Power distribution network multi-target feeder network reconstruction optimization method and device considering distributed power supply, equipment and storage medium

Abstract

The invention discloses a power distribution network multi-target feeder network reconstruction optimization method and device considering a distributed power supply, equipment and a storage medium, and the method comprises the steps: triggering a multi-mode reconstruction process through presetting an optimization target, a time scale and a safety constraint parameter; integrating real-time data and prediction data into a unified analysis section by adopting a dynamic weighted fusion algorithm; on the basis of load rate calculation and state evaluation results, a load unbalance degree, network loss and a cost function of operation times are considered to search a transfer path, a weight coefficient is dynamically adjusted, and an optimization strategy is generated; and generating a strategy combination through loop iteration, and executing and generating a report after multi-layer security check. According to the power distribution network multi-target feeder network reconstruction optimization method and device considering the distributed power supply, the equipment and the storage medium provided by the invention, multi-target optimization of load balance and line loss reduction of the transformer and the line is realized on the premise of meeting security constraints, and the economy and reliability of power distribution network operation are improved.

Description

TECHNICAL FIELD

[0001] The application relates to a power distribution network multi-objective feeder network reconstruction optimization method, device and equipment considering distributed power sources and a storage medium, and belongs to the technical field of power distribution network operation and optimization. BACKGROUND

[0002] As the terminal link of a power system, a power distribution network is directly related to the reliability and economy of user power supply. With high penetration rate of distributed power sources (photovoltaic, energy storage, etc.) and the complication of load characteristics, the operation state of the power distribution network presents strong dynamics, and traditional reconstruction methods based on a single target (such as only reducing network loss) or static topology have been difficult to meet the demand.

[0003] 1. Single target limitation: existing methods focus on minimizing the loss of the global power grid, ignoring the load balancing of specific transformers and line devices, which easily leads to local device overload and reduces power supply reliability.

[0004] 2. Insufficient dynamic adaptability: the influence of unbalanced three-phase load and asymmetric line parameters is not fully considered, and the adaptation to the characteristics of the distributed power source (such as photovoltaic output fluctuation and energy storage charging and discharging strategy) is poor, and the accuracy of power flow calculation is insufficient.

[0005] 3. Low decision efficiency: the transfer path search relies on artificial experience or fixed algorithms, and does not combine real-time load prediction and dynamic topology adjustment, which leads to a disconnection between the reconstruction strategy and the actual operation trend, shortens the timeliness of the optimization strategy, and increases the operation cost.

[0006] Therefore, there is an urgent need for a reconstruction method that can integrate multi-objective optimization, dynamic topology analysis and safety constraint verification to improve the economy and reliability of power distribution network operation. SUMMARY

[0007] Objective: In order to overcome the shortcomings in the prior art, the application provides a power distribution network multi-objective feeder network reconstruction optimization method, device, equipment and storage medium considering distributed power sources.

[0008] Technical solution: In order to solve the above technical problems, the technical solution adopted by the application is:

[0009] In a first aspect, a power distribution network multi-objective feeder network reconstruction optimization method considering distributed power sources is provided, which specifically includes:

[0010] Step 1: preset the core parameters of the power distribution network to be reconstructed and optimized.

[0011] Step 2: based on the core parameters, when the multi-mode trigger mechanism is triggered to enter step 3.

[0012] Step 3: Based on the core parameters, the future power data of the to-be-reconstructed optimal distribution network is predicted to obtain future prediction data, real-time power data of the to-be-reconstructed optimal distribution network is obtained as real-time data, a cross section corresponding to a time period is intercepted from the real-time data and the future prediction data to form a continuous multi-time cross section sequence, and the multi-time cross section sequence is weighted and fused to obtain a unified analysis cross section.

[0013] Step 4: Under the unified analysis cross section, the load rate of the to-be-reconstructed optimal distribution network is calculated, the load balancing degree, the network loss rate and the risk of exceeding the limit are evaluated according to the load rate of the to-be-reconstructed optimal distribution network, and when the load balancing degree, the network loss rate and the risk of exceeding the limit meet the iteration starting condition, step 5 is entered, otherwise step 3 is returned.

[0014] Step 5: Based on the network topology model of the to-be-reconstructed optimal distribution network, a transfer path is searched, based on the core parameters, the power flow and the network loss of the transfer path are calculated, and a path cost function under multiple optimization objectives is constructed, the path cost function is solved to obtain the load imbalance degree, the number of switch operations and the network loss reduction of the transfer path.

[0015] Step 6: When the load imbalance degree, the number of switch operations and the network loss reduction of the transfer path meet the iteration termination condition, the optimized distribution network is output, otherwise step 5 is returned.

[0016] Optionally, it further includes: step 7: performing safety checking, strategy execution and report generation according to the optimized distribution network.

[0017] Optionally, the core parameters include: optimization objective combination, time scale parameter, iteration constraint and safety constraint.

[0018] Optionally, the multi-mode triggering mechanism includes: timing scanning triggering, state prediction triggering and manual operation triggering.

[0019] Optionally, the step 3 specifically includes:

[0020] The combination model of the LSTM neural network and the grey prediction model is adopted to predict the future power data of the to-be-reconstructed optimal distribution network to obtain the future prediction data.

[0021] The number of cross sections is calculated according to the time scale parameter of the core parameters.

[0022] The real-time power data of the to-be-reconstructed optimal distribution network is obtained as real-time data, a cross section corresponding to a time period is intercepted from the real-time data and the future prediction data to form a continuous multi-time cross section sequence.

[0023] The multi-time section sequence is integrated into a unified analysis section by a weighted fusion algorithm, wherein the weight distribution rule of the weighted fusion algorithm is that the weight of real-time data is 0.6-0.8, the weight of short-term prediction data is 0.4-0.6, and the weight of medium-term prediction data is 0.2-0.4.

[0024] Optionally, the load rate of the power distribution network comprises a transformer load rate and a line load rate.

