A coordinated control method and device of a power system considering electric vehicles and fixed distributed resources, an electronic device and a storage medium

Abstract

The application discloses a kind of considering electric power system of electric automobile and fixed type distributed resource's coordinated control method, device, electronic equipment and storage medium, belong to electric power system and its automation technical field, the method includes: constructing system state vector;State vector input parameter dynamic adjustment model, generate equivalent model parameter adjustment quantity, and update initial equivalent model parameter, obtain target equivalent inertia constant, target equivalent frequency modulation coefficient and target equivalent response delay;According to target equivalent frequency modulation coefficient, determine aggregated equivalent up-regulation capacity and aggregated equivalent down-regulation capacity;With the minimum net cost as target, construct resource collaborative scheduling model and solve, obtain optimal scheduling instruction;According to optimal scheduling instruction, electric automobile and fixed type distributed resource are coordinated control. By implementing the present application, the technical problem that the control result of electric automobile and fixed type distributed resource coordinated control in the prior art is not accurate can be solved.

Description

Technical Field

[0001] This invention relates to the field of power systems and their automation technology, and more specifically to a coordinated control method, apparatus, electronic device, and storage medium for power systems that take into account electric vehicles and stationary distributed resources. Background Technology

[0002] In the operation of modern power systems, frequency stability is a core indicator for ensuring the security and reliability of power supply. With the widespread integration of new power sources into the grid, the large-scale participation of electric vehicles and stationary distributed resources such as distributed photovoltaic power sources and distributed energy storage devices has significantly altered the frequency response characteristics of the power system. Achieving coordinated control of electric vehicles and stationary distributed resources can effectively integrate decentralized regulation capabilities to compensate for the shortcomings of traditional frequency regulation resources. This is a necessary means to maintain power system frequency stability and improve the power system's immunity to disturbances, and it is also an important foundation for ensuring the economical and reliable operation of the power system.

[0003] Current technologies for coordinated control of electric vehicles and stationary distributed resources face the problem of inaccurate control results. This inaccuracy stems from the fact that most existing control methods rely on static modeling, which suffers from fixed model parameters. Specifically, existing methods typically use pre-set, fixed model parameters, failing to update these parameters based on operational factors such as ambient temperature, electric vehicle charging / discharging power, battery state of charge, and availability probability. This leads to discrepancies between the assessment of increased and decreased available capacity of aggregated resources and the actual situation. Due to this inaccurate aggregated capacity assessment, the control commands generated by the resource coordination scheduling model lack precision when calculating net costs using predicted clearing prices. Summary of the Invention

[0004] This invention provides a method, apparatus, electronic device, and storage medium for coordinated control of power systems that consider electric vehicles and stationary distributed resources, which can solve the technical problem of inaccurate control results in the coordinated control of electric vehicles and stationary distributed resources in the prior art.

[0005] An embodiment of the present invention provides a coordinated control method for a power system considering electric vehicles and stationary distributed resources, comprising: Acquire power system frequency operation data, ambient temperature, predicted clearing price, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources. The state vector is determined based on frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources. The state vector is input into a preset parameter dynamic adjustment model, so that the parameter dynamic adjustment model generates an equivalent model parameter adjustment amount based on the state vector; based on the equivalent model parameter adjustment amount, the preset initial equivalent model parameters are updated to generate target equivalent model parameters; wherein, the target equivalent model parameters include target equivalent inertia constant, target equivalent frequency modulation coefficient, and target equivalent response delay; based on the target equivalent frequency modulation coefficient, the aggregate equivalent up-adjustment capacity and aggregate equivalent down-adjustment capacity are determined; With the goal of minimizing net cost, the objective function and constraints of the resource collaborative scheduling model are constructed based on the objective equivalent inertia constant, objective equivalent response delay, aggregate equivalent up-adjustment capacity, aggregate equivalent down-adjustment capacity, and predicted clearing price. Under the constraints, the objective function of the resource collaborative scheduling model is solved to generate the minimum net cost and the optimal scheduling instruction. Based on minimum net cost, aggregated equivalent upsizing capacity, aggregated equivalent downsizing capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions, coordinated control is performed on electric vehicles and fixed distributed resources.

[0006] Furthermore, the predicted clearing price can be obtained through the following methods: Obtain the frequency regulation capacity requirements of the power system; The frequency operation data and frequency modulation capacity requirements are input into a preset price prediction model, so that the price prediction model generates a predicted clearing price based on the frequency operation data and frequency modulation capacity requirements. The preset price prediction model is trained in the following way: Obtain several price prediction training samples; wherein, the price prediction model is trained from several price prediction training samples; each price prediction training sample includes historical frequency regulation capacity demand, historical frequency operation data, and clearing price label; the clearing price label is the historical clearing price corresponding to the historical frequency regulation capacity demand and historical frequency operation data; Each price prediction training sample is sequentially input into the price prediction model to train the price prediction model until a preset first training iteration is reached. Each time a price prediction training sample is received, the price prediction model outputs the predicted clearing price corresponding to that sample. A first loss function value is calculated based on the predicted clearing price and its corresponding clearing price label. The price prediction model is then updated based on the first loss function value.

[0007] Furthermore, based on frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and the output power of fixed distributed resources, the state vector is determined, including: Based on the charging and discharging power of electric vehicles and the output power of fixed distributed resources, determine the total available capacity for aggregated resources to be increased and the total available capacity for aggregated resources to be decreased. The overall effective state of charge of the electric vehicle is determined based on the battery state of charge and the availability probability of the electric vehicle. A state vector is constructed based on frequency operation data, ambient temperature, total available capacity for upward adjustment of aggregated resources, total available capacity for downward adjustment of aggregated resources, and comprehensive effective state of charge.

[0008] Furthermore, the model is trained and dynamically adjusted using preset parameters in the following manner: Acquire several parameter adjustment training samples; wherein, the parameter dynamic adjustment model is trained from several parameter adjustment training samples; each parameter adjustment training sample includes a historical state vector and a target adjustment amount label; the target adjustment amount label is the historical equivalent model parameter adjustment amount corresponding to the historical state vector. Each parameter adjustment training sample is sequentially input into the parameter dynamic adjustment model to train the model until a preset second training iteration is reached. The parameter dynamic adjustment model outputs the corresponding predicted equivalent model parameter adjustment amount each time it receives a parameter adjustment training sample. A second loss function value is calculated based on the predicted equivalent model parameter adjustment amount and the corresponding target adjustment amount label. The parameter dynamic adjustment model is then updated based on the second loss function value.

[0009] Furthermore, the equivalent model parameter adjustment amounts include: equivalent inertia constant adjustment amount, equivalent frequency modulation coefficient adjustment amount, and equivalent response delay adjustment amount; The preset initial equivalent model parameters include: initial equivalent inertia constant, initial equivalent frequency modulation coefficient, and initial equivalent response delay; Based on the adjustment amount of the equivalent model parameters, the preset initial equivalent model parameters are updated to generate the target equivalent model parameters, including: The equivalent inertial constant adjustment is superimposed with the initial equivalent inertial constant to generate the target equivalent inertial constant; The equivalent frequency modulation coefficient adjustment is superimposed with the initial equivalent frequency modulation coefficient to generate the target equivalent frequency modulation coefficient; The equivalent response delay adjustment is added to the initial equivalent response delay to generate the target equivalent response delay; The target equivalent inertial constant, the target equivalent frequency modulation coefficient, and the target equivalent response delay are used as the target equivalent model parameters.

