Battery thermal risk dynamic assessment method under cloud-edge-end three-level architecture

Abstract

The invention relates to the technical field of thermal assessment, in particular to a battery thermal risk dynamic assessment method under a cloud-edge-end three-level architecture, which comprises the following steps: acquiring end-side thermal response data, extracting a burst fluctuation segment, associating time and path indexes to construct a paragraph set, extracting a repeated response segment to form an interference sequence, counting node changes, screening out disturbance points, and determining the thermal risk of a battery under the cloud-edge-end three-level architecture. And comparing the cloud record to identify the interruption position and generate an evaluation list. According to the method, through sudden thermal response section extraction and path index aggregation, a multilayer mapping relation between disturbance points and paths is constructed, the recognition capability of key nodes in a multi-path cross scene is enhanced, interference sequences are extracted through periodic response analysis, disturbance factor changes are tracked, and the traceability of abnormal paths is enhanced; a high-frequency disturbance track is screened out through node state statistics, an interference structure group with associated characteristics is formed, a transmission interruption chain segment is captured through cloud data comparison, and the path reconstruction capability and the risk chain identification depth under multi-point disturbance are enhanced.

Description

Technical Field

[0001] This invention relates to the field of thermal assessment technology, and in particular to a dynamic assessment method for battery thermal risk under a cloud-edge-device three-level architecture. Background Technology

[0002] The field of thermal assessment technology encompasses the quantitative analysis and modeling of the heat transfer behavior, thermal stability, heat diffusion paths, and thermal response characteristics of materials or devices under different thermal operating environments. The core of this technology involves physical modeling and thermal parameter measurement of the conduction, convection, and radiation processes of heat within physical structures. It is commonly used to determine the thermal risks of lithium-ion batteries, power modules, and high-heat-load components under specific operating conditions. Research in this field typically covers temperature acquisition and calibration, heat flux modeling, thermal runaway threshold setting, thermal propagation path reconstruction, and thermal response time analysis. Combined with thermophysical parameter calculation, structural layout optimization, and safety strategy formulation, it constitutes a complete thermal risk diagnosis system.

[0003] The dynamic assessment method for battery thermal risk under a three-tiered cloud-edge-device architecture refers to the dynamic quantification and path determination of thermal runaway risks induced by factors such as local overheating and abnormal temperature distribution in battery operation scenarios such as charging / discharging and static storage, based on the collaboration of cloud computing platforms, edge processing nodes, and terminal devices, and employing distributed thermal data acquisition, time series analysis, and state prediction mechanisms. This method encompasses periodic extraction and real-time transmission of thermal response data, threshold classification of temperature rise rate, multi-source cross-comparison of historical operating data, estimation of thermal diffusion range based on heat conduction laws, and time lag extrapolation and identification of thermal stress concentration areas under thermal critical states. The collaborative deployment of thermal risk diagnosis logic is achieved through a cloud-edge-device data linkage mechanism, combined with iterative feedback of thermal modeling parameters, to realize a continuous risk identification and extrapolation mechanism.

[0004] Existing technologies rely on single-path heat propagation modeling, which makes it difficult to accurately reconstruct the complete heat spread chain when thermal anomalies present multi-point disturbances or path intersections. This leads to ambiguity in node identification during path recognition. Periodic thermal response data processing lacks a mechanism to distinguish the recurring characteristics of multi-segment disturbances, making it difficult to track the evolution trajectory of disturbance sources. The identification of abnormal transmission states at the signal acquisition level heavily relies on continuous interruption performance, lacking structured path interference verification and linkage inference of the states of preceding and following nodes. This can easily lead to the hidden existence of interrupted chain segments, affecting the ability to identify potential thermal runaway areas in advance. In environments with high-density node deployment, the efficiency of risk tracing and the integrity of response are significantly limited. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a dynamic assessment method for battery thermal risk under a cloud-edge-device three-level architecture.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a dynamic assessment method for battery thermal risk under a cloud-edge-device three-level architecture, comprising the following steps:

[0007] S1: Acquire thermal response change data from the end-side temperature sensing signal, read the signal curve during the heating process, extract continuous fluctuation response segments, corresponding occurrence time and path identifiers, extract the index information of thermal change points, and associate the segment positions to obtain the end-side identifier sudden segment set.

[0008] S2: Based on the path identification information of the burst segment of the end-side identifier, retrieve the data segment of the path in the differentiated time period, identify the response segment that recurs within the period, extract the heat response state and time sequence of each segment, the associated change content and corresponding state information, and obtain the path interference sequence structure.

[0009] S3: Based on the path structure content in the path interference sequence structure, read the number of node state changes within the path segment, filter out the node positions of repeated records, extract the associated points, and obtain the continuously disturbed path node group.

[0010] S4: Based on the path segment data in the continuous disturbance path node group, read the signal status of the corresponding node in the cloud record, identify the location of transmission interruption and response loss, compare the connection status between the preceding and following nodes, and obtain the interrupted chain segment evaluation list.

[0011] As a further embodiment of the present invention, the end-side identifier burst segment set includes a path identifier mapping table, a thermal change point index sequence, and a time-related segment location table; the path interference sequence structure includes a path data link configuration table, interference timing status records, and a continuous change correlation matrix; the continuous disturbance path node group includes a repeating state node list, a path number matching index, and a disturbance frequency statistics table; and the interrupted link segment evaluation list includes a transmission interruption record sequence, a response missing state marker set, and a connection continuity comparison map.

[0012] As a further aspect of the present invention, the associated segment position refers to reading the time point of each thermal response in the temperature sensing data collected at the end side, corresponding to the time interval of the response and the path to which it belongs, and locating the position distribution of the thermal response segments on the path.