[0025] The over-limit risk comprises an over-limit state, wherein the over-limit state is determined when the load rate of the target device is greater than a preset threshold value and the duration is greater than a preset time length.

[0026] The load balancing degree comprises a load imbalance degree, wherein the load imbalance degree The expression is as follows:

[0027]

[0028] wherein, is the load rate of the i-th device in the analysis range, is the average load rate of all devices in the analysis range, is the number of devices.

[0029] The network loss rate comprises a ratio of section network loss to total network loss, wherein the total network loss The expression is as follows:

[0030]

[0031] wherein, is the feeder segment active loss, is the transformer no-load loss, is the transformer load loss.

[0032] The iteration starting condition comprises that when the over-limit state, the load imbalance degree and the network loss rate satisfy any one of the following conditions, the iteration is started.

[0033] i. There is an over-limit state main transformer or line.

[0034] ii. The load imbalance degree of the main transformer or line is greater than the iteration constraint in the core parameter.

[0035] iii. The network loss rate is higher than the historical same period benchmark value or the preset target value.

[0036] Optionally, the step 5 specifically comprises:

[0037] ​The network topology model of node-edge structure is constructed based on the to-be-reconstructed optimization power distribution network equipment model data, wherein the node is set as a node-type equipment, the edge is set as a connection-type equipment, the association relationship between the node and the edge is used to map the electrical connection state of the actual power grid in real time, and the switch state, the equipment parameter and the normalized load data are synchronized in real time, so as to ensure that the network topology model is consistent with the actual power grid state.

[0038] Based on the radial topology characteristics of the network topology model, the depth-first search is preferentially used, and the breadth-first search is used as an auxiliary to traverse the network node, the feeder load in the to-be-analyzed area is taken as a starting point, the node-type equipment is searched to the other side, and the dynamic association relationship of “node-power supply” is formed.

[0039] The switch whose current state is open and whose two sides belong to different “node-power supply” association relationships is a tie switch, and the original power trunk path equipment information and the transfer power trunk path equipment information are recorded synchronously.

[0040] The original power trunk path equipment and the transfer power trunk path equipment are subjected to feasibility verification, and the transfer path in which the current state of the node-type equipment is normal or the associated protection, fault or blocking action is normal and the tie switch has a remote control function is obtained.

[0041] The transfer path that meets the transfer margin constraint is selected from the transfer path.

[0042] The power flow and the network loss of the transfer path that meets the transfer margin constraint are calculated, and the path cost function under multiple optimization objectives is constructed, wherein the expression of the path cost function is as follows:

[0043]

[0044] Wherein, is a load imbalance improvement value, is a network loss reduction amount, is a switch operation number, is a weight coefficient.

[0045] The path cost function is solved, the solutions of the path cost function of the transfer path are sorted, the transfer path that violates the safety constraint in the core parameter is eliminated, the optimal solution is selected, the load imbalance degree, the switch operation number and the network loss reduction amount of the transfer path are selected, and the weight coefficient of the path cost function is dynamically adjusted.

[0046] Optionally, the iteration termination condition includes the following any case:

[0047] i. All optimization objective combinations in the core parameter are achieved.

[0048] ii. No new reconstruction optimization scheme is generated in the last iteration.

[0049] iii. The number of iterations reaches the upper limit of the iteration constraint configuration in the core parameters.

[0050] Optionally, the security check includes: node and permission check, guardian remote control check, topology error prevention check, remote control basic parameter check, network constraint check, and thermal conductivity safety check.

[0051] The strategy execution includes: pushing the final to-be-executed switch operation sequence to the operation ticket management application, supporting batch automatic or manual control issuance mode, and feeding back operation state information in real time during execution.

[0052] The report generation includes: automatically archiving and generating a reconstruction optimization report after successful execution, and supporting one-key export function.

[0053] In a second aspect, a power distribution network multi-objective feeder network reconstruction optimization device considering distributed power sources includes:

[0054] The parameter configuration module is configured to preset core parameters of the to-be-reconstructed and optimized power distribution network.

[0055] The trigger starting module is configured to enter the load prediction and data fusion module based on the core parameters when a multi-mode trigger mechanism is triggered.

[0056] The load prediction and data fusion module is configured to predict future power data of the to-be-reconstructed and optimized power distribution network based on the core parameters, obtain future prediction data, acquire real-time power data of the to-be-reconstructed and optimized power distribution network as real-time data, intercept cross sections of corresponding time periods from the real-time data and the future prediction data, form a continuous multi-time cross section sequence, and obtain a unified analysis cross section by weighted fusion of the multi-time cross section sequence.

[0057] The load rate calculation and state evaluation module is configured to calculate a load rate of the to-be-reconstructed and optimized power distribution network under the unified analysis cross section, evaluate a load balancing degree, a network loss rate, and an out-of-limit risk according to the load rate of the to-be-reconstructed and optimized power distribution network, and enter the dynamic topology optimization strategy generation module when the load balancing degree, the network loss rate, and the out-of-limit risk meet iteration starting conditions, or return to the load prediction and data fusion module.

[0058] The dynamic topology optimization strategy generation module is configured to search a transfer path based on a network topology model of the to-be-reconstructed and optimized power distribution network, calculate power flow and network loss of the transfer path based on the core parameters, construct a path cost function under multiple optimization objectives, solve the path cost function, and obtain a load imbalance degree, a switch operation number, and a network loss reduction amount of the transfer path.

[0059] The cyclic iteration analysis module is configured to output an optimized power distribution network when the load imbalance degree, the switch operation number, and the network loss reduction amount of the transfer path meet iteration termination conditions, or return to the dynamic topology optimization strategy generation module.

[0060] Optionally, further comprising: a security check and execution module, configured to perform security check, strategy execution and report generation according to the optimized power distribution network.

[0061] In a third aspect, a computer readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the method of any one of the first aspect.

[0062] In a fourth aspect, a computer device comprises:

[0063] a memory configured to store instructions.

[0064] a processor configured to execute the instructions to cause the computer device to perform operations of the method of any one of the first aspect.

