Electric vehicle cluster adjustable capacity evaluation method considering user response will

Abstract

The invention belongs to the technical field of fuel cells, and relates to an electric vehicle cluster adjustable capacity evaluation method considering user response willingness, which comprises the following steps: 1, establishing a unified energy storage model of electric vehicle EV individuals; 2, establishing a generalized energy storage (GES) resource general aggregation model of the electric vehicle EV cluster; 3, predicting an electric vehicle EV cluster aggregation model parameter set based on a bidirectional quantum long short-term memory neural network to obtain a theoretical demand response potential; 4, quantifying user side demand response willingness based on a TSK fuzzy system to obtain a demand response willingness coefficient; 5, the actual demand response potential is obtained by combining the data and the model driving result; according to the method, unified energy storage modeling is carried out on individuals by considering heterogeneity of the electric vehicles, and clusters of the electric vehicles are packaged into a generalized energy storage resource general aggregation model, so that the theoretical demand response potential of the electric vehicle cluster is accurately represented by a group of less parameters under the condition that the model precision is not influenced; and the actual demand response potential is obtained by considering the user response intention.

Description

Technical Field

[0001] The invention belongs to the technical field of fuel cells and relates to an adjustable capacity evaluation method for an electric vehicle cluster taking into account user response willingness. Background Art

[0002] As an important part of the new energy system, the new power system is developing rapidly. Distributed renewable energy, mainly wind power and photovoltaic power, is gradually replacing fossil energy to achieve large-scale development and utilization. While bringing clean and low-carbon, its strong randomness and volatility bring challenges to the safe and efficient operation of the power system. The "mobile energy storage" characteristics of electric vehicles make them an important demand response resource. As a new energy vehicle, its ownership continues to grow with the support of the country. It has huge demand response potential, can promote the consumption of renewable energy while ensuring user needs, and support the reliable and stable operation of the power system.

[0003] At present, the charging modes of electric vehicles (EVs) are divided into three types: fast charging, slow charging and battery swapping. Among them, the fast charging mode refers to the EV directly charging at the maximum charging power. Its charging power is uncontrollable and cannot provide flexibility. It belongs to conventional rigid load; the slow charging mode means that the EV adjusts the power size according to the actual needs of the power grid when time permits, and has the ability to load shift and reverse power supply. It belongs to mobile energy storage; the battery swapping mode means that the EV quickly replaces the battery to complete the charging, and the replaced battery participates in the dispatch in an orderly manner according to the needs of the power grid. However, the current battery swapping infrastructure is not yet perfect, and users have a low acceptance of the battery swapping mode, and the demand response potential is low. Therefore, the accurate evaluation of the actual adjustable capacity of the cluster of the EV slow charging mode can lay the foundation for the strong randomness and volatility of the stable new energy power generation mode.

[0004] However, the overall operating characteristics of the electric vehicle cluster and the willingness of users to participate in demand response are difficult to model effectively, and it is impossible to accurately evaluate the actual adjustable capacity of the electric vehicle cluster.

[0005] 1) The existing technical solutions do not consider the heterogeneity of electric vehicle clusters when aggregating them, and cannot establish an accurate aggregation model;

[0006] 2) The existing technical solutions do not consider the coupling relationship between different factors and various uncertainties when predicting the parameters of the electric vehicle cluster aggregation model, making it difficult to accurately evaluate its demand response potential;

[0007] 3) Existing technical solutions do not fully consider the user's demand response willingness when evaluating the adjustable capacity of electric vehicle clusters, and the evaluation results may not be feasible.

[0008] Therefore, an electric vehicle cluster adjustable capacity assessment method is urgently needed to calculate the demand response potential considering user response willingness, so as to support the safe and efficient operation of the power system. Summary of the invention

[0009] The technical solution adopted by the present invention to solve the technical problem is: a method for evaluating the adjustable capacity of an electric vehicle cluster considering the user's response willingness, comprising the following steps:

[0010] Step S1, establishing a unified energy storage model for individual electric vehicles EV;

[0011] Step S2, establishing a generalized aggregation model of generalized energy storage GES resources of electric vehicle EV clusters;

[0012] Step S3, predicting the electric vehicle EV cluster aggregation model parameter set based on the bidirectional quantum long short-term memory neural network to obtain the theoretical demand response potential;

[0013] Step S4, quantifying the user-side demand response willingness based on the TSK fuzzy system to obtain a demand response willingness coefficient;

[0014] Step S5: Combine the data and model-driven results to derive the corresponding potential of actual demand.

[0015] Preferably, in step S1, the general form of the individual model of the electric vehicle EV is expressed as:

[0016]

[0017] In formula (5), and They represent the grid-connected battery capacity of the i-th electric vehicle EV at time t and time t+1 respectively; and They represent the battery power of the i-th electric vehicle EV at the time of grid connection and grid disconnection respectively; and Respectively represent the on-grid and off-grid status of the i-th electric vehicle EV at time t and time t+1, which are Boolean variables, 0 represents off-grid and 1 represents on-grid; and They represent the charging and discharging efficiency of the i-th electric vehicle EV respectively; and They respectively represent the charging / discharging status of the i-th electric vehicle EV during the period from t to t+1, which are Boolean variables, 0 indicates that the electric vehicle EV is not charging / discharging, and 1 indicates that the electric vehicle EV is charging / discharging; and They respectively represent the charging / discharging power of the i-th electric vehicle EV in the period from t to t+1; Δt represents the unit time period of charging / discharging of the electric vehicle EV.

