A method and apparatus for assessing and optimizing the allocation of electric vehicle charging resources based on behavior perception.

Abstract

This invention discloses a behavior-aware method for assessing and optimizing the allocation of electric vehicle charging resources. The method includes: constructing a Logistic response probability model for each cluster based on electric vehicle charging behavior clustering results, considering two driving factors: flexibility margin and charging urgency, to obtain the response probability of each cluster; assessing the physical adjustability potential of each cluster from three dimensions: dynamic online scale, upper and lower bounds of base power, and response reliability, to obtain the response reliability index of each cluster; designing a differentiated incentive mechanism based on the response probability and response reliability index of each cluster, establishing a linear programming quota optimization model under budget and diversity constraints, solving for the optimal contracted quota, and generating a feasible aggregated power domain for connecting to the power grid dispatch center. This invention refines the response capabilities of each group based on behavioral characteristics, achieving differentiated and optimized resource allocation under budget constraints.

Description

Technical Field

[0001] This invention belongs to the field of power demand-side management and electric vehicle aggregation and scheduling technology, specifically relating to a method and device for assessing and optimizing the allocation of electric vehicle charging resource potential based on behavior perception. Background Technology

[0002] With the continuous increase in the penetration rate of electric vehicles, the challenges posed by large-scale disorderly charging to the safe operation of the power distribution network are becoming increasingly prominent. As a key intermediary connecting users and the power grid, charging aggregation operators play an important role in promoting orderly charging of electric vehicles and developing demand-side flexible resources. In order to effectively organize and utilize dispersed charging resources, aggregation operators need to accurately assess the adjustable charging potential of various user groups and achieve optimal resource allocation within a limited incentive budget.

[0003] However, existing research still has the following shortcomings in the assessment and allocation of charging resource potential:

[0004] First, existing methods generally treat electric vehicle charging load as a homogeneous adjustable resource, failing to fully consider the significant differences among different user groups in terms of charging duration, dwell time, initial state of charge, and willingness to participate. This homogeneous approach is prone to overestimating adjustability or making scheduling schemes difficult to implement due to neglecting user behavior constraints.

[0005] Secondly, traditional methods typically use the vehicle's current state of charge as the sole constraint to determine the lower limit of power when assessing the adjustable power range. This fails to adequately consider the user's departure time constraints and subsequent travel needs, leading to an overestimation of the lower limit of power, resulting in a collapse of the adjustable space and limiting the actual availability of flexible resources.

[0006] Third, existing charging resource allocation methods lack probabilistic modeling of users' willingness to respond, making it difficult to distinguish between highly flexible groups and highly urgent groups when participating in demand response. This results in a lack of targeted allocation of incentive costs and reduces the efficiency of resource allocation per unit cost. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a method and apparatus for assessing and optimizing the allocation of electric vehicle charging resources based on behavior perception. It refines the response capabilities of various groups based on behavioral characteristics, and achieves differentiated optimization of resource allocation under budget constraints.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0009] Firstly, this paper provides a behavior-aware method for assessing and optimizing the allocation of electric vehicle charging resources. This method includes: based on electric vehicle charging behavior clustering results, considering two driving factors—flexibility margin and charging urgency—a Logistic response probability model is constructed for each cluster to obtain the response probability of each cluster; based on the obtained response probabilities, the stability of adjustable resource delivery for each cluster is evaluated by combining dynamic online vehicle scale, baseline power upper and lower bounds, dwell time fluctuations, and state of charge fluctuations, resulting in a response reliability index characterizing the actual available capacity; based on the response probabilities and response reliability index of each cluster, a differentiated incentive mechanism is designed, a linear programming quota optimization model under budget and diversity constraints is established, the optimal contracted quota is solved, and a feasible aggregated power domain for connecting to the power grid dispatch center is generated.

[0010] Furthermore, the flexibility margin is defined as the difference between the remaining time after arrival and before departure and the minimum necessary charging time, used to reflect the user's acceptable charging time flexibility, and the calculation formula is as follows:

[0011] ;

[0012] in, For user i's flexibility margin, Let i be the expected departure time. Let i be the arrival time. The minimum necessary charging time required to meet the minimum charging needs of user i.

[0013] Furthermore, the charging urgency is defined as the power shortage per unit interval, used to characterize the user's charging rigidity, and the calculation formula is:

[0014] ;

[0015] in, The urgency of charging user i This represents the initial state of charge when user i arrives at the station. The time interval between the end of the current charge and the start of the next charge for user i.

