Day-ahead optimization scheduling method of light storage and charging integrated system

Abstract

The invention discloses a day-ahead optimization scheduling method for an optical storage and charging integrated system, and the method comprises the steps: obtaining the historical storage and charging data of the optical storage and charging integrated system and the time sequence data of node voltage, and carrying out the power balance constraint of the data; based on the Thevenin equivalence theorem, carrying out equivalence on the power distribution network by taking a grid-connected point as a port, and enabling the power distribution network to be equivalent to constant voltage source series equivalent impedance which changes along with time; constructing a neural network time sequence prediction model containing historical storage and charging data and equivalent impedance, collecting historical data to construct a data set, and training the neural network time sequence prediction model based on the data set; acquiring historical environment data and historical storage and charging data, and inputting the historical environment data and the historical storage and charging data into the trained neural network time sequence prediction model to obtain predicted storage and charging data and equivalent impedance; and S5, based on the predicted storage and charging data and the equivalent impedance, constructing an optical storage and charging integrated day-ahead optimization scheduling model. The invention provides a distribution network state estimation method. A prediction method based on a neural network is considered to predict system power and distribution network equivalent parameters.

Description

Technical Field

[0001] The present invention relates to the field of energy optimization technology, and more specifically to a day-ahead optimization scheduling method for a photovoltaic storage and charging integrated system. Background Art

[0002] Vigorously developing new energy power generation and new energy vehicles is an inevitable trend for the country to promote the "dual carbon" plan and take the path of sustainable development. At present, distributed clean energy is developing rapidly, photovoltaic policies continue to be favorable, the electric vehicle market is developing rapidly, and people's demand for charging piles is also increasing. By the beginning of 2024, the number of new energy vehicles in my country has exceeded 20 million, and the industry has begun to take shape. The integrated photovoltaic storage and charging power station combines photovoltaic power generation, energy storage and charging stations together, and uses the battery energy storage system to absorb electricity during the evening trough to support the power demand of fast charging stations during the daytime peak period. At the same time, the supplement of the photovoltaic power generation system can also effectively smooth the load fluctuations of the fast charging station, improve the system operation efficiency, and reduce the occupation of power system resources during the peak load period, thereby optimizing the operation mode and improving the economic efficiency of operation.

[0003] A large number of scholars have studied the optimization model of photovoltaic storage and charging integration from different perspectives, such as how to maximize the absorption of photovoltaic power generation, uncertainty analysis of day-ahead dispatch optimization, and cost control when considering load shifting. At present, uncertainty analysis for photovoltaic power generation is unavoidable. Most of them use a two-layer optimization dispatch model to robustly optimize the uncertainty of photovoltaic system power. The upper model aims to minimize the operating cost, and the lower model aims to eliminate the uncertainty of photovoltaic system power and smooth the power output of energy storage system. However, the disadvantage is that the analysis of uncertainty remains at the robust optimization level, and the results are conservative, which cannot cope with severe situations well and has a certain impact on the economy of the system. The design and dispatch strategy of photovoltaic storage and charging integrated microgrid system established an adaptive robust three-layer optimization dispatch model of photovoltaic storage and charging system based on the two-layer model. In view of the difficulty of directly solving the three-layer optimization model, representative scenarios are used to describe uncertainty. At the same time, auxiliary variables are introduced to replace the worst scenario operating cost, and the inner layer decoupling replacement is realized, which is converted into a single-layer robust optimization model with a limited number of scenarios, solving the problem of difficulty in evaluating and handling when the microgrid encounters severe situations. For the handling of load in the substation, the city's intelligent photovoltaic storage charging tower adaptive robust day-ahead optimization scheduling divides the load on the energy consumption side into three categories: non-dispatchable, interruptible, and transferable, and conducts fine modeling. Among them, car charging is classified as transferable load. It is worth noting that the additional costs incurred during load transfer should be specially considered, such as giving users certain preferential incentives to encourage users to charge more during normal and valley hours to achieve the purpose of promoting load transfer.

[0004] The publication number is CN111900727A, and the name is a PSO-based collaborative optimization scheduling method and device for a photovoltaic, storage, charging and swapping integrated charging station. The method includes the following steps: obtaining the measurement point information of each photovoltaic power generation module, energy storage module and electric vehicle charging and swapping module in the photovoltaic, storage, charging and swapping integrated charging station; establishing a collaborative scheduling strategy model based on the acquired information, and the collaborative scheduling strategy model includes an objective function and constraints for minimizing the operating cost of the photovoltaic, storage, charging and swapping integrated charging station; solving the collaborative scheduling strategy model through a particle swarm algorithm; and formulating a grid power input and energy storage output plan for the photovoltaic, storage, charging and swapping integrated charging station based on the optimal solution obtained.

