A micro-grid power tracking method based on cooperation of electric vehicles and temperature-controlled loads

Abstract

The application discloses a micro-grid power tracking method based on cooperation of electric vehicles and temperature control loads, and comprises the following steps: precise and efficient tracking is realized by constructing a hierarchical coordination control framework. Firstly, a price-power dynamic response model of electric vehicle clusters and a temperature-energy dynamic mathematical model of temperature control loads are established, and a geometric calculation method is adopted to aggregate the feasible operation domain of temperature control load clusters; taking the minimization of total control cost as an objective, an optimal incentive price sequence and an auxiliary control reference power sequence are solved through a main control model; aiming at the main control response error, a decentralized stochastic control algorithm is used to calculate the control signal and the equipment state switching probability, and temperature control loads are guided to compensate quickly. The method combines the advantages of the two types of resources, avoids the limitation of single control, does not need to collect private data of users, guarantees the comfort of users, takes into account the regulation capacity, response speed and economy, and effectively improves the power tracking accuracy and operation reliability of a micro-grid tie line.

Description

Technical Field

[0001] This invention relates to the field of microgrid power tracking technology, and in particular to a microgrid power tracking method based on the coordination of electric vehicles and temperature-controlled loads. Background Technology

[0002] Microgrids integrate a large number of heterogeneous distributed resources such as electric vehicles and temperature-controlled loads. Maintaining supply and demand balance and achieving accurate power point tracking (PPT) via tie lines are crucial for mitigating the impact on the main power grid. These distributed resources are characterized by their small scale, scattered distribution, diverse types, and uncertain behavior. How to efficiently manage these resources to provide reliable PPT services has become a core challenge in the field of microgrid control. Existing microgrid control methods are mainly divided into two categories: non-intrusive control and intrusive control, each targeting resources with different response characteristics. However, both methods have significant limitations and cannot individually meet the comprehensive requirements of microgrids for PPT in terms of response speed, regulation capacity, control accuracy, and economy. Therefore, there is an urgent need to construct a collaborative mechanism that integrates the advantages of both control methods.

[0003] The core shortcomings of existing technologies are mainly reflected in two aspects: First, the functional limitations of a single control method. Existing methods only use a single control mode to manage specific types of resources. Although price-based non-intrusive control can utilize the large-capacity regulation potential of resources such as electric vehicles, the response is lagging and highly uncertain due to user behavior. Direct control can achieve rapid response by leveraging temperature-controlled loads, but the model is complex, computationally burdensome, and there is a risk of user privacy leakage. It cannot simultaneously achieve large regulation capacity, high response speed, and high control accuracy. Second, there is a lack of a collaborative control mechanism. The complementary characteristics and interaction of the two types of control methods at different time scales are not considered. It is difficult to achieve the economically optimal power point tracking target of the microgrid while ensuring user privacy and user comfort. It cannot effectively resolve the contradiction between response performance and economy under a single control mode. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of existing technologies, this invention provides a microgrid power tracking method based on the coordination of electric vehicles and temperature-controlled loads.

[0005] The technical solution adopted in this invention is a microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads, comprising: Step S1: Construct a price-power dynamic response model for electric vehicle clusters within a microgrid. This model includes a response time constant, incentive price, steady-state target power determined by the price-power elasticity function, and additive power uncertainty terms, to characterize the dynamic response process of cluster aggregated power to changes in price signals. Step S2: Construct a temperature-energy dynamic mathematical model for a single temperature-controlled load, obtain a discrete-time linear model through equivalent transformation, and clarify the energy constraints and power adjustment constraints corresponding to temperature constraints and power constraints. Step S3: Based on the dynamic model of temperature-controlled load, the operating constraints of heterogeneous temperature-controlled loads in the microgrid are aggregated into a unified feasible operating domain using geometric calculation methods. The aggregated constraint boundary is solved by scaling and translating the basic homothetic polygon. Step S4: Construct a main control model based on price incentives, convert the electric vehicle dynamic response model into a discrete-time model, and minimize the total control cost. Combine incentive price constraints, electric vehicle instantaneous power constraints, and temperature control load aggregation response capability constraints to solve for the optimal incentive price sequence and auxiliary control reference power sequence. Step S5: Construct an auxiliary control model based on distributed stochastic control, calculate the control signal according to the auxiliary control reference power, solve the stochastic switching rate of the temperature-controlled load cluster through the temperature distribution probability density evolution equation, and determine the equipment state switching probability by combining the integral switching probability and the instantaneous switching probability and execute it. Step S6: Through the coordinated action of the electric vehicle cluster main control and the temperature-controlled load auxiliary control, microgrid tie-line power tracking is performed.

[0006] Furthermore, the electric vehicle cluster price-power dynamic response model in step S1 is as follows: , The price-power elasticity function is: , in, for The actual aggregate power of the electric vehicle cluster at any given time The discrete-time correlation coefficient. To incentivize prices, For the additive power uncertainty term, and These are the maximum discharge power and the maximum charging power, respectively. and These are the corresponding power thresholds. and These are the discharge and charge elasticity coefficients, respectively. and This represents the boundary of the elastic interval.

[0007] Furthermore, the temperature-energy dynamic mathematical model for a single temperature-controlled load in step S2 is as follows: , The discrete-time linear model is: , The energy constraint is: , in, for Constant indoor temperature Where R is the outdoor ambient temperature, R is the equivalent thermal resistance, and C is the equivalent heat capacity. Rated power, For cooling efficiency, For the switch state, The inertia coefficient, For time step, For energy state variables, For power adjustment variables, To input the correlation coefficient, It is half the temperature of the dead zone.

[0008] Furthermore, the feasible operating domain for temperature-controlled load cluster aggregation in step S3 is: , The approximate aggregation domain expression is: , The aggregation constraint boundary is: , in, For the number of temperature-controlled loads, For a single load feasible region, and Based on the basic polygon constraint matrix and vector, This is the total scaling factor. The total translation vector, To aggregate relevant parameters, These are the upper and lower limits for the polymerization power constraint. The upper and lower limits are defined as the energy constraints for polymerization. These are the parameters for the basic model.

[0009] Furthermore, the main control objective function in step S4 includes the electric vehicle incentive cost: , Temperature control load control cost: , in, At the starting time, To control the time domain, The cost coefficient for temperature control load control. To assist in power control, This represents the actual aggregate power of the electric vehicle cluster. To incentivize prices.

[0010] Furthermore, the random switching rate of the temperature-controlled load in step S5 is: , , The switching probability is:

[0011] , in, This is an intermediate variable for net switching flux. These are the average cooling rate and the heating rate, respectively. Net temperature change rate For the integral switching probability, This represents the instantaneous switching probability. Indoor temperature, For a moment.

