Transformer area flexible interconnection two-stage robust planning operation method

Abstract

The invention provides a transformer area flexible interconnection two-stage robust planning operation method. The method comprises the following steps: describing photovoltaic output and load fluctuation by using a multi-interval fluctuation uncertainty set; then, a two-stage robust planning operation model based on the FID is constructed, in the first stage, the FID investment cost serves as an optimization target to decide an FID planning scheme and configuration capacity, and in the second stage, the worst source load operation scene is given, and the annual operation cost is optimized; a column and constraint generation algorithm (Camp; and solving a two-stage robust model by using a strong duality theory. And finally, proving that the proposed method can accurately summarize a source load output operation scene within one year through 40 transformer area examples, and description of source load fluctuation is more practical, so that a more efficient and economic FID planning configuration scheme and a scheduling operation scheme can be provided by a two-stage robust model, and the economic benefit of the system is improved. Through introduction of correlation parameters in the source load fluctuation uncertainty model, a solution result obtained by planning operation can give consideration to both economical efficiency and robustness.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network planning, configuration and operation, and in particular to a two-stage robust planning and operation method for flexible interconnection of distribution areas. Background Art

[0002] With the acceleration and upgrading of global energy transformation, the proportion of AC and DC sources and loads connected to low-voltage distribution networks, represented by distributed wind power, photovoltaic power, energy storage and charging piles, has gradually increased. The volatility and intermittency of the output on the source side and the complexity and variability of the demand on the load side have brought severe challenges to the balance of power supply and demand and dispatching operation in the distribution area.

[0003] Most traditional substations are powered by a single line, and each substation is independent of each other, which makes it difficult to meet the current growing supply and demand balance and high reliability requirements. At the same time, the source-load operation characteristics of each substation vary greatly in time and space distribution. On the one hand, different types of loads such as industrial, commercial, and residential loads have complementary daily load characteristics. For example, the peak period of industrial and commercial loads corresponds to the trough period of residential loads, causing some substations to be overloaded during peak power consumption periods; on the other hand, after some substations are connected to distributed wind and solar energy and other new energy sources, there is surplus capacity while meeting the load demand, while other substations that have not been connected to new energy cannot share these surplus capacities, resulting in unbalanced loads between different substations.

[0004] To address the above problems, the low-voltage flexible interconnection device (FID) based on voltage source converter (VSC) has the advantages of bidirectional power flow and precise power flow control. It is also low-cost and can achieve interconnection and mutual supply between substations, thereby enhancing the power supply capacity of the substation, improving the load balancing rate of distribution transformers and the level of new energy consumption.

[0005] At present, there are relevant studies on the problem of flexible interconnection optimization configuration and operation scheduling for the uncertainty of source-side output and load-side demand. Most of them focus on robust optimization, and the reasonable modeling of source-load fluctuation uncertainty is the key to this type of robust optimization problem. Some literatures use single-interval box uncertainty sets to characterize the uncertainty of photovoltaic and load power. Since only single-interval fluctuations are considered, the uncertain variables can only take boundary values, which has great limitations; some literatures construct probability sets of wind, light, and load forecast errors through probability distribution and Wasserstein sphere. The flexible and adjustable radius of Wasserstein sphere can make the FID optimization scheme strike a balance between economy and robustness. However, due to the introduction of chance constraints and probability distribution, modeling is difficult and the solution time is long.

[0006] In addition, in the flexible interconnection planning problem of the substation, the selection of typical days of source-load data is the basic work of planning and design. The complexity and variability of source-load scenarios make it extremely difficult to traverse each operation scenario. Most literature uses the K-means clustering method to divide the original source-load data to obtain the typical days of distributed power output and load. However, the traditional K-means algorithm is simple in principle, and there are problems such as the difficulty in determining the number of clusters m in advance and the need to randomly select the initial cluster center. Therefore, a better algorithm is needed to improve the data clustering effect and obtain a more accurate and reasonable set of typical days.

[0007] In response to the above problems, this paper first implements the typical day division of source-load data based on the improved ISODATA clustering algorithm. This method does not need to determine the number of clusters m in advance, and can learn the high-dimensional characteristics of source-load time series data, thereby improving the clustering effect; secondly, the multi-interval box set is used to characterize the uncertain fluctuations of photovoltaic and load, avoiding the problem of strong conservatism of single-interval uncertain sets, so as to more accurately describe the source-load fluctuation scenario, fit the actual source-load fluctuation, and have a faster solution speed than chance constraints and probability distribution; then a two-stage robust planning operation model for low-voltage area flexible interconnection is established, which is a min-max-min framework. Among them, the first stage takes the lowest FID planning investment cost as the goal, and solves the FID flexible interconnection scheme and VSC configuration capacity; the second stage solves the lowest annual operating cost including power purchase cost and network loss under the worst source-load scenario, and optimizes the VSC transmission power and distribution transformer load rate. In the iterative solution process, the dynamic update of the area connectable set is considered to correct the selection of flexible interconnection planning scheme. Finally, 40 distribution network design examples and comparative experiments were used to verify that the method proposed in this paper can fully fit the actual situation of source-load fluctuations, accurately describe the operating scenarios of source-load fluctuations, and enhance the rationality and robustness of flexible interconnection planning and operation. Summary of the invention

[0008] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a two-stage robust planning and operation method for flexible interconnection of substations, which can more accurately describe the source-load output operation scenario within a year and carefully characterize the uncertainty of source-load fluctuations, making the source-load fluctuation scenario more realistic, thereby providing a more economical and reasonable FID planning and configuration plan and improving the economic benefits of the system.

[0009] The technical solution adopted by the present invention to solve the practical problem comprises the following steps:

[0010] A two-stage robust planning and operation method for flexible interconnection of substations, comprising:

[0011] Obtain data related to flexible interconnection devices and stations;

[0012] Construct a flexible interconnection planning and operation model for the substation area, the flexible interconnection planning and operation model for the substation area includes an objective function and constraints, wherein the objective function considers two-stage planning and operation, wherein the first stage takes the optimization of the investment and construction cost of the flexible interconnection device as the objective function, and the second stage takes the optimization of the annual operation cost of the system under the uncertain fluctuation of photovoltaic output and AC load as the objective function; the constraints include the operation constraints of the flexible interconnection device, the operation constraints of the substation area distribution transformer, the flow balance constraints of the medium-voltage feeder, the operation constraints of various types of equipment connected to the substation area except the source load, the source load uncertain fluctuation constraints and the safety constraints;

[0013] Based on the acquired flexible interconnection device and substation related data, the substation flexible interconnection planning and operation model is solved to obtain the planning configuration and scheduling operation plan of the flexible interconnection device.

[0014] Furthermore, the objective function includes:

[0015] The objective function C1 of the first stage is as follows:

[0016] ;

[0017] Where: is the investment cost of the flexible interconnection device;

[0018] The objective function C2 of the second stage is as follows:

[0019] ;

[0020] ;

[0021] Where: Cost of purchasing electricity from the main grid for the distribution network; The network loss cost includes medium voltage line loss, VSC operation loss and distribution transformer active power loss; The operating costs of the equipment and components connected to the substation, including photovoltaic PV, VSC, energy storage ESS and discrete capacitor CB; is the weight penalty function, , , and They represent the fluctuation parameters of photovoltaic and load at time t in the operation scenario k, and They respectively represent the weights of photovoltaic and load corresponding to the small interval q at time t in station i in operation scenario k.

