A multi-time scale loss reduction optimization method and system for distribution network

Abstract

The present invention discloses a multi-time scale loss reduction optimization method and system for a distribution network, which relates to the technical field of distribution network planning, and comprises the following steps: receiving electric vehicle driving related data, and calculating the charging power of the electric vehicle by using a Monte Carlo simulation method according to the electric vehicle driving related data; inputting the charging power of the electric vehicle into a pre-established dynamic time-of-use electricity price model, and outputting a dynamic time-of-use electricity price, and inputting the dynamic time-of-use electricity price into a pre-established multi-time scale optimization scheduling strategy model based on MPC before and during the day, taking network loss and operation cost as objective functions, and outputting an optimized scheduling result.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network planning, and in particular to a multi-time scale loss reduction optimization method and system for a distribution network. Background Art

[0002] With the explosive growth of new energy vehicles dominated by electric vehicles (EVs) in recent years, the large-scale charging of electric vehicles brought about by the high number of electric vehicles will cause the total load of the power grid to "peak on peak", load fluctuation, harmonic pollution, three-phase imbalance, voltage distortion and other problems. The resulting electric vehicle grid access technology (V2G) overcomes the limitations of "two-way communication and one-way energy transmission" of electric vehicles, making electric vehicles have the dual attributes of source and load. Through the interaction between the user side and the grid side, the orderly charging and discharging of electric vehicles is controlled, providing objective flexibility resources for the power system, achieving the comprehensive advantages of stabilizing the grid load, improving the consumption of renewable energy, and reducing network losses. Although there have been many achievements in the coordinated interactive optimization of electric vehicles and new energy, most of them have not considered the impact of model prediction of wind and solar output on different time scales, and have not solved the problem of reducing the loss of distribution networks due to prediction errors in the implementation of regulation. Summary of the invention

[0003] The technical problem to be solved by the present invention is how to solve the problem of reducing the loss of the distribution network caused by the prediction error in the implementation of regulation.

[0004] In the first aspect, the technical problem to be solved by the present invention is achieved by the following technical solution: a multi-time scale loss reduction optimization method for a distribution network, the method comprising the following steps:

[0005] Receiving driving data related to the electric vehicle, and calculating the charging and discharging power of the electric vehicle using a Monte Carlo simulation method according to the driving data related to the electric vehicle;

[0006] The charging and discharging power of electric vehicles is input into a pre-established dynamic time-of-use electricity price model, and the dynamic time-of-use electricity price is output. The dynamic time-of-use electricity price is input into a pre-established MPC-based day-ahead and intraday multi-time scale optimization scheduling strategy model, with network loss and operating cost as the objective function, and the optimized scheduling result is output.

[0007] In combination with the first aspect, in certain implementations of the first aspect, the method further includes: the electric vehicle driving related data includes: vehicle classification considering travel habits, battery characteristics, daily mileage l , Start charging time t 0 and initial state of charge SOC 0.

[0008] In combination with the first aspect, in certain implementations of the first aspect, the method further includes: l The probability function is:

[0009]

[0010] Where: is the mean mileage, is the standard deviation of mileage;

[0011] Probability distribution of electric vehicle starting charging time:

[0012] Electric vehicle charging start time t 0 follows a normal distribution, that is t 0~N( μt , s 2 t ), the probability density function is:

[0013]

[0014] Where: is the average of the initial charging time, is the standard deviation of the initial charging time;

[0015] Electric vehicle charging duration

[0016] The charging and discharging time of electric vehicles is related to the final battery state of charge. The initial state of charge of the wth vehicle is It is expressed as:

[0017]

[0018] Where: The power consumption per 100 km; is the battery capacity;

[0019] Charging time It is expressed as:

[0020]

[0021] Where: The actual charging time required for electric vehicles; is the expected SOC value of EV users; EV charging and discharging efficiency; is the charging power of a single electric vehicle w;

[0022] The total electric vehicle charging power is obtained by adding up the charging load of a single electric vehicle:

[0023]

[0024] Where: is the total number of electric vehicles, ( t ) is an electric car t The total charging power during the time period.

[0025] In combination with the first aspect, in certain implementations of the first aspect, the method further includes: the process of inputting the charging power of the electric vehicle into a pre-established dynamic time-of-use electricity price model and outputting a dynamic time-of-use electricity price is as follows:

[0026] When the equivalent load is higher than the mean s The peak period is when the value is times the sum of the peak-to-valley difference, which is lower than the mean. s The time when the difference between the peak and valley is double is the valley period, and the rest of the time is the normal period. The time-of-use electricity price model is expressed as:

[0027]

[0028]

[0029]

[0030]

[0031] Where: C(t) is the dynamic time-of-use electricity price, is the electricity price of equivalent load during normal hours, d The fluctuation range of electricity prices during peak and off-peak hours. is the equivalent load, is the basic load, is the mean value of equivalent load, is the peak-to-valley difference of equivalent load, , are the peak and valley values ​​of the equivalent load respectively.