[0012] Further, step S2 includes the following sub-steps: S21, based on the operating characteristics of the temperature-controlled load, establish a first-order thermodynamic model with the refrigeration equipment as the object. This model includes indoor temperature, outdoor ambient temperature, equivalent thermal resistance, equivalent heat capacity, rated power, refrigeration efficiency, and binary variable calibration parameters for switching state; S22, for the nonlinear characteristics of the thermodynamic model, through variable substitution and mathematical transformation, transform it into a discrete-time linear model that is convenient for aggregation calculation and stochastic control applications, clarifying the correspondence between the inertia coefficient, input gain coefficient parameters and original parameters in the model; S23, based on the temperature comfort requirements set by the user, determine the constraint range of indoor temperature, transform the temperature constraint into the constraint range of energy state variables, and at the same time, combine the rated power and reference power of the equipment to define the value boundary of the power adjustment variable; S24, introduce the energy state variable and power adjustment variable, rewrite the discrete-time linear model into an energy dynamic equation, forming a complete temperature-energy dynamic mathematical model, providing a foundation for subsequent aggregation analysis.

[0013] Further, step S3 includes the following sub-steps: S31, based on the discrete-time model obtained in step S2, a prediction time domain is set, and the energy dynamic equation of each temperature-controlled load in this time domain is represented in matrix form, clarifying the relationship between the coefficient matrix and the system parameters; S32, combining the power constraints and energy constraints of each temperature-controlled load, the feasible operating domain of a single load is represented as a polygon, and the constraint matrix and constraint vector corresponding to the polygon are determined; S33, a basic homothetic polygon is constructed, the parameters of which are the average values ​​of the corresponding parameters of all temperature-controlled loads. For each temperature-controlled load, the maximum scaling factor and translation vector are solved by linear programming, so that the scaled and translated basic polygon is completely included in the feasible domain of the load; S34, the scaling factors and translation vectors of all temperature-controlled loads are summarized, the total scaling factor and total translation vector are calculated, and the constraint matrix and vector of the basic homothetic polygon are combined to obtain the approximate aggregated feasible operating domain of the temperature-controlled load cluster, and the upper and lower limits of the aggregated power constraints and aggregated energy constraints of the cluster are determined.

[0014] Further, step S4 includes the following sub-steps: S41, at the set sampling time, convert the electric vehicle price-power dynamic response model constructed in step S1 into a discrete-time state equation, clarifying the expression form of each parameter in discrete time; S42, with the goal of minimizing the total control cost of the microgrid, construct an objective function including electric vehicle incentive cost, temperature control load control cost, and power tracking error penalty cost, and determine the optimization variables as incentive price and auxiliary control power; S43, set constraints, including the range constraint of incentive price, the upper and lower limits constraint of instantaneous electric vehicle power, and the response capability constraint that the auxiliary control power must be within the aggregate feasible operating domain calculated in step S3; S44, solve the optimization problem to obtain the optimal incentive price sequence and the auxiliary control reference power sequence, publish the optimal incentive price sequence as the main control signal, and use the auxiliary control reference power sequence as the tracking target of auxiliary control.

[0015] Further, step S5 includes the following sub-steps: S51, based on the steady-state operating characteristics of the temperature-controlled load, establish the correlation between average power consumption and the on / off duration and the off duration, define the control signal, and clarify the correspondence between the expected total power consumption of the temperature-controlled load cluster and the control signal; S52, based on the auxiliary control reference power determined in step S4 and the average power consumption of the temperature-controlled load cluster, calculate the control signal required at the current moment, providing a basis for subsequent switching rate calculation; S53, through the temperature distribution probability density evolution equation, combined with the average cooling rate, average heating rate, and net temperature change rate, solve the random switching rate of the temperature-controlled load from on to off and from off to on; S54, calculate the integral switching probability and the instantaneous switching probability respectively, determine the final switching probability of the temperature-controlled load in the current state based on the two types of switching probabilities, and after receiving the control signal, the equipment autonomously executes the state switching action according to the probability.

[0016] Beneficial Effects: This invention proposes a microgrid power point tracking method based on the collaboration between electric vehicles and temperature-controlled loads. By constructing a hierarchical coordinated control architecture for the collaboration between electric vehicles and temperature-controlled loads, and complementing price-incentivized main control with decentralized stochastic auxiliary control, it fully leverages the advantages of electric vehicles' large regulation capacity and low cost, and the fast response speed of temperature-controlled loads. This effectively solves the problem that a single control method cannot simultaneously achieve large regulation capacity, high response speed, and economy. By constructing a dynamic model that closely reflects actual response characteristics, it accurately characterizes the lag response and uncertainty of electric vehicles to price signals, overcoming the shortcomings of insufficient prediction accuracy in traditional static models. Through a decentralized stochastic control algorithm, it eliminates the need to collect user private data, guiding temperature-controlled loads to autonomously switch states only through broadcast control signals. This avoids the privacy leakage risks of centralized control and reduces the system's computational and communication burden. Simultaneously, by aggregating temperature-controlled load operating constraints through geometric calculation methods, it clarifies the boundary conditions of collaborative control. Combined with an optimization model aimed at minimizing cost, it achieves accurate tracking of microgrid tie-line power while ensuring user thermal comfort. This successfully resolves the contradiction between response performance, regulation capacity, and economy in existing technologies, improving the reliability and economy of microgrid operation. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the overall process of the method of the present invention. Figure 2 This is a flowchart of method step S2 of the present invention; Figure 3 This is a flowchart of method step S3 of the present invention; Figure 4 This is a flowchart of method step S4 of the present invention; Figure 5 This is a flowchart of step S5 of the method of the present invention. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] like Figure 1As shown, a microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads includes: Step S1, constructing a price-power dynamic response model of an electric vehicle cluster within the microgrid. This model includes a response time constant, an incentive price, a steady-state target power determined by a price-power elasticity function, and additive power uncertainties to characterize the dynamic response process of the cluster's aggregated power to price signal changes; Step S2, constructing a temperature-energy dynamic mathematical model of a single temperature-controlled load, obtaining a discrete-time linear model through equivalent transformation, and clarifying the energy constraints and power adjustment constraints corresponding to temperature and power constraints; Step S3, based on the temperature-controlled load dynamic model, using geometric calculation methods to aggregate the operating constraints of heterogeneous temperature-controlled loads within the microgrid into a unified feasible operating domain, through scaling of a basic homothetic polygon. Step S4: Construct a main control model based on price incentives, converting the electric vehicle dynamic response model into a discrete-time model. With the goal of minimizing the total control cost, combine incentive price constraints, electric vehicle instantaneous power constraints, and temperature-controlled load aggregation response capability constraints to solve for the optimal incentive price sequence and auxiliary control reference power sequence. Step S5: Construct an auxiliary control model based on decentralized stochastic control, calculate the control signal based on the auxiliary control reference power, solve for the stochastic switching rate of the temperature-controlled load cluster through the temperature distribution probability density evolution equation, and determine the equipment state switching probability by combining the integral switching probability and the instantaneous switching probability and execute it. Step S6: Perform microgrid tie-line power tracking through the synergistic effect of the electric vehicle cluster main control and the temperature-controlled load auxiliary control.

