A power grid scheduling method considering orderly charging of electric vehicles and new energy consumption

Abstract

The application discloses a kind of power grid scheduling methods considering orderly charging of electric vehicle and new energy consumption, including collecting power grid load historical data, weather information, electricity price information, fixed time acquisition new energy real-time output information, EV charging demand and charging state;The data collected are input into day-ahead scheduling model, and the day-ahead total expected cost function of power grid within time T is constructed;Constraint condition is constructed, and preliminary unit start-stop plan and EV charging day-ahead plan are obtained according to constraint condition, and dispatching instructions are issued according to plan;On the basis of cost function plan, based on the prediction results of wind and light generation and the change of EV charging load, a regulation mathematical model is constructed;Improved vole algorithm is used to optimize the variables in the regulation mathematical model, so as to minimize the expected cost.The application reduces the operation cost of power grid, guarantees the charging demand of users under the premise of considering the charging willingness of EV users, and provides an intelligent solution for the dispatching and operation of high-proportion renewable energy power grid.

Description

Technical Field

[0001] This invention relates to the field of smart grid optimized operation and scheduling technology, and in particular to a grid scheduling method that considers the orderly charging of electric vehicles and the consumption of new energy sources. Background Technology

[0002] With the deepening of my country's "dual-carbon" strategy, the installed capacity of new energy power generation, represented by wind power and photovoltaics, has continued to grow rapidly. However, the inherent volatility and intermittency of new energy power generation pose significant challenges to power system dispatch and operation. Currently, grid dispatch mainly adopts the traditional "source follows load" model, which has shown obvious inadequacy in scenarios with a high proportion of new energy. On the one hand, the regulation capacity of conventional thermal power units is insufficient to completely suppress the fluctuations in new energy output, leading to a large amount of wind and solar curtailment. Statistics show that in 2022, the amount of wind and solar curtailment in China exceeded 30 billion kilowatt-hours, and the curtailment rate in some areas rich in new energy was still as high as 10%. On the other hand, the existing dispatch system lacks a collaborative optimization mechanism with flexible loads, making it difficult to fully utilize the regulation potential of demand-side resources. Especially in grids with a high proportion of renewable energy, the limitations of this operating model are becoming increasingly prominent, severely restricting the efficient consumption of new energy and the safe and economical operation of the grid.

[0003] Meanwhile, the electric vehicle industry is experiencing explosive growth, and the spatiotemporal distribution characteristics of large-scale electric vehicle charging loads have brought new challenges and opportunities to the power system. Currently, the management of electric vehicle charging by the power grid is still in its early stages, mainly relying on simple orderly charging or time-of-use pricing strategies, failing to fully explore its regulatory potential as a distributed energy storage resource. More importantly, existing dispatching systems often optimize the generation and consumption sides separately, lacking an overall optimization framework for source-load coordination. Under the dual uncertainties of renewable energy output and electric vehicle charging demand, traditional deterministic dispatching models are no longer sufficient to meet the requirements of safe grid operation. For example, during the evening peak hours of a certain provincial power grid, load spikes caused by concentrated electric vehicle charging, coupled with a sudden drop in photovoltaic power generation, have repeatedly triggered local voltage exceedances and line overloads. These real-world problems urgently require the development of new power grid dispatching strategies to achieve coordinated optimization of renewable energy and electric vehicle charging loads. Summary of the Invention

[0004] In order to overcome the above-mentioned problems in the existing technology, the present invention proposes a power grid dispatching method that considers the orderly charging of electric vehicles and the consumption of new energy sources.

[0005] The technical solution adopted by this invention to solve its technical problem is: a power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources, comprising the following steps: Step 1: Collect historical data on grid load, meteorological information, and electricity price information; collect real-time power output information of new energy sources, EV charging demand, and charging status at fixed intervals. Step 2: Input the data collected in Step 1 into the day-ahead scheduling model to construct the day-ahead total expected cost function F1 of the power grid within time T; Step 3: Construct constraints, obtain preliminary unit start-up and shutdown plans and EV charging day plans based on the constraints, and issue dispatch instructions according to the plans; Step 4: Based on the results obtained from the cost function F1 in Step 2 and Step 3, and according to different operating scenarios, construct the regulation mathematical model F2 based on the prediction results of wind and solar power generation and the changes in EV charging load; Step 5: The improved lemming algorithm is used to optimize the variables in the control mathematical model F2 in step 4, thereby minimizing the expected cost.