[0010] Furthermore, based on the target equivalent frequency modulation coefficient, the aggregate equivalent up-modulation capacity and aggregate equivalent down-modulation capacity are determined, including: Based on the target equivalent frequency modulation coefficient, the total available capacity of aggregated resources is equivalently converted to generate the aggregated equivalent upward capacity. Based on the target equivalent frequency modulation coefficient, the total downgraded available capacity of aggregated resources is equivalently converted to generate the aggregated equivalent downgraded capacity.

[0011] Furthermore, based on minimum net cost, aggregated equivalent upsizing capacity, aggregated equivalent downsizing capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions, coordinated control is implemented for electric vehicles and stationary distributed resources, including: The total equivalent capacity of aggregation is determined based on the aggregate equivalent upward capacity and aggregate equivalent downward capacity. The unit equivalent capacity price is determined based on the minimum net cost and the aggregate equivalent total capacity; The target price is determined based on the unit equivalent capacity price, the predicted clearing price, frequency operation data, and the target equivalent frequency regulation coefficient; The aggregated equivalent upward capacity, aggregated equivalent downward capacity, and target price will be sent to the frequency modulation market for bidding, and the winning capacity will be determined. The capacity allocation coefficient is determined based on the winning bid capacity, the aggregate equivalent upward capacity, and the aggregate equivalent downward capacity. The optimal scheduling instruction is modified based on the capacity allocation coefficient to generate the actual execution instruction; Based on the actual execution instructions, the charging and discharging power of electric vehicles and the output power of stationary distributed resources are adjusted to achieve coordinated control of electric vehicles and stationary distributed resources.

[0012] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0013] One embodiment of the present invention provides a coordinated control device for a power system considering electric vehicles and stationary distributed resources, comprising: a data acquisition module, a state vector determination module, a capacity processing module, a scheduling model solving module, and a coordinated control execution module; The data acquisition module is used to acquire frequency operation data of the power system, ambient temperature, predicted clearing price, charging and discharging power of electric vehicles, battery state of charge of electric vehicles, availability probability of electric vehicles, and output power of fixed distributed resources. The state vector determination module is used to determine the state vector based on frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources. The capacity processing module is used to input the state vector into a preset parameter dynamic adjustment model, so that the parameter dynamic adjustment model generates an equivalent model parameter adjustment amount according to the state vector; based on the equivalent model parameter adjustment amount, the preset initial equivalent model parameters are updated to generate target equivalent model parameters; wherein, the target equivalent model parameters include a target equivalent inertia constant, a target equivalent frequency modulation coefficient, and a target equivalent response delay; based on the target equivalent frequency modulation coefficient, the aggregated equivalent upward capacity and aggregated equivalent downward capacity are determined. The scheduling model solving module is used to construct the objective function and constraints of the resource collaborative scheduling model with the goal of minimizing net cost, based on the objective equivalent inertia constant, objective equivalent response delay, aggregate equivalent up-adjustment capacity, aggregate equivalent down-adjustment capacity, and predicted clearing price; under the constraints, the objective function of the resource collaborative scheduling model is solved to generate the minimum net cost and the optimal scheduling instruction. The coordination control execution module is used to coordinate and control electric vehicles and fixed distributed resources based on minimum net cost, aggregate equivalent up-capacity, aggregate equivalent down-capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions.

[0014] Based on the above method embodiments, the present invention provides corresponding electronic device embodiments.

[0015] One embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the coordinated control method for a power system considering electric vehicles and stationary distributed resources as described in any of the above-described method embodiments.

[0016] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments.

[0017] One embodiment of the present invention provides a storage medium storing a computer program thereon, wherein, when the computer program is executed, it controls the device where the storage medium is located to perform the coordinated control method for a power system considering electric vehicles and stationary distributed resources as described in any of the above-described method embodiments.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a method, apparatus, electronic device, and storage medium for coordinated control of a power system considering electric vehicles and stationary distributed resources. The method acquires frequency operation data, ambient temperature, predicted clearing price, electric vehicle charging and discharging power, battery state of charge, availability probability, and output power of stationary distributed resources. Based on the frequency operation data, ambient temperature, electric vehicle operating status, and stationary distributed resource output power, a system state vector is constructed. The state vector is used as input parameters to dynamically adjust the model, generating equivalent model parameter adjustment amounts, and updating the initial equivalent model parameters to obtain the target equivalent inertia constant, target equivalent frequency regulation coefficient, and target equivalent response delay. The aggregated equivalent upward and downward capacity adjustments are determined based on the target equivalent frequency regulation coefficient. With minimizing net cost as the objective, a resource collaborative scheduling model is constructed and solved by combining the target equivalent inertia constant, target equivalent response delay, aggregated equivalent upward and downward capacity adjustments, and predicted clearing price to obtain the minimum net cost and optimal scheduling command. Based on the optimal scheduling command and relevant operating parameters, coordinated control of electric vehicles and stationary distributed resources is performed.

[0019] This invention determines the state vector by acquiring the ambient temperature of the power system, the charging and discharging power of electric vehicles, the battery state of charge of electric vehicles, and the availability probability of electric vehicles. It then dynamically adjusts the model parameters based on the state vector input parameters to generate the target equivalent model parameters, overcoming the shortcomings of pre-setting fixed model parameters. Based on the target equivalent frequency regulation coefficient in the target equivalent model parameters, it accurately determines the aggregated equivalent upward and downward capacity adjustments. Furthermore, by combining the predicted clearing price, the target equivalent inertia constant, and the target equivalent response delay, it constructs a resource collaborative scheduling model to generate optimal scheduling instructions. This solves the problem of inaccurate aggregated capacity assessment leading to a lack of precision in control instructions, achieving precise coordinated control of electric vehicles and fixed distributed resources. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a coordinated control method for a power system considering electric vehicles and stationary distributed resources, provided by an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the structure of a coordinated control device for a power system considering electric vehicles and stationary distributed resources, provided by an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] like Figure 1 As shown, to address the technical problem of inaccurate control results in the coordinated control of electric vehicles and stationary distributed resources in the prior art, an embodiment of the present invention provides a coordinated control method for a power system considering electric vehicles and stationary distributed resources, comprising at least the following steps: Step S1: Obtain the frequency operation data of the power system, ambient temperature, predicted clearing price, charging and discharging power of electric vehicles, battery state of charge of electric vehicles, availability probability of electric vehicles, and output power of fixed distributed resources.

[0024] Specifically, acquiring multi-dimensional state information provides the foundation for achieving coordinated control of the power system. Frequency operation data includes the power system's frequency deviation and frequency change rate. Ambient temperature is boundary condition data, which refers to external environmental variables that are not directly controlled but significantly affect the operating performance, aging rate, and frequency response characteristics of physical equipment and the entire power system. Multiple electric vehicles participate in the coordinated control of the power system. The battery state of charge (SBC) of each electric vehicle is its real-time SBC. The charging and discharging power of each electric vehicle is its current charging and discharging power. Stationary distributed resources include distributed photovoltaic (PV) power sources and distributed energy storage devices. The output power of stationary distributed resources includes the real-time output of distributed PV power sources and the charging and discharging power of each distributed energy storage device.

[0025] After acquiring the raw state data, preprocessing operations are performed on the data. These operations address missing, outlier, and duplicate values. Outliers include power fluctuations exceeding the physically possible range. The preprocessing process removes outliers and duplicates and interpolates missing values. Data from different sources and with different sampling frequencies are aggregated using interpolation methods to achieve the same timestamp. The preprocessed data is then stored in a real-time database, and a standardized data stream containing timestamps, subject identifiers, and various data fields is output.