[0032] In combination with the first aspect, in some implementations of the first aspect, the method further includes: a calculation process in which the day-ahead optimization model in the pre-established MPC-based day-ahead intra-day multi-time scale optimization scheduling strategy model takes the root node power purchase cost, the electric vehicle user charging and discharging cost, and the network loss as the objective function:

[0033] Objective function:

[0034] The root node has the lowest electricity purchase cost:

[0035]

[0036] Where: F 1 is the objective function of minimizing the power purchase cost of the root node, P buyInput active power for the root node, T is the total scheduling period;

[0037] The charging and discharging costs for electric vehicle users are minimal:

[0038]

[0039] Where: The objective function for minimizing the charging and discharging cost for electric vehicle users is: is the charging electricity price of electric vehicles at time t; To provide electric vehicle users with discharge incentive subsidies; and are the charging and discharging power of electric vehicles respectively;

[0040] Minimum network loss:

[0041]

[0042] Where: is the objective function of minimizing network loss, For branch ij Resistance; for t Time Branch ij The current;

[0043] The objective function is normalized by linear weighting as follows:

[0044]

[0045] Where: , and They are , and The weight coefficient of

[0046] The constraints include:

[0047] Single vehicle charging and discharging power constraints, electric vehicle charge state constraints, energy storage equipment power constraints, energy storage equipment charge constraints, photovoltaic constraints, transferable load constraints and distribution network related constraints;

[0048] Controllable unit u The prediction model expression is as follows:

[0049]

[0050] Puu ( k ) is a controllable unit u exist kThe initial output value of the time period; U is the total number of controllable units; ( k+i|k )for k Time period prediction future k +( i -1), k+i ]Controllable units within the time period u The output increment; pu ( k+i|k ) is the k-period prediction for the future k+i Output value of controllable unit u in time period; , is the climbing power value that the output of each controllable unit needs to meet in adjacent prediction time periods; N is the total number of time periods in the forecast time domain.

[0051] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the constraints of the intraday optimization model in the pre-established MPC-based day-ahead intraday multi-time scale optimization scheduling strategy model include AC power flow equation constraints, operation safety constraints, and active and reactive control equipment behavior constraints. There is a deviation in the output control sequence obtained by the actual and predictive control, so the feedback correction is as follows:

[0052]

[0053] In the formula, t=t0+ , t0 For the current moment, For the forecast sequence in the new round of rolling optimization scheduling, is the latest predicted sequence.

[0054] In a second aspect, in order to achieve the above-mentioned object, the present invention discloses a distribution network multi-time scale loss reduction optimization system, comprising:

[0055] A power calculation module is used to receive the driving data of the electric vehicle and calculate the charging power of the electric vehicle using the Monte Carlo simulation method according to the driving data of the electric vehicle;

[0056] The optimization scheduling module is used to input the charging power of electric vehicles into a pre-established dynamic time-of-use electricity price model, output the dynamic time-of-use electricity price, input the dynamic time-of-use electricity price into a pre-established MPC-based day-ahead and intraday multi-time scale optimization scheduling strategy model, take network loss and operating cost as the objective function, and output the optimized scheduling result.

[0057] In conjunction with the second aspect, in certain implementations of the second aspect, the vehicle classification, battery characteristics, and daily mileage are considered. l , Start charging time t0 and initial state of charge SOC 0;

[0058] Daily driving range of electric vehicles in power calculation module l The probability function is:

[0059]

[0060] Where: is the mean mileage, is the standard deviation of mileage;

[0061] Probability distribution of electric vehicle starting charging time:

[0062] Electric vehicle charging start time t 0 follows a normal distribution, that is t 0~N( μt , s 2 t ), the probability density function is:

[0063]

[0064] Where: is the average of the initial charging time, is the standard deviation of the initial charging time;

[0065] Electric vehicle charging duration

[0066] The charging and discharging time of electric vehicles is related to the final battery state of charge. The initial state of charge of the wth vehicle is It is expressed as:

[0067]

[0068] Where: The power consumption per 100 km; is the battery capacity;

[0069] Charging time t c，w It is expressed as:

[0070]

[0071] Where: The actual charging time required for electric vehicles; is the expected SOC value of EV users; EV charging and discharging efficiency; is the charging power of a single electric vehicle w;

[0072] The total electric vehicle charging power is obtained by adding up the charging load of a single electric vehicle:

[0073]

[0074] Where: is the total number of electric vehicles, ( t ) is an electric car t The total charging power during the time period.

[0075] The optimization scheduling module inputs the charging and discharging power of electric vehicles into the pre-established dynamic time-of-use electricity price model, and the process of outputting the dynamic time-of-use electricity price is as follows:

[0076] When the equivalent load is higher than the mean s The peak period is when the value is times the sum of the peak-to-valley difference, which is lower than the mean. s The time when the difference between the peak and valley is double is the valley period, and the rest of the time is the normal period. The time-of-use electricity price model is expressed as

[0077]

[0078]

[0079]

[0080]

[0081] Where: C(t) is the dynamic time-of-use electricity price, is the electricity price of equivalent load during normal hours, d The fluctuation range of electricity prices during peak and off-peak hours. is the equivalent load, is the basic load, is the mean value of equivalent load, is the peak-to-valley difference of equivalent load, , are the peak and valley values ​​of equivalent load respectively;

[0082] The calculation process of the day-ahead optimization model in the MPC-based day-ahead and intraday multi-time scale optimization scheduling strategy model pre-established in the optimization scheduling module takes the root node power purchase cost, electric vehicle user charging and discharging cost and network loss as the objective function:

[0083] Objective function:

[0084] The root node has the lowest electricity purchase cost:

[0085]

[0086] Where: F 1 is the objective function of minimizing the power purchase cost of the root node, P buy Input active power for the root node,T is the total dispatching cycle; the charging and discharging cost of electric vehicle users is the minimum:

[0087]

[0088] Where: The objective function for minimizing the charging and discharging cost for electric vehicle users is: is the charging electricity price of electric vehicles at time t; To provide electric vehicle users with discharge incentive subsidies; and are the charging and discharging power of electric vehicles respectively;

[0089] Minimum network loss:

[0090]

[0091] Where: is the objective function of minimizing network loss, For branch ij Resistance; for t Time Branch ij The current;

[0092] The objective function is normalized by linear weighting as follows:

[0093]

[0094] Where: , and They are , and The weight coefficient of

[0095] The constraints include:

[0096] Single vehicle charging and discharging power constraints, electric vehicle charge state constraints, energy storage equipment power constraints, energy storage equipment charge constraints, photovoltaic constraints, transferable load constraints and distribution network related constraints;

[0097] Controllable unit u The prediction model expression is as follows:

[0098]

[0099] P uo ( k ) is a controllable unit u exist k The initial output value of the time period; U is the total number of controllable units; ( k+i|k )for k Time period prediction future k +( i -1), k+i ]Controllable units within the time period u The output increment; p u ( k+i|k ) is the k-period prediction for the future k+i Output value of controllable unit u in time period; , is the climbing power value that the output of each controllable unit needs to meet in adjacent prediction time periods; N is the total number of time periods in the forecast time domain.

[0100] The constraints of the intraday optimization model in the MPC-based day-ahead and intraday multi-time scale optimization scheduling strategy model pre-established in the optimization scheduling module include AC power flow equation constraints, operation safety constraints, and active and reactive control equipment behavior constraints. There is a deviation in the output control sequence obtained by actual and predictive control, so the feedback correction is as follows:

[0101]

[0102] In the formula, t=t0+ , t0 For the current moment, For the forecast sequence in the new round of rolling optimization scheduling, is the latest predicted sequence.

[0103] Beneficial effects of the present invention:

[0104] The present invention proposes a multi-time scale loss reduction strategy for distribution networks that takes electric vehicles into consideration in order to address the uncertainty of wind and solar resource output, the disorderly charging of a large number of electric vehicles connected to the distribution network, which leads to the superposition of distribution network load peaks, increased network losses, and the deviation of scheduling results caused by single-time scale optimization scheduling relative to real-time changing DR resources. By setting different scenario algorithms in the embodiments for comparison, it is found that the proposed strategy has obvious effects in reducing losses and smoothing load fluctuations. While electric vehicles participate in grid interaction to reduce the impact of wind and solar grid connection, their flexibility compensates for the deviation in wind and solar output forecasts and maintains system balance. BRIEF DESCRIPTION OF THE DRAWINGS

[0105] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0106] Figure 1It is a flow chart of the multi-time scale loss reduction optimization method for distribution network of the present invention;

[0107] Figure 2 It is a multi-time scale dispatching system architecture of the multi-time scale loss reduction optimization method of the distribution network considering the flexible charging and discharging of electric vehicles in the present invention;

[0108] Figure 3 It is an improved IEEE33 node system diagram of the multi-time scale loss reduction optimization method of the distribution network considering the flexible charging and discharging of electric vehicles in the present invention;

[0109] Figure 4 It is a forecast diagram of wind and solar power output in the day before the present invention of the multi-time scale loss reduction optimization method of the distribution network considering the flexible charging and discharging of electric vehicles;

[0110] Figure 5 It is a diagram of the electric power balance of the multi-time scale loss reduction optimization method of the distribution network considering the flexible charging and discharging of electric vehicles of the present invention;

[0111] Figure 6 It is an equivalent load situation diagram under different scenarios of the multi-time scale loss reduction optimization method of the distribution network considering the flexible charging and discharging of electric vehicles in the present invention;

[0112] Figure 7 It is a network loss diagram under different scenarios of the multi-time scale loss reduction optimization method of the distribution network considering the flexible charging and discharging of electric vehicles in the present invention;

[0113] Figure 8 It is a flow chart of electric vehicle charging and discharging load generated by Monte Carlo simulation of the multi-time scale loss reduction optimization method of the distribution network considering the flexible charging and discharging of electric vehicles in the present invention;

[0114] Figure 9 It is a schematic diagram of the structure of the multi-time scale loss reduction optimization system of the distribution network of the present invention. DETAILED DESCRIPTION

[0115] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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.

[0116] Embodiment 1:

[0117] The following is an introduction to the relevant terms involved in the embodiments of the present application:

[0118] Electric vehicle: Electric vehicle (BEV) refers to a vehicle that uses an onboard power source as a power source, uses an electric motor to drive the wheels, and complies with various requirements of road traffic and safety regulations. As its impact on the environment is relatively small compared to traditional vehicles, its prospects are widely optimistic. Working principle: battery - current - power regulator - motor - power transmission system - drive the vehicle.

[0119] Distribution network: Distribution network refers to the power grid that receives electric energy from the transmission network or regional power plants and distributes it locally or step by step according to voltage to various users through distribution facilities. It is composed of overhead lines, cables, towers, distribution transformers, disconnectors, reactive power compensators and some ancillary facilities, and plays an important role in distributing electric energy in the power grid.

