Electric vehicle and central air conditioning load coordinated control method and system

Abstract

The present application relates to the technical field of power system demand side management, discloses a method and system for coordinated regulation of electric vehicles and central air conditioning loads, which solves the problem of insufficient grid regulation capacity and intensified load peak-valley difference under the background of high proportion of renewable energy access. The method comprises the following steps: collecting and preprocessing data; constructing a space-time distribution model of the electric vehicle group and a thermal dynamic model of the central air conditioning; establishing a multi-objective optimization model, the objective function of which comprehensively weighs the minimization of the grid side peak-valley difference, the minimization of user side discomfort and the minimization of system operation cost, the constraint conditions include grid safety constraints, equipment operation constraints and user comfort constraints; solving the coordinated regulation strategy; executing the regulation strategy and based on the strategy execution effect and real-time feedback data, using an online learning algorithm to rollingly update the model parameters and weight coefficients. The present application excavates the space-time regulation potential of electric vehicles and central air conditioning as flexible loads, and realizes the coordinated optimization of load resources across regions and time periods.

Description

Technical Field

[0001] This invention relates to the field of demand-side management technology for power systems, specifically to a system and method for coordinated control of electric vehicle and central air conditioning loads. Background Technology

[0002] With the increasing penetration of renewable energy sources (such as wind and solar power) in the power grid, the intermittency and volatility of their output pose significant challenges to the stable operation of the grid. On the one hand, peak renewable energy output may not match peak load, leading to wind and solar curtailment; on the other hand, the peak-to-valley load difference intensifies, increasing the pressure on the power grid for peak regulation. Traditional solutions mainly rely on regulation on the generation side, but their regulation capacity is limited and costs are high. Electric vehicles (EVs) and central air conditioning (ACs), as two important types of flexible loads, have enormous potential for controllability. EV charging behavior is flexible in time and space, while central air conditioning can be temporarily adjusted without affecting user comfort by utilizing the building's thermal inertia. However, most existing studies consider the regulation of EVs or central air conditioning independently, neglecting their spatiotemporal coupling characteristics. For example, the mobility of EVs causes their charging demand to change dynamically across different areas, while the load of central air conditioning is closely related to building thermal dynamics and outdoor meteorological conditions. If this spatiotemporal coupling characteristic is not fully considered, the optimization effect of load regulation strategies will be limited, failing to fully realize the synergistic benefits of aggregated resources. Therefore, there is an urgent need for a coordinated control method that can effectively characterize the spatiotemporal coupling characteristics of electric vehicles and central air conditioning, in order to tap their cross-regional and cross-time period regulation potential and support the grid's efficient absorption of renewable energy. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a time-space coupling system and method for coordinated control of electric vehicle and central air conditioning loads. The aim is to reduce the peak-valley difference of the power grid, improve the economic efficiency of operation, and ensure user comfort by coordinating and optimizing the time-space regulation capabilities of the two types of loads.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] A method for coordinated control of electric vehicle and central air conditioning load includes the following steps:

[0006] Data collection and preprocessing include electric vehicle operating status data, central air conditioning operating parameters, real-time power grid operating data, power grid forecast data, and meteorological data.

[0007] A spatiotemporal distribution model of electric vehicle groups and a thermal dynamic model of central air conditioning were constructed respectively.

[0008] A multi-objective optimization model is established, with the objective function comprehensively balancing the minimization of peak-valley difference on the power grid side, the minimization of user discomfort, and the minimization of system operating costs. The constraints include power grid security constraints, equipment operation constraints, and central air conditioning user temperature comfort constraints.

[0009] The coordinated control strategy is solved by using a distributed optimization algorithm to solve the multi-objective optimization model, and outputs the optimal charging power curve of electric vehicles and the optimal temperature setpoint curve of each central air conditioner in each power distribution area during the future scheduling cycle.

[0010] The system implements control strategies and, based on the effectiveness of these strategies and real-time feedback data, uses online learning algorithms to continuously update the parameters in the spatiotemporal distribution model of the electric vehicle population and the thermal dynamics model of the central air conditioning system, as well as the dynamic weight coefficients in the multi-objective optimization model.

[0011] Furthermore, the preprocessed data includes imputing missing data, identifying and correcting abnormal data, and standardizing and normalizing multi-source heterogeneous data.

[0012] Furthermore, the operating status data of the electric vehicle includes the maximum charging power of the electric vehicle v. Battery capacity Real-time state of charge during time period t Expected departure time Desired state of charge upon departure ;

[0013] The operating parameters of the central air conditioning system include at least the indoor temperature of central air conditioning unit i. Set temperature Rated power Comfortable temperature range User preferred temperature ;

[0014] The real-time operation data of the power grid is the power limit value of distribution area l. ];

[0015] The power grid forecast data includes the base load forecast within distribution area l. Renewable energy output forecast ;

[0016] The meteorological data refers to the outdoor temperature data within the power distribution area during time period t. .

[0017] Furthermore, the spatiotemporal distribution model of the electric vehicle population is described by constructing a spatiotemporal schedulable capacity matrix, and is expressed by the following formula:

[0018]

[0019] in, Indicates in Time-of-use power distribution area The total dispatchable charging capacity of the domestic electric vehicle population; As an indicator function, when the electric vehicle exist The time period is within the power distribution area It is 1 when the device is in a rechargeable state, and 0 otherwise. For electric vehicles exist Real-time state of charge over a given period of time; This represents the battery's maximum state of charge. For electric vehicles Maximum charging power; For charging efficiency; The preset time step; M represents the power distribution area. The total number of electric vehicles subject to internal regulation;

[0020] The thermal dynamics model of the central air conditioning system is established based on the first-order equivalent thermal parameter model, and its discretized form is expressed by the following formula:

[0021]

[0022] in, For the first The indoor temperature of the building where the central air conditioning is located during time period t. For the first The indoor temperature of the building where the central air conditioning is located during the t+1 time period; and These are the building's heat capacity and thermal resistance, respectively. For air conditioning cooling / heating efficiency; This represents the electrical power consumed by the air conditioner during time period t. The power distribution area where the building is located The outdoor temperature during time period t; The indoor heat disturbance during time period t; For time step.