[0006] In the aforementioned grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources, the day-ahead total expected cost function F1 of the grid in time T during step 2 is specifically as follows: in, t For a period of time; g For the first g One generator set; N G The number of generator sets; For the first g Start-up cost of a generator set; The downtime cost of the g-th generator unit; u g,t , v g,t Let ∈{0, 1} be a function, and let the generator set be... g During the period t Internal boot u g,t =1 Or shutdown v g,t =1, otherwise u g,t , v g,t All are 0; The marginal cost of generating electricity for unit g; P g,t For the unit g During the period t Those who contribute their efforts; C curt As a penalty factor for abandoning wind and solar power; , They are respectively tThe amount of wind and solar power curtailed during specific periods; C EV,i,t For EV car users i During the period t The electricity price for charging within the country; For the first time planned i EV t Charging power during a given period; C grid,t For time period t Electricity purchased from the main grid; P grid,t The power purchased from the main grid within time period t.

[0007] The above-mentioned grid dispatching method considering orderly charging of electric vehicles and consumption of new energy includes the following constraints in step 3: power balance constraints within time period t, upper and lower limits of generator output, generator ramping constraints, minimum start-stop time constraints, new energy power generation consumption constraints, EV charging constraints, user satisfaction constraints, fluctuation constraints of renewable energy output, and spatiotemporal constraints of EV charging load.

[0008] The aforementioned grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources, specifically defines the power balance constraint within time period t as follows: in, P load,t This represents the power of the grid's conventional load. P w,t The active power of wind power generation. P pv,t The active power of photovoltaic power generation P EV,i,t For the first i Taiwanese electric vehicles during the period t The charging power.

[0009] The aforementioned grid dispatching method considering orderly charging of electric vehicles and the consumption of new energy sources specifically includes the following EV charging constraints: No. i Taiwanese EV cars during the period t The charging demand must meet the following constraints: , In the formula, Δ t This refers to the charging time and the amount of charge required by the EV. The specific constraints on user satisfaction are: User satisfaction must meet the following constraints: In the formula, For the user's preferred time period, Δ t i2 This refers to the charging period. β i The minimum satisfaction rate required by the user of the i-th EV vehicle.

[0010] The aforementioned grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources, wherein the fluctuation constraint of renewable energy output is represented by the power flow of branch l, specifically: in, branch road nm Maximum allowable current Pn, Qn The first n Active and reactive power of each node N For the number of nodes, g nm , b nm These are the admittances of the branches, θ nm branch road nm The phase angle difference, I nm branch road nm The actual current value; The spatiotemporal constraints of the EV vehicle charging load are represented by the system frequency, specifically: Where, Δ f For frequency variation, Δ P This is due to a power deficit. K sys Δ is the system frequency regulation coefficient. f max This represents the maximum allowable frequency deviation.

[0011] In the aforementioned grid dispatching method that considers orderly charging of electric vehicles and the consumption of new energy sources, the control mathematical model F2 in step 4 is specifically as follows: in, E The predicted control costs for the first half hour of the day; In order to be in t The amount of renewable energy curtailed during a given period; For the first i EV t Actual charging power during the time period; For the first i The unit compensation cost for deviation in EV charging power; φ( i Let be a 0-1 function, representing the user's willingness to charge. If the first digit is 0... i The value is 1 if the EV user agrees to charging, otherwise it is 0. For the first t The unit cost of emergency power purchase during specific time periods; For the first t Emergency power purchase capacity during specific time periods; in, In order to be in t Load demand power during the period This refers to renewable energy power that can be absorbed by the power grid. For conventional units g Minimum output, P re,t In order to be in t Forecasted output of new energy sources during the period N EV For EV quantity, N G This refers to the number of conventional generating units.

