Method and system for drawing regional energy consumption map of electric vehicle

Abstract

The invention provides a method and a system for drawing a regional energy consumption map of an electric vehicle. Wherein the acquired driving data and external meteorological information comprise multiple parameters and temperature gradients in the GB32960 standard. A dynamic grid model is constructed based on administrative division and a road network, track points are mapped into space grid units by using an adaptive algorithm, and the units combine information such as gradient, temperature and traffic flow. A trajectory analysis technology is adopted to segment a driving segment crossing a grid boundary, and feature vectors are generated; by establishing a mapping relationship between the feature vector and the energy consumption contribution degree, the original energy consumption is calculated, and multi-factor correction is performed to generate a dynamic energy consumption reference value. And associating the distribution density of the charging piles with the traffic flow weight by using a space-time coupling diffusion model, and generating a three-dimensional dynamic map through a rendering engine. According to the technical scheme provided by the embodiment of the invention, the accuracy of electric vehicle energy consumption estimation and the scientificity of charging facility layout are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of new energy technologies, and in particular to a method and system for drawing an electric vehicle regional energy consumption map. Background Art

[0002] With the popularity of electric vehicles, city managers and transportation planners are facing new challenges in optimizing the layout of charging facilities and improving energy efficiency. Especially in large cities, understanding the energy consumption of electric vehicles in different areas is crucial for rationally planning the distribution of charging piles and reducing user mileage anxiety.

[0003] Currently, there are several methods and technologies that attempt to solve the above problems, including using GPS data to track vehicle trajectories, combining basic road slope information and simple energy consumption models to estimate energy consumption. In addition, some solutions provide basic energy consumption predictions by integrating limited external data sources (such as air temperature).

[0004] The main drawback of existing solutions is that they lack comprehensive consideration of complex variables, especially the key factors that affect the energy consumption of electric vehicles, such as road slope, temperature changes, and traffic flow. In addition, existing energy consumption maps are often based on a fixed grid system and cannot flexibly adapt to the distribution of data points of different densities, resulting in data overload in high-density areas and the loss of important details in low-density areas. At the same time, these solutions usually ignore the principle of cross-regional energy conservation, making it difficult to effectively analyze and correct driving segments that span multiple grids, thereby limiting the accuracy and practicality of energy consumption estimates. Summary of the invention

[0005] The embodiments of the present application provide a method and system for drawing an electric vehicle regional energy consumption map, so as to solve the problems of insufficient accuracy in electric vehicle energy consumption estimation and low scientificity in charging facility layout in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for drawing an electric vehicle regional energy consumption map, comprising: Acquire a multi-source data stream generated during the driving of the electric vehicle, and access an external meteorological interface to acquire the temperature gradient parameters of the environment, wherein the multi-source data stream includes longitude, latitude, cumulative mileage, state of charge, and total energy fields of the on-board energy storage device in the format of the national recommended standard GB 32960; A dynamic grid generation model is constructed based on the administrative division coding system and the road topology network, and an adaptive density adjustment algorithm is used to map trajectory points to hierarchically associated spatial grid cells, which are bound to the road slope distribution matrix, temperature gradient parameters, and traffic flow weight coefficients; Trajectory continuity analysis is used to perform multimodal segmentation on the driving segments that cross the grid boundaries in the spatial grid cells, and when the trajectory points enter different spatial grid cells, segment reconstruction with energy conservation constraints is triggered, and a feature vector including a grid code, a slope change rate, a temperature gradient vector correlation factor, and a traffic flow weight coefficient is generated for each sub-segment; Establishing a mapping relationship between the characteristic vector and the grid energy consumption contribution, calculating the original energy consumption value based on the difference between the total energy of the energy storage device and the accumulated mileage, performing nonlinear compensation correction in combination with the slope change rate, and superimposing the temperature gradient vector association factor and the traffic flow weight coefficient to generate the dynamic energy consumption benchmark value of multiple vehicle types in the spatial grid unit; A spatiotemporal coupled diffusion model is constructed, and the dynamic energy consumption benchmark value is correlated with the charging pile distribution density and the traffic flow weight coefficient. A three-dimensional dynamic map including an energy consumption thermal layer, a charging facility density layer and a traffic flow correlation layer is generated through a multi-resolution rendering engine.

[0007] Optionally, a mapping relationship between the characteristic vector and the grid energy consumption contribution is established, the original energy consumption value is calculated based on the difference between the total energy of the energy storage device and the accumulated mileage, a nonlinear compensation correction is performed in combination with the slope change rate, and the temperature gradient vector association factor and the traffic flow weight coefficient are superimposed to generate the dynamic energy consumption benchmark value of multiple vehicle types in the spatial grid unit, including: The original energy consumption value is calculated based on the difference between the total energy of the energy storage device and the accumulated mileage, the slope change rate data extracted from the road slope distribution matrix is ​​integrated, and the historical road topology data and the energy consumption deviation are back-propagated and trained using a multi-interval segmented optimization algorithm to generate a set of slope compensation coefficients; The ambient temperature, battery temperature rise gradient and thermal management power time series data in the temperature gradient parameters are integrated to construct a generation module of temperature influence factors, and a dynamic temperature correction coefficient is output using multi-dimensional feature fusion technology; Integrate the historical congestion pattern database of electric vehicles, and construct a traffic flow weight coefficient through an adaptive threshold iteration algorithm. The traffic flow weight coefficient includes a congestion intensity classification parameter and a speed fluctuation correction factor, and establishes a cross-dimensional correlation relationship with the road type code; Execute a multi-source parameter collaborative correction operation, perform a tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient, superimpose the traffic flow weight coefficient, output a dynamic energy consumption benchmark value, and use the vehicle dynamic characteristic parameters as a normalization adjustment factor to complete the dynamic calibration of the dynamic energy consumption benchmark value.

[0008] Optionally, a multi-source parameter collaborative correction operation is performed, the slope compensation coefficient set is subjected to a tensor product operation with the dynamic temperature correction coefficient, the traffic flow weight coefficient is superimposed, a dynamic energy consumption reference value is output, and a vehicle dynamic characteristic parameter is used as a normalization adjustment factor to complete the dynamic calibration of the dynamic energy consumption reference value, including: Constructing a three-dimensional tensor operation channel to perform a tensor product operation on a slope compensation coefficient set and a dynamic temperature correction coefficient, wherein the dimension of the two-dimensional matrix of the slope compensation coefficient set corresponds to the number of road slope classifications and the number of historical road topology data codes, and the dimension of the dynamic temperature correction coefficient matrix corresponds to the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications; Decomposing the traffic flow weight coefficient into a congestion intensity component and a speed fluctuation component, wherein the congestion intensity component is associated with the administrative division restriction strategy in the road type code, and the speed fluctuation component is bound to the speed-energy consumption mapping relationship table of the road topology network; Generate a normalized adjustment factor using the vehicle dynamic characteristic parameters, wherein the vehicle dynamic characteristic parameters include a total energy field of an energy storage device and a mass-energy consumption conversion matrix, and the normalized adjustment factor acts on a process of superimposing a result of the tensor product operation and a traffic flow weight component; Execute a closed-loop calibration operation of the dynamic energy consumption reference value, and the calibration operation output result is synchronously input into the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit, wherein the charging pile distribution density calculation unit is associated with the spatial coding attribute of the road topology network, and the energy consumption thermal value diffusion unit is bound to the temperature change cycle data of the meteorological interface; The slope compensation coefficient set, the dynamic temperature correction coefficient, the traffic flow weight coefficient and the vehicle dynamic characteristic parameters form a cross-level data pipeline in the spatiotemporal coupled diffusion model, and all intermediate calculation results act on the downstream computing module through the convolutional neural network channel.

[0009] Optionally, a three-dimensional tensor operation channel is constructed to perform a tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient, wherein the two-dimensional matrix dimension of the slope compensation coefficient set corresponds to the number of road slope classifications and the number of historical road topology data codes, and the dynamic temperature correction coefficient matrix dimension corresponds to the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications, including: A two-dimensional matrix of a slope compensation coefficient set is constructed by combining the number of road slope classifications and the number of historical road topology data codes, and a two-dimensional matrix of a dynamic temperature correction coefficient is constructed by combining the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications, so as to perform a tensor product operation to generate a three-dimensional compensation tensor structure, wherein the dimension of the three-dimensional compensation tensor structure includes a product combination of the number of road slope classifications, the number of temperature compensation intervals, and the number of battery chemical characteristic parameter classifications; Performing multi-level feature fusion processing on the three-dimensional compensation tensor structure, fusing the cross-validation results of the road slope classification number and the historical road topology data encoding, and injecting the battery chemical characteristic dynamic attenuation factor of the two-dimensional matrix of the dynamic temperature correction coefficient to generate a synergistic compensation tensor; Perform multi-source data superposition on the synergistic compensation tensor, the congestion intensity component and the speed fluctuation component of the traffic flow weight component to generate a dynamic multi-source fusion tensor; Performing a staged linear transformation on the dynamic multi-source fusion tensor using the mass-energy consumption conversion matrix in the vehicle dynamic characteristic parameters to generate a linear mapping image vector, wherein the staged linear transformation includes capacity attenuation compensation of the total energy field of the energy storage device and a gradient correction operation of the mass-energy consumption conversion matrix; Inputting the linear mapping image vector into the spatiotemporal coupled diffusion model for multi-scale feature calibration, and using the multi-scale feature calibration to associate the spatial coding attributes of the road topology network with the temperature change period data fluctuation curve of the meteorological interface, to generate a multi-dimensional output sequence carrying a dynamic energy consumption benchmark value; The multidimensional output sequence is transmitted to the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit through the convolutional neural network channel, wherein the charging pile distribution density calculation unit integrates the spatial coding attributes of the road topology network and the spatial distribution weight of the dynamic energy consumption benchmark value, and the energy consumption thermal value diffusion unit binds the temperature change cycle data and the attenuation correlation factor of the battery chemical characteristic parameter classification number.