[0013] The index information of the thermal change points refers to identifying the time period of continuous fluctuation during the analysis of thermal response changes, extracting the corresponding start and end time positions, and associating the paths corresponding to the time points.

[0014] The thermal response state refers to the temperature changes of the path and nodes in the time series, comparing the temperature trends of the previous and subsequent periods, and analyzing the behavioral characteristics of the thermal response on the time axis.

[0015] As a further aspect of the present invention, the associated changes refer to analyzing the fluctuation state of the thermal response data over time under each path, extracting the rising and falling trends of the response, comparing the magnitude of changes in each segment before and after time, and identifying the continuity and rhythm of change in the response process.

[0016] The thermal response state refers to the temperature changes of the path and nodes in the time series, comparing the temperature trends of the previous and subsequent periods, and analyzing the behavioral characteristics of the thermal response on the time axis.

[0017] As a further aspect of the present invention, the specific steps of S1 are as follows:

[0018] S101: Acquire thermal response change data from the end-side temperature sensing signal, read the path signal value change curve according to the data acquisition time sequence, detect the signal fluctuation segment during the heating process, extract the continuously changing time interval, match the path number corresponding to each segment, and obtain the continuous response segment association information set.

[0019] S102: Based on the continuous response segment association information set, extract the start and end time index values ​​of each thermal response segment, compare the distribution of path numbers within the time period, identify the alternating areas of thermal response changes between paths, associate the corresponding time points with path identifiers, and obtain thermal change index path comparison data.

[0020] S103: Based on the path identifier and time period index in the thermal change index path comparison data, the thermal change points and paragraph positions are matched according to the path order, and the matching records between the path number and the response segment are continued to obtain the end-side identifier sudden paragraph set.

[0021] As a further aspect of the present invention, the specific steps of S2 are as follows:

[0022] S201: Based on the path number information in the burst segment set of the terminal identifier, retrieve the differentiated time period data segment corresponding to each number, match each time interval with the path number, filter the path data records that appear repeatedly in the time period, and obtain the set of repeated path response segment intervals.

[0023] S202: Based on the thermal response value and time series of each data segment in the repeated path response segment interval set, calculate the slope of response change between adjacent data points, and obtain the path thermal response slope distribution result by corresponding to the average slope of change of each segment and the path identifier.

[0024] S203: Based on the correspondence between the change amplitude and continuous time period in the path thermal response slope distribution results, identify the path number that continuously fluctuates in thermal response during the period, and associate it with the information of continuously changing segments to obtain the path interference sequence structure.

[0025] As a further aspect of the present invention, the specific steps of S3 are as follows:

[0026] S301: Based on the path structure content in the path interference sequence structure, monitor the thermal response state of nodes in each path segment, identify nodes that respond within a continuous time period, match the thermal response state with the position of occurrence on the timeline, and obtain a set of node state changes.

[0027] S302: Based on the time index and state change records in the node state change set, calculate the slope of the thermal response change of each node in a continuous time period, associate the slope value with the node number, and obtain the response change trend slope set.

[0028] S303: Based on the node information in the slope set of the response change trend, identify the node segments with continuously changing slope values ​​in the path sequence, associate the path range and node position corresponding to the node segments, and obtain the continuous perturbation path node group.

[0029] As a further aspect of the present invention, the specific steps of S4 are as follows:

[0030] S401: Based on the path segment data in the continuous disturbance path node group, obtain the signal transmission sequence of the corresponding node in the cloud, examine the signal time sequence of the nodes in the path segment, locate the node position where the signal continuity is interrupted, and obtain the signal timing anomaly set.

[0031] S402: Based on the signal timing anomaly set, identify the time extension distribution between nodes, track the communication time series changes of nodes in the path connection structure, locate the path segments with missing signals between consecutive nodes, and obtain the node communication discontinuity distribution.

[0032] S403: Based on the intermittent distribution of node communication, check the continuity status of channel segments in the path structure, identify the node combinations that respond to breakpoints in the connection, analyze the interrupted segments according to the connection status of the path corresponding to the node, and obtain an evaluation list of interrupted chain segments.

[0033] As a further aspect of the present invention, the method further includes:

[0034] S5: Based on the path identifiers in the interrupted chain segment assessment list, analyze the state changes of the path thermal response trajectory under the period, extract the node sequence with continuous fluctuations, analyze the corresponding response amplitude and time interval between nodes, identify the path segments with risk change trends, and obtain dynamic thermal risk assessment results.

[0035] The dynamic thermal risk assessment results include a multi-period fluctuation node sequence, a thermal response amplitude distribution matrix, and a time-series risk feature vector.

[0036] As a further aspect of the present invention, the specific steps of S5 are as follows:

[0037] S501: Based on the path identifier in the interrupted chain segment evaluation list, scan the thermal response change trajectory of the node segment by segment according to the timeline, track the response state changes between adjacent nodes, locate the corresponding segment of the interrupted signal in the time series, and obtain the response discontinuity sequence set.

[0038] S502: Based on the node number sequence of each segment in the response discontinuity sequence set, synchronously identify the transmission path position corresponding to the node, analyze the fluctuation trend of the node thermal response amplitude in adjacent time windows on the path segment, and obtain the path discontinuity trend sequence.

[0039] S503: Based on the correspondence between the fluctuation amplitude of each segment and the path number in the path discontinuity trend sequence, extract the path segments in the periodic sequence where the thermal response fluctuation changes, rematch the path number with the continuous fluctuation characteristics, and obtain the dynamic thermal risk assessment result.

