Double-layer optimization configuration method for distributed energy storage system

Abstract

The invention relates to the technical field of power distribution networks, and discloses a distributed energy storage system double-layer optimization configuration method, which comprises the following steps: upper-layer optimization: constructing an upper-layer optimization model by taking energy storage comprehensive annual cost and annual income as objective functions; solving the upper-layer optimization model to obtain an optimal result of energy storage locating and sizing; lower-layer optimization: clustering the photovoltaic operation scenes of one year by adopting a K-means clustering algorithm to obtain a classic operation scene; a photovoltaic multi-mode control strategy is utilized to solve the voltage out-of-limit problem caused by photovoltaic power generation; and constructing a lower-layer optimization model by taking the photovoltaic annual income as a target function, and solving the lower-layer optimization model to obtain an optimal operation strategy of energy storage charging and discharging. The method has the advantages that the problem of voltage out-of-limit caused by distributed photovoltaic is solved by using the photovoltaic multi-mode control strategy in a classic photovoltaic operation scene, and compared with a photovoltaic droop control strategy, the method has a better voltage control effect and can improve the consumption capability of the distributed photovoltaic.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution network, and in particular to a double-layer optimization configuration method for a distributed energy storage system. Background Art

[0002] As the penetration rate of photovoltaic power generation in the distribution network increases, the complex power flow distribution will bring the risk of reverse heavy overload to the distribution network, resulting in voltage over-limit problems in the distribution network, affecting the safe and stable operation of the distribution network. At present, in order to solve the voltage over-limit problem caused by the integration of high-penetration distributed photovoltaic power generation into the distribution network, some people have proposed to adopt a new photovoltaic inverter control strategy to improve the photovoltaic inverter's power control ability, maintain the distribution network voltage, and always connect it to the grid with the maximum active power. Some people have proposed to use photovoltaic inverter droop control to adjust the output reactive power and active power to alleviate the voltage over-limit problem caused by photovoltaic power generation.

[0003] However, only using photovoltaic-related control strategies to solve the voltage over-limit problem caused by high-penetration distributed photovoltaic power generation will lead to the problem of abandoned light in the distribution network. The configuration of energy storage can not only coordinate the photovoltaic control strategy to solve the voltage over-limit problem, but also improve the absorption capacity of photovoltaic power generation and improve the economic efficiency of distribution network operation. However, the current energy storage configuration still has the following shortcomings: 1) The photovoltaic penetration rate in the distribution network is small, and the energy storage configuration is relatively fixed and not general; 2) Most of them only focus on the economic efficiency of energy storage and ignore the economic efficiency of photovoltaics; 3) Only using energy storage to solve the voltage over-limit and abandoned light problems without considering the impact of photovoltaic control strategies, resulting in redundant capacity of energy storage, which reduces the economic efficiency of energy storage configuration. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a distributed energy storage system double-layer optimization configuration method. The purpose of the present invention is achieved through the following technical solutions: A distributed energy storage system double-layer optimization configuration method, comprising the following steps:

[0005] Step S1, upper layer optimization, the specific steps are:

[0006] Step S11, taking the comprehensive annual cost and annual income of energy storage as the objective function, considering the ratio relationship between energy storage power and capacity, taking the energy storage investment cost, state of charge and operating state as constraints, constructing an upper optimization model;

[0007] Step S12, using the improved particle swarm algorithm to solve the upper optimization model to obtain the optimal result of energy storage site selection and capacity determination; Step S2, lower layer optimization, the specific steps are:

[0008] Step S21, clustering the photovoltaic operation scenarios for one year using the K-means clustering algorithm to obtain classic operation scenarios;

[0009] Step S22, using the photovoltaic multi-mode control strategy in the classic operation scenario to solve the voltage over-limit problem caused by photovoltaic power generation; Step S23, taking the photovoltaic annual income as the objective function, constructing a lower-level optimization model to coordinate the output cooperation between photovoltaics and energy storage, and solving the lower-level optimization model to obtain the optimal operation strategy for energy storage charging and discharging.

[0010] Specifically, the objective function of the upper optimization model in step S11 is:

[0011]

[0012] Where: objective function f 1,p The total cost of energy storage converted to daily cost; f 2,p is the daily income of energy storage; α is the weight coefficient of total energy storage cost; β is the weight coefficient of energy storage income; α+β=1; P p is the probability of a typical day of scenario p in a year; n p is the number of typical day scenes;

[0013]

[0014] Where: is the total investment cost of energy storage capacity and power; is the total operation and maintenance cost of energy storage; The charging and discharging benefits of energy storage; service fees charged for energy storage; To store energy and reduce the cost of photovoltaic power generation; T s It is the service life of energy storage.

[0015] Specifically, the total investment cost of energy storage capacity and power, the total operation and maintenance cost of energy storage, the charging and discharging income of energy storage, the service fee charged by energy storage, the income from reducing photovoltaic abandonment by energy storage, and the income from reducing photovoltaic abandonment by energy storage are calculated as follows:

[0016]

[0017] Where: c e is the unit capacity cost of energy storage; c p is the unit power cost of energy storage; E capi is the total capacity of energy storage; P capi is the total rated power of energy storage; c op is the annual operation and maintenance cost of energy storage; β is the number of years of energy storage operation; i r is the inflation rate; d r is the discount rate; ch is the time-of-use electricity price for energy storage charging; dis is the time-of-use electricity price of energy storage discharge;se Service fee charged for energy storage; P chi is the charging power of energy storage; P disi is the discharge power of energy storage; n pv is the total number of photovoltaics; n t is the total time of the scheduling cycle; n es is the total amount of energy storage; pvs The service fee charged for reducing the amount of photovoltaic curtailment due to energy storage; △P pv0 and △P pv are the amount of photovoltaic power abandonment when no energy storage is configured and the amount of photovoltaic power abandonment after energy storage is configured; T s It is the service life of energy storage.

