Dynamic Reconfiguration Method of AC / DC Hybrid Distribution Network with Battery Energy Storage and Demand-side Response

Abstract

The present invention provides a method for dynamic reconstruction of an AC / DC hybrid distribution network with battery energy storage and demand-side response. By sampling the operating area of ​​the charging and discharging power, the battery energy storage model under non-ideal conditions is reconstructed into a linear model in the sampling point set of the convex envelope. On this basis, cycle aging cost models based on cumulative throughput and discharge depth are established respectively, and an AC / DC hybrid distribution network reconstruction model with minimized comprehensive cost is constructed. The mixed integer second-order cone programming method is used to solve it, and the function of reducing the comprehensive operating cost is achieved. The present invention takes into account the cycle aging cost and demand-side response of battery energy storage. The charging and discharging process of battery energy storage directly affects its cycle aging, which avoids the drawbacks of the battery model under ideal conditions and is conducive to embedding into the reconstruction model of the AC / DC distribution network.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network reconstruction, and in particular relates to a method for dynamic reconstruction of an AC / DC hybrid distribution network with battery energy storage and demand-side response. Background Art

[0002] Distribution network reconstruction is to change the network topology by switching the states of tie switches and section switches, thereby achieving the goals of reducing network losses and improving power supply reliability. In recent years, with the large-scale grid connection of distributed photovoltaics and battery energy storage, the rapid development of AC / DC interconnection forms and load demand-side response, the reconstruction technology of AC / DC hybrid distribution networks faces many challenges.

[0003] Based on the existing research on the reconstruction of AC / DC hybrid distribution networks, although the three types of models, namely distributed power sources (sources), load demand side response (load), and battery energy storage (storage), are embedded in the AC / DC hybrid distribution network reconstruction model, there is a lack of research on the reconstruction of AC / DC hybrid distribution networks that simultaneously considers distributed power sources (sources), load demand side response (load), and battery energy storage (storage); secondly, the battery energy storage model used in the AC / DC hybrid distribution network reconstruction model is an ideal operating model, which does not consider the inherent electrochemical operating mechanism of battery energy storage, and does not consider the cycle aging cost.

[0004] The existing battery energy storage models are all operating models under ideal conditions, assuming that the battery's charging and discharging efficiency is a fixed value, which is contrary to the nonlinear equivalent circuit model of the battery unit. For this reason, some studies have established a nonlinear equivalent circuit model of the battery unit in combination with the internal charging and discharging process of the battery, and linearized it to obtain a non-ideal state model of battery energy storage. Some literature introduces the non-ideal state model of battery energy storage into the microgrid energy management system and further considers the uncertainty of distributed photovoltaics, proposing a random scheduling method for microgrids. However, almost no literature considers the non-ideal state model of battery energy storage in the research of AC / DC hybrid distribution network reconstruction. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a dynamic reconstruction method of an AC / DC hybrid distribution network with battery energy storage and demand-side response, so as to reduce the overall operating cost.

[0006] The technical solution adopted by the present invention to solve the above technical problems is: a method for dynamic reconstruction of an AC / DC hybrid distribution network with battery energy storage and demand-side response, comprising the following steps:

[0007] S1: Sampling the operating area of ​​the charge and discharge power, and reconstructing the battery energy storage model under non-ideal conditions into a linear model in the sampling point set of the convex envelope;

[0008] S2: construct a dynamic reconstruction model of the AC / DC distribution network, and couple the battery energy storage model constructed in step S1 with the power flow model;

[0009] S3: A mixed integer second-order cone programming method is used, Yalmip of the Matlab toolbox is used to build the model, and a commercial solver is called to solve the problem.

[0010] According to the above scheme, in step S1, the specific steps are:

[0011] S11: From the internal charging and discharging process of the battery, the upper and lower limits of the charging and discharging power are combined with the charging and discharging current and the input power P of the battery. in , the battery output power P out Establish relationships and build a linear model of battery energy storage under non-ideal conditions;

[0012] S12: Establish a battery energy storage cycle aging cost model.

[0013] Furthermore, in the step S11, the specific steps are:

[0014] S111: According to the discharge power P of the battery dis and S SOC Get the battery output power P out ,

[0015] According to the battery charging power P ch and S SOC Get the battery input power P in ;

[0016] S112: Fitting the output power P of the battery through the linear combination of sampling points out ; Sampling j samples to construct the polyhedron envelope, each sample group is used express;

[0017] Fit the battery input power P through the linear combination of sampling points in , sample k samples to construct the polyhedron envelope, and each sample group is express;

[0018] The black dots represent the sampling points, the connecting lines between the points represent the convex envelope connecting the sampling points, and the solid area represents the calculated values ​​based on the actual nonlinear equation;

[0019] S113: By establishing the battery input power P in , the battery output power P out The relationship between the charge and discharge power is used to obtain the linear model of battery energy storage under non-ideal conditions:

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] set up Respectively represent the input power and output power of the battery energy storage at node i at time t; the convex hull formed by the charge and discharge sampling points defines the operating area of ​​the battery energy storage: the discharge sample data of the discharge curve Used to define the discharge period variables Feasible domain, charging sample data of charging curve Used to define variables during charging The feasible domain of It is considered as a linear combination of battery energy storage samples;

[0026]

[0027] The above formula is the battery energy storage state constraint, which is composed of the battery charging state sample and the discharge state sample;

[0028] Without introducing 0-1 variables to identify the charge and discharge status of battery energy storage, additional conditions are introduced: each sampling set must have at least two sampling points of [0,0,0] and [0,1,0], representing the situation where the battery is not activated but fully discharged or fully charged, respectively;

[0029]

[0030]

[0031] The above formula is the charge-discharge association constraint, where x j,t and k,t It is a variable associated with the charge and discharge sample data.