[0020] Step S1 involves constructing a price-power dynamic response model for electric vehicle clusters within the microgrid. This process requires comprehensive incorporation of key technical parameters to accurately characterize the dynamic response of cluster aggregated power to price signal changes. During implementation, the core elements of the model are first defined, including the cluster response time constant reflecting the speed of response, the incentive price used to guide user electricity consumption behavior, the steady-state target power determined by the price-power elasticity function, and the additive power uncertainty term reflecting the uncertainty of user response. The value range of the additive power uncertainty term is set between -0.1 and +0.1. The price-power elasticity function is approximated using a piecewise linear form. The discharge elasticity coefficient and charging elasticity coefficient need to be estimated using historical market prices and electric vehicle power data from the microgrid. Simultaneously, the maximum discharge power, maximum charging power, and corresponding price thresholds must be defined, and the power output levels corresponding to different price ranges must be divided. The model construction process must strictly adhere to the positive and negative power definition rules: power consumption is positive when the incentive price is positive, and negative when the incentive price is negative. During the implementation phase, parameter calibration is required based on historical operating data from at least 1,000 electric vehicles to ensure that the model can accurately reflect the actual response characteristics of a large-scale electric vehicle cluster. It abandons the idealized assumptions of the traditional static model regarding immediate response, fully considers response lag and randomness, and provides accurate load response prediction for the subsequent main control phase. Its core significance lies in providing a reliable foundation for subsequent price incentive control through accurate modeling, and ensuring the economy and predictability of the adjustment capacity in the main control phase.

[0021] The purpose of step S2 is to construct a temperature-energy dynamic mathematical model for a single temperature-controlled load and complete the equivalent transformation and constraint definition. The implementation process needs to focus on the physical parameters and operational constraints of the equipment. First, a first-order thermodynamic model is established for the temperature-controlled load, clarifying the key parameters involved in the model, including indoor temperature, outdoor ambient temperature, equivalent thermal resistance, equivalent heat capacity, rated power, cooling efficiency, and on / off state binary variables. The nominal values ​​of equivalent thermal resistance and equivalent heat capacity are both set to 2, the nominal value of cooling efficiency is 2.5, the nominal value of rated power is 0.5, half of the temperature dead zone is set to 1, and the temperature setpoint is 25. Due to the nonlinear characteristics of the original thermodynamic model, it needs to be transformed into a discrete-time linear model through variable substitution and mathematical transformation, with a time step set to 10 seconds, to calculate the inertia coefficient and input gain coefficient. Then, based on the user's temperature comfort requirements, the indoor temperature constraint range is defined within the range of the temperature setpoint plus or minus half of the temperature dead zone, which is then transformed into the constraint range of the energy state variable. The energy state variable is calculated using the equivalent heat capacity, temperature setpoint, actual indoor temperature, and cooling efficiency. Simultaneously, by combining the rated power of the equipment with the baseline power required to maintain the set temperature, the boundary values ​​of the power adjustment variable are defined. The baseline power is calculated and determined by the outdoor ambient temperature, the temperature setpoint, and the input gain coefficient. During implementation, equipment heterogeneity must be considered. The specific parameters of each temperature-controlled load are multiplied by a uniformly distributed random number between 0.9 and 1.1 based on their nominal values. The initial temperature is randomly initialized within the set comfort range. The significance of this step lies in providing accurate individual model support for subsequent cluster aggregation analysis and control strategy design, ensuring that the operating status of the temperature-controlled load can be accurately quantified and described.

[0022] Step S3 is implemented based on the dynamic model of temperature-controlled loads. A geometric calculation method is used to aggregate the operational constraints of heterogeneous temperature-controlled loads within the microgrid into a unified feasible operating domain. The key lies in solving the aggregated constraint boundary through scaling and translation of a basic homothetic polygon. In implementation, firstly, based on the discrete-time linear model obtained in step S2, the prediction time domain length is set, and the energy dynamic equation of each temperature-controlled load within this time domain is transformed into matrix form, clarifying the coefficient matrix composed of system parameters. Then, for each temperature-controlled load, combined with its power and energy constraints, the feasible operating domain of a single load is represented as a polygon, and the constraint matrix and constraint vector corresponding to this polygon are determined. Next, a basic homothetic polygon is constructed, with its parameters taken as the average of the parameters corresponding to all temperature-controlled loads. For at least 5000 heterogeneous temperature-controlled loads within the microgrid, the maximum scaling factor and translation vector are solved one by one through linear programming to ensure that the scaled and translated basic polygon is completely included within the feasible domain of that load. Finally, by summing the scaling factors and translation vectors of all temperature-controlled loads, the total scaling factor and total translation vector are calculated. Combined with the constraint matrix and vectors of the basic homothetic polygon, an approximate aggregated feasible operating domain for the temperature-controlled load cluster is constructed. Furthermore, the upper and lower limits of the cluster's aggregated power and aggregated energy constraints are determined and provided to the main control stage as constraints. The significance of this step lies in solving the collaborative control challenge of large-scale heterogeneous temperature-controlled loads through aggregated modeling, providing clear constraint boundaries for the subsequent coordination of main and auxiliary control, and ensuring the feasibility and safety of the control strategy.

[0023] The key focus of step S4 is to construct a price-incentive-based master control model and solve for the optimal control signal. The implementation process revolves around model transformation, objective function construction, constraint setting, and optimization. First, at a set sampling time, the electric vehicle dynamic response model constructed in step S1 is converted into a discrete-time model. The sampling interval remains consistent with the time step in step S2, both being 10 seconds, clarifying the expression of each parameter in discrete time. Then, the objective function is to minimize the total control cost of the microgrid. This objective function includes three cost components: electric vehicle incentive cost, temperature-controlled load control cost, and power tracking error penalty cost. The control cost coefficient for the temperature-controlled load is set to 0.4 yuan per kilowatt-hour, and the penalty price coefficient for power mismatch is set to 0.5 yuan per kilowatt-hour. The control time domain is set to 24 hours. Constraints include the range of incentive price values, upper and lower limits of the instantaneous power of the electric vehicle, and the response capability constraint that the auxiliary control power must be within the aggregate feasible operating domain calculated in step S3. In the optimization process, the optimization variables are the incentive price and the auxiliary control power. By solving this optimization problem, the optimal incentive price sequence and the auxiliary control reference power sequence are obtained simultaneously. During implementation, it is necessary to ensure the computational efficiency of the optimization algorithm to adapt to a 10-second control time granularity. The optimal incentive price sequence is immediately released to electric vehicle users as the main control signal, and the auxiliary control reference power sequence serves as the power tracking target for the next auxiliary control step. The core significance of this step lies in achieving efficient scheduling of large-scale electric vehicle resources through economic optimization, obtaining economical regulation capacity, and providing primary power support for microgrid power tracking.