[0018] Preferably, in step S2, based on the unified energy storage model of individual electric vehicles EV, their clusters are encapsulated into a general aggregation model of generalized energy storage GES resources, and the model is:

[0019]

[0020] In formula (12), and They represent the power of generalized energy storage GES at time t and time t+1 respectively; P t GES,cha represents the charging power of the generalized energy storage GES from t to t+1; P t GES,dis represents the discharge power of the generalized energy storage GES during the period from t to t+1; represents the self-loss coefficient of the generalized energy storage GES during the period from t to t+1; represents the charging efficiency of the generalized energy storage GES during the period from t to t+1; It represents the inverse of the discharge efficiency of the generalized energy storage GES during the period from t to t+1; It represents the change in power caused by the connection and disconnection of electric vehicles EV to the grid during the period from t to t+1;

[0021] The demand response potential of an EV fleet is precisely characterized by a uniform set of parameters, which are as follows:

[0022]

[0023] In formula (16), the parameter α GES is the relationship between the power at the next moment and the power at the current moment, that is, the equivalent value of the power self-loss coefficient at the corresponding moment; parameter β GES is the equivalent value of the charging efficiency of all grid-connected electric vehicles EV at the corresponding moment; γ GES is the reciprocal of the equivalent value of the discharge efficiency of all grid-connected electric vehicles EV at the corresponding moment; δ GES is the sum of the power of off-grid and grid-connected electric vehicles EV at the corresponding moment; parameter and are the sum of the upper and lower limits of the charging / discharging power of all grid-connected electric vehicles EV at the corresponding moment; the parameters and They are the sum of the upper and lower limits of the power of all grid-connected electric vehicles (EV) at the corresponding moment.

[0024] Preferably, in step S3, the operating characteristics of the electric vehicle EV are divided into three categories: electric vehicle EV parameters, user travel plans and uncertainty factors; wherein the electric vehicle EV parameters include vehicle parameters and real-time power The user travel plan includes the expected plan, that is, the time when the electric vehicle (EV) user connects to the grid, the expected time when the user leaves the grid, and the power requirement Uncertain factors include: external environment, weather, season, date and time-of-use electricity price; a data-driven approach is used to predict the parameters of the electric vehicle EV cluster aggregation model.

[0025] Preferably, the data-driven method in step S3 uses a quantum machine learning method to predict parameters;

[0026] Step S3 includes the following sub-steps:

[0027] Step S3-1: Variational quantum circuit;

[0028] Quantum coding layer: encodes classical data into quantum states and converts the quantum bit state from the reference state |0 to the target state through the U(x) operation;

[0029] Variational layering: The V(θ) operation entangles the quantum bits and rotates them to the target state through multiple quantum gates with adjustable parameters;

[0030] Quantum measurement layer: measures the quantum state of some or all quantum bits based on quantum computing and outputs the calculation results of VQC;

[0031] Step S3-2: Quantum Long Short-Term Memory Neural Network

[0032] The classical neural network in the LSTM basic unit is replaced by VQC, thereby extending it to the quantum field. The main function of VQC is feature extraction and data compression; the calculation formula of QLSTM is as follows:

[0033] f t =σ(VQC1([h t-1 ,x t ])) (17)

[0034] i t =σ(VQC2([h t-1 ,x t ])) (18)

[0035] o t =σ(VQC4([h t-1 ,x t ])) (19)

[0036]

[0037]

[0038] h t =VQC5(o t *tanh(ct )) (twenty two)

[0039] y t =VQC6(o t *tanh(c t )) (twenty three)

[0040] In formulas (17) to (23), f t ,i t and t denote the output vectors of the forget gate, input gate, and output gate at time t respectively; σ denotes the sigmoid activation function, which maps the corresponding variables to the interval [0,1]; h t-1 represents the state vector of the hidden layer at time t-1; x t Represents the sequence input vector at time t; c% t represents the input unit state, i.e., the candidate state vector input to the memory unit at time t; tanh represents the hyperbolic tangent activation function; c t represents the state vector of the memory unit at time t; * represents the element-wise multiplication of the corresponding matrix; y t Represents the final output;

[0041] Step S3-3: Optimization process;

[0042] During the optimization process, the parameter offset rule is used to calculate the gradient of VQC, and the calculation expression is:

[0043]

[0044] In formula (24), x represents the input value; θ i represents the quantum parameters of the ith VQC circuit; f(·) represents the output with respect to the input and quantum parameters;

[0045] Iteratively optimize the classical-quantum hybrid computing model by minimizing the loss function by back-propagating gradients between the VQC and the classical neural network;

[0046] Step S3-4: Bidirectional quantum long short-term memory neural network

[0047] Select Bi-QLSTM as the framework for parameter prediction;

[0048] Bi-QLSTM consists of two QLSTMs with the same structure but opposite directions. The output of Bi-QLSTM contains past and future information in the input, and the weights are shared during the training process, which improves the generalization ability of the model without increasing the amount of data.