[0022] Furthermore, the flexible interconnection device includes a plurality of VSCs. Assuming that the power of the VSC flows out of the bus in a positive direction, the operation constraints of the flexible interconnection device include:

[0023] ;

[0024] ;

[0025] ;

[0026] ;

[0027] In the formula, represents the VSC set associated with the g-th flexible interconnection device; and are respectively the active power and reactive power of the VSC at the side of the grid i at time t in the operation scenario k, is the active power loss of the VSC at the side of the substation i at time t in the operation scenario k; is the VSC loss coefficient; is the capacity of the VSC on the i side of the substation; is the absolute value of the sine of the VSC power factor angle.

[0028] Furthermore, the distribution transformer operation constraints in the substation area include:

[0029] ;

[0030] Where: , are respectively the active power on the low-voltage side and high-voltage side of the distribution transformer on the side of the substation i at time t in the operation scenario k; , are the reactive powers of the low-voltage side and high-voltage side of the distribution transformer at the side of the transformer substation i at time t in the operation scenario k; , , and are the photovoltaic operating power, energy storage charging power, energy storage discharging power and discrete capacitor compensation power at the i side of the substation at time t in the operating scenario k; , are the active power and reactive power of the load at the i side of the substation at time t in the operation scenario k; and are the active power and reactive power of the VSC at the side of the substation i at time t in the operation scenario k; , , and are the no-load loss, rated load loss, no-load current percentage and short-circuit voltage percentage of the distribution transformer at side i of the substation respectively; is the load rate of the distribution transformer on the i side of the substation at time t in the operation scenario k; is the capacity of the distribution transformer on the i side of the substation.

[0031] Furthermore, the medium voltage feeder power flow balance constraint includes:

[0032] ;

[0033] Where: represents the set of downstream branches connected to the station area i, represents the set of upstream branches connected to area i; mi represents the line formed by area m flowing to area i, in represents the line formed by area i flowing to area n, ij represents the line formed by area i flowing to area j, is the high-voltage side active power of the distribution transformer at the side of the substation i at time t in the operation scenario k; is the reactive power on the high-voltage side of the distribution transformer on the i side of the substation at time t in the operation scenario k; and are the active power and reactive power injected by the medium voltage distribution network into the distribution transformer at station i at time t in operation scenario k, respectively; , and , line=[mi, in, ij] are the active power, reactive power and current of the corresponding line at time t in operation scenario k; and , line=[mi, ij] are the resistance and reactance of the corresponding line respectively; is the voltage amplitude of substation i / substation j at time t in operation scenario k; (i, j) represents branch ij; The branch set consisting of all medium voltage lines, Indicates the total number of stations.

[0034] Furthermore, the various types of equipment connected to the substation in addition to the source load are one or more of energy storage equipment and discrete capacitors.

[0035] Furthermore, the source load uncertainty fluctuation constraint includes:

[0036] ;

[0037] In the formula, are the power benchmark prediction values ​​of the source and load at the station i at time t in the operation scenario k, are the fluctuation deviation values ​​of the source and load at the station area i at time t in the operation scenario k, are the fluctuation parameters of the source and load at the station i at time t in the operation scenario k, p represents half of the number of divided small intervals, which is a positive integer variable; are the actual power values ​​of the source and load at the station i at time t after the fluctuation in the operation scenario k. The actual power values ​​after the fluctuation are used as endpoints and 0 to form a small interval according to the fluctuation parameter range from small to large. q represents the index number of the small interval formed from left to right according to the actual power values ​​after the fluctuation. Represents dependency parameters, including source dependency parameters Load correlation parameters , They represent the weights of PV output and load in the small interval q respectively; and are the lower and upper limits of the photovoltaic output in the small interval q respectively; f(r) represents the probability density function of photovoltaic output; It represents the number of samples in the load history data that fall within the small interval q, and N represents the total number of samples; The fluctuation parameter and interval weights The penalty function is composed of include and , the larger the interval weight of the fluctuation parameter, the smaller the value of the penalty function.

[0038] Furthermore, the total number of operating scenarios is obtained by clustering in the following manner:

[0039] (1) Randomly select a sample from the PV output and load data sets of all the substations as the first initial cluster center, then calculate the Euclidean distance from all the remaining samples to the current cluster center, and select the sample with the farthest distance as the next cluster center in turn until m0 initial cluster centers are determined;

[0040] (2) Calculate the distances from all samples in the original data set to the m0 cluster centers in the high-dimensional space, and assign each sample to the cluster corresponding to the cluster center closest to it;

[0041] (3) Given parameter α N , if the number of samples in a cluster is less than α N , then the cluster is discarded; after all clusters are determined, the number of current clusters is set to m and the distances from all samples in the original data set to the m cluster centers are recalculated in the high-dimensional space, and each sample is assigned to the cluster corresponding to the cluster center closest to it;

[0042] (4) Determine the current number of clusters. If the current number of clusters is , then the number of clusters is too small, and the split operation is started; if the number of current clusters is , the number of clusters is too large, and the merging operation is performed; the splitting operation is to set the splitting distance according to the maximum component of the standard deviation vector of the samples of the corresponding cluster, and then subtract and add the splitting distance from the original cluster center of the corresponding cluster to obtain two new cluster centers; the merging operation is to merge all the cluster centers whose distance between each two cluster centers is less than the threshold, and the merged new cluster center is the mean of the sum of the products of the original cluster centers of the corresponding two clusters and the number of samples in the cluster;

[0043] (5) When the clustering result no longer changes or the maximum number of iterations is reached, clustering ends and each cluster corresponds to a running scenario; otherwise, return to step (2) to continue iterating.

[0044] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a two-stage robust planning and operation method for flexible interconnection of substations is implemented.

[0045] A storage medium containing computer executable instructions, which, when executed by a computer processor, can implement a two-stage robust planning and operation method for flexible interconnection of metro areas.

[0046] Advantages and beneficial effects of the present invention:

[0047] Aiming at the uncertainty of source and load in the problem of flexible interconnection planning of substations, the present invention proposes a flexible interconnection two-stage robust planning and operation model of photovoltaic output-load fluctuation multi-interval uncertainty set. First, the multi-interval fluctuation uncertainty set is used to characterize the fluctuation of photovoltaic output and load; then, a two-stage robust planning and operation model based on FID is constructed, in which the first stage takes the FID investment cost as the optimization target to decide the FID planning scheme and configuration capacity, and the second stage gives the worst source-load operation scenario and optimizes the annual operation cost; and the column and constraint generation algorithm (C&CG) and strong duality theory are used to solve the two-stage robust model. In addition, the present invention also considers the improved clustering of source-load typical days, and adopts the improved ISODATA algorithm to realize the division of source-load typical days, so as to further improve the accuracy of the model. Finally, through the example of 40 substations, it is verified that the proposed method can accurately summarize the source-load output operation scenario within a year, and the description of source-load fluctuation is more realistic, which is conducive to the two-stage robust model to give a more efficient and economical FID planning configuration scheme and scheduling operation scheme, and improve the economic benefits of the system. The introduction of correlation parameters in the uncertain source-load fluctuation model can make the solution obtained in planning operation take into account both economy and robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of a two-stage robust planning and operation method for flexible interconnection of substations according to the present invention;

[0049] Figure 2 This is a schematic diagram of an exemplary low-voltage area interconnection system structure of the present invention;