[0032] Furthermore, in step S12, the specific steps are:

[0033] S121: Let q represent the percentage of battery energy storage capacity loss, B 1 and B 2 is the exponential factor obtained by fitting the experimental data, I C represents the battery charging rate, that is, the ratio of the current to the battery capacity during battery charging and discharging; then the cycle aging cost model based on the cumulative throughput power is:

[0034]

[0035] Let C B,0 is the initial investment cost of battery energy storage, η is the capacity percentage at the end of battery energy storage life; then the battery energy storage cycle aging cost C BESS for:

[0036]

[0037] S122: For a given cycle life at a given depth of discharge, the cycle aging cost model is:

[0038]

[0039] Let ρ t Indicates the percentage of life loss corresponding to a specific discharge depth in the cycle, ρ t =1 / Φ(D DOD ), then the cyclic aging cost per time step is for:

[0040]

[0041] Only when the battery energy storage is discharged in the next time step, ρ t Greater than ρ t-1 When D DOD When it increases, the loss of battery energy storage life increases; each discharge half cycle increases the aging cost, while the aging cost in the charging half cycle is 0;

[0042] Assume tp Representation and and ρ t The relevant weight factors, A is the adjacency matrix given by SOS2, b tp It is an additional binary variable to ensure the SOS2 property. SOS2 is used to perform piecewise linearization to express the fitting function of the discharge depth and the life loss percentage:

[0043]

[0044]

[0045]

[0046]

[0047]

[0048] According to the above scheme, in step S2,

[0049] Let C represent the comprehensive cost function of distribution network reconstruction; L represent the set of all branches of the distribution network; T represent the time set of distribution network reconstruction; Represents the node set of distributed photovoltaic access; α l,t is a 0-1 variable, indicating the disconnection status of branch l in the AC / DC distribution network at time t. 0 indicates disconnection and 1 indicates closing. ij,t , P ji,t They represent the active power flowing from node i to node j and the active power flowing from node j to node i in branch ij of the distribution network at time t respectively; represents the available power of distributed photovoltaic at node i at time t, represents the active power actually injected by distributed photovoltaic at node i at time t, cl oss 、c s 、c PV They represent the network loss cost coefficient, switch action cost coefficient, and abandoned light cost coefficient respectively; the AC / DC distribution network reconstruction target of minimizing the comprehensive cost including network loss cost, switch action cost, abandoned light cost, and battery energy storage cycle aging cost is:

[0050]

[0051] Furthermore, the step S2 includes the following constraints:

[0052] S21: Polar coordinate power flow constraint; They represent the active and reactive power injected by node i at time t respectively; represents the load of node i after implementing demand response at time t, They represent the reactive load of node i at time t; P i,t , Q i,t They represent the gateway power injected by node i at time t respectively; represents the reactive power compensated by CB at node i at time t; N(i) represents the set of all nodes connected to node i, I l Represents the current amplitude of branch l; at the same time, R l,t =V i,t V j,t cosθ ij,t , T l,t =V i,t V j,t sinθ ij,t , where V i,t 、V j,t They represent the voltage amplitude of nodes i and j at time t respectively; g l , b l represents the conductance and admittance to ground of line l, θ ij,trepresents the voltage phase angle at both ends of line ij at time t; due to u i,t Does not contain branch switch variable α l,t , introduce the branch switch variable To represent the variables jointly determined by the branch switch state and the branch terminal voltage; the second-order cone relaxation power flow model in the polar coordinate system is:

[0053]

[0054]

[0055]

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062]

[0063] When branch l is disconnected at time t, and are all 0, and when branch l is closed at time t, and and the corresponding node voltage u i,t and u j,t equal; in order to express and u i,t 、u j,t The relationship between Respectively represent the maximum voltage amplitude of nodes i and j, and add additional constraints as follows:

[0064]

[0065] set up represents the maximum current amplitude of branch l, and the node voltage and branch current constraints are:

[0066]

[0067]

[0068] S22: Set represents the reactive power and capacity of the distributed photovoltaic connected to the node i at time t, then the distributed photovoltaic operation constraints are:

[0069]

[0070]

[0071] S23: Constraints for on-load tap-changing transformers; transformer branch ij includes branches im and mj, branch mj only contains the tap-changing gear, assuming t ij Indicates the transformation ratio of transformer ij; T ij is the optimal gear of the transformer, which is an integer variable; Δt ij K is the change value of each level of transformer tap; ij is the number of gears of the transformer tap; are the upper and lower limits of the transformer ratio respectively; then the voltage relationship between the two ends of the transformer is:

[0072]

[0073]

[0074] 0≤T ij ≤K ij ,

[0075]

[0076] S24: AC / DC converter constraints; the converter branch ik includes the AC branch ij and the DC branch jk; suppose the active power of the converter AC side at time t is Reactive power is The reference point voltage is The active power on the DC side of the converter at time t is The reference point voltage is set up is a set of converter branches, then the branch power flow equation containing AC / DC converters is:

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083] Let Kc Indicates DC voltage utilization, M i represents the modulation degree, then the voltage amplitude relationship between the AC side and the DC side of the converter is:

[0084]

[0085] Assume η Conv represents the conversion efficiency, then the active power relationship between the AC side and the DC side of the converter is:

[0086]

[0087] S25: Radial topology constraint; introduction of β ij,t and β ji,t Two 0-1 variables determine whether the first and last nodes i and j of a branch are child nodes or parent nodes to ensure that the generated topology has a spanning tree connected to the root node, β ij,t =1 means node i is the parent node of node j, β ij,t =0 means no; N represents the set of all nodes in the distribution network, N s Represents the root node set; ST constraints are used to describe the radial topology constraints of the distribution network:

[0088]

[0089] S26: Let α l0 is the initial state of the tie switch of branch l before reconstruction, n SA It represents the maximum permissible value of the number of actions of all tie switches in the system, m is the total number of tie switches, and the number of tie switch actions is:

[0090]

[0091] S27: Set It represents the number of capacitor banks put into operation at node i at time t; Indicates the maximum number of capacitors in node i; C i,step represents the capacity of a single group of capacitors at node i; represents the set of all installed capacitor bank nodes; represents the maximum value of the change in the number of capacitor switching groups in adjacent time periods; then the group switching capacitor operation constraint is:

[0092]

[0093]

[0094] S28: Set i,t represents the electricity price elasticity coefficient of node i at time t; represents the change in electricity demand before and after node i implements demand response at time t; Δρt t represents the change in electricity price before and after node i implements demand response at time t; ρ t , They represent the electricity prices before and after the implementation of demand response at time t; They represent the upper and lower limits of electricity prices after implementing demand response; represents the active load of node i at time t; N i represents the node for demand-side response; based on the electricity price elasticity coefficient, the price incentive load is regulated, and the demand-side response model is:

[0095]

[0096]

[0097]

[0098]

[0099] Since the last formula belongs to nonlinear constraints, it is converted into a second-order cone form:

[0100]

[0101] A computer storage medium stores a computer program executable by a computer processor, wherein the computer program executes a method for dynamically reconfiguring an AC / DC hybrid distribution network with battery energy storage and demand-side response.

[0102] The beneficial effects of the present invention are:

[0103] 1. The method for dynamic reconstruction of an AC / DC hybrid distribution network with battery energy storage and demand-side response of the present invention samples the operating area of ​​the charging and discharging power, and reconstructs the battery energy storage model under a non-ideal state into a linear model in a set of sampling points of a convex envelope. On this basis, a cycle aging cost model based on the cumulative throughput power and the discharge depth is established respectively, and an AC / DC hybrid distribution network reconstruction model with minimized comprehensive cost is constructed, and a mixed integer second-order cone programming method is used to solve it; the function of reducing the comprehensive operating cost is achieved.

[0104] 2. The present invention takes into account the cycle aging cost and demand-side response of battery energy storage. The charging and discharging process of battery energy storage directly affects its cycle aging. The battery energy storage model under non-ideal conditions is essentially data-driven modeling based on sample measurements, which avoids the drawbacks of the battery model under ideal conditions and does not introduce binary variables to characterize the charging and discharging status, thereby reducing the complexity of the battery energy storage model and facilitating its embedding into the AC / DC distribution network reconstruction model.

[0105] 3. The present invention establishes a cycle aging cost model for battery energy storage based on the cumulative throughput and discharge depth, and constructs an AC / DC distribution network reconstruction target that minimizes comprehensive costs such as network loss cost, switch action cost, abandoned light cost, battery energy storage cycle aging cost, and demand-side response. The AC / DC hybrid distribution network flow constraints, network security constraints, topology constraints, and load demand-side response based on electricity price incentives are used as constraints, and the AC / DC hybrid distribution network reconstruction model is converted into a mixed integer second-order cone programming model for solution.

[0106] 4. Based on the second-order cone power flow model of AC / DC hybrid distribution network, the present invention considers the cycle aging cost of battery energy storage and the load demand side response, and proposes a dynamic reconstruction method for AC / DC hybrid distribution network. The main conclusions obtained include:

[0107] (1) The cycle aging cost model based on cumulative throughput and the cycle aging cost model based on discharge depth proposed in this invention

[0108] The cycle aging cost model based on the degree of discharge has better economy and effectiveness, and the cycle aging cost model based on the discharge depth has better economy than the cycle aging cost model based on the cumulative throughput capacity.