[0024] Step S5 involves constructing an auxiliary control model based on distributed stochastic control. This model achieves rapid compensation of the temperature-controlled load through precise calculation and signal broadcasting. The implementation process consists of four key stages: control signal calculation, switching rate solution, switching probability determination, and action execution. First, based on the steady-state operating characteristics of the temperature-controlled load, the correlation between average power consumption and the on / off duration is established, and the control signal is defined, clarifying the correspondence between the expected total power consumption of the temperature-controlled load cluster and the control signal. Then, based on the auxiliary control reference power determined in step S4 and the average power consumption of the temperature-controlled load cluster, the required control signal for the current moment is calculated. Next, using the temperature distribution probability density evolution equation, combined with the average cooling rate, average heating rate, and net temperature change rate, the stochastic switching rates of the temperature-controlled load from on to off and from off to on are solved. The average cooling rate and average heating rate are calculated based on the thermodynamic parameters of the temperature-controlled load. Determining the switching probability requires calculating both the integral switching probability and the instantaneous switching probability. The integral switching probability is calculated by multiplying the average switching rate of adjacent time points by the time step, while the instantaneous switching probability is determined based on the ratio of the temperature change rate of adjacent time points. Combining these two types of switching probabilities yields the final state switching probability. In implementation, the control center only needs to broadcast the control signal to the 5000 temperature-controlled load clusters, without needing to collect real-time status data from each device. After receiving the signal, the devices autonomously execute switching actions based on their own status and the final switching probability. The control time granularity remains at 10 seconds to ensure rapid response to tracking errors in the main control. The significance of this step lies in achieving rapid power compensation without privacy risks, filling the power gap caused by the lag in the main control response, and ensuring power tracking accuracy.

[0025] Step S6 is implemented through the coordinated action of the electric vehicle cluster main control and the temperature-controlled load auxiliary control to ultimately achieve power point tracking (PPT) on the microgrid tie line. The implementation process relies on the models and control signals from previous steps, focusing on the coordination mechanism and performance assurance. During implementation, the time scale matching principle of the coordinated control is first clarified: both the main control and auxiliary control adopt a 10-second time granularity to ensure consistent action. The electric vehicle cluster, as the main control resource, adjusts its charging or discharging power according to the optimal incentive price sequence published in step S4, undertaking the majority of the microgrid's power regulation tasks. Its large-scale regulation capacity ensures the economy and basic regulation capability of PPT. The temperature-controlled load cluster, as the auxiliary control resource, autonomously adjusts its switching state according to the control signal broadcast in step S5 and its own determined switching probability. It provides real-time compensation for dynamic tracking errors caused by the lag in the main control response, using its rapid response characteristics to fill power gaps. During implementation, the deviation between the actual exchange power and the reference power on the tie line needs to be monitored in real time, and the incentive price sequence of the main control and the reference power sequence of the auxiliary control are dynamically adjusted to ensure the deviation is controlled within the range of -100 kW to +100 kW. Simultaneously, the temperature trajectory of temperature-controlled loads is continuously monitored to ensure that the temperature of all temperature-controlled loads remains within the user-defined comfort range, guaranteeing user thermal comfort. This step requires coverage of a 24-hour continuous operation cycle and must achieve stable tracking of both stepped reference power signals and actual grid regulation signals, keeping the root mean square power tracking error within 7.07%. Its core significance lies in leveraging the complementary advantages of these two types of resources to achieve accurate, rapid, and economical tracking of microgrid tie-line power while protecting user rights, thereby improving the stability and reliability of microgrid operation.

[0026] Preferably, the electric vehicle cluster price-power dynamic response model in step S1 is as follows: , The price-power elasticity function is: , in, for The actual aggregate power of the electric vehicle cluster at any given time The discrete-time correlation coefficient. To incentivize prices, For the additive power uncertainty term, and These are the maximum discharge power and the maximum charging power, respectively. and These are the corresponding power thresholds. and These are the discharge and charge elasticity coefficients, respectively. and This represents the boundary of the elastic interval.

[0027] Specifically, step S1 defines the dynamic response model and price-power elasticity function of the electric vehicle cluster, clarifying the model's structure and parameter definitions to provide technical support for accurately characterizing the dynamic response of the cluster's aggregated power to price signals. This dynamic response model includes core parameters such as discrete-time correlation coefficients, incentive prices, additive power uncertainty terms, and the actual aggregated power of the cluster. The additive power uncertainty term is set to a range of -0.1 to +0.1 to reflect the randomness of user-response power adjustments. The price-power elasticity function adopts a piecewise linear form, dividing five different price intervals corresponding to different power output levels. It clarifies the maximum discharge power and maximum charging power, along with their corresponding price thresholds. Simultaneously, it sets parameters such as the discharge elasticity coefficient, charging elasticity coefficient, and elasticity interval boundaries. These coefficients are estimated statistically from historical market prices and electric vehicle power data of the microgrid. During implementation, historical operating data from at least 1,000 electric vehicles must first be collected, including charging and discharging power records under different price signals. The elasticity coefficients and price thresholds are determined through data fitting. Then, the dynamic response model is converted into a discrete-time form using discrete-time correlation coefficients to ensure that the model accurately reflects the lag response process of the cluster's aggregated power after price signal changes. By clearly defining the model structure and parameters, and abandoning the idealized assumptions of traditional static models regarding immediate response, the accuracy of power prediction in the main control stage is improved. This provides a reliable model foundation for subsequent optimal incentive price solutions, ensuring the economy and predictability of the electric vehicle cluster's regulation capacity.

[0028] Preferably, the temperature-energy dynamic mathematical model for a single temperature-controlled load in step S2 is as follows: , The discrete-time linear model is: , The energy constraint is: , in, for Constant indoor temperature Where R is the outdoor ambient temperature, R is the equivalent thermal resistance, and C is the equivalent heat capacity. Rated power, For cooling efficiency, For the switch state, The inertia coefficient, For time step, For energy state variables, For power adjustment variables, To input the correlation coefficient, It is half the temperature of the dead zone.

[0029] Specifically, step S2 involves the model construction and constraint definition of a single temperature-controlled load, clarifying the specific parameters and relationships of the thermodynamic model, discrete-time linear model, and energy constraints. This provides a basis for accurate modeling and subsequent aggregate analysis of the temperature-controlled load. The thermodynamic model includes parameters such as indoor temperature, outdoor ambient temperature, equivalent thermal resistance, equivalent heat capacity, rated power, cooling efficiency, and a binary variable representing on / off state. The nominal values ​​for equivalent thermal resistance and equivalent heat capacity are 2℃ / kWh, 2kWh / ℃, 2.5, and 0.5kW respectively. The binary variable representing on / off state takes only two values: 0 or 1, corresponding to the device's off and on states, respectively. To address the nonlinearity of the original model, a discrete-time linear model is obtained through variable substitution and mathematical transformation. The time step is set to 10 seconds, and the inertia coefficient is calculated from the time step and the equivalent thermal resistance and equivalent heat capacity. The input correlation coefficient is related to the equivalent thermal resistance and cooling efficiency. Energy constraints are set based on user temperature comfort requirements, with half of the temperature dead zone defined as 1°C. Combining equivalent heat capacity and cooling efficiency, the indoor temperature constraint range is transformed into a constraint range for energy state variables. Simultaneously, the boundaries of power adjustment variables are defined by rated power and reference power. The reference power is calculated from the outdoor ambient temperature, temperature setpoint, and input correlation coefficients. During implementation, the physical parameters of individual temperature-controlled loads must first be measured or determined based on nominal values. Considering equipment heterogeneity, the specific parameters of each load are multiplied by a uniformly distributed random number between 0.9 and 1.1 from the nominal value. Then, a discrete-time linear model and various constraints are obtained through mathematical transformation, converting the nonlinear model into a form easy to calculate, clarifying the boundary conditions for load operation, and providing accurate individual model support for subsequent cluster aggregation and control strategy design.