[0050] Figure 3 It is a schematic diagram of an exemplary source-load multi-interval uncertain set interval division of the present invention;

[0051] Figure 4 It is a flow chart of the two-stage robust planning operation algorithm of the flexible interconnection of the substations of the present invention;

[0052] Figure 5 This is a structural diagram of a 40-zone distribution network calculation example of an embodiment of the present invention;

[0053] Figure 6 A comparison chart of effectiveness indicators of various clustering algorithms in an embodiment of the present invention;

[0054] Figure 7 This is a result diagram of the FID transmission power between substations 8 and 29 and the transformer load rate of substation 29 under the corresponding operation scenario of a typical day 1 of an embodiment of the present invention;

[0055] Figure 8 This is a result diagram of the FID transmission power between substations 13 and 24 and the transformer load rate of substation 13 under the corresponding operation scenario of a typical day 1 of an embodiment of the present invention;

[0056] Fig. 9 It is a source-load output curve diagram of the station area 32 when the operation scene corresponding to the typical day 1 under the scene 3 and the scene 4 of the embodiment of the present invention;

[0057] Fig.10 1 is a curve diagram of the distribution transformer load rate of the substation 30 corresponding to the operation scenario on a typical day 1 under various scenarios of an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The present invention provides a two-stage robust planning and operation method for flexible interconnection of substations, which uses multi-interval box sets to characterize the uncertain fluctuations of photovoltaics and loads, avoiding the problem of strong conservatism of single-interval uncertain sets, thereby more accurately describing the source-load fluctuation scenario, fitting the actual source-load fluctuation, and having a faster solution speed than chance constraints and probability distributions; then establish a two-stage robust planning and operation model for flexible interconnection of low-voltage substations, which is a min-max-min framework. Among them, the first stage takes the lowest FID planning investment cost as the goal, and solves the FID flexible interconnection solution and VSC configuration capacity; the second stage solves the lowest annual operating cost including power purchase cost and network loss under the worst source-load scenario, and optimizes the VSC transmission power and distribution transformer load rate. Thereby, it can fully fit the actual situation of source-load fluctuations, accurately describe the operation scenario of source-load fluctuations, and enhance the rationality and robustness of flexible interconnection planning and operation. In order to make the structure and advantages of the present invention clearer, the content of the present invention will be further described below in conjunction with the accompanying drawings. A two-stage robust planning and operation method for flexible interconnection of substations of the present invention, such as Figure 1 As shown, the following steps are included:

[0059] Step 1: Obtain data related to the flexible interconnection device and the station area;

[0060] The method of the present invention is to set up a number of flexible interconnection devices between a number of substations in order to form an area interconnection system, thereby realizing the interconnection and mutual supply between substations, thereby enhancing the power supply capacity of the substation area, improving the load balancing rate of the distribution transformer and the level of new energy consumption. Figure 2 The figure shows an exemplary structure of a substation interconnection system. The entire substation interconnection system includes several substations, some or all of which are connected through flexible interconnection devices. Photovoltaic, energy storage, discrete capacitors, loads, etc. are connected to the substation, and the substation purchases electricity from the superior power grid through medium-voltage feeders. In a specific implementation scheme, the flexible interconnection device and substation-related data include at least the topological structure of each substation (such as the type of equipment connected and the access location, etc.), the system parameters of the flexible interconnection device and various types of equipment connected to the substation (including the capacity / rated value of various types of equipment, etc.), cost, historical operation data (photovoltaic output and load data sets of each substation, photovoltaic output probability density function, etc.), and different operation scenarios of typical days containing source-load data obtained based on historical operation data.

[0061] In a specific implementation plan, in order to select the source-load typical day that needs to be considered in the flexible interconnection planning of the substation area, an improved ISODATA clustering method is used to divide the source-load typical day, and then form different operation scenarios; the specific contents include:

[0062] (1) From the original data set, i.e., the PV output and load data set of all substations {x1, x2, …, x N}, N represents the total number of samples, one sample is randomly selected as the first initial cluster center c1, and then all the remaining samples x are calculated i The Euclidean distance d(x i ), select the sample with the farthest distance as the next cluster center until m0 initial cluster centers are determined.

[0063] (2) Calculate all samples x in the original data set in high-dimensional space i Go to these m0 cluster centers c respectively j The distance d(x i ,c j ), and assign each sample to the cluster corresponding to the cluster center closest to it, as shown in the following formula:

[0064] (1)

[0065] Where: represents a mapping function that maps sample points in the current space to a high-dimensional space. K(x,y) is the kernel function that satisfies ; represents the sample set of the jth cluster, represents the number of samples in the jth cluster, represents the cluster center of the jth cluster.

[0066] (3) Given parameter α N , if the number of samples in a cluster is less than α N , then discard the cluster and reclassify the samples of the cluster to the cluster corresponding to the nearest other cluster center according to the method in step (2), and let the current number of clusters be m.

[0067] (4) If the number of current clusters , then the number of clusters is too small, and the split operation is started; if the number of current clusters is , then the number of clusters is too large, and the merge operation is started. The specific steps of splitting and merging operations are as follows:

[0068] Split operation:

[0069] a) First calculate the standard deviation vector of samples in each cluster , for cluster The standard deviation vector of the sample, get its maximum component and the corresponding component positions;

[0070] b) Set a splitting parameter di, usually 0.5;

[0071] c) Order ,in The position of the cluster in step a) The location of the maximum component in the sample standard deviation vector;

[0072] d) Cluster Two new cluster centers split and

[0073] They are:

[0074] (2)

[0075] Merge Operation:

[0076] a) Calculate the distances between all cluster centers and arrange them in ascending order from small to large;

[0077] b) Given the minimum distance parameter , if the two cluster centers and The distance between , then the two cluster centers are merged into one cluster center, and the new cluster center is:

[0078] (3)

[0079] (5) When the clustering result no longer changes or the maximum number of iterations is reached, clustering ends; otherwise, return to step (2) to continue iterating.

[0080] Step 2: Construct a flexible interconnection planning and operation model for the substation, which includes an objective function and constraints. The objective function considers two-stage planning and operation, wherein the first stage takes the optimization of the investment and construction cost of the flexible interconnection device as the objective function, and the second stage takes the optimization of the annual operation cost of the system under the uncertain fluctuation of photovoltaic output and AC load as the objective function; the constraints include the operation constraints of the flexible interconnection device, the operation constraints of the substation distribution transformer, the flow balance constraints of the medium-voltage feeder, the operation constraints of various types of equipment connected to the substation except the source and load, the source and load uncertain fluctuation constraints and the safety constraints; in a specific implementation scheme, firstly, based on the flexible interconnection mode of the substation connected to FID, the source and load characteristics of photovoltaic, load, energy storage and discrete capacitors in the substation system are considered to establish a power flow mathematical model and operation constraints, which specifically include:

[0081] (a) FID operation constraints

[0082] A multi-terminal FID contains multiple VSCs. Assuming that the VSC power flows out of the bus in the positive direction, the g-th FID has the following constraints:

[0083] (4)

[0084] (5)

[0085] (6)

[0086] (7)

[0087] In the formula, represents the VSC set associated with the g-th FID; and are respectively the active power and reactive power of the VSC at the side of the grid i at time t in the operation scenario k, is the active power loss of the VSC at the side of the substation i at time t in the operation scenario k; is the VSC loss coefficient; is the capacity of the VSC at the i side of the substation; is the absolute value of the sine of the VSC power factor angle.

