CN-121981358-A - Dual-wheel vehicle battery analysis-based electricity-exchanging path management method

CN121981358ACN 121981358 ACN121981358 ACN 121981358ACN-121981358-A

Abstract

The invention relates to the technical field of energy management of electric vehicles and discloses a battery replacement path management method based on two-wheel vehicle battery analysis. The method comprises the steps of establishing mapping of battery operation data and space-time information, constructing a track tree describing state evolution, and extracting candidate track clusters meeting space aggregation and time continuous conditions simultaneously. And then, carrying out space-time slicing and feature fusion on each track cluster, aggregating into a global state feature pool, and inputting the global state feature pool into a path decision engine. The engine generates an alternative path set through matching degree comparison and smoothness constraint, and an effective path subset is obtained through pruning and conflict resolution. And finally, fusing real-time external constraints to form a composite constraint field, re-evaluating and adjusting the paths, and outputting a power conversion strategy. The method improves the capability of identifying the regular power change demand mode from the discrete data, and can carry out path planning according to the fine dynamic characteristics of the network state, thereby enhancing the prospective and adaptability of scheduling.

Inventors

Mao Kejin
PENG SHUIPING
WU LIANGFENG
Xia Zanhui

Assignees

湖南凌寒新能源科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (10)

1.A battery analysis-based battery replacement management method for a two-wheeled vehicle, the method comprising: Establishing a mapping relation between operation data and space-time information of a two-wheel vehicle battery pack to form a coding vector set; Constructing a track tree describing a battery state evolution process based on the coding vector set, and extracting candidate track clusters meeting specific space aggregation conditions and time continuity conditions by traversing the track tree; performing space-time slicing operation on the candidate track clusters, extracting continuous slice sets of each candidate track cluster in the time dimension, and performing feature fusion on each continuous slice set to generate corresponding slice feature expression; The slice feature expressions corresponding to all candidate track clusters are aggregated to form a global state feature pool, and the global state feature pool is input to a pre-trained path decision engine; In the path decision engine, generating a preliminary alternative path set by iteratively comparing the matching degree of different path assumptions and the global state feature pool and introducing path smoothness constraint; redundant path pruning and conflict path resolution are carried out on the preliminary alternative path set, and effective path subsets which are not in conflict with each other and meet preset capacity constraint are screened out; Integrating the effective path subset with real-time external constraint information, and forming a composite constraint field through dynamic weighting fusion; and carrying out cost reevaluation and sequence adjustment on each path in the effective path subset according to the composite constraint field, and outputting a power conversion strategy comprising a path network and a path track set.
2. The method for managing a battery replacement path based on analysis of a battery of a two-wheeled vehicle according to claim 1, wherein the establishing a mapping relationship between the operation data of the battery pack of the two-wheeled vehicle and the space-time information, forming the encoded vector set, comprises: establishing a layered mapping relation comprising battery identity, acquisition time and at least one operation parameter; Performing multi-dimensional grid coding of the battery state according to the hierarchical mapping relation, and generating a coding vector set taking the battery identity as an index and space-time coordinate grids as carriers; Each encoding vector in the set of encoding vectors includes an operating parameter feature code associated with the space-time coordinate grid; the establishing a mapping relation between the operation data and the space-time information of the two-wheel vehicle battery pack comprises the following steps: receiving an original operational data stream from a battery sensor, the original operational data stream containing at least a battery identification code, a time stamp, a voltage, a current, and a temperature; Cleaning the original operation data stream, wherein the cleaning comprises removing abnormal values exceeding a physical range, filling data loss caused by signal loss, and aligning and synchronizing time stamps; And carrying out association matching on the cleaned operation data stream and the space coordinates provided by the geographic position information system, and binding unique space-time coordinates for each battery identity at each acquisition time so as to form the hierarchical mapping relation.
3. The battery analysis-based battery replacement path management method according to claim 2, wherein the performing multi-dimensional grid coding of battery states according to hierarchical mapping relation comprises: defining a space-time coordinate grid system covering a target area, wherein the space-time coordinate grid system is divided into geographic grids in a space dimension and is divided into time windows with fixed duration in a time dimension; Assigning each data point in the hierarchical mapping relationship to a space-time coordinate grid to which the data point belongs; And carrying out statistical feature calculation on the operation parameters of all data points falling into the same space-time coordinate grid, wherein the statistical features at least comprise mean values, variances and change trends, and coding calculation results into a vector with a fixed dimension, wherein the vector is a coding vector corresponding to the space-time coordinate grid, and all the coding vectors form the coding vector set.