[0109] (2) The BESS cycle aging cost of AC / DC hybrid distribution network reconstruction decreases as the BESS investment cost decreases, while the network loss cost and switch operation cost are basically unrelated to the BESS investment cost;

[0110] (3) After applying the load demand side response, the original load curve is changed, which increases the load demand during the distributed photovoltaic peak period of 8h-16h, thereby reducing the abandonment of light, reducing the abandonment cost and the total cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0111] Figure 1 The battery energy storage of the embodiment of the present invention is (a) and (b) Operation area diagram.

[0112] Figure 2 4 is a fitting curve diagram of the discharge depth and the life loss percentage of the embodiment of the present invention.

[0113] Figure 3 Schematic diagram of a branch circuit including an on-load tap-changing transformer according to an embodiment of the present invention.

[0114] Figure 4 Schematic diagram of an AC / DC converter station according to an embodiment of the present invention.

[0115] Figure 5 is a load and distributed photovoltaic curve diagram of an embodiment of the present invention.

[0116] Figure 6 is a flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0117] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0118] See also Figure 6 The method for dynamic reconstruction of an AC / DC hybrid distribution network with battery energy storage and demand-side response according to an embodiment of the present invention comprises the following steps:

[0119] The first step is to sample the operating area of ​​the charging and discharging power, and reconstruct the battery energy storage model under non-ideal conditions into a linear model in the sampling point set of the convex envelope.

[0120] (1) Linearization model of battery energy storage under non-ideal conditions

[0121] Since the traditional battery energy storage model assumes that the charging and discharging power of the battery is constant, this will lead to the estimated state of charge (S SOC ) does not match the actual state of charge. Therefore, the present invention establishes a battery energy storage cycle aging cost model based on the cumulative throughput and discharge depth. Different from the traditional battery energy storage model, the battery energy storage linearization model under non-ideal conditions is to convert the upper and lower limits of the charge and discharge power into the charge and discharge current and the battery input power P during the charge and discharge process inside the battery. in , the battery output power P out The relationship is established, and by sampling the operating area of ​​the charge and discharge power, the reconstructed battery model is converted into a linear model in the sampling point set of the convex envelope. Therefore, the linearized battery energy storage model under non-ideal conditions is also called a measurement-based battery energy storage model. The modeling process of the linearized battery energy storage model under non-ideal conditions is described in detail below.

[0122] First, due to P out Depends on the battery discharge power P dis and S SOC , so according to P dis and S SOC Get P out Similarly, we can also calculate the charging power P of the battery. ch and S SOC Get P in .

[0123] Secondly, P out It can be approximated by a linear combination of sampling points. Therefore, the present invention constructs the polyhedron envelope by sampling j samples, and each sample group is express. and and The 3D diagram is as follows Figure 1 (a), where the solid area represents the calculated value based on the actual nonlinear equation. The sampling point j∈J is represented by a black dot, and the convex envelope connecting the sampling points is represented by the connecting lines between the points. Similarly, the input power P of the battery in It can also be approximated by a linear combination of sampling points, and and The 3D diagram is as follows Figure 1 (b) as shown.

[0124] Finally, by establishing the battery input power P in , the battery output power P out The relationship between the charging and discharging power gives the following non-ideal battery energy storage linearization model:

[0125]

[0126]

[0127]

[0128]

[0129]

[0130] In the formula, They represent the input and output power of the battery energy storage at node i at time t. Among them, the discharge sample data of the discharge curve Can be used to define discharge period variables Feasible domain of charging curve. Charging sample data of charging curve Can be used to define variables during charging In other words, the convex hull formed by the charging and discharging sampling points defines the operating area of ​​the battery energy storage. Therefore, the variable It can be regarded as a linear combination of battery energy storage samples.

[0131]

[0132] This formula is the battery energy storage state constraint, which is composed of battery charging state samples and discharging state samples. If this formula is considered alone, the battery energy storage can be charged and discharged at the same time at time t, which does not meet the battery energy storage operation requirements. Therefore, in order to identify the charging and discharging state of the battery energy storage without introducing 0-1 variables, an additional condition is introduced: each sampling set must have at least two sampling points of [0,0,0] and [0,1,0], representing the situation where the battery is not activated but fully discharged or fully charged, respectively.

[0133]

[0134]

[0135] This formula is the charge-discharge association constraint, where x j,t and k,t It is a variable associated with the charge and discharge sample data.

[0136] (2) Battery energy storage cycle aging cost model

[0137] A. Cycle aging cost model based on cumulative throughput

[0138] The cycle aging cost model based on the cumulative throughput is as follows:

[0139]

[0140] Where: q represents the percentage of battery energy storage capacity loss, B 1 and B 2 is the exponential factor obtained by fitting the experimental data. Here, taking lithium battery as an example, B 1 =0.0013, B 2 =0.3534,I C It indicates the battery charging rate, that is, the ratio of the current to the battery capacity when the battery is charging or discharging.

[0141] Therefore, the battery energy storage cycle aging cost C BESS as follows:

[0142]

[0143] In the formula, C B,0 is the initial investment cost of battery energy storage, and η is the capacity percentage at the end of battery energy storage life.