[0023] Furthermore, the objective function of the multi-objective optimization model is expressed by the following formula:

[0024]

[0025] in, For grid-side targets Dynamic weighting coefficients; For user-side discomfort target The dynamic weighting coefficients, For operating cost targets The dynamic weighting coefficients satisfy ;

[0026] Grid-side targets The square of the peak-to-valley difference of the total load of the distribution network power system during the dispatching cycle is expressed by the following formula:

[0027]

[0028] in, This represents the maximum total load within the distribution network power system over all time periods t. This represents the minimum total load within the distribution network power system over all time periods t.

[0029] User-side discomfort target The sum of discomfort experienced by all users can be expressed as follows:

[0030]

[0031] in, For central air conditioning User comfort sensitivity weight; Weighting of charging comfort sensitivity for electric vehicle (EV) users; The set temperature for the central air conditioning unit during time period t; For users of central air conditioning i, the preferred temperature; This represents the actual state of charge of the electric vehicle v. The desired SOC trajectory of electric vehicle v is based on the desired off-field state of charge of electric vehicle v. Linear interpolation generation;

[0032] Air conditioning comfort sensitivity weight The calculation formula is:

[0033]

[0034] in, The coefficient represents the building space type where the central air conditioning is located: 2.0 for enclosed spaces, 1.5 for open spaces, and 0.3 for narrow spaces. To control the building area for air conditioning; Used as the base area; The personnel density is 1.5 for full capacity, 1.0 for normal traffic flow, and 0.5 for low traffic flow.

[0035] Charging comfort sensitivity weight of electric vehicle v The calculation formula is:

[0036]

[0037] in, Battery health factor; Let V be the battery capacity of the vth electric vehicle; This serves as the baseline capacity for electric vehicles. As a time urgency factor, the closer to the expected departure time of electric vehicle v, the higher the urgency factor;

[0038] Operating cost target It can be expressed as follows:

[0039]

[0040] in, This is the summation of all time periods within the scheduling period; The purchase price of electricity from the power grid; The total power input from the grid during time period t; Price of electric vehicle charging service during time period t; The actual charging power of the vth electric vehicle during time period t; The electricity price for air conditioning during time period t; Let t be the operating power of the i-th central air conditioner during time period t.

[0041] Furthermore, the constraints include:

[0042] Power grid security constraints:

[0043]

[0044] in, Basic load forecasting; and Power limit for distribution area l; The actual charging power of the vth electric vehicle during time period t; Let be the operating power of the i-th air conditioner during time period t; Forecasting power output for renewable energy;

[0045] Electric vehicle constraints:

[0046] in, The vth electric vehicle during the time period The actual state of charge; and These are the upper and lower limits for battery safety. The target SOC required for electric vehicle v; The maximum allowed charging power for the vth electric vehicle. Let V be the battery capacity of the vth electric vehicle; For charging efficiency; No. The estimated departure time of the electric vehicles;

[0047] Central air conditioning constraints:

[0048] in, The operating power of the i-th central air conditioner during time period t; Let be the rated power of the i-th central air conditioner; and For comfortable temperature range; For the first The indoor temperature of the building where the central air conditioning is located during time period t;

[0049] Temperature comfort constraints for central air conditioning users:

[0050]

[0051] in, The maximum unidirectional temperature difference that a central air conditioning user is allowed the system to adjust based on their preferred temperature. Set the temperature for the central air conditioning unit; Set the user's preferred temperature.

[0052] Furthermore, the dynamic weighting coefficient The values ​​are 0.5, 0.3, and 0.2 respectively.

[0053] Furthermore, the distributed optimization algorithm employs the alternating direction multiplier method to decompose the global optimization problem into an electric vehicle optimization subproblem and a central air conditioning optimization subproblem, and solves iteratively using coordination variables. The augmented Lagrangian function of the global optimization problem is:

[0054]

[0055]

[0056]

[0057]

[0058] in, For electric vehicle decision variables, it includes the charging power of all electric vehicles at each time period. ; For central air conditioning decision variables, including the operating power of all air conditioners at each time period. ; As dual variables; This is a power aggregation matrix for electric vehicles. This is a power aggregation matrix for central air conditioning systems. The target vector for power grid supply; The penalty coefficient is k; the iteration number is k.

[0059] The objective function for the electric vehicle optimization subproblem, used to describe the economy and user comfort of charging power allocation among electric vehicle groups, is specifically defined as follows:

[0060]

[0061] in, , The optimal power supply reference target for time period t issued by the power grid dispatch center;

[0062] The objective function for the air conditioning optimization subproblem, used to describe the temperature control and energy consumption balance of the central air conditioning system group, is specifically defined as follows:

[0063]

[0064] in, ;

[0065] Through constraints Establish the power coupling relationship between electric vehicles and central air conditioning to achieve total load balance in the power distribution network; in each iteration, solve the subproblems of electric vehicles and air conditioning separately, and update the dual variables. Continue until the convergence condition is met;

[0066] When both the original residual and the dual residual satisfy the convergence condition, the system obtains a unique optimal solution:

[0067]

[0068] This is the optimal charging power curve for all electric vehicles in the next 24 hours. The optimal operating power / temperature setting curve for the central air conditioning system during the entire 24-hour period.

[0069] Furthermore, the online learning algorithm employs a recursive least squares method with a forgetting factor to update the dynamic parameter vector of the i-th central air conditioner. The update formula is expressed as follows:

[0070]

[0071]

[0072]

[0073] in, For the first Air conditioner Estimated parameter values ​​for the time period, including the reciprocal of thermal resistance. and energy efficiency coefficient ; The observed value of temperature change, i.e. ; The regression vector contains the outdoor-indoor temperature difference. With air conditioner power ; The forgetting factor is 0.95~0.99. Kalman gain; Let be the covariance matrix, representing the uncertainty of parameter estimation, and set to the identity matrix as the initial value.

[0074] A system for implementing the aforementioned method of coordinated control of electric vehicle and central air conditioning load includes:

[0075] Data acquisition and communication module: used to acquire electric vehicle operating status data, central air conditioning operating parameters, meteorological data, real-time power grid operating data and renewable energy output forecast data from smart meters, charging piles, building automation systems, meteorological stations and power grid dispatch centers, and to realize information exchange between systems;

[0076] Load modeling and prediction module: Constructs a spatiotemporal distribution model of electric vehicle groups and a thermal dynamic model of central air conditioning, respectively;

[0077] The collaborative optimization calculation module is used to establish a multi-objective optimization model. The objective function comprehensively balances minimizing the peak-valley difference on the power grid side, minimizing user discomfort on the user side, and minimizing system operating costs. The constraints include power grid security constraints, equipment operation constraints, and user comfort constraints. It also solves for the optimal collaborative control strategy.