[0049] Preferably, in step S4, the demand response willingness is added to the operating characteristics of the electric vehicle EV, and the demand response willingness includes: incentive mechanism and user psychology; wherein the incentive mechanism is the benefit of the user participating in the demand response, that is, the product of the amount of electricity interacting with the power grid and the time-of-use electricity price; the user psychology is the charge state of the electric vehicle EV and the user's recognition of the demand response; the user demand response willingness coefficient is calculated by model driving, so as to obtain the actual adjustable capacity of the electric vehicle EV cluster.

[0050] More preferably, in step S4, the model drives the calculation of the user demand response willingness coefficient, and the TSK fuzzy system is used to quantify the user's demand response willingness;

[0051] Step S4 includes the following sub-steps:

[0052] The steps of the user-side demand response willingness quantification model based on TSK fuzzy system are as follows:

[0053] Step S4-1: Given a d-dimensional input vector x=[x1, x2, …, x d ] T The fuzzy subset A corresponding to the i-th input vector in the k-th fuzzy rule k,i ; Where d=3, and the input vectors represent the battery SOC, incentive price, and demand response recognition degree, respectively, and their corresponding fuzzy subsets are "low SOC", "medium SOC", "high SOC", "low price", "medium price", "high price" and "low degree", "medium degree", "high degree" respectively;

[0054] Step S4-2: Select Gaussian distribution as the membership function of the input vector, and fuzzify the battery SOC, incentive price and demand response recognition degree respectively, that is, use their corresponding membership function values ​​as the input of TSK fuzzy rules; Gaussian membership function can be expressed as:

[0055]

[0056] In formula (25), represents the Gaussian distribution membership function of the i-th dimension input vector in the k-th fuzzy rule; c k,i and σ k,i Respectively represent the mean and standard deviation of the corresponding membership function;

[0057] Step S4-3: Perform fuzzy reasoning based on the fuzzy subset of the input vector and the "IF THEN" fuzzy rule base to determine the fuzzy set of the output variable; wherein K=27, and its general expression is:

[0058]

[0059] In formula (26), The function expression representing the kth fuzzy rule; q k,i represents the parameter corresponding to the i-th dimension input vector in the k-th fuzzy rule; q k,0 represents the bias parameter of the kth fuzzy rule;

[0060] Step S4-4: Calculate the final response willingness function according to the fuzzy set of output variables; the demand response willingness output of the TSK fuzzy system after defuzzification is expressed as:

[0061]

[0062] In formulas (27), (28), and (29), W EV It represents the quantitative results of the demand response willingness of electric vehicle (EV) users; and They respectively represent the membership function and normalized membership function corresponding to the kth fuzzy rule.

[0063] The beneficial effects of the present invention are:

[0064] 1. The present invention considers the heterogeneity of electric vehicles to uniformly model the energy storage of individuals, and encapsulates their clusters into a generalized aggregation model of generalized energy storage resources, thereby accurately characterizing the demand response potential of electric vehicle clusters with a smaller set of parameters without affecting the accuracy of the model.

[0065] 2. When predicting the parameters of the electric vehicle cluster aggregation model, the present invention takes into account the characteristics of massive high-dimensional, interactively coupled, dynamically changing and highly nonlinear operating data, adopts the Bi-QLSTM model to better represent the potential structure of high-dimensional data in the Hilbert space, and utilizes the superposition, entanglement and parallel characteristics of quantum states to enhance the learning and characterization capabilities of the complex relationships of coupled data and accelerate the training process, thereby improving the overall performance of the model.

[0066] 3. When evaluating the adjustable capacity of an electric vehicle cluster, the present invention fully considers the impact of the battery state of charge of the electric vehicle, the incentive price and the user's demand response recognition on the demand response willingness, selects the TSK fuzzy system with excellent nonlinear approximation ability, and accurately quantifies the user's demand response willingness. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 is a schematic diagram of an electric vehicle cluster adjustable capacity evaluation method considering user response willingness according to the present invention;

[0068] Figure 2 It is a general architecture diagram of the variational quantum circuit of the present invention;

[0069] Figure 3It is a variational quantum circuit architecture diagram of the QLSTM of the present invention;

[0070] Figure 4 It is a QLSTM neural network structure diagram of the present invention;

[0071] Figure 5 It is a Bi-QLSTM neural network structure diagram of the present invention;

[0072] Figure 6 It is the structural diagram of the TSK fuzzy system of the present invention. DETAILED DESCRIPTION

[0073] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the relevant technologies in the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] like Figures 1 to 6 As shown, this embodiment provides an electric vehicle cluster adjustable capacity evaluation method taking into account user response willingness.

[0075] 1. Unified energy storage model for individual electric vehicles

[0076] The charging mode of EV in the present invention is slow charging.