[0012] The aforementioned grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources, in which the improved lemming algorithm is used as the variable in the regulation mathematical model F2 as the lemming, the specific regulation process is as follows: Step A: Initialize algorithm parameters, set population size N, and maximum number of iterations T. MAX The upper and lower bounds of a variable; Step B: Randomly initialize the population positions, traverse all lemming positions, evaluate the fitness value of each lemming's initial position, find the individual with the lowest cost in the population, and record it as the current best position. Step C: At the beginning of each iteration, update the energy factor E. The larger the value of E, the more lemmings tend to explore; the smaller the value of E, the more lemmings tend to develop. Step D: Determine whether to explore or develop based on the current iteration's E value. If E > 1, choose the exploration behavior; if E < 1, choose the development behavior. Update the lemming position and constrain its upper and lower limits according to the constraints in Step 3. Calculate the fitness value of the new position based on the current lemming position and the day-ahead expected cost function F1 of the power grid in Time T in Step 2. If it is higher than the current optimal fitness value, update the global optimum. Step E: Determine if the maximum number of iterations has been reached. If it has, exit the loop and output the optimal position and corresponding best fitness value obtained through optimization. Otherwise, return to step C to continue iteratively solving the problem.

[0013] In the aforementioned grid dispatching method that considers orderly charging of electric vehicles and the consumption of new energy sources, the lemming position update formula in step D is as follows: in, For the first Only lemmings in the first Position in the next iteration This is the optimal position for the lemming at present. It serves as a directional marker, used to change the individual's search direction; It exhibits Brownian motion and follows a standard normal distribution; r1 is a random number uniformly distributed in the interval [-1, 1], r6 and r8 are random numbers in the interval [0, 2π], r9 is a random number following an exponential distribution of (0, ∞), and r5 and r7 are random numbers. , The location of an individual randomly selected within a lemming population; The formula for updating the lemming position in step D is as follows: in, r is the spiral shape factor. 10 G is a random number in the range [0, 1], G represents the escape coefficient, Levy represents the Levy flight function, and b represents the compression factor.

[0014] The beneficial effects of this invention are that it proposes a smart grid dispatching strategy that integrates orderly electric vehicle (EV) charging with the coordinated optimization of renewable energy consumption. Through a multi-timescale dynamic optimization framework, it treats EV charging load as a flexible and adjustable resource, achieving spatiotemporal matching with the output characteristics of wind and solar power. It innovatively proposes intraday EV charging prediction models and curtailment power prediction models, considering the impact of renewable energy output uncertainty on control costs. While ensuring the safe and stable operation of the grid, it significantly improves the renewable energy consumption rate. This strategy reduces grid operating costs through time-of-use pricing and charging demand-side management, while ensuring user charging needs are met while considering the charging intentions of EV users, providing an intelligent solution for the dispatching and operation of grids with a high proportion of renewable energy. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the process of this invention; Figure 2 This is a schematic diagram of the improved artificial lemming optimization algorithm of this invention. Detailed Implementation

[0016] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] This embodiment discloses a grid dispatching method that considers the orderly charging of electric vehicles (EVs) and the absorption of renewable energy. Through a multi-timescale dynamic optimization framework, it treats EV charging load as a flexible and adjustable resource, achieving spatiotemporal matching with the output characteristics of wind and solar power. It innovatively proposes intraday EV charging prediction models and curtailment power prediction models, considering the impact of renewable energy output uncertainty on control costs. While ensuring the safe and stable operation of the grid, it significantly improves the renewable energy absorption rate. This strategy reduces grid operating costs through time-of-use pricing and charging demand-side management, while ensuring user charging needs are met while considering the charging intentions of EV users, providing an intelligent solution for the dispatching and operation of grids with a high proportion of renewable energy.