[0018] Specifically, the constraints of the upper optimization model include:

[0019] Investment cost, access capacity and power constraints:

[0020]

[0021] Where: and The investment cost and investment cost cap for energy storage; and The upper and lower limits of the energy storage capacity accessed by each node; and N is the upper and lower limits for energy storage site selection; es Selecting a location for energy storage; The rated power of the node energy storage; is the rated capacity of the node energy storage; γ is the energy multiplier of the energy storage;

[0022] Charge and discharge power constraints:

[0023]

[0024] Where: is the rated power of the i-th energy storage; B chi and B disi is the charging state and discharging state of the i-th energy storage; operating state constraints:

[0025]

[0026] Where: E esi (t) is the charge state of the i-th energy storage at time t; η chi and η disi is the charging and discharging efficiency of the i-th energy storage.

[0027] Specifically, the objective function of the lower layer optimization model in step S23 is:

[0028]

[0029] Where: The daily dispatch cycle income of PV in the classic operation scenario; is the penalty fee for abandonment of photovoltaic power in the daily scheduling cycle of the classic operation scenario, n p is the number of typical daily scenes.

[0030] Specifically, the daily scheduling cycle income of photovoltaic power in the classic operation scenario and the daily scheduling cycle abandonment penalty fee of photovoltaic power in the classic operation scenario are calculated as follows:

[0031]

[0032] Where: pv The photovoltaic time-of-use electricity price; The government subsidizes the amount of photovoltaic power generated; lo is the light abandonment penalty coefficient; P mppj (t) is the active power generated by the jth photovoltaic at time t; P pvj (t) is the active power actually output by the jth photovoltaic to the distribution network at time t. Specifically, the constraints of the lower optimization model include:

[0033] Distribution network power balance constraints:

[0034]

[0035] Where: P s is the total power input to the distribution network; P loadn is the load power of n nodes at a certain moment; P esi is the active power output of the i-th energy storage at a certain moment; P pvj is the active power output of the jth photovoltaic at a certain moment; n bat is the number of distribution network nodes; distribution network flow constraints:

[0036]

[0037] Where: P SI (t), P pvI (t), P esI (t), P LI (t) are the active power flowing into node I at time t, the active output of photovoltaic power, the active output of full energy storage, and the active power consumed by node I; Q SI (t), Q pvI (t), Q LI (t) are the reactive power flowing into node I, the PV reactive output, and the reactive power consumed by node I at time t; V I (t), VJ (t) are the voltages of nodes I and J at time t; matrix R IJ and X IJ are the resistance and reactance between node I and node J respectively; δ IJ (t) is the phase angle difference between nodes I and J at time t;

[0038] Distribution network voltage forecast constraints:

[0039] The photovoltaic inverter obtains the output power at the next moment through the photovoltaic multi-mode control strategy, and predicts the node I voltage at time t+1, where P pvI =P pvj ,Q pvI =Q pvj ,

[0040]

[0041] Where: V I is the voltage of node I; △P is the active deviation between the photovoltaic output of node I and the load change; △Q is the reactive deviation between the photovoltaic output of node I and the load change; matrix R II and X II are the resistance and reactance of node I respectively;

[0042] Photovoltaic related constraints:

[0043]

[0044] Where: V min and V max is the upper and lower limits of the node voltage; P mppj is the active power generated by the jth photovoltaic; Q pvj and S pvj is the reactive power and capacity output by the jth PV inverter at a certain moment; is the power factor angle of the photovoltaic inverter; It is the lower limit of the power factor angle of the photovoltaic inverter.

[0045] Specifically, the photovoltaic multi-mode control strategy includes:

[0046] Mode A: Normal working mode

[0047] In normal working mode, the reactive power output by the photovoltaic inverter is:

[0048]

[0049] Where: S pvj is the apparent power of the jth PV inverter; P pvj (t) and Q pvj(t) is the active and reactive power output by the jth PV inverter at time t; is the jth photovoltaic power factor angle;

[0050] Mode B: Coordinated reactive power reduction mode

[0051] In the coordinated reactive power reduction mode, the reactive power output by the PV inverter is:

[0052]

[0053] Where: t DQ Represents the reactive power reduction time constant; Q max,pvj is the reactive power upper limit of the jth photovoltaic output; △t is the sampling time, unit: min; n DQ It is expressed as the number of times reactive power reduction lasts and satisfies t DQ -n DQ △t≥△t, when Q pvj (t) = -Q max,pvj (t) indicates that the reactive power regulation capability of the jth PV inverter has been fully utilized, and the PV inverter outputs inductive reactive power at this time;

[0054] Mode C: Coordinated Active Power Mode

[0055] In the coordinated active power mode, the active power output by the PV inverter is:

[0056]

[0057] Where: t DP Indicates the active power reduction time constant; V tm,pvj represents the voltage on the grid-connected side of the jth photovoltaic cell; t RP Indicates the active power recovery time constant; n DP and n RP They are respectively represented as the number of active power reduction durations and the number of active power recovery durations, and satisfy t DP -n DP △t≥△t and t RP -n RP △t≥△t;

[0058] Mode D: Coordinated reactive power recovery mode

[0059] In coordinated reactive power recovery mode, the inverter outputs capacitive reactive power:

[0060]

[0061] Where: t RQ Indicates the active power recovery time constant; n RQDenoted as the number of times of reactive power recovery, and satisfying t RQ -n RQ △t≥△t;

[0062] In the normal working mode, when V tm,pvj ≥V2 and the photovoltaic power generation increment △P pv >0, enter the coordinated reactive power reduction mode; in the coordinated reactive power reduction mode, when V tm,pvj <V2 and △P pvj <0, enter the coordinated reactive power recovery mode; when V tm,pvj ≥V max enter the coordinated active power mode;

[0063] In the coordinated active power mode, when V tm,pvj <V2, enter the coordinated reactive power recovery mode;

[0064] In the coordinated reactive power recovery mode, when V tm,pvj ≥V2 and △P pvj >0, return to mode B, otherwise, gradually return to the normal working mode.