[0077] Consider a cluster of N EVs. Due to the differences in the grid-connected periods of different EVs, the definition domains of different individuals are heterogeneous. The specific form can be expressed as:

[0078]

[0079] Three types of changes in the battery power of the grid-connected EV are considered. First, at the moment when the EV is connected to the grid, the change in the grid-connected battery power can be expressed as:

[0080]

[0081] in, and They represent the grid-connected battery power of the i-th EV before and after the grid-connected moment; dt represents the very short time period when the EV is connected to or off the grid; Represents the battery capacity of the i-th EV at the time of grid connection.

[0082] Then, during the normal EV grid-connected period, the change in grid-connected battery power can be expressed as:

[0083]

[0084] in, and They represent the grid-connected battery capacity of the i-th EV at time t and time t+1 respectively; and They represent the charging and discharging efficiency of the i-th EV respectively; and They represent the charging / discharging status of the i-th EV from t to t+1, respectively. They are Boolean variables, where 0 means the EV is not charging / discharging, and 1 means the EV is charging / discharging. and They represent the charging / discharging power of the i-th EV from t to t+1 respectively; Δt represents the unit time period of EV charging / discharging.

[0085] Finally, when the EV is off-grid, the change in the grid-connected battery power can be expressed as:

[0086]

[0087] in, and They represent the grid-connected battery power of the i-th EV before and after the off-grid moment; Represents the battery capacity of the i-th EV at the time of being off-grid.

[0088] Therefore, the general form of the EV individual model can be expressed as:

[0089]

[0090]

[0091] in, and Respectively represent the on-grid and off-grid status of the i-th EV at time t and time t+1, which are Boolean variables, 0 represents off-grid and 1 represents on-grid; and They represent the upper / lower limits of the charging power of the i-th EV respectively; and They represent the upper / lower limits of the discharge power of the i-th EV respectively; and They represent the upper / lower limits of the battery power of the i-th EV, It represents the expected battery capacity of the i-th EV when it is off-grid.

[0092] 2. Generalized Aggregation Model of Generalized Energy Storage Resources for Electric Vehicle Clusters

[0093] Based on the unified energy storage model of EV individuals, its cluster is encapsulated as a general aggregation model of generalized energy storage (GES) resources. The model is as follows:

[0094]

[0095] in, and Respectively represent the power of GES at time t and time t+1; P t GES,cha represents the charging power of GES from t to t+1; P t GES,dis represents the discharge power of GES during the period from t to t+1; It represents the power self-loss coefficient of GES in the period from t to t+1; represents the charging efficiency of GES from t to t+1; It represents the inverse of the discharge efficiency of GES during the period from t to t+1; It indicates the change in the amount of electricity caused by EV connection and disconnection from the GES during the period from t to t+1; and They represent the upper / lower limits of the charging power of GES in the period from t to t+1 respectively; and They represent the upper / lower limit of the discharge power of GES in the period from t to t+1 respectively; and Indicates the upper / lower limit of the power of GES during the period from t to t+1.

[0096] Therefore, the demand response potential of the EV cluster can be accurately characterized by a unified set of parameters. This aggregation method can characterize the demand response potential of EVs with fewer parameters without affecting the accuracy of the model, and the EV cluster can be called and controlled in a unified way. The parameter set is as follows:

[0097]

[0098] According to the individual model and cluster aggregation model of EV, the parameter α GES is the relationship between the power at the next moment and the power at the current moment, that is, the equivalent value of the power self-loss coefficient at the corresponding moment; parameter β GES is the equivalent value of the charging efficiency of all grid-connected EVs at the corresponding moment; γ GES is the inverse of the equivalent value of the discharge efficiency of all grid-connected EVs at the corresponding moment; δ GES is the sum of the power of off-grid and grid-connected EVs at the corresponding moment; parameter and are the sum of the upper and lower limits of the charging / discharging power of all grid-connected EVs at the corresponding moment; the parameters and They are the sum of the upper and lower limits of the power consumption of all grid-connected EVs at the corresponding moment.

[0099] 3. Evaluation method of adjustable capacity of electric vehicle cluster driven by bidirectional quantum long-short-term memory considering user response willingness

[0100] Accurate evaluation of the adjustable capacity of EV clusters is very important for the power dispatching department to formulate a reasonable dispatching plan according to the actual operation of the power grid. Therefore, the present invention also proposes a bidirectional quantum long short-term memory (Bidirectional Quantum Long Short-Term Memory, Bi-QLSTM) driven electric vehicle cluster adjustable capacity evaluation method considering user response willingness. This method comprehensively evaluates the actual demand response potential of EV clusters based on the operating characteristics of EV clusters and user willingness through model-data hybrid drive.