[0018] Due to the significant uncertainties in renewable energy output and EV charging demand, and the lack of uniformity in their timing, this invention employs model predictive control (MMC) technology to propose an optimized grid dispatch model. This model acquires real-time renewable energy output information (such as wind and solar power output) and EV charging status (including EV charging demand and status) every 30 minutes. While meeting the needs of EV users, it dynamically adjusts charging power matching to reduce wind and solar curtailment losses. The optimized model is divided into two layers: a day-ahead dispatch phase and an intraday optimization phase. The day-ahead model continuously adjusts the model to address existing uncertainties, predicting renewable energy output and EV charging power every half hour. Based on the day-ahead dispatch model (which fully considers the generation costs of conventional generators, renewable energy curtailment losses, EV charging dispatch compensation costs, and grid purchase costs), the optimized dispatch model is further refined based on the predictive model.

[0019] Specific scheduling methods are as follows: Figure 1 As shown, it includes the following steps: Step 1: Collect historical data on grid load, meteorological information, and electricity price information. Collect real-time power output information of new energy sources, EV charging demand, and charging status at fixed intervals.

[0020] Step 2: Input the data collected in Step 1 into the day-ahead scheduling model to construct the day-ahead total expected cost function F1 of the power grid within time T.

[0021] The day-ahead total expected cost function F1 of the power grid within time T is as follows: in, t For a period of time; g For the first g One generator set; NG The number of generator sets; For the first g Start-up cost of a generator set; The downtime cost of the g-th generator unit; u g,t , v g,t Let ∈{0, 1} be a function, and let the generator set be... g During the period t Internal boot u g,t =1 Or shutdown v g,t =1, otherwise u g,t , v g,t All are 0; The marginal cost of generating electricity for unit g; P g,t For the unit g During the period t Those who contribute their efforts; C curt As a penalty factor for abandoning wind and solar power; , They are respectively t The amount of wind and solar power curtailed during specific periods; C EV,i,t For EV car users i During the period t The electricity price for charging within the country; For the first time planned i EV t Charging power during a given period; C grid,t For time period t Electricity purchased from the main grid; P grid,t The power purchased from the main grid within time period t.

[0022] Step 3: Construct constraints, obtain preliminary unit start-up and shutdown plans and EV charging day plans based on the constraints, and issue dispatch instructions according to the plans.

[0023] When the power grid is operating, it should meet the following condition: Power purchased by the grid + output of new energy sources - curtailed wind and solar power = power consumed by conventional loads + power consumed by EV charging loads. t The power balance constraints within the time period are expressed as follows: In the formula: P load,t This represents the power of the grid's conventional load. P w,tThe active power of wind power generation. P pv,t The active power of photovoltaic power generation P EV,i,t For the first i Taiwanese electric vehicles during the period t The charging power, and other parameters are as described in equation (1).

[0024] The upper and lower limits of generator output are constrained as follows: In the formula: , These are the upper and lower limits of the output of generator set g, respectively, and other parameters are as described in equation (1).

[0025] The ramping constraint of the generator set is expressed as follows: In the formula: , These are the upper and lower limits of the climbing distance, respectively, and other parameters are as described in equation (1).

[0026] The minimum start / stop time constraint is as follows: In the formula: Constraints on renewable energy power generation and restrictions on curtailment are as follows: EV charging constraints, number i Taiwanese EV cars during the period t The charging demand must meet the following constraints: , In the formula: Δ t This refers to the charging time and the amount of charge required by the EV.

[0027] User satisfaction must meet the following constraints: In the formula: For the user's preferred time period, Δ t i2 This refers to the charging period. β i The minimum satisfaction rate required by the user of the i-th EV vehicle.

[0028] In scheduling models incorporating renewable energy injection and EV charging, grid security constraints must additionally consider the volatility of renewable energy output and the spatiotemporal uncertainty of EV charging load. (Branch) l The power flow satisfies the following constraint condition: In the formula: branch road nm Maximum allowable current Pn, Qn The first n Active and reactive power of each node N For the number of nodes, g nm , b nm These are the admittances of the branches, θ nm branch road nm The phase angle difference, I nm branch road nm The actual current value.

[0029] The system frequency must meet the following constraints In the formula: Δ f For frequency variation, Δ P This is due to power deficit (caused by fluctuations in new energy output or a sudden increase in EV charging load). K sys Δ is the system frequency regulation coefficient. f max This represents the maximum allowable frequency deviation.