[0065] Advantages: The method, device, equipment and storage medium for multi-objective feeder network reconstruction optimization of power distribution network considering distributed power supply provided by the application belong to the technical field of power distribution network operation and optimization, and a multi-objective feeder network reconstruction optimization method for power distribution network considering distributed power supply access is disclosed, which takes into account device load balancing, network loss reduction and operation cost. The method triggers a multi-mode reconstruction process by presetting optimization targets, time scales and safety constraint parameters. A dynamic weighted fusion algorithm is used to integrate real-time data and predicted data into a unified analysis section. Based on the load rate calculation and state evaluation results, a cost function considering the load imbalance degree, network loss and operation frequency is used to search for a transfer path, and the weight coefficient is dynamically adjusted to generate an optimization strategy. The strategy combination is generated through cyclic iteration, and the report is generated after the multi-layer safety check and execution. Under the premise of meeting the safety constraints, the method realizes multi-objective optimization of load balancing and line loss reduction of transformers and lines, and improves the economy and reliability of power distribution network operation. Compared with the prior art, the method has the following advantages:

[0066] 1. Multi-objective collaborative optimization: The dynamic weight cost function realizes the collaboration of load balancing, network loss reduction and operation cost, avoiding the limitations of single target.

[0067] 2. Strong dynamic adaptability: The real-time data and load prediction are integrated, and the dynamic topology modeling and phase-by-phase power flow calculation are combined to adapt to the distributed power supply access and three-phase unbalanced scenarios.

[0068] 3. Improved safety and reliability: The multi-layer safety check and transfer path feasibility check ensure that the reconstruction strategy does not have secondary over-limiting and reduces the operation risk.

[0069] 4. High engineering practicability: the method steps are clear, and the calculation formula and constraint parameters can be adjusted according to the characteristics of the power grid, facilitating the implementation by those skilled in the art. BRIEF DESCRIPTION OF DRAWINGS

[0070] Figure 1 A flowchart of a power distribution network multi-objective feeder network reconstruction optimization method considering distributed power sources in an embodiment of the present application.

[0071] Figure 2 A schematic diagram of a parameter configuration module in an embodiment of the present application.

[0072] Figure 3 A schematic diagram of a trigger starting module in an embodiment of the present application.

[0073] Figure 4 A flowchart of a load prediction and data fusion function in an embodiment of the present application.

[0074] Figure 5 A flowchart of a load rate calculation and state evaluation function in an embodiment of the present application.

[0075] Figure 6 A flowchart of a dynamic topology optimization strategy generation function in the present application.

[0076] Figure 7 A flowchart of a cycle iteration analysis function in an embodiment of the present application.

[0077] Figure 8 A flowchart of a load prediction and data fusion function in an embodiment of the present application.

[0078] Figure 9 A schematic diagram of a report generation module in an embodiment of the present application.

[0079] Figure 10 A schematic diagram of a power distribution network model structure in scenario 5 of an embodiment of the present application. DETAILED DESCRIPTION

[0080] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0081] The present application will be further described below with reference to specific embodiments.

[0082] Embodiment 1

[0083] This embodiment introduces a power distribution network multi-objective feeder network reconstruction optimization method considering distributed power sources, which is described with reference toOperation step , specifically comprising:

[0084] Step 1: Parameter configuration: preset core parameters of reconstruction optimization are stored in the system configuration library as the basis for subsequent trigger judgment, data processing and strategy generation.

[0085] Step 2: Trigger start: based on the configuration parameters, the reconstruction process is started through the multi-mode trigger mechanism.

[0086] Step 3: Load prediction and data fusion: predict future load data and fuse multi-period data to generate a unified analysis section.

[0087] Step 4: Load rate calculation and state evaluation: calculate the device load rate under the normalized data section, and evaluate the load balancing degree, network loss rate and out-of-limit risk.

[0088] Step 5: Dynamic topology optimization strategy generation: search for transfer paths based on network topology, calculate power flow and network loss, and generate reconstruction strategies.

[0089] Step 6: Recycle iteration analysis: repeat steps 4 to 5 until the iteration termination condition is met.

[0090] Step 7: Safety review and execution: review the strategy safety and execute and generate a report.

[0091] Further, the parameter configuration function in step 1 is mainly realized through the parameter configuration module in the implementation program, see Substation , specifically including the following configurable contents:

[0092] (a) Optimal target combination: based on the device load out-of-limit elimination as the basic target (transformer, line load rate ≤80%), additional targets can be selected as needed, including transformer load balancing (load imbalance degree ≤5%), line load balancing (load imbalance degree ≤10%), and network loss reduction (decrease ≥3%).

[0093] (b) Time scale parameters: future prediction time period is adjustable from 1 to 24 hours, and analysis time step is adjustable from 5 to 30 minutes, both of which are linked and adapted.

[0094] (c) Iteration constraints: the upper limit of optimization iteration times is adjustable from 10 to 20 times, the voltage deviation threshold of power flow calculation nodes is adjustable from 0.05 kV, and the upper limit of power flow calculation iteration times.

[0095] (d) Safety constraints: voltage deviation ≤±5%, transformer and line load rate ≤80%, and loop operation impact current ≤2 times rated value, all constraint values can be adjusted and configured by users according to the characteristics of the power grid.

[0096] Further, the trigger start function in step 2 is mainly realized by implementing a trigger start module in the program, see Line , and specifically includes the following optional trigger mode contents:

[0097] (a) Timing scanning trigger: automatically scan the system state according to a preset period (5-30 minutes adjustable), and trigger when it is detected that the measured value exceeds the optimization parameter, such as transformer or line load rate ≥ 80%, or load imbalance degree ≥ 10%, the scanning range can be configured as the whole network or a specified area.

[0098] (b) State prediction trigger: based on load prediction data, judge that the load rate ≥ 90% or the load imbalance degree ≥ 15% will occur in the future time period (30-90 minutes adjustable), and the prediction credibility needs to be ≥ 85%.