[0005] The publication number is CN117394318A, and the name is a method for integrated optimization and scheduling of photovoltaic storage and charging taking into account multiple operating modes, including: constructing a target optimization model, the target optimization model includes a first objective function, a second objective function and several constraints; the several constraints include market power purchase and sales constraints and power balance constraints; solving the optimal scheduling strategy and corresponding scheduling cost of the target optimization model under different operating modes of the microgrid; wherein, the first operating model is that electric vehicles participate in microgrid scheduling; the second operating model is that electric vehicles do not participate in microgrid scheduling; and selecting the optimal scheduling strategy with the smallest scheduling cost as the output scheduling strategy.

[0006] The disorderly charging of a large number of electric vehicles and the large-scale renewable energy power will have an adverse impact on the operation of the power system, so the optimization scheduling strategy of the integrated photovoltaic storage and charging system cannot ignore the impact of power fluctuations. However, the distribution network structure is complex and changeable, and it has its own randomness, uncertainty, and even unpredictable and incompletely measurable. Therefore, risk assessment based on distribution network flow is difficult to achieve in actual projects.

[0007] Therefore, how to provide a distribution network state estimation method and predict system power and distribution network equivalent parameters based on a neural network prediction method is an urgent problem that technicians in this field need to solve. Summary of the invention

[0008] In view of this, the present invention provides a day-ahead optimization scheduling method for an integrated photovoltaic storage and charging system, aiming to solve the technical problem that the distribution network structure is complex and changeable, and it has its own randomness, uncertainty, and even unpredictable and incompletely measurable nature. Therefore, risk assessment based on distribution network trends is difficult to achieve in actual engineering.

[0009] In order to achieve the above object, the present invention adopts the following technical solution:

[0010] A day-ahead optimization scheduling method for an integrated photovoltaic storage and charging system comprises the following steps:

[0011] S1. Obtain the historical storage and charging data of the integrated photovoltaic storage and charging system and the time series data of the node voltage, and impose power balance constraints on them;

[0012] S2. Based on the Thevenin equivalence theorem, the distribution network is equivalent to the constant voltage source that changes with time and is connected in series with an equivalent impedance.

[0013] S3, constructing a neural network time series prediction model including historical storage and charging data and equivalent impedance, collecting historical data to construct a data set, and training the neural network time series prediction model based on the data set;

[0014] S4, obtaining historical environmental data and historical storage and charging data, and inputting them into the trained neural network time series prediction model to obtain predicted storage and charging data and equivalent impedance;

[0015] S5. Construct a day-ahead optimization scheduling model for integrated photovoltaic storage and charging based on predicted storage and charging data and equivalent impedance.

[0016] Furthermore, the power balance constraint in S1 is specifically:

[0017] P Grid,k +P PV,k -P Load,k -P EV,k +P ESS,k =0

[0018] Where P Grid,k is the active power of the grid connection point at the time corresponding to the kth sampling point, and the direction injected from the grid into the photovoltaic storage and charging integrated system is the reference positive direction, P PV,k is the photovoltaic power at the kth sampling point at the corresponding moment, and the direction of photovoltaic power generation is taken as the reference positive direction, P Load,k is the load power at the kth sampling point, with the direction of load power consumption as the reference positive direction, P EV,k is the vehicle charging power at the kth sampling point, with the charging direction of the electric vehicle as the reference positive direction, P ESS,k is the charging and discharging power of the energy storage system at the time corresponding to the kth sampling point, with the discharging direction of the energy storage system as the reference positive direction.

[0019] Furthermore, the historical storage and charging data includes: historical photovoltaic power, historical load power, historical vehicle charging power and interactive power on the grid connection point side.

[0020] Furthermore, in S2, based on the Thevenin equivalence theorem, the distribution network is equivalent to the grid connection point as the port, and the distribution network is equivalent to a constant voltage source series equivalent impedance that changes with time. Specifically, based on the power flow equation between the equivalent power source and the grid connection point, an equivalent impedance equation group is established:

[0021]

[0022] Where k is an integer from 1 to n, V is the kth sampling point, and n is the total number of sampling points of historical data; Grid,k is the effective value of the voltage at the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, V Grid,k-1 is the effective value of the voltage at the k-1th sampling point of the photovoltaic storage and charging integrated system grid connection point, P Grid,k is the active power of the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, P Grid,k-1 is the active power of the k-1th sampling point of the photovoltaic storage and charging integrated system at the corresponding time, Q Grid,k is the reactive power of the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, Q Grid,k-1 is the reactive power of the k-1th sampling point of the photovoltaic storage and charging integrated system at the corresponding time, I ij,k is the effective value of the current flowing through the grid connection point at the corresponding moment of the kth sampling point; I ij,k-1 The effective value of the current flowing through the grid connection point at the corresponding moment of the k-1th sampling point; R k is the equivalent resistance of the kth sampling point of the distribution network at the corresponding time, X k is the equivalent reactance of the kth sampling point in the distribution network at the corresponding moment.