[0026] Cluster analysis was performed on the charging history records of each electric vehicle to calculate the availability probability for each vehicle at different time periods. The charging history records of electric vehicles include charging start time, charging duration, charging amount, and charging location. The availability probability is a value between zero and one, calculated by constructing a probability mapping function that includes an exponential decay factor. The specific calculation formula is as follows: In the formula, Indicates the availability probability. Indicates historical attendance rate. This indicates the time until the next departure (in hours). This represents a pre-set time penalty sensitivity constant used to adjust the severity of availability degradation as the departure time approaches. The smaller the difference between the current departure time and the next departure time, the larger the value of the exponential decay term. After deduction, this indicates a more limited supply of dispatchable energy for the corresponding electric vehicle, and a lower availability probability.

[0027] In a preferred embodiment, the predicted clearing price is obtained as follows: Obtain the frequency regulation capacity requirements of the power system; The frequency operation data and frequency modulation capacity requirements are input into a preset price prediction model, so that the price prediction model generates a predicted clearing price based on the frequency operation data and frequency modulation capacity requirements. The preset price prediction model is trained in the following way: Obtain several price prediction training samples; wherein, the price prediction model is trained from several price prediction training samples; each price prediction training sample includes historical frequency regulation capacity demand, historical frequency operation data, and clearing price label; the clearing price label is the historical clearing price corresponding to the historical frequency regulation capacity demand and historical frequency operation data; Each price prediction training sample is sequentially input into the price prediction model to train the price prediction model until a preset first training iteration is reached. Each time a price prediction training sample is received, the price prediction model outputs the predicted clearing price corresponding to that sample. A first loss function value is calculated based on the predicted clearing price and its corresponding clearing price label. The price prediction model is then updated based on the first loss function value.

[0028] Specifically, the predicted clearing price is obtained through the following methods: First, the frequency regulation capacity demand of the power system is obtained. This demand is based on the frequency regulation capacity demand data published by the power system operators. Second, the frequency operation data and the frequency regulation capacity demand are input into a pre-defined price prediction model, which generates the predicted clearing price based on these factors. The pre-defined price prediction model is constructed using a time-series prediction model based on a Long Short-Term Memory (LSTM) network. The input features of the LSM network include not only frequency operation data and frequency regulation capacity demand but also historical clearing price time-series data. The output of the LSM network is the predicted clearing price for the next time period. The predicted clearing price serves as an economic signal for lower-level resource optimization, guiding the allocation of various electric vehicles and stationary distributed resources towards high-yield scheduling methods.

[0029] The pre-defined price prediction model is trained as follows: A number of price prediction training samples are obtained. The price prediction model is trained using these samples. Each training sample includes historical FM capacity demand, historical frequency operation data, and a clearing price label. The clearing price label represents the historical clearing price corresponding to the historical FM capacity demand and historical frequency operation data. Each training sample is sequentially input into the price prediction model for training until a pre-defined first training iteration is reached. This first training iteration is a pre-set termination condition for network weight updates. Upon receiving a training sample, the price prediction model outputs the predicted clearing price corresponding to that sample. Based on the predicted clearing price and its corresponding clearing price label, a first loss function value is calculated. The formula for calculating the first loss function value is as follows: In the formula, This represents the value of the first loss function. Indicates the predicted clearing price. Indicates the clearance price tag.

[0030] The price prediction model is updated based on the first loss function value. The partial derivative gradient of the first loss function value with respect to the internal network weights of the price prediction model is calculated using the backpropagation algorithm. Then, the gradient descent algorithm is used to update the internal network weights of the price prediction model along the reverse direction of the gradient, ensuring that the predicted clearing price output by the price prediction model continuously approximates the historical objective clearing price label.

[0031] By accurately quantifying the availability probability of each electric vehicle and using long short-term memory networks to deeply predict price trends, this provides highly reliable data support for the subsequent construction of a resource collaborative scheduling model that integrates physical security and minimizes net cost objectives, thereby significantly improving the accuracy of frequency response control of the power system in response to complex fluctuations.

[0032] Step S2: Determine the state vector based on the frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources.

[0033] In a preferred embodiment, the state vector is determined based on frequency operating data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources, including: Based on the charging and discharging power of electric vehicles and the output power of fixed distributed resources, determine the total available capacity for aggregated resources to be increased and the total available capacity for aggregated resources to be decreased. The overall effective state of charge of the electric vehicle is determined based on the battery state of charge and the availability probability of the electric vehicle. A state vector is constructed based on frequency operation data, ambient temperature, total available capacity for upward adjustment of aggregated resources, total available capacity for downward adjustment of aggregated resources, and comprehensive effective state of charge.

[0034] Specifically, the state vector provides structured feature inputs for the subsequent parameter dynamic adjustment model to perceive the external power system operating environment and the internal resource status.

[0035] In one specific embodiment, a state vector is determined based on frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources. This includes determining the total upward adjustment available capacity and the total downward adjustment available capacity of aggregated resources based on the charging and discharging power of electric vehicles and the output power of fixed distributed resources. Aggregated resources include multiple electric vehicles and multiple fixed distributed resources. The total upward adjustment available capacity and the total downward adjustment available capacity represent the sum of effective power that aggregated resources can adjust upward and downward in response to power system dispatch instructions at the current moment, respectively. The calculation formulas for the total upward adjustment available capacity and the total downward adjustment available capacity are as follows: In the formula, This indicates the total increase in available capacity of aggregated resources. This represents the total number of electric vehicles. Indicates the serial number of the electric vehicle. Indicates the first Increased available capacity for electric vehicles This represents the total number of fixed distributed resources. Indicates the sequence number of the fixed distributed resource. Indicates the first The maximum output power of a fixed distributed resource. Indicates the first The output power of a fixed distributed resource This indicates the total reduced available capacity of aggregated resources. Indicates the first Reduced available capacity for electric vehicles Indicates the first Minimum output power of a fixed distributed resource.

[0036] No. The upward and downward adjustments to the available capacity of an electric vehicle are calculated using the current charging and discharging power of the electric vehicle and the maximum charging and discharging limit constraints preset by the physical equipment.

[0037] Specifically, the The formulas for calculating the upward and downward adjustment of available capacity for electric vehicles are as follows: In the formula, and They represent the first The maximum discharge power limit and the maximum charging power limit of an electric vehicle (charging power is recorded as a negative value, and discharge power is recorded as a positive value). This indicates the current charging and discharging power.

[0038] Fixed distributed photovoltaic power sources that do not have power regulation capabilities are not included in the calculation of available fixed distributed resource capacity.

[0039] The overall state of charge (SPC) of electric vehicles (EVs) is determined based on their battery state of charge (SPC) and availability probability. The SPC reflects the overall true dispatchable power margin of all EVs participating in power system frequency response control. The formula for calculating the SPC is as follows: In the formula, Indicates the overall effective state of charge. Indicates the first The state of charge of an electric vehicle's battery. Indicates the first The availability probability of an electric vehicle. By using the availability probability as a weight to perform a weighted summation of the state of charge of individual batteries, invalid battery charge values ​​that are about to be driven away can be completely filtered out, objectively restoring the true energy support base of a large-scale electric vehicle fleet.

