Power transmission and distribution network collaborative optimization method and system considering power and traffic multi-layer network coupling model

Abstract

The invention relates to the technical field of power transmission and distribution network collaborative optimization, in particular to a power transmission and distribution network collaborative optimization method and system considering a power and traffic multi-layer network coupling model, and aims to solve the problems that a power transmission and distribution network collaborative optimization model considering power-traffic network coupling is difficult to converge and time-consuming to solve. A McCormick envelope relaxation method is adopted to carry out relaxation of non-convex constraint, and a large M convex optimization balance method is adopted to determine an upper bound and a lower bound of a better variable. As the operation feasible region of the power distribution network at the single moment is influenced by the current state parameters of the system, the prediction of the operation feasible region of the power distribution network at the single moment is realized by adopting a deep learning method, and meanwhile, in order to improve the prediction accuracy, approximate processing is performed on the lower boundary of the operation feasible region of the power distribution network.

Description

Technical Field

[0001] The present invention relates to the technical field of coordinated optimization of transmission and distribution power grids, and particularly relates to a method and system for coordinated optimization of transmission and distribution power grids considering a multi-layer network coupling model of power and transportation. Background Art

[0002] For electric vehicle users, when driving in the transportation network, there will be a charging demand at a certain moment, and thus access the power grid through the charging facilities of the charging station at a certain node. That is to say, the charging behavior of electric vehicles is a key factor affecting the operation reliability of the power-transportation network coupling, and thus certain requirements are imposed on the coordinated optimization method.

[0003] The centralized solution method for coordinated optimization of transmission and distribution power grids considering power-transportation network coupling has numerous variables, is difficult to solve, requires collecting parameter information of all devices, and is difficult to protect privacy; while the distributed optimization model requires multiple iterations to complete the solution, resulting in difficult model convergence and time-consuming solution.

[0004] For example, the Chinese patent application with the publication number CN117649082A discloses a method and system for optimizing the operation reliability of a traffic-electricity coupling network. First, a traffic-electricity coupling network model is constructed; real-time operation data of the traffic-electricity coupling network is obtained based on the traffic-electricity coupling network model; traffic network reliability indicators and power network reliability indicators are calculated according to the real-time operation data of the traffic-electricity coupling network; face image information is collected, and the acceptable scheduling rational value of the user is calculated according to the face image information; an optimal scheduling scheme is generated and executed according to the traffic network reliability indicators, power network reliability indicators, and user acceptable scheduling rational value; the Chinese patent application with the publication number CN116720689A discloses a two-layer game model and equilibrium solution method for a power-transportation coupling network considering the elastic demand behavior of electric vehicles. It establishes a hybrid traffic flow model including the demand elasticity and charging behavior of electric vehicles to realize the coupling of urban power transportation networks; considering the optimization objectives and operation constraints of the coupling network equilibrium problem, a quasi-variational inequality framework is used to characterize the equilibrium state of elastic hybrid traffic users, and a mathematical optimization model of the urban power transportation coupling network is established; aiming at the non-convexity of the problem, the outer-layer hierarchical fixed-point iteration is used to decouple the problem into a two-layer convex optimization problem, and the inner-layer equilibrium problem is solved based on the viscous projection approximation algorithm to form a complete method for solving the equilibrium state of the two-layer game. However, the above-mentioned documents do not consider the processing method of the three-layer model and do not involve the relevant processing calculations of coordinated optimization. Summary of the Invention

[0005] Objective of the Invention: To overcome the deficiencies of the prior art, the present invention provides a collaborative optimization method for transmission and distribution networks considering the coupling model of power and transportation multi-layer networks. This method solves the problems of difficult privacy protection and time-consuming solution in traditional optimization methods. The present invention also provides a collaborative optimization system for transmission and distribution networks considering the coupling model of power and transportation multi-layer networks.

[0006] Technical Solution: According to the first aspect of the present invention, there is provided a collaborative optimization method for transmission and distribution networks considering the coupling model of power and transportation multi-layer networks, the method comprising: Construct a collaborative optimization model for transmission and distribution networks considering the coupling of power-transportation networks, which is a three-layer distributed structure, namely: the transmission network layer, the distribution network layer, and the transportation network layer; Construct the KKT conditions corresponding to the transportation network layer and project the KKT conditions into the distribution network layer, thereby reducing the three-layer optimization model to a two-layer optimization model; Use the large M convex optimization balancing method to set the distribution network layer after adding the KKT conditions as a convex model, and determine the upper and lower bounds of the Lagrange multipliers and traffic flows; Adopt the vertex search method to solve the operating feasible region of the above convex model, and use linear interpolation and the Douglas-Peucker algorithm to control the number of lower boundary points of the operating feasible regions obtained at different times, thereby obtaining training data with a unified dimension; Use the data set obtained from the training data to train a deep neural network, and then realize the collaborative optimization of the transmission and distribution networks based on the prediction results output by the neural network.

[0007] Further, it includes: The objective function of the collaborative optimization model for transmission and distribution networks considering the coupling of power-transportation networks is set to minimize the total cost of the transmission network and the distribution network. Specifically: the generator cost of the transmission network and the distribution network is minimized, the abandoned photovoltaic power is minimized, and the loss of the soft switch SOP in the distribution network is minimized; The constraint conditions considered in the transmission network layer include: node power balance constraint, DC power flow constraint, upper and lower limits of generator output constraint, unit ramp rate constraint, minimum start-stop time constraint of the unit, and photovoltaic constraint; The constraint conditions considered in the distribution network layer include: node power balance constraint, voltage drop constraint, current constraint, generator output constraint, photovoltaic constraint, voltage constraint, and transmission and distribution network boundary node constraint; The transportation network layer includes a charging power model and a traffic flow distribution model based on each path.

[0008] Further, it includes: The traffic flow distribution model is expressed as: The objective function of this model is to minimize the total driving time, charging time and the cost of charging electric vehicles, where: is the set of all paths, It is the collection of all charging stations. is the set of all OD pairs, It is an OD pair, where the OD pair is used to describe the traffic demand between the starting point and the end point. It's OD The set of all paths of Is the path Traffic on It's a charging station Traffic on Is the path The driving time on It's at the charging station The waiting time required for charging, and The paths are Basic driving time and charging stations Basic charging waiting time, and The road sections and charging stations capacity, It's OD path The cost of It's OD path The travel time on the road segment, It's OD path traffic flow, For OD travel needs, It's OD The demand function of It's OD The greatest travel demand, and OD pairs path Whether it passes through the road section and charging stations The flag variable, It's OD path of electric vehicles at charging stations The charging capacity, It's a charging station charging load.

[0009] Further, it includes: Constructing the KKT conditions corresponding to the transportation network layer and projecting the KKT conditions into the distribution network layer, including: Setting the KKT conditions corresponding to the traffic flow assignment model, expressed as:

[0010] wherein, is the OD pair path the sum of the total travel time, charging time, and cost on the path, and are respectively in the traffic flow assignment model and the corresponding Lagrange multipliers; Therefore, after canceling out it is expressed as: Adding the formulas (16)-(18) as constraint conditions to the constraint conditions considered in the distribution network layer.

[0011] Further, it includes: Using the large M convex optimization balancing method to set the distribution network layer after adding the KKT conditions as a convex model, including: According to the complementary slackness constraint it can be known that the traffic flow assignment model is a bilinear term, which results in the distribution network model after adding the KKT conditions being a non-convex model; the definition of the bilinear term involves two variables x and y , and their product form is x × y . Such product terms often appear in optimization models, especially when dealing with the interdependent relationships between variables. The bilinear term can be used to describe complex interactions.