[0040] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0041] In this invention, by extracting sudden thermal response segments and aggregating path indexes, a multi-layered mapping relationship between disturbance points and paths is constructed, enhancing the ability to identify key nodes in multi-path intersection scenarios. By extracting interference sequences through periodic response analysis and tracking changes in disturbance factors, the ability to trace the source of abnormal paths is strengthened. By statistically analyzing node states, high-frequency disturbance trajectories are screened out, forming a group of interference structures with associated characteristics. By comparing cloud data, transmission interruption segments are captured, enhancing the path reconstruction capability and the depth of risk chain identification under multi-point disturbances. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the main steps of the present invention;

[0043] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0044] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0045] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0046] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0047] Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0049] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0050] Please see Figure 1 This invention provides a technical solution: a dynamic assessment method for battery thermal risk under a cloud-edge-device three-level architecture, comprising the following steps:

[0051] S1: Obtain the thermal response change data transmitted by the end-side temperature sensing signal, read the signal value change curve in chronological order, extract the response segments that fluctuate continuously during the heating process, identify each response segment according to the path identifier corresponding to the occurrence time, extract the index sequence of thermal change points, associate the path identifier with the corresponding segment position, and obtain the end-side identifier burst segment set.

[0052] S2: Based on the path number information of the burst segment of the end-side identifier, retrieve the data segment corresponding to the path within the differentiated period, identify the segments that appear continuously in the record, extract the time sequence and thermal response change of the segments, configure the continuous change state and the path correspondence into the path data chain in time sequence, and obtain the path interference sequence structure.

[0053] S3: Based on the path structure content in the path interference sequence structure, read the number of state changes of nodes in each path segment, filter the node positions where the number of changes is repeated, match the node number with the corresponding path number, and obtain the continuous perturbation path node group.

[0054] S4: Based on the path segment data in the continuous disturbance path node group, read the signal transmission records of the associated path nodes in the cloud, analyze whether there are transmission interruptions and response missing states in the record sequence, compare the connection continuity between adjacent nodes, and obtain the interrupted chain segment evaluation list.

[0055] S5: Based on the path identifiers in the interrupted chain segment assessment list, perform sequence analysis on the path thermal response change trajectory, extract the node sequence whose state fluctuation duration covers multiple cycles, and compare the fluctuation characteristics of the path number with the distribution of thermal response amplitude and time interval between nodes to obtain the dynamic thermal risk assessment results.

[0056] The end-side identifier burst segment set includes a path identifier mapping table, a thermal change point index sequence, and a time-related segment location table. The path interference sequence structure includes a path data link configuration table, interference time sequence status records, and a continuous change correlation matrix. The continuous disturbance path node group includes a list of repeating state nodes, a path number matching index, and a disturbance frequency statistics table. The interrupted link segment assessment list includes a transmission interruption record sequence, a response missing state marker set, and a connection continuity comparison map. The dynamic thermal risk assessment results include a multi-cycle fluctuation node sequence, a thermal response amplitude distribution matrix, and a time-series risk feature vector.

[0057] Please see Figure 2 The specific steps of S1 are as follows:

[0058] S101: Acquire thermal response change data from the end-side temperature sensing signal, read the path signal value change curve according to the data acquisition time sequence, detect the signal fluctuation segment during the heating process, extract the continuously changing time interval, match the path number corresponding to each segment, and obtain the continuous response segment association information set.

[0059] First, in the node data receiving link, the instantaneous signal value sets of the temperature sensing units under each channel are called in the transmission order. The amplitude difference of each signal point value is compared with the data of adjacent points to construct a sequence of change curves within a continuous time period. Then, the frequency fluctuation trend of signal value increase or decrease within a unit sampling interval is identified. If the signal difference of every three consecutive points maintains the same direction of change in the positive or negative direction, it is judged as a single trend segment, and the boundary points of the segment are extracted according to the time index. Subsequently, the extracted multiple trend segments are aggregated and mapped under different path numbers. Each trend segment is looked up through the path correspondence. Find the physical connection identifier and establish the assignment channel relationship between the path segment and the thermal response segment. For example, under a certain path number A, the continuous signal segment is (0.12, 0.19, 0.28, 0.33), and the corresponding sampling sequence number is (5, 6, 7, 8). Then, this signal segment can be classified as a continuous upward trend, and a mapping is established between path number A and the sequence number interval (5-8). Then, the start and end positions of each trend interval are numbered and summarized. The path identifier, time start and end points, and signal change direction are stored as combined units in the interactive data table for subsequent temperature offset identification. Finally, the continuous response segment association information set is obtained.

[0060] S102: Based on the continuous response segment association information set, extract the start and end time index values ​​of each thermal response segment, compare the distribution of path numbers within the time period, identify the alternating areas of thermal response changes between paths, associate the corresponding time points with path identifiers, and obtain thermal change index path comparison data.

[0061] First, identify the thermal response curve corresponding to each data segment. Read the sampling sequence numbers corresponding to the start and end points of the curve as the start and end index values ​​for that segment. Then, match the corresponding path numbers. Establish a corresponding chain of time index and path identifier for each path. Next, detect the interaction relationship between overlapping index intervals of adjacent path numbers. If two or more path segments within the same time period are found to have index overlap, it is considered an alternating region, and the boundary of the alternating region is marked as a thermal response intervention point. This delineates the response interference range between paths. For example, the start and end indexes of the response segment for path number B... If path B is (100, 130) and path C is (120, 140), then the index interval (120, 130) is judged as an alternating region. If the direction of change of the node signal intensity is consistent within this time period, then path B and path C are set as interference pairs and associated and recorded as thermal change correlation pairs. After the path interference pairs are identified, the time index points under each interference pair need to be cross-verified to find out if the same node appears repeatedly under different paths, and mark it as a potential interference point. Finally, the correspondence between the path number and the alternating time point is formed, and thermal change index path comparison data is obtained.

[0062] S103: Based on the path identifier and time period index in the thermal change index path comparison data, the thermal change points and paragraph positions are matched according to the path order, and the matching records between the path number and the response segment are continued to obtain the end-side identifier sudden paragraph set.