[0088] (b) Operation constraints of distribution transformers in substations

[0089] In the area interconnection system, each area distribution transformer has the following power balance constraints:

[0090] (8)

[0091] Where: , are respectively the active power on the low-voltage side and high-voltage side of the distribution transformer on the side of the substation i at time t in the operation scenario k; , are the reactive powers of the low-voltage side and high-voltage side of the distribution transformer at the side of the transformer substation i at time t in the operation scenario k; , , and are the photovoltaic operating power, energy storage charging power, energy storage discharging power and discrete capacitor compensation power at the i side of the substation at time t in the operating scenario k; , are the active power and reactive power of the load at the i side of the substation at time t in the operation scenario k; , , and are the no-load loss, rated load loss, no-load current percentage and short-circuit voltage percentage of the distribution transformer at side i of the substation respectively; is the load rate of the distribution transformer on the i side of the substation at time t in the operation scenario k; is the capacity of the distribution transformer on the i side of the substation.

[0092] (c) Medium voltage feeder power flow balance constraints

[0093] For medium voltage feeders, the branch power flow equation based on Distflow is used to describe it. The phase angles of physical quantities such as voltage, current and power are removed by phase angle relaxation, and the quadratic equality constraints are converted into second-order cone constraints, resulting in the following power flow balance constraints:

[0094] (9)

[0095] Where: represents the set of downstream branches connected to the station area i, represents the set of upstream branches connected to area i; mi represents the line formed by area m flowing to area i, in represents the line formed by area i flowing to area n, ij represents the line formed by area i flowing to area j, and are the active power and reactive power injected by the medium voltage distribution network into the distribution transformer at station i at time t in operation scenario k, respectively; , and , line=[mi, in,ij] are the active power, reactive power and current of the corresponding line at time t in operation scenario k; and , line=[mi, ij] are the resistance and reactance of the corresponding line respectively; is the voltage amplitude of substation i / substation j at time t in operation scenario k; (i, j) represents branch ij; The branch set consisting of all medium voltage lines, Indicates the total number of stations.

[0096] (d) Energy storage operation constraints

[0097] Energy storage operation constraints include charging and discharging power constraints, capacity constraints, and power constraints, as shown in the following formulas:

[0098] (10)

[0099] Where: , are binary variables representing the charging and discharging states of the energy storage at the station area i at time t in the operation scenario k. By limiting the sum of the two to no more than 1, it is ensured that the energy storage cannot be in the charging and discharging states at any time; and are the maximum charging and discharging powers of the energy storage at side i of the station area, respectively; is the amount of energy stored on the side of substation i at time t in operation scenario k; is the capacity of energy storage on the i side of the substation; is the charging efficiency of the energy storage on the i side of the station; and are the minimum and maximum charge states of the energy storage on side i of the substation respectively; T represents the operation cycle.

[0100] (e) Discrete capacitor operation constraints

[0101] For discrete capacitors, there are constraints on the number of input groups and capacity, as shown in the following formulas:

[0102] (11)

[0103] In the formula, is the number of capacitor banks put into operation at time t in substation i in operation scenario k; is the reactive capacity of a single group of capacitors; The maximum number of capacitor groups that can be put into use.

[0104] (f) Constraints on source-load uncertainty fluctuations

[0105] Based on the theory of multi-interval uncertainty sets, the present invention establishes a multi-interval fluctuation uncertainty set of photovoltaic output and load, which is used to describe the source-load fluctuation in long-term planning scenarios.

[0106] First, divide the source and load multi-interval uncertainty set intervals into 2p small intervals, and use the following formula to describe the endpoint value of each small interval: :

[0107] (12)

[0108] Where: subscripts PV and Load represent photovoltaic output and load respectively. is the power benchmark prediction value of the photovoltaic or load at time t in the operation scenario k at the station i, is the fluctuation deviation value between individual cells, is the fluctuation parameter, which is an integer variable, representing The small interval after the fluctuation. Figure 3 As shown in the figure, the actual power value after the fluctuation is taken as the endpoint and (0) constitutes a small interval, and q represents the index number of the small interval formed from left to right according to the actual power value after fluctuation.

[0109] Considering that the photovoltaic output and load have their own time series correlation, that is, the difference in fluctuations in adjacent time periods will not be too large, the correlation parameters are introduced To characterize the strength of the source-load timing correlation, as shown in the following formula:

[0110] (13)

[0111] From the above formula, we can see that the correlation parameter The larger it is, the more times the source load fluctuates in adjacent time periods within the entire period T, that is, the weaker the correlation between adjacent time periods; otherwise, the stronger it is.

[0112] In order to fit the fluctuation of source load in reality, it is necessary to determine the weight of each small area where the fluctuation falls. Existing literature generally assumes that it follows a normal distribution. This approximate distribution is too ideal and does not make full use of historical data. Therefore, it is considered to discretize the probability density function of photovoltaic output and divide the photovoltaic output into multiple small areas. The weight of each small area can be obtained through integral operation. , as shown below:

[0113] (14)

[0114] Where: and are the lower and upper limits of the photovoltaic output r in the small interval q respectively. For load fluctuations, the number of samples in the load history data that fall in the small interval q is Divide by the total number of samples N as the weight of the cell Obviously, photovoltaic output and load tend to fluctuate to a small interval with a larger weight. Since the probability constraint is difficult to solve, a weight penalty function is introduced to characterize the tendency of source-load fluctuation, as shown in the following formula:

[0115] (15)

[0116] Where: Represented by the fluctuation parameter and interval weights The penalty function is constructed. The larger the interval weight of the fluctuation parameter, the smaller the value of the penalty function. Adding the above formula to the objective function of the two-stage robust optimization model can constrain the source load fluctuation to be more inclined to take the interval with a larger weight.

[0117] In summary, the model of source-load multi-interval uncertainty set is given as follows:

[0118] (16)

[0119] In the formula, are the power benchmark prediction values ​​of the source (photovoltaic) and load (load) at the i-th station in the operation scenario k at time t, is the fluctuation deviation value between the source and load at the station area i at time t in the operation scenario k, are the fluctuation parameter vectors of the source and load at the station i at time t in the operation scenario k, p represents half of the number of divided small intervals, which is a positive integer variable; are the actual power values ​​of the source and load at the station i at time t after the fluctuation in the operation scenario k. The actual power values ​​after the fluctuation are used as endpoints and 0 to form a small interval according to the fluctuation parameter range from small to large. q represents the index number of the small interval formed from left to right according to the actual power values ​​after the fluctuation. Represents correlation parameters, including source (photovoltaic) correlation parameters Load correlation parameters , and They represent the weights of PV output and load in the small interval q respectively; and are the lower and upper limits of the photovoltaic output in the small interval q respectively; f(r) represents the probability density function of photovoltaic output; It represents the number of samples in the load history data that fall within the small interval q, and N represents the total number of samples; The fluctuation parameter and interval weights The penalty function is composed of include and , the larger the interval weight of the fluctuation parameter, the smaller the value of the penalty function.

[0120] (g) Security constraints

[0121] Safety constraints include upper and lower limits of the transformer area voltage, upper and lower limits of the line current, and upper and lower limits of the transformer area power purchase. They are shown in the following formulas:

[0122] (17)

[0123] Where: and are the lower and upper limits of the voltage amplitude of the station area i respectively; is the upper limit of the current amplitude on line ij; is the maximum active power purchased by the station area i from the main grid; and are the minimum and maximum reactive power purchased from the main grid by substation i respectively.