[0023] Furthermore, the S3 is specifically:

[0024] Construct a neural network time series prediction model including photovoltaic power, vehicle charging power, load power, and equivalent impedance;

[0025] Historical environmental data and historical storage and charging data are collected, and the historical storage and charging data and the distribution network equivalent impedance at the next sampling point are used as the output label of each sample of historical data. The data set is constructed into a data set, which is divided into a training set and a test set. The neural network time series prediction model is trained through the data set.

[0026] Furthermore, the S4 is specifically:

[0027] The historical environmental data specifically include: historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical diffuse radiation, historical total radiation, historical rainfall, and historical time;

[0028] The historical storage and charging data specifically include: historical photovoltaic power, historical load power, and historical vehicle charging power;

[0029] The historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical scattered radiation, historical total radiation, historical rainfall, historical time, historical photovoltaic power, historical load power, historical vehicle charging power, and historical equivalent impedance of the current day and the previous day are input into the neural network time series prediction model trained by S3 to predict the photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance in the next 24 hours.

[0030] Furthermore, the S5 is specifically:

[0031] Based on the predicted storage and charging data and equivalent impedance, a day-ahead optimization scheduling model for integrated photovoltaic storage and charging is constructed. The constraints of the model are:

[0032]

[0033] Where k is an integer from 1 to N, representing the predicted kth scheduling time, N represents the total number of scheduling time points in the next 24 hours, Δt is the time interval between scheduling time points, is the planned value of the energy storage system charging and discharging power at the kth scheduling moment, with the discharge direction of the energy storage system as the reference positive direction, P dis,k is the planned value of the energy storage system discharge power at the kth scheduling moment in the future, P ch,k is the planned value of the energy storage system charging power at the kth scheduling moment in the future. The discharge direction of the energy storage system is taken as the reference positive direction. SOC min is the minimum value of the charge of the energy storage system, SOC max is the maximum charge of the energy storage system, SOC k is the planned value of the energy storage system charge at the kth scheduling moment, δ ESS is the self-discharge rate of the energy storage system, η is the charging and discharging efficiency of the energy storage system, u dis,k 、u ch,k are the planned states of the energy storage system charging and discharging at the kth scheduling moment in the future. When the value is 1, they are the discharging and charging states of the energy storage system. ESS,max is the maximum value of the charging and discharging power of the energy storage system, is the planned value of the grid connection point power at the kth scheduling moment, the direction of the grid injected into the photovoltaic storage and charging integrated system is the reference positive direction, P′ PV,k is the predicted value of photovoltaic power at the kth scheduling moment, with the direction of photovoltaic power generation as the reference positive direction, P′ Load,k is the predicted value of load power at the kth scheduling moment, with the direction of load power consumption as the reference positive direction, P′ EV,k is the predicted value of the vehicle charging power at the kth scheduling moment, taking the charging direction of the electric vehicle as the reference positive direction, P Grid,min , P Grid,max They are the minimum and maximum values ​​of the grid-connected point power respectively.

[0034] Furthermore, the S5 further includes:

[0035] A voltage evaluation index model is constructed based on the predicted equivalent impedance, active power and reactive power plan values ​​of the grid connection point to constrain the source-load fluctuation in the integrated photovoltaic storage and charging system.

[0036]

[0037] In the formula, is the effective value of the grid connection point voltage at the kth dispatching moment in the future, R ′ k , X ′ k are the predicted values ​​of the equivalent resistance and reactance of the distribution network at the kth dispatching moment in the future, are the planned values ​​of active power and reactive power of the grid connection point at the kth dispatching moment in the future, is the effective value of the current flowing through the grid connection point at the kth dispatching moment, V Grid,min 、V Grid,max They are the minimum and maximum effective values ​​of the grid connection point voltage respectively.

[0038] Furthermore, it also includes:

[0039] The day-ahead optimization scheduling model of the integrated photovoltaic storage and charging system is constructed with the goal of minimizing the total operating cost of the integrated photovoltaic storage and charging system in the next 24 hours:

[0040]

[0041] In the formula, C ESS,k is the operating cost of the energy storage system at the kth dispatching moment, C Grid,k is the cost of the distribution network at the kth scheduling time, K buy , K sell is the buying and selling price of distribution network, m k is the peak-valley electricity price coefficient at the kth dispatching time, C total The total cost of operating the integrated photovoltaic, storage and charging system for the next 24 hours.