[0040] Based on frequency operation data, ambient temperature, total available up-adjustment capacity of aggregated resources, total available down-adjustment capacity of aggregated resources, and comprehensive effective state of charge, a state vector is constructed. The specific method for constructing the state vector is to concatenate and combine multi-dimensional physical feature variables according to a pre-defined feature arrangement order to form a one-dimensional feature array. The pre-defined feature arrangement order is a fixed index position within the array of each physical feature. Using a fixed index position ensures the absolute stability of the input structure of the deep network model. The specific mathematical expression of the state vector is as follows: In the formula, Represents the state vector. Indicates the frequency deviation of the power system. Represents the rate of change of frequency in a power system. Indicates ambient temperature.

[0041] By constructing a state vector that includes the external environmental characteristics of the power system and the internal physical state characteristics of aggregated resources, the parameter dynamic adjustment model can accurately perceive the actual operating conditions, thus providing reliable data support for subsequent accurate updates of equivalent model parameters.

[0042] Step S3: Input the state vector into the preset parameter dynamic adjustment model so that the parameter dynamic adjustment model generates an equivalent model parameter adjustment amount based on the state vector; based on the equivalent model parameter adjustment amount, update the preset initial equivalent model parameters to generate target equivalent model parameters; wherein, the target equivalent model parameters include target equivalent inertia constant, target equivalent frequency modulation coefficient, and target equivalent response delay; determine the aggregate equivalent up-adjustment capacity and aggregate equivalent down-adjustment capacity based on the target equivalent frequency modulation coefficient.

[0043] In a preferred embodiment, a preset parameter dynamic adjustment model is trained in the following manner: Acquire several parameter adjustment training samples; wherein, the parameter dynamic adjustment model is trained from several parameter adjustment training samples; each parameter adjustment training sample includes a historical state vector and a target adjustment amount label; the target adjustment amount label is the historical equivalent model parameter adjustment amount corresponding to the historical state vector. Each parameter adjustment training sample is sequentially input into the parameter dynamic adjustment model to train the model until a preset second training iteration is reached. The parameter dynamic adjustment model outputs the corresponding predicted equivalent model parameter adjustment amount each time it receives a parameter adjustment training sample. A second loss function value is calculated based on the predicted equivalent model parameter adjustment amount and the corresponding target adjustment amount label. The parameter dynamic adjustment model is then updated based on the second loss function value.

[0044] In a preferred embodiment, the equivalent model parameter adjustment includes: equivalent inertia constant adjustment, equivalent frequency modulation coefficient adjustment, and equivalent response delay adjustment; The preset initial equivalent model parameters include: initial equivalent inertia constant, initial equivalent frequency modulation coefficient, and initial equivalent response delay; Based on the adjustment amount of the equivalent model parameters, the preset initial equivalent model parameters are updated to generate the target equivalent model parameters, including: The equivalent inertial constant adjustment is superimposed with the initial equivalent inertial constant to generate the target equivalent inertial constant; The equivalent frequency modulation coefficient adjustment is superimposed with the initial equivalent frequency modulation coefficient to generate the target equivalent frequency modulation coefficient; The equivalent response delay adjustment is added to the initial equivalent response delay to generate the target equivalent response delay; The target equivalent inertial constant, the target equivalent frequency modulation coefficient, and the target equivalent response delay are used as the target equivalent model parameters.

[0045] In a preferred embodiment, determining the aggregated equivalent up-modulation capacity and aggregated equivalent down-modulation capacity based on the target equivalent frequency modulation coefficient includes: Based on the target equivalent frequency modulation coefficient, the total available capacity of aggregated resources is equivalently converted to generate the aggregated equivalent upward capacity. Based on the target equivalent frequency modulation coefficient, the total downgraded available capacity of aggregated resources is equivalently converted to generate the aggregated equivalent downgraded capacity.

[0046] Specifically, the preset dynamic parameter adjustment model is constructed using a multilayer perceptron (MLP) deep neural network, whose network topology includes an input layer, several hidden layers, and an output layer. To achieve accurate mapping of the feature space, the number of neurons in the input layer strictly corresponds to the feature dimension of the constructed state vector; the number of neurons in the output layer is set to 3, corresponding to the adjustment of the equivalent inertia constant, the equivalent frequency modulation coefficient, and the equivalent response delay, respectively. By outputting equivalent model parameter adjustments rather than absolute parameter values, the dynamic parameter adjustment model effectively avoids numerical oscillations and ensures the absolute stability of the network iteration process.

[0047] In a specific embodiment, a preset parameter dynamic adjustment model is trained as follows: Several parameter adjustment training samples are obtained. The parameter dynamic adjustment model is trained using these training samples. Each parameter adjustment training sample includes a historical state vector and a target adjustment amount label. The target adjustment amount label is the objective, true, historically equivalent model parameter adjustment amount corresponding to the historical state vector. Each parameter adjustment training sample is sequentially input into the parameter dynamic adjustment model, and the model is iteratively trained until a preset second training iteration is reached. The preset second training iteration is a pre-defined termination condition for the iterative update of network parameters.

[0048] The parameter dynamic adjustment model outputs the corresponding predicted equivalent model parameter adjustment amount upon receiving each parameter adjustment training sample. Based on the predicted equivalent model parameter adjustment amount and the corresponding target adjustment amount label, the second loss function value is calculated. The mathematical expression for the second loss function value is constructed using the squared norm of a vector, as shown below: In the formula, This represents the value of the second loss function. This represents the vector of parameter adjustments for the predictive equivalent model. This represents the target adjustment label vector. Using the squared L2 norm, the absolute deviation of the prediction can be comprehensively quantified by parameters from three dimensions: inertia constant, frequency modulation coefficient, and response delay.

[0049] The parameter dynamic adjustment model is updated based on the value of the second loss function. Specifically, the backpropagation algorithm is used to calculate the partial derivative gradient of the second loss function value with respect to the internal network parameters of the parameter dynamic adjustment model, and the network parameter matrix is ​​gradually updated along the reverse direction of gradient descent, so that the output predicted equivalent model parameter adjustment amount gradually approaches the objective and true target adjustment amount label.

[0050] In one specific embodiment, the equivalent model parameter adjustments include adjustments to the equivalent inertia constant, the equivalent frequency regulation coefficient, and the equivalent response delay. The preset initial equivalent model parameters include the initial equivalent inertia constant, the initial equivalent frequency regulation coefficient, and the initial equivalent response delay. These preset initial equivalent model parameters are empirical benchmark values ​​pre-set based on historical normal operating conditions of the power system.

[0051] Based on the adjustment of equivalent model parameters, the preset initial equivalent model parameters are updated to generate target equivalent model parameters. Specific steps include superimposing the equivalent inertia constant adjustment with the initial equivalent inertia constant to generate the target equivalent inertia constant. The target equivalent inertia constant reflects the physical suppression capability of aggregated resources on the power system's frequency change rate. The equivalent frequency regulation coefficient adjustment is superimposed with the initial equivalent frequency regulation coefficient to generate the target equivalent frequency regulation coefficient. The target equivalent frequency regulation coefficient reflects the physical gain level of aggregated resources providing steady-state frequency support for the power system. The equivalent response delay adjustment is superimposed with the initial equivalent response delay to generate the target equivalent response delay. The target equivalent response delay reflects the average time span from receiving dispatch instructions to generating an actual effective physical response from the underlying control equipment. The target equivalent inertia constant, target equivalent frequency regulation coefficient, and target equivalent response delay are unified as the target equivalent model parameters.