[0124] (2) Taking FID investment cost and annual operation cost as optimization targets, a two-stage planning and operation model of flexible interconnection is established. The optimization target of the first stage is FID investment cost, which is used to give the optimal FID planning scheme and configuration capacity; the optimization target of the second stage is annual operation cost, including the operation cost of various equipment components, distribution network power purchase cost and network loss cost.

[0125] The flexible interconnection planning and operation model of this area aims to minimize the annual comprehensive cost consisting of FID investment and construction costs and system annual operation costs, among which the system annual operation costs include VSC, photovoltaic, energy storage and discrete capacitor operation costs, as well as the cost of power purchase from the distribution network to the upper main network, network loss costs, etc. The planning and operation model is solved under the constraints of satisfying the power balance of various components, upper and lower limits of parameters, and source-load fluctuations, so as to provide a flexible interconnection solution and VSC configuration capacity that fits the actual situation and improve the economy of FID planning and construction.

[0126] Considering two-stage planning and operation, the first stage optimizes the FID investment and construction cost, and the second stage considers the optimization of the system's annual operating cost under the uncertain fluctuations of PV output and AC load.

[0127] (a) Objective function of the first stage

[0128] The objective function of the first stage optimization is as follows:

[0129] (18)

[0130] Where: is the investment cost of FID, which is related to the configuration capacity, economic service life and unit capacity investment cost of VSC, as shown in the following formula:

[0131] (19)

[0132] Where: The VSC capacity installed for zone i, is the investment cost per unit capacity of VSC, year is the economic service life of VSC, and dr is the discount rate. Since VSC is a back-to-back type, the DC tie line is not considered.

[0133] (b) Second stage objective function

[0134] The second stage needs to consider the annual operating cost under the severe scenario caused by the volatility of photovoltaic output and load, as well as the source-load fluctuation weight penalty function. In a specific implementation scheme, the 365 days of the year are divided into several typical day sets to form each operation scenario by improving the ISODATA clustering algorithm in the above step 1. The annual operating cost is obtained by superimposing the costs of each operation scenario, i.e., each typical day set. Adding the weight penalty function, the overall objective function is as follows:

[0135] (20)

[0136] Where: Cost of purchasing electricity from the main grid for the distribution network; The network loss cost includes medium voltage line loss, VSC operation loss and distribution transformer active power loss; The operating costs of the substation components include photovoltaic PV, VSC, energy storage ESS and discrete capacitor CB; is the weight penalty function. As shown in the following formulas:

[0137] (twenty one)

[0138] (twenty two)

[0139] (twenty three)

[0140] (twenty four)

[0141] (25)

[0142] (26)

[0143] (27)

[0144] (28)

[0145] Where: represents the kth cluster in the typical day set, that is, the number of days included in the kth operation scenario, and n is the total number of operation scenarios; and They represent the total number of substations and the total number of medium voltage feeders, respectively, and T is the operation period; is the electricity price at time t; is the active power injected by the medium voltage distribution network into the distribution transformer at area i at time t in operation scenario k; is the line power loss of medium voltage feeder i at time t in operation scenario k; is the active power loss of the VSC at the side of the substation i at time t in the operation scenario k, are the operating costs of photovoltaic PV, VSC, energy storage ESS and discrete capacitor CB respectively; , , , and They are the PV operating power, VSC active power, energy storage charging power, energy storage discharging power and discrete capacitor compensation power at the i side of the substation at time t in the operating scenario k; , , and are the unit operating power costs of PV, VSC, energy storage, and discrete capacitors, respectively; is the unit time; for the weight penalty function , which means using the source-charge fluctuation parameter The value of is used to index the known weight sequence , which can constrain the fluctuation parameters when solving the severe scenario in the second stage max part Get as much weight as possible Larger value.

[0146] Step 3: Solve the flexible interconnection planning and operation model of the substation based on the acquired flexible interconnection device and substation related data to obtain the planning configuration and scheduling operation plan of the flexible interconnection device.

[0147] In step 1, the present invention establishes a two-stage robust optimization model for flexible interconnection of substations based on the consideration of uncertain fluctuations in multiple intervals of source and load. The main problem is the planning and configuration stage, and the goal is to solve the flexible interconnection plan and FID configuration capacity under the worst source and load operation scenario; the sub-problem is the actual operation stage, and according to the flexible interconnection planning and capacity configuration plan given by the main problem, the worst source and load operation scenario that fits the actual source and load fluctuation is found, and the scheduling operation plan that minimizes the annual operation cost under the source and load operation scenario is solved.

[0148] Furthermore, based on the two-stage robust planning operation model theory and the optimization objectives of the main problem and sub-problems, the flexible interconnection planning operation process can be described by a min-max-min two-stage three-level problem. The objective functions of the main problem and sub-problem are shown in equations (18) and (20), respectively. Combined with the constraints mentioned above, the robust model is constructed as shown in the following equation:

[0149] (29)

[0150] Where: x is the decision variable of the first stage; y is the operation variable of the second stage; d is the source-load operation scenario variable; X is the set of all feasible FID planning schemes after the area connectability analysis; Y is the source-load multi-interval fluctuation uncertainty set, as shown in formula (16); U(x, d) represents the feasible domain of the operation variable y when the decision variable x and the source-load scenario variable d are given, and its compact form is shown as follows:

[0151] (30)

[0152] In the formula: from top to bottom, they represent the inequality constraints, equality constraints and second-order cone constraints in the planning process; A, B, C, and D represent the coefficient matrices related to the variable x or y respectively; h and d correspond to the constant column vectors on the right side of the inequality constraints and equality constraints respectively; f represents the coefficient column vector related to the variable y in the second-order cone constraint; , , and denote the dual variables of each constraint.

[0153] For the two-stage robust optimization model established above, this paper uses the C&CG algorithm to solve it. Based on equations (29) and (30), the main problem form of the two-stage robust model is as follows:

[0154] (31)

[0155] In the formula, and are the coefficient column vectors of the objective functions of the main problem and sub-problem respectively; Represents the intermediate variable used to replace the objective function of the sub-problem; is the current iteration number; is the solution of the subproblem after L iterations; It represents the worst source-load operation scenario in L iterations, that is, the operation scenario where the random fluctuation of source-load leads to the maximum operation cost C2.

[0156] The sub-problem form is as follows:

[0157] (32)

[0158] The max-min double-layer form of the subproblem is an NP-hard problem and is difficult to solve with traditional planning algorithms. Therefore, according to the strong duality theory and the variable correspondence of formula (29), the min form of the inner layer of the subproblem is converted into the max form, and after merging it with the outer layer max, the following dual form can be obtained:

[0159] (33)

[0160] In the above formula, the objective function has a bilinear term , which makes the problem non-convex and difficult to solve. Therefore, based on the source-load uncertainty fluctuation model (16), the specific expression of the source-load scenario variable d can be obtained as follows:

[0161] (34)

[0162] In the formula, and are the fluctuation parameters of photovoltaic output and load, respectively, which are integer variables. Substituting the source-load scenario variable d corresponding to the above formula into the objective function of formula (32), the product of integer variables and continuous variables will appear. Therefore, the large M method and auxiliary variables are used to linearize it, which is convenient for solving, and we can get:

[0163] (35)

[0164] In the formula, , represents the power prediction benchmark value of photovoltaic output and load; ; , u and is an auxiliary variable introduced; M is a sufficiently large positive number.