[0030] Preferably, the feasible operating domain for temperature-controlled load cluster aggregation in step S3 is: , The approximate aggregation domain expression is: , The aggregation constraint boundary is: , in, For the number of temperature-controlled loads, For a single load feasible region, and Based on the basic polygon constraint matrix and vector, This is the total scaling factor. The total translation vector, To aggregate relevant parameters, These are the upper and lower limits for the polymerization power constraint. The upper and lower limits are defined as the energy constraints for polymerization. These are the parameters for the basic model.

[0031] Specifically, the method for constructing the aggregated feasible operating domain of the temperature-controlled load cluster in step S3 is defined, clarifying the mathematical expression of the aggregated domain and the calculation method of the aggregated constraint boundary, thus solving the problem of collaborative control modeling for large-scale heterogeneous temperature-controlled loads. This aggregated feasible operating domain is obtained through the Minkowski sum of the feasible domains of all individual temperature-controlled loads. The feasible domain of an individual load is represented as a polygon, including a constraint matrix and constraint vector, determined by the power and energy constraints of each load. In implementation, a basic homothetic polygon is first constructed, with its parameters taken as the average of the parameters corresponding to 5000 heterogeneous temperature-controlled loads within the microgrid. For each load, the maximum scaling factor and translation vector are solved using linear programming to ensure that the scaled and translated basic polygon is completely included within the feasible domain of that load. Then, the scaling factors and translation vectors of all loads are summarized to calculate the total scaling factor and total translation vector. Combining the constraint matrix and vector of the basic homothetic polygon, an approximate aggregated feasible operating domain for the cluster is constructed. Further derivation of the upper and lower limits of the aggregated power constraint, the upper and lower limits of the aggregated energy constraint, and related aggregated parameters is obtained. The calculation of the aggregated constraint boundary requires combining the coefficient matrix of the basic model with the input correlation coefficients. The implementation uses geometric calculation methods and linear programming algorithms to first represent the feasible region of a single load as a polygon, and then obtains a unified feasible region at the cluster level through scaling, translation and aggregation operations. The core significance is to transform the complex constraints of large-scale heterogeneous loads into unified boundary conditions, providing a clear basis for the temperature control load response capability constraints in the subsequent main control model, and ensuring the feasibility and safety of the collaborative control strategy.

[0032] Preferably, the main control objective function in step S4 includes the electric vehicle incentive cost: , Temperature control load control cost: , in, At the starting time, To control the time domain, The cost coefficient for temperature control load control. To assist in power control, This represents the actual aggregate power of the electric vehicle cluster. To incentivize prices.

[0033] Specifically, the core parameters of the main control objective function in step S4 define the calculation methods for electric vehicle incentive costs and temperature-controlled load control costs, providing a clear basis for optimization solutions aimed at minimizing total control costs. This objective function includes three cost components: the electric vehicle incentive cost is determined by the integral of the incentive price and the actual aggregated power of the cluster over the control time domain; the temperature-controlled load control cost is determined by the integral of the temperature-controlled load control cost coefficient and the auxiliary control power over the same control time domain. The control time domain is set to 24 hours, and the temperature-controlled load control cost coefficient is set to 0.4 yuan per kilowatt-hour. During implementation, the start time and the time range of the control time domain must be clearly defined first. Historical cost data of microgrid operation must be collected, including electric vehicle incentive price standards and temperature-controlled load operating costs, to determine the control cost coefficient and the penalty price coefficient for power mismatch. The penalty price coefficient is set to 0.5 yuan per kilowatt-hour. The incentive cost of electric vehicles, the control cost of temperature-controlled loads, and the penalty cost of power point tracking (PPT) error are integrated into the overall control cost objective function. The optimization variables are set as incentive price and auxiliary control power. A complete optimization problem is constructed by combining incentive price constraints, instantaneous power constraints of electric vehicles, and aggregated response capability constraints of temperature-controlled loads. This problem is solved using a numerical optimization algorithm to obtain the optimal incentive price sequence and the auxiliary control reference power sequence. The optimal incentive price sequence needs to be immediately released to electric vehicle users, and the auxiliary control reference power sequence serves as the tracking target for subsequent auxiliary control. By clearly defining the composition and parameters of the objective function, the trade-off between control cost and PPT accuracy is quantified, ensuring that the main control stage, while acquiring economical regulation capacity, provides a clear tracking target for auxiliary control, thus guaranteeing the economy and accuracy of microgrid power point tracking.

[0034] Preferably, the random switching rate of the temperature control load in step S5 is: , , The switching probability is:

[0035] , in, This is an intermediate variable for net switching flux. These are the average cooling rate and the heating rate, respectively. Net temperature change rate For the integral switching probability, This represents the instantaneous switching probability. Indoor temperature, For a moment.

[0036] Specifically, the calculation method for the random switching rate and switching probability of the temperature-controlled load in step S5 clarifies the definition and calculation logic of each parameter, providing technical support for the implementation of the distributed random control strategy. This includes two types of random switching rates, corresponding to the state transitions of the temperature-controlled load from off to on and from on to off, respectively. The calculation of the switching rate involves parameters such as the net switching throughput intermediate variable, the average cooling rate, the average heating rate, and the net temperature change rate. The average cooling rate and average heating rate are calculated based on physical parameters of the temperature-controlled load, such as its equivalent thermal resistance, equivalent heat capacity, rated power, and cooling efficiency. The net temperature change rate is determined by the correlation between these two parameters and the control signal. The switching probability is divided into integral switching probability and instantaneous switching probability. The integral switching probability is calculated by multiplying the average switching rate at adjacent moments by a 10-second time step. The instantaneous switching probability is determined based on the ratio of the temperature change rates at adjacent moments. The final state switching probability is obtained by multiplying the two types of switching probabilities. During implementation, the control center first calculates the control signal based on the auxiliary control reference power and the average power consumption of the temperature-controlled load cluster. Then, it calculates the average cooling rate, average heating rate, and net switching throughput intermediate variables by combining outdoor ambient temperature, indoor temperature, and equipment physical parameters. This allows for the determination of two types of switching rates. Subsequently, the integral switching probability and instantaneous switching probability are calculated using time steps and parameters at adjacent times. Finally, the control signal is broadcast to 5000 temperature-controlled loads. Upon receiving the signal, the equipment autonomously executes switching actions based on its current state and the final switching probability. By clearly defining the calculation logic for switching rates and probabilities, decentralized control is achieved without collecting user private data. This protects user privacy, reduces communication and computational burden, and ensures that the temperature-controlled load cluster can quickly respond to the auxiliary control reference power, compensating for the dynamic tracking error of the main control.