[0012] Adopting the McCormick envelope relaxation method to transform the bilinear term into a linear form, thereby transforming the above non-convex model into a convex model, expressed as follows: ; ; ;

[0013] wherein, and are respectively the lower bound and upper bound of the Lagrange multiplier , and are respectively the traffic flow Lower and upper bounds.

[0014] Furthermore, it includes: The determination of the Lagrange multiplier and the upper and lower bounds of traffic flow includes: Using the Big M method to transform non-convex constraints into linear constraints, a mixed-integer programming problem is obtained as follows: ;

[0015] Wherein, is a very large positive number, is an auxiliary binary variable. By adjusting the value of the auxiliary binary variable , the problem is decomposed into convex optimization sub-problems; By changing the variable parameters in the traffic flow distribution model and respectively recording the range of the results obtained by using the Big M method and , within each recorded and interval range, keeping the value of the binary variable fixed. At this time, all and within the interval are between and , thereby determining the and lower and upper bounds.

[0016] Furthermore, it includes: The solution of the above convex model using the vertex search method for the operating feasible region includes: Determining the vertex search model according to the constraints of the distribution network itself, determining the initial vertex according to the dimension of the coupling variable, thereby forming an initial convex hull, and storing the initial vertex in the initialized vertex set; Searching for new vertices along the outer normal direction of each edge of the convex hull through the inner loop, and recording the corresponding improvement ratio, and the improvement ratio gradually decreases with the number of iterations; Until after the last iteration, determining the error between the current convex hull and the true projection through the outer loop is less than the allowable error, and terminating the loop, thereby obtaining the operating feasible region corresponding to the current moment.

[0017] Furthermore, it includes: The determination of the vertex search model according to the constraints of the distribution network itself, determining the initial vertex according to the dimension of the coupling variable, thereby forming an initial convex hull, and storing the initial vertex in the initialized vertex set includes: Determining the vertex search model

[0018] Among them, is the objective function for vertex search, the value of the vector represents the direction of vertex search, and are the coefficient matrices of the coupling variable and the internal variable respectively, is the right - hand constant coefficient vector; The number of initial vertices is at least one more than the dimension corresponding to the coupling variable and the initial vertices are determined along the axis.

[0019] Furthermore, it includes: Searching for new vertices along the outer normal direction of each edge of the convex hull through the inner loop, and recording the corresponding improvement ratio, where the improvement ratio gradually decreases with the number of iterations, including: If the convex hull formed by the existing vertices generated in the th loop is , assuming is surrounded by planes, the new vertex is searched along the outer normal vector direction of the plane of ; The outer normal vector of the th plane is , then the newly identified vertex is the optimal solution obtained by solving after making ; Indicates the improvement ratio of the new vertex to the current convex hull, which is:

[0020] Among them, is the outer normal vector of the th plane, is the right - hand constant term of the corresponding hyperplane equation; if , it indicates that is outside , add to the vertex set , making closer to the actual feasible region, and record the current improvement ratio. As the number of iterations increases, the improvement ratio gradually decreases. Furthermore, it includes: After the last iteration, determining the error between the current convex hull and the true projection through the outer loop, and terminating the loop when the error is less than the allowable error, so as to obtain the corresponding operating feasible region at the current moment, including: evaluating the error between the current convex hull and the true projection using the Hausdorff distance , expressed as:

[0021] Among them, is the convex hull of the approximate feasible region composed of existing nodes the set of all planes; When the error is less than the allowable error, the vertex search method terminates the iteration and outputs the convex hull composed of existing nodes as the final result, otherwise, enter the next outer loop.

[0022] Furthermore, it includes: The number of lower boundary points of the operation feasible region obtained at different times is controlled and processed by using linear interpolation and the Douglas - Peucker algorithm, including: Determine the number of lower boundary points of the operation feasible region of the distribution network at the current moment. When the current number of points is less than the specified number, use linear interpolation to add new lower boundary points between the existing points to depict the feasible region in detail without changing its shape; when the number of points exceeds the specified number, apply the Douglas - Peucker algorithm to simplify the number of lower boundary points to reduce the number of lower boundary points while minimizing the impact on the shape of the feasible region.

[0023] Furthermore, it includes: The application of the Douglas - Peucker algorithm to simplify the number of lower boundary points to reduce the number of lower boundary points while minimizing the impact on the shape of the feasible region includes: Connect the first point and the last point of the current lower boundary points, so that the head and tail points are connected to form an initial line segment, and mark the head and tail points as reserved points; Find the point among all internal points that is the farthest from the initial line segment , if is within the tolerance , the simplification is completed and all internal points are discarded; if it is not within the tolerance , then must be retained and the new part is processed recursively; finally, the output curve is the set of all retained points, where is the maximum distance from

[0024] On the other hand, the present invention also provides a coordinated optimization system for transmission and distribution networks considering the coupling model of power and transportation multi - layer networks, and this system includes: A model construction module for constructing a coordinated optimization model of transmission and distribution networks considering the coupling of power - transportation networks. This model is a three - layer distributed structure, namely: the transmission network layer, the distribution network layer, and the transportation network layer; The model dimension reduction module is used to construct the KKT conditions corresponding to the transportation network layer and project the KKT conditions into the distribution network layer, thereby reducing the three-layer optimization model to a two-layer optimization model; The model conversion module is used to set the distribution network layer after adding the KKT conditions as a convex model by using the large M convex optimization balancing method, and determine the upper and lower bounds of the Lagrange multipliers and traffic flow; The feasible region calculation module is used to solve the operating feasible region of the above convex model by using the vertex search method, and perform control processing on the number of lower boundary points of the operating feasible region obtained at different times by using linear interpolation and the Douglas-Peucker algorithm, so as to obtain training data with a unified dimension; The neural network training module is used to train the deep neural network by using the data set obtained from the training data, and then realize the coordinated optimization of the transmission and distribution network based on the prediction results output by the neural network.

[0025] Beneficial effects: Compared with the prior art, the present invention has the following advantages: (1) Aiming at the problems of difficult privacy protection and time-consuming solution in traditional optimization methods, the present invention adopts an equivalent projection method to use the KKT conditions to map the operating feasible region of the distribution network from a high-dimensional space to a low-dimensional space only containing coupling variables, protecting privacy while avoiding iterative solution, thereby realizing the acceleration of the solution.

[0026] (2) The present invention adopts the McCormick envelope relaxation method to relax non-convex constraints, and adopts the large M convex optimization balancing method to determine the optimal variable upper and lower bounds, which enables to find a balance between feasibility and accuracy, select the upper and lower bounds that can ensure the model feasibility and reduce errors, and the accuracy of the corresponding model is higher.

[0027] (3) In order to improve the prediction accuracy, the present invention adopts the vertex search method to collect the data set of the operating feasible region of the distribution network and approximate the lower boundary of the operating feasible region. On this basis, in order to solve the problem of inconsistent neural network input dimensions caused by the change in the number of lower boundary points in the distribution network operation analysis, an optimization method is proposed, which combines linear interpolation and the Douglas-Peucker algorithm, thereby improving data consistency, optimizing the training and prediction performance of the neural network model, and verifying its effectiveness in maintaining the model accuracy and generalization ability through experiments.

[0028] (4) Since the operating feasible region of the distribution network at a single moment is affected by the current state parameters of the system, the present invention adopts a deep learning method to realize the prediction of the operating feasible region of the distribution network at a single moment.