[0010] Optionally, a multi-level feature fusion process is performed on the three-dimensional compensation tensor structure, the cross-validation result of the road slope classification number and the historical road topology data encoding is fused, and the dynamic attenuation factor of the battery chemical characteristics of the two-dimensional matrix of the dynamic temperature correction coefficient is injected to generate a synergistic compensation tensor, including: A cross-validation matrix is ​​constructed based on the road slope classification number and the data encoded by the historical road topology data, and a correlation feature extraction process is performed on the cross-validation matrix using a multi-channel convolutional layer to generate an initial fusion weight matrix that matches the subsequent tensor operation dimension; According to the battery chemical characteristic parameter classification dimension of the dynamic temperature correction coefficient matrix, the initial fusion weight matrix is ​​subjected to attenuation factor vector decomposition processing, and the dynamic attenuation coefficient calculation processing is performed on each temperature compensation interval in combination with the time series attenuation model to generate an attenuation factor tensor carrying a temperature gradient correlation factor; Performing an element-by-element product operation on the initial fusion weight matrix and the attenuation factor tensor by using a tensor broadcast mechanism to generate a coupling tensor including a four-dimensional structure; Based on the multi-dimensional attention mechanism, cross-dimensional feature alignment processing is performed on the coupling tensor, a query vector is generated using the feature of the road slope classification number, a key vector is generated using the feature of the historical road topology data encoding, and a value vector is generated using the temperature attenuation factor, and the query vector, key vector and value vector are normalized and weighted to generate an attention weight matrix; The three-dimensional compensation tensor and the attention weight matrix are subjected to residual connection processing, and a batch normalization layer is used to perform feature fusion processing on the connection result to generate a collaborative compensation tensor containing a four-dimensional product combination.

[0011] Optionally, the dynamic multi-source fusion tensor is subjected to a staged linear transformation using a mass-energy consumption conversion matrix in the vehicle dynamic characteristic parameters to generate a linear mapping image vector, wherein the staged linear transformation includes a capacity attenuation compensation of the total energy field of the energy storage device and a gradient correction operation of the mass-energy consumption conversion matrix, including: A dynamic capacity correction tensor is constructed based on the capacity attenuation compensation operation of the total energy field of the energy storage device, and a multi-dimensional parameter fusion process is performed on the compensation operation using the time series attenuation characteristics of the historical state of charge data and the dynamic correlation factor of the temperature gradient parameter to generate a dynamic capacity correction matrix carrying capacity correction characteristics; According to the road slope classification number and the temperature compensation interval parameter of the dynamic multi-source fusion tensor, a gradient correction operation is performed on the mass-energy consumption conversion matrix, and a weight vector associated with the vehicle mass parameter and the slope classification number is extracted to generate a corrected mass-energy consumption conversion matrix; The dynamic capacity correction matrix and the corrected mass-energy consumption conversion matrix are subjected to feature coupling processing by using a tensor product operation to generate a mass-energy consumption correlation tensor including a capacity attenuation compensation feature and a road slope-mass coupling feature; Based on the staged linear transformation, multi-dimensional feature fusion processing is performed on the mass-energy consumption correlation tensor. In the first stage, the superposition characteristics of the road slope classification number and the traffic flow speed fluctuation component are used to generate a primary mapping vector. In the second stage, the capacity attenuation compensation parameter of the energy storage device is injected into the primary mapping vector to generate an intermediate mapping tensor using a dynamic residual connection. In the third stage, a batch normalization layer is used to perform gradient correction processing on the intermediate mapping tensor to generate a linear mapping image vector.

[0012] Optionally, a cross-dimensional feature alignment process is performed on the coupling tensor based on a multi-dimensional attention mechanism, a query vector is generated using the feature of the road slope classification number, a key vector is generated using the feature of the historical road topology data encoding, and a value vector is generated using the temperature attenuation factor, and the query vector, the key vector and the value vector are normalized and weighted to generate an attention weight matrix, including: Using the road slope classification number, a multi-level feature decomposition process is performed on the height dimension of the four-dimensional coupling tensor, and the decomposed coarse-grained classification features and fine-grained fluctuation components are transmitted to the channel segmentation processing module to generate a query sub-vector group carrying multi-level slope features; Based on the historical road topology data encoding, spatial pooling processing is performed on the width dimension of the four-dimensional coupling tensor, and a dynamic weight adjustment is performed on the extracted local correlation features and global distribution features in combination with a temporal decay model to generate a key vector block; Performing a dynamic mask operation on the temperature gradient dimension of the four-dimensional coupling tensor according to the temperature attenuation factor, fusing the dynamic attenuation curve of the battery chemical characteristic parameter to generate a value vector sequence carrying a temperature compensation gradient feature; Performing group attention calculation processing on the query sub-vector group, the key vector block and the value vector sequence, and adjusting the correlation strength between different road slope classification numbers and temperature ranges through a dynamic weight allocation mechanism to generate a group attention weight matrix; The grouped attention weight matrix is ​​bidirectionally concatenated along the channel dimension, and the space-temperature feature aggregation process is performed on the concatenated tensor using the hierarchical fusion mechanism of the multi-scale convolution kernel to generate a space-temperature coupling tensor; Constructing a dynamic temperature decay mask based on the time series dimension of the decay factor tensor, and performing an element-wise product operation on the space-temperature coupling tensor to generate an enhanced attention weight matrix; The enhanced attention weight matrix is ​​residually connected with the four-dimensional coupling tensor, and the compensated tensor is subjected to cross-dimensional feature fusion through a batch normalization layer to generate an attention weight matrix that binds multi-source features.

[0013] In a second aspect, an embodiment of the present application provides a system for drawing an electric vehicle regional energy consumption map, comprising: An acquisition module is used to acquire multi-source data streams generated during the driving of the electric vehicle, and access an external meteorological interface to acquire temperature gradient parameters of the environment, wherein the multi-source data streams include longitude, latitude, cumulative mileage, state of charge, and total energy fields of the on-board energy storage device in the format of the national recommended standard GB 32960; A generation module is used to construct a dynamic grid generation model based on the administrative division coding system and the road topology network, and use an adaptive density adjustment algorithm to map trajectory points to spatial grid cells with hierarchical associations, wherein the spatial grid cells are bound to a road slope distribution matrix, a temperature gradient parameter, and a traffic flow weight coefficient; an execution module, configured to perform multimodal segmentation on the driving segments crossing the grid boundary in the spatial grid unit by using trajectory continuity analysis, triggering segment reconstruction with energy conservation constraints when the trajectory points enter different spatial grid units, and generating a feature vector including a grid code, a slope change rate, a temperature gradient vector correlation factor, and a traffic flow weight coefficient for each sub-segment; A calculation module, used to establish a mapping relationship between the characteristic vector and the grid energy consumption contribution, calculate the original energy consumption value based on the difference between the total energy of the energy storage device and the accumulated mileage, perform nonlinear compensation correction in combination with the slope change rate, and superimpose the temperature gradient vector association factor and the traffic flow weight coefficient to generate the dynamic energy consumption benchmark value of multiple vehicle types in the spatial grid unit; The analysis module is used to construct a spatiotemporal coupled diffusion model, perform correlation analysis on the dynamic energy consumption benchmark value, the charging pile distribution density and the traffic flow weight coefficient, and generate a three-dimensional dynamic map including an energy consumption thermal layer, a charging facility density layer and a traffic flow correlation layer through a multi-resolution rendering engine.

[0014] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for drawing an electric vehicle regional energy consumption map as described in the first aspect above.

[0015] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a method for drawing an electric vehicle regional energy consumption map as described in the first aspect.

[0016] In an embodiment of the present application, a multi-source data stream generated during the driving of an electric vehicle is obtained, and an external meteorological interface is connected to obtain the temperature gradient parameters of the environment, wherein the multi-source data stream includes longitude, latitude, cumulative mileage, state of charge and total energy fields of the on-board energy storage device in the format of the national recommended standard GB 32960; a dynamic grid generation model is constructed based on the administrative division coding system and the road topology network, and an adaptive density adjustment algorithm is used to map trajectory points to spatial grid units with hierarchical associations, and the spatial grid units are bound to the road slope distribution matrix, temperature gradient parameters and traffic flow weight coefficients; trajectory continuity analysis is used to perform multimodal segmentation on the driving segments that cross the grid boundaries in the spatial grid units, and when the trajectory points enter different spatial grid units, the segment reconstruction with energy conservation constraints is triggered, and a grid code, slope change rate, temperature and other parameters are generated for each sub-segment. The characteristic vector of the gradient vector association factor and the traffic flow weight coefficient is established; a mapping relationship between the characteristic vector and the grid energy consumption contribution is established, the original energy consumption value is calculated based on the difference between the total energy of the energy storage device and the accumulated mileage, and a nonlinear compensation correction is performed in combination with the slope change rate, and the temperature gradient vector association factor and the traffic flow weight coefficient are superimposed to generate the dynamic energy consumption benchmark value of multiple models in the spatial grid unit; a spatiotemporal coupled diffusion model is constructed, and the dynamic energy consumption benchmark value is correlated with the charging pile distribution density and the traffic flow weight coefficient. A three-dimensional dynamic map containing an energy consumption thermal layer, a charging facility density layer and a traffic flow correlation layer is generated through a multi-resolution rendering engine.

[0017] The technical solution of this application has the following beneficial effects: This application uses multi-source data streams, environmental parameters and advanced algorithm models to achieve accurate quantification and visualization of electric vehicle energy consumption, which not only helps to improve the travel experience of electric vehicle users, but also guides urban planners to rationally configure charging facilities and promote the popularization and development of new energy vehicles. In addition, by adjusting and optimizing the energy consumption benchmark value in real time, the practicality and foresight of the system are further enhanced, which is of great significance for promoting green transportation.

[0018] Furthermore, by establishing a mapping relationship between the feature vector and the grid energy consumption contribution, the method first calculates the original energy consumption value based on the difference between the total energy of the energy storage device and the accumulated mileage, and performs nonlinear compensation correction in combination with the slope change rate. At the same time, the ambient temperature, battery temperature rise gradient and thermal management power time series data are integrated, and the dynamic temperature correction coefficient is generated using multi-dimensional feature fusion technology. In addition, a traffic flow weight coefficient containing congestion intensity classification parameters and speed fluctuation correction factors is constructed through an adaptive threshold iteration algorithm and associated with the road type code. Finally, the slope compensation coefficient set, dynamic temperature correction coefficient and traffic flow weight coefficient are collaboratively corrected using a tensor product operation, and a dynamic energy consumption benchmark value is output, and the vehicle dynamic characteristic parameters are used as normalization adjustment factors for dynamic calibration.

[0019] The above method can significantly improve the accuracy and reliability of electric vehicle energy consumption estimation. By comprehensively considering multiple influencing factors such as slope, temperature change, and traffic flow, and using advanced algorithm models to quantify and compensate for these factors, the energy consumption benchmark value is closer to the actual driving situation. Furthermore, this accurate energy consumption estimation not only helps to optimize the layout of charging facilities, but also effectively improves the travel experience of electric vehicle users, reduces mileage anxiety, and promotes the popularization and development of new energy vehicles. In addition, by dynamically adjusting and optimizing the energy consumption benchmark value, the practicality and foresight of the system are further enhanced.

