation is recorded as the remaining active power change according to its regulation direction, and the active power change generated by the reverse regulation is recorded according to its regulation direction, arranged sequentially according to the sampling time to form a bidirectional regulation action sequence. Thus, even if the total active power change of the distribution cabinet is small, the active power changes corresponding to the two regulation directions can be obtained separately.

[0027] In this embodiment, the previous round of control and reverse control act on different load branches within the same distribution cabinet. Each branch only participates in the calculation based on the control record currently executed by the load controller. When the same branch receives replacement control, the calculation is still based on the original record before the feedback confirmation of replacement is executed. After the replacement is confirmed, the original record ends and a new record is established based on the actual measured active power at the time of replacement. The difference in remaining control power is not calculated between different records.

[0028] like Figure 1 As shown, after completing the correlation analysis of the previous round of regulation and reverse regulation, the following steps are performed: S2. Based on the bidirectional regulation sequence, the change in the total active power of the distribution cabinet is compared with the change in active power corresponding to the two regulation directions to determine whether the distribution cabinet is in the false stability stage of regulation cancellation.

[0029] like Figure 1 As shown, in step S2, within the historical valid records where the distribution cabinet has not performed load control and the power supply connection and load branch affiliation remain consistent, the difference in adjacent total active power with the same current unified sampling period is obtained. The upper limit of this difference is superimposed with the upper limit of the active power error of the adjacent sampling difference on the incoming side, serving as the upper limit of the preset power fluctuation range. The upper limit of the single active power error is determined based on the metering device's technical documents and actual measurement conditions; the upper limit of the error of adjacent sampling differences is the sum of the upper limits of the errors of two consecutive samplings; the upper limit of the error after summing multiple branches is the sum of the upper limits of the corresponding branch difference errors. The historical limit is used to characterize the actual power fluctuations that exist when the distribution cabinet has not performed load control, while the upper limit of the active power error is used to cover the metering deviation in the current adjacent sampling comparison. Both are used together to determine whether the change in the total active power of the distribution cabinet is within the preset power fluctuation range.

[0030] Based on the bidirectional control sequence, within adjacent sampling intervals where the branch affiliation, participating branches, and target control power remain unchanged, the remaining control power at the previous moment of the previous control round is subtracted from the current remaining control power. Similarly, the change in current active power generated during reverse control is subtracted from the change in active power generated at the previous moment, and these changes are recorded separately in their respective directions. The absolute values ​​of the changes in the two directions are summed to obtain the change in control power within the distribution cabinet. The absolute value of the difference in the total active power of the distribution cabinet within the same interval is taken as the change in the total active power of the distribution cabinet. When the control record changes, a new sequence is formed using the measured power corresponding to the new record; the target control power adjustment is not counted as an actual power change.

[0031] When the two control directions of each adjacent interval within a continuous sampling window maintain opposite power changes, the absolute values ​​of the changes in both directions are greater than the corresponding upper limit of the summarization error, the change in control power within the distribution cabinet is greater than the change in the total active power of the distribution cabinet, and the change in the total active power is within the preset power fluctuation range, the control cancellation false stability stage is determined. The number of continuous sampling periods is the number determined during model construction and remains unchanged during the comparison process.

[0032] like Figure 3 As shown, the predicted active power of the distribution cabinet after the cancellation is determined based on the false stability phase of the control cancellation, specifically including: S3, based on the false stability phase of the control cancellation, comparing the expected end times corresponding to the two control directions, determining the control direction that ends first, and determining the change in the remaining active power corresponding to the end time of the other control direction, so as to obtain the predicted active power of the distribution cabinet after the cancellation is determined.

[0033] like Figure 3 As shown, step S3 further includes steps S301 to S303. Step S301: The expected end time corresponding to the control direction that ends first is taken as the power change correction time; Step S302: Based on the expected active power changes generated by the two control directions between the current time and the power change correction time, the total active power of the current distribution cabinet is corrected to obtain the predicted active power of the distribution cabinet corresponding to the power change correction time; Step S303: Combining the remaining active power change of the other control direction after the power change correction time, the predicted active power of the distribution cabinet after the cancellation is obtained.

[0034] Specifically, the expected end time of the previous round of regulation is determined according to step S1; the same method is used to determine the expected end time of the reverse regulation. The two expected end times are compared, and the earlier one is taken as the power change correction time, and the corresponding regulation direction is determined as the regulation direction that ends first.