[0016] Furthermore, the arrival time of the charging urgency is transformed using a sine-cosine periodic encoding method to eliminate the periodic truncation effect of the time variable:

[0017] ;

[0018] ;

[0019] in, The sine code value for the arrival time. The cosine encoded value of the arrival time;

[0020] After periodic encoding, with The original linear values ​​of arrival times are used to replace the linear values ​​of arrival times in subsequent feature calculations to ensure that the Euclidean distance between adjacent times in the feature space is continuous and bounded.

[0021] Furthermore, a Logistic response probability model is constructed for each cluster group, specifically as follows:

[0022] Calculate the mean flexibility margin of each member in cluster k. With the average charging urgency Based on the theory of stochastic utility maximization, a logistic response probability model is established with flexibility margin as the promoting factor and charging urgency as the inhibiting factor.

[0023] ;

[0024] in, Let k be the response probability of the cluster group. , These are the flexibility promotion coefficient and the urgency inhibition coefficient for group k, respectively.

[0025] Furthermore, the method for calculating the dynamic online scale is as follows:

[0026] For cluster k, the expected number of online vehicles at time t is estimated using a convolution method based on its historical arrival time distribution and dwell time distribution.

[0027] ;

[0028] in, Let k be the expected number of online vehicles in cluster k. Let k be the total number of vehicles in cluster k. The probability density function of vehicle arrival times for cluster k. Let be the cumulative distribution function of the dwell time of cluster k. Let be the arrival time variable in the convolution integral, representing the historical time at which the vehicle may arrive at the charging station, and , Indicates the time of arrival of the vehicle From the current moment The length of time the stay has been.

[0029] Furthermore, the calculation method for the basic power upper and lower bounds is as follows: For online vehicles in cluster group k, based on the constraints of state of charge and charging cutoff time, their power lower bound and power upper bound are determined respectively; the power lower bound reflects the minimum charging power required for the vehicle to ensure minimum travel demand.

[0030] ;

[0031] The upper limit of the charging power is determined by the vehicle's rated charging power and the current state of charge.

[0032] ;

[0033] in, This is the lower bound of the basic power of cluster group k. This represents the minimum off-site state of charge requirement for cluster group k. The mean of the initial state of charge of cluster group k upon arrival at the station. Let $k$ be the mean battery capacity of cluster k. For charging efficiency, Let be the mean of the expected departure times of vehicles in cluster k, and t be the current scheduling time. This is the upper bound of the basic power of cluster group k. Let be the average rated charging power of vehicles in cluster k. This is the scheduling time step.

[0034] Furthermore, the response reliability is calculated using the following formula:

[0035] ;

[0036] in, In response to reliability, The response probability of cluster k. The coefficient of variation of the duration of stay. is the coefficient of variation of the state of charge.

[0037] Furthermore, a differentiated incentive mechanism is designed, a linear programming quota optimization model under budget and diversity constraints is established, the optimal contracted quota is solved, and the feasible region of aggregated power for connecting to the power grid dispatch center is generated, specifically including:

[0038] Design a differentiated incentive mechanism, expressed as:

[0039] ;

[0040] ;

[0041] ;

[0042] in, As the benchmark discount factor, As the benchmark premium factor, As a regulating factor, The maximum urgency to charge each cluster group. Let $\frac{ ... , These are the discount factor and premium factor corresponding to cluster group k, respectively. The benchmark incentive unit price, Let $\frac{k}{k}$ be the unit quota incentive cost for cluster group $k$.

[0043] Establish a linear programming quota optimization model under budget and diversity constraints, with the following expression:

[0044] ;

[0045] Constraints:

[0046] ;

[0047] ;

[0048] ;

[0049] in, For the signing quota of cluster group k, In response to reliability, Let B be the amount of flexible resources available for cluster group k, where K is the total number of clusters and B is the budget constraint. This is the upper limit of the contracted quota. For indicator functions, This represents the minimum number of participants required to meet diversity constraints.

[0050] Construct a feasible domain for aggregated power to connect to the power grid dispatch center:

[0051] ;

[0052] ;

[0053] in, The feasible range of aggregated power at time t constitutes a flexible resource boundary for grid-side dispatch. Let k be the total number of vehicles in cluster k. This is the lower bound of the basic power of cluster group k. This is the upper bound of the basic power of cluster group k.