[0063] First, the path identifiers are categorized and sorted. A mapping relationship is established between the start and end time intervals of thermal changes corresponding to each path group according to their index order. Each node segment corresponding to each time interval is confirmed point by point. Nodes whose thermal response change rate exceeds a fixed amplitude ratio are identified as thermal change points. During execution, for path number A, the start and end indices of thermal change points are 20 to 55. This segment is mapped to path A. Simultaneously, the interval indices of path B (40 to 75) are retrieved, and it is confirmed that the segment with indices 40 to 55 in path B also contains change points with the same response change trend direction. Based on these overlapping path segments, a many-to-one correspondence chain between paths and time periods is constructed. The chain list retains the path number, thermal change start and end index, and corresponding... The process involves three aspects: paragraph location, path segment location, and corresponding node sequence location. The completeness of the matching relationship between the path segment and the corresponding node sequence is then determined. If the number of changed points within a path segment exceeds 75% of the total number of nodes in the segment, the segment is identified as a sudden response segment. In actual operation, for example, if path number C corresponds to hot change point segments 100 to 130, and this segment contains a total of 20 nodes, with 15 nodes exhibiting hot change point characteristics, then its proportion is 75%, meeting the criteria for a sudden response segment. Such segments need to be further linked to the path number and marked with a time index range. Data writing operations are performed on each matched segment to complete the sequential correspondence between the path number and the sudden response segment location, ultimately obtaining the end-side identified sudden segment set.

[0064] Please see Figure 3 The specific steps of S2 are as follows:

[0065] S201: Based on the path number information in the burst segment set of the end-side identifier, retrieve the differentiated time period data segment corresponding to each number, match each time interval with the path number, filter the path data records that appear repeatedly in the time period, and obtain the set of repeated path response segment intervals.

[0066] First, each path number is used as an independent identifier to retrieve the relevant response time interval within the current period. A mapping relationship is then established between the path number and the corresponding data segment time interval. For path number A, if the corresponding time interval is between 12 and 45, the index interval corresponding to path A in the thermal response data is queried to confirm whether the thermal response amplitude of the data segment within the interval meets the response segment fluctuation threshold range. It is also checked whether path number A and records with the same thermal response pattern appear consecutively multiple times in the sampled data. In multiple periods, when the time interval corresponding to path number A repeatedly appears in the sequence where the response fluctuation amplitude changes exceed the set boundary... In the column, the response data segment under path number A is included in the scope of repeated response segment determination. Then, the repeated matching process is carried out for all path numbers according to the period dimension. For example, if the response segment interval of path number B is 80 to 120 in period 1 and 85 to 125 in period 2, and the difference between the thermal response change amplitude and the response time interval is less than 5 units, it can be determined that path B has a repeated thermal response segment in the period. Then, according to the correspondence between path number and time period, all repeated thermal response data segments that meet the conditions are formed into a segment set and assigned to the corresponding path number, and finally the repeated path response segment interval set is obtained.

[0067] S202: Based on the thermal response value and time series of each data segment in the repeated path response segment interval set, calculate the slope of response change between adjacent data points, the average slope of change of each segment and the path identifier, and obtain the path thermal response slope distribution result.

[0068] The specific formula for calculating the slope of the response change between adjacent data points is as follows:

[0069]

[0070] in, This represents the slope of the response change between the j-th neighboring data points in the i-th segment of the r-th path. This represents the temperature increment at the j-th data point in the i-th segment of the r-th path. This represents the time increment for the corresponding data point. This represents the mean of the i-th segment of the time series along the r-th path. This represents the mean of the temperature sequence of the i-th segment of the r-th path. γ represents the standard deviation of the i-th time series segment in the r-th path. (r) η is a coefficient used in the r-th path to adjust the effect of time fluctuation intensity on the slope of the thermal response. (r) This is a factor used in the r-th path to adjust the degree of influence of temperature deviation on the slope denominator;

[0071] The data with path r=1, segment number i=2, and sampling point number j=3 is as follows:

[0072] Temperature point Previous temperature point but

[0073] Sampling time Previous time point but

[0074] The average time interval is Standard deviation

[0075] The average temperature is

[0076] Set γ (1) =0.75, η (1) =0.1;

[0077] Substitute into the formula:

[0078]

[0079]

[0080] Interpretation of results and numerical significance: The results show that the current response slope is approximately 0.493 K / s, which is close to the set warning threshold of 0.5 K / s, indicating that a strong thermal disturbance has already developed in the thermal path;

[0081] Explanation of the innovative aspects of the formula:

[0082] The advantage of the formula is that by adding standard deviation, temperature deviation term and adjustment factor to establish a composite response change structure, it has the ability to control the stability of time sampling and the sensitivity of temperature fluctuation under dynamic disturbance conditions, thereby improving the forward depth of risk identification in the characterization of thermal change trend during system operation.

[0083] S203: Based on the correspondence between the change amplitude and continuous time period in the path thermal response slope distribution results, identify the path number that continuously fluctuates in thermal response during the period, and associate it with the information of continuously changing segments to obtain the path interference sequence structure.

[0084] First, obtain the change value of the thermal response slope for each path number in different time periods. Then, arrange these changes in chronological order using a time index to define the response trend of the path number in each period. If the slope values ​​for consecutive time periods t1 to t5 in path number A are 2.1, 2.4, 2.3, 2.5, and 2.2 respectively, and the direction of change remains unchanged, then the path is determined to be in a state of continuous thermal response fluctuation within that period. In this case, path number A is associated with the time interval t1 to t5 as continuously changing segment data. Subsequently, the same operation is performed on all path numbers in sequence. A list of corresponding paths and continuous fluctuation intervals is formed. Then, the periodic sequence is compared according to each path number. For path number B, if its slope changes within ±0.5 in period 1 and period 2 and maintains this fluctuation trend in five consecutive time indices, it can be determined as a continuous fluctuation path of thermal response within the period. The path number and its corresponding change segment information in different periods are collected and summarized to form a continuous fluctuation identifier set of thermal response. Finally, based on the arrangement order of each path number in the continuous fluctuation identifier set of thermal response and the corresponding index of the change segment, the path interference sequence structure is obtained.