[0035] When the expected end times for two directions are the same, the predicted active power of the distribution cabinet at the common end time is calculated based on the process of each branch reaching the target control power, and the remaining changes after the common end time are no longer included. When a direction contains multiple branches, its latest expected end time indicates that the entire direction is completed, but does not mean that the branches have always completely canceled out before; the power maintenance process after each branch is completed ahead of schedule is still calculated branch by branch according to step S3.

[0036] For another control direction expected to end later, determine the active power change that has occurred from the time the control was issued to the current time for each branch. Then, calculate the expected change from the current time to the power change correction time based on the current power change rate; the latter should not exceed the current remaining control power. Add the two parts to obtain the cumulative active power change up to the power change correction time. Subtract this cumulative change from the absolute value of the difference between the active power sample value at the time the control was issued and the target control power to obtain the remaining active power change of each branch after the power change correction time. Then, accumulate the changes for the same control direction. Each branch is first limited before being aggregated to avoid completed branches continuing to contribute to the expected change.

[0037] Subsequently, between the current sampling time and the power change correction time, each controlled load branch extends according to its current power change rate, and after reaching the target controlled power, it no longer increases the expected power change; load branches that did not participate in the previous round of control and reverse control remain unchanged with their current active power within this prediction interval. The predicted active power of the distribution cabinet corresponding to the power change correction time is determined according to the following formula: ; In the formula, This represents the total active power of the distribution cabinet at the current sampling time. This is the current sampling time; The time for power change correction; in this formula Traverse all load branches involved in the previous round of control and reverse control within the distribution cabinet, counting each branch only once; For the first The remaining control power that has not yet been completed in each load branch is taken as a non-negative value; This is the absolute value of the current rate of change of power in this load branch; To determine the control direction coefficient for the corresponding branch, when increasing the total active power of the distribution cabinet, take... When reducing the total active power of the distribution cabinet, take Minimum value calculation is used to ensure that the expected change in active power does not exceed the remaining control power of the corresponding branch. All active power values ​​use the same unit, the time unit is consistent with the time unit in the power change rate, and the direction coefficient is dimensionless.

[0038] Taking into account the remaining active power change in the other control direction after the power change correction time, the predicted active power of the distribution cabinet when the other control direction is completed after the cancellation is achieved is obtained, and determined according to the following formula: ; In the formula, To offset the predicted active power of the distribution cabinet when the other control direction is completed after the cancellation; Predicted active power of the distribution cabinet at the time of power change correction; To adjust the directional coefficient, when another control direction is used to increase the total active power of the distribution cabinet, Pick When another control direction is used to reduce the total active power of the distribution cabinet, Pick ; This represents the change in residual active power after the power change correction time in the other control direction; where, , and Using the same active power unit, It is a dimensionless quantity.

[0039] like Figure 4 As shown, the solid red line represents the total active power of the distribution cabinet, the dashed blue line represents the remaining control power of the previous round, that is, the power change that has not yet been completed in the previous round of control; the dotted green line represents the cumulative power increment in the reverse direction, that is, the active power change that has been generated after the reverse control is issued. This refers to the time when the previous round of regulations was issued. For the timing of reverse regulation issuance, This refers to the time when the previous round of regulation was completed and the offsetting effects were lifted. This is the time when reverse regulation is completed. In this example, the active power of other loads remains unchanged. The previous round of regulation was used to reduce the total active power of the distribution cabinet, and the reverse regulation is used to increase the total active power of the distribution cabinet. to During this period, the total active power decreased following the previous round of regulation; to During this period, the reduction in remaining control power corresponding to the previous round of regulation is close to the increase in active power change already generated by reverse regulation. The actual power changes in the two directions cancel each other out, and the total active power change remains within the preset power fluctuation range. to During this period, the remaining control power corresponding to the previous round of control is zero, and the reverse control continues to increase the power. The total active power increases as the amount of active power change generated by the reverse control increases. correspond The predicted active power at time 1, after taking into account the change in remaining active power after that time, yields... Predicted active power at time .

[0040] like Figure 1 As shown, after obtaining the predicted active power of the distribution cabinet after the cancellation is resolved, the following cluster energy efficiency simulation process is performed: S4, input the predicted active power of the distribution cabinet into the digital twin model to perform cluster energy efficiency simulation, determine the load distribution of each distribution cabinet according to the simulation results, and control the corresponding load branch.