[0042] Furthermore, the linear programming algorithm is used to solve the day-ahead optimization scheduling model of the integrated photovoltaic storage and charging system. k is taken from 1 to N, and the grid connection point interaction power from 1 to N scheduling points in the next 24 hours is obtained.

[0043] The above technical solution includes at least the following technical effects:

[0044] Through the above technical solutions, it can be known that the present invention discloses a method for optimizing the day-ahead scheduling of an integrated photovoltaic storage and charging system. Compared with the prior art, the present invention uses the grid connection point as the port to perform an equivalent operation on the distribution network based on the Thevenin equivalence theorem, and the distribution network is equivalent to a simplified circuit model of a constant voltage source in series with an equivalent impedance that changes with time. According to the power flow equation between the equivalent power source and the grid connection point, a group of equations for solving the equivalent impedance of the distribution network can be established. The time series data of historical photovoltaic power data, historical load power, historical vehicle charging power, historical grid connection point active power, and historical grid connection point voltage are collected, and the historical voltage effective value, active power, and reactive power are substituted into the group of equations for solving the equivalent impedance of the distribution network to obtain the historical distribution network equivalent impedance parameters. A recurrent neural network time series prediction model with five dimensions, including photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance, is constructed. By training historical data, the historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical scattered radiation, historical total radiation, historical rainfall, historical time, historical photovoltaic power, historical load power, historical vehicle charging power, and historical equivalent impedance of the current day and the previous period are used to input the trained neural network time series prediction model to obtain the photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance data for the next 24 hours. Then, by solving the day-ahead optimization scheduling model of the integrated photovoltaic storage and charging system, the voltage distribution of the system grid connection point in the next 24 hours and the active power of the system grid connection point in the next 24 hours are obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0046] Figure 1 The figure is a flow chart of an optimization scheduling method of the present invention.

[0047] Figure 2 The figure is a schematic diagram of the structure of a distribution network equivalent circuit of the present invention.

[0048] Figure 3 The present invention is a schematic diagram of the waveform structure of the active power historical time series data of the grid connection point of the integrated photovoltaic storage and charging system.

[0049] Figure 4 The present invention is a schematic diagram of the waveform structure of the voltage history time series data of the grid connection point of the integrated photovoltaic storage and charging system.

[0050] Figure 5 This is a schematic diagram of the day-ahead optimization scheduling results of a photovoltaic storage and charging integrated system of the present invention. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] See attached Figure 1-5 , obtain the historical storage and charging data of the integrated photovoltaic storage and charging system and the time series data of the node voltage, and obtain the power balance constraint;

[0053] P Grid，k +P PV，k -P Load，k -P EV，k +P ESS，k =0

[0054] Where P Grid,k is the active power of the grid connection point at the time corresponding to the kth sampling point, and the direction injected from the grid into the photovoltaic storage and charging integrated system is the reference positive direction, P PV,k is the photovoltaic power at the kth sampling point at the corresponding moment, and the direction of photovoltaic power generation is taken as the reference positive direction, P Load,k is the load power at the kth sampling point, with the direction of load power consumption as the reference positive direction, P EV,k is the vehicle charging power at the kth sampling point, with the charging direction of the electric vehicle as the reference positive direction, P ESS,k is the charging and discharging power of the energy storage system at the time corresponding to the kth sampling point, with the discharging direction of the energy storage system as the reference positive direction.

[0055] In this example, the photovoltaic power comes from data collected every 15 minutes in a certain area of ​​Australia for 7 days, the load power comes from data collected every 15 minutes in Nanjing for 7 days, and the car charging power comes from data collected every 15 minutes in a certain park for 7 days. The active power of the grid connection point and the time series data of the grid connection point voltage are also collected from a photovoltaic storage charging system every 15 minutes for 7 days. The waveform of the time series data is shown in the figure below. Figure 3 , 4 shown.

[0056] According to Thevenin equivalence theorem, the distribution network can be equivalent to a simplified circuit model of a constant voltage source that varies with time and is connected in series with an equivalent impedance. The equivalent circuit is as follows: Figure 2 shown.

[0057] According to the power flow equation between the equivalent power source and the grid connection point, the equation group can be established:

[0058]

[0059] Where k is an integer from 1 to n, is the sampling point, and n is the total number of sampling points of historical data; V Grid,k is the effective value of the voltage at the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, V Grid,k-1 is the effective value of the voltage at the k-1th sampling point of the photovoltaic storage and charging integrated system grid connection point, P Grid,k is the active power of the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, P Grid,k-1 is the active power of the k-1th sampling point of the photovoltaic storage and charging integrated system at the corresponding time, Q Grid,k is the reactive power of the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, Q Grid,k-1 is the reactive power of the k-1th sampling point of the photovoltaic storage and charging integrated system at the corresponding time, I ij , k is the effective value of the current flowing through the grid connection point at the corresponding moment of the kth sampling point; I ij,k-1 The effective value of the current flowing through the grid connection point at the corresponding moment of the k-1th sampling point; R k is the equivalent resistance of the kth sampling point of the distribution network at the corresponding time, X k is the equivalent reactance of the kth sampling point in the distribution network at the corresponding moment.