[0037] Preferred, such as Figure 2 As shown, step S2 includes the following sub-steps: S21. Based on the operating characteristics of temperature-controlled load, a first-order thermodynamic model is established with refrigeration equipment as the object. The model includes indoor temperature, outdoor ambient temperature, equivalent thermal resistance, equivalent heat capacity, rated power, refrigeration efficiency, and binary variable calibration parameters of switching state. S22, the nonlinear characteristics of the thermodynamic model are transformed into a discrete-time linear model that is easy to aggregate calculation and stochastic control application through variable substitution and mathematical transformation, clarifying the correspondence between the inertia coefficient, input gain coefficient and original parameters in the model; S23, based on the user's set temperature comfort requirements, determines the indoor temperature constraint range, transforms the temperature constraint into the energy state variable constraint range, and defines the value boundary of the power adjustment variable by combining the equipment's rated power and reference power. S24 introduces energy state variables and power adjustment variables, rewriting the discrete-time linear model into an energy dynamic equation, forming a complete temperature-energy dynamic mathematical model, which provides a foundation for subsequent aggregate analysis.

[0038] Specifically, the implementation process of step S2 clarifies the complete construction process of the temperature-energy dynamic mathematical model of the temperature-controlled load.

[0039] Based on the operating characteristics of temperature-controlled loads, S21 establishes a first-order thermodynamic model with refrigeration equipment as the object. The model incorporates key parameters such as indoor temperature, outdoor ambient temperature, equivalent thermal resistance, equivalent heat capacity, rated power, refrigeration efficiency, and on / off state variables. The nominal value of equivalent thermal resistance is 2℃ / kWh, the nominal value of equivalent heat capacity is 2kWh / ℃, the nominal value of refrigeration efficiency is 2.5, the nominal value of rated power is 0.5kW, and the on / off state variables only take two values: 0 or 1.

[0040] S22 addresses the nonlinear characteristics of the thermodynamic model by using variable substitution and mathematical transformation to convert it into a discrete-time linear model that is easy for aggregate calculations and stochastic control applications. The time step is set to 10 seconds, and the quantitative correspondence between parameters such as the inertia coefficient and input gain coefficient in the model and the original parameters such as equivalent thermal resistance, equivalent heat capacity, time step, and refrigeration efficiency is clearly defined.

[0041] Based on the user's set temperature comfort requirements, S23 defines the indoor temperature constraint range within the range of the set temperature value of 25℃ plus or minus half of the temperature dead zone of 1℃. This transforms the temperature constraint into the constraint range of the energy state variable. At the same time, it combines the rated power of the equipment with the reference power required to maintain the set temperature to define the value boundary of the power adjustment variable.

[0042] S24 introduces energy state variables and power adjustment variables, further rewriting the discrete-time linear model into an energy dynamic equation, forming a complete temperature-energy dynamic mathematical model, providing accurate individual model support for the cluster aggregation analysis in the subsequent step S3. The entire step-by-step implementation process must fully consider the heterogeneity of the equipment, and the specific parameters of each temperature-controlled load are determined by multiplying the nominal value by a uniformly distributed random number between 0.9 and 1.1.

[0043] Preferred, such as Figure 3 As shown, step S3 includes the following sub-steps: S31. Based on the discrete-time model obtained in step S2, the prediction time domain is set, and the energy dynamic equation of each temperature-controlled load in this time domain is represented in matrix form, clarifying the relationship between the coefficient matrix and the system parameters. S32, combining the power constraints and energy constraints of each temperature-controlled load, the feasible operating domain of a single load is represented as a polygon, and the constraint matrix and constraint vector corresponding to the polygon are determined; S33, construct a basic homothetic polygon whose parameters are the average of the parameters corresponding to all temperature-controlled loads. For each temperature-controlled load, solve the maximum scaling factor and translation vector through linear programming so that the scaled and translated basic polygon is completely included in the feasible domain of the load. S34. Summarize the scaling factors and translation vectors of all temperature-controlled loads, calculate the total scaling factor and total translation vector, and combine the constraint matrix and vector of the basic homothetic polygon to obtain the approximate aggregate feasible operating domain of the temperature-controlled load cluster, and determine the upper and lower limits of the aggregate power constraint and aggregate energy constraint of the cluster.

[0044] Specifically, step S3, the construction of the aggregated feasible operating domain, clarifies the implementation logic and technical details of each step. S31, based on the discrete-time linear model obtained in step S2, sets a reasonable prediction time domain length and represents the energy dynamic equation of each of the 5000 heterogeneous temperature-controlled loads in the microgrid within this time domain in matrix form. It clearly defines the correlation between the coefficient matrix and system parameters such as equivalent thermal resistance, equivalent heat capacity, and time step, ensuring that the matrix form accurately reflects the dynamic operating characteristics of the load. S32, combining the power constraints and energy constraints of each temperature-controlled load (where power constraints are determined based on rated power and reference power, and energy constraints are derived from the temperature comfort range), represents the feasible operating domain of a single load as a polygon. Through mathematical operations, it determines the constraint matrix and constraint vector corresponding to this polygon, clarifying the quantification boundaries of each constraint. S33 constructs a basic homomorphic polygon, where all parameters are the average of the corresponding parameters of all temperature-controlled loads. For each temperature-controlled load, a linear programming algorithm is used to solve for the maximum scaling factor and translation vector, ensuring that the scaled and translated basic polygon is completely included within the feasible domain of that load, guaranteeing the integrity of the constraints for a single load. S34 summarizes the scaling factors and translation vectors of all 5000 temperature-controlled loads, calculates the total scaling factor and total translation vector, and combines the constraint matrix and vectors of the basic homothetic polygons to obtain the approximate aggregated feasible operating domain of the temperature-controlled load cluster through geometric aggregation operations. It further derives and determines the upper and lower limits of the aggregated power constraint and the upper and lower limits of the aggregated energy constraint of the cluster, providing clear constraint inputs for the main control model in step S4.

[0045] Preferred, such as Figure 4 As shown, step S4 includes the following sub-steps: S41, at the set sampling time, the electric vehicle price-power dynamic response model constructed in step S1 is converted into a discrete-time state equation, and the expression form of each parameter in discrete time is clarified. S42, with the goal of minimizing the total control cost of the microgrid, constructs an objective function that includes electric vehicle incentive cost, temperature control load control cost and power tracking error penalty cost, and determines the incentive price and auxiliary control power as the optimization variables; S43, set constraints, including the range of incentive price, the upper and lower limits of instantaneous power of electric vehicle, and the response capability constraint that the auxiliary control power must be within the aggregate feasible operating domain calculated in step S3. S44. Solve the optimization problem to obtain the optimal incentive price sequence and the auxiliary control reference power sequence. The optimal incentive price sequence is issued as the main control signal, and the auxiliary control reference power sequence is used as the tracking target of the auxiliary control.