[0041] Specifically, the digital twin model records the upper limit of the allowable active power of the distribution cabinet, the load branch assignment, and the upper and lower limits of the active power of the adjustable load branches. The upper limit of the allowable active power of the distribution cabinet is determined based on the equipment operating limits; the upper and lower limits of the active power of the adjustable load branches are determined based on the intersection of the equipment's allowable adjustment range and the active power range that meets the current production requirements. Branches that were not regulated in the previous round maintain their original target regulation power in load allocation, non-adjustable branches maintain their projected active power, and the remaining adjustable branches use the target regulation power as the solution variable. The load rate is defined as the ratio of the total active power of each distribution cabinet to the upper limit of the allowable active power of the distribution cabinet. The objective is to minimize the sum of the squares of the differences between each load rate and its arithmetic mean. Under the above constraints, the load allocation is solved. When the current production requirement is to maintain the active power demand of the cluster unchanged, the upper and lower limits of the active power of each load branch are summarized and the intersection with the active power range that meets the current production requirements is taken to determine the allowable total active power range of the cluster. In both simulations, the sum of the active power of each load branch remains within the allowable total active power range of the cluster, avoiding the need to achieve load balance by changing the power consumption tasks.

[0042] The aforementioned secondary planning uses the difference between the current measured total active power of the distribution cabinet and the current measured active power of the branches participating in the control as a fixed part, and then includes the active power of these branches after the control is completed, forming the load rate calculation formula. Branches that did not complete control in the previous round use the original target control power, while branches under reverse control determine fixed values ​​or adjustable ranges according to the corresponding deduction method. Other adjustable branches are solved within their active power upper and lower limits. When multiple optimal allocations exist, the allocation with the smallest sum of squares of adjustment amounts relative to the current target control power for each adjustable branch is prioritized. The predicted active power in step S3 is used for deduction when the original target control power continues to be executed; after the target control power changes, the branch change amount is recalculated according to the new target control power, without superimposing the original predicted active power again. The power change rate of newly participating branches is determined based on the control records of the most recent control with the same control direction, unchanged target control power, and continuous and effective execution feedback; if there are no available records, no new adjustments will be arranged for that branch in this cycle.

[0043] This embodiment reduces unnecessary reverse regulation by identifying bidirectional control cancellation, predicting power changes after cancellation removal, and comparing feasible load allocations. Load factor balancing is used to evaluate the load distribution of the distribution cabinet cluster and is not directly used as evidence of reduced total power consumption or line losses.

[0044] For distribution cabinets in the pseudo-steady phase of control offsetting, their projected active power is determined based on whether reverse control continues or stops. For other distribution cabinets, the real-time total active power is used as the initial projected value, and the changes in active power caused by incomplete control and current load allocation are taken into account. The two projections maintain consistency in the power supply connection, load branch affiliation, and upper and lower limits of active power for each distribution cabinet. When reverse control stops, the branch power change before the stop action takes effect is calculated first, and then the branch power after the stop action is fixed at the active power expected at the time of the stop action. When reverse control continues, the corresponding load branch is allowed to adjust the target control power within the original reverse control direction and the upper and lower limits of the branch active power.

[0045] Both simulations begin from the current sampling time. When reverse control continues, the simulation is performed according to the active power change determined in step S3. When reverse control stops, the reverse power change before the stop action takes effect and the power change not yet completed in the previous round are retained. For branches that need to change the target control power this time, the original target control power is used for calculation before the action takes effect, and the remaining control power is updated according to the new target control power after the action takes effect. The calculation is performed according to the change rate and target limit method in steps S1 and S3 to avoid double-counting the remaining change of the original target control power. The expected effective delay is taken as the upper limit of the effective time interval from issuance to execution feedback confirmation in the historical record of the controller corresponding to this control. The expected stop effective time is determined accordingly. If there is no available execution record, or the controller does not have the ability to stop continued adjustment and maintain power, the stop execution is not considered as an executable plan for this cycle. After the actual issuance, the effective time and maintained power are updated with execution feedback and measured power.

[0046] Both simulations start from the current sampling time and calculate the latest expected completion time for all participating control branches under both scenarios; branches that complete ahead of schedule maintain their power at completion. The total active power of each distribution cabinet is calculated using a uniform sampling period, and power is also calculated at the expected effective time of control and the expected time each branch reaches its target control power. When each branch extends in segments according to its rate of change, the upper limit of the allowable active power of the distribution cabinet, the upper and lower limits of the active power of the branch, and the allowable range of the total active power of the cluster are checked at these turning points and interval endpoints. After the upper limit of the allowable active power of the distribution cabinet, the upper and lower limits of the active power of the branch, and the allowable range of the total active power of the cluster are all satisfied, the absolute value of the difference in the total active power of the distribution cabinet at each adjacent simulation sampling time is then judged to be within the preset power fluctuation range, in order to limit the change in the total active power of the distribution cabinet at adjacent sampling times during the simulation. Only when all conditions are met can the candidate scheme be executed. If the secondary programming has no feasible solution, lacks necessary expected times, or the current candidate does not meet the above conditions, the candidate shall not be judged as meeting the requirements; the current candidate not meeting the above conditions does not mean that there are no other feasible allocations.