[0020] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 A flow chart showing a method for drawing an electric vehicle regional energy consumption map provided by the present application is shown; Figure 2 A schematic diagram of the structure of a system for drawing an electric vehicle regional energy consumption map provided by the present application is shown; Figure 3 A schematic diagram of the structure of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0024] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0025] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0026] Figure 1 A flowchart of a method for drawing an electric vehicle regional energy consumption map is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes: 101. Obtain a multi-source data stream generated during the driving of the electric vehicle, and access an external meteorological interface to obtain a temperature gradient parameter of the environment, wherein the multi-source data stream includes longitude, latitude, cumulative mileage, state of charge, and total energy fields of the on-board energy storage device in the format of the national recommended standard GB 32960; Among them, multi-source data streams: are various types of data collected from the driving process of electric vehicles; including location information, speed, battery status, etc.; used to analyze vehicle operating conditions and energy consumption; External meteorological interface: a technical means to access data from meteorological services and provide real-time or historical weather information; Temperature gradient parameter: refers to the temperature changes in different geographical locations or time periods; it is used to correct environmental factors in energy consumption calculations.

[0027] In actual operation, a basic data set for energy consumption assessment was preliminarily constructed by accessing data in the format of the national recommended standard GB 32960 and combining it with an external meteorological interface to obtain ambient temperature information.

[0028] For example, in a city traffic management system, the GPS location (longitude, latitude), cumulative mileage, state of charge, and total energy of the on-board energy storage device of all registered electric vehicles during daily driving are first collected. At the same time, through cooperation with the local weather station, real-time temperature gradient data covering the entire city is obtained. These data are integrated into a central database and prepared for subsequent energy consumption analysis.

[0029] 102. A dynamic grid generation model is constructed based on the administrative division coding system and the road topology network, and the trajectory points are mapped to spatial grid units with hierarchical associations using an adaptive density adjustment algorithm. The spatial grid units are bound to the road slope distribution matrix, temperature gradient parameters, and traffic flow weight coefficients; Among them, the administrative division coding system is a system of numbering according to geographical areas; it is used to identify different administrative regions; Road topology network: refers to a structured map consisting of a road network; it is used to describe the connection relationship between roads; Dynamic grid generation model: a spatial partitioning method based on the above two types of information that can adapt to data distributions of different densities; used to improve data processing efficiency; Adaptive density adjustment algorithm: a technique that automatically adjusts the grid size according to the density of data points; ensuring that trajectory points can be accurately mapped to corresponding grid cells; Spatial grid unit: It is a small area with hierarchical association that is finally formed; it is bound to a specific road slope, temperature gradient and traffic flow weight coefficient.

[0030] In actual operation, a dynamic grid model is established using administrative divisions and road topology networks, and the grid size is optimized through an adaptive algorithm to ensure that trajectory points can be accurately mapped to the corresponding grid cells.

[0031] Based on the data collected in the above examples, a coding system based on administrative districts was defined within the city, and a detailed topological network containing major roads was established. The network and adaptive algorithm were used to divide the city into multiple small grid cells, each of which was bound to a specific road slope, temperature gradient, and traffic flow weight coefficient, laying the foundation for subsequent analysis.

[0032] 103. Perform multimodal segmentation on the driving segments that cross the grid boundary in the spatial grid unit by using trajectory continuity analysis, trigger segment reconstruction with energy conservation constraints when the trajectory points enter different spatial grid units, and generate a feature vector including a grid code, a slope change rate, a temperature gradient vector correlation factor, and a traffic flow weight coefficient for each sub-segment; Among them, trajectory continuity analysis: is the study of the consistency and coherence of the vehicle's driving path; it is used to detect and handle trajectory interruptions; Driving segments that cross grid boundaries: refers to the driving segments when the vehicle moves from one grid to another; requires special processing to ensure data consistency; Multimodal segmentation: a method to classify paths according to different driving modes; used to more accurately simulate actual driving behavior; Fragment reconstruction of energy conservation constraints: to ensure the accuracy of energy calculation when switching between different grids; to ensure the correctness of energy consumption estimation; Feature vector: Contains various parameters that describe the characteristics of each driving segment; such as grid code, slope change rate, etc.

[0033] In actual operation, by analyzing the trajectory points, the situations crossing the grid boundary are identified and handled to ensure that each crossing can correctly reflect the change in energy consumption and generate a detailed feature vector for each sub-segment.

[0034] Continuing with the above embodiment, the system monitors an electric vehicle driving in the urban area. Whenever it enters a new grid cell, the system automatically performs an energy conservation check and updates the feature vector according to the characteristics of the current road section, including key information such as grid code and slope change rate, so as to prepare for more accurate energy consumption prediction.

[0035] 104. Establish a mapping relationship between the characteristic vector and the grid energy consumption contribution, calculate the original energy consumption value based on the difference between the total energy of the energy storage device and the accumulated mileage, perform nonlinear compensation correction in combination with the slope change rate, and superimpose the temperature gradient vector association factor and the traffic flow weight coefficient to generate the dynamic energy consumption benchmark value of multiple vehicle models in the spatial grid unit; Among them, grid energy consumption contribution refers to the influence of each grid unit on the overall energy consumption; it is used to evaluate the energy consumption of each area; Original energy consumption value: It is the basic energy consumption index calculated based on the difference between the total energy of the energy storage device and the accumulated mileage; it serves as the basis for energy consumption estimation; Non-linear compensation correction: takes into account the actual impact of factors such as slope on energy consumption; used to adjust energy consumption estimation results; Dynamic energy consumption benchmark value: a more accurate energy consumption estimate obtained after comprehensive consideration of multiple factors; reflects the energy consumption level under actual driving conditions.

[0036] In actual operation, by establishing the relationship between the characteristic vector and the grid energy consumption contribution, a variety of correction methods are used to calculate the dynamic energy consumption benchmark value that is closer to the actual situation.

[0037] Based on the information collected in the above embodiment, the system starts to calculate the energy consumption contribution of each vehicle in each grid unit, and makes corrections based on the slope, temperature and other factors to obtain an accurate energy consumption benchmark value, which helps to better understand the actual energy consumption of electric vehicles under different conditions.

[0038] 105. Construct a spatiotemporal coupled diffusion model, perform correlation analysis on the dynamic energy consumption benchmark value, the charging pile distribution density and the traffic flow weight coefficient, and generate a three-dimensional dynamic map including an energy consumption thermal layer, a charging facility density layer and a traffic flow correlation layer through a multi-resolution rendering engine.

[0039] Among them, the spatiotemporal coupled diffusion model is a method that combines time and space dimensions to analyze data propagation; it is used to handle complex data distribution; Charging pile distribution density: reflects the number of charging facilities in the area; used to evaluate the rationality of the layout of charging facilities; Multi-resolution rendering engine: a technology that can generate high-quality visual output; used to display information such as energy consumption thermal layer, charging facility density layer, etc. Three-dimensional dynamic map: It is the final result; it displays information on multiple aspects such as energy consumption thermal layer, charging facility density layer and traffic flow correlation layer.

[0040] In actual operation, the space-time coupled diffusion model is used, combined with the distribution density of charging piles and the traffic flow weight coefficient, and a three-dimensional dynamic map that comprehensively displays the energy consumption is generated through a multi-resolution rendering engine.

[0041] Continuing with the above embodiment, finally, the system integrates the data from all previous steps and generates a three-dimensional energy consumption map covering the entire city using the spatiotemporal coupled diffusion model. This map not only shows the energy consumption hotspots in each area, but also marks the location distribution of charging piles, helping city planners make more informed decisions, such as where to add new charging stations to relieve pressure during peak hours.

[0042] Through the implementation of steps 101 to 105, the method effectively improves the accuracy of electric vehicle energy consumption estimation and enhances the understanding of energy consumption under complex driving conditions. It not only provides a scientific basis for optimizing the layout of charging facilities, but also significantly improves the practicality and foresight of the system through refined spatial grid unit division and fragment reconstruction of energy conservation constraints. The application of this method makes the management of new energy vehicles more efficient and promotes the development of green transportation.

[0043] In order to solve the complexity of multiple factors in energy consumption estimation and further improve the accuracy and adaptability of the calculation of the energy consumption benchmark value of electric vehicles, in some embodiments, in step 104, a mapping relationship between the characteristic vector and the grid energy consumption contribution is established, the original energy consumption value is calculated based on the difference between the total energy of the energy storage device and the accumulated mileage, a nonlinear compensation correction is performed in combination with the slope change rate, and the temperature gradient vector association factor and the traffic flow weight coefficient are superimposed to generate the dynamic energy consumption benchmark value of multiple models in the spatial grid unit, including: The original energy consumption value is calculated based on the difference between the total energy of the energy storage device and the accumulated mileage, the slope change rate data extracted from the road slope distribution matrix is ​​integrated, and the historical road topology data and the energy consumption deviation are back-propagated and trained using a multi-interval segmented optimization algorithm to generate a set of slope compensation coefficients; the ambient temperature, battery temperature rise gradient and thermal management power time series data in the temperature gradient parameters are integrated to construct a generation module of temperature influencing factors, and a dynamic temperature correction coefficient is output using a multi-dimensional feature fusion technology; the historical congestion mode database of electric vehicles is integrated, and a traffic flow weight coefficient is constructed through an adaptive threshold iteration algorithm. The traffic flow weight coefficient includes a congestion intensity classification parameter and a speed fluctuation correction factor, and a cross-dimensional association relationship is established with the road type code; a multi-source parameter collaborative correction operation is performed, the slope compensation coefficient set and the dynamic temperature correction coefficient are tensor-multiplied, the traffic flow weight coefficient is superimposed, a dynamic energy consumption benchmark value is output, and the vehicle power characteristic parameter is used as a normalization adjustment factor to complete the dynamic calibration of the dynamic energy consumption benchmark value.