[0052] In a specific embodiment, the aggregated equivalent upward capacity and aggregated equivalent downward capacity are determined based on the target equivalent frequency modulation coefficient. The specific steps include: equivalently converting the total available upward capacity of aggregated resources according to the target equivalent frequency modulation coefficient to generate the aggregated equivalent upward capacity; and equivalently converting the total available downward capacity of aggregated resources according to the target equivalent frequency modulation coefficient to generate the aggregated equivalent downward capacity. The equivalent conversion process first requires calculating the dynamic response capability coefficient to quantify the dynamic contribution of the resources. The formula for calculating the dynamic response capability coefficient is as follows: In the formula, Indicates the dynamic response capability coefficient. This represents the target equivalent frequency modulation coefficient. This represents the reference equivalent frequency modulation factor when the resource is in its basic state.

[0053] After calculating the dynamic response capability coefficient, the total upward adjustment of available capacity and the total downward adjustment of available capacity are respectively converted using the dynamic response capability coefficient. The specific mathematical expression formula is as follows: In the formula, This indicates an equivalent increase in capacity through aggregation. This indicates an equivalent reduction in capacity due to aggregation.

[0054] By dynamically predicting parameter adjustments and updating the initial baseline parameters, and completing the equivalent conversion of available capacity, the time-varying physical characteristics of electric vehicles and stationary distributed resources are accurately characterized, thereby significantly improving the accuracy of the power system's assessment of the real frequency response capability of massive heterogeneous resources.

[0055] Step S4: With the goal of minimizing net cost, construct the objective function and constraints of the resource collaborative scheduling model based on the objective equivalent inertia constant, objective equivalent response delay, aggregate equivalent up-adjustment capacity, aggregate equivalent down-adjustment capacity, and predicted clearing price; under the constraints, solve the objective function of the resource collaborative scheduling model to generate the minimum net cost and optimal scheduling instruction.

[0056] In one specific embodiment, constructing a resource collaborative scheduling model requires comprehensive consideration of both physical security boundaries and the principle of maximizing economic benefits. Based on the physical characteristics of the underlying devices, the constraints are divided into response delay compensation constraints, equivalent inertial support constraints, energy capacity security constraints, and power allocation balance equation constraints.

[0057] A response delay compensation constraint is constructed based on the target equivalent response delay. To address the inherent time difference between the underlying physical device receiving the command and the actual output power, the control logic needs to issue scheduling commands in advance to ensure that the physical output power arrives precisely at the expected time. The mathematical expression for the response delay compensation constraint is as follows: In the formula, This represents the actual response power at the current moment. This represents the optimal scheduling instruction. Represents a time variable. This indicates the target equivalent response delay.

[0058] Equivalent inertial support constraints are constructed based on the target equivalent inertial constant. These constraints mandate that the power change rate of aggregated resources must be greater than or equal to a specific physical control threshold, thereby ensuring that the underlying resources possess sufficient physical support capabilities to withstand disturbances. The formula for the equivalent inertial support constraints is as follows: In the formula, This represents the overall power change rate of aggregated resources. Represents the target's equivalent inertial constant. This indicates the frequency deviation of the power system.

[0059] Energy capacity safety constraints are constructed based on aggregated equivalent upward and downward capacity adjustments. The purpose of these constraints is to prevent overcharging and discharging of underlying physical devices, which could lead to accelerated hardware lifespan degradation. The specific mathematical expression for the energy capacity safety constraints is as follows: In the formula, This represents the energy supported by inertia. This represents the pre-set grid security energy threshold. The pre-set grid security energy threshold is the minimum energy reserve baseline determined by the power system dispatch center based on long-term stable operating condition tests.

[0060] The objective function is designed to quantify the overall economic cost incurred by underlying devices in participating in scheduling and regulation. A pre-defined internal cost function is used to calculate the discounted costs corresponding to the physical devices performing control actions.

[0061] For electric vehicles, the pre-defined internal cost function covers battery depreciation due to natural aging, routine maintenance and operating expenses, and economic penalties for deviating from the user's predetermined goals. Considering the differences in the operating conditions of each electric vehicle, the overall internal cost is the sum of the individual costs of all electric vehicles. The formula for calculating the overall internal cost function for electric vehicles is as follows: In the formula, This represents the total internal cost of the electric vehicle group. Indicates assignment to the first Optimal scheduling instructions for electric vehicles. , , They represent the first The electric vehicle is pre-set with battery aging cost coefficients, daily operating cost coefficients, and deviation penalty cost coefficients. Indicates the time after the end of the scheduling control cycle. The final state of charge of the battery of an electric vehicle. Indicates the first The target battery state of charge set by the electric vehicle user. All pre-set cost coefficients are statistical empirical constants obtained by fitting the factory parameters of physical components with past operating records.

[0062] In order for the resource collaborative scheduling model to calculate the deviation penalty cost, the solution algorithm needs to rely on the state transition equation to evaluate the actual impact of scheduling commands on battery capacity. Final state of charge of an electric vehicle The state transition calculation formula is as follows: In the formula, Indicates the first The initial state of charge of the battery of an electric vehicle (included in the state vector). This indicates the preset duration of a single scheduling control cycle. Indicates the first The rated capacity of an electric vehicle battery. Indicates the first The overall charging and discharging efficiency coefficient of an electric vehicle. This state transition equation establishes a mathematical closed loop between decision variables and individual physical states, making the optimization process feasible.

[0063] For stationary distributed resources, the preset internal cost function mainly considers the charging and discharging losses and operation and maintenance depreciation of distributed energy storage devices. The calculation formula for the internal cost function corresponding to stationary distributed resources is as follows: In the formula, This represents the total internal cost of a fixed distributed resource. and They represent the first The pre-set charging / discharging aging cost coefficient and operation and maintenance cost coefficient for each fixed distributed resource. Indicates assignment to the first Optimal scheduling instructions for a fixed distributed resource.

[0064] At this point, the optimal scheduling instruction described in the response delay compensation constraint satisfy: .

[0065] To ensure that the optimal scheduling instructions received by the underlying physical devices fully cover the expected equivalent capacity, a power allocation balance equation constraint is constructed. This equation constraint mandates that the sum of the optimal scheduling instructions allocated to all individual devices must equal the expected total scheduling demand of the aggregated resources. The specific mathematical expression is as follows: In the formula, This represents the expected total scheduling demand. When aggregated resources are ready to participate in the up-frequency adjustment response, When aggregated resources are ready to participate in the down-frequency response, This constraint serves as a hard equality boundary for the resource collaborative scheduling model, preventing trivial solutions where the overall output of aggregated resources is zero due to simply pursuing cost minimization.

[0066] By combining the predicted clearing price and a pre-defined internal cost function, a resource collaborative scheduling model objective function is constructed with the global goal of minimizing net cost. The objective function compares the total scheduling overhead with the direct benefits of participating in the frequency regulation market, guiding the control strategy to find the decision point with the maximum profit margin. The mathematical expression of the objective function is as follows: In the formula, This represents the net cost objective function.

[0067] After successfully constructing a complete mathematical scheduling model, the solution algorithm is executed. Under the premise of satisfying the above-mentioned stringent physical constraints, the solution algorithm continuously optimizes and finally outputs the optimal scheduling instruction that minimizes the net cost objective function.

[0068] By establishing a mathematical collaborative scheduling model that deeply integrates physical security constraints and dynamic market price signals, and performing optimization within a strict control rule framework, the potential grid oscillation risk caused by disordered responses from multiple stakeholders at the underlying level is completely eliminated, achieving a deep unification of physical security and scheduling economic benefits in large-scale frequency regulation tasks.