4. The two-wheeled vehicle battery analysis-based power-change path management method according to claim 3, wherein the constructing a trajectory tree describing a battery state evolution process based on the set of encoding vectors includes: The root node of the track tree corresponds to an initial space-time grid, and branch nodes are generated according to state transition probabilities; Taking a specific space grid at the initial moment as a root node of the track tree; Starting from the root node, calculating the probability that the node is likely to be transferred to each adjacent space grid in the next time window according to the state transfer model, and adding the grid with the probability exceeding a threshold value into a track tree as a child node; Recursively repeating the transfer process for each newly added node until a preset time depth is reached or the transfer probability is lower than a survival threshold value, thereby completing the construction of the track tree; And backtracking from leaf nodes to root nodes of the constructed track tree to form a plurality of complete paths, and clustering path segments which are adjacent in spatial position and continuous in time to form the candidate track cluster.
5. The battery analysis-based power conversion management method according to claim 1, wherein the performing a space-time slicing operation on the candidate track clusters, extracting a set of consecutive slices of each candidate track cluster in a time dimension, comprises: For each candidate track cluster, segmenting along the time axis thereof with a fixed slice length and a sliding step length; Grouping all the coding vectors in each time slice according to the space positions of the coding vectors, and extracting fusion characteristics in each space position group; And splicing the fusion characteristics of all the spatial position groups in the same time slice according to a preset spatial topological sequence to generate slice characteristic expressions corresponding to the time slice, wherein the characteristic expressions of all the continuous time slices form a continuous slice set of the candidate track cluster.
6. The method for managing a power-exchanging path based on battery analysis of two-wheeled vehicles according to claim 5, wherein the aggregating slice feature expressions corresponding to all candidate track clusters to form a global state feature pool comprises: Collecting all slice characteristic expressions in a continuous slice set generated by all candidate track clusters into a characteristic storage structure; carrying out standardized processing on all slice feature expressions in the feature storage structure, and eliminating dimension differences of different track clusters and different time slices; And attaching a track cluster identifier and a time slice identifier of the source of each slice feature expression after the standardized processing to form the global state feature pool with traceability.
7. The two-wheeled vehicle battery analysis-based power conversion path management method according to claim 1, wherein generating, inside the path decision engine, a preliminary set of alternative paths by iteratively comparing matching degrees of different path hypotheses with the global state feature pool and introducing a path smoothness constraint, comprises: The path decision engine generates a plurality of initial path hypotheses according to the starting point and the ending point of the power change request; For each initial path hypothesis, calculating the similarity between the coding vector corresponding to the passed space-time grid sequence and the related feature in the global state feature pool, and accumulating to obtain the total matching degree score of the initial path hypothesis; In the matching degree calculation process, penalty is applied to state transition amplitude between adjacent grids in a path hypothesis, and the state transition amplitude is used as a path smoothness constraint to adjust the total matching degree score of the path hypothesis; and screening path assumptions with the total matching degree score higher than a preset threshold value to form the preliminary alternative path set.
8. The two-wheeled vehicle battery analysis-based power conversion path management method according to claim 1, wherein the performing redundant path pruning and collision path resolution on the preliminary alternative path set includes: Calculating the spatial overlapping degree of any two paths in the preliminary alternative path set, if the overlapping degree exceeds an overlapping threshold value, judging the path to be a redundant path, and only reserving one path with higher matching degree score; And checking the number of paths passing through the same space-time grid, if the number exceeds the bearing capacity of the same space-time grid, judging the paths as conflict paths, and sequentially removing the conflict paths according to the priority and the matching degree score of the paths until the capacity constraint is met, so as to obtain the effective path subset.
9. The method for managing a battery-operated path change based on a two-wheeled vehicle according to claim 1, wherein the integrating the effective path subset with real-time external constraint information forms a composite constraint field by dynamic weighted fusion, comprising: The external constraint information at least comprises traffic flow data and power exchange station load data; acquiring traffic flow data updated in real time, and converting the traffic flow data into a delay constraint field for the path transit time; acquiring load data of a power exchange station updated in real time, and converting the load data into a waiting constraint field for power exchange service time; and respectively distributing dynamic weights for the delay constraint field and the waiting constraint field, wherein the dynamic weights are dynamically adjusted according to the time attribute and the path criticality, and the weighted two constraint fields are overlapped to generate the composite constraint field.
10. The method for managing a power conversion path based on a battery analysis of a two-wheeled vehicle according to claim 1, wherein the performing a cost reevaluation and a sequence adjustment on each path in the subset of effective paths according to the composite constraint field comprises: correcting the original cost of each path in the subset of effective paths using the composite constraint field, the correction being manifested by adding additional cost as the path passes through the high constraint grid; re-ordering paths in the effective path subset according to the corrected path cost; and checking whether infeasible path segments caused by high constraint exist in the reordered paths, and carrying out local rescheduling on the path segments so as to bypass a high constraint area, and finally outputting an adjusted path network and path track set as the power conversion strategy.