[0029] (5) The deep learning method of this application can accurately predict the shape of the lower boundary of the operation feasible region of the distribution network, while ensuring the optimality of the solution result and the rapidity of the solution. Case studies show that the collaborative optimization prediction results obtained in this application are highly consistent with the actual values in the overall trend. Especially in the 0 - 2 MW interval, although there are minor errors, the region enclosed by the predicted lower boundary is almost exactly the same as the actual feasible region, indicating that the model can effectively capture key features and has strong robustness. Its stable performance in different power intervals demonstrates its good prediction ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is the topology diagram of the T57D33U12 test system described in the embodiment of the present invention; Figure 2 is the parameter curve diagram of the test system described in the embodiment of the present invention; Figure 3 is the comparison diagram of the solution results described in the embodiment of the present invention; Figure 4 is the binary variable variation trend with the traffic load ratio coefficient; Figure 5 is the variation trend of the maximum value of electric vehicle trips and the maximum value of Lagrange multipliers described in the embodiment of the present invention; Figure 6 is the comparison of the solution results of three schemes described in the embodiment of the present invention; Figure 7 is the process of solving the operation feasible region by the vertex search method described in the embodiment of the present invention; Figure 8 is the improvement ratio diagram of the vertex search method described in the embodiment of the present invention; Figure 9 is the schematic diagram of the approximate treatment of the lower boundary of the feasible region described in the embodiment of the present invention; Figure 10 is the training error and test error diagram described in the embodiment of the present invention; Figure 11 is the comparison between the feasible region prediction result and the actual value described in the embodiment of the present invention; Figure 12 is the comparison of the distribution network feasible regions at different times described in the embodiment of the present invention; Figure 13 is the flowchart of the non - iterative solution method for the collaborative optimization of the transmission and distribution network considering the power - traffic network coupling described in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0032] Embodiment 1: As Figure 13 shown, the non-iterative solution method for the coordinated optimization of the transmission and distribution network considering the power-transportation network coupling in the embodiment of the present invention first forms a coordinated optimization model of the transmission and distribution network considering the power-transportation network coupling. The original model is time-consuming to solve and difficult to converge. Secondly, the McCormick envelope relaxation method is used to process the non-convex constraints and transform them into a convex optimization model. Thirdly, the vertex search method is used to collect the data set of the operation feasible region of the distribution network and approximately process the lower boundary of the operation feasible region. Finally, the training data set is used to train the deep neural network, and the prediction results output by the neural network are used to achieve the fast solution of the coordinated optimization of the transmission and distribution network, which mainly includes the following steps: The first step: Form a coordinated optimization model of the transmission and distribution network considering the power-transportation network coupling, and add the KKT conditions of the transportation network to the distribution network. This model is huge in scale, has a large number of variables, and complex constraints, making it difficult to solve.

[0033] Generally, the objective function of the coordinated optimization of the transmission and distribution network is set to minimize the total cost of the transmission and distribution network, mainly including minimizing the generator cost of the transmission and distribution network, minimizing the abandoned photovoltaic power, and minimizing the loss of soft open points (SOPs) in the distribution network. For example: ; ;

[0034] Among them, and are the operating costs of the transmission network and the distribution network respectively, is the set of all time scales, and are the sets of all unit nodes in the transmission network and the distribution network respectively, and are the sets of all photovoltaic nodes in the transmission network and the distribution network respectively, is the set of all SOP nodes in the distribution network, and are the unit power generation costs of the generators in the transmission network and the distribution network respectively, and are the unit penalty costs of the abandoned photovoltaic power in the transmission network and the distribution network respectively, is the loss conversion cost coefficient of SOP in the distribution network, and are the active power injection powers from the transmission network and distribution network generators to the nodes respectively, and are the abandoned photovoltaic powers in the transmission network and distribution network respectively, is the loss of SOP in the distribution network.

[0035] The constraints considered in the transmission network part are: node power balance constraint, DC power flow constraint, generator output upper and lower limits constraint, unit ramp rate constraint, unit minimum start-up and shut-down time constraints, and photovoltaic constraint, as follows:

[0036] Among them, is the set of all nodes in the transmission network, and are the sets of the predecessor nodes and successor nodes of node respectively, is the active power on branch , is the active power injected by the photovoltaic system into the node, is the active load of the node in the transmission network, is the set of all branches in the transmission network, is the reactance of branch in the transmission network, is the phase angle of the node, is the start-up and shut-down flag of the unit, and are the lower limit and upper limit of the generator output respectively, and are the minimum rate and maximum rate of the unit ramp rate respectively, and are the minimum on-time and minimum off-time of the unit respectively, is the upper limit of the active power output of the photovoltaic system.

[0037] The constraints considered in the distribution network part are node power balance constraint, voltage drop constraint, current constraint, generator output constraint, photovoltaic constraint, voltage constraint, and transmission and distribution network boundary node constraint, such as: ; ; ; ; ; ;

[0038] Among them, is the set of all nodes of the distribution network, and are the set of predecessor nodes and the set of successor nodes of the node respectively, and are the active power and reactive power injected by the distribution network generator into the node respectively, and are the active power and reactive power on the distribution network branch respectively. is the square of the amplitude of the current flowing through the distribution network branch respectively. and are the resistance and reactance of the distribution network branch respectively. and are the active power and reactive power injected by the photovoltaic system into the node respectively, and are the active load and reactive load of the distribution network node respectively, is the charging power of the distribution network charging station, is the square of the node voltage of the distribution network. Assuming that all photovoltaic systems operate in the constant power factor operation mode, is the maximum active power output of the distribution network photovoltaic system, is the power factor of photovoltaic operation, and are the minimum value and the maximum value of the distribution network node voltage respectively.

[0039] The boundary node constraints of the transmission and distribution network are:

[0040] Among them, is the active power output of the distribution network balancing node.

[0041] The losses of each part of the SOP in the distribution network are as follows: ; ;

[0042] Among them, , and are the losses of the AC-DC rectifier 1, the AC-DC rectifier 2 and the DC-DC module respectively, and are the loss coefficients of the AC-DC rectifier, and is the loss coefficient of the DC-DC module, , , , , , are auxiliary variables for calculating losses and are defined as follows:

[0043] where, and are the capacities of the AC-DC rectifier and the DC-DC module respectively, and are the active power and reactive power of the port of the SOP respectively.

[0044] The power balance constraint of the SOP is as follows: ; ; ; For the urban traffic network, first enumerate the offline routes for all OD pairs and calculate the charging power required for each route.

[0045] The model for calculating the charging power of each path is: ; ; ; ; ; ; ; ; where, is the charging power of the path of the OD pair at the charging station , is the set of all charging stations in all traffic networks, is the set of all road segments in the traffic network, is the remaining power of the electric vehicle at the node of the OD pair , is the distance of the road segment of the traffic network, is the unit mileage power attenuation rate of the electric vehicle, is the OD pair Binary flag variable indicating whether a section has been passed , which is a very large positive number is a very large positive number is the set of all nodes in the transportation network is a node in the transportation network Flag variable indicating whether charging is needed is the maximum battery capacity of the electric vehicle is the mileage anxiety coefficient of the electric vehicle user

[0046] The traffic flow assignment model is as follows: ; ; ; ; ; ; ; ;

[0047] The objective function is to minimize the total travel time, charging time, and the cost of electric vehicle charging. Among them, is the set of all paths is the set of all charging stations is the set of all OD pairs is the OD pair the set of all paths of is the path the traffic flow on is the charging station the traffic flow on is the path the travel time on is the waiting time required for charging at the charging station and are respectively the basic travel time of the path and the basic charging waiting time of the charging station and are respectively the capacity of the section and the capacity of the charging station is the OD pair the path the cost on​​​ is the travel time of the section on the path for OD pair is the flow of the path for OD pair is the travel demand for OD pair ; is the demand function for OD pair ; is the maximum travel demand for OD pair ; and are the flag variables indicating whether the path for OD pair passes through the section and the charging station respectively; is the charging power of the electric vehicle on the path for OD pair at the charging station ;