[0065] Specifically, the following steps are used to solve the upper-layer optimization model and the lower-layer optimization model:

[0066] Step S110: Set the particle swarm parameters, photovoltaic and energy storage parameters, and at the same time load the load data and the photovoltaic classic operation scenario data obtained by the K-means clustering algorithm, and randomly generate an initial population P with a scale of 100 individuals. Each individual contains the positions and rated capacities of 10 randomly generated energy storages;

[0067] Step S120: Use the photovoltaic multi-mode control strategy to obtain the initial scheduled photovoltaic output power in the clustered classic operation scenarios, solve the lower-layer optimization model through the quadratic programming algorithm, obtain the optimal charge and discharge strategies of the energy storage and the optimal photovoltaic output power of each individual in the initial population P within the scheduling period of the classic operation scenario, and then calculate the lower-layer optimization objective function values in each classic operation scenario;

[0068] Step S130: Transmit the energy storage charge and discharge strategies and photovoltaic output power obtained by the lower-layer optimization in each classic operation scenario to the upper-layer optimization model to calculate the upper-layer objective function values;

[0069] Step S140: Calculate the fitness values of the individuals and the overall population, search for the global optimum and the individual optimum, update the particle swarm, obtain the offspring population, update the inertia weight, speed, etc. of the particle swarm algorithm, and at the same time update the number, position and rated capacity of the energy storage; Step S150: Iterative optimization;

[0070] Step S160, repeating steps S130 to S150 until the termination condition is met, and obtaining the final energy storage optimization configuration result;

[0071] Step S170, output the result, output the data such as the number, location, rated capacity and optimal charging and discharging strategy of energy storage corresponding to the optimal population.

[0072] The present invention has the following advantages:

[0073] 1. The present invention uses the photovoltaic multi-mode control strategy to solve the voltage over-limit problem caused by distributed photovoltaics in the clustered photovoltaic classic operation scenario. By comparing with the photovoltaic droop control strategy, the photovoltaic multi-mode control strategy has a better voltage control effect and can improve the absorption capacity of distributed photovoltaics;

[0074] 2. The double-layer optimization configuration method successfully solves the voltage over-limit problem caused by distributed photovoltaic power generation and peak power consumption in the distribution network, and improves the distribution network's ability to absorb distributed photovoltaic power;

[0075] 3. The two-layer optimization configuration method, compared with only using distributed energy storage to solve the voltage over-limit problem caused by distributed photovoltaic power generation, adding photovoltaic control strategy can reduce the redundancy of distributed energy storage configuration capacity and improve the economy of configured energy storage. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 It is a schematic diagram of the double-layer optimization model architecture of the present invention;

[0077] Figure 2 A photovoltaic operation scenario diagram for one year of the present invention;

[0078] Figure 3 This is a photovoltaic classic operation scenario diagram of the present invention;

[0079] Figure 4 The photovoltaic multi-mode switching diagram of the present invention;

[0080] Figure 5 It is a flow chart for solving the double-layer optimization model of the present invention;

[0081] Figure 6 IEEE 34-node test system diagram of the present invention;

[0082] Figure 7 It is a diagram showing the total photovoltaic and total load changes of the present invention;

[0083] Figure 8 It is the voltage rate diagram of each node of the distribution network without control measures of the present invention;

[0084] Fig. 9It is a voltage diagram of each node of the distribution network under the photovoltaic droop control of the present invention;

[0085] Fig.10 It is a voltage diagram of each node of the distribution network under the photovoltaic multi-mode control of the present invention;

[0086] Fig.11 A comparison chart of the annual abandoned photovoltaic power under the two control measures of the present invention;

[0087] Fig.12 The energy storage site selection diagram of the present invention;

[0088] Fig.13 The voltage diagram of each node of the distribution network of the optimal energy storage configuration of the present invention;

[0089] Fig.14 This is a diagram of energy storage power and capacity changes for the optimal configuration of the present invention. DETAILED DESCRIPTION

[0090] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the embodiments described are only part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0091] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0092] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including a..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0093] The present invention is further described below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the following description. Figures 1 to 14 As shown, a distributed energy storage system double-layer optimization configuration method includes the following steps:

[0094] Step S1, upper layer optimization. The upper layer optimization model is mainly responsible for solving the problem of site selection and capacity determination of energy storage, with the purpose of reducing the investment cost of energy storage and increasing the operating income of energy storage, and improving the distribution network's ability to absorb photovoltaic power generation. In order to make energy storage perform better in terms of economic benefits and charging and discharging efficiency, the present invention uses a linear weighting method to convert the multi-objective function into a single objective function. In this process, the upper layer optimization target mainly considers the lowest annual cost of energy storage, while taking into account the maximum annual income;

[0095] The specific steps are:

[0096] Step S11, taking the comprehensive annual cost and annual income of energy storage as the objective function, considering the ratio relationship between energy storage power and capacity, taking the energy storage investment cost, state of charge and operating state as constraints, constructing an upper optimization model;

[0097] The objective function of the upper optimization model is:

[0098]

[0099] Where: objective function f 1,p The total cost of energy storage converted to daily cost; f 2,p is the daily income of energy storage; α is the weight coefficient of total energy storage cost; β is the weight coefficient of energy storage income; α+β=1; P p is the probability of a typical day of scenario p in a year; n p is the number of typical day scenes;

[0100]

[0101] Where: is the total investment cost of energy storage capacity and power; is the total operation and maintenance cost of energy storage; The charging and discharging benefits of energy storage; service fees charged for energy storage; To store energy and reduce the cost of photovoltaic power generation; T s The service life of energy storage;

[0102]

[0103] Where: c e is the unit capacity cost of energy storage; c p is the unit power cost of energy storage; E capi is the total capacity of energy storage; P capiis the total rated power of energy storage; c op is the annual operation and maintenance cost of energy storage; β is the number of years of energy storage operation; i r is the inflation rate; d r is the discount rate; ch is the time-of-use electricity price for energy storage charging; dis is the time-of-use electricity price of energy storage discharge; se Service fee charged for energy storage; P chi is the charging power of energy storage; P disi is the discharge power of energy storage; n pv is the total number of photovoltaics; n t is the total time of the scheduling cycle; n es is the total amount of energy storage; pvs The service fee charged for reducing the amount of photovoltaic curtailment due to energy storage; △P pv0 and △P pv are the amount of photovoltaic power abandonment when no energy storage is configured and the amount of photovoltaic power abandonment after energy storage is configured; T s The service life of energy storage;

[0104] The constraints of the upper optimization model include:

[0105] Investment cost, access capacity and power constraints:

[0106]

[0107] Where: and The investment cost and investment cost cap for energy storage; and The upper and lower limits of the energy storage capacity accessed by each node; and N is the upper and lower limits for energy storage site selection; es Selecting a location for energy storage; The rated power of the node energy storage; is the rated capacity of the node energy storage; γ is the energy multiplier of the energy storage;

[0108] Charge and discharge power constraints:

[0109]

[0110] Where: is the rated power of the i-th energy storage; B chi and B disi is the charging state and discharging state of the i-th energy storage; operating state constraints:

[0111]

[0112] Where: E esi(t) is the charge state of the i-th energy storage at time t; η chi and η disi is the charging and discharging efficiency of the i-th energy storage; step S12, using the improved particle swarm algorithm to solve the upper optimization model to obtain the optimal result of energy storage site selection and capacity determination; step S2, lower layer optimization, the specific steps are:

[0113] Step S21, clustering the photovoltaic operation scenarios for one year using the K-means clustering algorithm to obtain classic operation scenarios;

[0114] Step S22: Use the photovoltaic multi-mode control strategy in the classic operation scenario to solve the voltage over-limit problem caused by photovoltaic power generation; Step S23: Take the annual photovoltaic income as the objective function, build a lower-level optimization model to coordinate the output between photovoltaic and energy storage, and solve the lower-level optimization model to obtain the optimal operation strategy for energy storage charging and discharging. The lower-level optimization first uses the K-means clustering algorithm to cluster the photovoltaic operation scenarios to obtain four classic operation scenarios. The core idea of ​​the K-means clustering algorithm is to cluster e vectors x a (a=1,…,e), divided into E groups G b (b=1,…,E), and find the cluster center of each group so that the value function of the distance index is minimized. When the Euclidean distance is selected as group G b The vector x a and the corresponding cluster center c b When the dissimilarity index is used, the value function J can be defined as:

[0115]

[0116] Based on the actual photovoltaic data of a certain region in a year, such as Figure 2 As shown in the figure, each curve represents the daily power of photovoltaic power generation. Assuming that the prediction error satisfies the normal distribution, the Monte Carlo method is used to randomly generate photovoltaic operation scenarios for one year. The sampling period of photovoltaic power generation is 5 minutes. The target cluster number is set to 4. The classic operation scenarios obtained after clustering are as follows: Figure 3 As shown in Table 1, the annual probability of occurrence of classic photovoltaic operation scenarios is as follows:

[0117] Table 1 Annual occurrence probability of classic operation scenarios

[0118]

[0119] The lower optimization model uses the photovoltaic multi-mode control strategy to regulate the output power of photovoltaics in the four classic operation scenarios of K-means clustering. While ensuring that the voltage of the distribution network does not exceed the limit, it optimizes the charging and discharging operation strategy of distributed energy storage to achieve the maximum absorption of distributed photovoltaics. The lower optimization objective function is to maximize the annual income of photovoltaics.

[0120] The objective function of the lower optimization model is:

[0121]

[0122] Where: The daily dispatch cycle income of PV in the classic operation scenario; is the penalty fee for abandonment of photovoltaic power in the daily scheduling cycle of the classic operation scenario, n p is the number of typical day scenes;

[0123]

[0124] Where: pv The photovoltaic time-of-use electricity price; The government subsidizes the amount of photovoltaic power generated; lo is the light abandonment penalty coefficient; P mppj (t) is the active power generated by the jth photovoltaic at time t; P pvj (t) is the active power actually output by the jth photovoltaic to the distribution network at time t; the constraints of the lower optimization model include:

[0125] Distribution network power balance constraints:

[0126]

[0127] Where: P s is the total power input to the distribution network; P loadn is the load power of n nodes at a certain moment; P esi is the active power output of the i-th energy storage at a certain moment; P pvj is the active power output of the jth photovoltaic at a certain moment; n bat is the number of distribution network nodes.

[0128] Distribution network power flow constraints

[0129]

[0130] Where: P SI (t), P pvI (t), P esI (t), P LI (t) are the active power flowing into node I at time t, the active output of photovoltaic power, the active output of full energy storage, and the active power consumed by node I; Q SI (t), Q pvI (t), Q LI (t) are the reactive power flowing into node I, the PV reactive output, and the reactive power consumed by node I at time t; V I (t), V J (t) are the voltages of nodes I and J at time t; matrix RIJ and X IJ are the resistance and reactance between node I and node J respectively; δ IJ (t) is the phase difference between nodes I and J at time t.

[0131] Distribution network voltage forecast constraints:

[0132] The photovoltaic inverter obtains the output power at the next moment through the photovoltaic multi-mode control strategy, and predicts the node I voltage at time t+1, where P pvI =P pvj ,Q pvI =Q pvj .

[0133]

[0134] Where: V I is the voltage of node I; △P is the active deviation between the photovoltaic output of node I and the load change; △Q is the reactive deviation between the photovoltaic output of node I and the load change; matrix R II and X II are the resistance and reactance of node I respectively.

[0135] Photovoltaic related constraints:

[0136]

[0137] Where: V min and V max is the upper and lower limits of the node voltage; P mppj is the active power generated by the jth photovoltaic; Q pvj and S pvj is the reactive power and capacity output by the jth PV inverter at a certain moment; is the power factor angle of the photovoltaic inverter; It is the lower limit of the power factor angle of the photovoltaic inverter.