[0144] B. Cycle aging cost model based on discharge depth

[0145] For a given cycle life at a given depth of discharge, the cycle aging cost model is as follows:

[0146]

[0147] In the formula, is the cyclic aging cost for each time step, which is calculated as follows:

[0148]

[0149] In the formula, ρ t Indicates the percentage of life loss corresponding to a specific discharge depth in the cycle, ρ t =1 / Φ(D DOD ), attached Figure 2 The fitting curve between the depth of discharge and the percentage of life loss is given.

[0150] Only when the battery energy storage is discharged in the next time step, ρ t will be greater than ρ t-1 . Because when D DOD When the discharge depth increases, the loss of battery energy storage life will increase. Therefore, each discharge half cycle will increase the aging cost, while the aging cost in the charging half cycle is 0. From the above formula, it can be seen that the cycle aging cost function based on the discharge depth is related to the discharge depth at the beginning and the discharge depth at the end of each cycle, and has nothing to do with the time step; even when the discharge time of the two cycles is different, if their starting and ending discharge depths are the same, there will be the same cycle aging cost.

[0151] In order to express the fitting function of the discharge depth and the life loss percentage, the present invention adopts SOS2 (a special order set (SOS) constraint of type 2) for piecewise linearization, as shown below:

[0152]

[0153]

[0154]

[0155]

[0156]

[0157] In the formula, ξ tp Representation and and ρt t The relevant weight factors, A is the adjacency matrix given by SOS2, b tp is an additional binary variable that guarantees the SOS2 property.

[0158] Step 2: Construct a dynamic reconstruction model of the AC / DC distribution network and couple the battery energy storage model constructed in the first step with the power flow model.

[0159] The present invention considers the AC / DC distribution network reconstruction goal of minimizing comprehensive costs such as network loss cost, switch operation cost, abandoned light cost, and battery energy storage cycle aging cost, namely:

[0160]

[0161] Where, C represents the comprehensive cost function of distribution network reconstruction; L represents the set of all branches of the distribution network; T represents the time set of distribution network reconstruction; Represents the node set of distributed photovoltaic access; α l,t is a 0-1 variable, indicating the disconnection status of branch l in the AC / DC distribution network at time t. 0 indicates disconnection and 1 indicates closing. ij,t , P ji,t They represent the active power flowing from node i to node j and the active power flowing from node j to node i in branch ij of the distribution network at time t respectively; represents the available power of distributed photovoltaic at node i at time t, represents the active power actually injected by distributed photovoltaic at node i at time t, c loss 、c s 、c PV They represent network loss cost coefficient, switch action cost coefficient and abandoned light cost coefficient respectively.

[0162] (1) Polar coordinate power flow constraints

[0163] The second-order cone relaxation power flow model in polar coordinate system is as follows:

[0164]

[0165]

[0166]

[0167]

[0168]

[0169]

[0170]

[0171]

[0172]

[0173]

[0174] In the formula, They represent the active and reactive power injected by node i at time t respectively; represents the load of node i after implementing demand response at time t, They represent the reactive load of node i at time t. i,t , Q i,t They represent the gateway power injected by node i at time t respectively; represents the reactive power compensated by CB at node i at time t. N(i) represents the set of all nodes connected to node i, Il Represents the current amplitude of branch l. At the same time, R l,t =V i,t V j,t cosθ ij,t , T l,t =V i,t V j,t sinθ ij,t , where V i,t 、V j,t They represent the voltage amplitude of nodes i and j at time t respectively; g l , b l represents the conductance and admittance to ground of line l, θ ij,t represents the voltage phase angle at both ends of line ij at time t. At the same time, due to u i,t Does not contain branch switch variable α l,t Therefore, we further introduce the branch switch variable To represent the variable determined by the branch switch state and the branch terminal voltage.

[0175] Since when branch l is disconnected at time t, and are all 0, and when branch l is closed at time t, and and the corresponding node voltage u i,t and u j,t Equal. In order to express and u i,t 、u j,t Therefore, we add additional constraints:

[0176]

[0177] In the formula, Respectively represent the maximum voltage amplitude of nodes i and j. At the same time, the node voltage and branch current constraints are:

[0178]

[0179]

[0180] In the formula, Indicates the maximum current amplitude of branch l.

[0181] (2) Distributed PV operation constraints

[0182]

[0183]

[0184] In the formula, Represents the reactive power and capacity of the distributed photovoltaic connected to the i node at time t.

[0185] (3) Constraints on on-load tap-changing transformers

[0186] The branch diagram containing the on-load tap-changing transformer is shown in the attached figure. Figure 3 The transformer branch ij can be divided into two parts: branch im and branch mj. Branch mj only contains the voltage adjustment position. The voltage relationship between the two ends of the transformer is as follows:

[0187]

[0188]

[0189] 0≤T ij ≤K ij

[0190]

[0191] Where, t ij represents the transformation ratio of transformer ij, T ij is the optimal gear of the transformer, which is an integer variable; Δt ij K is the change value of each level of transformer tap; ij is the number of gears of the transformer tap; They are the upper and lower limits of the transformer ratio respectively.