[0101] The theoretical demand response potential of an EV cluster can be accurately characterized by a set of unified parameters (Equation 16) of its GES resource general aggregation model (Equation 12). According to the physical meaning of the parameter set, it is closely related to the operating characteristics of EVs. The present invention divides EV operating characteristics into three categories: EV parameters, user travel plans, and uncertainty factors. Among them, EV parameters include vehicle parameters and real-time power The user travel plan includes the expected plan, i.e. the EV user's grid connection time, expected grid disconnection time and power requirements At the same time, various uncertain factors are considered, such as the actual plan, i.e., the actual off-grid time of EV users, and the external environment, i.e., weather, season, date, and time-of-use electricity price. Since the operating characteristics of EV clusters are complex and difficult to accurately model, the present invention adopts a data-driven method to predict the parameters of the EV cluster aggregation model.

[0102] In addition, the adjustable capacity obtained by predicting the aggregation model parameters in the data-driven link does not take into account the user's willingness to respond to demand, and the result is the theoretical adjustable capacity of the EV cluster. In order to accurately characterize the impact of the user's willingness to respond to demand on the adjustable capacity, the present invention considers the influence of two major factors: incentive mechanism and user psychology. Among them, the incentive mechanism is the benefit of users participating in demand response, that is, the product of the amount of electricity interacting with the grid and the time-of-use electricity price; user psychology is the charge state of the EV and the user's recognition of demand response. Since the factors that affect the user's willingness to respond to demand are relatively clear, the present invention calculates the user's willingness to respond to demand through model-driven calculation, thereby obtaining the actual adjustable capacity of the EV cluster.

[0103] 3.1. Aggregate model parameter prediction model based on bidirectional quantum long short-term memory

[0104] EV operation characteristics are typical time series data, and the demand response potential of EV clusters is closely related to the nonlinear characteristics in time. Deep learning models can fully exploit the underlying features of data by utilizing the nonlinear transformation of hidden layers, extracting abstract and easily distinguishable high-level features, and thus have excellent feature representation, data fitting and generalization capabilities. Among them, the Long Short-Term Memory (LSTM) neural network, as an improved model of the Recurrent Neural Network (RNN), introduces a gate control unit based on the self-feedback of the RNN. It filters redundant and misleading features in the time series through its unique memory and forgetting mode, effectively controls the gradient flow, and thus achieves accurate prediction of time series of arbitrary length. However, in actual applications, the operation data of EV clusters presents the characteristics of massive high-dimensionality, interactive coupling, dynamic changes and high nonlinearity. The LSTM model may find it difficult to represent and learn the complex relationships in the data, resulting in unstable training results, which in turn affects the performance of the model.

[0105] As an emerging computing paradigm, quantum computing has powerful parallel capabilities and is considered to be an effective way to improve the speed and performance of classical machine learning. Therefore, the data-driven link of the present invention adopts quantum machine learning methods for parameter prediction. The hardware technology in the current quantum computing era is in the noisy medium-scale quantum stage, characterized by a limited number of quantum bits and a short decoherence time. Variational quantum algorithms have been proven to have natural noise resistance and can even benefit from noise, which is particularly suitable for recent quantum devices. Therefore, a quantum-classical hybrid computing framework is built through a variational quantum circuit (VQC) to realize quantum long short-term memory (Quantum Long Short-Term Memory, QLSTM).

[0106] The core of the quantum-classical hybrid computing framework is to transform the computing task into an optimization problem. The algorithm combines the computing power of quantum computers with the optimization methods of classical computers. It processes data through VQC on the quantum computer and calculates the loss function through quantum measurement. It optimizes the adjustable parameter θ on the classical computer and completes the computing task when the loss function is minimized / maximum.

[0107] QLSTM can better represent the potential structure of high-dimensional data in Hilbert space, and by utilizing the superposition, entanglement and parallel characteristics of quantum states, it can reduce the number of training parameters and accelerate the training process while improving the ability to learn and represent complex relationships in data, thereby improving the overall performance of the model. Therefore, the QLSTM model can more accurately predict the adjustable capacity of EV clusters and provide a basis for the scheduling and management of EVs.

[0108] 1) Variational quantum circuits

[0109] Quantum circuit is a kind of quantum computing model, which consists of multiple quantum bits, quantum gates and quantum measurement modules. Its computing process is that multiple quantum bits store input data in the form of quantum information, then operate the state of the quantum bits through a series of quantum gates, and finally convert it into classical information through the quantum measurement link and output the calculation result. VQC is a special type of quantum circuit, which is characterized by variable parameters and can be continuously optimized by iteration. Its architecture is as follows Figure 2 , 3 As shown, it mainly consists of three parts.

[0110] Quantum coding layer: Encodes classical data into quantum states and converts the quantum bit state from the reference state |0 to the target state through the U(x) operation. Usually one quantum bit encodes one classical input feature.

[0111] Variational layer: The V(θ) operation entangles the quantum bits and rotates them to the target state through multiple quantum gates with adjustable parameters. This part is the actual learnable part of VQC, and the parameters are continuously optimized during the iteration process.

[0112] Quantum measurement layer: Based on quantum computing, the quantum state of some or all quantum bits is measured and the calculation result of VQC is output. The result of this part of the measurement is the expected value of each quantum bit.