[0138] Photovoltaic multi-mode control strategy, the goal of the control algorithm is to give priority to adjusting the reactive power output of the photovoltaic inverter, and to use active power regulation as a last resort to solve the voltage limit problem caused by high-penetration distributed photovoltaic power generation. Compared with the photovoltaic droop control strategy, the multi-mode control strategy has clear logic, can minimize the active power loss of photovoltaics, and can change the power regulation rate according to actual needs. This control strategy is implemented on the photovoltaic inverter, and the switching logic between the modes is as follows Figure 4 As shown, the photovoltaic multi-mode control strategy includes:

[0139] Mode A: Normal working mode. Under normal working conditions, the voltage of all photovoltaic grid-connected sides in the distribution network is lower than the maximum voltage upper limit allowed by the distribution network. The active power output by the photovoltaic inverter operates according to the maximum power point tracking (MPPT) logic, that is, P pvj =P mppj ;

[0140] In normal working mode, the reactive power output by the photovoltaic inverter is adjusted according to the following formula:

[0141]

[0142] Where: S pvj is the apparent power of the jth PV inverter; P pvj (t) and Q pvj (t) is the active and reactive power output by the jth PV inverter at time t; is the jth photovoltaic power factor angle;

[0143] Mode B: Coordinated reactive power reduction mode. In the early stage of photovoltaic power generation, the photovoltaic inverter outputs capacitive reactive power, which helps to increase the photovoltaic grid-connected side voltage V tm,pvj As the active power generated by photovoltaics gradually increases, the capacitive reactive power output by the photovoltaic inverter will also decrease. The control target of this mode is V tm,pvj ≥V2 and photovoltaic power generation increment △P pv >0, the photovoltaic inverter gradually transitions to output inductive reactive power to ease the voltage rise rate on the photovoltaic grid-connected side. In this mode, the active power output by the photovoltaic inverter still operates according to the MPPT logic. After adjusting the time t DQ , so that the reactive power output by the photovoltaic inverter reaches the target value as soon as possible -Q max,pvj ;

[0144] In the coordinated reactive power reduction mode, the reactive power output by the PV inverter is:

[0145]

[0146] Where: t DQ Represents the reactive power reduction time constant; Q max,pvj Indicates the reactive power upper limit of the jth photovoltaic output; △t is the sampling time, unit: min; n DQ It is expressed as the number of times reactive power reduction lasts and satisfies t DQ -n DQ △t≥△t, when Q pvj (t) = -Q max,pvj(t) indicates that the reactive power regulation capacity of the j-th photovoltaic inverter has been fully utilized, and at this time, the photovoltaic inverter outputs inductive reactive power; if V tm,pvj < V2 and △P pvj < 0, it indicates that the photovoltaic power generation has decreased due to weather reasons; Mode C: Coordinated active power mode. When V tm,pvj ≥ V max , at this time, the active power output by the photovoltaic inverter will not follow the MPPT logic operation, but the active power output by the photovoltaic will be set according to the voltage over-limit degree on the photovoltaic grid-connected side, that is, P pv = P set . The control objective of this mode is to ensure that the voltage on the photovoltaic grid-connected side of the distribution network does not exceed the limit while reducing the reduction of the active power output by the photovoltaic. If V tm,pvj ≥ V max , the active power output by the photovoltaic is reduced. If V tm,pvj < V max , the active power output by the photovoltaic is restored, and it is dynamically adjusted to maintain it within the range of V max and below;

[0147] In the coordinated active power mode, the active power output by the photovoltaic inverter is:

[0148]

[0149] In the formula: t DP represents the active power reduction time constant; V tm,pvj is the voltage on the j-th photovoltaic grid-connected side; t RP represents the active power recovery time constant; n DP and n RP respectively represent the active power reduction continuous times and the active power recovery continuous times, and satisfy t DP - n DP △t ≥ △t and t RP - n RP △t ≥ △t;

[0150] Mode D: Coordinated reactive power recovery mode. There are two ways to enter this mode. The first is to enter Mode D from Mode B. The photovoltaic inverter always follows the MPPT logic operation and always satisfies V tm,pvj < V max . The second is to enter Mode D from Mode C. When V tm,pvj < V2, it indicates that the active power output by the photovoltaic has not been in the peak power generation interval, and there will be no situation where V tm,pvj ≥ V maxIn the case of alleviating the rate of decline of the terminal voltage of the photovoltaic, the inverter is adjusted to output capacitive reactive power. In the coordinated reactive power recovery mode, the inverter outputs capacitive reactive power:

[0151]

[0152] where: t RQ represents the active power recovery time constant; n RQ represents the number of times of reactive power recovery continuation, and satisfies t RQ -n RQ △t≥△t;

[0153] In the normal operation mode, when V tm,pvj ≥V2 and the photovoltaic power generation increment △P pv >0, it enters the coordinated reactive power reduction mode; in the coordinated reactive power reduction mode, when V tm,pvj <V2 and △P pvj <0, it enters the coordinated reactive power recovery mode; when V tm,pvj ≥V max it enters the coordinated active power mode;

[0154] In the coordinated active power mode, when V tm,pvj <V2, it enters the coordinated reactive power recovery mode;

[0155] In the coordinated reactive power recovery mode, when V tm,pvj ≥V2 and △P pvj >0, it returns to mode B, otherwise, it gradually returns to the normal operation mode.

[0156] The upper-layer optimization model and the lower-layer optimization model are solved by combining the improved particle swarm optimization (PSO), K-means clustering algorithm, photovoltaic multi-mode control strategy and quadratic programming algorithm. The specific steps are as follows:

[0157] Step S110, initialization, set the particle swarm parameters, photovoltaic and energy storage parameters, and at the same time load the load data and the photovoltaic classic operation scenario data obtained by the K-means clustering algorithm. Use the particle swarm algorithm to generate an initial population P with a scale of 100 individuals, and each individual includes the number of energy storage, position and rated capacity;

[0158] Step S120, the lower layer optimization model, with 24 hours a day as the daily scheduling cycle, the lower layer first uses the photovoltaic multi-mode control strategy in the four classic operation scenarios obtained by clustering to obtain the initial scheduling photovoltaic output power, and solves the lower layer optimization model through the quadratic programming algorithm to obtain the optimal energy storage charging and discharging strategy and the optimal photovoltaic output power of each individual in the initial population P within the scheduling cycle of the four classic operation scenarios, and then calculates the energy storage charging and discharging power and photovoltaic output power in each classic operation scenario in the lower layer;

[0159] Step S130, the upper optimization model transfers the energy storage charging and discharging strategy and photovoltaic output power in each classic operation scenario to the upper optimization model to calculate the upper objective function value;

[0160] Step S140, calculate the fitness values ​​of the individuals and the whole of the population, find the global optimum and the individual optimum, update the particle swarm, obtain the offspring population, update the inertia weight, speed, etc. of the particle swarm algorithm, and update the number, location and rated capacity of the energy storage at the same time; Step S150, iterative optimization;

[0161] Step S160, repeating steps S130 to S150 until the termination condition is met, and obtaining the final energy storage optimization configuration result;

[0162] Step S170, output the result, output the data such as the number, location, rated capacity and optimal charging and discharging strategy of energy storage corresponding to the optimal population.