[0120] like Figure 1 As shown, a multi-time scale loss reduction optimization method for a distribution network comprises the following steps:

[0121] S101: receiving driving data related to the electric vehicle, and calculating the charging power of the electric vehicle using a Monte Carlo simulation method according to the driving data related to the electric vehicle;

[0122] Electric vehicle driving related data includes: vehicle classification based on travel habits, battery characteristics, daily mileage l , Start charging time t 0 and initial state of charge SOC 0. Through survey statistics, the relevant factors approximately satisfy some probability distributions, and then the charging power of electric vehicles is randomly generated by using the Monte Carlo simulation method.

[0123] Electric vehicles are characterized and classified from different perspectives. The first type is mainly used for commuting between home and workplace, that is, commuting twice a day; the second type is owned by companies or unemployed people, who may make multiple trips a day. This method only considers the fast charging behavior of electric vehicles. The second type of vehicle can be charged by a slow charging device at home due to its flexible travel time, so it is not considered here.

[0124] Assuming battery capacity C bat Normal distribution:

[0125]

[0126] Where: is the mean value of battery capacity, is the standard deviation of the battery capacity.

[0127] According to statistical data, daily mileage l It approximately follows a log-normal distribution, that is, S ~ Log -N( m s , s 2 s ), whose probability function is:

[0128]

[0129] Where: is the mean mileage, is the standard deviation of mileage.

[0130] Probability distribution of electric vehicle starting charging time:

[0131] Electric vehicle charging start time t 0 follows a normal distribution, that is t 0~N( μt , s 2 t ), the probability density function is:

[0132]

[0133] Where: is the average of the initial charging time, is the standard deviation of the initial charging time.

[0134] Electric vehicle charging duration:

[0135] The charging time of an electric vehicle is related to the final battery state of charge. The initial state of charge of the wth vehicle is It is expressed as:

[0136]

[0137] Where: The power consumption per 100 km; is the battery capacity.

[0138] Charging time It is expressed as:

[0139]

[0140] Where: The actual charging time required for electric vehicles; is the expected SOC value of EV users; EV charging and discharging efficiency; is the charging power of a single electric vehicle w.

[0141] The total electric vehicle charging power is obtained by adding up the charging load of a single electric vehicle:

[0142]

[0143] Where: is the total number of electric vehicles, ( t ) is an electric car t The total charging power during the time period.

[0144] S102: Input the charging power of the electric vehicle into a pre-established dynamic time-of-use electricity price model, and output the dynamic time-of-use electricity price C(t). Input the dynamic time-of-use electricity price C(t) into a pre-established MPC-based day-ahead and intraday multi-time scale optimization scheduling strategy model, with the root node electricity purchase cost, electric vehicle user charging and discharging cost and network loss as the objective function, and output the optimized scheduling result.

[0145] A multi-time-scale dispatching system architecture for the distribution network considering electric vehicles is constructed. According to the uncertainty of wind and solar power output, a feedback rolling optimization strategy based on model predictive control (MPC) is established. Dynamic time-of-use electricity prices are introduced, and a mathematical model is constructed with the root node electricity purchase cost, electric vehicle user charging and discharging cost, and network loss as objective functions considering the charging and discharging characteristics of electric vehicles.

[0146] Model predictive control is also called rolling horizon optimization control. Compared with open-loop optimization control, model predictive control can achieve closed-loop optimization control for a period of time in the future, and through feedback correction, timely correct the deviation of control results caused by prediction errors and random factors, thereby greatly improving the robustness and anti-interference of the system. Based on this idea, a multi-time scale control strategy for distribution network considering electric vehicle charging and discharging based on model predictive control is established, which mainly includes the day-ahead optimization scheduling layer and the intra-day rolling scheduling layer. The day-ahead optimization model starts from the perspective of the root power purchase cost of the distribution network, and according to the equivalent load after adding electric vehicle charging, a reasonable time-of-use electricity price is formulated to guide users to control the charging and discharging power, so as to achieve the purpose of improving the load and reducing the loss of the distribution network.

[0147] A dynamic time-of-use electricity price model is established based on equivalent load. s The peak period is when the value is times the sum of the peak-to-valley difference, which is lower than the mean. s The time when the difference between the peak and valley is double is the valley period, and the rest of the time is the normal period. The time-of-use electricity price model is expressed as:

[0148]

[0149]

[0150]

[0151]

[0152] Where: C(t) is the dynamic time-of-use electricity price, is the electricity price of equivalent load during normal hours, d The fluctuation range of electricity prices during peak and off-peak hours. is the equivalent load, is the basic load, is the mean value of equivalent load, is the peak-to-valley difference of equivalent load, , are the peak and valley values ​​of the equivalent load respectively.

[0153] Objective function:

[0154] The root node has the lowest electricity purchase cost:

[0155]

[0156] Where: F 1 is the objective function of minimizing the power purchase cost of the root node; P buy Input active power for the root node, T is the total scheduling period.