[0030] In studying dispatch strategies for EV charging and renewable energy injection into the power grid, spatiotemporal coupling constraints must be considered. Taking into account the spatiotemporal distribution characteristics of EV charging loads, the simultaneous charging power of EVs on the same feeder must be limited to avoid local overload. Therefore, the line must satisfy the following constraints: In the formula: ε b To be connected to the same feeder b A collection of EV charging vehicles For feeder b The maximum allowable charging power.

[0031] Step 4: Based on the results obtained from the cost function F1 in Step 2 and Step 3, and according to different operating scenarios, construct the regulation mathematical model F2 based on the prediction results of wind and solar power generation and the changes in EV charging load.

[0032] Due to the strong uncertainty in new energy power output and EV charging load, scheduling strategies need to be adjusted according to real-time conditions to minimize the expected total cost. However, traditional scheduling strategies are mostly overly conservative with high redundancy, resulting in high operating costs. To reduce redundancy and address the problem of strong uncertainty, this invention describes uncertainty based on probabilistic scenarios, thereby minimizing the expected cost. Based on the day-ahead scheduling model, adjustments are made in different time periods according to different operating scenarios, based on the predicted results of wind and solar power generation and changes in EV charging load. The proposed mathematical model is expressed as follows: In the formula: E This refers to the predicted control costs for the first half hour of the day.

[0033] The mathematical model for the cost of corrective control within half an hour is expressed as follows: in, E The predicted control costs for the first half hour of the day; In order to be in t Curtailment of renewable energy during a given period (including curtailment of wind power and photovoltaic power). For the first i EV t Actual charging power during the time period; For the first i The unit compensation cost for deviation in EV charging power; φ ( i Let be a 0-1 function, representing the user's willingness to charge. If the first digit is 0... i The value is 1 if the EV user agrees to charging, otherwise it is 0. For the first t The unit cost of emergency power purchase during specific time periods; For the first t Emergency power purchase capacity during specific time periods. in, In order to be in t Load demand power during the period This refers to renewable energy power that can be absorbed by the power grid. For conventional units g Minimum output, P re,t In order to be in t Forecasted output of new energy sources during the period (including wind power output and photovoltaic output). N EV For EV quantity, N G This refers to the number of conventional generating units.

[0034] Step 5: The improved lemming algorithm is used to optimize the variables in the control mathematical model F2 in step 4, thereby minimizing the expected cost.

[0035] The Artificial Lemming Optimization Algorithm (ALA) simulates four typical behaviors of lemmings in nature: long-distance migration, burrowing, foraging, and evading predators, to solve complex optimization problems. ALA effectively addresses challenges such as premature convergence, insufficient exploration, and lack of robustness in high-dimensional spaces while maintaining computational efficiency, and boasts advantages such as fast convergence speed.

[0036] ① Long-distance migration behavior. Lemmings exhibit periodic large-scale migration behavior. When habitat resources are scarce or population density is too high, they undertake long-distance collective migrations to find more suitable living environments, updating their individual locations by combining Brownian motion and random orientation adjustments. The mathematical model for this behavior is as follows: In the formula, For the first Only lemmings in the first Position in the next iteration This is the optimal position for the lemming at present. It serves as a directional marker and can be used to change the search direction of an individual. It exhibits Brownian motion and follows a standard normal distribution. It is a random number that is uniformly distributed in the interval [-1, 1]. The location of an individual randomly selected within the lemming population.

[0037] ② Burrowing behavior. Lemmings update their position based on the current optimal solution and random individual perturbations, avoiding local optima through periodic perturbations. They randomly dig complex tunnel and burrow systems underground in their habitat to help them avoid predators and extreme weather; this behavior helps them discover potential optimal solutions for the algorithm. The position update formula is as follows: In the formula, It is a random number related to the current iteration number. The location of an individual randomly selected within the lemming population.