[0078] Strategy Execution and Control Module: This module receives the optimal control strategy output by the collaborative optimization calculation module and sends control commands, including electric vehicle charging power setpoints and central air conditioning temperature setpoints, to the terminal equipment in each power distribution area to achieve strategy execution and closed-loop control.

[0079] Performance evaluation and learning module: used to monitor the control effect, execute control strategies, and based on the strategy execution effect and real-time feedback data, use online learning algorithms to continuously update the parameters in the spatiotemporal distribution model of electric vehicle groups and the thermal dynamic model of central air conditioning, as well as the dynamic weight coefficients in the multi-objective optimization model.

[0080] The present invention has the following beneficial effects:

[0081] Unlike traditional methods for controlling the load of electric vehicles and central air conditioning, this invention precisely quantifies the controllable potential of EV and AC loads in time and space dimensions through a "spatiotemporal schedulable capacity matrix" and a "thermal dynamic model." It comprehensively considers three major objectives: power grid security, user comfort, and economy. Furthermore, it flexibly adjusts the optimization guidance through dynamic weight coefficients and employs distributed algorithms such as ADMM to effectively solve the "curse of dimensionality" problem caused by large-scale load aggregation, ensuring computational efficiency and privacy protection. The introduction of an online learning mechanism enables the rolling updates of model parameters and weights based on actual operating data, ensuring that the control strategy always matches the actual state of the system and improving the robustness and long-term effectiveness of the system. Attached Figure Description

[0082] Figure 1 This is a flowchart of the method for coordinated control of electric vehicle and central air conditioning load according to an embodiment of the present invention;

[0083] Figure 2 This is a schematic diagram of the solution process using the alternating direction multiplier method in this embodiment of the invention. Detailed Implementation

[0084] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. These embodiments are implemented based on the technical solution of the present invention, and provide detailed implementation methods and processes. However, the scope of protection of the present invention is not limited to the following embodiments.

[0085] A method for coordinated control of electric vehicle and central air conditioning load includes the following steps:

[0086] S1: Data collection and preprocessing, including electric vehicle operating status data, central air conditioning operating parameters, real-time power grid operating data, power grid forecast data, and meteorological data;

[0087] S2: Construct a spatiotemporal distribution model of electric vehicle groups and a thermal dynamic model of central air conditioning, respectively;

[0088] S3: Establish a multi-objective optimization model. The objective function comprehensively balances minimizing the peak-valley difference on the power grid side, minimizing user discomfort on the user side, and minimizing system operating costs. The constraints include power grid security constraints, equipment operation constraints, and user comfort constraints.

[0089] S4: Solving the coordinated control strategy, using a distributed optimization algorithm to solve the multi-objective optimization model, outputting the optimal charging power curve of electric vehicles in each power distribution area and the optimal temperature setpoint curve of each central air conditioner in the future scheduling cycle;

[0090] S5: Execute the control strategy, and based on the strategy execution effect and real-time feedback data, use online learning algorithms to continuously update the parameters in the spatiotemporal distribution model of electric vehicle groups and the thermal dynamic model of central air conditioning, as well as the dynamic weight coefficients in the multi-objective optimization model.

[0091] Specifically, in S1, data preprocessing includes imputing missing data, identifying and correcting outlier data, and standardizing and normalizing multi-source heterogeneous data.

[0092] The operating status data of electric vehicles includes the maximum charging power of electric vehicle v. Battery capacity Real-time state of charge during time period t Expected departure time Desired state of charge upon departure ;

[0093] The operating parameters of a central air conditioning system should include at least the indoor temperature of the central air conditioning unit. Set temperature Rated power Comfortable temperature range User preferred temperature ;

[0094] Real-time power grid operation data is the power limit value for distribution area l. ];

[0095] Grid forecast data includes base load forecasts within distribution area l. Renewable energy output forecast ;

[0096] The meteorological data refers to the outdoor temperature data within the power distribution area during time period t. .

[0097] For example, in step S1, five areas (l=1...5) under the jurisdiction of a city's 10kV distribution network are selected as the control targets. The dispatch window is for the next 24 hours, the time resolution is set to Δt=15min, and there are a total of 96 time periods. The monitoring targets include 2,000 electric vehicles (v=1…2000) and 1,000 central air conditioning units (i=1…1000).

[0098] Specifically, in S2, the spatiotemporal distribution model of the electric vehicle population is described by constructing a spatiotemporal schedulable capacity matrix, and is expressed by the following formula:

[0099]

[0100] in, Indicates the time period Power distribution area The total dispatchable charging capacity of the domestic electric vehicle fleet Number the time period; As an indicator function, when the electric vehicle During the period Located in the power distribution area It is 1 when the device is in a rechargeable state, and 0 otherwise. For electric vehicles During the period Real-time state of charge; This represents the battery's maximum state of charge. For electric vehicles Maximum charging power; For charging efficiency, 15 minutes is preferred under daily rolling optimization scheduling, and 5 minutes is preferred under extreme weather conditions. M represents the preset time step; M represents the total number of electric vehicles participating in the control within the power distribution area l.

[0101] In this embodiment, region 3 is designated as a vehicle with the number […]. Taking electric vehicles as an example, their parameters are as follows: , , , , ;

[0102] Substitute into the schedulable capacity matrix model:

[0103]

[0104] This indicates that the electric vehicle can provide approximately 0.964 kWh of charging regulation capacity to the control system during the current period.

[0105] Area 3 has a total of For each adjustable electric vehicle, the system calculates the parameters for each vehicle according to the formula described above. And sum them:

[0106]

[0107] In the data collected in this embodiment, the SOC, charging power, availability, and other parameters of electric vehicles in the power distribution area vary. The system calculates these parameters one by one to obtain the following results:

[0108]

[0109] This indicates that distribution area 3 is in the time period The load adjustment potential of electric vehicles is approximately .

[0110] Specifically, in S2, the thermal dynamic model of the central air conditioning system is established based on the first-order equivalent thermal parameter model, and its discretized form is expressed by the following formula:

[0111]

[0112] in, For the first The indoor temperature of the building where the central air conditioning is located during time period t. For the first The indoor temperature of the building where the central air conditioning is located during the t+1 time period; and These are the building's heat capacity and thermal resistance, respectively. For air conditioning cooling / heating efficiency; This represents the electrical power consumed by the air conditioner during time period t. The power distribution area where the building is located The outdoor temperature during time period t; The indoor heat disturbance during time period t; Let t be the time step, and t+1 represent consecutive time steps.