[0099] (c) Manual operation trigger: support dispatchers to manually start by selecting target area, trigger reason and optimization target through human-computer interface, and the trigger reason includes planned maintenance, load surge, power supply demand, etc.

[0100] Further, the load prediction and data fusion function in step 3 is mainly realized by implementing a load prediction and data fusion module in the program, see Operation device , and specifically includes the following contents:

[0101] (a) Load prediction: use a combination model of LSTM neural network and gray prediction model to generate future 1-24 hour load data, with a prediction error ≤ 8%, and the combination model weight can be dynamically updated according to historical prediction accuracy to improve the prediction adaptability in different scenarios.

[0102] (b) Multi-section construction: according to the time step configured in step 1, the corresponding section in the time period is intercepted from real-time data and future prediction data according to the rule "section number = total prediction time / time step", forming a continuous multi-time section sequence, and the section data includes node power, voltage and current parameters.

[0103] (c) Normalization processing: integrate multi-section data into unified analysis section through weighted fusion algorithm, and the weight distribution rule is: real-time data weight 0.6-0.8, short-term prediction data weight 0.4-0.6, medium-term prediction data weight 0.2-0.4, and the specific weight can be dynamically adjusted according to data credibility, which is negatively correlated with prediction error.

[0104] Further, the load rate calculation and state evaluation function in step 4 is mainly realized by implementing a load rate calculation and state evaluation module in the program, see Operation mode , and specifically includes the following contents:

[0105] (a) Load rate calculation:

[0106] i. Calculate the transformer load rate:

[0107]

[0108] Wherein, is the active value of the transformer winding, is the rated power of the transformer winding.

[0109] ii. Calculate the line load rate:

[0110]

[0111] Wherein, is the current value of the feeder load, is the current rating of the feeder load.

[0112] (b) State evaluation:

[0113] i. Determine the out-of-limit state:

[0114] When the load rate of the target device is greater than the preset threshold (80% adjustable), and the duration is greater than the preset time (10-30 minutes adjustable), it is determined that the out-of-limit state.

[0115] ii. Calculate the load imbalance degree:

[0116]

[0117] Wherein, is the load rate of the device in the analysis range, is the average load rate of all devices in the analysis range, is the number of devices.

[0118] Further, the dynamic topology optimization strategy generation function in step 5 is mainly realized by implementing the dynamic topology optimization strategy generation module in the program, see Operation result , specifically including the following contents:

[0119] (a) Dynamic topology modeling:

[0120] i. Based on the distribution network equipment model data, a network topology model of node-edge structure is constructed, the nodes include line, load and other node class devices, the edges include transformers, feeder sections, switches, circuit breakers and other connection class devices, and the association relationship between nodes and edges is real-time mapping of the actual electrical connection state of the power grid.

[0121] ii. Real-time synchronization of switch state (on / off), device parameters (including line resistance, reactance, device operating capacity), and normalized load (active, current) data to ensure that the topology model is consistent with the actual grid state.

[0122] (b) Transfer path search:

[0123] i. Feasible transfer path search: search for the target device's operating contact switch and corresponding opposite power supply device, backbone path device set, etc. under the current topology state.

[0124] The implementation process of the above feasible transfer path search specifically includes the following content:

[0125] 1. Dynamic power supply range analysis: based on the radial topology characteristics of the distribution network, preferentially use depth-first search (DFS) supplemented by breadth-first search (BFS) to traverse network nodes, take the feeder load in the analysis area as the starting point, search along the switch (considering the real-time state), feeder segment, etc. to the other side, forming a "node-power supply source" dynamic association relationship;

[0126] 2. Contact point identification: search for switches whose current state is open and whose nodes on both sides belong to different "node-power supply source" association relationships, and record the original power supply backbone path device information and the transfer power supply backbone path device information;

[0127] 3. Path feasibility check: the devices involved in the power supply path on both sides have no abnormal current state or associated protection, fault, or locking action, and the contact switch has remote control function.

[0128] ii. Path screening needs to meet the transfer margin constraint: the remaining capacity of the opposite device is greater than the amount of load to be transferred (does not cause the opposite device to be overrated again), and the operating device is a controllable (has remote control function, reasonable remote signaling state) switch device.

[0129] (c) Power flow and network loss calculation:

[0130] i. Power flow calculation considers three-phase load imbalance, line parameter asymmetry, and distributed resource (photovoltaic, energy storage, etc.) access, and ends when the node voltage deviation is less than the configured threshold (0.05 kV adjustable) or the iteration count reaches the upper limit value (50 times configurable).

[0131] The implementation process of the above power flow calculation specifically includes the following content:

[0132] 1. Node and parameter phase separation: three-phase loads are modeled separately according to A, B, and C phases, node injection power is divided into active and reactive power for each phase, line parameters consider resistance and reactance for each phase, and node admittance matrix is constructed separately for each phase.

[0133] 2. Distributed resource access rules: photovoltaic / wind power is accessed as a split-phase PQ node, active power is allocated according to the predicted value in step 3, and reactive power is adjusted within a certain active power range; energy storage is accessed as a split-phase PV node, and the voltage of each phase is maintained within a certain threshold range (±5% adjustable) of the rated value, and the active power is adjusted according to the charging and discharging plan;

[0134] 3. Power flow solution and verification: an improved forward-backward substitution method is used to calculate the voltage and current of each phase iteratively, and the convergence criterion is that the maximum deviation of three-phase voltage is less than a threshold value (0.05kV adjustable), and the result output includes phase current, phase voltage, phase active power, phase reactive power and other information.

[0135] ii. Network loss calculation is to calculate the active loss of each feeder section and the loss of each transformer (including no-load loss and load loss), and the total network loss is the sum of the losses of all devices.