[0038] ③ Foraging behavior. Lemmings move short distances around known food sources within familiar habitats, adjusting their movement paths to search for richer resources within the known area. This behavior, mimicking the efficient search behavior of lemmings through a spiral search, allows for detailed exploration of local areas, contributing to improved solution quality. The location update formula is as follows: In the formula, The spiral shape factor represents the spiral shape of the random search during foraging.

[0039] ④ Predator avoidance behavior. When lemmings spot a predator, they quickly issue a warning and hide, escaping danger through underground tunnels and by running. This behavior, combined with levy flight to update position, enhances the algorithm's ability to search for the global optimum and avoids getting trapped in local optima. The position update formula is as follows: In the formula, It is the evasion coefficient. It is the Lévy flight function. It represents the maximum number of iterations.

[0040] To balance the exploration and development phases, an energy factor was designed. The expression is as follows: when When lemmings are in a certain phase, they will engage in an exploration phase, which involves migration or burrowing; otherwise, they will engage in a development phase, which involves foraging or escaping predators.

[0041] Although the artificial lemming optimization algorithm has a relatively simple structure and fewer parameters, allowing for a wide search range within the search space, it struggles to obtain high-precision optimal solutions in local searches. It is prone to getting trapped in local optima in the later stages of the search, leading to decreased population diversity and premature convergence. The specific improvements to the lemming optimization algorithm in this embodiment are as follows: ① A sine / cosine optimization strategy is introduced during the exploration phase. This strategy dynamically adjusts the search direction and step size of each individual by alternating the action of sine and cosine functions. The introduction of periodic perturbations during the search process helps expand the search range in the early stages, enabling the algorithm to escape local optima and enhancing its flexibility. The improved position update formula for the exploration phase is as follows: ② Introducing the Cauchy inverse cumulative distribution operator to the behavior of avoiding predators. The Cauchy inverse cumulative distribution operator is the inverse function of the Cauchy cumulative distribution function. In this invention, we take... It can generate both small local perturbation step sizes and large jump step sizes, which helps to explore distant regions of the search space, enabling individuals to escape the current search area, explore new solution spaces, and enhance population activity. The improved position update formula for predator avoidance behavior is as follows: In summary, the flowchart of the improved artificial lemming optimization algorithm is as follows: Figure 2 As shown.

[0042] When implementing the improved lemming algorithm, the variables in the mathematical model F2 are used as lemmings, and the specific implementation process is as follows: Step A: Initialize algorithm parameters, set population size N, and maximum number of iterations T. MAX The upper and lower bounds of a variable; Step B: Randomly initialize the population positions, traverse all lemming positions, evaluate the fitness value of each lemming's initial position, find the individual with the lowest cost in the population, and record it as the current best position. Step C: At the beginning of each iteration, update the energy factor E. The larger the value of E, the more lemmings tend to explore; the smaller the value of E, the more lemmings tend to develop. Step D: Determine whether to explore or develop based on the current iteration's E value. If E > 1, choose the exploration behavior; if E < 1, choose the development behavior. Update the lemming position and constrain its upper and lower limits according to the constraints in Step 3. Calculate the fitness value of the new position based on the current lemming position and the day-ahead expected cost function F1 of the power grid in Time T in Step 2. If it is higher than the current optimal fitness value, update the global optimum. Step E: Determine if the maximum number of iterations has been reached. If it has, exit the loop and output the optimal position and corresponding best fitness value obtained through optimization. Otherwise, return to step C to continue iteratively solving the problem.

[0043] The formula for updating the lemming position in step D is as follows: in, For the first Only lemmings in the first Position in the next iteration This is the optimal position for the lemming at present. It serves as a directional marker, used to change the individual's search direction; It exhibits Brownian motion and follows a standard normal distribution; r1 is a random number uniformly distributed in the interval [-1, 1], r6 and r8 are random numbers in the interval [0, 2π], r9 is a random number following an exponential distribution of (0, ∞), and r5 and r7 are random numbers. , The location of an individual randomly selected within a lemming population; The formula for updating the lemming position in step D is as follows: in, r is the spiral shape factor. 10 Let G be a random number in the range [0, 1], G represent the escape coefficient, Levy represent the Levy flight function, and b represent the compression factor.