[0047] When the candidate scheme corresponding to the cessation of reverse regulation meets the above conditions, in order to reduce redundant regulation, this scheme is executed first, the corresponding load branch is stopped from continuing to adjust along the reverse regulation direction, and the corresponding target regulation power is issued to other branches participating in the adjustment in this scheme; if only the candidate scheme corresponding to the continued execution of reverse regulation meets the above conditions, the target regulation power of all branches participating in the adjustment in this scheme is executed. If neither of the two candidate schemes meets the above conditions, the following handling method for those that have not passed the deduction is followed, and steps S1 to S4 are re-executed after updating the bidirectional regulation action sequence in the next effective sampling period.

[0048] If both candidate schemes fail, the power process of continuing to execute the original target control power for each branch will be checked again. If the original target control power still meets the above conditions, the original target control power will be maintained. If the original target control power is expected to exceed the limit or lacks necessary input and cannot be verified, the addition of new energy efficiency optimization control will be stopped, and the corresponding distribution cabinet, branch, and unmet condition prompts will be output. The existing protection and manual handling will be used to handle the anomaly. The model will continue to collect and execute feedback, but this situation will not be recorded as optimization completion.

[0049] Finally, based on the selected scheme, the load controllers of each participating branch are located, and corresponding control commands are issued to either control the target control power or stop further adjustment. The issuance time and control direction are recorded. Reverse control stop execution means stopping the power change along the original reverse control direction. After taking effect, operation continues according to the stop-effective power fed back by the controller; this operation does not mean cutting off the branch power. Each branch continues to be included in its original control effect from issuance to actual effectiveness; issuance is not equivalent to achieving the target. In the next sampling cycle, the control record is updated with the measured power and execution feedback, the remaining control power, power change rate, and expected end time are recalculated, and steps S1 to S4 are executed again.

[0050] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the present invention's digital twin-based power distribution cabinet cluster energy efficiency optimization control method and its inventive concept, should be covered within the scope of protection of the present invention.

Claims

1. A method for optimizing the energy efficiency control of a distribution cabinet cluster based on digital twins, characterized in that, Includes the following steps: The system obtains the total active power of the distribution cabinet, the active power of the load branch, the target control power, the control issuance time, the execution feedback and the control direction, determines the remaining control power corresponding to the previous round of control for different load branches in the same distribution cabinet and the expected end time of the two control directions, and forms a two-way control action sequence by combining the active power change generated by the reverse control. When the change in the total active power of the distribution cabinet is within the preset power fluctuation range, and the active power change directions corresponding to the two control directions are opposite, the control cancellation false stability stage is determined. Compare the expected end times of the two control directions, take the earlier expected end time as the first end time, and obtain the predicted active power of the distribution cabinet after the cancellation based on the change in the remaining active power of the other control direction after the first end time. In the digital twin model, cluster energy efficiency simulations are performed for both the continuation and cessation of reverse regulation. Based on the simulation results, the target regulation power is determined, and the corresponding load branches are controlled.

2. The method for optimizing energy efficiency control of a distribution cabinet cluster based on digital twins according to claim 1, characterized in that, Methods for obtaining the predicted active power of the distribution cabinet after cancellation include: The expected end time corresponding to the first ending control direction is taken as the power change correction time; Based on the expected change in active power from the current time to the power change correction time in the two control directions, the total active power of the current distribution cabinet is corrected to obtain the corresponding predicted active power of the distribution cabinet. By combining the change in residual active power after the power change correction time in the other control direction, the predicted active power of the distribution cabinet after the cancellation is obtained.

3. The method for energy efficiency optimization control of a distribution cabinet cluster based on digital twins according to claim 2, characterized in that, The determination of the change in residual active power includes: Based on the measured active power changes and current power change rates of each load branch in the other control direction since the control was issued, the cumulative active power change at the power change correction time is determined. The remaining active power change of each branch is obtained by subtracting the absolute value of the difference between the active power sample value corresponding to the control of each load branch and the target control power, and then accumulating them in the same control direction. The cumulative active power change does not exceed the absolute value.