[0136] Among them, the implementation process of the above network loss calculation specifically includes the following contents:

[0137] 1. Calculate the active loss of the feeder section:

[0138]

[0139] Where, is the phase current of the feeder section, is the resistance of the feeder section, is the loss calculation period (the time step configured in step 1);

[0140] 2. Calculate the transformer no-load loss:

[0141]

[0142] Where, is the rated no-load loss of the transformer, is the loss calculation period (the time step configured in step 1);

[0143] 3. Calculate the transformer load loss:

[0144]

[0145] Where, is the rated load loss of the transformer, is the current apparent power of the transformer (the result of the power flow calculation), is the rated capacity of the transformer, is the loss calculation period (the time step configured in step 1);

[0146] 4. Calculate the total network loss as the sum of the active loss of each feeder section and the no-load loss and load loss of all transformers, that is:

[0147]

[0148] (d) Calculate the path cost function under multi-optimization objectives according to the following formula:

[0149]

[0150] wherein, is the load imbalance improvement value, is the network loss reduction amount, is the number of switch operations, is the weight coefficient (dynamically allocated according to the optimization objective combination configured in step 1).

[0151] Further, the loop iteration analysis function in step 6 is mainly realized by implementing a loop iteration analysis module in the program, as shown in Success , and specifically includes the following contents:

[0152] (a) Iteration start condition: based on the load rate, load imbalance degree and network loss rate data output by step 4, the iteration is started when any of the following conditions is met.

[0153] i. There is an out-of-limit state main transformer or line.

[0154] ii. The load imbalance degree of the main transformer or line is greater than the threshold set in step 1.

[0155] iii. The network loss rate is higher than the historical same period benchmark value or the preset target value (such as 5%).

[0156] (b) Iteration termination condition: the iteration is terminated when any of the following conditions is met, and the current optimal strategy combination is output:

[0157] i. All selected optimization objectives (out-of-limit state elimination, device load imbalance reduction, network loss rate reduction) are achieved.

[0158] ii. The last iteration does not generate a new reconstruction optimization scheme.

[0159] iii. The number of iterations reaches the upper limit configured in step 1.

[0160] (c) Strategy iteration analysis:

[0161] i. Based on the power flow network loss calculation result of step 5, the generated scheme is comprehensively judged (the cost function is calculated) and sorted, the invalid scheme that violates the safety constraint is eliminated, and the optimal strategy is selected to join the overall reconstruction optimization strategy set.

[0162] ii. According to the improvement trend of the optimization index corresponding to the generated strategy, the weight coefficient of the cost function in step 5 is dynamically adjusted.

[0163] If an optimization target has not improved for two consecutive iterations, its corresponding weight is increased by 10%-50%.

[0164] If the target has been achieved, its weight is reduced to below 50% of the initial value.

[0165] Thus, the optimization targets that have not met the standard are preferentially strengthened.

[0166] iii. After each iteration, according to the position switch information in the latest generated strategy content, incremental dynamic topology search is performed, and the load rate and unbalance degree of related devices are recalculated as the basis data for the next iteration.

[0167] Further, the safety check and execution function in step 7 is mainly realized through the safety check and execution module in the implementation program, as shown in Failure , which specifically includes the following contents:

[0168] (a) Safety check:

[0169] i. Node and permission verification: The machine nodes and related control permissions of the optimization scheme to be executed are verified.

[0170] ii. Monitoring and remote control verification: The relevant remote control execution sequence is pushed to the remote control guardian, and the relevant guardian verifies and confirms.

[0171] iii. Topology anti-misoperation verification: The optimization scheme to be executed is verified for topology anti-misoperation, including topology anti-misoperation locking judgment and verification functions for switches, knife switches, grounding knife switches, and nameplate devices.

[0172] iv. Remote control basic parameter verification: The control devices in the optimization scheme to be executed are verified for remote control basic parameters.

[0173] v. Network constraint verification: Whether there is a secondary constraint overrun problem after power transfer is checked, including various constraint conditions such as feeder current, device voltage, transformer capacity, etc.

[0174] vi. Thermal conduction safety verification: Whether the thermal stability current during loop closing operation is lower than the device tolerance limit is verified.

[0175] (b) Strategy execution:

[0176] The final switch operation sequence to be executed is pushed to the operation ticket management application, supporting batch automatic or manual control issuance mode, and real-time feedback of operation state information (success / failure / timeout) during execution.

[0177] (c) Report generation: automatically archive the generation of restructuring optimization report after successful execution, and support one-key export function.

[0178] The report generation function is mainly realized by the report generation module in the implementation program, see Success The specific content includes:

[0179] i. The operation switch sequence included in the restructuring optimization strategy, i.e. the switch combination of operations required for restructuring.

[0180] ii. The power flow distribution before and after restructuring optimization.

[0181] iii. The user range affected by the restructuring optimization strategy, i.e. the power supply change of distribution transformers and distribution network loads in the analysis range.

[0182] iv. Comparison of device operating parameters before and after the execution of the restructuring strategy, including: comparison information of bus voltage state, device load distribution, network loss, etc.

[0183] Embodiment 2:

[0184] This embodiment introduces a power distribution network multi-objective feeder network restructuring optimization device considering distributed power supply, which includes:

[0185] The parameter configuration module is configured to: preset the core parameters of restructuring optimization, stored in the system configuration library as the basis for subsequent trigger judgment, data processing and strategy generation.

[0186] The trigger starting module is configured to: based on the configuration parameters, start the restructuring process through a multi-mode trigger mechanism.

[0187] The load prediction and data fusion module is configured to: predict future load data and fuse multi-period data to generate a unified analysis section.

[0188] The load rate calculation and state evaluation module is configured to: calculate the device load rate and evaluate the load balancing degree, network loss rate and risk of exceeding the limit under the normalized data section.

[0189] The dynamic topology optimization strategy generation module is configured to: search for transfer paths based on network topology, calculate power flow and network loss, and generate restructuring strategies.

[0190] The cyclic iteration analysis module is configured to: repeatedly call the load rate calculation and state evaluation module and the dynamic topology optimization strategy generation module until the iteration termination condition is met.

[0191] The safety check and execution module is configured to: execute and generate a report after checking the safety of the strategy.