[0044] This embodiment is the first to model the adjustable charging potential of EV clusters as a spatiotemporally flexible resource, dynamically matching it with wind / solar power output, and proposing a game model of "charging demand satisfaction - grid absorption benefits" to balance user willingness and grid demand.

[0045] The scheduling strategy employs a two-stage stochastic programming approach to address the uncertainties in new energy and EV forecasting. Based on day-ahead forecasts, an intraday time-segmented correction model is introduced, which not only improves the accuracy of the control strategy but also effectively addresses the uncertainties in wind and solar power output and EV charging.

[0046] An analytical model for the curtailment of wind and solar power is established, and a "curtailment-charging" correlation penalty function is proposed and quantified in the objective function.

[0047] The artificial lemming algorithm struggles to achieve high-precision optimal solutions in local searches, easily getting trapped in local optima in the later stages, leading to decreased population diversity and premature convergence. Therefore, this invention proposes an improved artificial lemming optimization algorithm that introduces a sine-cosine optimization strategy during the exploration phase. This strategy dynamically adjusts the individual's search direction and step size through the alternating action of sine and cosine functions. Furthermore, a Cauchy inverse cumulative distribution operator is introduced to address predator avoidance behavior, helping individuals escape the current search area, explore new solution spaces, and enhance population activity.

[0048] The above embodiments are merely exemplary embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art can make various modifications or equivalent substitutions to the present invention within its scope and spirit, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of the present invention.

Claims

1. A power grid dispatching method considering the orderly charging of electric vehicles and the consumption of new energy sources, characterized in that, Includes the following steps: Step 1: Collect historical data on grid load, meteorological information, and electricity price information; collect real-time power output information of new energy sources, EV charging demand, and charging status at fixed intervals. Step 2: Input the data collected in Step 1 into the day-ahead scheduling model to construct the day-ahead total expected cost function F1 of the power grid within time T; Step 3: Construct constraints, obtain preliminary unit start-up and shutdown plans and EV charging day plans based on the constraints, and issue dispatch instructions according to the plans; Step 4: Based on the results obtained from the cost function F1 in Step 2 and Step 3, and according to different operating scenarios, construct the regulation mathematical model F2 based on the prediction results of wind and solar power generation and the changes in EV charging load; Step 5: The improved lemming algorithm is used to optimize the variables in the control mathematical model F2 in step 4, thereby minimizing the expected cost.

2. The power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources according to claim 1, characterized in that, In step 2, the day-ahead total expected cost function F1 of the power grid within time T is specifically as follows: in, t For a period of time; g For the first g One generator set; N G The number of generator sets; For the first g Start-up cost of a generator set; The downtime cost of the g-th generator unit; u g,t , v g,t Let ∈{0, 1} be a function, and let the generator set be... g During the period t Internal boot u g,t =1 Or shutdown v g,t =1, otherwise u g,t , v g,t All are 0; The marginal cost of generating electricity for unit g; P g,t For the unit g During the period t Those who contribute their efforts; C curt As a penalty factor for abandoning wind and solar power; , They are respectively t The amount of wind and solar power curtailed during specific periods; C EV,i,t For EV car users i During the period t The electricity price for charging within the country; For the first time planned i EV t Charging power during a given period; C grid,t For time period t Electricity purchased from the main grid; P grid,t The power purchased from the main grid within time period t.

3. The power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources according to claim 1, characterized in that, The constraints in step 3 include: power balance constraints within time period t, upper and lower limits of generator output constraints, generator ramping constraints, minimum start-stop time constraints, new energy power generation absorption constraints, EV charging constraints, user satisfaction constraints, fluctuation constraints of renewable energy output, and spatiotemporal constraints of EV charging load.

4. A power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources according to claim 3, characterized in that, The power balance constraint within time period t is specifically as follows: in, P load,t This represents the power of the grid's conventional load. P w,t The active power of wind power generation. P pv,t The active power of photovoltaic power generation P EV,i,t For the first i Taiwanese electric vehicles during the period t The charging power.