[0054] Secondly, a behavior-aware electric vehicle charging resource potential assessment and optimization allocation device is provided, comprising: a response probability calculation module, used to construct a Logistic response probability model for each cluster group based on the clustering results of electric vehicle charging behavior, taking into account two driving factors: flexibility margin and charging urgency, to obtain the response probability of each cluster group; a reliability calculation module, used to evaluate the stability of adjustable resource delivery for each cluster group based on the obtained response probabilities of each cluster group, combined with dynamic online vehicle scale, basic power upper and lower bounds, dwell time fluctuations, and state of charge fluctuations, to obtain a response reliability index used to characterize the actual callable capacity; and an optimization module, used to design a differentiated incentive mechanism based on the response probabilities and response reliability index of each cluster group, establish a linear programming quota optimization model under budget and diversity constraints, solve for the optimal contracted quota, and generate an aggregated power feasible domain for connecting to the power grid dispatch center.

[0055] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0056] (1) Based on the clustering results of electric vehicle charging behavior, this invention takes into account two types of driving factors: flexibility margin and charging urgency. A Logistic response probability model is constructed for each cluster group to obtain the response probability of each cluster group. Based on the obtained response probability of each cluster group, the stability of adjustable resource delivery of each cluster group is evaluated by combining dynamic online vehicle scale, basic power upper and lower bounds, dwell time fluctuation and state of charge fluctuation, and a response reliability index is obtained to characterize the actual callable capacity. Based on the response probability and reliability index, a differentiated incentive mechanism is designed, a linear programming quota optimization model under budget and diversity constraints is established, the optimal contract quota is solved, and the aggregated power feasible domain for connecting to the power grid dispatch center is generated. The response capacity of each group is refined from the behavioral characteristics, and the differentiated optimization allocation of resources is realized under budget constraints.

[0057] (2) Starting from the heterogeneity of electric vehicle user behavior, this invention introduces two types of driving variables: flexibility margin and charging urgency. By establishing a group-level response probability model through the random utility maximization framework, it realizes the probabilistic mapping from static behavior profile to quantitative representation of participation intention, effectively overcoming the limitation of traditional methods that regard all users as homogeneous adjustable resources.

[0058] (3) In the power boundary characterization, the present invention explicitly incorporates the user departure time constraint and subsequent travel demand into the power lower bound calculation, which effectively avoids the problem of adjustable space collapse caused by overestimation of the power lower bound in the traditional method, and ensures the feasibility of the assessed adjustable potential.

[0059] (4) The present invention designs a differentiated incentive mechanism of flexibility discount and urgency premium. Under the dual constraints of budget and diversity, the contract quota of each group is optimized by linear programming. Compared with the baseline schemes such as uniform allocation and full participation, it significantly improves the reliable flexibility energy under unit incentive cost and improves the efficiency of charging resource allocation.

[0060] (5) The aggregated power feasible domain output by the present invention has temporal continuity and physical feasibility, and can be directly connected to the power grid dispatch center, providing accurate flexible resource boundary constraints for subsequent master-slave game collaborative dispatch, and supporting the multi-objective collaborative operation of power grid and aggregated operators. Attached Figure Description

[0061] Figure 1 This is a schematic diagram of the main process of a method for assessing and optimizing the allocation of electric vehicle charging resources based on behavior perception, provided in an embodiment of the present invention.

[0062] Figure 2 This is a schematic diagram of the implementation process of the method for assessing and optimizing the charging resource potential of electric vehicles based on behavior perception in an embodiment of the present invention;

[0063] Figure 3 This is a detailed flowchart illustrating the user response probability modeling and cluster-level adjustable potential assessment in an embodiment of the present invention. Detailed Implementation

[0064] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0065] like Figures 1-3 As shown, a behavior-aware method for assessing and optimizing the allocation of electric vehicle charging resources includes: based on the clustering results of electric vehicle charging behavior, considering two driving factors—flexibility margin and charging urgency—a Logistic response probability model is constructed for each cluster to obtain the response probability of each cluster; based on the obtained response probabilities of each cluster, the stability of adjustable resource delivery for each cluster is evaluated by combining dynamic online vehicle scale, upper and lower bounds of base power, dwell time fluctuations, and state of charge fluctuations, resulting in a response reliability index characterizing the actual available capacity; based on the response probabilities and response reliability index of each cluster, a differentiated incentive mechanism is designed, a linear programming quota optimization model under budget and diversity constraints is established, the optimal contracted quota is solved, and an aggregated power feasible region for connecting to the power grid dispatch center is generated.