[0192] Embodiment 3:

[0193] The embodiment introduces a device, comprising,

[0194] a memory.

[0195] a processor.

[0196] and

[0197] a computer program;

[0198] wherein the computer program is stored in the memory and configured to be executed by the processor to implement the power distribution network multi-objective feeder network reconstruction optimization method considering distributed power supply according to the embodiment 1.

[0199] In some embodiments, the execution subject of the power distribution network multi-objective feeder network reconstruction optimization method considering distributed power supply access according to the embodiment is generally a computer device system with certain computing capacity, including but not limited to a power distribution automation master station system, a master distribution integrated system, and a regional new generation dispatching technical support system. In possible implementation manners, the above reconstruction optimization method can be realized by the way that the processor calls the computer program stored in the memory, and the specific operation modes include but are not limited to front-end interface operation or background automatic operation.

[0200] Embodiment 4:

[0201] The embodiment introduces a storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the power distribution network multi-objective feeder network reconstruction optimization method considering distributed power supply according to the embodiment 1.

[0202] Embodiment 5

[0203] The embodiment is further to illustrate the technical solution of the present application, and the specific application of the above feeder reconstruction optimization method is described in detail below in combination with the actual operation scene of a certain 10kV residential area power distribution network. The power distribution network model structure in this scene can be referred to Failure , and the initial parameters and states are specifically configured as:

[0204] Device parameters: three transformers (T1, T2, T3) running in parallel, with a rated capacity of 630kVA; three 10kV feeders (L1, L2, L3) on the low-voltage side of the transformer through the bus, with a rated current of 400A; including a distributed photovoltaic PV1 with a capacity of 200kW (connected to L1, with a real-time output of 150kW), and a storage PS1 with a charging capacity of 100kWh (connected to L2, with a current SOC of 30%);

[0205] Initial operating state: T1 load rate 85%, T2 load rate 60%, T3 load rate 55%; L1 load 340A, L2 load 200A, L3 load 180A; total network loss 12kW; three-phase load imbalance: L1 A-phase current 130A, B-phase 110A, C-phase 100A.

[0206] After that, the reconstruction optimization analysis process is started in the following steps:

[0207] (1) Parameter configuration: based on the scene demand, the preset core parameters are as follows:

[0208] (a) Optimization target combination: eliminate device out-of-limit (load rate ≤80%), transformer load imbalance degree ≤5%, network loss reduction ≥3%;

[0209] (b) Time scale: future prediction period 1 hour, analysis time step 15 minutes;

[0210] (c) Iteration constraint: upper limit of iteration number 15 times, node voltage deviation threshold for power flow calculation 0.05kV;

[0211] (d) Safety constraint: voltage deviation ≤±5%, closed-loop impact current ≤2 times rated value.

[0212] (2) Trigger start:

[0213] The reconstruction analysis is started through the state prediction trigger mechanism:

[0214] According to the load prediction display, the L1 current will rise to 380A (load rate rises to 95%) in the next 30 minutes, the T1 load rate rises to 92%, and the load of other devices basically remains stable. The prediction credibility is 90% (≥85%), which meets the trigger condition of future load rate ≥90%, so the reconstruction analysis process is started.

[0215] (3) Load prediction and data fusion:

[0216] (a) Load prediction: combined prediction using LSTM neural network (weight 0.7) and gray prediction model (weight 0.3), generating L1 current data for 1 hour in the future: 340A, 360A, 380A, 370A (prediction error 6%, meeting the condition of prediction error ≤8%);

[0217] (b) Multi-section construction: 5 sections are intercepted according to 15-minute steps (real-time + 4 predictions), and the section data contains node current and power;

[0218] (c) Normalization processing: real-time data weight 0.7, short-term prediction (15-30 minutes) weight 0.5, fused section data: T1 average load 560kW, L1 average current 365A.

[0219] (4) Load rate calculation and state assessment:

[0220] (a) Load rate calculation (only show the prediction of the device calculation process):

[0221] T1 load rate:

[0222] L1 load rate:

[0223] (b) State assessment:

[0224] Overrun judgment: T1 load rate 89%>80%, L1 load rate 91.25%>80%, and predicted for more than 30 minutes, judged as T1, L1 overrun state;

[0225] Load imbalance calculation:

[0226] Line load imbalance:(>5%, need to optimize)

[0227] Net loss assessment: Normalized section total net loss 12.5kW (higher than historical same period benchmark value 11kW);

[0228] (5) Dynamic topology optimization strategy generation:

[0229] (a) Dynamic topology modeling: Build node-edge model, synchronize switch state (K8, K12 are tie switches, currently open, other switches are closed) and line resistance 0.2Ω / km;

[0230] (b) Transfer path search (only show part of the path and operating switch information):

[0231] Path 1: Line L1 transfers 100A load to line L2 through tie switch K8 (L2 remaining capacity 190A≥100A);

[0232] Path 2: Line L1 transfers 80A load to line L3 through tie switch K12 (L3 remaining capacity 210A≥80A);

[0233] (c) Power flow and network loss calculation:

[0234] Path 1: After transfer, L1 current value drops to 265A, load rate drops to 66.25%, L2 current value rises to 310A, load rate rises to 77.5%, T1 load rate drops to 68%, T2 load rate rises to 75%; Net loss reduces to 9.2kW;

[0235] Path 2: After transfer, L1 current value decreases to 285A, load rate decreases to 71.25%, L3 current value increases to 270A, load rate increases to 67.5%, T1 load rate decreases to 72%, T3 load rate increases to 70%; network loss decreases to 10.1kW;

[0236] (d) Cost function calculation (preset ):

[0237] Path 1:

[0238] Path 2:

[0239] (6) Iterative analysis:

[0240] (a) Start iteration: there are out-of-limit devices T1 (load rate 89%>80%) and L1 load rate (91.25%>80%), which meet the start iteration conditions;

[0241] (b) Iterative process (only the core process is briefly shown below):