5. A power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources according to claim 3, characterized in that, The specific charging constraints for EV vehicles are as follows: No. i Taiwanese EV cars during the period t The charging demand must meet the following constraints: ， In the formula, Δ t This refers to the charging time and the amount of charge required by the EV. The specific constraints on user satisfaction are: User satisfaction must meet the following constraints: In the formula, For the user's preferred time period, Δ t i2 This refers to the charging period. β i The minimum satisfaction rate required by the user of the i-th EV vehicle.

6. A power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources according to claim 3, characterized in that, The volatility constraint of the renewable energy output is represented by the power flow of branch l, specifically: in, branch road nm Maximum allowable current Pn, Qn The first n Active and reactive power of each node N For the number of nodes, g nm , b nm These are the admittances of the branches, θ nm branch road nm The phase angle difference, I nm branch road nm The actual current value; The spatiotemporal constraints of the EV vehicle charging load are represented by the system frequency, specifically: Where, Δ f For frequency variation, Δ P This is due to a power deficit. K sys Δ is the system frequency regulation coefficient. f max This represents the maximum allowable frequency deviation.

7. A power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources according to claim 1, characterized in that, The specific steps for adjusting the mathematical model F2 in step 4 are as follows: in, E The predicted control costs for the first half hour of the day; In order to be in t Curtailment of renewable energy during a given period (including curtailment of wind power and photovoltaic power). For the first i EV t Actual charging power during the time period; For the first i The unit compensation cost for deviation in EV charging power; φ ( i Let be a 0-1 function, representing the user's willingness to charge. If the first digit is 0... i The value is 1 if the EV user agrees to charging, otherwise it is 0. For the first t The unit cost of emergency power purchase during specific time periods; For the first t Emergency power purchase capacity during specific time periods; in, In order to be in t Load demand power during the period This refers to renewable energy power that can be absorbed by the power grid. For conventional units g Minimum output, P re,t In order to be in t Forecasted output of new energy sources during the period N EV For EV quantity, N G This refers to the number of conventional generating units.

8. A power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources according to claim 1, characterized in that, When implementing the improved lemming algorithm, the variables in the mathematical model F2 are used as lemmings, and the specific implementation process is as follows: Step A: Initialize algorithm parameters, set population size N, and maximum number of iterations T. MAX The upper and lower bounds of a variable; Step B: Randomly initialize the population positions, traverse all lemming positions, evaluate the fitness value of each lemming's initial position, find the individual with the lowest cost in the population, and record it as the current best position. Step C: At the beginning of each iteration, update the energy factor E. The larger the value of E, the more lemmings tend to explore; the smaller the value of E, the more lemmings tend to develop. Step D: Determine whether to explore or develop based on the current iteration's E value. If E > 1, choose the exploration behavior; if E < 1, choose the development behavior. Update the lemming position and constrain its upper and lower limits according to the constraints in Step 3. Calculate the fitness value of the new position based on the current lemming position and the day-ahead expected cost function F1 of the power grid in Time T in Step 2. If it is higher than the current optimal fitness value, update the global optimum. Step E: Determine if the maximum number of iterations has been reached. If it has, exit the loop and output the optimal position and corresponding best fitness value obtained through optimization. Otherwise, return to step C to continue iteratively solving the problem.

9. A power grid dispatching method considering orderly charging of electric vehicles and consumption of new energy sources according to claim 8, characterized in that, The formula for updating the lemming position in step D is as follows: in, For the first Only lemmings in the first Position in the next iteration This is the optimal position for the lemming at present. It serves as a directional marker, used to change the individual's search direction; It exhibits Brownian motion and follows a standard normal distribution; r1 is a random number uniformly distributed in the interval [-1, 1], r6 and r8 are random numbers in the interval [0, 2π], r9 is a random number following an exponential distribution of (0, ∞), and r5 and r7 are random numbers. , The location of an individual randomly selected within a lemming population; The formula for updating the lemming position in step D is as follows: in, r is the spiral shape factor. 10 G is a random number in the range [0, 1], G represents the escape coefficient, Levy represents the Levy flight function, and b represents the compression factor.

Images