[0157] The charging and discharging costs for electric vehicle users are minimal:

[0158]

[0159] Where: The objective function for minimizing the charging and discharging cost for electric vehicle users is: is the charging electricity price of electric vehicles at time t; To provide electric vehicle users with discharge incentive subsidies; and are the charging and discharging power of electric vehicles respectively;

[0160] Minimum network loss:

[0161]

[0162] Where: is the objective function of minimizing network loss, For branch ij Resistance; for t Time branch ij The current;

[0163] The objective function is normalized by linear weighting as follows:

[0164]

[0165] Where: , and They are , and The weight coefficient of

[0166] Constraints:

[0167] Electric vehicle charging and discharging

[0168] Based on the charging power obtained by the Monte Carlo simulation method, the charging and discharging power boundary of electric vehicle users is obtained through the proportional coefficient of participating in the vehicle-grid interaction. And it must meet the requirement that the sum of the charging and discharging power in the whole cycle is equal to the sum of the charging power requirements; Single car charging and discharging power constraints:

[0169]

[0170] Electric vehicle state of charge constraints:

[0171]

[0172] Where: is the state of charge of electric vehicle i.

[0173] Energy storage constraints:

[0174] Energy storage equipment power constraints:

[0175]

[0176] Energy storage equipment charge constraints:

[0177]

[0178] In the formula, SOC ES is the SOC value of the battery; and are the battery charging and discharging power respectively; and They are the charge and discharge status marks of the battery respectively, satisfying the mutual exclusion constraint.

[0179]

[0180] Photovoltaic constraints:

[0181]

[0182] Where: yes t Actual photovoltaic output at all times; and are the minimum and maximum photovoltaic output respectively.

[0183] Transferable load constraints:

[0184]

[0185] Where: is the load after the node is transferred; is the base load of the node.

[0186] Distribution network related constraints:

[0187]

[0188]

[0189]

[0190]

[0191]

[0192]

[0193]

[0194] Where: P Indicates line disconnection; and Node j Department t Active and reactive power injected at all times; and They are t Time branch i,j Active power and reactive power on and Branch i,j resistance and reactance; and For Node i,j The node voltage of and They are t The minimum and maximum values ​​of node voltage at the moment; For branch t The maximum current at that moment.

[0195] Due to the deviation between the day-ahead optimization scheduling plan and the actual operation caused by the long-term forecast and climate environment, in the intraday scheduling stage, the rolling time domain comprehensive operation cost and its network loss are minimized as the optimization target to build an MPC intraday time domain rolling optimization scheduling model based on electric vehicle flexible charging and discharging. The day-ahead scheduling plan is revised in time to minimize the scheduling deviation. The scheduling cycle is set every time within the day for 15 minutes. That is, during the intraday operation t According to the real-time forecast of wind and solar power output, the distributed power supply, energy storage, and electric vehicle charging and discharging in the future N periods are optimized and dispatched every other period. The incremental result of the first period in the time domain is added to the feedback correction link as a real-time control instruction, and the model prediction output is corrected through the actual measurement value to form a closed-loop control. Repeat this process until the next period t+1. Prediction model:

[0196] The output increment of the root node power purchase, battery and electric vehicle charging and discharging power is assumed to be the controllable variable, and the load, wind power output and photovoltaic output are the disturbance variables. Through rolling optimization, the optimal output value of each controllable distributed power source, battery and electric vehicle charging and discharging in the future limited time domain can be obtained. In the above optimization model, the controllable unit of the MPC model u The prediction model expression is as follows:

[0197]

[0198] Puu ( k ) is a controllable unit u exist k The initial output value of the time period; U is the total number of controllable units; ( k+i|k ) period predicts the future[ k +( i -1), k+i ]Controllable units within the time period u The output increment; pu ( k+i|k ) is the k-period prediction for the future k+i Output value of controllable unit u in time period; , is the climbing power value that the output of each controllable unit needs to meet in adjacent prediction time periods; N is the total number of time periods in the forecast time domain.

[0199] Intraday objective function and constraints:

[0200] The transferable load is a slow dispatching device, which will not change in the intraday stage after being determined in the day-ahead stage. The intraday objective function and constraints are similar to those of the day-ahead optimization model, including AC power flow equation constraints, operation safety constraints, and active and reactive control equipment behavior constraints.

[0201] Feedback correction:

[0202] There is a deviation in the output control sequence obtained by actual and predictive control, so it is necessary to introduce a feedback correction link to correct the model output of the system with the actual measurement value, and use the actual measurement value as the initial value of the next round of rolling optimization, thus forming a closed-loop optimization control.

[0203]

[0204] In the formula, t=t0+ , t0 For the current moment, For the forecast sequence in the new round of rolling optimization scheduling, is the latest predicted sequence.

[0205] Specifically, the present invention is further described below by way of embodiments:

[0206] Example

[0207] The architecture of multi-timescale scheduling system is as follows Figure 2 As shown, the modified IEEE33 node system is as follows Figure 3 As shown in Figure 1, its rated voltage is 12.66kv, rated power is 10MW, nodes 7, 8, 24, and 25 are daily peak load nodes, with a certain proportion of transferable loads set; node 18 is an energy storage node; node 28 is a photovoltaic power station; node 32 is an electric vehicle charging station, and the number of electric vehicles available for dispatch is set to 500. The disordered charging and discharging power of electric vehicles generated by Monte Carlo is shown in Figure 1. Figure 8 shown.

[0208] Figure 5 The balance of electric power. By coordinating wind, solar and storage vehicles to smooth load fluctuations, the peak-to-valley difference of the distribution network load is reduced, making the overall load level flat, and the optimization effect is obvious.