[0048] The coordinated optimization model of the transmission and distribution network considering the coupling of the power - traffic network is a three - layer distributed - structure model, which is difficult to solve directly. Therefore, the KKT conditions are adopted to transform the traffic network into the distribution network, reducing the three - layer optimization model to a two - layer optimization model. The KKT conditions (Karush - Kuhn - Tucker conditions) are a set of necessary conditions in optimization theory, applicable to solving nonlinear programming problems with equality and inequality constraints. When the objective function and the constraint conditions are convex, the KKT conditions are also sufficient conditions for finding the optimal solution. Specifically, in this embodiment, the traffic flow assignment model (including the minimization of travel time, charging waiting time, and cost, as well as constraints such as flow conservation and non - negativity) is transformed into equivalent KKT conditions and then projected onto the distribution network layer. Through these conditions, the original traffic - layer flow variables and Lagrange multipliers are eliminated and replaced with the aggregated charging load variables in the distribution network and the corresponding linear complementary constraints, thus simplifying the original optimization model coupled by the transmission network, distribution network, and traffic network into a two - layer model that only contains the transmission network layer and the distribution network with the traffic - layer KKT constraints embedded in the distribution network model, achieving model dimension reduction, integrated solution, and privacy protection.

[0049] In this embodiment, the KKT conditions of the optimization model of the urban traffic network are as follows: ; ; ; ;

[0050] wherein, is the sum of the total travel time, charging time, and cost on the path , and and are the Lagrange multipliers of equations (1-46) and (1-47), respectively. Eliminating the Lagrange multipliers , it can be further written as: ; ;

[0051] In summary, the objective function of the centralized algorithm for the coordinated optimization of the transmission and distribution network considering the power-transportation network coupling is (1-1), and the constraint conditions are the transmission network-related constraints (1-4)-(1-10), the distribution network-related constraints (1-11)-(1-17), the transmission and distribution network boundary node constraints (1-18), the SOP-related constraints (1-19)-(1-30), the KKT conditions of the transportation network (1-54)-(1-56), and the decision variables include a large number of binary variables and continuous variables in the transmission network, distribution network, and transportation network.

[0052] The centralized algorithm for the coordinated optimization of the transmission and distribution network is a mixed-integer nonlinear programming problem (MINLP) with numerous variables and complex constraints. Its solution time increases significantly with the increase in the problem scale, and it cannot even be solved using existing commercial solvers. In addition, since the centralized program needs to collect the parameter information of all devices in the transmission and distribution network, it involves privacy protection issues.

[0053] In Figure 1 the example system shown, the transmission network adopts the IEEE 57-node network. The access of new energy sources such as wind power and photovoltaic power is not considered in the transmission network. There are a total of seven units in the transmission network. The minimum start-stop time of the units is set to 3 hours, and the ramp rate is set to 20%. Among them, node 50 is connected to the IEEE 33-node distribution network.

[0054] A total of 7 distributed photovoltaic systems are set in the distribution network. The photovoltaic penetration rate is set to 0.5. An SOP is connected between node 7 and node 8. The SOP is connected with a DC load, and distributed generators are connected to node 6 and node 11. The transportation network is a 12-node urban transportation network with a total of 20 paths and 4 OD pairs. The OD pair information is shown in Table 1-1. The battery capacity of the electric vehicle is set to 100 kWh, the battery power attenuation coefficient is set to 0.2 kWh / km, the initial battery power when the electric vehicle departs is set to 30 kWh, and the user range anxiety coefficient is set to 0.2.

[0055] Table 1-1 OD Pair Information of Node Transportation Network

[0056] First, list the more likely paths for each OD pair and calculate the charging power of electric vehicles for each path.

[0057] Secondly, take the charging power of electric vehicles as a known value and input it into the traffic flow distribution program. Finally, write out the KKT conditions of the traffic flow distribution model and include them in the coordinated optimal scheduling model of the transmission and distribution network to complete the solution.

[0058] The purpose of possible path selection is to reduce the computational complexity. Enumerating all off-line paths may take a large amount of time and computing power. In addition, in practice, electric vehicle users will only choose several paths with better comprehensive costs considering time, distance, and required charging power. Therefore, it is unnecessary to perform off-line enumeration for all paths.

[0059] When selecting a path, first determine the shortest path that does not require electric vehicle charging. Due to the influence of traffic congestion effects, too many electric vehicles choosing the same path will lead to an extended travel time for electric vehicles. Therefore, the traffic flow distribution model will distribute the traffic flow to other paths. However, electric vehicles need to be charged when passing through longer paths, otherwise they cannot complete the whole journey. Therefore, paths passing through charging stations need to be selected.

[0060] The parameters of the test system are as Figure 2 shown. The research scenario is the day-ahead market. To adapt to the high volatility of new energy in the power system and improve market flexibility, the time interval is selected as 15 minutes. As Figure 2 shown in a, the system load is at a peak from 16:00 to 20:00. As Figure 2 shown in b, the output of the photovoltaic system reaches its peak from 11:00 to 12:00. As Figure 2 shown in c, the DC load simulates the rail transit load and reaches its peak around 8:00 am and 20:00 pm. As Figure 2 shown in d, the travel demand curve of electric vehicles reaches its peak at 7:00 and 17:00.

[0061] The solution results obtained by the centralized algorithm and the distributed algorithm are shown in Table 1-2. The objective function value and the cost of the power transmission and distribution network obtained by the two algorithms are very close, and the value obtained by the distributed algorithm is slightly higher. At the same time, the solution time of the distributed algorithm is about 2.3 times that of the centralized algorithm, and it takes 322.70 s to complete the solution. Among them, it takes 161.25 s to complete the solution each time the distribution network is optimized. This is because the distribution network takes into account the KKT conditions of the traffic network and the constraints related to SOP, and also includes integer variables. The variables are complex and there are many constraints, making it difficult to solve. When using the distributed algorithm to solve, the scale of each sub-problem is small, which is conducive to rapid solution. However, it requires multiple iterations to converge, and the total solution time is still very long.

[0062] Table 1-2 Comparison of solution results of centralized method and distributed method

[0063] The cost curves at each moment calculated by the two algorithms are as Figure 3 shown in a, b, c, and d in it. Among them, the cost of the power transmission network and the distribution network at each moment calculated by the distributed algorithm is slightly higher than that of the centralized algorithm. The cost of the traffic network completely coincides. The loss of SOP of the distributed algorithm is higher, which reflects the characteristic that the centralized algorithm can calculate the global optimal solution. The costs calculated by using the two algorithms show the same law. For example, the cost curve of the power transmission network is similar to the shape of the load curve. The cost of the distribution network is affected by various factors such as load and photovoltaic. At the moment when the load level is high but the photovoltaic output is low (8:00-24:00), the cost of the distribution network is high.

[0064] Step 2: When using the vertex search method to solve the operation feasible region, it is required that the model is a convex optimization model. Therefore, the McCormick envelope relaxation method is used to process the non-convex constraints, and the large M convex optimization balancing method is used to determine the optimal upper and lower bounds of the variables.

[0065] Specifically, when using the vertex search method to solve the operation feasible region of the distribution network, it is required that the optimization model of the distribution network is a convex model. Because the feasible region of a convex model is a convex set, its local optimal solution is the global optimal solution, and the vertices can accurately describe the boundary of the feasible region, thus simplifying the solution process and improving the calculation efficiency. The feasible region of a non-convex model may contain multiple local optimal solutions or discontinuous regions, resulting in the vertex search method being difficult to effectively find the global optimal solution and even possibly falling into a local optimum.