[0046] Specifically, step S4, the construction and solution process of the main control model, clarifies the core tasks and implementation methods of each stage. In S41, at the set sampling time, the sampling interval is consistent with the time step of step S2, which is 10 seconds. The electric vehicle price-power dynamic response model constructed in step S1 is converted into a discrete-time state equation. Through mathematical discretization, the specific expressions of parameters such as the response time constant, incentive price, and additive power uncertainty term in discrete time are clarified, ensuring that the discretized model can accurately reflect the dynamic response characteristics of the electric vehicle cluster. In S42, with the goal of minimizing the total control cost of the microgrid, an objective function is constructed that includes the electric vehicle incentive cost, the temperature-controlled load control cost, and the power tracking error penalty cost. The control cost coefficient of the temperature-controlled load is set to 0.4 yuan per kilowatt-hour, the penalty price coefficient of power mismatch is set to 0.5 yuan per kilowatt-hour, the control time domain is set to 24 hours, and the optimization variables are clarified as the incentive price and the auxiliary control power. S43 sets three types of constraints, including a reasonable range of incentive price values, upper and lower limits of the maximum charging and discharging power of the electric vehicle's instantaneous power, and a response capability constraint that the auxiliary control power must be within the aggregate feasible operating domain calculated in step S3, ensuring the physical feasibility of the optimization problem. S44 uses an efficient numerical optimization algorithm to solve the above optimization problem, obtaining the optimal incentive price sequence and the auxiliary control reference power sequence. The optimal incentive price sequence is immediately published to electric vehicle users as the main control signal, while the auxiliary control reference power sequence serves as the power tracking target for the auxiliary control in the next step S5, providing an economically optimal main control basis for coordinated control.

[0047] Preferred, such as Figure 5 As shown, step S5 includes the following sub-steps: S51, based on the steady-state operation characteristics of temperature-controlled loads, establishes the correlation between average power consumption and on / off duration, defines control signals, and clarifies the correspondence between the expected total power consumption of the temperature-controlled load cluster and the control signals; S52, based on the auxiliary control reference power and the average power consumption of the temperature-controlled load cluster determined in step S4, calculate the control signal required at the current moment to provide a basis for subsequent switching rate calculation; S53 uses the temperature distribution probability density evolution equation, combined with the average cooling rate, average heating rate and net temperature change rate, to solve the random switching rate of the temperature-controlled load from on to off and from off to on. S54 calculates the integral switching probability and the instantaneous switching probability respectively, and determines the final switching probability of the temperature-controlled load in the current state based on the two types of switching probabilities. After receiving the control signal, the equipment autonomously performs the state switching action according to the probability.

[0048] Specifically, the distributed stochastic auxiliary control implementation process in step S5 is divided into four sub-steps, clarifying the complete process of control signal generation, switching rate calculation, switching probability calculation, and action execution. S51, based on the steady-state operating characteristics of the temperature-controlled load, establishes a quantitative correlation between average power consumption and the on / off duration and duration. The control signal is defined through mathematical derivation, clarifying the correspondence between the expected total power consumption of the temperature-controlled load cluster and the control signal, providing a theoretical basis for subsequent control signal calculation. S52, based on the auxiliary control reference power determined in step S4 and the average power consumption of the temperature-controlled load cluster obtained through real-time monitoring, uses preset calculation logic to solve for the control signal required at the current moment, ensuring that the control signal accurately matches the tracking error compensation requirements of the main control, providing key input for subsequent switching rate calculation. S53 uses the temperature distribution probability density evolution equation, combined with the average cooling rate and average heating rate calculated from parameters such as equivalent thermal resistance, equivalent heat capacity, rated power, and cooling efficiency, and the net temperature change rate obtained by combining the two with the control signal, to mathematically solve the random switching rate of the temperature-controlled load from "on" to "off" and from "off" to "on," quantifying the dynamic characteristics of load state transitions. S54 calculates the integral switching probability and the instantaneous switching probability respectively. The integral switching probability is obtained by multiplying the average switching rate of adjacent moments by a 10-second time step, while the instantaneous switching probability is determined based on the ratio of the temperature change rate of adjacent moments. Based on the two types of switching probabilities, the final switching probability of the temperature-controlled load in the current state is determined by multiplication. After receiving the control signal broadcast by the control center, the temperature-controlled load autonomously executes the state switching action according to its current on / off state and the final switching probability, achieving rapid error compensation.

[0049] A microgrid power point tracking (PPT) method based on the coordination of electric vehicles (EVs) and temperature-controlled loads is proposed. This method constructs a hierarchical coordinated control architecture that organically combines price-incentivized main control with decentralized stochastic auxiliary control. It fully leverages the advantages of EVs—large regulation capacity and low control cost—and the characteristics of temperature-controlled loads—fast response speed and flexible switching—to achieve complementary advantages of these two heterogeneous resources at different time scales. This approach addresses the multiple requirements of microgrid PPT for large regulation capacity, high response speed, and economic efficiency. The dynamic response model designed for EVs fully considers response lag and behavioral uncertainty, making it more closely aligned with actual operating scenarios compared to traditional static models, thus improving prediction accuracy and control effectiveness during the main control phase. The decentralized stochastic control strategy for temperature-controlled loads does not require the collection of user private data. It guides equipment to make autonomous decisions through broadcast control signals, protecting user privacy while reducing system communication pressure and computational complexity, and maintaining user thermal comfort.

[0050] This method addresses the issue that a single control approach cannot simultaneously balance response speed and regulation capacity. It resolves the contradiction between response lag and insufficient accuracy through a collaborative mechanism where the main control handles the primary power regulation task, while the auxiliary control compensates for dynamic errors. To address the response lag and uncertainty issues of traditional price-based control, it utilizes a dynamic response model to accurately characterize the actual response process of electric vehicles, improving the predictability of power point tracking. To address the complexity and privacy risks of centralized control models, it employs distributed stochastic control and geometric aggregation methods to simplify the calculation process and avoid uploading user data. Furthermore, through a cost optimization model under multiple constraints, it achieves economically optimal operation while ensuring control performance, solving the problem of balancing performance and economy in existing technologies.