4. The method for optimizing the energy efficiency control of a distribution cabinet cluster based on digital twins according to claim 1, characterized in that, The determination of remaining control power and expected termination time includes: The remaining control power of the load branch corresponding to the previous round of control is determined based on the execution feedback, the sampled value of the branch power metering device, and the target control power. The current power change rate is determined based on the active power change in the most recent continuous effective sampling period of the load branch, and the expected end time is determined based on the remaining control power and the current power change rate; if the current power change rate is zero and the control process has not been completed, the record of incomplete control is retained and recalculated in the next effective sampling period. When multiple load branches are involved, the latest expected end time is taken as the expected end time of the previous round of regulation.

5. The method for energy efficiency optimization control of a distribution cabinet cluster based on digital twins according to claim 1, characterized in that, The formation of the bidirectional regulatory sequence includes: The remaining control power corresponding to the previous round of control is recorded as the change in remaining active power according to the control direction; After the feedback confirmation of the start of reverse control and the change of active power in the load branch along the reverse control direction, the amount of active power change that has occurred is determined based on the active power sample value corresponding to the control issuance and the current active power. Record the changes in remaining active power and the changes in active power already generated according to the control direction and sampling time to form a two-way control sequence.

6. The method for energy efficiency optimization control of a distribution cabinet cluster based on digital twins according to claim 1, characterized in that, The determination of the adjustment offsetting the false stability phase includes: The active power change in the two control directions within the adjacent sampling time interval is determined based on the bidirectional control action sequence, and the absolute values ​​are accumulated to obtain the control power change inside the distribution cabinet. When the active power changes corresponding to the two control directions are opposite, the change in control power inside the distribution cabinet is greater than the change in total active power of the distribution cabinet, and the change in total active power of the distribution cabinet is within the preset power fluctuation range during the continuous sampling period, the control cancellation false stability stage is determined.

7. The method for optimizing energy efficiency control of a distribution cabinet cluster based on digital twins according to claim 6, characterized in that, The number of consecutive sampling periods is determined by rounding down the ratio of the shortest control duration in the historical normal control record where the target control power and control direction remain unchanged, the execution feedback is continuously effective, and continuous adjacent comparisons can be formed to a uniform sampling period, and the sampling values ​​covering the consecutive sampling periods are used to form adjacent sampling differences. The absolute values ​​of the active power changes corresponding to the two control directions are greater than the sum of the upper limits of the active power error of the adjacent sampling differences of the load branches involved in each control direction.

8. The method for energy efficiency optimization control of a distribution cabinet cluster based on digital twins according to claim 1, characterized in that, The change in total active power of the distribution cabinet is the absolute value of the difference in total active power of the distribution cabinet at adjacent sampling times. The upper limit of the preset power fluctuation range is the sum of the historical upper limit of the absolute value of the difference in total active power between adjacent samples in the historical valid records where the distribution cabinet has not performed load control and the power supply connection and load branch affiliation are consistent, and the upper limit of the active power error of the difference between adjacent samples on the incoming line side.

9. The method for energy efficiency optimization control of a distribution cabinet cluster based on digital twins according to claim 1, characterized in that, Cluster energy efficiency simulation includes: For distribution cabinets in the pseudo-stable phase of regulation offsetting, the active power is calculated and extrapolated according to whether the reverse regulation continues to be executed or stops; for other distribution cabinets, the real-time total active power is used as the initial value, and the changes in active power due to incomplete regulation and the current load distribution are included. When reverse control continues to be implemented, the total active power of the current distribution cabinet is used as the basis for calculation based on the predicted active power of the distribution cabinet after the offset is removed; when reverse control stops, the total active power of the current distribution cabinet is used as the basis for calculation, taking into account the changes in active power before the stop action takes effect and the changes in active power that have not yet been completed in the previous round of control. The two simulations keep the power supply connection of the distribution cabinet and the load branch affiliation unchanged. The load allocation is determined within the upper limit of the allowable active power of the distribution cabinet and the upper and lower limits of the active power of the branch. The power fluctuation is judged according to the preset power fluctuation range.

10. The method for energy efficiency optimization control of a distribution cabinet cluster based on digital twins according to claim 9, characterized in that, When the reverse control stops executing the corresponding deduction to meet the upper limit of the allowable active power of the distribution cabinet, the upper and lower limits of the active power of the branch and the preset power fluctuation range, the corresponding load branch stops adjusting the active power along the reverse control direction, and the target control power of the participating branches other than the corresponding load branch is executed. Otherwise, when the reverse regulation continues to execute the corresponding deduction and the above conditions are met, the target regulation power of the load branch participating in the regulation is adjusted according to the load allocation. If the above conditions are not met in both simulations, and the above conditions are met when the original target control power is continued to be implemented, then the original target control power will be maintained; otherwise, no new energy efficiency optimization control will be issued, and the simulation will be repeated in the next effective sampling period.