[0192] (4) AC / DC converter station constraints

[0193] Attached Figure 4 The schematic diagram of the AC / DC converter station is given. The converter branch ik can be divided into an AC branch ij and a DC branch jk. Assume that the active power of the converter AC side at time t is Reactive power is The reference point voltage is The active power on the DC side at time t is The reference point voltage is Similar to the AC power flow in the polar coordinate system, the branch power flow equation containing the AC / DC converter can be written as:

[0194]

[0195]

[0196]

[0197]

[0198]

[0199]

[0200] In the formula, It is the set of converter branches.

[0201] The relationship between the voltage amplitudes on the AC and DC sides of the converter station is:

[0202]

[0203] In the formula, K c Indicates DC voltage utilization, M i Indicates the modulation system.

[0204] The active power relationship between the AC side and the DC side of the converter station is:

[0205]

[0206] Where η Conv Indicates conversion efficiency.

[0207] (5) Radial topology constraints

[0208] The present invention adopts ST constraint to describe the radial topology constraint of the distribution network, as shown in the following formula:

[0209]

[0210] By introducing β ij,t and β ji,t Two 0-1 variables are used to determine whether the first and last nodes i and j of a branch are child nodes or parent nodes to ensure that the generated topology has a spanning tree connected to the root node, β ij,t =1 means node i is the parent node of node j, β ij,t = 0, then it is not. N represents the set of all nodes in the distribution network, N s Represents a root node collection.

[0211] (6) Number of contact switch operations

[0212]

[0213] In the formula, α l0 is the initial state of the tie switch of branch l before reconstruction, n SA It represents the maximum allowed number of actions of all tie switches in the system, and m is the total number of tie switches.

[0214] (7) Operation constraints of group switching capacitors

[0215]

[0216]

[0217] In the formula, It represents the number of capacitor banks put into operation at node i at time t; Indicates the maximum number of capacitors in node i; C i,step represents the capacity of a single group of capacitors at node i; Represents the set of all installed capacitor bank nodes. Indicates the maximum value of the change in the number of capacitor groups in adjacent time periods.

[0218] (8) Demand-side response model based on electricity price elasticity coefficient

[0219] The present invention regulates price incentive loads based on the electricity price elasticity coefficient, and the demand side response model is described by the following formula:

[0220]

[0221]

[0222]

[0223]

[0224] In the formula, ξ i,t represents the electricity price elasticity coefficient of node i at time t; represents the change in electricity demand before and after node i implements demand response at time t; Δρ t represents the change in electricity price before and after node i implements demand response at time t; ρ t , They represent the electricity prices before and after the implementation of demand response at time t; They represent the upper and lower limits of electricity prices after implementing demand response; N represents the active load of node i at time t. i Indicates the node that performs demand-side response.

[0225] Since the last formula belongs to nonlinear constraints, it needs to be converted into a second-order cone form, namely:

[0226]

[0227]

[0228] Step 3: Use the mixed integer second-order cone programming method, use Yalmip of the Matlab toolbox to build a model, and call the commercial solver to solve it.

[0229] In order to verify the effectiveness of the present invention, the present invention uses an IEEE-33 node example for verification. The wiring diagram of the improved IEEE-33 node system is shown in the attached figure. Figure 5 As shown in Table 1, the parameter settings of PV, BESS, CB, and OLTC connected to the system are given. The system base voltage is 12.66kV, and the base capacity is 10MVA. The active power range of the gateway is 0-2000kW, the reactive power range is 0-2000kvar, the branch current amplitude upper limit is 0.5pu, and the node voltage amplitude is 0.95pu-1.05pu. Figure 5 The expected curves of load and distributed photovoltaic are given, and it is assumed that the output of two distributed photovoltaics is the same, and only the loads in residential and commercial areas are responded to on the demand side. At the start, except for the disconnection of the tie switch line, all lines are closed, the switch cost is 1 yuan / time, the network loss cost coefficient is 0.2 yuan / kWh, and the abandoned light cost coefficient is 0.2 yuan / kWh.

[0230] In order to analyze the effectiveness of the battery energy storage model under non-ideal conditions established by the present invention, four strategies are established by setting whether to consider the cycle aging cost of battery energy storage, namely:

[0231] Strategy 1: Traditional battery energy storage model without considering the energy storage cycle aging cost;

[0232] Strategy 2: Ignoring the aging cost of energy storage cycles, and using a non-ideal battery energy storage model;

[0233] Strategy 3: Cycle aging cost model based on cumulative throughput power;

[0234] Strategy 4: Cycle aging cost model based on discharge depth;

[0235] The reconstruction results of the AC / DC hybrid distribution network of the four strategies are shown in Table 1, where the energy storage investment cost of strategies 3 and 4 is 1 million yuan.