Description

Dual-wheel vehicle battery analysis-based electricity-exchanging path management method Technical Field The invention relates to the technical field of energy management of electric vehicles, in particular to a battery replacement path management method based on two-wheel vehicle battery analysis. Background Currently, path planning in two-wheel vehicle battery-changing operation mainly depends on two technologies. Space statistics based on historical order data, such as identifying high density areas of battery change demand by thermodynamic diagrams, and static site planning or fixed path generation is performed accordingly. And the other is to perform responsive scheduling based on the real-time low-power state of the vehicle, and dispatch the battery-powered vehicle according to the single-point battery alarm information. These methods generally treat the need for a change of power as points that are spatially discrete, or as temporally isolated transient events. These solutions have drawbacks. The static method based on spatial clustering cannot distinguish the requirements of temporary aggregation, regularity and continuous generation, and omits the process of continuous attenuation of the battery state along with time. The real-time response mode is completely passive, and lacks predictability of the evolution trend of the overall demand of the area. Both have no effective modeling of the space-time correlation contained in the battery operation data, so that the capturing of the demand dynamics stays on the surface, and the planned path is often not matched with the space-time evolution track of the actual demand, so that the scheduling efficiency is low. There is a need for a method that automatically identifies, from discrete battery data, those battery state change laws that continue to appear spatially and that continue to develop temporally. The space-time law needs to be converted into a characteristic representation capable of finely describing the dynamic state of the network in different time periods and different areas so as to support the path planning system to make more prospective and adaptive decisions. Disclosure of Invention The invention aims to provide a battery replacement path management method based on two-wheel vehicle battery analysis, which aims to solve the problems in the background technology. In order to achieve the above object, the present invention provides a battery replacement management method based on two-wheeled vehicle battery analysis, the method comprising: Establishing a mapping relation between operation data and space-time information of a two-wheel vehicle battery pack to form a coding vector set; Constructing a track tree describing a battery state evolution process based on the coding vector set, and extracting candidate track clusters meeting specific space aggregation conditions and time continuity conditions by traversing the track tree; performing space-time slicing operation on the candidate track clusters, extracting continuous slice sets of each candidate track cluster in the time dimension, and performing feature fusion on each continuous slice set to generate corresponding slice feature expression; The slice feature expressions corresponding to all candidate track clusters are aggregated to form a global state feature pool, and the global state feature pool is input to a pre-trained path decision engine; In the path decision engine, generating a preliminary alternative path set by iteratively comparing the matching degree of different path assumptions and the global state feature pool and introducing path smoothness constraint; redundant path pruning and conflict path resolution are carried out on the preliminary alternative path set, and effective path subsets which are not in conflict with each other and meet preset capacity constraint are screened out; Integrating the effective path subset with real-time external constraint information, and forming a composite constraint field through dynamic weighting fusion; and carrying out cost reevaluation and sequence adjustment on each path in the effective path subset according to the composite constraint field, and outputting a power conversion strategy comprising a path network and a path track set. Preferably, the establishing a mapping relationship between the operation data and the space-time information of the two-wheel vehicle battery pack, and forming the encoded vector set includes: establishing a layered mapping relation comprising battery identity, acquisition time and at least one operation parameter; Performing multi-dimensional grid coding of the battery state according to the hierarchical mapping relation, and generating a coding vector set taking the battery identity as an index and space-time coordinate grids as carriers; Each encoding vector in the set of encoding vectors includes an operating parameter feature code associated with the space-time coordinate grid; the establishing a m