[0163] In order to verify the effectiveness of the dual-layer optimization configuration method for distributed energy storage systems of the present invention, the IEEE34 node system is used as a test system for simulation verification. Figure 6 In order to reflect the high penetration rate of distributed photovoltaics in the distribution network, no photovoltaics are installed at node 0, and photovoltaic power generation equipment is installed at the remaining nodes. The capacity of each photovoltaic power generation equipment is set to 60kW, and the power factor is The energy storage optimization configuration is carried out in the clustering of the photovoltaic classic operation scenarios. The photovoltaic changes and total load fluctuations in each scenario are shown in the following figure. Figure 7 As shown;

[0164] Photovoltaic and energy storage use a time-of-use electricity price mechanism, which is specifically divided into three time periods: off-peak period, flat period and peak period. The electricity price for each period is different to encourage more use of electricity during off-peak periods and reduce electricity use during peak periods. The specific time-of-use electricity prices are shown in Table 2. The electric energy generated by distributed photovoltaics is first sent to the distribution network to obtain revenue. Distributed energy storage uses time-of-use electricity prices to realize peak-valley cashing and assist photovoltaic power generation, ensuring that there is no voltage over-limit problem in the distribution network, and maximizing the benefits of both. The daily scheduling time of the classic operation scenario is 24h, and the relevant parameters of photovoltaics and energy storage are shown in Table 3;

[0165] Table 2 PV and energy storage electricity price parameters

[0166]

[0167] Table 3 Parameters related to photovoltaic and energy storage

[0168]

[0169]

[0170] The simulation cases are set up as follows: 1) Case 1: No PV control measures are adopted. 2) Case 2: Comparison of voltage control effects of droop control strategy and multi-mode control strategy, without energy storage optimization configuration. 3) Case 3: Energy storage optimization configuration under multi-mode control.

[0171] 1) No photovoltaic control measures are adopted

[0172] Figure 8 The following is a simulation result diagram of the voltage change of each node in the distribution network without the photovoltaic control strategy in Scheme 1. By analyzing the figure, it can be concluded that when the distributed photovoltaic power generation in the distribution network is in scenario 1, the problem of the voltage exceeding the upper limit in the distribution network is relatively serious. At about 13h, the peak voltage reached 1.15pu, while in scenario 2, the peak voltage reached 1.10pu. If photovoltaic control measures are not taken, this will lead to a decline in the power supply quality of the distribution network and affect the stable operation of the distribution network. In addition, when the distribution network is in the peak power consumption period and the photovoltaic power generation is low, the voltage of some nodes in the distribution network is below the lower limit, which may affect some users in the distribution network. Therefore, in order to maintain the stable operation of the distribution network, the photovoltaic control strategy and energy storage are adopted to solve the problem of voltage exceeding the upper limit caused by high-penetration distributed photovoltaic power generation. The configuration of energy storage can also solve the problem of voltage exceeding the lower limit caused by the peak power consumption period of the distribution network.

[0173] 2) Comparative analysis of droop control and multi-mode control

[0174] Fig. 9 and Fig.10 The voltage diagrams of each node of the distribution network are obtained by adopting the photovoltaic droop control strategy and the photovoltaic multi-mode control strategy respectively under the same system. By analyzing and comparing the two diagrams, it is concluded that both control strategies can solve the voltage over-limit problem of distributed photovoltaic power generation in scenarios 1 and 2, but the droop control strategy still has some voltage over-limit situations when the node voltage reaches the voltage upper limit, while the multi-mode control strategy has a better voltage control effect and does not have the voltage over-limit problem. This is because the strategy has begun to adjust the photovoltaic output power before the node voltage reaches the voltage upper limit. In summary, the photovoltaic multi-mode control strategy is superior to the droop control strategy in voltage control.

[0175] Fig.11 This is a comparison chart of the total abandoned light amount of distributed photovoltaic power generation under two control measures in one year. The annual abandoned light amount of the photovoltaic droop control strategy is 1135.111MW, and the annual abandoned light amount of the photovoltaic multi-mode control strategy is 969.638MW. In comparison, the multi-mode control strategy successfully increased the photovoltaic power generation consumption by 165.473MW. In summary, the photovoltaic multi-mode control strategy can reduce the abandoned light amount of photovoltaic power generation.

[0176] 3) Optimal configuration of energy storage under photovoltaic multi-mode control

[0177] Although the two photovoltaic control strategies can effectively solve the problem of voltage over-limit, in the scenario of high-penetration distributed photovoltaic grid-connected, relying solely on the photovoltaic control strategy, there will be a problem of abandoned light in the distribution network, resulting in a waste of resources. In addition, the ideal control effect is not achieved in terms of voltage control of the distribution network. Therefore, based on the photovoltaic multi-mode control strategy with better voltage control effect, the distributed energy storage of the distribution network is optimized. The number of distributed energy storage configurations is iteratively optimized starting from 10, and the final optimized configuration result is Scheme 1. Among them, Scheme 2 is a non-optimal configuration result. The two schemes are compared and analyzed, and the energy storage configuration results are shown in Table 4.

[0178] Table 4 Energy storage configuration results

[0179]

[0180] From the analysis of Table 4, it can be concluded that in terms of the comprehensive cost of distributed energy storage, Scheme 2 has a higher comprehensive cost than Scheme 1, and the energy storage income of Scheme 2 is also less than that of Scheme 1, which means that Scheme 2 has high costs but low income, and the configured distributed energy storage capacity is redundant and not fully utilized. In terms of photovoltaic income, there is little difference between Scheme 1 and Scheme 2. From the perspective of energy storage investors, while ensuring the stability of the distribution network voltage, it can not only improve the comprehensive economy of the distribution network, but also reduce the comprehensive cost of distributed energy storage, so Scheme 1 is the optimal result. In the four classic operating scenarios of distributed photovoltaics, the annual abandoned light of distributed photovoltaics is 969.638MW when relying solely on photovoltaic control strategies. After adopting the optimized configuration of energy storage, the annual abandoned light of distributed photovoltaics is reduced to 7.469MW, reducing the abandoned light of photovoltaics by 99.23%. In summary, the double-layer optimization configuration method proposed in this paper improves the absorption capacity of distributed photovoltaic power generation while taking into account the economic efficiency of photovoltaics and energy storage. Fig.12 This is the energy storage site selection map, and the hexagon represents the energy storage site location.