[0085] Please see Figure 4 The specific steps of S3 are as follows:

[0086] S301: Based on the path structure content in the path interference sequence structure, monitor the thermal response state of nodes in each path segment, identify the nodes that respond within a continuous time period, match the thermal response state with the position of occurrence on the time line, and obtain the node state change set.

[0087] First, the path number and its associated node set are clearly defined. The node number, arrangement order, and position index in each path are broken down segment by segment. The node number mapping relationship is then synchronized using previously acquired path response segments. Based on this, a data stream sequence within a continuous response period is selected, and the status identifier of each node number is read to determine if there are characteristic values ​​of thermal response changes within the continuous time period. During this process, a baseline range for judging node status changes needs to be set. This range can be taken as ±2 standard deviations centered on the average response amplitude. Node numbers exceeding this range in the corresponding temperature sensing data are considered to have responded. For example, in the sequence with path number P-17, if node N-11 is in the 22nd to 23rd position... If the sensing value of segment 9 increases by more than 2.8℃ compared to the previous cycle and is maintained for more than 3 consecutive segments, then the node is added to the response node list, and the time index corresponding to its location is written into the thermal response matching chain. Then, the same process is performed on other nodes. After the entire path data is processed, the state changes of each node have been marked. Then, the index numbers of these state change nodes are synchronized with the path numbers in which they appear, and their appearance order and response state type are arranged according to the path order. That is, those above the set response range are positive responses, those below the range are negative responses, and nodes without response are not processed. Finally, all nodes that have thermal response states and their time indices are compared in the path dimension to obtain the node state change set.

[0088] S302: Based on the time index and state change records in the node state change set, calculate the slope of the thermal response change of each node in a continuous time period, associate the slope value with the node number, and obtain the response change trend slope set.

[0089] The specific formula for calculating the slope of the thermal response change at each node over a continuous time period is as follows:

[0090]

[0091] in, This represents the slope of the thermal response change of the e-th node within the time interval v and immediately following time v′. This represents the average thermal response energy of the e-th node during the v-th time period. This represents the average thermal response energy of the e-th node within the immediate time period v′. This represents the average thermal response amplitude of the e-th node during the v-th time period. This represents the median value of the thermal response amplitude of the e-th node during the v-th time period. This represents the average prediction residual value of the e-th node during the v-th time period. denoted as the average amplitude of the state transition disturbance of the e-th node during the v-th time period, where e represents the node number, v represents the current time period number, and v′ represents the next consecutive time number adjacent to v in the time series.

[0092] In the example, suppose we consider the node with node number e = 3:

[0093] The thermal response energy sequence measured at node v=6 (10:00-10:10) is as follows:

[0094] 122.3, 124.1, 123.7, 125.0, 122.8 kJ, calculate the average value as follows:

[0095]

[0096] The thermal response energy collected in the next consecutive time period (10:10-10:20) is:

[0097] 127.6, 128.1, 127.4, 128.5, 127.9 kJ, therefore:

[0098]

[0099] The response amplitudes were recorded as 21.2, 21.7, 20.9, 21.0, and 21.5W during the 6th time period, with an average value of:

[0100]

[0101] The median value after sorting is 21.2W.

[0102]

[0103] The predicted residuals were obtained by comparing the actual measured values ​​(21.2, 21.7, 20.9, 21.0, 21.5) with the predicted values ​​(21.0, 21.4, 20.7, 20.9, 21.3), resulting in residuals of (0.2, 0.3, 0.2, 0.1, 0.2). The average of these residuals was then calculated.

[0104]

[0105] The rate of change of the abrupt change points in the perturbation amplitude monitoring was (0.4, 0.6, 0.5, 0.7, 0.3), with a mean of:

[0106]

[0107] Substituting the above values ​​into the formula, the numerator is:

[0108] |127.9-123.58|·(21.26-21.2)·0.2=4.32·0.06·0.2=0.05184;

[0109] The denominator is:

[0110]

[0111] The final result is:

[0112]

[0113] Interpretation of Results and Numerical Significance: Slope Value Below the dynamic risk benchmark threshold τ crit =0.005, based on the statistical setting of the 95% confidence interval of the node group fluctuation. This result indicates that node 3 is in a stable response trend state during the time period and does not trigger the intervention mechanism.

[0114] Explanation of the innovative aspects of the formula:

[0115] The advantage of the formula lies in its ability to introduce the predicted residuals. and disturbance response factor This allows the formula to reflect the changes in nodal thermal response in terms of actual offset, modeling error, and external disturbances, thus characterizing the response trend of each nodal.

[0116] S303: Based on the node information in the slope set of the response change trend, identify the node segments with continuously changing slope values ​​in the path sequence, associate the path range and node position corresponding to the node segments, and obtain the continuous perturbation path node group.