[0113] 2) Quantum long short-term memory neural network

[0114] The present invention replaces the classical neural network in the LSTM basic unit with VQC, thereby extending it to the quantum field. The main function of VQC is feature extraction and data compression. The basic structure of QLSTM is as follows Figure 4 The calculation formula is as follows:

[0115] f t =σ(VQC1([h t-1 ,x t ])) (17)

[0116] i t =σ(VQC2([h t-1 ,x t ])) (18)

[0117] o t =σ(VQC4([h t-1 ,x t ])) (19)

[0118]

[0119] ht =VQC5(o t *tanh(c t )) (twenty two)

[0120] y t =VQC6(o t *tanh(c t )) (twenty three)

[0121] Among them, f t ,i t and t denote the output vectors of the forget gate, input gate, and output gate at time t respectively; σ denotes the sigmoid activation function, which maps the corresponding variables to the interval [0,1]; h t-1 represents the state vector of the hidden layer at time t-1; x t Represents the sequence input vector at time t; represents the input unit state, i.e., the candidate state vector input to the memory unit at time t; tanh represents the hyperbolic tangent activation function; c t represents the state vector of the memory unit at time t; * represents the element-wise multiplication of the corresponding matrix; y t Represents the final output.

[0122] 3) Optimization process

[0123] During the optimization process, the present invention uses the parameter offset rule to calculate the gradient of VQC, and the calculation expression is:

[0124]

[0125] Among them, x represents the input value; θ i represents the quantum parameters of the ith VQC circuit; f(·) represents the output with respect to the input and quantum parameters.

[0126] Therefore, the classical-quantum hybrid computing model can be iteratively optimized by minimizing the loss function by back-propagating gradients between the VQC and the classical neural network.

[0127] 4) Bidirectional quantum long short-term memory neural network

[0128] The operating characteristics of the EV cluster depend not only on historical data, but also on future data, and have the properties of a bidirectional time series. In order to better learn and represent the complex relationships in the data series, the present invention selects Bi-QLSTM as the framework for parameter prediction.

[0129] Bi-QLSTM consists of two QLSTMs with the same structure and opposite directions. Figure 5As shown in the figure. The basic idea of ​​this method is to process the input through the forward and reverse QLSTMs respectively, and then concatenate the bidirectional processing results to obtain the final output. The output of Bi-QLSTM contains the past and future information in the input, and the weights are shared during the training process, which improves the generalization ability of the model without increasing the amount of data.

[0130] 3.2. Quantification model of user-side demand response willingness based on TSK fuzzy system

[0131] Due to its excellent nonlinear approximation ability, the TSK fuzzy system is widely used in the fields of system identification, pattern recognition and data processing. The present invention considers the influence of EV battery state of charge (SOC), incentive price and user demand response recognition on demand response willingness, and selects this method to quantify the user's demand response willingness, thereby evaluating the actual adjustable capacity of the EV cluster. Figure 6 It can be seen that the steps of the user-side demand response willingness quantification model based on the TSK fuzzy system are:

[0132] Step 1: First, given a d-dimensional input vector x = [x1, x2, …, x d ] T The fuzzy subset A corresponding to the i-th input vector in the k-th fuzzy rule k,i In the present invention, d=3, and the input vectors represent the battery SOC, incentive price, and demand response recognition degree, respectively, and their corresponding fuzzy subsets are "low SOC", "medium SOC", "high SOC", "low price", "medium price", "high price" and "low degree", "medium degree", "high degree", respectively.

[0133] Step 2: Select Gaussian distribution as the membership function of the input vector, and fuzzify the battery SOC, incentive price and demand response recognition degree respectively, that is, use their corresponding membership function values ​​as the input of TSK fuzzy rules. The Gaussian membership function can be expressed as:

[0134]

[0135] in, represents the Gaussian distribution membership function of the i-th dimension input vector in the k-th fuzzy rule; c k,i and σ k,i represent the mean and standard deviation of the corresponding membership function respectively.

[0136] Step 3: Perform fuzzy reasoning based on the fuzzy subsets of the input vector and the "IF THEN" fuzzy rule base to determine the fuzzy set of the output variable. According to the dimension of the input vector and the number of fuzzy subsets, the number of inference rules of the TSK fuzzy system in the present invention is K=27, and its general expression is:

[0137]

[0138] in, The function expression representing the kth fuzzy rule; q k,i represents the parameter corresponding to the i-th dimension input vector in the k-th fuzzy rule; q k,0 Represents the bias parameter of the kth fuzzy rule.

[0139] Step 4: Calculate the final response willingness function based on the fuzzy set of output variables. The demand response willingness output of the TSK fuzzy system after defuzzification can be expressed as:

[0140]

[0141] Among them, W EV It represents the quantitative results of EV users’ demand response willingness; and They respectively represent the membership function and normalized membership function corresponding to the kth fuzzy rule.

[0142] This implementation method considers the heterogeneity of electric vehicles to uniformly model individual energy storage, and accordingly establishes a generalized energy storage resource aggregation model;

[0143] This implementation method establishes an electric vehicle aggregation model parameter prediction model based on the Bi-QLSTM model, which uses the superposition, entanglement and parallel characteristics of quantum states to better characterize and learn the ability of complex relationships in data and accelerate training efficiency, thereby improving the overall performance of the model;

[0144] This implementation method establishes a user-side demand response willingness quantification model based on the TSK fuzzy system, thereby achieving accurate quantification of user demand response willingness.