[0066] Step A: Based on the clustering results of electric vehicle charging behavior, and taking into account two driving factors, flexibility margin and charging urgency, construct a Logistic response probability model for each cluster group to obtain the response probability of each cluster group.

[0067] In this invention, the clustering results of electric vehicle charging behavior are obtained using a soft clustering method based on a Gaussian mixture model. Each charging event is assigned to a cluster group in the form of a posterior probability vector, preserving the uncertainty information of sample attribution. The posterior probability vector is incorporated into the weight calculation of the Logistic response probability model as a quantitative basis for sample uncertainty.

[0068] Step A1: Calculate the flexibility margin for each charging event. The flexibility margin is defined as the difference between the remaining time after arrival and departure and the minimum necessary charging time, reflecting the user's acceptable charging time flexibility. The flexibility margin for user i is... The calculation formula is:

[0069] ;

[0070] in, Let i be the expected departure time. Let i be the arrival time. The minimum necessary charging time required to meet the minimum charging needs of user i.

[0071] Step A2: Calculate the charging urgency for each charging event. Charging urgency is defined as the power shortfall per unit interval, used to characterize the user's charging rigidity. The charging urgency for user i is... The calculation formula is:

[0072] ;

[0073] in, This represents the initial state of charge when user i arrives at the station. The time interval between the end of the current charge and the start of the next charge for user i.

[0074] For arrival time characteristics, sine-cosine periodic encoding must be performed first:

[0075] ;

[0076] ;

[0077] in, The sine code value for the arrival time. The cosine encoded value for the arrival time.

[0078] This encoding method eliminates the periodic truncation effect of time variables during the midnight period, ensuring that the Euclidean distance between adjacent times in the feature space remains continuous and bounded. After periodic encoding, the... The original linear values ​​of arrival times are used to replace the linear values ​​of arrival times in subsequent feature calculations to ensure that the Euclidean distance between adjacent times in the feature space is continuous and bounded.

[0079] Step A3: Calculate the mean flexibility margin of each member in cluster k. With the average charging urgency Based on the theory of stochastic utility maximization, a logistic response probability model is established with flexibility margin as the promoting factor and charging urgency as the inhibiting factor.

[0080] ;

[0081] in, Let k be the response probability of the cluster group. The larger the value, the stronger the willingness of the cluster to respond to scheduling incentives. , These are the flexibility promotion coefficient and urgency inhibition coefficient of group k, respectively, obtained by maximum likelihood estimation of historical response data. , These are the parameters of the Logistic response probability model for cluster group k, reflecting the strength of the influence of flexibility and urgency on the willingness to participate, respectively.

[0082] Step B: Based on the response probabilities of each cluster, and combined with the dynamic online vehicle scale, the upper and lower bounds of the base power, the fluctuation of dwell time and the fluctuation of state of charge, the stability of the adjustable resource delivery of each cluster is evaluated to obtain a response reliability index used to characterize the actual callable capacity.

[0083] Step B1: Perform dynamic online scale modeling of the charging vehicles.

[0084] For cluster k, its historical arrival time distribution is used. Distribution of stay duration The expected number of online vehicles at time t is estimated using a convolution method.

[0085] ;

[0086] in, Let k be the expected number of online vehicles in cluster k. Let k be the total number of vehicles in cluster k. Let k be the probability density function of vehicle arrival times for cluster k. Let be the cumulative distribution function of the dwell time of cluster k. Let be the arrival time variable in the convolution integral, representing the historical time at which the vehicle may arrive at the charging station, and , Indicates the time of arrival of the vehicle From the current moment The length of time the stay has been. The probability density function representing the dwell time of cluster group k is denoted by , and The relationship satisfies:

[0087] ;

[0088] The dynamic online scale of charging vehicles reflects the time-varying dynamic characteristics of charging resources.

[0089] Step B2: Characterization of the basic power boundary under state of charge constraints. Combining the minimum off-site state of charge requirement and the charging cutoff time, the lower and upper power bounds for online vehicles in cluster k are determined respectively.