[0066] Since the KKT conditions of the urban traffic network optimization model are considered in the distribution network optimization model, the complementary slackness constraints It is a bilinear term, which makes the distribution network optimization model a non-convex model. Therefore, it is necessary to perform feasible relaxation on it. The commonly used method to handle bilinear terms is the McCormick Envelope Relaxation (MCE) method, which transforms bilinear terms into linear forms, such as: ; ; ;

[0067] where, and are the lower and upper bounds of the Lagrange multiplier respectively, and and are the lower and upper bounds of the traffic flow respectively.

[0068] Using the MCE method, non-convex bilinear terms can be relaxed into linear constraints, transforming the distribution network optimization model into a convex model. Then, the vertex search method can be used to solve the operating feasible region of the distribution network. However, the selection of the upper and lower bounds of the variables when using the MCE method has a very important impact on the solution results. Looser upper and lower bounds ensure the feasibility of the model but may lead to larger relaxation errors and less accurate results; tighter upper and lower bounds make the solution results more accurate but may make the model infeasible. Therefore, it is necessary to find a balance between feasibility and accuracy and select upper and lower bounds that can ensure the feasibility of the model and reduce errors.

[0069] This embodiment proposes the Big-M Convex Optimization Balancing Method (BCOBM) to determine accurate variable upper and lower bounds, uses the Big-M Method to handle non-convex constraints, and decomposes the problem into convex optimization sub-problems by adjusting the values of binary variables. The core of this method lies in transforming non-convex problems into convex problems by introducing auxiliary variables and the Big-M method, and at the same time determining the upper and lower bounds of variables by recording and analyzing the solution results, so as to reduce errors while ensuring the feasibility of the model.

[0070] Table 1-3 Steps of BCOBM for Determining Variable Upper and Lower Bounds

[0071] First, use the Big-M method to transform non-convex constraints into linear constraints, obtaining a mixed integer programming problem (MIP), as shown below: ;

[0072] wherein, is a very large positive number, is an auxiliary binary variable. By adjusting the value of the binary variable , the problem is decomposed into convex optimization sub-problems.

[0073] By changing the variable parameters in the transportation network, such as the travel demand of electric vehicles, record the range of the results obtained by using the big M method and . In each recorded and interval, the binary variable remains unchanged, and the model is a convex optimization model. Therefore, all and within the interval are between and , so that the upper and lower bounds of and can be determined. The specific steps are shown in Table 1-3.

[0074] Considering a heavier traffic load, the travel demands of four OD pairs are 1850, 1450, 14,00 and 1945 respectively. By changing the proportionality coefficient of the traffic load, the image of the binary variable changing with the proportionality coefficient of the traffic load is as shown in Figure 4 .

[0075] According to what is shown by a and b in Figure 4 , when the travel demand coefficient increases from 0.20 to 0.29, the value of the binary variable corresponding to OD pair 4 path 2 changes from 1 to 0. Constrained by the big method, the corresponding Lagrange multiplier becomes 0, and the traffic flow changes from 0 to a positive number, indicating that as the traffic demand increases, the travel demand of OD pair 4 begins to allocate traffic flow to path 2.

[0076] When the travel demand coefficient is between 0.20 and 0.29, the value of the binary variable is fixed, and at this time the model is relaxed to a convex optimization model. Record the values of the variables at this time and . Then all variables and within the travel demand range of 0.20 to 0.29 are between and , and the upper and lower bounds of the variables within all interval ranges are determined accordingly.

[0077] Since Figure 4 In [figure], gray represents that the traffic flow is 0 and the Lagrange multiplier is not 0, and white represents that the Lagrange multiplier is 0 and the traffic flow is not 0. This figure shows that as the travel demand coefficient changes within a certain range, the upper and lower bounds of the Lagrange multiplier and traffic flow in this range can be determined. This embodiment exemplifies the range of 0.2 - 0.29. Figure 4 In c, d, e, and f, they are all similar to the above analysis and will not be elaborated here.

[0078] As the traffic load increases, the change trends of the maximum values of electric vehicles and the maximum values of Lagrange multipliers for all paths of all OD pairs are respectively as Figure 5 shown. The maximum value of the Lagrange multiplier gradually decreases as the traffic load increases, and the number of electric vehicles shows a trend of first increasing and then decreasing. This is because as the proportion of traffic load increases, the total travel demand of electric vehicles increases, resulting in an increase in the maximum value of electric vehicles. When it increases to a certain extent, affected by the traffic congestion effect, the cost corresponding to the travel time of the path increases, resulting in the traffic flow distribution model starting to distribute to other paths, resulting in a decrease in the maximum value of electric vehicles allocated to a single path.

[0079] Compare the results of the three schemes: (1) MCE1: Solve using the big M method; (2) MCE2: Determine the upper and lower bounds using the BCOBM method and then solve using MCE after relaxation; (3) MCE3: Select relatively loose upper and lower bounds, that is, always select the maximum and minimum values that appear in Figure 3-4 the [figure].

[0080] The solution results of the three solution schemes are as Figure 6 shown. As shown in a of Figure 6 , in terms of the comprehensive cost of OD pairs, the cost of MCE3 for all OD pairs is relatively close to that of MCE1; as shown in b of Figure 6 , in the comparison of the charging station load, the traffic loads of MCE3 and MCE1 are concentrated on charging stations 5 and 6, while MCE2 has charging loads on all charging stations, indicating that the selection of the upper and lower bounds is too loose; as shown in c of Figure 6 , in the comparison of the path traffic flow, the traffic flow of MCE3 on all paths is relatively close to that of MCE1, while MCE2 also has traffic flow on path numbers 4 and 6, indicating that the error calculated by using the MCE2 scheme is relatively large. Considering all indicators comprehensively, the difference between the MCE3 scheme and the MCE1 scheme in all aspects is relatively small. Therefore, it can be considered that the MCE3 scheme performs better in approaching the MCE1 scheme and shows higher accuracy.

[0081] Step 3: Use the vertex search method to collect the data set of the operation feasible region of the distribution network, and approximate the lower boundary of the operation feasible region to improve the prediction accuracy.

[0082] Common methods for solving the operation feasible region include the vertex search method and the multi-parameter programming method. The vertex search method characterizes the feasible region by traversing its vertices from the inside to the outside of the feasible region, and the multi-parameter programming solves by analyzing the changes of the objective function and constraint conditions with parameters. In this embodiment, the vertex search method is used to solve the VPP feasible region. The process combining with the above text embodiment includes: Operation feasible region prediction process: Large M convex optimization processing model → Vertex search method to solve the feasible region → Obtain boundary points → Linear interpolation / Douglas-Peucker algorithm simplification → Construct neural network data set → Network prediction. The large M method processes the constraint upper and lower bounds of the optimization variables to ensure the stable operation of the vertex search method; the vertex search method collects the boundary data of the feasible region to obtain the lower boundary point set; linear interpolation and Douglas-Peucker further regularize the boundary point set to form data of a unified dimension and input it to the neural network; finally, non-iterative and efficient collaborative optimization prediction is achieved through deep learning.

[0083] (1) Determine the vertex search model

[0084] In the formula, is the objective function of vertex search, The value of the vector represents the direction of vertex search, and are the coefficient matrices of the coupling variable and the internal variable respectively, is the right-end constant coefficient vector.

[0085] For solving the operation feasible region of the distribution network, its constraint conditions are the constraints of the distribution network itself and the constraints related to the distribution network cost (1 - 64), where, is a very large positive number. In this embodiment, the very large positive number can be represented by "positive infinity", that is, "+∞".