[0044] In this embodiment, the slope change rate is data extracted from the road slope distribution matrix; it is used to describe the degree of inclination of different road sections; it has a direct impact on vehicle energy consumption; Multi-interval segmentation optimization algorithm: a method that divides data into multiple intervals for optimization processing; used to reduce energy consumption deviation; generates a set of slope compensation coefficients through back-propagation training; Temperature impact factor: a module built by integrating ambient temperature, battery temperature rise gradient, and thermal management power timing data; used to correct energy consumption differences caused by temperature changes; Dynamic temperature correction coefficient: It is the result of the multi-dimensional feature fusion technology output; it is used to adjust the actual energy consumption of the vehicle under different temperature conditions; Traffic flow weight coefficient: includes congestion intensity classification parameters and speed fluctuation correction factors; establishes a cross-dimensional correlation with the road type code; reflects the impact of traffic flow on energy consumption under different road conditions; Tensor product operation: a mathematical operation method; used to process the interaction between multidimensional data; in this scenario, it combines the slope compensation coefficient set and the dynamic temperature correction coefficient; Normalization adjustment factor: generated based on vehicle dynamic characteristic parameters; used to ensure the comparability of dynamic energy consumption benchmark values ​​between different models.

[0045] In the embodiment of the present application, firstly, the original energy consumption value is calculated based on the difference between the total energy of the energy storage device and the accumulated mileage.

[0046] Then, the road slope data is fused using a multi-interval segmented optimization algorithm to generate a set of slope compensation coefficients, and the dynamic temperature correction coefficient is generated by combining the temperature data with multi-dimensional feature fusion technology.

[0047] Next, the historical congestion pattern database is integrated and the traffic flow weight coefficient is generated using an adaptive threshold iterative algorithm.

[0048] Finally, a multi-source parameter collaborative correction operation is performed to combine all these correction coefficients and weight factors to generate the final dynamic energy consumption benchmark values ​​of multiple vehicle models within the spatial grid unit.

[0049] Here is a specific example: In the intelligent traffic management system of a large city, the system first calculates the basic energy consumption value of each electric vehicle based on its driving data (such as the difference between the total energy of the energy storage device and the accumulated mileage). Next, by analyzing the slope change rate on each road in the city and applying a multi-interval segmented optimization algorithm, the system can generate corresponding slope compensation coefficients for different sections. At the same time, the system accesses the real-time temperature data provided by the meteorological station, and combines the vehicle battery temperature rise gradient and the working status of the thermal management system to calculate the dynamic temperature correction coefficient to accurately reflect the impact of temperature on energy consumption.

[0050] In addition, the system integrates a detailed historical congestion pattern database and generates traffic flow weight coefficients based on traffic conditions at different time periods and road sections through an adaptive threshold iteration algorithm. These weight coefficients not only take into account the degree of road congestion, but also combine the speed fluctuations of vehicles and the specific attributes of the road type.

[0051] Finally, the system combines all the above correction coefficients and weight factors through tensor product operations to generate a dynamic energy consumption benchmark value within each spatial grid unit. In order to ensure the consistency of this value among different types of vehicles, the system will also use the vehicle's dynamic characteristic parameters as a normalized adjustment factor for final adjustment. In this way, urban planners can optimize the layout of charging facilities based on these precise energy consumption benchmark values ​​and improve the level of green traffic management in the entire city.

[0052] In order to solve the problem of complexity of multi-source parameter collaborative correction operation and further improve the accuracy and adaptability of dynamic energy consumption benchmark value calculation, in the above embodiment, multi-source parameter collaborative correction operation is performed, the slope compensation coefficient set and the dynamic temperature correction coefficient are subjected to tensor product operation, the traffic flow weight coefficient is superimposed, the dynamic energy consumption benchmark value is output, and the vehicle dynamic characteristic parameter is used as a normalization adjustment factor to complete the dynamic calibration of the dynamic energy consumption benchmark value, including: Construct a three-dimensional tensor operation channel to perform tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient, wherein the two-dimensional matrix dimension of the slope compensation coefficient set corresponds to the number of road slope classifications and the number of historical road topology data codes, and the dynamic temperature correction coefficient matrix dimension corresponds to the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications; decompose the traffic flow weight coefficient into a congestion intensity component and a speed fluctuation component, wherein the congestion intensity component is associated with the administrative division restriction strategy in the road type code, and the speed fluctuation component is bound to the speed-energy consumption mapping relationship table of the road topology network; use the vehicle dynamic characteristic parameters to generate a normalized adjustment factor, wherein the vehicle dynamic characteristic parameters include the total energy field of the energy storage device and the mass The energy consumption conversion matrix is ​​obtained by converting the energy consumption of the charging pile into an energy-consumption conversion matrix, and the normalization adjustment factor acts on the result of the tensor product operation and the superposition process of the traffic flow weight component; the closed-loop calibration operation of the dynamic energy consumption benchmark value is performed, and the output result of the calibration operation is synchronously input into the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit, wherein the charging pile distribution density calculation unit is associated with the spatial coding attribute of the road topology network, and the energy consumption thermal value diffusion unit is bound to the temperature change cycle data of the meteorological interface; the slope compensation coefficient set, the dynamic temperature correction coefficient, the traffic flow weight coefficient and the vehicle dynamic characteristic parameter form a cross-level data pipeline in the spatiotemporal coupling diffusion model, and all intermediate operation results act on the downstream calculation module through the convolutional neural network channel.

[0053] In this embodiment, the three-dimensional tensor operation channel: is a mathematical operation method for processing the interaction between multi-dimensional data; is used to perform a complex tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient; Slope compensation coefficient set: its two-dimensional matrix dimension corresponds to the number of road slope classifications and the number of historical road topology data codes; it is used to adjust the energy consumption estimation under different road sections; Congestion intensity component: It is a part of the traffic flow weight coefficient; it is related to the administrative division restriction strategy in the road type code; it reflects the impact of congestion level on energy consumption; Speed ​​fluctuation component: also part of the traffic flow weight coefficient; bound to the speed-energy consumption mapping relationship table of the road topology network; reflects the impact of speed changes on energy consumption; Closed-loop calibration operation: a self-correction mechanism; used to dynamically adjust and optimize the energy consumption baseline value of the final output; Charging pile distribution density calculation unit: associates the spatial coding attributes of the road topology network; used to evaluate the rationality of the layout of charging facilities in the region; Energy consumption thermal value diffusion unit: binds the temperature change cycle data of the meteorological interface; used to display energy consumption hot spots and their changing trends over time.

[0054] In the embodiment of the present application, first, a three-dimensional tensor operation channel is constructed to perform a tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient.

[0055] Then, the traffic flow weight coefficient is decomposed into a congestion intensity component and a speed fluctuation component, and adjusted in combination with the road type coding and the speed-energy consumption mapping relationship table.

[0056] Next, the vehicle dynamic characteristic parameters are used to generate a normalized adjustment factor, which acts on the superposition process of the tensor product operation result and the traffic flow weight component.

[0057] Finally, a closed-loop calibration operation of the dynamic energy consumption benchmark value is performed, and the output result is simultaneously input into the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit to form a cross-level data pipeline, and acts on the downstream computing module through the convolutional neural network channel.

[0058] Here is a specific example: In an intelligent traffic management system for a large city, the system first constructed a three-dimensional tensor operation channel, which performed complex tensor product operations on the slope compensation coefficient set (based on the number of road slope classifications and historical road topology data encoding) and the dynamic temperature correction coefficient (based on the number of temperature compensation intervals and battery chemical characteristic parameter classification) of each electric vehicle on different road sections. Then, the system decomposed the traffic flow weight coefficient into a congestion intensity component and a speed fluctuation component. The former takes into account the restriction strategy of the road type, and the latter is adjusted according to the speed-energy consumption mapping relationship table.

[0059] In order to ensure the comparability of energy consumption benchmarks between different models, the system uses the vehicle's dynamic characteristic parameters to generate normalized adjustment factors. These parameters include the total energy field of the energy storage device and the mass-energy consumption conversion matrix, which are applied to the superposition process of the result of the tensor product operation and the traffic flow weight component to complete the final dynamic energy consumption benchmark calculation.

[0060] Subsequently, the system performed a closed-loop calibration operation, and the output results were not only used to update the charging pile distribution density calculation unit, but also sent to the energy consumption thermal value diffusion unit, which combined with the temperature change cycle data provided by the meteorological station to display the energy consumption hotspots and their changing trends in various areas of the city. This precise energy consumption benchmark value calculation method greatly improves the efficiency of the new energy vehicle management system, helps urban planners better layout charging facilities, and promotes the development of green transportation.

[0061] In order to solve the problems of high complexity and insufficient precision in the collaborative processing of multi-source parameters and further improve the accuracy and adaptability of the calculation of the dynamic energy consumption benchmark value, in the above-mentioned multiple embodiments, a three-dimensional tensor operation channel is constructed to perform a tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient, wherein the two-dimensional matrix dimension of the slope compensation coefficient set corresponds to the number of road slope classifications and the number of historical road topology data codes, and the dynamic temperature correction coefficient matrix dimension corresponds to the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications, including: The road slope classification number and the historical road topology data encoding number are used to construct a two-dimensional matrix of the slope compensation coefficient set, and the temperature compensation interval number and the battery chemical characteristic parameter classification number are used to construct a two-dimensional matrix of the dynamic temperature correction coefficient, so as to perform tensor product operation to generate a three-dimensional compensation tensor structure, wherein the dimension of the three-dimensional compensation tensor structure includes the product combination of the road slope classification number, the temperature compensation interval number and the battery chemical characteristic parameter classification number; the three-dimensional compensation tensor structure is subjected to multi-level feature fusion processing, the cross-validation results of the road slope classification number and the historical road topology data encoding are fused, and the battery chemical characteristic dynamic attenuation factor of the two-dimensional matrix of the dynamic temperature correction coefficient is injected to generate a collaborative compensation tensor; the collaborative compensation tensor is subjected to multi-source data superposition with the congestion intensity component and the speed fluctuation component of the traffic flow weight component to generate a dynamic multi-source fusion tensor; the mass-energy in the vehicle dynamic characteristic parameters is used to generate a dynamic multi-source fusion tensor. The energy consumption conversion matrix performs a phased linear transformation on the dynamic multi-source fusion tensor to generate a linear mapping image vector, wherein the phased linear transformation includes the capacity attenuation compensation of the total energy field of the energy storage device and the gradient correction operation of the mass-energy consumption conversion matrix; the linear mapping image vector is input into the spatiotemporal coupled diffusion model for multi-scale feature calibration, and the multi-scale feature calibration is used to associate the spatial coding attributes of the road topology network with the temperature change periodic data fluctuation curve of the meteorological interface to generate a multi-dimensional output sequence carrying a dynamic energy consumption benchmark value; the multi-dimensional output sequence is transmitted to the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit through the convolutional neural network channel, wherein the charging pile distribution density calculation unit integrates the spatial coding attributes of the road topology network and the spatial distribution weight of the dynamic energy consumption benchmark value, and the energy consumption thermal value diffusion unit binds the temperature change periodic data with the attenuation correlation factor of the battery chemical characteristic parameter classification number.