[0090] For online vehicles in cluster k, the lower bound of power reflects the minimum charging power required to ensure minimum travel demand.

[0091] ;

[0092] The upper limit of the charging power is determined by the vehicle's rated charging power and the current state of charge.

[0093] ;

[0094] in, This is the lower bound of the basic power of cluster group k. This represents the minimum off-site state of charge requirement for cluster group k. The mean of the initial state of charge of cluster group k upon arrival at the station. Let $k$ be the mean battery capacity of cluster k. For charging efficiency, Let be the mean of the expected departure times of vehicles in cluster k, and t be the current scheduling time. This is the upper bound of the basic power of cluster group k. Let be the average rated charging power of vehicles in cluster k. This is the scheduling time step.

[0095] The lower limit of power is determined by the minimum off-site state of charge requirement and the remaining on-site time, fully considering the energy replenishment rigidity of users' subsequent travel, and avoiding the simplistic use of SOC in traditional methods. start The problem of overestimating the adjustable space caused by directly estimating the lower limit of power.

[0096] Step B3: Reliable metric of response incorporating multidimensional uncertainty. (In response probability...) Based on this, combined with the coefficient of variation of stay duration With the coefficient of variation of the state of charge By penalizing uncertainty, a group reliability index is obtained. :

[0097] ;

[0098] in, For response reliability.

[0099] Taking into account three types of uncertainties—response probability, dwell time uncertainty, and state of charge fluctuation—a comprehensive evaluation of the stability of the adjustable potential of the cluster population is formed.

[0100] Step C: Based on the response probability obtained in Step A and the response reliability index obtained in Step B, design a differentiated incentive mechanism, establish a linear programming quota optimization model under budget and diversity constraints, solve for the optimal contracted quota, and generate the aggregated power feasible region for connecting to the power grid dispatch center.

[0101] Step C1: Design a differentiated incentive mechanism, adopting a dual incentive structure of flexibility discount and urgency premium, and design differentiated incentive parameters for each group, expressed as follows:

[0102] ;

[0103] ;

[0104] ;

[0105] in, As the benchmark discount factor, As the benchmark premium factor, As a regulating factor, The maximum urgency to charge each cluster group. Let $\frac{ ... , These are the discount factor and premium factor corresponding to cluster group k, respectively. The benchmark incentive unit price, Let $\frac{k}{k}$ be the unit quota incentive cost for cluster group $k$.

[0106] Clusters with low charging urgency and high flexibility margins can be incentivized with discounts to reduce call costs, while clusters with high charging urgency can be incentivized with premiums to increase participation, thereby maximizing resource utilization efficiency within a limited budget.

[0107] Step C2: Establish a linear programming quota optimization model under budget and diversity constraints, with the following expression:

[0108] ;

[0109] Constraints:

[0110] ;

[0111] ;

[0112] ;

[0113] in, For the signing quota of cluster group k, In response to reliability, Let K be the amount of flexible resources available to cluster group k, defined as the adjustable energy or adjustable power potential that the group can provide during the scheduling period, where K is the total number of clusters and B is the budget constraint. This is the upper limit of the contracted quota. For indicator functions, This represents the minimum number of participants required to meet diversity constraints.

[0114] With reliable, flexible, and energy-efficient performance as the optimization objective, and considering budget constraint B and contracted quota cap constraint... and diversity constraints Solving for the optimal contract quota vector This linear programming problem is convex and can be solved efficiently using the simplex method or the interior point method.

[0115] Step C3: Construct the aggregated power feasible region for connecting to the power grid dispatch center:

[0116] ;

[0117] ;

[0118] in, The feasible range of aggregated power at time t constitutes a flexible resource boundary for grid-side dispatch. Let k be the total number of vehicles in cluster k. This is the lower bound of the basic power of cluster group k. This is the upper bound of the basic power of cluster group k.

[0119] Output by time period The interval forms a full-time-domain aggregated scheduling boundary, which can be directly reported to the power grid dispatch center for collaborative scheduling constraint input.