[0165] Based on the above derivation, the nonlinear robust problem that was difficult to solve before can be transformed into a linear mixed integer main problem (31) and subproblem (32), which are in min form and max form respectively, and can be solved by alternating iterations using the C&CG algorithm. The solution process is as follows: Figure 4 As shown, the description is as follows:

[0166] (1) Import the distribution network operation parameters and source-load related operation data, and initialize the lower bound LB of the annual comprehensive operation cost as , the upper bound UB is , set the number of iterations L=1.

[0167] (2) Initialize a set of source and load scenario variables , substitute into the main problem model to solve the optimal FID planning result , and the optimal comprehensive cost given by the main problem As the lower bound LB.

[0168] (3) The optimal planning result in step (2) Substitute it into the sub-problem model to solve it and get the optimal operating cost of the sub-problem And the updated worst-case source load scenario , and Used to update the upper bound UB.

[0169] (4) Given a convergence threshold , if the difference between the upper and lower bounds No more than , the iteration stops and the current optimal solution is returned , and the worst scenario ; Otherwise, let L=L+1, and add the following variables and constraints to the main problem model:

[0170] (36)

[0171] Then return to step (2) to continue solving the main problem and iterate backward until convergence stops.

[0172] The present invention uses a multi-interval fluctuation uncertainty set to characterize the fluctuations of photovoltaic output and load; then, a two-stage robust planning and operation model based on FID is constructed, in which the first stage takes the FID investment cost as the optimization target to decide the FID planning scheme and configuration capacity, and the second stage gives the worst source-load operation scenario and optimizes the annual operation cost; and uses the column and constraint generation algorithm (C&CG) and strong duality theory to solve the two-stage robust model. This method can accurately summarize the source-load output operation scenario within a year, and the description of the source-load fluctuation is more realistic, which is conducive to the two-stage robust model to give a more efficient and economical FID planning configuration scheme and scheduling operation scheme, thereby improving the economic benefits of the system. The introduction of correlation parameters in the source-load fluctuation uncertainty model enables the solution of the planning operation to take into account both economy and robustness.

[0173] The effect of the present invention is further described below in conjunction with a specific embodiment:

[0174] by Figure 2 The low-voltage area interconnected system structure is used as the calculation example, photovoltaic, energy storage, discrete capacitors and different types of loads are connected in the corresponding area, and the uncertain fluctuations of photovoltaic output and load are considered. The effect of the improved method proposed in the present invention is verified through this example:

[0175] The example of the present invention adopts a 40-station distribution network structure, such as Figure 5 As shown, the type classification of each substation and the access location of photovoltaic, energy storage and discrete capacitors are shown. The voltage level on the high-voltage side is 10kV, the voltage level on the low-voltage side is 380V, and the power reference value is 1MVA. Considering the differences in the types of loads connected to different substations, the loads of the other 39 substations in the distribution network except the root substation (substation 1) are divided into three types, including industrial loads, commercial loads and residential loads. The actual output of photovoltaic power does not exceed the given rated value; the maximum charging and discharging power of energy storage is 100kW, and the capacity is 500kWh. In order to prevent excessive charging and discharging of energy storage, its state of charge is limited to between 10% and 90%, the initial state of charge is 50%, the charging efficiency is 0.9, and the discharge efficiency is 1.11; the capacity of a single discrete capacitor is 25kVar, and the maximum gear is 8, that is, the discrete capacitor group can compensate for a maximum of 200kVar reactive power.

[0176] The service life of the VSC in the FID device is set to 20 years, the discount rate dr is 0.06, the maximum capacity of a single VSC is 500kV·A, and the loss coefficient is 0.02. MATLAB is used to call Gurobi to solve, and the running memory is 32G.

[0177] The simulation analysis is as follows:

[0178] (1) Comparison of typical day clustering algorithms

[0179] In order to compare the clustering effects of different typical day clustering algorithms, the traditional K-means algorithm, ISODATA algorithm and the improved ISODATA algorithm proposed in one embodiment of the present invention are compared in terms of clustering effectiveness indicators. Figure 6 shown.

[0180] As can be seen from the figure, the improved ISODATA algorithm has a smaller DBI index and a larger DI index. Therefore, the samples within the cluster are more similar, and the samples between clusters are more different, showing a better clustering effect and being able to more accurately describe the source load sample data within a year.

[0181] (2) Analysis of results of two-stage robust planning operation model

[0182] After solving the two-stage robust planning operation model proposed in the present invention, the flexible interconnection planning schemes of the substations are obtained as follows: (8, 29), (8, 31), (12, 18), (12, 20), (13, 24) and (24, 30), where the numbers represent the substation numbers; the VSC capacity configuration schemes are shown in Table 1.

[0183] Table 1 VSC capacity configuration scheme

[0184]

[0185] According to the planning results, photovoltaic access is available in both substations 8 and 30, and interconnection with other substations that are not connected to photovoltaics can achieve full photovoltaic absorption; the interconnection between other substations that are not connected to photovoltaics is mainly for the timing complementary characteristics of different types of loads, so as to smooth out the load rate differences between distribution transformers in each substation, such as the interconnection between residential load substation 12 and commercial load substation 20, and the interconnection between commercial load substation 24 and industrial load substation 13.

[0186] The comparison of the annual comprehensive cost and corresponding cost composition before and after the flexible interconnection planning is shown in Table 2 below:

[0187] Table 2 Cost results before and after flexible interconnection planning

[0188]

[0189] As can be seen from the table above, after the flexible interconnection planning, the annual comprehensive cost of the system has been greatly reduced, reducing by 3.24756 million yuan, reflecting that the access of flexible interconnection devices can significantly optimize the economic benefits of the system. The annual electricity purchase cost has been effectively reduced by 26.67%; the annual network loss cost has been reduced by 115,140 yuan, which is because the access of FID has optimized the active power flow distribution of the system and reduced active power loss; the annual operating cost has increased, and the increase is mainly due to the loss cost of the connected FID device during operation.

[0190] The two substation interconnection schemes (8, 29) and (13, 24) are selected for operation analysis. Figure 7 It shows the changes in the FID transmission power between areas 8 and 29 and the load factor of the distribution transformer in area 29 before and after the interconnection under the corresponding operation scenario on typical day 1; Figure 8 It shows the changes in the FID transmission power between areas 13 and 24 and the distribution transformer load factor of area 13 before and after the interconnection under the corresponding operating scenario on a typical day 1.

[0191] Depend on Figure 7 It can be seen that since substation 8 is connected to photovoltaic power, while substation 29 is not, there is a large difference in net load between the two substations. It is necessary to transmit the excess power of substation 8 to substation 29 through FID to avoid abandonment of light and promote photovoltaic consumption. Therefore, the transmission power of FID is basically negative, and the maximum transmission power in a single period reaches 112.1kW. With the reactive support function of FID, the load factor of the distribution transformer in substation 29 is significantly reduced, especially from 9 to 18 o'clock when the photovoltaic output is large. The load factor of the distribution transformer in substation 29 is below 0.2, which effectively reduces the risk of continuous overload of the distribution transformer in the substation. Figure 8 In the case of the power grid, although neither area 13 nor area 24 is connected to photovoltaic power, the different load types make the load rates of the two areas quite different. Therefore, the area interconnection through FID can fully take into account the differences in the timing characteristics of different types of loads, and make up for this difference in timing characteristics through the flexible flow of power between the stations. Specifically, from 1 to 7 in the morning and from 19 to 24 at night, factories are generally in a shutdown state, while some shopping malls are still in operation. Therefore, area 13 supplies a small amount of electricity to area 24 through FID to support the operation of commercial loads, and the load rate of distribution transformers has increased. During the day, that is, from 8 to 18 o'clock, the start of factory operation has greatly increased the industrial load, exceeding the commercial load in the same period. Therefore, area 24 transmits electricity to area 13 through FID, which greatly alleviates the burden of distribution transformers in area 13 and significantly reduces its load rate.