[0062] In this embodiment, the three-dimensional compensation tensor structure is a data structure generated by tensor product operation; its dimensions include the product combination of the number of road slope classifications, the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications; it is used to comprehensively consider multiple influencing factors; Multi-level feature fusion processing: a method of integrating data features from different sources; used to enhance the expressiveness and accuracy of data; in this scenario, the cross-validation results of road slope classification and historical road topology data encoding are fused; Synergistic compensation tensor: a data structure generated after multi-level feature fusion processing; a dynamic attenuation factor of battery chemical characteristics in a two-dimensional matrix injected with a dynamic temperature correction coefficient; used to more accurately reflect the impact of various factors on energy consumption; Dynamic multi-source fusion tensor: It is the result of superimposing the synergy compensation tensor with the traffic flow weight component (congestion intensity component and speed fluctuation component); it is used to comprehensively consider multiple influencing factors; Mass-energy consumption conversion matrix: a matrix generated based on vehicle dynamic characteristic parameters; used to describe the relationship between vehicle mass and energy consumption; in this process, capacity attenuation compensation and gradient correction operations are performed; Multi-scale feature calibration: a self-adjusting mechanism; used to optimize the energy consumption benchmark value of the final output; associate the spatial encoding attributes of the road topology network with the temperature change cycle data fluctuation curve of the meteorological interface; Multidimensional output sequence: is the result after being processed by the spatiotemporal coupled diffusion model; carries dynamic energy consumption benchmark values; is used for subsequent analysis and decision support.

[0063] In the embodiment of the present application, first, a three-dimensional tensor operation channel is constructed, and a tensor product operation is performed on the slope compensation coefficient set and the dynamic temperature correction coefficient to generate a three-dimensional compensation tensor structure.

[0064] Then, the structure is subjected to multi-level feature fusion processing, and the dynamic attenuation factor of battery chemical characteristics is injected to generate a synergistic compensation tensor.

[0065] Then, the synergy compensation tensor is superimposed with the traffic flow weight component to generate a dynamic multi-source fusion tensor. The dynamic multi-source fusion tensor is linearly transformed in stages using the mass-energy consumption conversion matrix to generate a linear mapping image vector.

[0066] Finally, the vector is input into the spatiotemporal coupled diffusion model for multi-scale feature calibration to generate a multi-dimensional output sequence, which is then transmitted to the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit through the convolutional neural network channel.

[0067] Here is a specific example: In an intelligent traffic management system for a large city, the system first built a three-dimensional tensor operation channel, which performs complex tensor product operations on the slope compensation coefficient set of each electric vehicle on different road sections (based on the number of road slope classifications and historical road topology data encoding) and the dynamic temperature correction coefficient (based on the number of temperature compensation intervals and battery chemical characteristic parameter classification) to generate a three-dimensional compensation tensor structure. Then, the system performs multi-level feature fusion processing on this structure, combines the cross-validation results of the road slope classification number and the historical road topology data encoding, and injects the dynamic attenuation factor of the battery chemical characteristics to generate a synergistic compensation tensor.

[0068] Subsequently, the system superimposes the synergistic compensation tensor with the traffic flow weight component (including the congestion intensity component and the speed fluctuation component) to generate a dynamic multi-source fusion tensor. In order to further refine it, the system uses the mass-energy consumption conversion matrix generated by the vehicle's dynamic characteristic parameters to perform a phased linear transformation on the dynamic multi-source fusion tensor to generate a linear mapping image vector. This process includes the capacity attenuation compensation of the total energy field of the energy storage device and the gradient correction operation of the mass-energy consumption conversion matrix.

[0069] Finally, the system inputs the generated linear mapping image vector into the spatiotemporal coupled diffusion model for multi-scale feature calibration, associates the spatial encoding attributes of the road topology network with the temperature change periodic data fluctuation curve of the meteorological interface, and generates a multi-dimensional output sequence with dynamic energy consumption benchmark values. These output sequences are transmitted to the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit through the convolutional neural network channel, helping urban planners to better layout charging facilities and display energy consumption hotspots and their changing trends in various areas of the city. This precise energy consumption benchmark value calculation method has greatly improved the efficiency of the new energy vehicle management system and promoted the development of green transportation.

[0070] In order to solve the complexity and precision problems in the multi-source data fusion process and further improve the accuracy and adaptability of the dynamic energy consumption benchmark value calculation, in the above-mentioned multiple embodiments, the three-dimensional compensation tensor structure is subjected to multi-level feature fusion processing, the cross-validation results of the road slope classification number and the historical road topology data encoding are fused, and the dynamic attenuation factor of the battery chemical characteristics of the two-dimensional matrix of the dynamic temperature correction coefficient is injected to generate a collaborative compensation tensor, including: A cross-validation matrix is ​​constructed based on the data encoded by the road slope classification number and the historical road topology data, and a multi-channel convolution layer is used to extract the associated features of the cross-validation matrix to generate an initial fusion weight matrix that matches the subsequent tensor operation dimension; according to the battery chemical characteristic parameter classification dimension of the dynamic temperature correction coefficient matrix, the initial fusion weight matrix is ​​decomposed into an attenuation factor vector, and a dynamic attenuation coefficient calculation process is performed on each temperature compensation interval in combination with a time series attenuation model to generate an attenuation factor tensor carrying a temperature gradient associated factor; the initial fusion weight matrix is ​​combined with the attenuation factor tensor using a tensor broadcast mechanism. The sub-tensors are processed by element-by-element product operation to generate a coupled tensor containing a four-dimensional structure; cross-dimensional feature alignment processing is performed on the coupled tensor based on the multi-dimensional attention mechanism, and a query vector is generated by using the features of the road slope classification number, a key vector is generated by using the features of the historical road topology data encoding, and a value vector is generated by using the temperature attenuation factor; the query vector, the key vector and the value vector are normalized and weighted to generate an attention weight matrix; the three-dimensional compensation tensor is residually connected with the attention weight matrix, and feature fusion processing is performed on the connection result using a batch normalization layer to generate a collaborative compensation tensor containing a four-dimensional product combination.

[0071] In this embodiment, the cross-validation matrix: a matrix constructed based on the data encoded by the road slope classification number and the historical road topology data; used to verify the correlation between different road sections; Multi-channel convolutional layer: A deep learning technique used to extract relevant features from the cross-validation matrix; generate an initial fusion weight matrix that matches the dimensions of subsequent tensor operations; Attenuation factor vector decomposition processing: method for decomposing the initial fusion weight matrix; classifying dimensions according to battery chemical characteristic parameters; calculating the dynamic attenuation coefficient of each temperature compensation interval in combination with the time series attenuation model; Attenuation factor tensor: a data structure containing temperature gradient correlation factors; combined with the initial fusion weight matrix through element-by-element product operation; generating a coupling tensor; Tensor broadcast mechanism: a mathematical operation method used to expand data between different dimensions; in this scenario, it implements the element-by-element product of the initial fusion weight matrix and the attenuation factor tensor; The coupled tensor of the four-dimensional structure is a data structure generated after the element-by-element product operation; it contains the information of the four-dimensional combination; it is used for cross-dimensional feature alignment; Multi-dimensional attention mechanism: a self-adjusting mechanism used to enhance feature representation on a specific dimension; in this scenario, query vector, key vector and value vector are used for normalized weighting; Attention weight matrix: is the result generated by the multi-dimensional attention mechanism; it is used to perform feature fusion on the coupled tensor; Residual connection processing: A neural network technique used to preserve the original input information; in this case, the 3D compensation tensor is combined with the attention weight matrix; Batch Normalization Layer: A technique used to normalize data; optimizing the feature fusion results in the process; generating the final co-compensation tensor.

[0072] In an embodiment of the present application, first, a cross-validation matrix is ​​constructed based on the data encoded by the road slope classification number and the historical road topology data, and a multi-channel convolutional layer is used to perform feature extraction to generate an initial fusion weight matrix.

[0073] Next, according to the battery chemical characteristic parameter classification dimension of the dynamic temperature correction coefficient matrix, the initial fusion weight matrix is ​​decomposed into attenuation factor vectors, and the dynamic attenuation coefficient of each temperature compensation interval is calculated in combination with the time series attenuation model to generate the attenuation factor tensor.

[0074] Then, the tensor broadcast mechanism is used to perform element-by-element multiplication of the initial fusion weight matrix and the attenuation factor tensor to generate a coupling tensor with a four-dimensional structure.

[0075] Next, based on the multi-dimensional attention mechanism, cross-dimensional feature alignment is performed on the coupling tensor to generate an attention weight matrix.

[0076] Finally, the three-dimensional compensation tensor is residually connected with the attention weight matrix, and the batch normalization layer is used to perform feature fusion processing on the connection result to generate a collaborative compensation tensor.

[0077] Here is a specific example: In a large city’s intelligent traffic management system, the system first built a cross-validation matrix based on the number of road slope classifications and historical road topology data encoding for each electric vehicle’s driving path. In order to extract key features from this data, the system used a multi-channel convolutional layer to generate an initial fusion weight matrix that matched the dimensions of subsequent tensor operations.

[0078] Next, the system performs attenuation factor vector decomposition on the initial fusion weight matrix according to the battery chemical characteristic parameter classification dimension of the dynamic temperature correction coefficient matrix. Combined with the time series attenuation model, the system calculates the dynamic attenuation coefficient in each temperature compensation interval and generates an attenuation factor tensor carrying the temperature gradient correlation factor. Subsequently, the system uses the tensor broadcast mechanism to perform an element-by-element product operation on the initial fusion weight matrix and the attenuation factor tensor, generating a coupling tensor containing a four-dimensional structure.

[0079] To further enhance the expressiveness of the data, the system performs cross-dimensional feature alignment on the coupled tensor based on a multi-dimensional attention mechanism. In this process, the system uses the features of the road slope classification number to generate a query vector, the features of the historical road topology data encoding to generate a key vector, and the temperature attenuation factor to generate a value vector. By normalizing and weighting the query vector, key vector, and value vector, the system generates an attention weight matrix.