[0209] Four scenarios are set up to analyze and compare the impact of the proposed strategy on the participation of electric vehicles in charging in the distribution system. The scheduling results are shown below.

[0210] Scenario 1: Under the traditional time-of-use electricity price, the single-time-scale optimization strategy of V2G is not considered, and electric vehicles are charged in an unordered manner.

[0211] Scenario 2: Under the traditional time-of-use electricity price, consider the single time scale optimization strategy of V2G.

[0212] Scenario 3: Consider the multi-time scale optimization strategy of V2G under traditional time-of-use electricity prices.

[0213] Scenario 4: The present invention proposes an optimization strategy that takes into account V2G and dynamic time-of-use electricity prices.

[0214] Comparison of equivalent load and network loss results under different scenarios Figure 6 , Figure 7 As shown in the figure, in scenario 1, with traditional electricity prices and disorderly charging of electric vehicles, a large number of electric vehicles do not need to be charged between 17:00 and 21:00, which makes the load curve appear "peak on peak", increases the peak-to-valley difference, causes higher network losses, and increases the support burden of the power grid and the charging costs of users. Compared with disorderly charging, scenario 3 reduces a certain equivalent load peak in the V2G mode, smoothes load fluctuations and reduces network losses to a certain extent, but it is still an optimization process on a single time scale, and does not consider the fluctuations of wind and solar power in a short period of time and the limitations of fixed electricity prices on the optimization effect, resulting in its "peak shaving and valley filling" and loss reduction effects are not significant.

[0215] Scenarios 3 and 4 are multi-time scale optimization scheduling. Compared with single time scale, during the peak load periods of 17:00-21:00 and 10:00-13:00, the charging load of electric vehicles is significantly reduced and transferred to the load valley period, and the network loss is greatly reduced compared with scenarios 1 and 2. In order to better encourage users to participate in scheduling, scenario 4 adopts a dynamic time-of-use electricity price optimization strategy, which has smaller load peak-valley difference and network loss than scenario 3.

[0216] Embodiment 2: In the second aspect, as Figure 9 As shown, in order to achieve the above-mentioned purpose, the present invention discloses a multi-time scale loss reduction optimization system for a distribution network, comprising:

[0217] The power calculation module 11 is used to receive the driving data of the electric vehicle, and calculate the charging power of the electric vehicle using the Monte Carlo simulation method according to the driving data of the electric vehicle;

[0218] The optimization scheduling module 12 is used to input the charging power of the electric vehicle into a pre-established dynamic time-of-use electricity price model, and output the dynamic time-of-use electricity price C(t). The dynamic time-of-use electricity price C(t) is input into a pre-established MPC-based day-ahead and intraday multi-time scale optimization scheduling strategy model, with the root node electricity purchase cost, electric vehicle user charging and discharging cost and network loss as the objective function, and the output is the optimization scheduling result.

[0219] In conjunction with the second aspect, in some implementations of the second aspect, the system further includes: the electric vehicle driving related data in the power calculation module 11 includes: vehicle classification considering travel habits, battery characteristics, daily mileage l , Start charging time t 0 and initial state of charge SOC0. Through survey statistics, the relevant factors approximately satisfy some probability distributions, and then the charging power of electric vehicles is randomly generated by using the Monte Carlo simulation method.

[0220] Electric vehicles are characterized and classified from different perspectives. The first type is mainly used for commuting between home and workplace, that is, commuting twice a day; the second type is owned by companies or unemployed people, who may make multiple trips a day. This method only considers the fast charging behavior of electric vehicles. The second type of vehicle can be charged by a slow charging device at home due to its flexible travel time, so it is not considered here.

[0221] Assuming battery capacity C bat Normal distribution:

[0222]

[0223] is the mean value of battery capacity, is the standard deviation of the battery capacity.

[0224] According to statistical data, daily mileage l It approximately follows a log-normal distribution, that is, S ~ Log -N( m s , s 2 s ), whose probability function is:

[0225]

[0226] Where: is the mean mileage, is the standard deviation of mileage.

[0227] Probability distribution of electric vehicle starting charging time:

[0228] Electric vehicle charging start time t 0 follows a normal distribution, that is t 0~N( μt , s 2 t ), the probability density function is:

[0229]

[0230] Where: is the average of the initial charging time, is the standard deviation of the initial charging time.

[0231] Electric vehicle charging duration:

[0232] The charging time of an electric vehicle is related to the final battery state of charge. The initial state of charge of the wth vehicle is It is expressed as:

[0233]

[0234] Where: The power consumption per 100 km; is the battery capacity.

[0235] Charging time t c,w It is expressed as:

[0236]

[0237] Where: The actual charging time required for electric vehicles; is the expected SOC value of EV users; EV charging and discharging efficiency; is the charging power of a single electric vehicle w.

[0238] The total electric vehicle charging power is obtained by adding up the charging load of a single electric vehicle:

[0239]

[0240] Where: is the total number of electric vehicles, ( t ) is an electric car t The total charging power during the time period.