[0086] (2) Initialization: At least initial vertices need to be determined to form an initial convex hull, where, is the dimension of the coupling variable. In this embodiment, the dimension of the coupling variable is 2, and at least 3 initial vertices need to be determined for initialization. These initial vertices can be determined along the axis, that is, when solving, let , is the The standard basis vectors with one element being 1 and the other elements being 0. Store the initial vertices in the initialized vertex set .

[0087] In this embodiment, the variables related to the operation and dispatch of the distribution network are defined into two categories: The first category: Coupling variables (coordination variables) ; Among them, and are the active power exchanged and the operating cost at the boundary nodes of the distribution network during time period respectively. The coupling variables are submitted to the distribution network for collaborative optimization clearing, and the optimization results are used as the boundary conditions for the optimized dispatch of the distribution network to fix the variables.

[0088] The boundary nodes in this embodiment are the common nodes of the transmission and distribution networks, which are both in the topology of the transmission network and in the topology of the distribution network. Therefore, the coupling variables are the variables used to describe the interaction between the distribution network and the transmission network, and are the variables belonging to both networks at the same time.

[0089] The second category: Internal variables , among which, includes all decision variables of the distribution network except the coupling variables. For the fixed coupling variables , the internal variables can be determined by the self-scheduling optimization of the distribution network, that is, given , the internal variables are obtained by solving the optimized sub-problem of the distribution network.

[0090] Let the variable operation feasible region that satisfies the distribution network constraints during the time period of the distribution network be , and define the distribution network operation feasible region as: for any coupling variable on the operation feasible region , there exists a set of feasible internal variables that do not violate the operation constraints of the distribution network, that is, the following formula: Among them, is the set of all feasible solutions of the distribution network during the time period .

[0091] (3) Inner loop: As Figure 8 shown, the inner loop of the vertex search method constructs the convex hull formed by the existing vertices and identifies the new vertices located outside the convex hull. The convex hull formed by the existing vertices generated in the -th loop is , assuming that is surrounded by planes, the new vertices are searched along the outer normal vector direction of the plane of . In the The outer normal vector of a plane is , then the newly identified vertex and the optimal solution obtained later. Calculate the distance from the vertex to of the plane using Equation (1-65), which is always non-negative. In the equation, is the improvement ratio of the new vertex to the current convex hull, is the th outer normal vector of the plane, is the right-hand side constant term of the corresponding hyperplane equation. If , it indicates that is outside , add it to the vertex set , making closer to the actual feasible region and record the improvement ratio.

[0092] (4) Outer loop: The outer loop evaluates the error of the current approximation and compares it with the allowable error to decide whether to terminate the algorithm. Use the Hausdorff distance to evaluate the error between the current convex hull and the true projection : In the equation, is the convex hull of the approximate feasible region composed of existing nodes of all planes.

[0093] When is less than the allowable error, the vertex search method terminates and outputs the convex hull composed of existing nodes as the final result, otherwise enter the next outer loop.

[0094] The process of using the vertex search method to solve the operation feasible region of the distribution network and the improvement ratio IR are respectively as Figure 7 shown.

[0095] The vertex search method first generates three initial vertices through initialization to form an initial convex hull; then searches for new vertices along the outer normal direction of each edge of the convex hull through the inner loop, and the improvement ratio gradually decreases with the number of iterations; until the last iteration, it is determined through the outer loop that is less than the allowable error and the loop is terminated. As shown in Figure 7 , a to i in it are the entire search process, Figure 7 , the operation feasible region of the distribution network in (i) in it. Any point in the feasible region contains a set of feasible decision variable values, while no set of decision variable values that satisfy the operation constraints of the distribution network can be found for any point outside the feasible region.

[0096] The operating feasible region of the distribution network at each moment is affected by the current system state parameters of the distribution network. The inconsistent parameters of the load level, photovoltaic output, DC load, and traffic load of the distribution network at each moment lead to changes in the operating feasible region of the distribution network. The operating feasible region of the distribution network needs to be solved again at each moment. Therefore, in this embodiment, an attempt is made to use the method of deep learning and use a neural network to predict the operating feasible region of the distribution network at each moment to avoid the process of repeated solution. In addition, the lower boundary of the operating feasible region of the distribution network already contains all the information of the operating feasible region of the distribution network. Therefore, it is possible to only predict the coordinates of the lower boundary points of the operating feasible region of the distribution network.

[0097] In the actual data collection process, it is noted that the number of lower boundary points of the operating feasible region of the distribution network at different time points is inconsistent. To solve this problem, the usual approach is to use the zero-padding technique, that is, to ensure the unity of the data dimension by adding zero values. Although zero-padding can achieve dimensional consistency, it does not provide additional effective information for the model, but only formally extends the length of the data. This approach may cause the model to overfit the padded zero-value part, thereby affecting the accuracy of model prediction.

[0098] In addition, it is found that only a certain number of points can roughly depict the outline of the operating feasible region of the distribution network, and too many points often only describe some local details of the feasible region, which is redundant in many cases. Therefore, in order to improve the performance of the model and avoid unnecessary computational overhead, it is necessary to find a more efficient method to handle this problem of inconsistent data dimensions, rather than simply relying on zero-padding.

[0099] To solve the problem of inconsistent input dimensions of the neural network caused by the change in the number of lower boundary points in the operation analysis of the distribution network, this embodiment proposes an optimization method, as Figure 9 shown. This method combines linear interpolation and the Douglas-Peucker algorithm. When the number of points is less than the specified number, linear interpolation is used to add new points between the existing points to depict the feasible region in detail without changing its shape; when the number of points exceeds the specified number, the Douglas-Peucker algorithm is applied to simplify the points to reduce the number of points while minimizing the impact on the shape of the feasible region. Through this method, it is aimed to improve data consistency, optimize the training and prediction performance of the neural network model, and verify its effectiveness in maintaining the accuracy and generalization ability of the model through experiments.

[0100] The Douglas-Peucker algorithm is a method for simplifying a polyline. It smooths the polyline composed of line segments by reducing the number of points. The degree of simplification of this algorithm is controlled by the parameter and Defines the maximum distance between the original points and the simplified curve. The algorithm first connects the start and end points to form an initial approximation and marks the start and end points as retained points. Then, the algorithm finds the point among all the interior points that is farthest from the line segment , if (i.e., the maximum distance to the line segment) is within the tolerance , the simplification is complete and all interior points can be discarded; if not within the tolerance, then must be retained and the algorithm recursively processes the new segment. Finally, the output curve is the set of all retained points.

[0101] Specifically, in this embodiment, given the starting curve as an ordered set of points, and the distance dimension , the algorithm first calculates the first approximation as the line segment ; Next, the algorithm checks all interior points to find the point farthest from the line segment , whose distance is ; If , the simplification is complete and all interior points can be discarded. Otherwise,[[]] must be retained and the algorithm recursively processes the new segments and .

[0102] Step 4: Use the training data set to train the deep neural network, and use the prediction results output by the neural network to achieve the fast solution of the coordinated optimization of the power transmission and distribution network.

[0103] Use the neural network to predict the operation feasible region of the above power distribution network. In this embodiment, a fully connected neural network is used, and the model parameters and training parameters of the neural network are shown in Table 1-4. There are a total of 10,000 data sets, and the data sets are divided according to the ratio of 70% training set, 15% validation set, and 15% test set. The training error is as Figure 10 shown:[[]] Table 1-4 Model parameters and training parameters of the fully connected neural network

[0104] According to Figure 11As shown in the predicted results of the feasible operating region of the distribution network, the neural network model performs excellently in terms of prediction accuracy, generalization ability, and practical application value. The predicted results are highly consistent with the actual values in the overall trend. Especially in the range of 0 - 2 MW, although there are minor errors, the region enclosed by the predicted lower boundary is almost exactly the same as the actual feasible region, indicating that the model can effectively capture key features and has strong robustness. Its stable performance in different power ranges demonstrates its good prediction ability.