[0080] Finally, the system performs residual connection processing on the three-dimensional compensation tensor and the attention weight matrix, and uses the batch normalization layer to perform feature fusion processing on the connection result to generate a synergistic compensation tensor. This synergistic compensation tensor not only contains detailed energy consumption influencing factors, but also improves the overall performance of the system through precise data processing, allowing urban planners to more accurately evaluate and optimize the layout of charging facilities and promote the development of green transportation.

[0081] In order to solve the problem of the complexity of vehicle dynamic characteristic parameters in the process of calculating the dynamic energy consumption reference value and further improve the accuracy and adaptability of the linear mapping image vector generation, as another embodiment, the mass-energy consumption conversion matrix in the vehicle dynamic characteristic parameters is used to perform a staged linear transformation on the dynamic multi-source fusion tensor to generate a linear mapping image vector, wherein the staged linear transformation includes the capacity attenuation compensation of the total energy field of the energy storage device and the gradient correction operation of the mass-energy consumption conversion matrix, including: A dynamic capacity correction tensor is constructed based on the capacity attenuation compensation operation of the total energy field of the energy storage device, and the compensation operation is subjected to multi-dimensional parameter fusion processing using the time series attenuation characteristics of the historical state of charge data and the dynamic correlation factor of the temperature gradient parameter to generate a dynamic capacity correction matrix carrying capacity correction characteristics; a gradient correction operation is performed on the mass-energy consumption conversion matrix according to the road slope classification number and the temperature compensation interval parameter of the dynamic multi-source fusion tensor, and the associated weight vector of the vehicle mass parameter and the slope classification number is extracted to generate a corrected mass-energy consumption conversion matrix; the dynamic capacity correction matrix is ​​converted into The modified mass-energy consumption conversion matrix is ​​subjected to feature coupling processing to generate a mass-energy consumption correlation tensor including a capacity attenuation compensation feature and a road slope-mass coupling feature; the mass-energy consumption correlation tensor is subjected to multi-dimensional feature fusion processing based on a staged linear transformation. In the first stage, a primary mapping vector is generated by utilizing the superposition feature of the road slope classification number and the traffic flow speed fluctuation component. In the second stage, a dynamic residual connection is used to inject the capacity attenuation compensation parameter of the energy storage device into the primary mapping vector to generate an intermediate mapping tensor. In the third stage, a batch normalization layer is used to perform gradient correction processing on the intermediate mapping tensor to generate a linear mapping image vector.

[0082] In this embodiment, the dynamic capacity correction tensor: a data structure constructed based on the capacity decay compensation operation of the total energy field of the energy storage device; a multi-dimensional parameter fusion process is performed using the time series decay characteristics of the historical state of charge data and the dynamic correlation factor of the temperature gradient parameter; used to compensate for the decay of the battery capacity over time; Capacity Correction Feature: Information contained in the dynamic capacity correction matrix; used to adjust for the reduction in energy storage capacity due to battery aging; Road slope classification number: data indicating the degree of inclination of different road sections; used to adjust the mass-energy consumption conversion matrix to reflect the impact of road slope on energy consumption; Temperature compensation interval parameter: represents data in different temperature ranges; used to adjust the mass-energy consumption conversion matrix to reflect the impact of temperature changes on energy consumption; Mass-energy consumption correlation tensor: a data structure generated after tensor multiplication operation; contains capacity attenuation compensation features and road slope-mass coupling features; used to comprehensively consider multiple influencing factors; Phased linear transformation: a step-by-step optimization method; divided into three stages: primary mapping vector generation, dynamic residual connection and batch normalization layer processing; finally generating a linear mapping image vector; Primary mapping vector: the result generated in the first stage; using the superposition characteristics of the road slope classification number and the traffic flow speed fluctuation component; Intermediate mapping tensor: the result generated in the second stage; the energy storage device capacity decay compensation parameters are injected into the primary mapping vector through dynamic residual connection; Linearly mapped image vector: The final result; gradient correction is performed on the intermediate mapped tensor through a batch normalization layer.

[0083] In the embodiment of the present application, firstly, a dynamic capacity correction tensor is constructed based on the capacity decay compensation operation of the total energy field of the energy storage device, and multi-dimensional parameter fusion processing is performed using the time series decay characteristics of the historical state of charge data and the dynamic correlation factor of the temperature gradient parameter.

[0084] Next, according to the road slope classification number and temperature compensation interval parameters of the dynamic multi-source fusion tensor, a gradient correction operation is performed on the mass-energy consumption conversion matrix to generate a corrected mass-energy consumption conversion matrix.

[0085] Then, the dynamic capacity correction matrix and the corrected mass-energy consumption conversion matrix are characteristically coupled using tensor product operations to generate a mass-energy consumption correlation tensor.

[0086] Finally, multi-dimensional feature fusion processing is performed on the mass-energy consumption correlation tensor based on the staged linear transformation to generate the primary mapping vector, the intermediate mapping tensor and the linear mapping image vector in sequence.

[0087] Here is a specific example: In an intelligent traffic management system for a large city, the system first constructs a dynamic capacity correction tensor based on the total energy field of the energy storage device of each electric vehicle. In order to compensate for the decay of battery capacity over time, the system uses the time series decay characteristics of historical state of charge data and the dynamic correlation factors of temperature gradient parameters to perform multi-dimensional parameter fusion processing and generate a dynamic capacity correction matrix with capacity correction characteristics.

[0088] Next, the system performs a gradient correction operation on the mass-energy conversion matrix based on the road slope classification number and temperature compensation interval parameters of the dynamic multi-source fusion tensor. By extracting the associated weight vector of the vehicle mass parameter and the slope classification number, the system generates a corrected mass-energy conversion matrix. Subsequently, the system uses tensor product operations to feature-couple the dynamic capacity correction matrix with the corrected mass-energy conversion matrix, generating a mass-energy association tensor.

[0089] In order to further optimize this tensor, the system performs multi-dimensional feature fusion processing based on the staged linear transformation method. In the first stage, the system uses the superposition features of the road slope classification number and the traffic flow speed fluctuation component to generate a primary mapping vector. In the second stage, the system injects the capacity attenuation compensation parameters of the energy storage device into the primary mapping vector through a dynamic residual connection to generate an intermediate mapping tensor. In the third stage, the system uses a batch normalization layer to perform gradient correction processing on the intermediate mapping tensor, and finally generates a linear mapping image vector.

[0090] This precise linear mapping image vector not only reflects the impact of battery capacity attenuation and road slope on energy consumption, but also improves the overall performance of the system through multi-stage optimization processing. Urban planners can use this data to more accurately evaluate and optimize the layout of charging facilities, improve the management efficiency of new energy vehicles, and promote the development of green transportation.

[0091] In order to solve the complexity and precision problems of feature alignment in the process of multi-source data fusion and further improve the accuracy and adaptability of dynamic energy consumption benchmark value calculation, as another embodiment, cross-dimensional feature alignment processing is performed on the coupling tensor based on a multi-dimensional attention mechanism, a query vector is generated using the feature of the road slope classification number, a key vector is generated using the feature of the historical road topology data encoding, and a value vector is generated using the temperature attenuation factor. The query vector, the key vector and the value vector are normalized and weighted to generate an attention weight matrix, including: The height dimension of the four-dimensional coupling tensor is subjected to multi-level feature decomposition processing by using the road slope classification number, and the decomposed coarse-grained classification features and fine-grained fluctuation components are passed to the channel segmentation processing module to generate a query sub-vector group carrying multi-level slope features; spatial pooling processing is performed on the width dimension of the four-dimensional coupling tensor based on the historical road topology data encoding, and the extracted local correlation features and global distribution features are dynamically adjusted in combination with the time series attenuation model to generate a key vector block; dynamic masking operation is performed on the temperature gradient dimension of the four-dimensional coupling tensor according to the temperature attenuation factor, and the dynamic attenuation curve of the battery chemical characteristic parameters is integrated to generate a value vector sequence carrying a temperature compensation gradient feature; grouping is performed on the query sub-vector group, the key vector block and the value vector sequence Attention calculation processing, adjusting the correlation strength between different road slope grades and temperature ranges through a dynamic weight allocation mechanism to generate a grouped attention weight matrix; performing bidirectional splicing processing on the grouped attention weight matrix along the channel dimension, and using the hierarchical fusion mechanism of the multi-scale convolution kernel to perform space-temperature feature aggregation processing on the spliced ​​tensor to generate a space-temperature coupling tensor; constructing a dynamic temperature attenuation mask based on the time series dimension of the attenuation factor tensor, and performing element-level multiplication operation on the space-temperature coupling tensor to generate an enhanced attention weight matrix; performing residual connection processing on the enhanced attention weight matrix and the four-dimensional coupling tensor, and performing cross-dimensional feature fusion on the compensated tensor through a batch normalization layer to generate an attention weight matrix that binds multi-source features.

[0092] In this embodiment, multi-level feature decomposition processing: a method for decomposing complex features into coarse-grained and fine-grained features; in this scenario, it is used for the number of road slope classifications; Query sub-vector group: a set consisting of coarse-grained hierarchical features and fine-grained fluctuation components; used to generate query vectors; Spatial pooling: a technique for extracting local correlation features and global distribution features; based on historical road topology data encoding; Key vector block: a data structure containing local correlation features and global distribution features; used for feature alignment; Dynamic masking: a technique used to adjust the characteristics of a specific dimension; in this case, the temperature gradient dimension applied to the 4D coupled tensor; Value vector sequence: data structure carrying temperature compensation gradient characteristics; dynamic decay curve generation combined with battery chemical characteristic parameters; Grouped attention computation: A method to optimize the strength of associations between different features; achieved through a dynamic weight assignment mechanism; Group attention weight matrix: It is the result generated after group attention calculation and processing; it is used for subsequent feature fusion; Bidirectional stitching: A method of merging data from different directions; performed along the channel dimension; Spatial-Temperature Feature Aggregation Processing: A technique for integrating spatial and temperature features; using a multi-scale convolution kernel hierarchical fusion mechanism; Dynamic temperature decay mask: built on the time series dimension of the decay factor tensor; used for element-wise product operations; Enhanced attention weight matrix: the result after element-wise product operation; used for residual connection processing with the four-dimensional coupling tensor; Cross-dimensional feature fusion: A method for combining features from multiple dimensions; implemented via batch normalization layers.

[0093] In the embodiment of the present application, firstly, a multi-level eigendecomposition process is performed on the height dimension of the four-dimensional coupling tensor using the road slope classification number to generate a query sub-vector group.

[0094] Next, spatial pooling is performed on the width dimension of the four-dimensional coupling tensor based on the historical road topology data encoding, and a key vector block is generated in combination with the temporal decay model.