[0120] Example 2

[0121] Based on the behavior-aware electric vehicle charging resource potential assessment and optimization allocation method described in Embodiment 1, this embodiment provides a behavior-aware electric vehicle charging resource potential assessment and optimization allocation device, including: a response probability calculation module, used to construct a Logistic response probability model for each cluster group based on the clustering results of electric vehicle charging behavior, taking into account two types of driving factors: flexibility margin and charging urgency, to obtain the response probability of each cluster group; a reliability calculation module, used to evaluate the adjustable resource delivery stability of each cluster group based on the obtained response probabilities of each cluster group, combined with dynamic online vehicle scale, basic power upper and lower bounds, dwell time fluctuations, and state of charge fluctuations, to obtain a response reliability index used to characterize the actual callable capacity; and an optimization module, used to design a differentiated incentive mechanism based on the response probabilities and response reliability index of each cluster group, establish a linear programming quota optimization model under budget and diversity constraints, solve for the optimal contracted quota, and generate an aggregated power feasible domain for connecting to the power grid dispatch center.

[0122] This invention is based on the clustering results of electric vehicle charging behavior. Relying on the theory of maximizing random utility, it constructs a Logistic response probability model for each cluster group, using flexibility margin as a promoting factor and charging urgency as an inhibiting factor. Then, it conducts a low-level evaluation of the physical adjustability potential of each cluster group from three dimensions: dynamic online scale, upper and lower bounds of basic power, and response reliability. On this basis, it designs a flexibility discount-urgency premium incentive mechanism that reflects the differences between groups, and establishes a linear programming quota optimization model with the goal of maximizing reliable flexibility energy and constraints of budget and diversity. This generates an aggregated power feasible region and contracted quota that can be directly connected to the power grid dispatch center.

[0123] This invention, starting from the user behavior side, effectively improves the adjustable space collapse problem caused by power lower bound estimation in traditional methods, significantly improves the resource allocation efficiency under limited incentive budget, and provides accurate flexible resource boundaries for grid-charging operator collaborative scheduling.

[0124] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0125] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0126] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0127] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0128] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for assessing and optimizing the allocation of electric vehicle charging resource potential based on behavior perception, characterized in that, include: Based on the clustering results of electric vehicle charging behavior, and taking into account two driving factors, flexibility margin and charging urgency, a Logistic response probability model is constructed for each cluster to obtain the response probability of each cluster. Based on the response probabilities of each cluster, and combined with the dynamic online vehicle scale, the upper and lower bounds of the base power, the fluctuation of dwell time and the fluctuation of state of charge, the stability of the adjustable resource delivery of each cluster is evaluated, and a response reliability index is obtained to characterize the actual callable capacity. Based on the response probability and response reliability indices of each cluster, a differentiated incentive mechanism is designed, a linear programming quota optimization model under budget and diversity constraints is established, the optimal contracted quota is solved, and the aggregated power feasible region for connecting to the power grid dispatch center is generated.

2. The method for assessing and optimizing the potential of electric vehicle charging resources based on behavior perception according to claim 1, characterized in that, The flexibility margin is defined as the difference between the remaining time after arrival and before departure and the minimum necessary charging time, reflecting the user's acceptable charging time flexibility. The calculation formula is as follows: ； in, For user i's flexibility margin, Let i be the expected departure time. Let i be the arrival time. The minimum necessary charging time required to meet the minimum charging needs of user i.

3. The method for assessing and optimizing the potential of electric vehicle charging resources based on behavior perception according to claim 1, characterized in that, The charging urgency is defined as the power shortage per unit interval time, used to characterize the user's charging rigidity, and the calculation formula is: ； in, The urgency of charging user i This represents the initial state of charge of user i upon arrival at the station. The time interval between the end of the current charge and the start of the next charge for user i.

4. The method for assessing and optimizing the potential of electric vehicle charging resources based on behavior perception according to claim 3, characterized in that, The arrival time of the charging urgency is transformed using a sine-cosine periodic encoding method to eliminate the periodic cutoff effect of the time variable. ； ； in, The sine code value for the arrival time. The cosine encoded value of the arrival time; After periodic encoding, with The original linear values ​​of arrival times are used to replace the linear values ​​of arrival times in subsequent feature calculations to ensure that the Euclidean distance between adjacent times in the feature space is continuous and bounded.

5. The method for assessing and optimizing the potential of electric vehicle charging resources based on behavior perception according to claim 1, characterized in that, For each cluster, a Logistic response probability model is constructed as follows: Calculate the mean flexibility margin of each member in cluster k. With the average charging urgency Based on the theory of stochastic utility maximization, a logistic response probability model is established with flexibility margin as the promoting factor and charging urgency as the inhibiting factor. ； in, Let k be the response probability of the cluster group. , These are the flexibility promotion coefficient and the urgency inhibition coefficient for group k, respectively.