[0060] V Grid，k 、V Grid，k-1 , P Grid，k , P Grid，k-1 , Q Grid，k , Q Grid，k-1 ,I ij，k ,I ij，k-1 is the known quantity in the equations, R k , X k is the unknown quantity in the equations. k , X k They are all unknown variables, so when solving, at least two adjacent sampling points (k-1, k) are used to construct the distribution network equivalent impedance solution equations. Since the constructed equations are nonlinear equations, particle swarm optimization or heuristic algorithms can be used to solve them, and the distribution network equivalent resistance reactance R corresponding to the kth moment can be obtained. k , X k .

[0061] k takes values ​​from 1 to 672 respectively. Substituting the historical voltage effective value, active power and reactive power at the kth sampling point into the solution equations of the distribution network equivalent impedance, the distribution network equivalent resistance R at the sampling points 1 to n can be obtained. k and reactance X k (k=1,2…672).

[0062] Construct a recurrent neural network time series prediction model such as a long short-term memory network (LSTM) or a gated recurrent neural network (GRU) with five dimensions of photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance. In the embodiment, a gated recurrent neural network GRU is used; collect data including historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical scattered radiation, historical total radiation, historical rainfall, historical time, etc., and construct samples of historical data sets in combination with historical photovoltaic power, historical load power, historical vehicle charging power, and historical equivalent impedance; use the photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance at the corresponding moment of the next sampling point as the output label of each sample of historical data; thereby construct a data set and divide it into a training set and a test set, which are used for training and testing of the neural network time series prediction model respectively. A neural network time series prediction model with five dimensions including photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance can be obtained.

[0063] The historical environmental data include: historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical diffuse radiation, historical total radiation, historical rainfall, and historical time;

[0064] The historical storage and charging data include: historical photovoltaic power, historical load power, and historical vehicle charging power;

[0065] The trained neural network time series prediction model is input with the historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical scattered radiation, historical total radiation, historical rainfall, historical time, historical photovoltaic power, historical load power, historical vehicle charging power, and historical equivalent impedance of the current day and the previous period. This embodiment uses the historical data of the past 7 days to predict the photovoltaic power, vehicle charging power, load power, and distribution network equivalent resistance and reactance in the next 24 hours. The historical data and predicted data are shown in Table 1, Table 2, and Table 3 respectively:

[0066] Table 1 Historical meteorological data for the past 7 days

[0067]

[0068] Table 2 Historical power and equivalent data for the past 7 days

[0069]

[0070]

[0071]

[0072] Table 3 Five-dimensional data of 24 hours a day output by the prediction model

[0073] time Photovoltaic power (MW) Load power(MW) Car charging power (MW) Equivalent resistance (MΩ) Equivalent reactance(MΩ) 00:00 0 0.038 0.170 0.474 0.211 01:00 0 0.038 0.167 0.478 0.214 02:00 0 0.039 0.092 0.476 0.215 03:00 0 0.039 0.256 0.476 0.217 04:00 0 0.041 0.299 0.477 0.213 05:00 0 0.044 0.532 0.173 0.914 06:00 0.146 0.049 0.639 0.0007 0.999 07:00 0.301 0.052 0.847 0.0002 0.989 08:00 0.418 0.054 1.272 0.001 0.838 09:00 0.487 0.055 0.551 0.001 0.640 10:00 0.519 0.049 1.120 0.002 0.563 11:00 0.520 0.051 1.624 0.003 0.659 12:00 0.480 0.052 0.758 0.001 0.685 13:00 0.404 0.052 1.711 0.003 0.665 14:00 0.289 0.053 1.085 0.001 0.863 15:00 0.138 0.057 1.192 0.001 0.983 16:00 0.0001 0.058 1.319 0.001 0.998 17:00 0 0.056 1.191 0.221 0.986 18:00 0 0.055 0.779 0.513 0.250 19:00 0 0.053 0.570 0.512 0.228 20:00 0 0.049 0.058 0.496 0.224 21:00 0 0.045 0.224 0.498 0.221 22:00 0 0.044 0.054 0.482 0.217 23:00 0 0.041 0.140 0.479 0.213

[0074] The photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance at each time interval predicted in the next 24 hours are used as input to establish a day-ahead optimization scheduling model for photovoltaic storage and charging integration. The constraints of the model are:

[0075]

[0076] In the formula, k is an integer ranging from 1 to N, representing the predicted scheduling time, N represents the total number of scheduling time points in the next 24 hours, which is 24 in this example, and Δt is the time interval between scheduling time points, which is 1 hour in this example. Pdis is the planned value of the energy storage system charging and discharging power at the kth scheduling moment, with the discharge direction of the energy storage system as the reference positive direction, ,k is the planned value of the energy storage system discharge power at the kth scheduling moment in the future, Pch ,k is the planned value of the energy storage system charging power at the kth scheduling moment in the future, with the direction of energy storage system discharge as the reference positive direction, SOCmin is the minimum value of the energy storage system charge, SOCmax is the maximum value of the energy storage system charge, in this case they are 0.1 and 1 respectively, SOCk is the planned value of the energy storage system charge at the kth scheduling moment, δESS is the energy storage system self-discharge rate, in this case it is 0.02 per month, η is the energy storage system charge and discharge efficiency, in this case it is 0.9, u dis,k 、u ch,k are the planned states of the energy storage system charging and discharging at the kth scheduling moment in the future. When the value is 1, they are the discharging and charging states of the energy storage system. ESS,max is the maximum value of the charging and discharging power of the energy storage system, is the planned value of the grid connection point power at the kth scheduling moment, the direction of the grid injected into the photovoltaic storage and charging integrated system is the reference positive direction, P′ PV,k is the predicted value of photovoltaic power at the kth scheduling moment, with the direction of photovoltaic power generation as the reference positive direction, P′ Load,k is the predicted value of load power at the kth scheduling moment, with the direction of load power consumption as the reference positive direction, P′ EV,k is the predicted value of the vehicle charging power at the kth scheduling moment, taking the charging direction of the electric vehicle as the reference positive direction, P Grid,min , P Grid,max They are the minimum and maximum values ​​of the grid-connected point power, respectively -1.5MW and 1.5MW in this example.

[0077] In order to take into account the impact of source-load changes on the grid-connected point voltage in the next 24-hour dispatch plan, a voltage evaluation index model is constructed using the predicted equivalent resistance reactance, active power, and reactive power plan values ​​of the grid-connected point, thereby establishing voltage safety constraints to constrain the impact of source-load fluctuations in the integrated photovoltaic storage and charging system:

[0078]

[0079] In the formula, is the effective value of the grid connection point voltage at the kth dispatching moment in the future, R ′ k , X ′ k are the predicted values ​​of the equivalent resistance and reactance of the distribution network at the kth dispatching moment in the future, are the planned values ​​of active power and reactive power of the grid connection point at the kth dispatching moment in the future, is the effective value of the current flowing through the grid connection point at the kth dispatching moment, V Grid,min 、V Grid,max The minimum and maximum effective values ​​of the grid connection point voltage are 0.95 MV and 1.05 MV in this example. The effective values ​​of the grid connection point voltage in the next 24 hours are shown in Table 4:

[0080] Table 4 Effective value of grid connection point voltage in the next 24 hours

[0081]

[0082]

[0083] The day-ahead optimization scheduling model of the integrated photovoltaic storage and charging system is constructed with the goal of minimizing the total operating cost of the integrated photovoltaic storage and charging system in the next 24 hours:

[0084]

[0085] In the formula, C ESS,k is the operating cost of the energy storage system at the kth dispatching moment, C Grid,k is the cost of the distribution network at the kth scheduling time, K buy , K sell is the electricity price of the distribution network, which is 1.648 yuan and 0.824 yuan per kWh respectively in this example, m k is the peak-valley electricity price coefficient at the kth dispatching time. 10:00-13:00 and 17:00-22:00 are the peak electricity consumption periods. k The power consumption is 1.71, the previous day 23:00 to the next day 7:00 is the low power consumption period, which is 0.36, and the rest of the period is the level power consumption period, m k is 1, C total The total cost of operating the integrated photovoltaic, storage and charging system for the next 24 hours.

[0086] The operating cost of the future 24-hour photovoltaic storage and charging integrated system is shown in Table 5:

[0087] Table 5 Operating costs of the future 24-hour photovoltaic storage and charging integrated system

[0088] Energy storage cost (yuan) Grid connection cost (yuan) Total cost (yuan) 0.122 23.964 24.086

[0089] Since the constructed photovoltaic storage and charging integrated day-ahead optimization scheduling model is a linear programming optimization model, the linear programming algorithm can be used to solve the model. k is taken from 1 to 24, and the grid connection point interaction power from 1 to 24 scheduling points in the next 24 hours can be obtained. The results are shown in the table. The active power of the grid connection point in the next 24 hours is shown in Table 6. The optimized dispatch results of the integrated photovoltaic storage and charging system are shown in Table 6. Figure 4 shown.