[0069] Step S5: Based on the minimum net cost, aggregated equivalent up-capacity, aggregated equivalent down-capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions, coordinate and control electric vehicles and fixed distributed resources.

[0070] In a preferred embodiment, coordinated control of electric vehicles and stationary distributed resources is performed based on minimum net cost, aggregated equivalent upsizing capacity, aggregated equivalent downsizing capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions, including: The total equivalent capacity of aggregation is determined based on the aggregate equivalent upward capacity and aggregate equivalent downward capacity. The unit equivalent capacity price is determined based on the minimum net cost and the aggregate equivalent total capacity; The target price is determined based on the unit equivalent capacity price, the predicted clearing price, frequency operation data, and the target equivalent frequency regulation coefficient; The aggregated equivalent upward capacity, aggregated equivalent downward capacity, and target price will be sent to the frequency modulation market for bidding, and the winning capacity will be determined. The capacity allocation coefficient is determined based on the winning bid capacity, the aggregate equivalent upward capacity, and the aggregate equivalent downward capacity. The optimal scheduling instruction is modified based on the capacity allocation coefficient to generate the actual execution instruction; Based on the actual execution instructions, the charging and discharging power of electric vehicles and the output power of stationary distributed resources are adjusted to achieve coordinated control of electric vehicles and stationary distributed resources.

[0071] Specifically, coordinated control is a crucial execution step that drives the underlying physical equipment to its actual power output based on the results of previous optimization solutions. Coordinated control is performed on electric vehicles and stationary distributed resources based on minimum net cost, aggregated equivalent upward capacity, aggregated equivalent downward capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions. This includes determining the aggregated equivalent total capacity based on the aggregated equivalent upward and downward capacity. The aggregated equivalent total capacity represents the overall schedulable capacity currently capable of participating in the frequency regulation market. The mathematical formula for calculating the aggregated equivalent total capacity is as follows: In the formula, This represents the equivalent total capacity of the aggregate.

[0072] The unit equivalent capacity price is determined based on the minimum net cost and the aggregated equivalent total capacity. The unit equivalent capacity price serves as the basic price boundary covering the cost of accessing the underlying physical resources. The mathematical expression for calculating the unit equivalent capacity price is as follows: In the formula, This indicates the price quoted per unit of equivalent capacity. This represents the minimum net cost.

[0073] After obtaining the basic price boundary, the target price is determined based on the unit equivalent capacity bid, the predicted clearing price, frequency operation data, and the target equivalent frequency regulation coefficient. The target price is calculated by deeply integrating the bid boundary value, the expected economic conditions of the frequency regulation market, and the operating status of the power system. The mathematical formula for calculating the target price is as follows.

[0074] In the formula, Indicates the target price. This represents the pre-set market strategy coefficient. This represents the pre-set frequency-sensitive gain coefficient. Frequency operation data includes frequency deviation data. The pre-set market strategy coefficient is a correction parameter pre-configured by operators based on historical bidding data. The main function of the pre-set market strategy coefficient is to appropriately reduce its value in a highly competitive market environment to improve the overall success rate of bidding for physical equipment. The pre-set frequency-sensitive gain coefficient is a fixed multiplier constant pre-set according to the power system frequency response requirements. The pre-set frequency-sensitive gain coefficient allows the final target price to increase synchronously with the degree of power system frequency deterioration and the enhancement of the underlying equipment's physical response capabilities.

[0075] The aggregated equivalent upward capacity, aggregated equivalent downward capacity, and target price are sent to the frequency regulation market for bidding, and the winning capacity is determined. The frequency regulation market performs unified clearing and settlement based on the bid data of all participants and issues the winning bid feedback results in real time. After obtaining the winning capacity, the capacity allocation coefficient is determined based on the winning capacity, aggregated equivalent upward capacity, and aggregated equivalent downward capacity. The mathematical expression of the capacity allocation coefficient is as follows: In the formula, Represents the capacity allocation factor. Indicates the winning bid capacity. This represents the matched equivalent capacity parameter determined based on the direction of the FM market directive. When the winning bid directive is for an upward adjustment, the matched equivalent capacity parameter equals the aggregated equivalent upward capacity; when the winning bid directive is for a downward adjustment, it equals the aggregated equivalent downward capacity. The capacity allocation coefficient objectively reflects the proportional conversion relationship between the actual capacity quota recognized in the FM market clearing and the expected reported capacity quota.

[0076] The optimal scheduling instruction is modified based on the capacity allocation coefficient to generate the actual execution instruction. The generation process involves proportionally scaling the optimal scheduling instruction using the capacity allocation coefficient. The corresponding mathematical formula is as follows: In the formula, This indicates the actual execution of the instruction.

[0077] Based on the actual execution instructions, the charging and discharging power of electric vehicles and the output power of stationary distributed resources are adjusted to achieve coordinated control of electric vehicles and stationary distributed resources. The coordination and management center sends the actual execution instructions one by one to the underlying hardware controllers such as smart charging piles and photovoltaic inverters through the underlying communication network, so that the underlying hardware controllers strictly adjust the real-time charging and discharging status and real-time active power output amplitude of the corresponding hardware according to the actual execution instructions.

[0078] By combining market economy clearing rules with the dynamic physical response capabilities of underlying equipment to generate a multi-dimensional bidding strategy, and issuing power execution instructions with allocation coefficient correction to hardware equipment after determining the actual winning bid amount, the risk of response delay failure of underlying hardware during instruction execution is effectively avoided, and precise closed-loop physical control of massive distributed energy resources in the power system under complex frequency regulation conditions is realized.

[0079] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0080] like Figure 2As shown, an embodiment of the present invention provides a coordinated control device for a power system considering electric vehicles and stationary distributed resources, including: a data acquisition module, a state vector determination module, a capacity processing module, a scheduling model solving module, and a coordinated control execution module; The data acquisition module is used to acquire frequency operation data of the power system, ambient temperature, predicted clearing price, charging and discharging power of electric vehicles, battery state of charge of electric vehicles, availability probability of electric vehicles, and output power of fixed distributed resources. The state vector determination module is used to determine the state vector based on frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources. The capacity processing module is used to input the state vector into a preset parameter dynamic adjustment model, so that the parameter dynamic adjustment model generates an equivalent model parameter adjustment amount according to the state vector; based on the equivalent model parameter adjustment amount, the preset initial equivalent model parameters are updated to generate target equivalent model parameters; wherein, the target equivalent model parameters include a target equivalent inertia constant, a target equivalent frequency modulation coefficient, and a target equivalent response delay; based on the target equivalent frequency modulation coefficient, the aggregated equivalent upward capacity and aggregated equivalent downward capacity are determined. The scheduling model solving module is used to construct the objective function and constraints of the resource collaborative scheduling model with the goal of minimizing net cost, based on the objective equivalent inertia constant, objective equivalent response delay, aggregate equivalent up-adjustment capacity, aggregate equivalent down-adjustment capacity, and predicted clearing price; under the constraints, the objective function of the resource collaborative scheduling model is solved to generate the minimum net cost and the optimal scheduling instruction. The coordination control execution module is used to coordinate and control electric vehicles and fixed distributed resources based on minimum net cost, aggregate equivalent up-capacity, aggregate equivalent down-capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions.

