Battery tray aluminum profile full life cycle performance prediction method and system

Abstract

The application discloses a battery tray aluminum profile full life cycle performance prediction method and system, and particularly relates to the field of battery tray aluminum profile performance prediction, and comprises the following steps: obtaining edge edge internet of things collection data, link identification, collection time and profile identification of the battery tray aluminum profile in the corresponding edge of extrusion, aging, machining, welding and shaping links, sequentially writing according to the profile identification, collection time and length direction position coding to obtain a previous position record set; the application extracts the position deviation formed by the previous link, and combines the assembly position, the connection relationship and the load transmission path to carry out mapping, accumulation and risk positioning, so as to solve the problem that whether the previous subtle deviation of the battery tray aluminum profile is amplified in the subsequent process and the amplified position is difficult to identify in the prior art.

Description

Technical Field

[0001] This invention relates to the field of performance prediction technology for aluminum profiles for battery trays, and more specifically, to a method and system for predicting the full life-cycle performance of aluminum profiles for battery trays. Background Technology

[0002] In the field of full life-cycle performance prediction for aluminum profiles for battery trays, existing technologies mainly aim to address whether the performance of aluminum profiles will continue to decline and when risks will arise during production, assembly, and subsequent use. A common approach is to rely on edge computing to collect data on dimensions, hardness, weld condition, deformation, vibration, and load information at stages such as extrusion, aging, machining, welding, straightening, assembly, and vehicle operation. Then, based on the data collected at each stage, threshold judgments, rule comparisons, or model calculations are performed to provide performance prediction results. This method can be effective when the impact of a single process is significant and the relationship between changes before and after is direct. However, in the actual application of aluminum profiles for battery trays, a significant abnormality in a particular process does not necessarily manifest as an immediate performance problem. For example, in the process of welding and straightening the aluminum profiles before they enter the tray assembly and are finally installed on the chassis, localized hardness differences, slight springback deviations, boundary weakening, or assembly position differences formed in the previous process often do not reach a significant abnormality at the time, and on-site testing may show that everything is basically normal. However, under the subsequent clamping and fixing, battery pack support, and continuous torsional loads on the vehicle, these seemingly insignificant deviations will gradually be amplified at specific locations, eventually manifesting as more concentrated localized deformation, easier fatigue accumulation, and more concentrated repair locations. The problem with existing technologies is that they mostly treat the data from each stage as separate results, only showing whether a certain stage exceeded limits at the time, but making it difficult to further determine whether the deviations left from the previous stages will be amplified later, at what location, and whether they will truly translate into performance risks. Therefore, the technical problem to be solved by this application is how to identify the process by which subtle performance deviations formed in the preceding stages of battery tray aluminum profile production are gradually amplified during subsequent assembly and use under edge IoT conditions, thereby improving the accuracy of full life cycle performance prediction. Summary of the Invention

[0003] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method and system for predicting the full life cycle performance of aluminum profiles for battery trays. By extracting the positional deviations formed in the preceding stages and combining them with the assembly position, connection relationship, and load transfer path to perform mapping, accumulation, and risk location, the problems mentioned in the background art can be solved.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for predicting the full life-cycle performance of aluminum profiles for battery trays, comprising: S1. Obtain edge IoT data, process identifier, acquisition time and profile identifier corresponding to the extrusion, aging, machining, welding and straightening of the aluminum profile of the battery tray. Write them in order according to the profile identifier, acquisition time and length direction position code to obtain the previous position record set. S2. Extract the collected values ​​of adjacent links under the same position code in the preceding position record set, the collected values ​​of adjacent position codes in the same link, and the collected values ​​of consecutive collection times of the same position code, calculate the corresponding differences and changes, and obtain the preceding deviation record set. S3. Obtain the assembly position data, connection position data and load input position data of the battery tray aluminum profile in the assembly process, establish the mapping relationship between the previous deviation record set and the assembly position data, and determine the transfer order between each assembly position to obtain the position transfer relationship set. S4. For each assembly position, read the previous deviation record of the mapping, and accumulate the difference, change value and connection crossing number under the corresponding position code one by one along the position transmission relationship to obtain the deviation accumulation result; S5. Extract the assembly locations in the cumulative deviation results where the cumulative deviation value continues to increase, the number of connection crossings is greater than one, and the assembly locations are located on the transmission path between the load input location and the connection location. These locations are then identified as target risk locations, and a set of target risk locations is obtained. S6. Generate the full life cycle performance prediction results of the aluminum profile for the battery tray according to the magnitude of the cumulative deviation value, the order of the growth of the cumulative deviation value, and the order of the location transmission of the target risk location set.