[0051] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0052] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads, characterized in that, include: Step S1: Construct a price-power dynamic response model for electric vehicle clusters within a microgrid. This model includes a response time constant, incentive price, steady-state target power determined by the price-power elasticity function, and additive power uncertainty terms, to characterize the dynamic response process of cluster aggregated power to changes in price signals. Step S2: Construct a temperature-energy dynamic mathematical model for a single temperature-controlled load, obtain a discrete-time linear model through equivalent transformation, and clarify the energy constraints and power adjustment constraints corresponding to temperature constraints and power constraints. Step S3: Based on the dynamic model of temperature-controlled load, the operating constraints of heterogeneous temperature-controlled loads in the microgrid are aggregated into a unified feasible operating domain using geometric calculation methods. The aggregated constraint boundary is solved by scaling and translating the basic homothetic polygon. Step S4: Construct a main control model based on price incentives, convert the electric vehicle dynamic response model into a discrete-time model, and minimize the total control cost. Combine incentive price constraints, electric vehicle instantaneous power constraints, and temperature control load aggregation response capability constraints to solve for the optimal incentive price sequence and auxiliary control reference power sequence. Step S5: Construct an auxiliary control model based on distributed stochastic control, calculate the control signal according to the auxiliary control reference power, solve the stochastic switching rate of the temperature-controlled load cluster through the temperature distribution probability density evolution equation, and determine the equipment state switching probability by combining the integral switching probability and the instantaneous switching probability and execute it. Step S6: Through the coordinated action of the electric vehicle cluster main control and the temperature-controlled load auxiliary control, microgrid tie-line power tracking is performed.

2. The microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, The electric vehicle cluster price-power dynamic response model in step S1 is as follows: ， The price-power elasticity function is: ， in, for The actual aggregate power of the electric vehicle cluster at any given time The discrete-time correlation coefficient. To incentivize prices, For the additive power uncertainty term, and These are the maximum discharge power and the maximum charging power, respectively. and These are the corresponding power thresholds. and These are the discharge and charge elasticity coefficients, respectively. and This represents the boundary of the elastic interval.

3. The microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, The temperature-energy dynamic mathematical model for a single temperature-controlled load in step S2 is as follows: ， The discrete-time linear model is: ， The energy constraint is: ， in, for Constant indoor temperature Where R is the outdoor ambient temperature, R is the equivalent thermal resistance, and C is the equivalent heat capacity. Rated power, For cooling efficiency, For the switch state, The inertia coefficient, For time step, For energy state variables, For power adjustment variables, To input the correlation coefficient, It is half the temperature of the dead zone.

4. A microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, The feasible operating domain for temperature-controlled load cluster aggregation in step S3 is: ， The approximate aggregation domain expression is: ， The aggregation constraint boundary is: , in, For the number of temperature-controlled loads, For a single load feasible region, and Based on the basic polygon constraint matrix and vector, This is the total scaling factor. The total translation vector, To aggregate relevant parameters, These are the upper and lower limits for the polymerization power constraint. The upper and lower limits are defined as the energy constraints for polymerization. These are the parameters for the basic model.

5. A microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, The main control objective function in step S4 includes the electric vehicle incentive cost: ， Temperature control load control cost: ， in, At the starting time, To control the time domain, The cost coefficient for temperature control load control. To assist in power control, This represents the actual aggregate power of the electric vehicle cluster. To incentivize prices.

6. A microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, The random switching rate of the temperature control load in step S5 is: ， , The switching probability is: ， in, This is an intermediate variable for net switching flux. These are the average cooling rate and the heating rate, respectively. Net temperature change rate For the integral switching probability, This represents the instantaneous switching probability. Indoor temperature, For a moment.

7. A microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, Step S2 includes the following sub-steps: S21. Based on the operating characteristics of temperature-controlled load, a first-order thermodynamic model is established with refrigeration equipment as the object. The model includes indoor temperature, outdoor ambient temperature, equivalent thermal resistance, equivalent heat capacity, rated power, refrigeration efficiency, and binary variable calibration parameters of switching state. S22, the nonlinear characteristics of the thermodynamic model are transformed into a discrete-time linear model that is easy to aggregate calculation and stochastic control application through variable substitution and mathematical transformation, clarifying the correspondence between the inertia coefficient, input gain coefficient and original parameters in the model; S23, based on the user's set temperature comfort requirements, determines the indoor temperature constraint range, transforms the temperature constraint into the energy state variable constraint range, and defines the value boundary of the power adjustment variable by combining the equipment's rated power and reference power. S24 introduces energy state variables and power adjustment variables, rewriting the discrete-time linear model into an energy dynamic equation, forming a complete temperature-energy dynamic mathematical model, which provides a foundation for subsequent aggregate analysis.

8. A microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, Step S3 includes the following sub-steps: S31. Based on the discrete-time model obtained in step S2, the prediction time domain is set, and the energy dynamic equation of each temperature-controlled load in this time domain is represented in matrix form, clarifying the relationship between the coefficient matrix and the system parameters. S32, combining the power constraints and energy constraints of each temperature-controlled load, the feasible operating domain of a single load is represented as a polygon, and the constraint matrix and constraint vector corresponding to the polygon are determined; S33, construct a basic homothetic polygon whose parameters are the average of the parameters corresponding to all temperature-controlled loads. For each temperature-controlled load, solve the maximum scaling factor and translation vector through linear programming so that the scaled and translated basic polygon is completely included in the feasible domain of the load. S34. Summarize the scaling factors and translation vectors of all temperature-controlled loads, calculate the total scaling factor and total translation vector, and combine the constraint matrix and vector of the basic homothetic polygon to obtain the approximate aggregate feasible operating domain of the temperature-controlled load cluster, and determine the upper and lower limits of the aggregate power constraint and aggregate energy constraint of the cluster.

9. A microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, Step S4 includes the following sub-steps: S41, at the set sampling time, the electric vehicle price-power dynamic response model constructed in step S1 is converted into a discrete-time state equation, and the expression form of each parameter in discrete time is clarified. S42, with the goal of minimizing the total control cost of the microgrid, constructs an objective function that includes electric vehicle incentive cost, temperature control load control cost and power tracking error penalty cost, and determines the incentive price and auxiliary control power as the optimization variables; S43, set constraints, including the range of incentive price, the upper and lower limits of instantaneous power of electric vehicle, and the response capability constraint that the auxiliary control power must be within the aggregate feasible operating domain calculated in step S3. S44. Solve the optimization problem to obtain the optimal incentive price sequence and the auxiliary control reference power sequence. The optimal incentive price sequence is issued as the main control signal, and the auxiliary control reference power sequence is used as the tracking target of the auxiliary control.

10. A microgrid power point tracking method based on the coordination of electric vehicles and temperature-controlled loads according to claim 1, characterized in that, Step S5 It includes the following steps: S51, based on the steady-state operation characteristics of temperature-controlled loads, establishes the correlation between average power consumption and on / off duration, defines control signals, and clarifies the correspondence between the expected total power consumption of the temperature-controlled load cluster and the control signals; S52, based on the auxiliary control reference power and the average power consumption of the temperature-controlled load cluster determined in step S4, calculate the control signal required at the current moment to provide a basis for subsequent switching rate calculation; S53 uses the temperature distribution probability density evolution equation, combined with the average cooling rate, average heating rate and net temperature change rate, to solve the random switching rate of the temperature-controlled load from on to off and from off to on. S54 calculates the integral switching probability and the instantaneous switching probability respectively, and determines the final switching probability of the temperature-controlled load in the current state based on the two types of switching probabilities. After receiving the control signal, the equipment autonomously performs the state switching action according to the probability.

Images