[0113] For example, taking a single central air conditioner (number 66) located in a commercial building within power distribution zone 3 as an example, its parameters are as follows: building heat capacity... Building thermal resistance Indoor temperature Outdoor temperature Air conditioner power Refrigeration efficiency Internal disturbances , ;

[0114] Substitute into the thermal dynamics model of the central air conditioning system:

[0115]

[0116] This indicates that the temperature change is gradual during this period, and the model can accurately predict the thermal inertial response.

[0117] Specifically, in S3, the objective function of the multi-objective optimization model is expressed as follows:

[0118]

[0119] in, For grid-side targets Dynamic weighting coefficients; For user-side discomfort target The dynamic weighting coefficients, For operating cost targets The dynamic weighting coefficients satisfy .

[0120] Preferred dynamic weighting coefficients The values ​​are 0.5, 0.3, and 0.2 respectively.

[0121] To prioritize the control of peak-valley differences and over-limit risks, the dynamic weighting coefficient α is set to 0.5. At the same time, user comfort (β=0.3) and economy (γ=0.2) are also taken into account. The ratio of the three factors is consistent with the actual focus of demand-side management, which can ensure the stability and interpretability of optimization.

[0122] Grid-side targets The square of the peak-to-valley difference of the total load of the distribution network power system during the dispatching cycle is expressed by the following formula:

[0123]

[0124] in, This represents the maximum total load within the distribution network power system over all time periods t. This represents the minimum total load within the distribution network power system over all time periods t.

[0125] For example, taking typical weekday data as an example, peak power Valley power .

[0126] Substitute into the formula for the square of the peak-to-valley difference:

[0127]

[0128] Specifically, in S3, the user-side discomfort target The sum of discomfort experienced by all users can be expressed as follows:

[0129]

[0130] in, For central air conditioning User comfort sensitivity weight; Weighting of charging comfort sensitivity for electric vehicle (EV) users; The set temperature for the central air conditioning unit during time period t; For users of central air conditioning i, the preferred temperature; This represents the actual state of charge of the electric vehicle v. The desired SOC trajectory of electric vehicle v is based on the desired off-field state of charge of electric vehicle v. Linear interpolation generation;

[0131] Air conditioning comfort sensitivity weight The calculation formula is:

[0132]

[0133] in, The coefficient represents the building space type where the central air conditioning is located: 2.0 for enclosed spaces, 1.5 for open spaces, and 0.3 for narrow spaces. To control the building area for air conditioning; Used as the base area; The personnel density is 1.5 for full capacity, 1.0 for normal flow, and 0.5 for low flow.

[0134] For example, taking an office area in a commercial building located within power distribution zone 3 as an example, select the first day of a typical office day. The time period (i.e., 10:00-10:15 on the same day) is used as an example time period. This period has passed the morning rush hour, with almost all personnel present and the building entering a stable energy consumption phase. While the personnel density is high, it does not fluctuate drastically, making it suitable as a benchmark for calculating the comfort weight of air conditioning. One typical air conditioner (denoted as number...) is used... For example, its parameter is: air conditioner type coefficient. (Enclosed office space), controlled area , benchmark area population density Substituting the values, we can obtain the time period for this air conditioner. Comfort sensitivity weight:

[0135]

[0136] In actual operation, the 100 central air conditioning units in distribution area 3 are not entirely the same in terms of space type, controlled area, and personnel density. Each air conditioning unit... Given monitoring and configuration data respectively, the system calculates the results for each unit. And combined with their respective temperature deviation terms Accumulate.

[0137] After monitoring all 100 air conditioners in distribution area 3 during the specified time periods The user discomfort was calculated and summed one by one to obtain the air conditioning comfort target for this power distribution area during this time period:

[0138]

[0139] This indicates that the overall discomfort level of air conditioning load in distribution area 3 during the period from 10:00 to 10:15 is approximately... (Dimensionless) provides a quantitative evaluation basis for subsequent collaborative optimization.

[0140] Charging comfort sensitivity weight of electric vehicle v The calculation formula is:

[0141]

[0142] in, Battery health factor; Let v be the battery capacity of the vth vehicle; This is the baseline capacity for electric vehicles, typically 60 kWh. As a time urgency factor, the closer to the expected departure time of the electric vehicle v, the higher the urgency factor.

[0143] For example, consider a typical electric vehicle (denoted as number 3) in the parking lot of distribution area 3. Taking the example of an air conditioner as an example, let's select the first one... 40 time slots ( (This is used as the calculation period.) The parameters of the electric vehicle are as follows:

[0144] Battery capacity Battery health Current distance from the estimated departure time of the electric vehicle Substituting the values, we can obtain the electric vehicle's time period. Charging comfort sensitivity weight:

[0145]

[0146] Further calculate its charging deviation term. It is known that the electric vehicle operates during the time period... State of charge

[0147]

[0148] Its expected SOC trajectory is determined by the user requirements of electric vehicles. The result obtained through linear interpolation is:

[0149]

[0150] The contribution of a single vehicle to the comfort deviation is:

[0151]

[0152] In actual operation, the 400 electric vehicles within distribution area 3 exhibit significant differences in battery capacity, SOC, estimated departure time, and battery health. All data are provided by real-time monitoring data, and the system calculates the results for each vehicle according to the above formula. And sum them one by one, taking into account their SOC deviation terms.

[0153] In this embodiment, the data collected covers all 400 electric vehicles during the specified time period. By calculating and summing the user discomforts one by one, the comfort target for electric vehicles is obtained as follows:

[0154]

[0155] This value represents the overall user discomfort level caused by the deviation between the current SOC and the expected SOC for all electric vehicles during the period of 10:00-10:15. (Dimensionless) provides a quantitative basis for collaborative optimization on the electric vehicle side.

[0156] In S3, operating cost target It can be expressed as follows:

[0157]

[0158] in, This is the summation of all time periods within the scheduling period; The purchase price of electricity from the power grid; Let t be the total power input from the grid during time period t; Price of electric vehicle charging service during time period t; The actual charging power of the vth electric vehicle during time period t; The electricity price for air conditioning during time period t; Let t be the operating power of the i-th central air conditioner during time period t.