[0081] In a preferred embodiment, the data acquisition module obtains the predicted clearing price in the following manner: Obtain the frequency regulation capacity requirements of the power system; The frequency operation data and frequency modulation capacity requirements are input into a preset price prediction model, so that the price prediction model generates a predicted clearing price based on the frequency operation data and frequency modulation capacity requirements. The preset price prediction model is trained in the following way: Obtain several price prediction training samples; wherein, the price prediction model is trained from several price prediction training samples; each price prediction training sample includes historical frequency regulation capacity demand, historical frequency operation data, and clearing price label; the clearing price label is the historical clearing price corresponding to the historical frequency regulation capacity demand and historical frequency operation data; Each price prediction training sample is sequentially input into the price prediction model to train the price prediction model until a preset first training iteration is reached. Each time a price prediction training sample is received, the price prediction model outputs the predicted clearing price corresponding to that sample. A first loss function value is calculated based on the predicted clearing price and its corresponding clearing price label. The price prediction model is then updated based on the first loss function value.

[0082] In a preferred embodiment, the state vector determination module determines the state vector based on frequency operating data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources, including: Based on the charging and discharging power of electric vehicles and the output power of fixed distributed resources, determine the total available capacity for aggregated resources to be increased and the total available capacity for aggregated resources to be decreased. The overall effective state of charge of the electric vehicle is determined based on the battery state of charge and the availability probability of the electric vehicle. A state vector is constructed based on frequency operation data, ambient temperature, total available capacity for upward adjustment of aggregated resources, total available capacity for downward adjustment of aggregated resources, and comprehensive effective state of charge.

[0083] In a preferred embodiment, the capacity processing module trains a preset parameter dynamic adjustment model in the following manner: Acquire several parameter adjustment training samples; wherein, the parameter dynamic adjustment model is trained from several parameter adjustment training samples; each parameter adjustment training sample includes a historical state vector and a target adjustment amount label; the target adjustment amount label is the historical equivalent model parameter adjustment amount corresponding to the historical state vector. Each parameter adjustment training sample is sequentially input into the parameter dynamic adjustment model to train the model until a preset second training iteration is reached. The parameter dynamic adjustment model outputs the corresponding predicted equivalent model parameter adjustment amount each time it receives a parameter adjustment training sample. A second loss function value is calculated based on the predicted equivalent model parameter adjustment amount and the corresponding target adjustment amount label. The parameter dynamic adjustment model is then updated based on the second loss function value.

[0084] In a preferred embodiment, the capacity processing module includes the following adjustment amounts for the equivalent model parameters: equivalent inertia constant adjustment, equivalent frequency modulation coefficient adjustment, and equivalent response delay adjustment. The preset initial equivalent model parameters include: initial equivalent inertia constant, initial equivalent frequency modulation coefficient, and initial equivalent response delay; Based on the adjustment amount of the equivalent model parameters, the preset initial equivalent model parameters are updated to generate the target equivalent model parameters, including: The equivalent inertial constant adjustment is superimposed with the initial equivalent inertial constant to generate the target equivalent inertial constant; The equivalent frequency modulation coefficient adjustment is superimposed with the initial equivalent frequency modulation coefficient to generate the target equivalent frequency modulation coefficient; The equivalent response delay adjustment is added to the initial equivalent response delay to generate the target equivalent response delay; The target equivalent inertial constant, the target equivalent frequency modulation coefficient, and the target equivalent response delay are used as the target equivalent model parameters.

[0085] In a preferred embodiment, the capacity processing module determines the aggregated equivalent up-regulation capacity and aggregated equivalent down-regulation capacity based on the target equivalent frequency modulation coefficient, including: Based on the target equivalent frequency modulation coefficient, the total available capacity of aggregated resources is equivalently converted to generate the aggregated equivalent upward capacity. Based on the target equivalent frequency modulation coefficient, the total downgraded available capacity of aggregated resources is equivalently converted to generate the aggregated equivalent downgraded capacity.

[0086] In a preferred embodiment, the coordination control execution module performs coordinated control on electric vehicles and stationary distributed resources based on minimum net cost, aggregated equivalent up-capacity, aggregated equivalent down-capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions, including: The total equivalent capacity of aggregation is determined based on the aggregate equivalent upward capacity and aggregate equivalent downward capacity. The unit equivalent capacity price is determined based on the minimum net cost and the aggregate equivalent total capacity; The target price is determined based on the unit equivalent capacity price, the predicted clearing price, frequency operation data, and the target equivalent frequency regulation coefficient; The aggregated equivalent upward capacity, aggregated equivalent downward capacity, and target price will be sent to the frequency modulation market for bidding, and the winning capacity will be determined. The capacity allocation coefficient is determined based on the winning bid capacity, the aggregate equivalent upward capacity, and the aggregate equivalent downward capacity. The optimal scheduling instruction is modified based on the capacity allocation coefficient to generate the actual execution instruction; Based on the actual execution instructions, the charging and discharging power of electric vehicles and the output power of stationary distributed resources are adjusted to achieve coordinated control of electric vehicles and stationary distributed resources.

[0087] It should be noted that the embodiments of the device described above correspond to the embodiments of the present invention described above, and are capable of implementing the coordinated control method for power systems considering electric vehicles and fixed distributed resources as described in any one of the above embodiments of the present invention. Furthermore, the embodiments of the device described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Additionally, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between modules indicates that they have a communication connection, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without creative effort.

[0088] Based on the above-described method embodiments of the present invention, a corresponding embodiment of an electronic device is provided.

[0089] An embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the coordinated control method of a power system considering electric vehicles and stationary distributed resources as described in any one of the present invention, or, when the processor executes the computer program, it implements the functions of each module in the above-described device embodiments.

[0090] For example, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the terminal device.

[0091] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0092] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0093] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, applications required for at least one function, etc.; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0094] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments; Another embodiment of the present invention provides a storage medium including a stored computer program, wherein, when the computer program is executed, the device where the storage medium is located executes any of the above-described coordinated control methods of the present invention for power systems considering electric vehicles and stationary distributed resources.

[0095] The aforementioned storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0096] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0097] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A coordinated control method for a power system considering electric vehicles and stationary distributed resources, characterized in that, include: Acquire power system frequency operation data, ambient temperature, predicted clearing price, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources. The state vector is determined based on frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources. The state vector is input into a preset parameter dynamic adjustment model, so that the parameter dynamic adjustment model generates an equivalent model parameter adjustment amount based on the state vector; based on the equivalent model parameter adjustment amount, the preset initial equivalent model parameters are updated to generate target equivalent model parameters; wherein, the target equivalent model parameters include target equivalent inertia constant, target equivalent frequency modulation coefficient, and target equivalent response delay; based on the target equivalent frequency modulation coefficient, the aggregate equivalent up-adjustment capacity and aggregate equivalent down-adjustment capacity are determined; With the goal of minimizing net cost, the objective function and constraints of the resource collaborative scheduling model are constructed based on the objective equivalent inertia constant, objective equivalent response delay, aggregate equivalent up-adjustment capacity, aggregate equivalent down-adjustment capacity, and predicted clearing price. Under the constraints, the objective function of the resource collaborative scheduling model is solved to generate the minimum net cost and the optimal scheduling instruction. Based on minimum net cost, aggregated equivalent upsizing capacity, aggregated equivalent downsizing capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions, coordinated control is performed on electric vehicles and fixed distributed resources.