[0241] The constraints of the intraday optimization model in the multi-time scale optimization scheduling strategy model based on MPC established in the optimization scheduling module 12 include AC power flow equation constraints, operation safety constraints, and active and reactive control equipment behavior constraints. There is a deviation in the output control sequence obtained by actual and predictive control, so it is necessary to introduce a feedback correction link to correct the model output of the system with the actual measurement value, and use the actual measurement value as the initial value of the next round of rolling optimization, thereby forming a closed-loop optimization control. That is:

[0242]

[0243] In the formula, t=t 0 + , t 0 is the current time, For the forecast sequence in the new round of rolling optimization scheduling, is the latest predicted sequence.

[0244] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0245] The above shows and describes the basic principles, main features and advantages of the present disclosure. Those skilled in the art should understand that the present disclosure is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present disclosure. Without departing from the spirit and scope of the present disclosure, the present disclosure may have various changes and improvements, and these changes and improvements fall within the scope of the present disclosure to be protected.

Claims

1. A multi-time scale loss reduction optimization method for distribution network, characterized in that: The method comprises the following steps: Receiving driving data related to the electric vehicle, and calculating the charging power of the electric vehicle using a Monte Carlo simulation method according to the driving data related to the electric vehicle; The charging power of electric vehicles is input into a pre-established dynamic time-of-use electricity price model, and the dynamic time-of-use electricity price is output. The dynamic time-of-use electricity price is input into a pre-established multi-time scale optimization scheduling strategy model based on MPC in the day ahead, and the root node power purchase cost, electric vehicle user charging and discharging cost and network loss are used as the objective function, and the optimized scheduling result is output; The calculation process of the day-ahead optimization model in the pre-established MPC-based day-ahead intraday multi-time scale optimization dispatch strategy model takes the root node power purchase cost, electric vehicle user charging and discharging cost and network loss as the objective function: Objective function: The root node has the lowest electricity purchase cost: Where: F 1 is the objective function of minimizing the power purchase cost of the root node, P buy Input active power for the root node, T is the total scheduling period; The charging and discharging costs for electric vehicle users are minimal: Where: The objective function for minimizing the charging and discharging cost for electric vehicle users is: is the charging electricity price of electric vehicles at time t; To provide electric vehicle users with discharge incentive subsidies; and are the charging and discharging power of electric vehicles respectively; Minimum network loss: Where: is the objective function of minimizing network loss, For branch ij The resistance of for t Time branch ij The current; The objective function is normalized by linear weighting as follows: Where: , and They are , and The weight coefficient of The constraints include: Single vehicle charging and discharging power constraints, electric vehicle charge state constraints, energy storage equipment power constraints, energy storage equipment charge constraints, photovoltaic constraints, transferable load constraints and distribution network related constraints; Controllable unit u The prediction model expression is as follows: P uo ( k ) is a controllable unit u exist k The initial output value of the time period; U is the total number of controllable units; ( k+i|k )for k Time period prediction future k +( i -1), k+i ]Controllable units within the time period u The output increment; p u ( k+i|k ) is the k-period prediction for the future k+i Output value of controllable unit u in time period; , is the climbing power value that the output of each controllable unit needs to meet in adjacent prediction time periods; N is the total number of time periods in the forecast time domain.

2. A distribution network multi-time scale loss reduction optimization method according to claim 1, characterized in that: The electric vehicle driving related data includes: classification of vehicles considering travel habits, battery characteristics, daily mileage l , Start charging time t 0 and initial state of charge .

3. A distribution network multi-time scale loss reduction optimization method according to claim 2, characterized in that: Daily driving range of the electric vehicle l The probability function is: Where: is the mean mileage, is the standard deviation of mileage; Probability distribution of electric vehicle starting charging time: Electric vehicle charging start time t 0 follows a normal distribution, that is t 0~N( μt , σ 2 t ), the probability density function is: Where: is the average of the initial charging time, is the standard deviation of the initial charging time; Electric vehicle charging duration The charging and discharging time of electric vehicles is related to the final battery state of charge. The initial state of charge of the wth vehicle is It is expressed as: Where: The power consumption per 100 km; is the battery capacity; Charging time t c,w It is expressed as: Where: The actual charging time required for electric vehicles; is the expected SOC value of EV users; EV charging and discharging efficiency; is the charging power of a single electric vehicle w; The total electric vehicle charging power is obtained by adding up the charging load of a single electric vehicle: Where: is the total number of electric vehicles, ( t ) is an electric car t The total charging power during the time period.

4. A distribution network multi-time scale loss reduction optimization method according to claim 1, characterized in that: The process of inputting the charging power of the electric vehicle into the pre-established dynamic time-of-use electricity price model and outputting the dynamic time-of-use electricity price is as follows: When the equivalent load is higher than the mean σ The peak period is when the value is times the sum of the peak-to-valley difference, which is lower than the mean. σ The time when the difference between the peak and valley is double is the valley period, and the rest of the time is the normal period. The time-of-use electricity price model is expressed as: Where: C(t) is the dynamic time-of-use electricity price, is the electricity price of equivalent load during normal hours, δ The fluctuation range of electricity prices during peak and off-peak hours. is the equivalent load, is the basic load, is the mean value of equivalent load, is the peak-to-valley difference of equivalent load, , are the peak and valley values ​​of the equivalent load respectively.