[0159] This is the summation of all time periods within a scheduling cycle, such as 24 hours, with a time period length of... A total of 96 time slots; The electricity purchase price is from the power grid, and the data source is the power company's time-of-use electricity price list; For time period Total power input from the grid; For time period Electric vehicle charging service prices are based on data from charging station operators. For time period No. The actual charging power of a vehicle; For time period Electricity price for air conditioning; For time period No. The operating power of the air conditioner.

[0160] For example, the first power distribution area 3 is selected. The time period (10:00-10:15) is used as an example period. This period coincides with the stable operation range of the commercial office area, with a typical power consumption structure and stable load changes, making it suitable as a sample for operating cost analysis.

[0161] Based on data provided by the dispatch center and operators, the price and load data for the example time period are as follows:

[0162] Grid-side time-of-use pricing:

[0163]

[0164] Power purchased by the power grid (summary calculation results for the distribution area):

[0165]

[0166] Electric vehicle charging service price (operator's price for the day):

[0167]

[0168] The total charging power of 400 electric vehicles within distribution area 3, after individual vehicle-by-vehicle statistics:

[0169]

[0170] Electricity price for air conditioning (commercial building electricity price):

[0171]

[0172] Total operating power of 100 central air conditioners in distribution area 3 during time period 40:

[0173]

[0174] Substitute the above data into the objective function for operating costs:

[0175]

[0176] Calculate item by item:

[0177]

[0178] Therefore, the operating cost for this period is:

[0179]

[0180] That is, the total operating cost of distribution area 3 during the period from 10:00 to 10:15, after considering the purchase of electricity from the power grid, charging of electric vehicles and electricity consumption of central air conditioning, is about RMB 3.363, which provides a quantitative basis for the economic evaluation in multi-objective optimization.

[0181] In S3, the constraints include:

[0182] Power grid security constraints:

[0183]

[0184] in, Basic load forecasting; and Power limit for distribution area l; The actual charging power of the vth electric vehicle during time period t; Let be the operating power of the i-th air conditioner during time period t; Forecasting of renewable energy output.

[0185] For example, the data for distribution area 3 during the time period t = 40 (10:00–10:15) is as follows: basic load forecast. =9.5MW, renewable energy forecast =1.8MW, total power of electric vehicles =2.4MW, total power of central air conditioning =3.0MW, permissible operating range

[0186] Substitute power grid security constraints:

[0187]

[0188] get:

[0189]

[0190] The results show that the load during the current scheduling period is within the safe range.

[0191] Electric vehicle constraints:

[0192] in, The vth electric vehicle during the time period The actual state of charge; and These are the upper and lower limits for battery safety. The target SOC required for electric vehicle v; The maximum allowed charging power for the vth electric vehicle. Let V be the battery capacity of the vth electric vehicle; For charging efficiency; No. The estimated departure time of the electric vehicles.

[0193] For example, a typical electric vehicle in distribution area 3 For example, the parameters of this electric vehicle are as follows: SOC=0.55 in the previous period, battery capacity... =60kWh, maximum allowable power =7kW, target SOC =0.8, charging efficiency Time step =0.25h;

[0194] If charging is performed at the maximum power of 7.0 kW during this period, then:

[0195]

[0196] Estimated remaining time to target SOC:

[0197]

[0198] This indicates that, provided the upper limit power constraint is met, the electric vehicle can reach the user's expected SOC before the expected departure.

[0199] Equipment operation constraints include electric vehicle constraints and central air conditioning constraints.

[0200] Central air conditioning constraints:

[0201] in, The operating power of the i-th air conditioner during time period t; Let be the rated power of the i-th air conditioner; and For comfortable temperature range.

[0202] For example, consider a central air conditioner in power distribution area 3. For example, rated power Current operating power =28kW, Comfort Temperature Range If the current room temperature Equipment operating power Then the following conditions are met:

[0203]

[0204]

[0205] It meets temperature and power constraints.

[0206] User temperature comfort constraints:

[0207] Temperature comfort constraints for central air conditioning users:

[0208]

[0209] in, The maximum unidirectional temperature difference that a central air conditioning user is allowed the system to adjust based on their preferred temperature is generally set to 2℃. Set the temperature for the central air conditioning unit; Set the user's preferred temperature.

[0210] User comfort constraints:

[0211]

[0212] in, To allow for temperature differences; To set the temperature; Set the user's preferred temperature.

[0213] For example, consider a central air conditioner in power distribution area 3. For example, user preferences Optimized temperature setting Permissible temperature difference ;

[0214]

[0215] This indicates that the adjustment range is within an acceptable comfort zone.

[0216] Specifically, in step S4, the distributed optimization algorithm uses the alternating direction multiplier method to decompose the global optimization problem into an electric vehicle optimization subproblem and a central air conditioning optimization subproblem, and iteratively solves the problem using coordination variables. The solution process is as follows: Figure 2 As shown, the original global optimization problem is first decomposed into an electric vehicle subproblem and a central air conditioning subproblem based on the augmented Lagrangian function. A coordination variable is set to decouple the two types of decisions. In each iteration, the air conditioning variable z and its dual variable y are fixed, and the electric vehicle subproblem is solved to obtain a new x. Then, the updated x and y are fixed, and the air conditioning subproblem is solved to obtain a new z. Subsequently, the dual variable y is updated based on the current power matching residual to strengthen the constraint on global consistency. These three steps are executed iteratively until both the original residual and the dual residual satisfy the convergence condition.

[0217] In S4, the distributed optimization algorithm employs the alternating direction multiplier method to decompose the global optimization problem into electric vehicle optimization subproblems and central air conditioning optimization subproblems, and solves them iteratively using coordination variables. The augmented Lagrangian function of the global optimization problem is:

[0218]

[0219]

[0220]

[0221]

[0222] in, For electric vehicle decision variables, it includes the charging power of all electric vehicles at each time period. ; For central air conditioning decision variables, including the operating power of all air conditioners at each time period. ; As dual variables; This is a power aggregation matrix for electric vehicles. This is a power aggregation matrix for central air conditioning systems. The target vector for power grid supply; The penalty coefficient is k; the iteration number is k.

[0223] The objective function for the electric vehicle optimization subproblem, used to describe the economy and user comfort of charging power allocation among electric vehicle groups, is specifically defined as follows:

[0224]

[0225] in, , This is the reference target for the optimal power supply during time period t, issued by the power grid dispatch center.