[0181] Fig.13 This is the voltage change of each node in the distribution network after the energy storage is optimized. Fig. 9 and Fig.10A comparative analysis shows that the configuration of distributed energy storage solves the problem of the voltage of the distribution network exceeding the upper limit caused by distributed photovoltaic power generation and the problem of the voltage of the distribution network exceeding the lower limit caused by peak power consumption.

[0182] Fig.14 The figure shows the changes in power and capacity of distributed energy storage in four classic operating scenarios. As can be seen from the figure, in scenarios 1 and 2, distributed photovoltaics are in peak power generation. While part of the power is consumed by other power facilities in the distribution network, distributed energy storage is also absorbing the active power generated by photovoltaics. Distributed energy storage is charged from the distribution network during the low power consumption period and discharged to the distribution network during the peak power consumption period, thereby obtaining corresponding profits. By analyzing the four classic scenarios, it can be concluded that distributed energy storage can basically meet the daily power cycle needs.

[0183] The above is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any technician familiar with the art can make many possible changes and modifications to the technical solution of the present invention by using the above-mentioned technical content without departing from the scope of the technical solution of the present invention, or modify it into an equivalent embodiment of equivalent changes. Therefore, any changes, modifications, equivalent changes and modifications made to the above embodiments based on the technology of the present invention without departing from the content of the technical solution of the present invention belong to the protection scope of the present technical solution.

Claims

1. A two-layer optimization configuration method for a distributed energy storage system, characterized in that: The following steps are involved: Step S1, upper layer optimization, the specific steps are: Step S11, taking the comprehensive annual cost and annual income of energy storage as the objective function, considering the ratio relationship between energy storage power and capacity, taking the energy storage investment cost, state of charge and operating state as constraints, constructing an upper optimization model; Step S12, using the improved particle swarm algorithm to solve the upper optimization model to obtain the optimal result of energy storage site selection and capacity determination; Step S2, lower layer optimization, the specific steps are: Step S21, clustering the photovoltaic operation scenarios for one year using the K-means clustering algorithm to obtain classic operation scenarios; Step S22: Using a photovoltaic multi-mode control strategy in a classic operation scenario to solve the voltage over-limit problem caused by photovoltaic power generation; Step S23: Taking the annual income of photovoltaic power generation as the objective function, construct a lower-level optimization model to coordinate the output cooperation between photovoltaic power generation and energy storage, and solve the lower-level optimization model to obtain the optimal operation strategy for energy storage charging and discharging.

2. A distributed energy storage system double-layer optimization configuration method according to claim 1, characterized in that: The objective function of the upper optimization model in step S11 is: Where: f 1,p The total cost of energy storage converted to daily cost; f 2,p is the daily income of energy storage; α is the weight coefficient of total energy storage cost; β is the weight coefficient of energy storage income; α+β=1; P p is the probability of a typical day of scenario p in a year; n p is the number of typical day scenes; Where: is the total investment cost of energy storage capacity and power; is the total operation and maintenance cost of energy storage; The charging and discharging benefits of energy storage; service fees charged for energy storage; To store energy and reduce the cost of photovoltaic power generation; T s It is the service life of energy storage.

3. A distributed energy storage system double-layer optimization configuration method according to claim 2, characterized in that: The total investment cost of energy storage capacity and power, the total operation and maintenance cost of energy storage, the charging and discharging income of energy storage, the service fee charged by energy storage, the income from energy storage to reduce photovoltaic abandonment and the income from energy storage to reduce photovoltaic abandonment are calculated as follows: Where: c e is the unit capacity cost of energy storage; c p is the unit power cost of energy storage; E capi is the total capacity of energy storage; P capi is the total rated power of energy storage; c op is the annual operation and maintenance cost of energy storage; β is the number of years of energy storage operation; i r is the inflation rate; d r is the discount rate; ch is the time-of-use electricity price for energy storage charging; dis is the time-of-use electricity price of energy storage discharge; se Service fee charged for energy storage; P chi is the charging power of energy storage; P disi is the discharge power of energy storage; n pv is the total number of photovoltaics; n t is the total time of the scheduling cycle; n es is the total amount of energy storage; pvs The service fee charged for reducing the amount of photovoltaic curtailment due to energy storage; △P pv0 and △P pv are the amount of photovoltaic power abandonment when no energy storage is configured and the amount of photovoltaic power abandonment after energy storage is configured; T s It is the service life of energy storage.

4. A distributed energy storage system double-layer optimization configuration method according to claim 3, characterized in that: The constraints of the upper optimization model include: Investment cost, access capacity and power constraints: Where: and The investment cost and investment cost cap for energy storage; and The upper and lower limits of the energy storage access capacity of each node; and N is the upper and lower limits for energy storage site selection; es Selecting a location for energy storage; The rated power of the node energy storage; is the rated capacity of the node energy storage; γ is the energy multiplier of the energy storage; Charge and discharge power constraints: Where: is the rated power of the i-th energy storage; B chi and B disi is the charging state and discharging state of the i-th energy storage; Running state constraints: Where: E esi (t) is the charge state of the i-th energy storage at time t; η chi and η disi is the charging and discharging efficiency of the i-th energy storage.

5. A distributed energy storage system double-layer optimization configuration method according to claim 1, characterized in that: The objective function of the lower optimization model in step S23 is: Where: The daily dispatch cycle income of PV in the classic operation scenario; is the penalty fee for abandonment of photovoltaic power in the daily scheduling cycle of the classic operation scenario, n p is the number of typical daily scenes.