[0145] In summary, the present invention considers the heterogeneity of electric vehicles to uniformly model individual energy storage, and encapsulates their clusters into a generalized energy storage resource general aggregation model, thereby accurately characterizing the demand response potential of electric vehicle clusters with a small set of parameters without affecting the accuracy of the model. Therefore, the present invention has broad application prospects in the strong randomness and volatility of stable new energy power generation modes.

[0146] It should be emphasized that the above are only preferred embodiments of the present invention and do not limit the present invention in any form. Any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A method for evaluating the adjustable capacity of an electric vehicle cluster considering user response willingness, characterized in that: The following steps are involved: Step S1, establishing a unified energy storage model for individual electric vehicles EV; Step S2, establishing a generalized aggregation model of generalized energy storage GES resources of electric vehicle EV clusters; Step S3, predicting the electric vehicle EV cluster aggregation model parameter set based on the bidirectional quantum long short-term memory neural network to obtain the theoretical demand response potential; Step S4, quantifying the user-side demand response willingness based on the TSK fuzzy system to obtain a demand response willingness coefficient; Step S5: Combining data and model-driven results to derive actual demand response potential.

2. The method for evaluating the adjustable capacity of an electric vehicle cluster considering user response willingness according to claim 1, characterized in that: In step S1, the general form of the individual model of the electric vehicle EV is expressed as: In formula (5), and They represent the grid-connected battery capacity of the i-th electric vehicle EV at time t and time t+1 respectively; and They represent the battery power of the i-th electric vehicle EV at the time of grid connection and grid disconnection respectively; and Respectively represent the on-grid and off-grid status of the i-th electric vehicle EV at time t and time t+1, which are Boolean variables, 0 represents off-grid and 1 represents on-grid; and They represent the charging and discharging efficiency of the i-th electric vehicle EV respectively; and They respectively represent the charging / discharging status of the i-th electric vehicle EV during the period from t to t+1, which are Boolean variables, 0 indicates that the electric vehicle EV is not charging / discharging, and 1 indicates that the electric vehicle EV is charging / discharging; and They respectively represent the charging / discharging power of the i-th electric vehicle EV in the period from t to t+1; Δt represents the unit time period of charging / discharging of the electric vehicle EV.

3. The method for evaluating the adjustable capacity of an electric vehicle cluster considering user response willingness according to claim 1, characterized in that: In step S2, based on the unified energy storage model of individual electric vehicles EV, the cluster is encapsulated into a general aggregation model of generalized energy storage GES resources, and the model is: In formula (12), and They represent the power of generalized energy storage GES at time t and time t+1 respectively; P t GES,cha represents the charging power of the generalized energy storage GES from t to t+1; P t GES,dis represents the discharge power of the generalized energy storage GES during the period from t to t+1; represents the self-loss coefficient of the generalized energy storage GES during the period from t to t+1; represents the charging efficiency of the generalized energy storage GES during the period from t to t+1; It represents the inverse of the discharge efficiency of the generalized energy storage GES during the period from t to t+1; It represents the change in power caused by the connection and disconnection of electric vehicles EV to the grid during the period from t to t+1; The demand response potential of an EV fleet is precisely characterized by a uniform set of parameters, which are as follows: In formula (16), the parameter α GES is the relationship between the power at the next moment and the power at the current moment, that is, the equivalent value of the power self-loss coefficient at the corresponding moment; parameter β GES is the equivalent value of the charging efficiency of all grid-connected electric vehicles EV at the corresponding moment; γ GES is the reciprocal of the equivalent value of the discharge efficiency of all grid-connected electric vehicles EV at the corresponding moment; δ GES is the sum of the power of off-grid and grid-connected electric vehicles EV at the corresponding moment; parameter and are the sum of the upper and lower limits of the charging / discharging power of all grid-connected electric vehicles EV at the corresponding moment; the parameters and They are the sum of the upper and lower limits of the power of all grid-connected electric vehicles (EV) at the corresponding moment.

4. The method for evaluating the adjustable capacity of an electric vehicle cluster considering user response willingness according to claim 1, characterized in that: In step S3, the operating characteristics of the electric vehicle EV are divided into three categories: electric vehicle EV parameters, user travel plans and uncertainty factors; wherein the electric vehicle EV parameters include vehicle parameters and real-time power The user travel plan includes an expected plan, i.e., the time when the electric vehicle (EV) user connects to the grid, the expected time when the user disconnects from the grid, and the power requirement The uncertain factors include: external environment, weather, season, date and time-of-use electricity price; a data-driven method is used to predict the parameters of the electric vehicle EV cluster aggregation model.