[0192] The above analysis shows that flexible interconnection of substations can alleviate the overload phenomenon of distribution transformers in the substations, optimize the distribution of distribution network flow to reduce network losses, increase the absorption of distributed photovoltaic power generation, and improve operating efficiency.

[0193] (3) Analysis of results of different planning and operation schemes

[0194] This paper sets up four example scenarios to analyze and compare the solution results under different planning and operation schemes. Each scenario is as follows:

[0195] Scenario 1: Deterministic planning is adopted, and the optional area interconnection set is not updated;

[0196] Scenario 2: Deterministic planning is adopted, and the interconnection set update of the substation area is optional;

[0197] Scenario 3: single-interval uncertainty planning is adopted, and the interconnection set of substations can be updated;

[0198] Scenario 4: The solution of the present invention, i.e., multi-interval uncertainty planning, and optional update of the interconnection set of substations.

[0199] Among them, the deterministic planning in scenarios 1 and 2 refers to not considering the volatility of photovoltaic output and load in the two-stage robust operation, but using the benchmark forecast value as the typical daily data. The single-interval uncertainty planning in scenario 3 uses the traditional box uncertainty set, and the fluctuation of the uncertainty can only take values ​​at the two endpoints of the uncertainty range.

[0200] The planning and operation results under various scenarios are shown in Table 3 below:

[0201] Table 3 Planning and operation results of each scenario

[0202]

[0203] Table It represents the annual standard deviation of the load rate of the distribution transformer in the substation area, and is used to measure the load balance of the distribution transformers in each substation area. The calculation formula is as follows:

[0204] (37)

[0205] In the formula, It represents the average value of the distribution transformer load rate in each substation at time t in operation scenario k.

[0206] As shown in Table 3, compared with scenario 1, the annual comprehensive cost of scenario 2 increased by 720,800 yuan, and each single cost increased. This is because the dynamic update of the optional area interconnection set expanded the optional range of FID planning schemes and increased the configuration capacity of FID to better meet the requirements of power complementarity and balanced load between stations, so the FID investment cost increased by 163,300 yuan. Although some economic benefits were sacrificed, the load balance of the distribution transformer in the area increased by 4.457%. Compared with scenario 2, the annual comprehensive costs of scenarios 3 and 4 were greatly reduced, by 1,834,580 yuan and 2,425,810 yuan respectively, and the FID investment cost and annual operating cost were significantly compressed. This is because the introduction of the photovoltaic output-load uncertainty fluctuation and weight penalty function makes the planning and operation process more in line with the actual source and load situation, can make full use of photovoltaic output to smooth the peak load, optimize the selection of FID planning configuration schemes and the dispatching and operation of distribution transformers, energy storage and other equipment, thereby reducing network losses and improving operating efficiency.

[0207] The actual source-load output curve at the station 32 in the typical day 1 corresponding to the operation scenario under scenario 3 and scenario 4 is as follows: Fig. 9As shown. It can be seen that compared with single-interval uncertain fluctuations, multi-interval uncertain fluctuations can take into account the internal range of the upper and lower limits of the source and load output, avoiding the problem that the single-interval uncertain fluctuation can only take values ​​at the boundary, causing the optimization result to be too conservative. Therefore, compared with scenario 3, the annual comprehensive cost and each single cost of scenario 4 continue to be lower, and It decreased by 12.214%, and the load balancing effect of the distribution transformer was the best in the four scenarios.

[0208] Take area 30 as an example. Fig.10 The load factor changes of distribution transformers in the corresponding operation scenarios of typical day 1 without flexible interconnection planning and four planned operation scenarios are shown. As can be seen from the figure, when FID is not connected, the load factor of distribution transformers in substation 30 is always stable at a high level of about 0.63, with the risk of heavy load and overload. In the scenarios connected to FID, the load factor of distribution transformers from 8 am to 8 pm is significantly reduced, and the lowest level can reach below 0.3; the load factor in other periods has increased, which is to alleviate the heavy load problem of distribution transformers in the opposite substations connected to substation 30, but it has not exceeded 0.8, so the risk of continuous heavy load of distribution transformers is effectively suppressed as a whole. From 9 to 18 o'clock, compared with scenarios 1 and 2, the load factors of distribution transformers in scenarios 3 and 4 are lower and the fluctuations are more drastic. This is because the net load difference can be more fully utilized after considering the uncertain fluctuations of source load, and the power mutual assistance of FID can more effectively support the power supply of the substation; at the same time, due to the existence of source load fluctuations, the operation and dispatch of distribution transformers in the substation fluctuate to track the uncertain changes of source load. The load factor fluctuation in scenario 4 is smoother than that in scenario 3, thanks to the fact that the multi-interval uncertainty set of source load is more accurate and realistic than the single-interval uncertainty set.

[0209] In summary, the two-stage robust planning and operation method proposed in the present invention takes into account the typical day division of source and load based on the improved ISODATA algorithm and the uncertain fluctuation of source and load in multiple intervals. It can more accurately summarize the source and load operation scenarios within a year and describe the source and load fluctuations, making the source and load scenario changes in the planning and operation process more in line with reality. The FID planning and operation scheme given by the two-stage robust model is more reasonable and accurate, thereby improving the system economy and operation efficiency and reducing various costs and losses.

[0210] Corresponding to the aforementioned embodiment of a two-stage robust planning and operation method for flexible interconnection of substations, the present invention also provides an electronic device, including one or more processors, for implementing a two-stage robust planning and operation method for flexible interconnection of substations in the above embodiment.

[0211] The device embodiment can be implemented by software, hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by the processor of any device with data processing capability in which it is located to read the corresponding computer program instructions in the non-volatile memory into the memory for execution. From the hardware level, it mainly includes a processor, memory, network interface, and non-volatile memory. In addition, any device with data processing capability in which the device in the embodiment is located can also include other hardware according to the actual function of the device with data processing capability, which will not be described in detail.

[0212] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0213] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of the present invention. Ordinary technicians in this field can understand and implement it without paying creative work.

[0214] An embodiment of the present invention further provides a computer-readable storage medium on which a program is stored. When the program is executed by a processor, a two-stage robust planning operation method for flexible interconnection of substations in the above embodiment is implemented.

[0215] The computer-readable storage medium may be an internal storage unit of any device with data processing capability described in any of the aforementioned embodiments, such as a hard disk or a memory. The computer-readable storage medium may also be any device with data processing capability, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), an SD card, a flash card (Flash Card), etc. equipped on the device. Furthermore, the computer-readable storage medium may also include both an internal storage unit of any device with data processing capability and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capability, and may also be used to temporarily store data that has been output or is to be output.

[0216] Those skilled in the art to which the present invention relates may make various modifications or additions to the specific embodiments described or replace them in a similar manner, but they will not deviate from the spirit of the present invention or exceed the scope defined by the attached claims.