[0090] Table 6 Active power of the grid connection point in the next 24 hours

[0091]

[0092]

[0093] According to the day-ahead optimization scheduling method of a photovoltaic storage and charging integrated system designed by the present invention, the distribution network is equivalent to a simplified circuit model of a constant voltage source in series with equivalent impedance that changes with time, and a group of equations for solving the equivalent impedance of the distribution network can be established according to the power flow equation between the equivalent power source and the grid connection point. The historical photovoltaic power data, historical load power, historical vehicle charging power, historical grid connection point active power, and historical grid connection point voltage time series data are collected, and the historical voltage effective value, active power, and reactive power are substituted into the group of equations for solving the distribution network equivalent impedance to obtain the historical distribution network equivalent impedance parameters. A recurrent neural network time series prediction model with five dimensions, including photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance, is constructed. By training historical data, the historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical scattered radiation, historical total radiation, historical rainfall, historical time, historical photovoltaic power, historical load power, historical vehicle charging power, and historical equivalent impedance of the current day and the previous period are used to input the trained neural network time series prediction model to obtain the photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance data for the next 24 hours. Then, by solving the day-ahead optimization scheduling model of the integrated photovoltaic storage and charging system, the voltage distribution of the system grid connection point in the next 24 hours and the active power of the system grid connection point in the next 24 hours are obtained.

[0094] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0095] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A day-ahead optimization scheduling method for an integrated photovoltaic storage and charging system, characterized in that: The following steps are involved: S1. Obtain the historical storage and charging data of the integrated photovoltaic storage and charging system and the time series data of the node voltage, and impose power balance constraints on them; S2. Based on the Thevenin equivalence theorem, the distribution network is equivalent to the constant voltage source that changes with time and is connected in series with an equivalent impedance. S3, constructing a neural network time series prediction model including historical storage and charging data and equivalent impedance, collecting historical data to construct a data set, and training the neural network time series prediction model based on the data set; S4, obtaining historical environmental data and historical storage and charging data, and inputting them into the trained neural network time series prediction model to obtain predicted storage and charging data and equivalent impedance; S5. Construct a day-ahead optimization scheduling model for integrated photovoltaic storage and charging based on predicted storage and charging data and equivalent impedance.

2. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 1, characterized in that: The power balance constraint in S1 is specifically: P Grid，k +P PV，k -P Load，k -P EV，k +P ESS，k =0 Where P Grid，k is the active power of the grid connection point at the time corresponding to the kth sampling point, and the direction injected from the grid into the photovoltaic storage and charging integrated system is the reference positive direction, P PV，k is the photovoltaic power at the kth sampling point at the corresponding moment, and the direction of photovoltaic power generation is taken as the reference positive direction, P Load，k is the load power at the kth sampling point, with the direction of load power consumption as the reference positive direction, P EV，k is the vehicle charging power at the kth sampling point, with the charging direction of the electric vehicle as the reference positive direction, P ESS，k is the charging and discharging power of the energy storage system at the time corresponding to the kth sampling point, with the discharging direction of the energy storage system as the reference positive direction.

3. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 1, characterized in that: The historical storage and charging data include: historical photovoltaic power, historical load power, historical vehicle charging power and interactive power on the grid connection point side.

4. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 1, characterized in that: In S2, based on the Thevenin equivalence theorem, the distribution network is equivalent to the grid connection point as the port, and the distribution network is equivalent to a constant voltage source that changes with time and is connected in series with an equivalent impedance. Specifically, based on the power flow equation between the equivalent power source and the grid connection point, an equivalent impedance equation group is established: Where k is an integer from 1 to n, V is the kth sampling point, and n is the total number of sampling points of historical data; Grid，k is the effective value of the voltage at the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, V Grid，k-1 is the effective value of the voltage at the k-1th sampling point of the photovoltaic storage and charging integrated system grid connection point, P Grid，k is the active power of the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, P Grid，k-1 is the active power of the k-1th sampling point of the photovoltaic storage and charging integrated system at the corresponding time, Q Grid，k is the reactive power of the kth sampling point of the photovoltaic storage and charging integrated system at the corresponding time, Q Grid，k-1 is the reactive power of the k-1th sampling point of the photovoltaic storage and charging integrated system at the corresponding time, I ij，k is the effective value of the current flowing through the grid connection point at the corresponding moment of the kth sampling point; I ij，k-1 The effective value of the current flowing through the grid connection point at the corresponding moment of the k-1th sampling point; R k is the equivalent resistance of the kth sampling point of the distribution network at the corresponding time, X k is the equivalent reactance of the kth sampling point in the distribution network at the corresponding moment.

5. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 1, characterized in that: The S3 is specifically: Construct a neural network time series prediction model including photovoltaic power, vehicle charging power, load power, and equivalent impedance; Historical environmental data and historical storage and charging data are collected, and the historical storage and charging data and the distribution network equivalent impedance at the next sampling point are used as the output label of each sample of historical data. The data set is constructed into a data set, which is divided into a training set and a test set. The neural network time series prediction model is trained through the data set.

6. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 1, characterized in that: The S4 is specifically: The historical environmental data specifically include: historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical diffuse radiation, historical total radiation, historical rainfall, and historical time; The historical storage and charging data specifically include: historical photovoltaic power, historical load power, and historical vehicle charging power; The historical wind speed, historical temperature, historical relative humidity, historical direct radiation, historical scattered radiation, historical total radiation, historical rainfall, historical time, historical photovoltaic power, historical load power, historical vehicle charging power, and historical equivalent impedance of the current day and the previous day are input into the neural network time series prediction model trained by S3 to predict the photovoltaic power, vehicle charging power, load power, distribution network equivalent resistance and reactance in the next 24 hours.

7. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 1, characterized in that: The S5 is specifically: Based on the predicted storage and charging data and equivalent impedance, a day-ahead optimization scheduling model for integrated photovoltaic storage and charging is constructed. The constraints of the model are: Where k is an integer from 1 to N, representing the predicted kth scheduling time, N represents the total number of scheduling time points in the next 24 hours, Δt is the time interval between scheduling time points, is the planned value of the energy storage system charging and discharging power at the kth scheduling moment, with the discharge direction of the energy storage system as the reference positive direction, P dis，k is the planned value of the energy storage system discharge power at the kth scheduling moment in the future, P ch，k is the planned value of the energy storage system charging power at the kth scheduling moment in the future. The discharge direction of the energy storage system is taken as the reference positive direction. SOC min is the minimum value of the charge of the energy storage system, SOC max is the maximum charge of the energy storage system, SOC k is the planned value of the energy storage system charge at the kth scheduling moment, δ ESS is the self-discharge rate of the energy storage system, η is the charging and discharging efficiency of the energy storage system, u dis，k 、u ch，k are the planned states of the energy storage system charging and discharging at the kth scheduling moment in the future. When the value is 1, they are the discharging and charging states of the energy storage system. ESS，max is the maximum value of the charging and discharging power of the energy storage system, is the planned value of the grid connection point power at the kth scheduling moment, the direction of the grid injected into the photovoltaic storage and charging integrated system is the reference positive direction, P′ PV，k is the predicted value of photovoltaic power at the kth scheduling moment, with the direction of photovoltaic power generation as the reference positive direction, P′ Load，k is the predicted value of load power at the kth scheduling moment, with the direction of load power consumption as the reference positive direction, P′ EV，k is the predicted value of the vehicle charging power at the kth scheduling moment, taking the charging direction of the electric vehicle as the reference positive direction, P Grid，min , P Grid，max They are the minimum and maximum values ​​of the grid-connected point power respectively.

8. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 7, characterized in that: The S5 further includes: A voltage evaluation index model is constructed based on the predicted equivalent impedance, active power and reactive power plan values ​​of the grid connection point to constrain the source-load fluctuation in the integrated photovoltaic storage and charging system. In the formula, is the effective value of the grid connection point voltage at the kth dispatching moment in the future, R′ k , X′ k are the predicted values ​​of the equivalent resistance and reactance of the distribution network at the kth dispatching moment in the future, are the planned values ​​of active power and reactive power of the grid connection point at the kth dispatching moment in the future, is the effective value of the current flowing through the grid connection point at the kth dispatching moment, V Grid，min 、V Grid，max They are the minimum and maximum effective values ​​of the grid connection point voltage respectively.

9. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 8, characterized in that: Also includes: The day-ahead optimization scheduling model of the integrated photovoltaic storage and charging system is constructed with the goal of minimizing the total operating cost of the integrated photovoltaic storage and charging system in the next 24 hours: In the formula, C ESS，k is the operating cost of the energy storage system at the kth dispatching moment, C Grid，k is the cost of the distribution network at the kth scheduling time, K buy , K sell is the distribution network buying and selling price, m k is the peak-valley electricity price coefficient at the kth dispatching time, C total The total cost of operating the integrated photovoltaic, storage and charging system for the next 24 hours.

10. The method for optimizing the day-ahead scheduling of a photovoltaic storage and charging integrated system according to claim 9, characterized in that: Based on the linear programming algorithm, the optimal scheduling model of the integrated photovoltaic storage and charging system is solved. k is taken from 1 to N respectively, and the grid connection point interaction power from 1 to N scheduling points in the next 24 hours is obtained.

Images