[0105] The feasible operating regions of the distribution network obtained by prediction at different times are as Figure 12 shown. Among them, the transmission power reaches the minimum value around 4:00, that is, the lower boundary of the feasible operating region is at the leftmost end. At this time, the load level in the system is low, and the active power deficit in the system is small. The cost of the distribution network reaches the maximum value around 20:00. At this time, the photovoltaic output is 0, and the loads of the distribution network, DC load, and traffic load are all high, resulting in the maximum cost of the system. The costs of the distribution network at around 0:00 and 12:00 are both low. This is because the system load level is low at 0:00, and the photovoltaic output of the system is the largest at 12:00. The active power deficit in the system is small, resulting in a low transmission power and a low cost of the distribution network.

[0106] The solution results of the coordinated optimization of the transmission and distribution network using three methods are shown in Table 1 - 5. Thanks to the accuracy of the neural network prediction, the equivalent projection method used in this application has little difference in both the objective function value and the cost of the transmission and distribution network compared with the centralized method. At the same time, its solution time is only 95.12 s. Among them, the optimization time of the transmission network is 0.18 s, and the solution time of the distribution network is 94.94 s. The acceleration effect compared with the centralized method can reach 31.72%, and the acceleration effect compared with the distributed method can reach 70.52%.

[0107] Table 1 - 5 Comparison Table of Solution Results

[0108] Embodiment 2: The present invention also provides a coordinated optimization system for the transmission and distribution network considering the coupling model of the power and transportation multi - layer network. The system includes: A model construction module for constructing a coordinated optimization model of the transmission and distribution network considering the power - transportation network coupling. This model is a three - layer distributed structure, namely: the transmission network layer, the distribution network layer, and the transportation network layer; A model dimension reduction module for constructing the KKT conditions corresponding to the transportation network layer and projecting the KKT conditions into the distribution network layer, thereby reducing the three - layer optimization model to a two - layer optimization model; A model conversion module for setting the distribution network layer after adding the KKT conditions as a convex model using the large - M convex optimization balancing method and determining the upper and lower bounds of the Lagrange multipliers and traffic flows; The feasible region calculation module is used to solve the operation feasible region of the above convex model by using the vertex search method, and perform control processing on the number of lower boundary points of the operation feasible regions obtained at different times by using linear interpolation and the Douglas-Peucker algorithm, so as to obtain training data of a unified dimension; The neural network training module is used to train a deep neural network by using the data set obtained from the training data, and further realize the collaborative optimization of the transmission and distribution network based on the prediction result output by the neural network.

[0109] Other technical features of the transmission and distribution network collaborative optimization system according to this embodiment, which considers the coupling model of the power and transportation multi-layer network, are similar to the corresponding transmission and distribution network collaborative optimization method that considers the coupling model of the power and transportation multi-layer network, and will not be elaborated here.

[0110] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The meaning of "a plurality" is two or more, unless otherwise specifically defined.

[0111] In the present invention, unless otherwise clearly specified and defined, the terms "installed", "connected", "connected", "fixed", etc. should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or integrated; it may be a mechanical connection or an electrical connection; it may be directly connected, or indirectly connected through an intermediate medium, and may be the internal connection of two components or the interaction relationship between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0112] In the present invention, unless otherwise clearly specified and defined, the first feature being "on" or "under" the second feature may be that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on" the second feature may be that the first feature is directly above or obliquely above the second feature, or merely indicates that the first feature is at a higher horizontal height than the second feature. The first feature being "under", "below" and "beneath" the second feature may be that the first feature is directly below or obliquely below the second feature, or merely indicates that the first feature is at a lower horizontal height than the second feature.

[0113] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0114] Any process or method description represented in a flowchart or described in other ways herein can be understood to represent a module, segment, or portion of code including one or more executable instructions for implementing a specific logical function or process, and the scope of the preferred embodiments of the present invention includes additional implementations, where the functions can be executed in a manner not shown or discussed, including in a substantially simultaneous manner according to the functions involved or in a reverse order, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0115] The logic and / or steps represented in a flowchart or described in other ways herein, for example, can be considered as a sequenced list of executable instructions for implementing a logical function, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in connection with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of the computer-readable medium include the following: an electrical connection portion having one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0116] It should be understood that each part of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0117] Those of ordinary skill in the art can understand that all or part of the steps carried by the methods of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0118] In addition, each functional unit in various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. When the above integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0119] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A collaborative optimization method for transmission and distribution networks considering the coupling model of power and transportation multi-layer networks, characterized in that, The method includes: Constructing a collaborative optimization model of the transmission and distribution power grid considering the coupling of the power-transportation network, which is a three-layer distributed structure, namely: the transmission power grid layer, the distribution power grid layer, and the transportation network layer; Constructing the KKT conditions corresponding to the transportation network layer and projecting the KKT conditions into the distribution power grid layer, thereby reducing the three-layer optimization model to a two-layer optimization model; Using the large M convex optimization balancing method to set the distribution power grid layer after adding the KKT conditions as a convex model, and determining the upper and lower bounds of the Lagrange multipliers and traffic flows; Adopting the vertex search method to solve the operating feasible region of the above convex model, and using linear interpolation and the Douglas-Peucker algorithm to control the number of lower boundary points of the operating feasible regions obtained at different times, thereby obtaining training data of a unified dimension; Training a deep neural network using the dataset obtained from the training data, and then realizing the collaborative optimization of the transmission and distribution power grid based on the prediction results output by the neural network.

2. The collaborative optimization method for the transmission and distribution network considering the coupling model of the multi-layer network of power and transportation according to claim 1, characterized in that The objective function of the collaborative optimization model of the transmission and distribution power grid considering the coupling of the power-transportation network is set to minimize the total cost of the transmission and distribution power grids. Specifically: minimizing the generator cost of the transmission and distribution power grids, minimizing the abandoned photovoltaic power, and minimizing the soft-switching SOP loss in the distribution power grid; The constraint conditions considered by the transmission power grid layer include: node power balance constraint, DC power flow constraint, generator output upper and lower limit constraint, unit ramp constraint, unit minimum start-stop time constraint, and photovoltaic constraint; The constraint conditions considered by the distribution power grid layer include: node power balance constraint, voltage drop constraint, current constraint, generator output constraint, photovoltaic constraint, voltage constraint, and transmission and distribution power grid boundary node constraint; The transportation network layer includes a charging power model and a traffic flow distribution model based on each path.

3. The collaborative optimization method for the transmission and distribution network considering the coupling model of the multi-layer network of power and transportation according to claim 2, characterized in that The traffic flow distribution model is expressed as: The objective function corresponding to this model is to minimize the total travel time, charging time, and the cost of electric vehicle charging. Among them, is the set of all paths, is the set of all charging stations, is the set of all OD pairs, represents a certain OD pair, is the set of all paths of OD pair , is the path on the flow, is the charging station on the flow, is the path on the travel time, is at the charging station charging required waiting time, and are the path basic travel time and charging station basic charging waiting time, and are the section and charging station capacity, is the OD pair path on the cost, is the OD pair path on the section travel time, is the OD pair path traffic flow, is the OD pair travel demand, is the OD pair demand function, is the OD pair maximum travel demand, and are the OD pair [[ID= {80}]]path [[ID=8 {2}]]whether through the section and charging station flag variable, is the OD pair path electric vehicle at the charging station charging power, is the charging station charging load.