Claims

1. A two-stage robust planning and operation method for flexible interconnection of substations, characterized in that: include: Obtain data related to flexible interconnection devices and stations; Construct a flexible interconnection planning and operation model for the substation area, the flexible interconnection planning and operation model for the substation area includes an objective function and constraints, wherein the objective function considers two-stage planning and operation, wherein the first stage takes the optimization of the investment and construction cost of the flexible interconnection device as the objective function, and the second stage takes the optimization of the annual operation cost of the system under the uncertain fluctuation of photovoltaic output and AC load as the objective function; the constraints include the operation constraints of the flexible interconnection device, the operation constraints of the substation area distribution transformer, the flow balance constraints of the medium-voltage feeder, the operation constraints of various types of equipment connected to the substation area except the source load, the source load uncertain fluctuation constraints and the safety constraints; Based on the acquired flexible interconnection device and substation related data, the substation flexible interconnection planning and operation model is solved to obtain the planning configuration and scheduling operation plan of the flexible interconnection device.

2. The method according to claim 1, characterized in that The objective function includes: The objective function C1 of the first stage is as follows: ； Where: is the investment cost of the flexible interconnection device; The objective function C2 of the second stage is as follows: ； ； Where: Cost of purchasing electricity from the main grid for the distribution network; The network loss cost includes medium voltage line loss, VSC operation loss and distribution transformer active power loss; The operating costs of the equipment and components connected to the substation, including photovoltaic PV, VSC, energy storage ESS and discrete capacitor CB; is the weight penalty function, , , and They represent the fluctuation parameters of photovoltaic and load at time t in the operation scenario k, and They respectively represent the weights of photovoltaic and load corresponding to the small interval q at time t in station i in operation scenario k.

3. The method according to claim 1, characterized in that The flexible interconnection device includes multiple VSCs. Assuming that the power of the VSC flows out of the bus in the positive direction, the operation constraints of the flexible interconnection device include: ； ； ； ； In the formula, represents the VSC set associated with the g-th flexible interconnection device; and are respectively the active power and reactive power of the VSC at the side of the grid i at time t in the operation scenario k, is the active power loss of the VSC at the side of the substation i at time t in the operation scenario k; is the VSC loss coefficient; is the capacity of the VSC at the i side of the substation; is the absolute value of the sine of the VSC power factor angle.

4. The method according to claim 1, characterized in that: The distribution transformer operation constraints in the substation area include: ； Where: , are respectively the active power on the low-voltage side and high-voltage side of the distribution transformer on the side of the substation i at time t in the operation scenario k; , are the reactive powers of the low-voltage side and high-voltage side of the distribution transformer at the side of the transformer substation i at time t in the operation scenario k; , , and are the photovoltaic operating power, energy storage charging power, energy storage discharging power and discrete capacitor compensation power at the i side of the substation at time t in the operating scenario k; , are the active power and reactive power of the load at the i side of the substation at time t in the operation scenario k; and are the active power and reactive power of the VSC at the side of the substation i at time t in the operation scenario k; , , and are the no-load loss, rated load loss, no-load current percentage and short-circuit voltage percentage of the distribution transformer at side i of the substation respectively; is the load rate of the distribution transformer on the i side of the substation at time t in the operation scenario k; is the capacity of the distribution transformer on the i side of the substation.

5. The method according to claim 1, characterized in that The medium voltage feeder power flow balance constraints include: ； Where: represents the set of downstream branches connected to the station area i, represents the set of upstream branches connected to area i; mi represents the line formed by area m flowing to area i, in represents the line formed by area i flowing to area n, ij represents the line formed by area i flowing to area j, is the high-voltage side active power of the distribution transformer at the side of the substation i at time t in the operation scenario k; is the reactive power on the high-voltage side of the distribution transformer on the i side of the substation at time t in the operation scenario k; and are the active power and reactive power injected by the medium voltage distribution network into the distribution transformer at station i at time t in operation scenario k, respectively; , and , line=[mi, in, ij] are the active power, reactive power and current of the corresponding line at time t in operation scenario k; and , line=[mi, ij] are the resistance and reactance of the corresponding line respectively; is the voltage amplitude of substation i / substation j at time t in operation scenario k; (i, j) represents branch ij; The branch set consisting of all medium voltage lines, Indicates the total number of stations.

6. The method according to claim 1, characterized in that The various types of equipment connected to the substation area in addition to the source load are one or more of energy storage equipment and discrete capacitors.

7. The method according to claim 1, characterized in that The source load uncertainty fluctuation constraints include: ； In the formula, are the power benchmark prediction values ​​of the source and load at the station i at time t in the operation scenario k, are the fluctuation deviation values ​​of the source and load at the station area i at time t in the operation scenario k, are the fluctuation parameters of the source and load at the station i at time t in the operation scenario k, p represents half of the number of divided small intervals, which is a positive integer variable; are the actual power values ​​of the source and load at the station i at time t after the fluctuation in the operation scenario k. The actual power values ​​after the fluctuation are used as endpoints and 0 to form a small interval according to the fluctuation parameter range from small to large. q represents the index number of the small interval formed from left to right according to the actual power values ​​after the fluctuation. Represents dependency parameters, including source dependency parameters Load correlation parameters , They represent the weights of PV output and load in the small interval q respectively; and are the lower and upper limits of the photovoltaic output in the small interval q respectively; f(r) represents the probability density function of photovoltaic output; It represents the number of samples in the load history data that fall within the small interval q, and N represents the total number of samples; Represented by the fluctuation parameter and interval weights The penalty function is composed of include and , the larger the interval weight of the fluctuation parameter, the smaller the value of the penalty function.

8. The method according to claim 1, characterized in that The total number of running scenarios is obtained by clustering in the following way: (1) Randomly select a sample from the PV output and load data sets of all the substations as the first initial cluster center, then calculate the Euclidean distance from all the remaining samples to the current cluster center, and select the sample with the farthest distance as the next cluster center in turn until m0 initial cluster centers are determined; (2) Calculate the distances from all samples in the original data set to the m0 cluster centers in the high-dimensional space, and assign each sample to the cluster corresponding to the cluster center closest to it; (3) Given parameter α N , if the number of samples in a cluster is less than α N , then the cluster is discarded; after all clusters are determined, the number of current clusters is set to m and the distances from all samples in the original data set to the m cluster centers are recalculated in the high-dimensional space, and each sample is assigned to the cluster corresponding to the cluster center closest to it; (4) Determine the current number of clusters. If the current number of clusters is , then the number of clusters is too small, and the split operation is started; if the number of current clusters is , the number of clusters is too large, and the merging operation is performed; the splitting operation is to set the splitting distance according to the maximum component of the standard deviation vector of the samples of the corresponding cluster, and then subtract and add the splitting distance from the original cluster center of the corresponding cluster to obtain two new cluster centers; the merging operation is to merge all the cluster centers whose distance between each two cluster centers is less than the threshold, and the merged new cluster center is the mean of the sum of the products of the original cluster centers of the corresponding two clusters and the number of samples in the cluster; (5) When the clustering result no longer changes or the maximum number of iterations is reached, clustering ends and each cluster corresponds to a running scenario; otherwise, return to step (2) to continue iterating.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements a two-stage robust planning and operation method for flexible interconnection of substations as described in any one of claims 1-8.

10. A storage medium comprising computer executable instructions, which, when executed by a computer processor, implement a two-stage robust planning and operation method for flexible interconnection of substations as described in any one of claims 1 to 8.

Images