[0095] Then, a dynamic masking operation is performed on the temperature gradient dimension of the four-dimensional coupling tensor according to the temperature attenuation factor to generate a sequence of value vectors.

[0096] Next, the grouped attention calculation is performed on the query sub-vector group, key vector block, and value vector sequence to generate a grouped attention weight matrix. The grouped attention weight matrix is ​​bidirectionally concatenated along the channel dimension, and the hierarchical fusion mechanism of the multi-scale convolution kernel is used to aggregate the space-temperature features to generate a space-temperature coupling tensor. A dynamic temperature attenuation mask is constructed based on the time series dimension of the attenuation factor tensor, and an element-wise product operation is performed on the space-temperature coupling tensor to generate an enhanced attention weight matrix.

[0097] Finally, the enhanced attention weight matrix is ​​residually connected with the four-dimensional coupling tensor, and cross-dimensional feature fusion is performed through a batch normalization layer to generate the final attention weight matrix.

[0098] Here is a specific example: In an intelligent traffic management system for a large city, the system first uses the road slope classification number to perform multi-level feature decomposition on the height dimension of the four-dimensional coupling tensor, generating a query sub-vector group containing coarse-grained classification features and fine-grained fluctuation components. These features are passed to the channel segmentation processing module to generate a query sub-vector group carrying multi-level slope features.

[0099] Next, the system performs spatial pooling on the width dimension of the four-dimensional coupling tensor based on the historical road topology data encoding, extracts local correlation features and global distribution features in combination with the time series decay model, and performs dynamic weight adjustment to generate key vector blocks. At the same time, a dynamic mask operation is performed on the temperature gradient dimension of the four-dimensional coupling tensor according to the temperature decay factor, and the dynamic decay curve of the battery chemical characteristic parameters is integrated to generate a value vector sequence carrying the temperature compensation gradient feature.

[0100] Subsequently, the system performs group attention calculations on the query sub-vector group, key vector block, and value vector sequence, and adjusts the correlation strength between different road slope classification numbers and temperature ranges through a dynamic weight allocation mechanism to generate a group attention weight matrix. These matrices are bidirectionally concatenated along the channel dimension, and the space-temperature feature aggregation is performed using the hierarchical fusion mechanism of multi-scale convolution kernels to generate a space-temperature coupling tensor.

[0101] To further optimize this tensor, the system constructs a dynamic temperature decay mask based on the time series dimension of the decay factor tensor, and performs element-wise product operations on the space-temperature coupling tensor to generate an enhanced attention weight matrix. Finally, the system performs residual connection processing on the enhanced attention weight matrix and the four-dimensional coupling tensor, and performs cross-dimensional feature fusion through a batch normalization layer to generate an attention weight matrix that binds multi-source features.

[0102] This precise attention weight matrix not only reflects the impact of road slope and temperature changes on energy consumption, but also improves the overall performance of the system through multi-stage optimization processing. Urban planners can use this data to more accurately evaluate and optimize the layout of charging facilities, improve the management efficiency of new energy vehicles, and promote the development of green transportation.

[0103] Figure 2 A schematic diagram of the structure of a device (or system) for drawing an electric vehicle regional energy consumption map is provided in an embodiment of the present application, such as Figure 2 As shown, the device comprises: The acquisition module 21 is used to acquire the multi-source data stream generated during the driving of the electric vehicle, and access the external meteorological interface to acquire the temperature gradient parameters of the environment, wherein the multi-source data stream includes the longitude, latitude, cumulative mileage, state of charge and total energy fields of the on-board energy storage device in the format of the national recommended standard GB 32960; A generation module 22 is used to construct a dynamic grid generation model based on the administrative division coding system and the road topology network, and use an adaptive density adjustment algorithm to map the trajectory points to spatial grid cells with hierarchical associations, wherein the spatial grid cells are bound to a road slope distribution matrix, a temperature gradient parameter, and a traffic flow weight coefficient; An execution module 23 is used to perform multimodal segmentation on the driving segments that cross the grid boundary in the spatial grid unit by using trajectory continuity analysis, trigger the segment reconstruction of energy conservation constraints when the trajectory points enter different spatial grid units, and generate a feature vector including a grid code, a slope change rate, a temperature gradient vector correlation factor, and a traffic flow weight coefficient for each sub-segment; A calculation module 24 is used to establish a mapping relationship between the characteristic vector and the grid energy consumption contribution, calculate the original energy consumption value based on the difference between the total energy of the energy storage device and the accumulated mileage, perform nonlinear compensation correction in combination with the slope change rate, and superimpose the temperature gradient vector association factor and the traffic flow weight coefficient to generate a dynamic energy consumption benchmark value of multiple vehicle types in the spatial grid unit; The analysis module 25 is used to construct a spatiotemporal coupled diffusion model, perform correlation analysis on the dynamic energy consumption benchmark value, the charging pile distribution density and the traffic flow weight coefficient, and generate a three-dimensional dynamic map including an energy consumption thermal layer, a charging facility density layer and a traffic flow correlation layer through a multi-resolution rendering engine.

[0104] Figure 2 The device for drawing an electric vehicle regional energy consumption map can execute Figure 1 The implementation principle and technical effect of the method for drawing an electric vehicle regional energy consumption map described in the embodiment are not described in detail. The specific way in which each module and unit performs operations in the device for drawing an electric vehicle regional energy consumption map in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.

[0105] In one possible design, Figure 2 The device for drawing an electric vehicle regional energy consumption map in the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0106] The processing component 32 is used for the above Figure 1 The embodiment provides a method for drawing an electric vehicle regional energy consumption map.

[0107] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0108] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0109] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0110] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0111] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0112] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0113] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 A method for drawing a regional energy consumption map of electric vehicles in the illustrated embodiment.

[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0115] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0116] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for drawing an electric vehicle regional energy consumption map, characterized in that: include: Acquire a multi-source data stream generated during the driving of the electric vehicle, and access an external meteorological interface to acquire the temperature gradient parameters of the environment, wherein the multi-source data stream includes longitude, latitude, cumulative mileage, state of charge, and total energy fields of the on-board energy storage device in the format of the national recommended standard GB 32960; A dynamic grid generation model is constructed based on the administrative division coding system and the road topology network, and an adaptive density adjustment algorithm is used to map trajectory points to hierarchically associated spatial grid cells, which are bound to the road slope distribution matrix, temperature gradient parameters, and traffic flow weight coefficients; Trajectory continuity analysis is used to perform multimodal segmentation on the driving segments that cross the grid boundaries in the spatial grid cells, and when the trajectory points enter different spatial grid cells, segment reconstruction with energy conservation constraints is triggered, and a feature vector including a grid code, a slope change rate, a temperature gradient vector correlation factor, and a traffic flow weight coefficient is generated for each sub-segment; Establishing a mapping relationship between the characteristic vector and the grid energy consumption contribution, calculating the original energy consumption value based on the difference between the total energy of the energy storage device and the accumulated mileage, performing nonlinear compensation correction in combination with the slope change rate, and superimposing the temperature gradient vector association factor and the traffic flow weight coefficient to generate the dynamic energy consumption benchmark value of multiple vehicle types in the spatial grid unit; A spatiotemporal coupled diffusion model is constructed, and the dynamic energy consumption benchmark value is correlated with the charging pile distribution density and the traffic flow weight coefficient. A three-dimensional dynamic map including an energy consumption thermal layer, a charging facility density layer and a traffic flow correlation layer is generated through a multi-resolution rendering engine.

2. The method according to claim 1, characterized in that A mapping relationship between the characteristic vector and the grid energy consumption contribution is established, the original energy consumption value is calculated based on the difference between the total energy of the energy storage device and the accumulated mileage, a nonlinear compensation correction is performed in combination with the slope change rate, and the temperature gradient vector association factor and the traffic flow weight coefficient are superimposed to generate the dynamic energy consumption benchmark value of multiple vehicle types in the spatial grid unit, including: The original energy consumption value is calculated based on the difference between the total energy of the energy storage device and the accumulated mileage, the slope change rate data extracted from the road slope distribution matrix is ​​integrated, and the historical road topology data and the energy consumption deviation are back-propagated and trained using a multi-interval segmented optimization algorithm to generate a set of slope compensation coefficients; The ambient temperature, battery temperature rise gradient and thermal management power time series data in the temperature gradient parameters are integrated to construct a generation module of temperature influence factors, and a dynamic temperature correction coefficient is output using multi-dimensional feature fusion technology; Integrate the historical congestion pattern database of electric vehicles, and construct a traffic flow weight coefficient through an adaptive threshold iteration algorithm. The traffic flow weight coefficient includes a congestion intensity classification parameter and a speed fluctuation correction factor, and establishes a cross-dimensional correlation relationship with the road type code; Execute a multi-source parameter collaborative correction operation, perform a tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient, superimpose the traffic flow weight coefficient, output a dynamic energy consumption benchmark value, and use the vehicle dynamic characteristic parameters as a normalization adjustment factor to complete the dynamic calibration of the dynamic energy consumption benchmark value.

3. The method according to claim 2, characterized in that Executing a multi-source parameter collaborative correction operation, performing a tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient, superimposing the traffic flow weight coefficient, outputting a dynamic energy consumption reference value, and using the vehicle power characteristic parameter as a normalization adjustment factor to complete the dynamic calibration of the dynamic energy consumption reference value, including: Constructing a three-dimensional tensor operation channel to perform a tensor product operation on a slope compensation coefficient set and a dynamic temperature correction coefficient, wherein the dimension of the two-dimensional matrix of the slope compensation coefficient set corresponds to the number of road slope classifications and the number of historical road topology data codes, and the dimension of the dynamic temperature correction coefficient matrix corresponds to the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications; Decomposing the traffic flow weight coefficient into a congestion intensity component and a speed fluctuation component, wherein the congestion intensity component is associated with the administrative division restriction strategy in the road type code, and the speed fluctuation component is bound to the speed-energy consumption mapping relationship table of the road topology network; Generate a normalized adjustment factor using the vehicle dynamic characteristic parameters, wherein the vehicle dynamic characteristic parameters include a total energy field of an energy storage device and a mass-energy consumption conversion matrix, and the normalized adjustment factor acts on a process of superimposing a result of the tensor product operation and a traffic flow weight component; Execute a closed-loop calibration operation of the dynamic energy consumption reference value, and the calibration operation output result is synchronously input into the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit, wherein the charging pile distribution density calculation unit is associated with the spatial coding attribute of the road topology network, and the energy consumption thermal value diffusion unit is bound to the temperature change cycle data of the meteorological interface; The slope compensation coefficient set, the dynamic temperature correction coefficient, the traffic flow weight coefficient and the vehicle dynamic characteristic parameters form a cross-level data pipeline in the spatiotemporal coupled diffusion model, and all intermediate calculation results act on the downstream computing module through the convolutional neural network channel.