[0226] The objective function for the air conditioning optimization subproblem, used to describe the temperature control and energy consumption balance of the central air conditioning system group, is specifically defined as follows:

[0227]

[0228] in, ;

[0229] Through constraints Establish the power coupling relationship between electric vehicles and central air conditioning to achieve total load balance in the power distribution network; in each iteration, solve the subproblems of electric vehicles and air conditioning separately, and update the dual variables. Continue until the convergence condition is met;

[0230] Through continuous iteration, the power distribution of electric vehicles is optimized. and central air conditioning operating power and utilize dual variables By strengthening the power balance constraint, the system obtains a unique optimal solution when both the original residual and the dual residual simultaneously satisfy the convergence condition:

[0231]

[0232] This is the optimal charging power curve for all electric vehicles in the next 24 hours. The optimal operating power / temperature setting curve for the central air conditioning system during the entire 24-hour period.

[0233] For example, the power distribution area 3 is selected during the t=40 time period (10:00–10:15) on a typical office day as an example time period to illustrate the collaborative control solution process based on ADMM of the present invention. The basic parameters are as follows: number of electric vehicles: 400 vehicles, number of central air conditioners: 100 units.

[0234] Grid-side reference power (given by safety constraints and intraday dispatch):

[0235]

[0236] Aggregated power of electric vehicles within distribution area 3 (statistical value before regulation):

[0237]

[0238] Central air conditioning aggregate power within distribution area 3 (statistical value before control):

[0239]

[0240] The user discomfort index for this period has been calculated through step S3:

[0241]

[0242] The multi-objective weights are selected as follows:

[0243]

[0244] Time-of-use electricity pricing and service prices:

[0245]

[0246]

[0247]

[0248] Under the above configuration, power distribution area 3 during the time period The single-time values ​​of the EV sub-target and AC sub-target can be expressed as:

[0249]

[0250] In this embodiment, for ease of explanation, the power distribution area 3 is defined as the time period. The combined EV and AC power are considered as one-dimensional decision quantities, and are denoted as follows:

[0251]

[0252]

[0253] The target value for EV+AC flexible power during this period is set as follows:

[0254]

[0255] This target value is determined by the upper-level dispatcher based on the power grid operating status and base load. Renewable energy output The combined assessment of over-limit risks provides a "flexible output level" that represents the overall acceptable level of EV and AC for this period.

[0256] In this simplified scenario, it can be made This indicates that the combined power of EVs and AC directly participates in the flexible power balance constraint:

[0257]

[0258] Penalty parameter selection The initial value of the dual variable is set to .

[0259] At the initial moment (before optimization), the combined EV and AC power of distribution area 3 in time period 40 is:

[0260]

[0261]

[0262] At this point, the flexible power balance residual is:

[0263]

[0264] This indicates that under the current scheduling scheme, The total flexible power is 0.6MW lower than the target, and the gap needs to be made up by increasing EV or AC output, or a combination of both.

[0265] In each round of ADMM iteration, this invention updates the decision variables and dual variables of distribution area 3 and time period 40 in the following order:

[0266] Given the air conditioning power of the previous round With dual variables Under the premise of solving the EV subproblem:

[0267]

[0268] This step, while keeping the AC power constant, adjusts the EV charging power to achieve the target on the EV side. The penalty function combined with the flexible power balance constraint is optimal.

[0269] Updated EV aggregate power With dual variables Next, solve the AC subproblem:

[0270]

[0271] This step, under the condition of fixed EV power, adjusts the AC power to make the AC side target... The penalty function combined with the flexible power balance constraint is optimal.

[0272] Update the dual variables based on the current flexible power balance residuals:

[0273]

[0274] Under the premise of satisfying the upper and lower limits of the power of EV and AC respectively and the user comfort constraints, the distribution area 3 can be obtained through numerical solution. The ADMM iteration table for each time period is shown in Table 1 (simplified to a 3-round iteration diagram).

[0275] Table 1

[0276] Iteration round k Flexible power residual 0 (Initial) 2.40 3.00 -0.6 1 2.55 3.15 -0.3 2 2.70 3.30 0.00

[0277] While ensuring that the power constraints of a single electric vehicle and a single air conditioner, the SOC safety range, and the indoor temperature comfort range are all satisfied, the ADMM algorithm reduces the flexible power residual from -0.60MW to 0 through two rounds of iteration, ultimately obtaining:

[0278]

[0279] At this time, power distribution area 3 is in The combined EV and AC power during the time period satisfies:

[0280]

[0281] It also takes into account the goals of the power grid side. User discomfort target Operating cost target The trade-offs. Compared with the initial plan ( Compared to the previous version, the optimized EV charging power and AC operating power have achieved a synergistic increase, effectively compensating for the flexible power gap, while also improving the weighting. In a multi-target configuration, the sub-targets in a single time period are made possible. Overall, the cost is reduced, achieving the optimal balance between power grid security, user comfort, and economy.

[0282] By analyzing all time periods By repeating the ADMM collaborative solution process described above, this invention can obtain the optimal charging power curves for electric vehicles in distribution area 3 and even the entire distribution network within the entire 24-hour window. Curve of optimal temperature setting value for central air conditioning This provides a stable and executable control strategy for the strategy distribution and online learning updates in step S5.

[0283] Specifically, in S5, the online learning algorithm uses recursive least squares with a forgetting factor to update the dynamic parameter vector of the i-th central air conditioner. The update formula is expressed as follows:

[0284]

[0285]

[0286]

[0287] in, For the first Air conditioner Estimated parameter values ​​for the time period, including the reciprocal of thermal resistance. and energy efficiency coefficient ; The observed value of temperature change, i.e. ; The regression vector contains the outdoor-indoor temperature difference. With air conditioner power ; The forgetting factor is 0.95~0.99. Kalman gain; Let be the covariance matrix, representing the uncertainty of parameter estimation, and set to the identity matrix as the initial value.

[0288] For example, select a typical air conditioner (numbered) in power distribution area 3. ), its in The monitoring data for the time period (10:00-10:15) is as follows:

[0289] Indoor temperature:

[0290]

[0291] Outdoor temperature:

[0292]

[0293] Air conditioner operating power:

[0294]

[0295] Therefore, the regression vector is:

[0296]

[0297] Temperature change:

[0298]

[0299] Initialization parameters:

[0300]

[0301] Calculate the Kalman gain:

[0302]

[0303] Prediction error:

[0304]

[0305] Update parameters:

[0306]

[0307] Reciprocal estimation of building thermal resistance:

[0308]

[0309] This indicates that the building's thermal inertia is slightly lower than initially estimated.