[0117] First, the data in the slope set is categorized by node number. The trend curve of each node's thermal response across multiple path segments is identified. The thermal response slope values ​​of each node are arranged in time index order, forming a slope variation trajectory for continuous time periods. Based on this, the direction of fluctuation in adjacent time periods of the slope curve is detected to see if there is a continuously rising or falling trend. For example, if the slope values ​​of node N-04 in path P-13 are 1.2, 1.6, 2.1, 2.3, 2.5, and 2.8 for six consecutive time periods, its direction of change can be determined as a continuously rising trend. Then, the value of this node in the entire path sequence is extracted. The start and end time indices of the node are matched with the path number and location number to form the node segment boundary. Then, the same steps are performed on all nodes with continuous trends. The node segments with continuous change trends are set according to the start and end time periods and the path number, and their distribution density is screened to determine whether there are multiple nodes with continuous trends clustered within a certain path number range. If this happens, the node segments are matched according to the path number. Finally, the node number, occurrence time period, and path number of the node with continuous slope change trend and path range concentration characteristics are jointly output to obtain the continuous disturbance path node group.

[0118] Please see Figure 5 The specific steps of S4 are as follows:

[0119] S401: Based on the path segment data in the continuous disturbance path node group, obtain the signal transmission sequence of the corresponding node in the cloud, examine the signal time sequence of the nodes in the path segment, locate the node position where the signal continuity is interrupted, and obtain the signal timing anomaly set.

[0120] First, the path numbers and their corresponding response time periods for each node are listed sequentially. For each path segment, the signal response records of each node during the thermal disturbance are extracted, and all records are concatenated into a continuous time series according to the acquisition time. Next, the signal time sequence relationship between two adjacent nodes on each path is identified, and the continuity of signal time between adjacent nodes is detected. If the signal response time of the later node is earlier than that of the earlier node, or if there is a non-continuous time jump between the two nodes, i.e., a discontinuity in the response time, the node is marked as a node with an abnormal signal sequence, such as nodes A and B in path P-7. The response times of node D, C, and D are 3, 7, 8, and 5 respectively. Since the response time of node D is less than the time of its predecessor node C (8), it is considered a timing anomaly. The above-mentioned abnormal nodes are then sorted by their positions within the path to observe whether the anomaly occurs at the beginning, middle, or end of the path, in order to determine the path segment range affected by the anomaly. At the same time, invalid interruptions caused by node duplication or data noise are filtered out. If two nodes have the same response time, no anomaly is determined. Finally, the node number, path identifier, and corresponding time index of each path segment where signal continuity interruption occurs are jointly collected to obtain the signal timing anomaly set.

[0121] S402: Based on the signal timing anomaly set, identify the time extension distribution between nodes, track the communication time series changes of nodes in the path connection structure, locate the path segments with missing signals between consecutive nodes, and obtain the discontinuous distribution of node communication.

[0122] First, obtain the node pairs with abnormal responses and their corresponding time index values ​​in each path. These node pairs are categorized and arranged by path number, and a time sequence chain is established according to the original structural order of the nodes in the path. Then, an extension analysis is performed on the time index interval between two adjacent nodes in the path. The response time difference of each pair of adjacent nodes is extracted and compared horizontally with the time differences between other nodes in the path. If the time difference between a pair of nodes is found to be more than twice the average difference in the path, it is determined that there is communication extension in that path segment, for example, path P-5. The response interval between nodes C and D is 0.9 seconds, while the intervals for the remaining segments are all less than 0.4 seconds. Therefore, this segment is identified as an extension segment. Further analysis of the path connection structure data is conducted to determine whether the segment is a physically continuous segment. If two nodes have a direct connection relationship in the structure graph and there is no alternative branch, they are considered as a continuous node connection segment. Under the dual conditions of structural continuity and abnormal delay, this segment is located as a communication interruption segment. Subsequently, the segment numbers that meet the conditions are summarized, and the node information, corresponding path, time difference, and connection relationship are combined and listed to finally obtain the distribution of node communication interruptions.

[0123] S403: Based on the intermittent distribution of node communication, check the continuity status of channel segments in the path structure, identify the node combinations that respond to the breakpoints in the connection, analyze the interrupted segments according to the connection status of the path corresponding to the node, and obtain an evaluation list of interrupted chain segments.

[0124] First, all paths in the path structure are sequentially numbered, and the original connection order of each node in the path is listed according to the path topology. Then, combined with the breakpoint nodes involved in the communication interruption, the connection situation before and after them is matched in the path structure. The relative position of each pair of nodes in the path segment in the number list is extracted. If it is found that a pair of breakpoint nodes are not directly adjacent in the number list and there are other nodes between them that have not experienced communication anomalies, it is determined that the breakpoint combination does not meet the structural continuity standard and is removed. Only the node pairs that are directly continuous in the connection structure and appear in the distribution of communication interruptions are retained. Then, the node numbers that appear repeatedly in the path are filtered out from these node pairs. It is checked whether the path number has a unique channel direction in the structure diagram. If there is a split in the channel branch structure, all nodes involving branch intersections under the path are marked and recorded separately as indeterminate segments. Finally, all nodes that meet the structural continuity, response interruption existence and path identification are combined according to the path number, start and end node numbers, connection segment length and other information. The channel segment continuity judgment result is used as a label item to obtain the interrupted chain segment evaluation list.

[0125] Please see Figure 6 The specific steps of S5 are as follows:

[0126] S501: Based on the path identifier in the interrupted chain segment evaluation list, scan the thermal response change trajectory of the node segment by segment according to the timeline, track the response state changes between adjacent nodes, locate the corresponding segment of the interrupted signal in the time series, and obtain the response discontinuity sequence set.