4. The method according to claim 3, characterized in that Construct a three-dimensional tensor operation channel to perform tensor product operation on the slope compensation coefficient set and the dynamic temperature correction coefficient, wherein the two-dimensional matrix dimension of the slope compensation coefficient set corresponds to the number of road slope classifications and the number of historical road topology data codes, and the dynamic temperature correction coefficient matrix dimension corresponds to the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications, including: A two-dimensional matrix of a slope compensation coefficient set is constructed by combining the number of road slope classifications and the number of historical road topology data codes, and a two-dimensional matrix of a dynamic temperature correction coefficient is constructed by combining the number of temperature compensation intervals and the number of battery chemical characteristic parameter classifications, so as to perform a tensor product operation to generate a three-dimensional compensation tensor structure, wherein the dimension of the three-dimensional compensation tensor structure includes a product combination of the number of road slope classifications, the number of temperature compensation intervals, and the number of battery chemical characteristic parameter classifications; Performing multi-level feature fusion processing on the three-dimensional compensation tensor structure, fusing the cross-validation results of the road slope classification number and the historical road topology data encoding, and injecting the battery chemical characteristic dynamic attenuation factor of the two-dimensional matrix of the dynamic temperature correction coefficient to generate a synergistic compensation tensor; Perform multi-source data superposition on the synergistic compensation tensor, the congestion intensity component and the speed fluctuation component of the traffic flow weight component to generate a dynamic multi-source fusion tensor; Performing a staged linear transformation on the dynamic multi-source fusion tensor using the mass-energy consumption conversion matrix in the vehicle dynamic characteristic parameters to generate a linear mapping image vector, wherein the staged linear transformation includes capacity attenuation compensation of the total energy field of the energy storage device and a gradient correction operation of the mass-energy consumption conversion matrix; Inputting the linear mapping image vector into the spatiotemporal coupled diffusion model for multi-scale feature calibration, and using the multi-scale feature calibration to associate the spatial coding attributes of the road topology network with the temperature change period data fluctuation curve of the meteorological interface, to generate a multi-dimensional output sequence carrying a dynamic energy consumption benchmark value; The multidimensional output sequence is transmitted to the charging pile distribution density calculation unit and the energy consumption thermal value diffusion unit through the convolutional neural network channel, wherein the charging pile distribution density calculation unit integrates the spatial coding attributes of the road topology network and the spatial distribution weights of the dynamic energy consumption benchmark value, and the energy consumption thermal value diffusion unit binds the temperature change cycle data and the attenuation correlation factor of the battery chemical characteristic parameter classification number.

5. The method according to claim 4, characterized in that The three-dimensional compensation tensor structure is subjected to multi-level feature fusion processing, the cross-validation result of the road slope classification number and the historical road topology data encoding is fused, and the dynamic attenuation factor of the battery chemical characteristics of the two-dimensional matrix of the dynamic temperature correction coefficient is injected to generate a synergistic compensation tensor, including: A cross-validation matrix is ​​constructed based on the road slope classification number and the data encoded by the historical road topology data, and a correlation feature extraction process is performed on the cross-validation matrix using a multi-channel convolutional layer to generate an initial fusion weight matrix that matches the subsequent tensor operation dimension; According to the battery chemical characteristic parameter classification dimension of the dynamic temperature correction coefficient matrix, the initial fusion weight matrix is ​​subjected to attenuation factor vector decomposition processing, and the dynamic attenuation coefficient calculation processing is performed on each temperature compensation interval in combination with the time series attenuation model to generate an attenuation factor tensor carrying a temperature gradient correlation factor; Performing an element-by-element product operation on the initial fusion weight matrix and the attenuation factor tensor by using a tensor broadcast mechanism to generate a coupling tensor including a four-dimensional structure; Based on the multi-dimensional attention mechanism, cross-dimensional feature alignment processing is performed on the coupling tensor, a query vector is generated using the feature of the road slope classification number, a key vector is generated using the feature of the historical road topology data encoding, and a value vector is generated using the temperature attenuation factor, and the query vector, key vector and value vector are normalized and weighted to generate an attention weight matrix; The three-dimensional compensation tensor and the attention weight matrix are subjected to residual connection processing, and a batch normalization layer is used to perform feature fusion processing on the connection result to generate a collaborative compensation tensor containing a four-dimensional product combination.

6. The method according to claim 4, characterized in that The dynamic multi-source fusion tensor is subjected to a staged linear transformation using the mass-energy consumption conversion matrix in the vehicle dynamic characteristic parameters to generate a linear mapping image vector, wherein the staged linear transformation includes a capacity attenuation compensation of the total energy field of the energy storage device and a gradient correction operation of the mass-energy consumption conversion matrix, including: A dynamic capacity correction tensor is constructed based on the capacity attenuation compensation operation of the total energy field of the energy storage device, and a multi-dimensional parameter fusion process is performed on the compensation operation using the time series attenuation characteristics of the historical state of charge data and the dynamic correlation factor of the temperature gradient parameter to generate a dynamic capacity correction matrix carrying capacity correction characteristics; According to the road slope classification number and the temperature compensation interval parameter of the dynamic multi-source fusion tensor, a gradient correction operation is performed on the mass-energy consumption conversion matrix, and a weight vector associated with the vehicle mass parameter and the slope classification number is extracted to generate a corrected mass-energy consumption conversion matrix; The dynamic capacity correction matrix and the corrected mass-energy consumption conversion matrix are subjected to feature coupling processing by using a tensor product operation to generate a mass-energy consumption correlation tensor including a capacity attenuation compensation feature and a road slope-mass coupling feature; Based on the staged linear transformation, multi-dimensional feature fusion processing is performed on the mass-energy consumption correlation tensor. In the first stage, the superposition characteristics of the road slope classification number and the traffic flow speed fluctuation component are used to generate a primary mapping vector. In the second stage, the capacity attenuation compensation parameter of the energy storage device is injected into the primary mapping vector using dynamic residual connection to generate an intermediate mapping tensor. In the third stage, a batch normalization layer is used to perform gradient correction processing on the intermediate mapping tensor to generate a linear mapping image vector.

7. The method according to claim 5, characterized in that Based on the multi-dimensional attention mechanism, cross-dimensional feature alignment processing is performed on the coupling tensor, a query vector is generated using the feature of the road slope classification number, a key vector is generated using the feature of the historical road topology data encoding, and a value vector is generated using the temperature attenuation factor. The query vector, the key vector and the value vector are normalized and weighted to generate an attention weight matrix, including: Using the road slope classification number, a multi-level feature decomposition process is performed on the height dimension of the four-dimensional coupling tensor, and the decomposed coarse-grained classification features and fine-grained fluctuation components are transmitted to the channel segmentation processing module to generate a query sub-vector group carrying multi-level slope features; Based on the historical road topology data encoding, spatial pooling processing is performed on the width dimension of the four-dimensional coupling tensor, and a dynamic weight adjustment is performed on the extracted local correlation features and global distribution features in combination with a temporal decay model to generate a key vector block; Performing a dynamic mask operation on the temperature gradient dimension of the four-dimensional coupling tensor according to the temperature attenuation factor, fusing the dynamic attenuation curve of the battery chemical characteristic parameter to generate a value vector sequence carrying a temperature compensation gradient feature; Performing group attention calculation processing on the query sub-vector group, the key vector block and the value vector sequence, and adjusting the correlation strength between different road slope classification numbers and temperature ranges through a dynamic weight allocation mechanism to generate a group attention weight matrix; The grouped attention weight matrix is ​​bidirectionally concatenated along the channel dimension, and the space-temperature feature aggregation process is performed on the concatenated tensor using the hierarchical fusion mechanism of the multi-scale convolution kernel to generate a space-temperature coupling tensor; Constructing a dynamic temperature decay mask based on the time series dimension of the decay factor tensor, and performing an element-wise product operation on the space-temperature coupling tensor to generate an enhanced attention weight matrix; The enhanced attention weight matrix is ​​residually connected with the four-dimensional coupling tensor, and the compensated tensor is subjected to cross-dimensional feature fusion through a batch normalization layer to generate an attention weight matrix that binds multi-source features.

8. A system for drawing regional energy consumption maps of electric vehicles, characterized in that: include: An acquisition module is used to acquire multi-source data streams generated during the driving of the electric vehicle, and access an external meteorological interface to acquire temperature gradient parameters of the environment, wherein the multi-source data streams include longitude, latitude, cumulative mileage, state of charge, and total energy fields of the on-board energy storage device in the format of the national recommended standard GB32960; A generation module is used to construct a dynamic grid generation model based on the administrative division coding system and the road topology network, and use an adaptive density adjustment algorithm to map trajectory points to spatial grid cells with hierarchical associations, wherein the spatial grid cells are bound to a road slope distribution matrix, a temperature gradient parameter, and a traffic flow weight coefficient; an execution module, configured to perform multimodal segmentation on the driving segments crossing the grid boundary in the spatial grid unit by using trajectory continuity analysis, triggering segment reconstruction with energy conservation constraints when the trajectory points enter different spatial grid units, and generating a feature vector including a grid code, a slope change rate, a temperature gradient vector correlation factor, and a traffic flow weight coefficient for each sub-segment; A calculation module, used to establish a mapping relationship between the characteristic vector and the grid energy consumption contribution, calculate the original energy consumption value based on the difference between the total energy of the energy storage device and the accumulated mileage, perform nonlinear compensation correction in combination with the slope change rate, and superimpose the temperature gradient vector association factor and the traffic flow weight coefficient to generate the dynamic energy consumption benchmark value of multiple vehicle types in the spatial grid unit; The analysis module is used to construct a spatiotemporal coupled diffusion model, perform correlation analysis on the dynamic energy consumption benchmark value, the charging pile distribution density and the traffic flow weight coefficient, and generate a three-dimensional dynamic map including an energy consumption thermal layer, a charging facility density layer and a traffic flow correlation layer through a multi-resolution rendering engine.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for drawing an electric vehicle regional energy consumption map as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a method for drawing an electric vehicle regional energy consumption map as described in any one of claims 1 to 7 is implemented.

Images