[0236] Table 1 System configuration parameters

[0237]

[0238]

[0239] Table 2 Comparison of various costs of AC / DC hybrid distribution network reconstruction using four strategies

[0240]

[0241] It is not difficult to see from Table 2 that: in terms of network loss cost, strategy 1 is the smallest, and strategies 2, 3, and 4 are basically equal; in terms of abandoned light cost, strategy 4 < strategy 3 < strategy 2 < strategy 1, and the abandoned light cost reflects the absorption capacity of AC / DC hybrid distribution network reconstruction for new energy. The smaller the abandoned light cost means that more new energy is absorbed. Therefore, the adoption of strategy 4 (cycle aging cost model based on discharge depth) to reconstruct the AC / DC distribution network has the best new energy absorption capacity. In terms of the total cost of AC / DC hybrid distribution network reconstruction, although strategies 1 and 2 do not consider the cycle aging cost of battery energy storage, strategy 4 < strategy 3 < strategy 2 < strategy 1. Therefore, the cycle aging cost model based on cumulative throughput and the cycle aging cost model based on discharge depth proposed in the present invention have better economy and effectiveness, and the cycle aging cost model based on discharge depth has better economy than the cycle aging cost model based on cumulative throughput.

[0242] The above embodiments are only used to illustrate the design ideas and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design ideas disclosed by the present invention are within the protection scope of the present invention.

Claims

1. A method for dynamic reconstruction of an AC / DC hybrid distribution network with battery energy storage and demand-side response, characterized in that: The following steps are involved: S1: Sampling the operating area of ​​the charge and discharge power, and reconstructing the battery energy storage model under non-ideal conditions into a linear model in the sampling point set of the convex envelope; the specific steps are: S11: From the internal charging and discharging process of the battery, the upper and lower limits of the charging and discharging power are combined with the charging and discharging current and the input power P of the battery. in , the battery output power P out Establish relationships and build a linear model of battery energy storage under non-ideal conditions; S12: Establish a battery energy storage cycle aging cost model; S2: construct a dynamic reconstruction model of the AC / DC distribution network, and couple the battery energy storage model constructed in step S1 with the power flow model; Let C represent the comprehensive cost function of distribution network reconstruction; L represent the set of all branches of the distribution network; T represent the time set of distribution network reconstruction; Represents the node set of distributed photovoltaic access; α l,t is a 0-1 variable, indicating the disconnection status of branch l in the AC / DC distribution network at time t. 0 indicates disconnection and 1 indicates closing. ij,t , P ji,t They represent the active power flowing from node i to node j and the active power flowing from node j to node i in branch ij of the distribution network at time t respectively; represents the available power of distributed photovoltaic at node i at time t, represents the active power actually injected by distributed photovoltaic at node i at time t, c loss 、c s 、c PV They represent the network loss cost coefficient, switch action cost coefficient, and abandoned light cost coefficient respectively; the AC / DC distribution network reconstruction target of minimizing the comprehensive cost including network loss cost, switch action cost, abandoned light cost, and battery energy storage cycle aging cost is: S3: A mixed integer second-order cone programming method is used, Yalmip of the Matlab toolbox is used to build the model, and a commercial solver is called to solve the problem.

2. The method for dynamic reconstruction of AC / DC hybrid distribution network with battery energy storage and demand-side response according to claim 1 is characterized in that: In the step S11, the specific steps are: S111: According to the discharge power P of the battery dis and S SOC Get the battery output power P out , According to the battery charging power P ch and S SOC Get the battery input power P in ; S112: Fitting the output power P of the battery through the linear combination of sampling points out ; Sampling j samples to construct the polyhedron envelope, each sample group is used express; Fit the battery input power P through the linear combination of sampling points in , sample k samples to construct the polyhedron envelope, and each sample group is express; The black dots represent the sampling points, the connecting lines between the points represent the convex envelope connecting the sampling points, and the solid area represents the calculated values ​​based on the actual nonlinear equation; S113: By establishing the battery input power P in , the battery output power P out The relationship between the charge and discharge power is used to obtain the linear model of battery energy storage under non-ideal conditions: set up Respectively represent the input power and output power of the battery energy storage at node i at time t; the convex hull formed by the charge and discharge sampling points defines the operating area of ​​the battery energy storage: the discharge sample data of the discharge curve Used to define the discharge period variables Feasible domain, charging sample data of charging curve Used to define variables during charging The feasible domain of h is regarded as a linear combination of battery energy storage samples; The above formula is the battery energy storage state constraint, which is composed of the battery charging state sample and the battery discharging state sample. Without introducing 0-1 variables to identify the charging and discharging state of the battery energy storage, an additional condition is introduced: each sampling set must have at least two sampling points of [0,0,0] and [0,1,0], which represent the situation where the battery is not activated but fully discharged or fully charged respectively. The above formula is the charge-discharge association constraint, where x j,t and k,t It is a variable associated with the charge and discharge sample data.