[0127] First, all path numbers in the list are read sequentially, and a corresponding heat response record segment is established for each path in the time series data. When processing a path segment, the heat response values ​​of all associated nodes in the path appearing in the time series are extracted into a linearly arranged state flow dataset. Then, along the time axis, the response states of adjacent nodes in the path are scanned sequentially in fixed-length time windows. Within each window, the node states at the current time point and the previous time point are called to compare whether there is a sudden change in their heat response state from high or low amplitude to no response. The time index corresponding to the no-response state and the numbers of the adjacent nodes before and after it are recorded. Then, the trend of the heat response state changes of two adjacent nodes is compared. The system determines whether the state transition lasts for more than two time segments. If the continuity condition is met, it is considered a potential location of the interruption signal. The combination of nodes with the abnormal state, the time segment location, and the path number are paired and summarized. For node segments with multiple response interruptions in the same path, it is necessary to check whether there is a hot response recovery segment. If the response does not recover in the subsequent time period, it is continuously included in the interrupted state segment. If it recovers to a valid response within two time segments after the interruption, it is considered a brief fluctuation and is not treated as an interruption. Finally, all state interruption segments and their node location information scanned in each path structure are recorded according to the path number, response start and end time, number of nodes involved, etc., to obtain a response interruption sequence set.

[0128] S502: Based on the node number sequence of each segment in the response discontinuity sequence set, synchronously identify the transmission path position corresponding to the node, analyze the fluctuation trend of the node thermal response amplitude in adjacent time windows on the path segment, and obtain the path discontinuity trend sequence.

[0129] First, each segment of discontinuity information is traversed sequentially to extract a complete list of node numbers contained in that segment. Each node number is then associated with its transmission sequence information along its path using a path mapping relationship. During this process, it is necessary to verify whether the number appears in multiple path branches. If it does, the path segment closest to the discontinuity time is selected as the primary path match based on the chronological order, thus completing the unique association between the node and the path segment. Subsequently, using the node segments identified as discontinuous in each path as the scope, the thermal response values ​​of all nodes covered by that path segment are collected within multiple consecutive time windows. The width of each time window can be set to 10 time points, and multiple adjacent windows are generated using a fixed-step sliding method. Within each window, the thermal response amplitude of all nodes on that path segment is extracted, and the values ​​are then analyzed according to the physical connections of the nodes. The sequential arrangement forms a fluctuation vector sequence. The response amplitude changes of the same node are compared between two adjacent time windows, and their absolute differences are calculated. Further, it is determined whether the positive and negative directions of the changes are consistent. When the difference changes continuously increase or decrease in multiple consecutive time windows, it can be regarded as a continuous segment of fluctuation trend. The condition for this trend determination is that the difference of the response value of any node in three or more adjacent windows is greater than the preset floating benchmark of 0.2, and the fluctuations are in the same direction. For example, if the thermal response values ​​of node n12 on path P8 are 1.0, 1.3, and 1.5 in three adjacent time windows, its fluctuation trend is judged to be continuously increasing. If most nodes in this path segment have similar trend directions, then the entire path segment is judged to have a continuous thermal response change trend, and finally, a path discontinuous trend sequence is obtained.

[0130] S503: Based on the correspondence between the fluctuation amplitude of each segment and the path number in the path discontinuity trend sequence, extract the path segments in the periodic sequence where the thermal response fluctuation changes, rematch the path number with the continuous fluctuation characteristics, and obtain the dynamic thermal risk assessment results.

[0131] First, each record in the sequence is grouped according to its path number. The thermal response fluctuation amplitudes of multiple time periods corresponding to the same path number are then expanded sequentially, and the amplitude variation range of each segment is extracted. The period position number of each segment is recorded. Next, a correspondence table between path segments and period positions is constructed using the path number as the primary key index. The thermal response fluctuation amplitude of each record is evaluated. The evaluation method is that if the amplitude change exceeds 0.25 response units twice or more in a continuous time slice, the path segment is determined to be a high-fluctuation segment. For example, if the fluctuation amplitudes of path number P7 in period 3 and period 5 are 0.32 and 0.45 respectively, then this path is marked as a high-fluctuation path segment in these two period periods. Subsequently, the time slices of the path segments determined to be high-fluctuation are connected segment by segment, ensuring the continuity of the time sequence numbers. Under this premise, all path segments in the connection relationship are aggregated and identified. If a path number shows high fluctuation characteristics in three consecutive periods, the entire path segment is marked as a dynamic fluctuation path. Then, based on the above identification results, each dynamic fluctuation path is compared with its original thermal response fluctuation data again to extract the minimum amplitude fluctuation node and the maximum amplitude fluctuation node in the same period segment, calculate the change ratio in adjacent periods, and determine whether the fluctuation direction is consistent. If the fluctuation direction is consistent in three consecutive periods, the path segment is further confirmed to have continuous dynamic thermal response characteristics. Finally, combining the mapping relationship between the path number, time slice number, and fluctuation trend direction, a complete list including number index, thermal response dynamic flag, and trend label is output to obtain the dynamic thermal risk assessment result.

[0132] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A dynamic assessment method for battery thermal risk under a cloud-edge-device three-tier architecture, characterized in that, Includes the following steps: S1: Acquire thermal response change data from the end-side temperature sensing signal, read the signal curve during the heating process, extract continuous fluctuation response segments, corresponding occurrence time and path identifiers, extract the index information of thermal change points, and associate the segment positions to obtain the end-side identifier sudden segment set. S2: Based on the path identification information of the burst segment of the end-side identifier, retrieve the data segment of the path in the differentiated time period, identify the response segment that recurs within the period, extract the heat response state and time sequence of each segment, the associated change content and corresponding state information, and obtain the path interference sequence structure. S3: Based on the path structure content in the path interference sequence structure, read the number of node state changes within the path segment, filter out the node positions of repeated records, extract the associated points, and obtain the continuously disturbed path node group. S4: Based on the path segment data in the continuous disturbance path node group, read the signal status of the corresponding node in the cloud record, identify the location of transmission interruption and response loss, compare the connection status between the preceding and following nodes, and obtain the interrupted chain segment evaluation list.

2. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture as described in claim 1, characterized in that, The end-side identifier burst segment set includes a path identifier mapping table, a thermal change point index sequence, and a time-related segment location table. The path interference sequence structure includes a path data link configuration table, interference timing status records, and a continuous change correlation matrix. The continuous disturbance path node group includes a repeating state node list, a path number matching index, and a disturbance frequency statistics table. The interrupted link segment evaluation list includes a transmission interruption record sequence, a response missing state marker set, and a connection continuity comparison map.

3. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture as described in claim 1, characterized in that, The associated segment position refers to the time point of each thermal response occurrence in the temperature sensing data collected at the end side, the corresponding time interval of the response occurrence and the path to which it belongs, and the location distribution of the thermal response segment on the path. The index information of the thermal change points refers to identifying the time period of continuous fluctuation during the analysis of thermal response changes, extracting the corresponding start and end time positions, and associating the paths corresponding to the time points.

4. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture as described in claim 1, characterized in that, The related changes refer to analyzing the fluctuation of thermal response data over time under each path, extracting the upward and downward trends of the response, comparing the magnitude of changes in each segment before and after time, and identifying the continuity and rhythm of change in the response process. The thermal response state refers to the temperature changes of the path and nodes in the time series, comparing the temperature trends of the previous and subsequent periods, and analyzing the behavioral characteristics of the thermal response on the time axis.

5. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture as described in claim 1, characterized in that, The specific steps of S1 are as follows: S101: Acquire thermal response change data from the end-side temperature sensing signal, read the path signal value change curve according to the data acquisition time sequence, detect the signal fluctuation segment during the heating process, extract the continuously changing time interval, match the path number corresponding to each segment, and obtain the continuous response segment association information set. S102: Based on the continuous response segment association information set, extract the start and end time index values ​​of each thermal response segment, compare the distribution of path numbers within the time period, identify the alternating areas of thermal response changes between paths, associate the corresponding time points with path identifiers, and obtain thermal change index path comparison data. S103: Based on the path identifier and time period index in the thermal change index path comparison data, the thermal change points and paragraph positions are matched according to the path order, and the matching records between the path number and the response segment are continued to obtain the end-side identifier sudden paragraph set.

6. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture as described in claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the path number information in the burst segment set of the terminal identifier, retrieve the differentiated time period data segment corresponding to each number, match each time interval with the path number, filter the path data records that appear repeatedly in the time period, and obtain the set of repeated path response segment intervals. S202: Based on the thermal response value and time series of each data segment in the repeated path response segment interval set, calculate the slope of response change between adjacent data points, and obtain the path thermal response slope distribution result by corresponding to the average slope of change of each segment and the path identifier. S203: Based on the correspondence between the change amplitude and continuous time period in the path thermal response slope distribution results, identify the path number that continuously fluctuates in thermal response during the period, and associate it with the information of continuously changing segments to obtain the path interference sequence structure.

7. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture as described in claim 1, characterized in that, The specific steps of S3 are as follows: S301: Based on the path structure content in the path interference sequence structure, monitor the thermal response state of nodes in each path segment, identify nodes that respond within a continuous time period, match the thermal response state with the position of occurrence on the timeline, and obtain a set of node state changes. S302: Based on the time index and state change records in the node state change set, calculate the slope of the thermal response change of each node in a continuous time period, associate the slope value with the node number, and obtain the response change trend slope set. S303: Based on the node information in the slope set of the response change trend, identify the node segments with continuously changing slope values ​​in the path sequence, associate the path range and node position corresponding to the node segments, and obtain the continuous perturbation path node group.

8. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture as described in claim 1, characterized in that, The specific steps of S4 are as follows: S401: Based on the path segment data in the continuous disturbance path node group, obtain the signal transmission sequence of the corresponding node in the cloud, examine the signal time sequence of the nodes in the path segment, locate the node position where the signal continuity is interrupted, and obtain the signal timing anomaly set. S402: Based on the signal timing anomaly set, identify the time extension distribution between nodes, track the communication time series changes of nodes in the path connection structure, locate the path segments with missing signals between consecutive nodes, and obtain the node communication discontinuity distribution. S403: Based on the intermittent distribution of node communication, check the continuity status of channel segments in the path structure, identify the node combinations that respond to breakpoints in the connection, analyze the interrupted segments according to the connection status of the path corresponding to the node, and obtain an evaluation list of interrupted chain segments.

9. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture as described in claim 1, characterized in that, The method further includes: S5: Based on the path identifiers in the interrupted chain segment assessment list, analyze the state changes of the path thermal response trajectory under the period, extract the node sequence with continuous fluctuations, analyze the response amplitude and the time interval between nodes, identify the path segments with risk change trends, and obtain the dynamic thermal risk assessment results. The dynamic thermal risk assessment results include a multi-period fluctuation node sequence, a thermal response amplitude distribution matrix, and a time-series risk feature vector.

10. The method for dynamic assessment of battery thermal risk under a cloud-edge-device three-tier architecture according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Based on the path identifier in the interrupted chain segment evaluation list, scan the thermal response change trajectory of the node segment by segment according to the timeline, track the response state changes between adjacent nodes, locate the corresponding segment of the interrupted signal in the time series, and obtain the response discontinuity sequence set. S502: Based on the node number sequence of each segment in the response discontinuity sequence set, synchronously identify the transmission path position corresponding to the node, analyze the fluctuation trend of the node thermal response amplitude in adjacent time windows on the path segment, and obtain the path discontinuity trend sequence. S503: Based on the correspondence between the fluctuation amplitude of each segment and the path number in the path discontinuity trend sequence, extract the path segments in the periodic sequence where the thermal response fluctuation changes, rematch the path number with the continuous fluctuation characteristics, and obtain the dynamic thermal risk assessment result.

Images