[0310] Energy efficiency ratio parameters:

[0311]

[0312] An increase of approximately [amount] compared to the original value This indicates a slight improvement in cooling efficiency.

[0313] The updated parameter vector is The reciprocal of thermal resistance This indicates that the building's thermal inertia has been reduced, and the energy efficiency coefficient has improved. =0.0174 indicates that the cooling efficiency has improved by approximately 9% compared to the initial state. The parameter update cycle is consistent with the scheduling step size (Δt=15min), and adaptive modeling of different building thermal characteristics is achieved through rolling updates. The comprehensive results show that ADMM co-optimization rapidly reduced the power error in multiple power distribution areas from 7.3MW to 0.02MW, RLS online learning reduced the temperature prediction error of the thermal dynamics model from ±0.45°C to ±0.12°C, improved the overall system control stability by 17%, and shortened the regulation response delay by 12%.

[0314] A system for implementing the aforementioned method of coordinated control of electric vehicle and central air conditioning load includes:

[0315] Data acquisition and communication module: used to acquire electric vehicle operating status data, central air conditioning operating parameters, meteorological data, real-time power grid operating data and renewable energy output forecast data from smart meters, charging piles, building automation systems, meteorological stations and power grid dispatch centers, and to realize information exchange between systems;

[0316] Load modeling and prediction module: Constructs a spatiotemporal distribution model of electric vehicle groups and a thermal dynamic model of central air conditioning, respectively;

[0317] The collaborative optimization calculation module is used to establish a multi-objective optimization model. The objective function comprehensively balances minimizing the peak-valley difference on the power grid side, minimizing user discomfort on the user side, and minimizing system operating costs. The constraints include power grid security constraints, equipment operation constraints, and user comfort constraints. It also solves for the optimal collaborative control strategy.

[0318] Strategy Execution and Control Module: This module receives the optimal control strategy output by the collaborative optimization calculation module and sends control commands, including electric vehicle charging power setpoints and central air conditioning temperature setpoints, to the terminal equipment in each power distribution area to achieve strategy execution and closed-loop control.

[0319] Performance evaluation and learning module: used to monitor the effect of regulation, execute regulation strategies, and update model parameters and weight coefficients on a rolling basis using online learning algorithms based on the effect of strategy execution and real-time feedback data.

[0320] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for coordinated control of electric vehicle and central air conditioning load, characterized in that, Includes the following steps: Data collection and preprocessing include electric vehicle operating status data, central air conditioning operating parameters, real-time power grid operating data, power grid forecast data, and meteorological data. A spatiotemporal distribution model of electric vehicle groups and a thermal dynamic model of central air conditioning were constructed respectively. A multi-objective optimization model is established, with the objective function comprehensively balancing the minimization of peak-valley difference on the power grid side, the minimization of user discomfort, and the minimization of system operating costs. The constraints include power grid security constraints, equipment operation constraints, and central air conditioning user temperature comfort constraints. The coordinated control strategy is solved by using a distributed optimization algorithm to solve the multi-objective optimization model, and outputs the optimal charging power curve of electric vehicles and the optimal temperature setpoint curve of each central air conditioner in each power distribution area during the future scheduling cycle. Implement control strategies, and based on the strategy implementation effect and real-time feedback data, use online learning algorithms to continuously update the parameters in the spatiotemporal distribution model of electric vehicle groups and the thermal dynamic model of central air conditioning, as well as the dynamic weight coefficients in the multi-objective optimization model. The spatiotemporal distribution model of the electric vehicle group is described by constructing a spatiotemporal schedulable capacity matrix, and is expressed by the following formula: ； in, Indicates in Time-of-use power distribution area The total dispatchable charging capacity of the domestic electric vehicle population; As an indicator function, when the electric vehicle exist The time period is within the power distribution area It is 1 when the device is in a rechargeable state, and 0 otherwise. For electric vehicles exist Real-time state of charge over a given period of time; This represents the battery's maximum state of charge. For electric vehicles Maximum charging power; For charging efficiency; The preset time step; M represents the power distribution area. The total number of electric vehicles subject to internal regulation; The thermal dynamics model of the central air conditioning system is established based on the first-order equivalent thermal parameter model, and its discretized form is expressed by the following formula: ； in, For the first The indoor temperature of the building where the central air conditioning is located during time period t. For the first The indoor temperature of the building where the central air conditioning is located during the t+1 time period; and These are the building's heat capacity and thermal resistance, respectively. For air conditioning cooling / heating efficiency; This represents the electrical power consumed by the air conditioner during time period t. The power distribution area where the building is located The outdoor temperature during time period t; The indoor heat disturbance during time period t; For time step.

2. The method for coordinated control of electric vehicle and central air conditioning load according to claim 1, characterized in that, The preprocessed data includes imputing missing data, identifying and correcting abnormal data, and standardizing and normalizing multi-source heterogeneous data.

3. The method for coordinated control of electric vehicle and central air conditioning load according to claim 1, characterized in that, The operating status data of the electric vehicle includes the maximum charging power of electric vehicle v. Battery capacity Real-time state of charge during time period t Expected departure time Desired state of charge upon departure ; The operating parameters of the central air conditioning system include at least the indoor temperature of central air conditioning unit i. Set temperature Rated power Comfortable temperature range User preferred temperature ; The real-time operation data of the power grid is the power limit value of distribution area l. ]; The power grid forecast data includes the base load forecast within distribution area l. Renewable energy output forecast ; The meteorological data refers to the outdoor temperature data within the power distribution area during time period t. .