[0005] In a preferred embodiment, S1 includes: S1-1. Extract the profile identifier, process identifier, and acquisition time from the edge IoT acquisition data. First, group them according to the profile identifier, then sort them in ascending order according to the acquisition time. If the acquisition time is the same, sort them in the process order of extrusion, aging, machining, welding, and straightening to obtain the process time sequence queue. S1-2. Extract the length direction position code corresponding to each edge IoT acquisition data in the time sequence queue of the extraction process. Write the data with the same position code under the same profile identifier into the same position record in sequence. Write the data with the position code arranged in consecutive numerical order into adjacent position records in sequence to obtain the previous position record set.

[0006] In a preferred embodiment, S2 includes: S2-1. Group the previous position record set by profile identification, extract two adjacent collection values ​​arranged in the link sequence under the same position code, two adjacent collection values ​​arranged in the position code sequence under the same link, and two consecutive collection values ​​arranged in the collection time sequence under the same position code. Subtract the next collection value from the previous collection value to obtain the corresponding difference value, and form the deviation calculation result. S2-2. Write the inter-stage difference, inter-position difference, and time change value in the deviation calculation results into the corresponding positions according to the profile identifier, stage identifier, collection time, and position code. Then, summarize the inter-stage difference, inter-position difference, and time change value corresponding to the same position code into the same deviation record to obtain the previous deviation record set.

[0007] In a preferred embodiment, S3 includes: S3-1. Read the profile identifier, position code and acquisition time from the previous deviation record set. Read the profile identifier, assembly position identifier, start position code, end position code and assembly sequence from the assembly position data. Write the previous deviation records with the same profile identifier and position code within the range of start position code and end position code under the corresponding assembly position identifier. Output the position mapping table. S3-2. Read the acquisition time and assembly order from the position mapping table, retain the records whose acquisition time is earlier than the corresponding time of the assembly order, delete the records whose acquisition time is not earlier than the corresponding time of the assembly order, and output the valid position mapping table.

[0008] In a preferred embodiment, S3 further includes: S3-3. Read the connection start point, connection end point and connection sequence from the connection position data, read the load position and load direction from the load input position data, take the connection record where the connection start point is equal to the load position as the first connection record, and search for the connection records where the connection end point is the same as the connection start point of the next connection record in the connection sequence, write the search results into the transfer chain in order, and output the position transfer chain list. S3-4. Read the assembly position identifier in the valid position mapping table, read the connection start point, connection end point and intra-chain order in the position transfer chain, write the assembly position identifier into the intra-chain order position between the corresponding connection start point and connection end point, generate the previous transfer position and the next transfer position according to the intra-chain order, and output the initial position transfer table.

[0009] In a preferred embodiment, S3 further includes: S3-5. Read multiple assembly position identifiers corresponding to the same preceding deviation record in the initial position transfer table, calculate the absolute value of the difference between the position code and the starting position code corresponding to each assembly position, and the absolute value of the difference between the position code and the ending position code corresponding to each assembly position. Keep the assembly position identifiers whose sum of absolute differences is the lower limit and delete the rest of the assembly position identifiers. Output the single mapping position table. S3-6. Read multiple position transfer chains corresponding to the same assembly position identifier in the single mapping position table, count the number of preceding and following connection records that are connected to the assembly position identifier in each position transfer chain, retain the position transfer chains whose sum of preceding and following connection records is the upper limit, delete the remaining position transfer chains, and output the position transfer relationship set.

[0010] In a preferred embodiment, S4 includes: S4-1. Read the position code, previous transfer position and subsequent transfer position corresponding to each assembly position in the position transfer relationship set. Read the inter-link difference, inter-position difference and time change value under the corresponding position code in the previous deviation record set. Write the inter-link difference, inter-position difference and time change value corresponding to each assembly position under the corresponding assembly position identifier. Output the position deviation table. S4-2. Read the inter-link difference, inter-position difference, and time change value corresponding to each assembly position identifier in the position deviation table. Accumulate the inter-link difference, inter-position difference, and time change value one by one in the order from the previous transfer position to the next transfer position. Increment the connection crossing number by one each time a position transfer relationship is passed, and output the position accumulation table. S4-3. Read the difference accumulation result, change value accumulation result, and connection span count corresponding to each assembly position identifier in the position accumulation table. Write the difference accumulation result, change value accumulation result, and connection span count under the corresponding assembly position identifier, and output the deviation accumulation result.