6. The method for assessing and optimizing the potential of electric vehicle charging resources based on behavior perception according to claim 1, characterized in that, The method for calculating the dynamic online scale is as follows: For cluster k, the expected number of online vehicles at time t is estimated using a convolution method based on its historical arrival time distribution and dwell time distribution. ； in, Let k be the expected number of online vehicles in cluster k. Let k be the total number of vehicles in cluster k. Let k be the probability density function of vehicle arrival times for cluster k. Let be the cumulative distribution function of the dwell time of cluster k. Let be the arrival time variable in the convolution integral, representing the historical time at which the vehicle may arrive at the charging station, and , Indicates the time of arrival of the vehicle From the current moment The length of time the stay has been.

7. The method for assessing and optimizing the potential of electric vehicle charging resources based on behavior perception according to claim 1, characterized in that, The calculation method for the lower and lower bounds of the basic power is as follows: For online vehicles in cluster k, based on the constraints of state of charge and charging cutoff time, their lower and upper power bounds are determined respectively; the lower power bound reflects the minimum charging power required for the vehicle to ensure minimum travel demand. ； The upper limit of the charging power is determined by the vehicle's rated charging power and the current state of charge. ； in, This is the lower bound of the basic power of cluster group k. This represents the minimum off-site state of charge requirement for cluster group k. The mean of the initial state of charge of cluster group k upon arrival at the station. Let $k$ be the mean battery capacity of cluster k. For charging efficiency, Let be the mean of the expected departure times of vehicles in cluster k, and t be the current scheduling time. This is the upper bound of the basic power of cluster group k. Let be the average rated charging power of vehicles in cluster k. This is the scheduling time step.

8. The method for assessing and optimizing the potential of electric vehicle charging resources based on behavior perception according to claim 1, characterized in that, The response reliability is calculated using the following formula: ； in, In response to reliability, The response probability of cluster k. The coefficient of variation of the duration of stay. is the coefficient of variation of the state of charge.

9. The method for assessing and optimizing the potential of electric vehicle charging resources based on behavior perception according to claim 1, characterized in that, Design a differentiated incentive mechanism, establish a linear programming quota optimization model under budget and diversity constraints, solve for the optimal contracted quota, and generate the aggregated power feasible region for connecting to the power grid dispatch center, specifically including: Design a differentiated incentive mechanism, expressed as: ； ； ； in, As the benchmark discount factor, As the benchmark premium factor, As a regulating factor, The maximum urgency to charge each cluster group. Let $\frac{ ... , These are the discount factor and premium factor corresponding to cluster group k, respectively. The benchmark incentive unit price, Let $\frac{k}{k}$ be the unit quota incentive cost for cluster group $k$. Establish a linear programming quota optimization model under budget and diversity constraints, with the following expression: ； Constraints: ； ； ； in, For the signing quota of cluster group k, In response to reliability, Let B be the amount of flexible resources available for cluster group k, where K is the total number of clusters and B is the budget constraint. This is the upper limit of the contracted quota. For indicator functions, The minimum number of participating groups required by diversity constraints; Construct a feasible domain for aggregated power to connect to the power grid dispatch center: ； ； in, The feasible range of aggregated power at time t constitutes a flexible resource boundary for grid-side dispatch. Let k be the total number of vehicles in cluster k. This is the lower bound of the basic power of cluster group k. This is the upper bound of the basic power of cluster group k.

10. A device for assessing and optimizing the allocation of electric vehicle charging resource potential based on behavior perception, comprising: The response probability calculation module is used to construct a Logistic response probability model for each cluster based on the clustering results of electric vehicle charging behavior, taking into account two driving factors: flexibility margin and charging urgency, and to obtain the response probability of each cluster. The reliability calculation module is used to evaluate the stability of adjustable resource delivery for each cluster group based on the response probability of each cluster group, combined with dynamic online vehicle scale, basic power upper and lower bounds, dwell time fluctuations and state of charge fluctuations, and obtain a response reliability index to characterize the actual callable capacity. The optimization module is used to design differentiated incentive mechanisms based on the response probability and response reliability indices of each cluster, establish a linear programming quota optimization model under budget and diversity constraints, solve for the optimal contracted quota, and generate the aggregated power feasible region for connecting to the power grid dispatch center.

Images