5. The method for evaluating the adjustable capacity of an electric vehicle cluster considering user response willingness according to claim 4, characterized in that: The data-driven method in step S3 uses quantum machine learning methods to predict parameters; The step S3 comprises the following sub-steps: Step S3-1: Variational quantum circuit; Quantum coding layer: encodes classical data into quantum states and converts the quantum bit state from the reference state |0 to the target state through the U(x) operation; Variational layering: The V(θ) operation entangles the quantum bits and rotates them to the target state through multiple quantum gates with adjustable parameters; Quantum measurement layer: measures the quantum state of some or all quantum bits based on quantum computing and outputs the calculation results of VQC; Step S3-2: Quantum Long Short-Term Memory Neural Network The classical neural network in the LSTM basic unit is replaced by VQC, thereby extending it to the quantum field. The main function of VQC is feature extraction and data compression; the calculation formula of QLSTM is as follows: f t =σ(VQC1([h t-1 ,x t ])) (17) I t =σ(VQC2([h t-1 ,x t ])) (18) the t =σ(VQC4([h t-1 ,x t ])) (19) h t =VQC5(o t *tanh(c t )) (22) y t =VQC6(o t *tanh(c t )) (23) In formulas (17) to (23), f t ,i t and t denote the output vectors of the forget gate, input gate, and output gate at time t respectively; σ denotes the sigmoid activation function, which maps the corresponding variables to the interval [0,1]; h t-1 represents the state vector of the hidden layer at time t-1; x t Represents the sequence input vector at time t; represents the input unit state, i.e., the candidate state vector input to the memory unit at time t; tanh represents the hyperbolic tangent activation function; c t represents the state vector of the memory unit at time t; * represents the element-wise multiplication of the corresponding matrix; y t Represents the final output; Step S3-3: Optimization process; During the optimization process, the parameter offset rule is used to calculate the gradient of VQC, and the calculation expression is: In formula (24), x represents the input value; θ i represents the quantum parameters of the ith VQC circuit; f(·) represents the output with respect to the input and quantum parameters; Iteratively optimize the classical-quantum hybrid computing model by minimizing the loss function by back-propagating gradients between the VQC and the classical neural network; Step S3-4: Bidirectional quantum long short-term memory neural network Select Bi-QLSTM as the framework for parameter prediction; Bi-QLSTM consists of two QLSTMs with the same structure but opposite directions. The output of Bi-QLSTM contains past and future information in the input, and the weights are shared during the training process, which improves the generalization ability of the model without increasing the amount of data.

6. The method for evaluating the adjustable capacity of an electric vehicle cluster considering user response willingness according to claim 1, characterized in that: In the step S4, the demand response willingness is added to the operating characteristics of the electric vehicle EV, and the demand response willingness includes: incentive mechanism and user psychology; wherein the incentive mechanism is the benefit of the user participating in the demand response, that is, the product of the amount of electricity interacting with the power grid and the time-of-use electricity price; the user psychology is the charge state of the electric vehicle EV and the user's recognition of the demand response; the user demand response willingness coefficient is calculated by model driving, so as to obtain the actual adjustable capacity of the electric vehicle EV cluster.

7. The method for evaluating the adjustable capacity of an electric vehicle cluster considering user response willingness according to claim 6, characterized in that: In the step S4, the model drives the calculation of the user demand response willingness coefficient, and uses the TSK fuzzy system to quantify the user's demand response willingness; The step S4 comprises the following sub-steps: The steps of the user-side demand response willingness quantification model based on TSK fuzzy system are as follows: Step S4-1: Given a d-dimensional input vector x=[x1, x2, …, x d ] T The fuzzy subset A corresponding to the i-th input vector in the k-th fuzzy rule k,i ; Where d=3, and the input vectors represent the battery SOC, incentive price and demand response recognition degree respectively, and their corresponding fuzzy subsets are "low SOC", "medium SOC", "high SOC", "low price", "medium price", "high price" and "low degree", "medium degree", "high degree" respectively; Step S4-2: Select Gaussian distribution as the membership function of the input vector, and fuzzify the battery SOC, incentive price and demand response recognition degree respectively, that is, use their corresponding membership function values ​​as the input of TSK fuzzy rules; Gaussian membership function can be expressed as: In formula (25), represents the Gaussian distribution membership function of the i-th dimension input vector in the k-th fuzzy rule; c k,i and σ k,i Respectively represent the mean and standard deviation of the corresponding membership function; Step S4-3: Perform fuzzy reasoning based on the fuzzy subset of the input vector and the "IF THEN" fuzzy rule base to determine the fuzzy set of the output variable; wherein K=27, and its general expression is: In formula (26), The function expression representing the kth fuzzy rule; q k,i represents the parameter corresponding to the i-th dimension input vector in the k-th fuzzy rule; q k,0 represents the bias parameter of the kth fuzzy rule; Step S4-4: Calculate the final response willingness function according to the fuzzy set of output variables; the demand response willingness output of the TSK fuzzy system after defuzzification is expressed as: In formulas (27), (28), and (29), W EV It represents the quantitative results of the demand response willingness of electric vehicle (EV) users; and They respectively represent the membership function and normalized membership function corresponding to the kth fuzzy rule.

Images