[0242] First iteration:

[0243] Sort and select path 2 as the preferred reconstruction optimization strategy. After execution, the transformer imbalance degree does not meet the standard (7.2%>5%), and the network loss meets the standard (decreases by 19.2%>3%);

[0244] Weight adjustment: because the "load balancing" target has not improved, adjust the cost function parameters from 0.5 to 0.7, from 0.3 to 0.1;

[0245] Second iteration:

[0246] New reconstruction optimization strategy: transfer 50A load from line L2 to line L3. After execution, the transformer imbalance degree meets the standard (T1 load rate 72%, T2 load rate 65%, T3 load rate 75%, imbalance degree 3.5%<5%), and the network loss meets the standard (9.8kW, decreases by 21.6%>3%);

[0247] (c) Terminate iteration: achieve the preset optimization target (no device out of limit, transformer load imbalance degree ≤5%, network loss reduction ≥3%);

[0248] (6) Safety review and execution:

[0249] (a) Safety review:

[0250] Node and permission check: the operation node and login user information are valid and have control permissions, passed;

[0251] Remote control check: the remote control sequence is confirmed by the guardian, and passed;

[0252] Topology error prevention check: the target switch to be controlled has no lockout, and the two sides have no ground switch closed position, passed;

[0253] Remote control basic parameter check: the remote control parameters of the target switch to be controlled are correctly configured, passed;

[0254] Network constraint check: after power transfer, there is no device load rate exceeding limit (all ≤80%), voltage deviation 1.8% (≤±5%), passed;

[0255] Thermal conductivity safety check: the loop closing impact current 650A is less than the maximum withstand current of the feeder 800A (2 times the rated value), passed;

[0256] Check whether the thermal stability current during loop closing operation is lower than the device tolerance limit;

[0257] (b) Strategy execution:

[0258] K12 is closed after prepositioning, successful;

[0259] K4 is opened after prepositioning, successful;

[0260] K8 is closed after prepositioning, successful;

[0261] K10 is opened after prepositioning, successful;

[0262] (c) Report generation (only part of the content is shown):

[0263] Switch operation sequence:

[0264] Transformer Substation Load value (before) Load rate (before) Load value (after) Load rate (after) 1 S3 L3 K12 560.7 kW 453.6 kW 2 S1 L1 K4 390.6 kW 409.5 kW 3 S2 L2 K8 365.4 kW 472.5 kW 4 S2 L2 K10 Line Substation

[0265] Transformer load distribution:

[0266] Load value (before) Load rate (before) Load value (after) Load rate (after) Figure 1 Figure 1 T1 S1 Figure 1 89% Figure 1 72% T2 S2 Figure 1 62% Figure 1 65% T3 S3 ​ 58% ​ 75%

[0267] Feeder load distribution:

[0268] ​ ​ ​ ​ ​ ​ L1 S1 365 A 91.25% 285 A 71.25% L2 S2 210 A 52.5% 160 A 40% L3 S3 190 A 47.5% 320 A 80%

[0269] Those skilled in the art will appreciate that embodiments of the present application can be provided as methods, systems, or computer program products. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage media, etc.) having computer usable program code embodied in the medium.

[0270] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart or flowsheet block or blocks. ​ one or more flowcharts and / or blocks ​ means for functionally implementing the

[0271] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the ​ one or more flowcharts and / or blocks ​ means for functionally implementing the

[0272] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart or flowsheet block or blocks. ​ one or more flowcharts and / or blocks ​ means for functionally implementing the

[0273] The above only is the preferred embodiment of the present application, it should be pointed out that: for the ordinary skilled in the art, without departing from the principles of the present application, can also make a number of improvements and refinements, these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A multi-objective feeder network reconfiguration optimization method for a power distribution network considering distributed generation, characterized in that: Specifically comprising: Step 1: preset the core parameters of the to-be-reconstructed optimal distribution network; Step 2: based on the core parameters, when the multi-mode triggering mechanism is triggered, enter step 3; Step 3: based on the core parameters, predict the future power data of the to-be-reconstructed optimal distribution network to obtain future prediction data, obtain the real-time power data of the to-be-reconstructed optimal distribution network as real-time data, intercept the cross section of the corresponding period from the real-time data and the future prediction data, form a continuous multi-time cross section sequence, and perform weighted fusion on the multi-time cross section sequence to obtain a unified analysis cross section; Step 4: under the unified analysis cross section, calculate the load rate of the to-be-reconstructed optimal distribution network, evaluate the load balancing degree, network loss rate and risk of exceeding the limit according to the load rate of the to-be-reconstructed optimal distribution network, and when the load balancing degree, network loss rate and risk of exceeding the limit meet the iteration start condition, enter step 5, otherwise return to step 3; Step 5: based on the network topology model of the to-be-reconstructed optimal distribution network, search for a transfer path, calculate the power flow and network loss of the transfer path based on the core parameters, and construct a path cost function under multiple optimization objectives, solve the path cost function, and obtain the load imbalance degree, switch operation times and network loss reduction of the transfer path; Step 6: when the load imbalance degree, switch operation times and network loss reduction of the transfer path meet the iteration termination condition, output the optimized distribution network, otherwise return to step 5. 2.The method of claim 1, wherein the method further comprises: Also includes: Step 7: perform safety check, strategy execution and report generation according to the optimized distribution network. 3.The method of claim 1, wherein: The step 3 specifically comprises: A combination model of an LSTM neural network and a gray prediction model is used to predict the future power data of the to-be-reconstructed optimal distribution network to obtain future prediction data; According to the time scale parameter of the core parameters, the number of cross sections is calculated; The real-time power data of the to-be-reconstructed optimal distribution network is obtained as real-time data, and the cross section of the corresponding period is intercepted from the real-time data and the future prediction data to form a continuous multi-time cross section sequence; The multi-time cross section sequence is integrated into a unified analysis cross section through a weighted fusion algorithm, wherein the weight distribution rule of the weighted fusion algorithm is that the weight of the real-time data is 0.6-0.8, the weight of the short-term prediction data is 0.4-0.6, and the weight of the medium-term prediction data is 0.2-0.