4. The collaborative optimization method for the transmission and distribution network considering the coupling model of the power and transportation multi-layer network according to claim 3, characterized in that The construction of the KKT conditions corresponding to the transportation network layer and the projection of the KKT conditions into the distribution power grid layer include: Setting the KKT conditions corresponding to the traffic flow distribution model, expressed as: Among them, is the sum of the total travel time, charging time, and cost on the path , and are the Lagrange multipliers corresponding to and in the traffic flow assignment model, respectively. Therefore, after offsetting, it is expressed as: Adding the formulas (16)-(18) as constraint conditions to the constraint conditions considered by the distribution power grid layer.

5. The collaborative optimization method for the transmission and distribution network considering the coupling model of the multi-layer network of power and transportation according to claim 4, wherein The use of the large M convex optimization balancing method to set the distribution power grid layer after adding the KKT conditions as a convex model includes: According to the complementary slackness constraints It can be seen that the traffic flow distribution model is a bilinear term, which results in the distribution network model after adding the KKT conditions being a non-convex model; Adopting the McCormick envelope relaxation method to transform the bilinear terms into linear forms, thereby transforming the above non-convex model into a convex model, expressed as follows: Among them, and are the lower and upper bounds of the Lagrange multiplier respectively, and are the lower and upper bounds of the traffic flow respectively.

6. The collaborative optimization method of the power transmission and distribution network considering the coupling model of the power and transportation multi-layer network according to claim 5, characterized in that The determination of the upper and lower bounds of the Lagrange multipliers and traffic flows includes: Using the large M method to transform non-convex constraints into linear constraints to obtain a mixed-integer programming problem, as follows: Among them, is a very large positive number, is an auxiliary binary variable. By adjusting the value of the auxiliary binary variable , the problem is decomposed into convex optimization sub-problems; By changing the variable parameters in the traffic flow distribution model and separately recording the ranges of the results solved using the Big M method and , within each recorded and interval range, keep the binary variable value fixed. At this time, all the and within the interval are between and , thereby determining the upper and lower bounds of and .

7. The collaborative optimization method for the transmission and distribution network considering the coupling model of the multi-layer network of power and transportation according to claim 2, wherein The adoption of the vertex search method to solve the operating feasible region of the above convex model includes: Determining a vertex search model according to the constraint conditions of the distribution power grid itself, and determining an initial vertex according to the dimension of the coupling variables, thereby forming an initial convex hull, and storing the initial vertex in the initialized vertex set; the coupling variables therein are variables used to describe the interaction between the distribution power grid and the transmission power grid, and are variables that belong to both networks at the same time; Searching for new vertices along the outer normal direction of each edge of the convex hull through an inner loop, and recording the corresponding improvement ratio, and the improvement ratio gradually decreases with the number of iterations; After the last iteration, the error between the current convex hull and the true projection is determined through the outer loop. If it is less than the allowable error, the loop is terminated, thereby obtaining the operating feasible region corresponding to the current moment.

8. The collaborative optimization method for the transmission and distribution network considering the coupling model of the multi-layer network of electric power and transportation according to claim 7, characterized in that Determine the vertex search model according to the constraints of the distribution network itself, determine the initial vertices according to the dimension of the coupling variables, thereby forming an initial convex hull, and store the initial vertices in the initialized vertex set, including: Determine the vertex search model Among them, is the objective function of vertex search, the value of the vector represents the direction of vertex search, and are the coefficient matrices of the coupling variable and the internal variable respectively, is the right-end constant coefficient vector; The number of initial vertices is at least the coupling variable The corresponding dimension is incremented by one, and the initial vertices are determined along the axis.

9. The collaborative optimization method for the transmission and distribution network considering the coupling model of the multi-layer network of power and transportation according to claim 8, wherein Search for new vertices along the outer normal direction of each edge of the convex hull through the inner loop, and record the corresponding improvement ratio, where the improvement ratio gradually decreases with the number of iterations, including: If the convex hull formed by the existing vertices generated in the th cycle is , assuming that is enclosed by planes, the new vertex is searched along the direction of the outer normal vector of the plane; The outer normal vector of the th plane is Then the newly recognized vertex is the optimal solution obtained by solving after making Indicate new vertices The improvement ratio for the current convex hull is as follows: Among them, is the outer normal vector of the -th plane, and is the right-hand side constant term of the corresponding hyperplane equation; if , it indicates that is outside , then add to the vertex set , so that is closer to the actual feasible region, and record the current improvement ratio. As the number of iterations increases, the improvement ratio gradually decreases.

10. The collaborative optimization method for the transmission and distribution network considering the coupling model of the multi-layer network of power and transportation according to claim 9, characterized in that, After the last iteration, the error between the current convex hull and the true projection is determined through the outer loop If it is less than the allowable error, the loop is terminated, and the operation feasible region corresponding to the current moment is obtained, including: Use the Hausdorff distance to evaluate the error between the current convex hull and the true projection , expressed as: Among them, is the convex hull of the approximate feasible region composed of existing nodes and is the set of all planes; When the error is less than the allowable error, the vertex search method terminates the iteration and outputs the convex hull formed by the existing nodes as the final result. Otherwise, it enters the next outer loop.

11. The collaborative optimization method of the power transmission and distribution network considering the coupling model of the multi-layer network of power and transportation according to claim 7, wherein Control and process the number of lower boundary points of the operation feasible region obtained at different times using linear interpolation and the Douglas-Peucker algorithm, including: Determine the number of lower boundary points of the operation feasible region of the distribution network at the current moment. When the current number of points is less than the specified number, use linear interpolation to add new lower boundary points between the existing points to depict the feasible region without changing its shape; when the number of points exceeds the specified number, apply the Douglas-Peucker algorithm to simplify the number of lower boundary points to reduce the number of lower boundary points while minimizing the impact on the shape of the feasible region.

12. The collaborative optimization method for the transmission and distribution network considering the coupling model of the multi-layer network of electric power and transportation according to claim 11, characterized in that Apply the Douglas-Peucker algorithm to simplify the number of lower boundary points to reduce the number of lower boundary points while minimizing the impact on the shape of the feasible region, including: Connect the first point and the last point of the current lower boundary points, so that the first and last points are connected to form an initial line segment, and mark the first and last points as reserved points; Find the point among all internal points that is farthest from the line segment , if is within the tolerance , the simplification is complete and all internal points are discarded; if not within the tolerance , then must be retained and the new part is processed recursively; finally, the output curve is the set of all retained points, where is the maximum distance to the line segment.

13. A collaborative optimization system for transmission and distribution networks considering the coupling model of power and transportation multi-layer networks, characterized in that, The system includes: A model construction module for constructing a coordinated optimization model of the transmission and distribution network considering the coupling of the power-transportation network. This model is a three-layer distributed structure, namely: the transmission network layer, the distribution network layer, and the transportation network layer; A model dimension reduction module for constructing the KKT conditions corresponding to the transportation network layer and projecting the KKT conditions into the distribution network layer, thereby reducing the three-layer optimization model to a two-layer optimization model; A model conversion module for setting the distribution network layer after adding the KKT conditions as a convex model using the large M convex optimization balancing method, and determining the upper and lower bounds of the Lagrange multiplier and traffic flow; A feasible region calculation module for solving the operation feasible region of the above convex model using the vertex search method, and controlling and processing the number of lower boundary points of the operation feasible region obtained at different times using linear interpolation and the Douglas-Peucker algorithm, thereby obtaining training data with a unified dimension; A neural network training module for training a deep neural network using the dataset obtained from the training data, and then realizing the coordinated optimization of the transmission and distribution network based on the prediction results output by the neural network.

Images