5. A distribution network multi-time scale loss reduction optimization method according to claim 1, characterized in that: The constraints of the intraday optimization model in the pre-established MPC-based day-ahead intraday multi-time scale optimization dispatch strategy model include AC power flow equation constraints, operation safety constraints, and active and reactive control equipment behavior constraints. There is a deviation in the output control sequence obtained by actual and predictive control, so the feedback correction is as follows: In the formula, t=t0+ , t0 For the current moment, For the forecast sequence in the new round of rolling optimization scheduling, is the latest predicted sequence.

6. A multi-time scale loss reduction optimization system for distribution network, characterized in that: include: A power calculation module is used to receive the driving data of the electric vehicle and calculate the charging power of the electric vehicle using the Monte Carlo simulation method according to the driving data of the electric vehicle; The optimization scheduling module is used to input the charging power of the electric vehicle into a pre-established dynamic time-of-use electricity price model, and output the dynamic time-of-use electricity price. The dynamic time-of-use electricity price is input into a pre-established multi-time scale optimization scheduling strategy model based on MPC in the day before and within the day, and the root node power purchase cost, the electric vehicle user charging and discharging cost and the network loss are used as the objective function to output the optimization scheduling result. In the optimization scheduling module, the day-ahead optimization model in the pre-established MPC-based day-ahead intraday multi-time scale optimization scheduling strategy model takes the root node power purchase cost, electric vehicle user charging and discharging cost and network loss as the objective function of the calculation process: Objective function: The root node has the lowest electricity purchase cost: Where: F 1 is the objective function of minimizing the power purchase cost of the root node, P buy Input active power for the root node, T is the total dispatching cycle; the charging and discharging cost of electric vehicle users is the minimum: Where: The objective function for minimizing the charging and discharging cost for electric vehicle users is: is the charging electricity price of electric vehicles at time t; To provide electric vehicle users with discharge incentive subsidies; and are the charging and discharging power of electric vehicles respectively; Minimum network loss: Where: is the objective function of minimizing network loss, For branch ij The resistance of for t Time branch ij The current; The objective function is normalized by linear weighting as follows: Where: , and They are , and The weight coefficient of The constraints include: Single vehicle charging and discharging power constraints, electric vehicle charge state constraints, energy storage equipment power constraints, energy storage equipment charge constraints, photovoltaic constraints, transferable load constraints and distribution network related constraints; Controllable unit u The prediction model expression is as follows: P uo ( k ) is a controllable unit u exist k The initial output value of the time period; U is the total number of controllable units; ( k+i|k )for k Time period prediction future k +( i -1), k+i ]Controllable units within the time period u The output increment; p u ( k+i|k ) is the k-period prediction for the future k+i Output value of time-controlled unit u , is the climbing power value that the output of each controllable unit needs to meet in adjacent prediction time periods; N is the total number of time periods in the forecast time domain.

7. A distribution network multi-time scale loss reduction optimization system according to claim 6, characterized in that: The electric vehicle driving related data in the power calculation module includes: vehicle classification considering travel habits, battery characteristics, daily mileage l , Start charging time t 0 and initial state of charge SOC 0; Daily driving range of electric vehicles in power calculation module l The probability function is: Where: is the mean mileage, is the standard deviation of mileage; Probability distribution of electric vehicle starting charging time: Electric vehicle charging start time t 0 follows a normal distribution, that is t 0~N( μt , σ 2 t ), the probability density function is: Where: is the average of the initial charging time, is the standard deviation of the initial charging time; Electric vehicle charging duration The charging and discharging time of electric vehicles is related to the final battery state of charge. The initial state of charge of the wth vehicle is It is expressed as: Where: The power consumption per 100 km; is the battery capacity; Charging time It is expressed as: Where: The actual charging time required for electric vehicles; is the expected SOC value of EV users; EV charging and discharging efficiency; is the charging power of a single electric vehicle w; The total electric vehicle charging power is obtained by adding up the charging load of a single electric vehicle: Where: is the total number of electric vehicles, ( t ) is an electric car t The total charging power during the time period.

8. A distribution network multi-time scale loss reduction optimization system according to claim 7, characterized in that: In the optimization scheduling module, the charging power of electric vehicles is input into the pre-established dynamic time-of-use electricity price model, and the process of outputting the dynamic time-of-use electricity price is as follows: When the equivalent load is higher than the mean σ The peak period is when the value is times the sum of the peak-to-valley difference, which is lower than the mean. σ The time when the difference between the peak and valley is double is the valley period, and the rest of the time is the normal period. The time-of-use electricity price model is expressed as Where: C(t) is the dynamic time-of-use electricity price, is the electricity price of equivalent load during normal hours, δ The fluctuation range of electricity prices during peak and off-peak hours. is the equivalent load, is the basic load, is the mean value of equivalent load, is the peak-to-valley difference of equivalent load, , are the peak and valley values ​​of the equivalent load respectively.

9. A distribution network multi-time scale loss reduction optimization system according to claim 6, characterized in that: The constraints of the intraday optimization model in the MPC-based day-ahead and intraday multi-time scale optimization scheduling strategy model pre-established in the optimization scheduling module include AC power flow equation constraints, operation safety constraints, and active and reactive control equipment behavior constraints. There is a deviation in the output control sequence obtained by actual and predictive control, so the feedback correction is as follows: In the formula, t=t0+ , t0 For the current moment, For the forecast sequence in the new round of rolling optimization scheduling, is the latest predicted sequence.

Images