4.

4. The method of claim 1, wherein the method further comprises: The load rate of the distribution network includes: transformer load rate, line load rate; The risk of exceeding the limit includes an exceeding state, and the exceeding state is that when the load rate of the target device is greater than a preset threshold and the duration is greater than a preset time length, the exceeding state is determined; load balancing, including load unbalancing The expression is as follows: ； in, For the analysis scope of the first The load rate of each device To analyze the average load rate of all devices within the scope, For the number of devices; The net loss rate includes: the ratio of the cross-section net loss to the total net loss, wherein the total net loss The expression is as follows: ； wherein, is the feeder segment active loss, is the transformer no-load loss, is the transformer load loss; The iteration start condition includes: when the exceeding state, the load imbalance degree and the network loss rate meet any one of the following conditions, the iteration is started; i. there is an exceeding state main transformer or line; ii. the load imbalance degree of the main transformer or line is greater than the iteration constraint in the core parameters; iii. the network loss rate is higher than the historical same period benchmark value or the preset target value.

5. The method of claim 1, wherein the method further comprises: The step 5 specifically comprises: A network topology model of node-edge structure is constructed based on the device model data of the to-be-reconstructed optimization power distribution network, wherein the nodes are set as node-type devices, the edges are set as connection-type devices, the association relationship between the nodes and the edges is used to map the electrical connection state of the actual power grid in real time, and the switch state, the device parameters and the normalized load data are synchronized in real time, so as to ensure that the network topology model is consistent with the actual power grid state; Based on the radial topology characteristics of the network topology model, the network nodes are traversed preferentially by using the depth-first search and supplemented by the breadth-first search, the feeder load in the to-be-analyzed area is taken as the starting point, the node-type devices are searched along to the other side, and a dynamic association relationship of "node-power supply" is formed; The switches whose current state is open and whose two sides belong to different "node-power supply" association relationships are the tie switches, and the original power backbone path device information and the backup power backbone path device information are recorded synchronously; The original power backbone path devices and the backup power backbone path devices are subjected to feasibility verification, and a backup path is obtained, in which the current state of the node-type devices is normal or is associated with protection, fault or blocking action, and the tie switch has a remote control function; The backup path that meets the backup margin constraint is selected from the backup path; The power flow and the network loss of the backup path that meets the backup margin constraint are calculated, a path cost function under multiple optimization objectives is constructed, and the expression of the path cost function is as follows: ； wherein, is a load imbalance improvement value, is a network loss reduction amount, is a number of switch operations, is a weight coefficient; The path cost function is solved, the solutions of the path cost function of the backup path are sorted, the backup path that violates the safety constraint in the core parameters is eliminated, the optimal solution is selected as the load imbalance degree, the number of switch operations and the network loss reduction amount of the backup path, and the weight coefficient of the path cost function is dynamically adjusted.

6. The method of claim 1, wherein the method further comprises: The iteration termination condition includes that the iteration is terminated when any of the following conditions is met: i. All optimization objective combinations in the core parameters are achieved; ii. No new reconstruction optimization scheme is generated in the last iteration; iii. The number of iterations reaches the upper limit of the iteration constraint configuration in the core parameters.

7. The method of claim 2, wherein the method further comprises: The safety check includes node and permission check, guardianship remote control check, topology anti-misoperation check, remote control basic parameter check, network constraint check and thermal guide safety check; The strategy execution includes pushing the final switch operation sequence to be executed to an operation ticket management application, supporting batch automatic or manual control of each control mode, and feeding back operation state information in real time during execution; The report generation includes automatically archiving and generating a reconstruction optimization report after successful execution, and supporting a one-key export function.

8. A power distribution network multi-objective feeder network reconfiguration optimization device considering distributed power sources, characterized by: Specifically, it includes: A parameter configuration module is configured to preset core parameters of a to-be-reconstructed optimization power distribution network; A trigger starting module is configured to trigger a multi-mode trigger mechanism based on the core parameters to enter a load prediction and data fusion module; The load prediction and data fusion module is configured to predict future power data of the to-be-reconstructed optimization power distribution network based on the core parameters, obtain future prediction data, acquire real-time power data of the to-be-reconstructed optimization power distribution network as real-time data, intercept cross sections of corresponding time periods from the real-time data and the future prediction data, form a continuous multi-time cross section sequence, and perform weighted fusion on the multi-time cross section sequence to obtain a unified analysis cross section. A load rate calculation and state evaluation module is configured to calculate a load rate of the to-be-reconstructed optimal distribution network under the unified analysis section, evaluate a load balancing degree, a network loss rate and an out-of-limit risk according to the load rate of the to-be-reconstructed optimal distribution network, and enter a dynamic topology optimization strategy generation module when the load balancing degree, the network loss rate and the out-of-limit risk meet an iteration starting condition, or return to the load prediction and data fusion module. The dynamic topology optimization strategy generation module is configured to search a transfer path based on a network topology model of the to-be-reconstructed optimal distribution network, calculate power flow and network loss of the transfer path based on core parameters, construct a path cost function under multiple optimization objectives, solve the path cost function, and obtain a load imbalance degree, a switch operation number and a network loss reduction amount of the transfer path. The loop iteration analysis module is configured to output an optimized distribution network when the load imbalance degree, the switch operation number and the network loss reduction amount of the transfer path meet an iteration termination condition, or return to the dynamic topology optimization strategy generation module.

9. A computer-readable storage medium, characterized in that: A computer program is stored on the computer, and when the computer program is executed by a processor, the method for multi-objective feeder network reconstruction optimization of a distribution network considering distributed power sources is realized.

10. A computer device, comprising: The computer device comprises: a memory configured to store instructions; a processor configured to execute the instructions, so that the computer device performs operations of the method for multi-objective feeder network reconstruction optimization of a distribution network considering distributed power sources.

Images