3. The method for dynamic reconstruction of AC / DC hybrid distribution network with battery energy storage and demand-side response according to claim 2 is characterized in that: In the step S12, the specific steps are: S121: Let q represent the percentage of battery energy storage capacity loss, B1 and B2 are exponential factors obtained by fitting experimental data, I C represents the battery charging rate, that is, the ratio of the current to the battery capacity during battery charging and discharging; then the cycle aging cost model based on the cumulative throughput power is: Let C B,0 is the initial investment cost of battery energy storage, η is the capacity percentage at the end of battery energy storage life; then the battery energy storage cycle aging cost C BESS for: S122: For a given cycle life at a given depth of discharge, the cycle aging cost model is: Let ρ t Indicates the percentage of life loss corresponding to a specific discharge depth in the cycle, ρ t =1 / Φ(D DOD ), then the cyclic aging cost per time step is for: Only when the battery energy storage is discharged in the next time step, ρ t Greater than ρ t-1 ; When D DOD When it increases, the loss of battery energy storage life increases; each discharge half cycle increases the aging cost, while the aging cost in the charging half cycle is 0; Assume tp Representation and and ρ t The relevant weight factors, A is the adjacency matrix given by SOS2, b tp It is an additional binary variable to ensure the SOS2 property. SOS2 is used to perform piecewise linearization to express the fitting function of the discharge depth and the life loss percentage:

4. The method for dynamic reconstruction of AC / DC hybrid distribution network with battery energy storage and demand-side response according to claim 1 is characterized in that: The step S2 includes the following constraints: S21: Polar coordinate power flow constraint; They represent the active and reactive power injected by node i at time t respectively; represents the load of node i after implementing demand response at time t, They represent the reactive load of node i at time t; P i,t , Q i,t They represent the gateway power injected by node i at time t respectively; represents the reactive power compensated by CB at node i at time t; N(i) represents the set of all nodes connected to node i, I l Represents the current amplitude of branch l; at the same time, R l,t =V i,t V j,t cosθ ij,t , T l,t =V i,t V j,t sinθ ij,t , where V i,t 、V j,t They represent the voltage amplitude of nodes i and j at time t respectively; g l , b l represents the conductance and admittance to ground of line l, θ ij,t represents the voltage phase angle at both ends of line ij at time t; due to u i,t Does not contain branch switch variable α l,t , introduce the branch switch variable To represent the variables jointly determined by the branch switch state and the branch terminal voltage; the second-order cone relaxation power flow model in the polar coordinate system is: When branch l is disconnected at time t, and are all 0, and when branch l is closed at time t, and and the corresponding node voltage u i,t and u j,t Equal; to express and u i,t 、u j,t The relationship between i max , Respectively represent the maximum voltage amplitude of nodes i and j, and add additional constraints as follows: set up represents the maximum current amplitude of branch l, and the node voltage and branch current constraints are: S22: Set represents the reactive power and capacity of the distributed photovoltaic connected to the node i at time t, then the distributed photovoltaic operation constraints are: S23: Constraints for on-load tap-changing transformers; transformer branch ij includes branches im and mj, branch mj only contains the tap-changing gear, assuming t ij Indicates the transformation ratio of transformer ij; T ij is the optimal gear of the transformer, which is an integer variable; Δt ij K is the change value of each level of transformer tap; ij is the number of gears of the transformer tap; are the upper and lower limits of the transformer ratio respectively; then the voltage relationship between the two ends of the transformer is: 0≤T ij ≤K ij , S24: AC / DC converter constraints; the converter branch ik includes the AC branch ij and the DC branch jk; suppose the active power of the converter AC side at time t is Reactive power is The reference point voltage is The active power on the DC side of the converter at time t is The reference point voltage is set up is a set of converter branches, then the branch power flow equation containing AC / DC converters is: Let K c Indicates DC voltage utilization, M i represents the modulation degree, then the voltage amplitude relationship between the AC side and the DC side of the converter is: Assume η Conv represents the conversion efficiency, then the active power relationship between the AC side and the DC side of the converter is: S25: Radial topology constraint; introduction of β ij,t and β ji,t Two 0-1 variables determine whether the first and last nodes i and j of a branch are child nodes or parent nodes to ensure that the generated topology has a spanning tree connected to the root node, β ij,t =1 means node i is the parent node of node j, β ij,t =0 means no; N represents the set of all nodes in the distribution network, N s Represents the root node set; ST constraints are used to describe the radial topology constraints of the distribution network: S26: Let α l0 is the initial state of the tie switch of branch l before reconstruction, n SA It represents the maximum permissible number of actions of all tie switches in the system, m is the total number of tie switches, and the number of tie switch actions is: S27: Set It represents the number of capacitor banks put into operation at node i at time t; Indicates the maximum number of capacitors in node i; C i,step represents the capacity of a single group of capacitors at node i; represents the set of all installed capacitor bank nodes; represents the maximum value of the change in the number of capacitor switching groups in adjacent time periods; then the group switching capacitor operation constraint is: S28: Set i,t represents the electricity price elasticity coefficient of node i at time t; represents the change in electricity demand before and after node i implements demand response at time t; Δρ t represents the change in electricity price before and after node i implements demand response at time t; ρ t , They represent the electricity prices before and after the implementation of demand response at time t; They represent the upper and lower limits of electricity prices after implementing demand response; represents the active load of node i at time t; N i represents the node for demand-side response; based on the electricity price elasticity coefficient, the price incentive load is regulated, and the demand-side response model is: Since the last formula belongs to nonlinear constraints, it is converted into a second-order cone form:

5. A computer storage medium, characterized in that: A computer program executable by a computer processor is stored therein, and the computer program executes the method for dynamic reconstruction of an AC / DC hybrid distribution network with battery energy storage and demand-side response as described in any one of claims 1 to 4.

Images