6. A distributed energy storage system double-layer optimization configuration method according to claim 5, characterized in that: The calculation of the daily scheduling cycle income of photovoltaic power in the classic operation scenario and the daily scheduling cycle abandonment penalty fee of photovoltaic power in the classic operation scenario is as follows: Where: pv The photovoltaic time-of-use electricity price; The government subsidies for photovoltaic power generation; lo is the light abandonment penalty coefficient; P mppj (t) is the active power generated by the jth photovoltaic at time t; P pvj (t) is the active power actually output by the jth photovoltaic unit to the distribution network at time t.

7. A distributed energy storage system double-layer optimization configuration method according to claim 6, characterized in that: The constraints of the lower optimization model include: Distribution network power balance constraints: Where: P s is the total power input to the distribution network; P loadn is the load power of n nodes at a certain moment; P esi is the active power output of the i-th energy storage at a certain moment; P pvj is the active power output of the jth photovoltaic at a certain moment; n bat is the number of distribution network nodes; distribution network flow constraints: Where: P SI (t), P pvI (t), P esI (t), P LI (t) are the active power flowing into node I at time t, the active output of photovoltaic power, the active output of full energy storage, and the active power consumed by node I; Q SI (t), Q pvI (t), Q LI (t) are the reactive power flowing into node I, the PV reactive output, and the reactive power consumed by node I at time t; V I (t), V J (t) are the voltages of nodes I and J at time t; matrix R IJ and X IJ are the resistance and reactance between node I and node J respectively; δ IJ (t) is the phase angle difference between nodes I and J at time t; Distribution network voltage forecast constraints: The photovoltaic inverter obtains the output power at the next moment through the photovoltaic multi-mode control strategy, and predicts the node I voltage at time t+1, where P pvI =P pvj ,Q pvI =Q pvj , Where: V I is the voltage of node I; △P is the active deviation between the photovoltaic output of node I and the load change; △Q is the reactive deviation between the photovoltaic output of node I and the load change; matrix R II and X II are the resistance and reactance of node I respectively; Photovoltaic related constraints: Where: V min and V max is the upper and lower limits of the node voltage; P mppj is the active power generated by the jth photovoltaic; Q pvj and S pvj is the reactive power and capacity output by the jth PV inverter at a certain moment; is the power factor angle of the photovoltaic inverter; It is the lower limit of the power factor angle of the photovoltaic inverter.

8. A distributed energy storage system double-layer optimization configuration method according to claim 1, characterized in that: The photovoltaic multi-mode control strategy includes: Mode A: Normal working mode In normal working mode, the reactive power output by the photovoltaic inverter is: Where: S pvj is the apparent power of the jth PV inverter; P pvj (t) and Q pvj (t) is the active and reactive power output by the jth PV inverter at time t; is the jth photovoltaic power factor angle; Mode B: Coordinated reactive power reduction mode In the coordinated reactive power reduction mode, the reactive power output by the PV inverter is: Where: t DQ Represents the reactive power reduction time constant; Q max,pvj is the reactive power upper limit of the jth photovoltaic output; △t is the sampling time, unit: min; n DQ It is expressed as the number of times reactive power reduction lasts and satisfies t DQ -n DQ △t≥△t, when Q pvj (t) = -Q max,pvj (t) indicates that the reactive power regulation capability of the jth PV inverter has been fully utilized, and the PV inverter outputs inductive reactive power at this time; Mode C: Coordinated Active Power Mode In the coordinated active power mode, the active power output by the PV inverter is: Where: t DP Represents the active power reduction time constant; V tm,pvj represents the voltage on the grid-connected side of the jth photovoltaic cell; t RP Indicates the active power recovery time constant; n DP and n RP They are respectively represented as the number of active power reduction durations and the number of active power recovery durations, and satisfy t DP -n DP △t≥△t and t RP -n RP △t≥△t; Mode D: Coordinated reactive power recovery mode In coordinated reactive power recovery mode, the inverter outputs capacitive reactive power: Where: t RQ Indicates the active power recovery time constant; n RQ It is expressed as the number of times reactive power recovery lasts and satisfies t RQ -n RQ △t≥△t; In the normal working mode, when V tm,pvj ≥ V2 and the photovoltaic power generation increment △P pv > 0, it enters the coordinated reactive power reduction mode; in the coordinated reactive power reduction mode, when V tm,pvj < V2 and △P pvj < 0, it enters the coordinated reactive power recovery mode; when V tm,pvj ≥ V max it enters the coordinated active power mode; In the coordinated active power mode, when V tm,pvj < V2, enter the coordinated reactive power recovery mode; In coordinated reactive power recovery mode, when V tm,pvj ≥V2 and △P pvj >0, returns to mode B, otherwise, gradually returns to normal working mode.

9. A distributed energy storage system double-layer optimization configuration method according to claim 1, characterized in that: The upper optimization model and the lower optimization model are solved by the following steps: Step S110, setting particle swarm parameters, photovoltaic and energy storage parameters, and loading load data and photovoltaic classic operation scenario data obtained by K-means clustering algorithm, randomly generating an initial population P with a size of 100 individuals, each of which contains 10 randomly generated energy storage locations and rated capacities; Step S120, using the photovoltaic multi-mode control strategy in the classic operation scenario obtained by clustering to obtain the initial scheduling photovoltaic output power, solving the lower-level optimization model through the quadratic programming algorithm, obtaining the optimal energy storage charging and discharging strategy and the optimal photovoltaic output power of each individual in the initial population P within the scheduling period of the classic operation scenario, and then calculating the lower-level optimization objective function value in each classic operation scenario; Step S130, transferring the energy storage charging and discharging strategy and photovoltaic output power obtained by the lower-level optimization in each classic operation scenario to the upper-level optimization model to calculate the upper-level objective function value; Step S140, calculate the fitness values ​​of the individuals and the whole of the population, find the global optimum and the individual optimum, update the particle swarm, obtain the offspring population, update the inertia weight, speed, etc. of the particle swarm algorithm, and update the number, location and rated capacity of the energy storage at the same time; Step S150, iterative optimization; Step S160, repeating steps S130 to S150 until the termination condition is met, and obtaining the final energy storage optimization configuration result; Step S170, output the result, output the energy storage number, location, rated capacity and energy storage optimal charging and discharging strategy data corresponding to the optimal population.

Images