2. The coordinated control method for power systems considering electric vehicles and stationary distributed resources as described in claim 1, characterized in that, The predicted clearing price can be obtained through the following methods: Obtain the frequency regulation capacity requirements of the power system; The frequency operation data and frequency modulation capacity requirements are input into a preset price prediction model, so that the price prediction model generates a predicted clearing price based on the frequency operation data and frequency modulation capacity requirements. The preset price prediction model is trained in the following way: Obtain several price prediction training samples; wherein, the price prediction model is trained from several price prediction training samples; each price prediction training sample includes historical frequency regulation capacity demand, historical frequency operation data, and clearing price label; the clearing price label is the historical clearing price corresponding to the historical frequency regulation capacity demand and historical frequency operation data; Each price prediction training sample is sequentially input into the price prediction model to train the price prediction model until a preset first training iteration is reached. Each time a price prediction training sample is received, the price prediction model outputs the predicted clearing price corresponding to that sample. A first loss function value is calculated based on the predicted clearing price and its corresponding clearing price label. The price prediction model is then updated based on the first loss function value.

3. The coordinated control method for power systems considering electric vehicles and stationary distributed resources as described in claim 2, characterized in that, Based on frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources, the state vector is determined, including: Based on the charging and discharging power of electric vehicles and the output power of fixed distributed resources, determine the total available capacity for aggregated resources to be increased and the total available capacity for aggregated resources to be decreased. The overall effective state of charge of the electric vehicle is determined based on the battery state of charge and the availability probability of the electric vehicle. A state vector is constructed based on frequency operation data, ambient temperature, total available capacity for upward adjustment of aggregated resources, total available capacity for downward adjustment of aggregated resources, and comprehensive effective state of charge.

4. The coordinated control method for power systems considering electric vehicles and stationary distributed resources as described in claim 3, characterized in that, The model is trained by dynamically adjusting the preset parameters in the following way: Acquire several parameter adjustment training samples; wherein, the parameter dynamic adjustment model is trained from several parameter adjustment training samples; each parameter adjustment training sample includes a historical state vector and a target adjustment amount label; the target adjustment amount label is the historical equivalent model parameter adjustment amount corresponding to the historical state vector. Each parameter adjustment training sample is sequentially input into the parameter dynamic adjustment model to train the model until a preset second training iteration is reached. The parameter dynamic adjustment model outputs the corresponding predicted equivalent model parameter adjustment amount each time it receives a parameter adjustment training sample. A second loss function value is calculated based on the predicted equivalent model parameter adjustment amount and the corresponding target adjustment amount label. The parameter dynamic adjustment model is then updated based on the second loss function value.

5. The coordinated control method for a power system considering electric vehicles and stationary distributed resources as described in claim 4, characterized in that, The equivalent model parameter adjustment amounts include: equivalent inertia constant adjustment amount, equivalent frequency modulation coefficient adjustment amount, and equivalent response delay adjustment amount; The preset initial equivalent model parameters include: initial equivalent inertia constant, initial equivalent frequency modulation coefficient, and initial equivalent response delay; Based on the adjustment amount of the equivalent model parameters, the preset initial equivalent model parameters are updated to generate the target equivalent model parameters, including: The equivalent inertial constant adjustment is superimposed with the initial equivalent inertial constant to generate the target equivalent inertial constant; The equivalent frequency modulation coefficient adjustment is superimposed with the initial equivalent frequency modulation coefficient to generate the target equivalent frequency modulation coefficient; The equivalent response delay adjustment is added to the initial equivalent response delay to generate the target equivalent response delay; The target equivalent inertial constant, the target equivalent frequency modulation coefficient, and the target equivalent response delay are used as the target equivalent model parameters.

6. The coordinated control method for a power system considering electric vehicles and stationary distributed resources as described in claim 5, characterized in that, Based on the target equivalent frequency modulation coefficient, determine the aggregate equivalent up-modulation capacity and aggregate equivalent down-modulation capacity, including: Based on the target equivalent frequency modulation coefficient, the total available capacity of aggregated resources is equivalently converted to generate the aggregated equivalent upward capacity. Based on the target equivalent frequency modulation coefficient, the total downgraded available capacity of aggregated resources is equivalently converted to generate the aggregated equivalent downgraded capacity.

7. The coordinated control method for power systems considering electric vehicles and stationary distributed resources as described in claim 6, characterized in that, Based on minimum net cost, aggregated equivalent upsizing capacity, aggregated equivalent downsizing capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions, coordinated control is implemented for electric vehicles and stationary distributed resources, including: The total equivalent capacity of aggregation is determined based on the aggregate equivalent upward capacity and aggregate equivalent downward capacity. The unit equivalent capacity price is determined based on the minimum net cost and the aggregate equivalent total capacity; The target price is determined based on the unit equivalent capacity price, the predicted clearing price, frequency operation data, and the target equivalent frequency regulation coefficient; The aggregated equivalent upward capacity, aggregated equivalent downward capacity, and target price will be sent to the frequency modulation market for bidding, and the winning capacity will be determined. The capacity allocation coefficient is determined based on the winning bid capacity, the aggregate equivalent upward capacity, and the aggregate equivalent downward capacity. The optimal scheduling instruction is modified based on the capacity allocation coefficient to generate the actual execution instruction; Based on the actual execution instructions, the charging and discharging power of electric vehicles and the output power of stationary distributed resources are adjusted to achieve coordinated control of electric vehicles and stationary distributed resources.

8. A coordinated control device for a power system considering electric vehicles and stationary distributed resources, characterized in that, include: The system comprises a data acquisition module, a state vector determination module, a capacity processing module, a scheduling model solving module, and a coordination control execution module. The data acquisition module is used to acquire frequency operation data of the power system, ambient temperature, predicted clearing price, charging and discharging power of electric vehicles, battery state of charge of electric vehicles, availability probability of electric vehicles, and output power of fixed distributed resources. The state vector determination module is used to determine the state vector based on frequency operation data, ambient temperature, electric vehicle charging and discharging power, electric vehicle battery state of charge, electric vehicle availability probability, and output power of fixed distributed resources. The capacity processing module is used to input the state vector into a preset parameter dynamic adjustment model, so that the parameter dynamic adjustment model generates an equivalent model parameter adjustment amount according to the state vector; based on the equivalent model parameter adjustment amount, the preset initial equivalent model parameters are updated to generate target equivalent model parameters; wherein, the target equivalent model parameters include a target equivalent inertia constant, a target equivalent frequency modulation coefficient, and a target equivalent response delay; based on the target equivalent frequency modulation coefficient, the aggregated equivalent upward capacity and aggregated equivalent downward capacity are determined. The scheduling model solving module is used to construct the objective function and constraints of the resource collaborative scheduling model with the goal of minimizing net cost, based on the objective equivalent inertia constant, objective equivalent response delay, aggregate equivalent up-adjustment capacity, aggregate equivalent down-adjustment capacity, and predicted clearing price; under the constraints, the objective function of the resource collaborative scheduling model is solved to generate the minimum net cost and the optimal scheduling instruction. The coordination control execution module is used to coordinate and control electric vehicles and fixed distributed resources based on minimum net cost, aggregate equivalent up-capacity, aggregate equivalent down-capacity, predicted clearing price, frequency operation data, target equivalent frequency regulation coefficient, and optimal scheduling instructions.

9. An electronic device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the coordinated control method for a power system taking into account electric vehicles and stationary distributed resources as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the storage medium is located to perform the coordinated control method for a power system taking into account electric vehicles and stationary distributed resources as described in any one of claims 1 to 7.

Images