4. The method for coordinated control of electric vehicle and central air conditioning load according to claim 1, characterized in that, The objective function of the multi-objective optimization model is expressed by the following formula: ； in, For grid-side targets Dynamic weighting coefficients; For user-side discomfort target The dynamic weighting coefficients, For operating cost targets The dynamic weighting coefficients satisfy ; Grid-side targets The square of the peak-to-valley difference of the total load of the distribution network power system during the dispatching cycle is expressed by the following formula: ； in, This represents the maximum total load within the distribution network power system over all time periods t. This represents the minimum total load within the distribution network power system over all time periods t. User-side discomfort target The sum of discomfort experienced by all users can be expressed as follows: ； in, For central air conditioning User comfort sensitivity weight; Weighting of charging comfort sensitivity for electric vehicle (EV) users; The set temperature for the central air conditioning unit during time period t; For users of central air conditioning i, the preferred temperature; This represents the actual state of charge of the electric vehicle v. The desired SOC trajectory of electric vehicle v is based on the desired off-field state of charge of electric vehicle v. Linear interpolation generation; Air conditioning comfort sensitivity weight The calculation formula is: ； in, The coefficient represents the building space type where the central air conditioning is located: 2.0 for enclosed spaces, 1.5 for open spaces, and 0.3 for narrow spaces. To control the building area for air conditioning; Used as the base area; The personnel density is 1.5 for full capacity, 1.0 for normal traffic flow, and 0.5 for low traffic flow. Charging comfort sensitivity weight of electric vehicle v The calculation formula is: ； in, Battery health factor; Let V be the battery capacity of the vth electric vehicle; This serves as the baseline capacity for electric vehicles. As a time urgency factor, the closer to the expected departure time of electric vehicle v, the higher the urgency factor; Operating cost target It can be expressed as follows: ； in, This is the summation of all time periods within the scheduling period; The purchase price of electricity from the power grid; The total power input from the grid during time period t; Price of electric vehicle charging service during time period t; The actual charging power of the vth electric vehicle during time period t; The electricity price for air conditioning during time period t; Let t be the operating power of the i-th central air conditioner during time period t.

5. The method for coordinated control of electric vehicle and central air conditioning load according to claim 1, characterized in that, The constraints include: Power grid security constraints: ； in, Basic load forecasting; and Power limit for distribution area l; The actual charging power of the vth electric vehicle during time period t; Let be the operating power of the i-th air conditioner during time period t; Forecasting power output for renewable energy; Electric vehicle constraints: ; in, The vth electric vehicle during the time period The actual state of charge; and These are the upper and lower limits for battery safety. The target SOC required for electric vehicle v; The maximum allowed charging power for the vth electric vehicle. Let V be the battery capacity of the vth electric vehicle; For charging efficiency; No. The estimated departure time of the electric vehicles; Central air conditioning constraints: ; in, The operating power of the i-th central air conditioner during time period t; Let be the rated power of the i-th central air conditioner; and For comfortable temperature range; For the first The indoor temperature of the building where the central air conditioning is located during time period t; Temperature comfort constraints for central air conditioning users: ； in, The maximum unidirectional temperature difference that a central air conditioning user is allowed the system to adjust based on their preferred temperature. Set the temperature for the central air conditioning unit; Set the user's preferred temperature.

6. The method for coordinated control of electric vehicle and central air conditioning load according to claim 4, characterized in that, The dynamic weighting coefficient The values ​​are 0.5, 0.3, and 0.2 respectively.

7. The method for coordinated control of electric vehicle and central air conditioning load according to claim 4, characterized in that, The distributed optimization algorithm employs the alternating direction multiplier method to decompose the global optimization problem into electric vehicle optimization subproblems and central air conditioning optimization subproblems, and solves them iteratively using coordination variables. The augmented Lagrangian function of the global optimization problem is: ； ； ； ； in, For electric vehicle decision variables, it includes the charging power of all electric vehicles at each time period. ; For central air conditioning decision variables, including the operating power of all air conditioners at each time period. ; As dual variables; This is a power aggregation matrix for electric vehicles. This is a power aggregation matrix for central air conditioning systems. The target vector for power grid supply; The penalty coefficient is k; the iteration number is k. The objective function for the electric vehicle optimization subproblem, used to describe the economy and user comfort of charging power allocation among electric vehicle groups, is specifically defined as follows: ； in, , The optimal power supply reference target for time period t issued by the power grid dispatch center; The objective function for the air conditioning optimization subproblem, used to describe the temperature control and energy consumption balance of the central air conditioning system group, is specifically defined as follows: ； in, ; Through constraints Establish the power coupling relationship between electric vehicles and central air conditioning to achieve overall load balance in the power distribution network; in each iteration, solve the subproblems of electric vehicles and air conditioning separately, and update the dual variables. Continue until the convergence condition is met; When both the original residual and the dual residual satisfy the convergence condition, the system obtains a unique optimal solution: ； This is the optimal charging power curve for all electric vehicles in the next 24 hours. The optimal operating power / temperature setting curve for the central air conditioning system during the entire 24-hour period.

8. The method for coordinated control of electric vehicle and central air conditioning load according to claim 1, characterized in that, The online learning algorithm employs a recursive least squares method with a forgetting factor to update the dynamic parameter vector of the i-th central air conditioner. The update formula is expressed as follows: ； ； ； in, For the first Air conditioner Estimated parameter values ​​for the time period, including the reciprocal of thermal resistance. and energy efficiency coefficient ; The observed value of temperature change, i.e. ; The regression vector contains the outdoor-indoor temperature difference. With air conditioner power ; The forgetting factor is 0.95~0.

99. Kalman gain; Let be the covariance matrix, representing the uncertainty of parameter estimation, and set to the identity matrix as the initial value.

9. A system for implementing the method for coordinated control of electric vehicle and central air conditioning load as described in any one of claims 1-8, characterized in that, include: Data acquisition and communication module: used to acquire electric vehicle operating status data, central air conditioning operating parameters, meteorological data, real-time power grid operating data and renewable energy output forecast data from smart meters, charging piles, building automation systems, meteorological stations and power grid dispatch centers, and to realize information exchange between systems; Load modeling and prediction module: Constructs a spatiotemporal distribution model of electric vehicle groups and a thermal dynamic model of central air conditioning, respectively; The collaborative optimization calculation module is used to establish a multi-objective optimization model. The objective function comprehensively balances minimizing the peak-valley difference on the power grid side, minimizing user discomfort on the user side, and minimizing system operating costs. The constraints include power grid security constraints, equipment operation constraints, and user comfort constraints. It also solves for the optimal collaborative control strategy. Strategy Execution and Control Module: This module receives the optimal control strategy output by the collaborative optimization calculation module and sends control commands, including electric vehicle charging power setpoints and central air conditioning temperature setpoints, to the terminal equipment in each power distribution area to achieve strategy execution and closed-loop control. Performance evaluation and learning module: used to monitor the control effect, execute control strategies, and based on the strategy execution effect and real-time feedback data, use online learning algorithms to continuously update the parameters in the spatiotemporal distribution model of electric vehicle groups and the thermal dynamic model of central air conditioning, as well as the dynamic weight coefficients in the multi-objective optimization model.

Images