[0011] In a preferred embodiment, S5 includes: S5-1. Read the cumulative deviation value and connection span number of each assembly position in the cumulative deviation result. Extract three or more consecutive assembly positions according to the transmission order of the same transmission path in the position transmission relationship set. Calculate the difference between the cumulative deviation value of the next assembly position and the cumulative deviation value of the previous assembly position segment by segment. Keep the assembly positions where the difference of each segment is greater than zero and the connection span number is greater than one. Output the candidate risk position table. S5-2. Read the assembly position in the candidate risk position table, read the load position in the load input position data, read the connection start point and connection end point in the connection position data, retrieve the connection record with the load position as the start point and the assembly position in the candidate risk position table as the end point one by one, write the connection record with the connection end point and the connection start point of the next connection record into the corresponding assembly position in the transmission order, retain the assembly position on the continuous transmission path between the load position and the connection position, and output the path hit position table. S5-3. Read the assembly positions in the path hit position table, summarize each assembly position and write it into the target risk position set, and output the target risk position set.

[0012] In a preferred embodiment, S6 includes: S6-1. Read the cumulative deviation value, the order of growth of the cumulative deviation value, and the order of location transmission for each target risk location in the target risk location set. Sort them from largest to smallest according to the cumulative deviation value. For target risk locations with the same cumulative deviation value, sort them from front to back according to the order of growth of the cumulative deviation value. For target risk locations with the same cumulative deviation value and the same order of growth of the cumulative deviation value, sort them from front to back according to the order of location transmission. Output the prediction sorting table. S6-2. Read each target risk location in the prediction sorting table, write the target risk location, cumulative deviation value, cumulative deviation value growth order and location transmission order in sequence according to the sorting order, generate the deterioration sequence number according to the writing order, and output the full life cycle performance prediction result.

[0013] In a preferred embodiment, the battery tray aluminum profile full life cycle performance prediction system includes: The location recording module is used to acquire edge IoT data, process identifiers, acquisition times and profile identifiers corresponding to the extrusion, aging, machining, welding and straightening processes of the battery tray aluminum profile. The data is written sequentially according to the profile identifier, acquisition time and length direction position code to obtain the previous position record set. The deviation extraction module is used to extract the collected values ​​of adjacent links under the same position code in the preceding position record set, the collected values ​​of adjacent position codes of the same link, and the collected values ​​of consecutive collection times of the same position code, calculate the corresponding difference and change value, and obtain the preceding deviation record set. The relationship mapping module is used to obtain the assembly position data, connection position data and load input position data of the battery tray aluminum profile in the assembly process, establish the mapping relationship between the previous deviation record set and the assembly position data, and determine the transfer order between each assembly position to obtain the position transfer relationship set. The deviation accumulation module is used to read the previous deviation record of the mapping for each assembly position, and accumulate the difference, change value and connection span number under the corresponding position code one by one along the position transmission relationship to obtain the deviation accumulation result; The risk location module is used to extract assembly locations in the deviation accumulation results that have continuously increasing cumulative deviation values, more than one connection span, and are located on the transmission path between the load input position and the connection position, and determine them as target risk locations to obtain a set of target risk locations; The results generation module is used to generate full life cycle performance prediction results for battery tray aluminum profiles based on the magnitude of the cumulative deviation value, the order of the cumulative deviation value growth, and the order of location transfer in the target risk location set.

[0014] The technical effects and advantages of this invention are as follows: 1. By integrating the processing of preceding position records, assembly position mapping, and position transfer relationships, it is possible to continuously identify the amplification process of subtle deviations in the preceding process during subsequent assembly and load transfer, thereby relatively improving the accuracy of full life cycle performance prediction. 2. By writing multi-stage edge IoT data sequentially according to profile identification, acquisition time, and location code, the impact of data cross-referencing, reversal, and mispairing at different workstations on subsequent calculation results can be reduced; 3. By extracting the inter-stage difference under the same location code, the inter-location difference under the same stage, and the time change value under the same location code, the scattered collected values ​​can be transformed into a unified deviation expression, improving the consistency of subsequent mapping and accumulation. 4. By mapping the previous deviation record to the assembly position and combining it with the load position, connection start point and connection end point to construct the transmission chain, it is possible to establish a correspondence between the deviation source position and the subsequent structural propagation path; 5. By accumulating the difference, change value, and number of connection crossings along the positional transmission relationship, the cumulative deviation result of the assembly position level is formed, reflecting the continuous accumulation of deviation in the transmission path; 6. By screening out assembly locations where the cumulative deviation value continues to increase and hits the load transmission path, and generating sorting results, the targeting of risk location identification can be improved, and a basis for prioritizing rework and control can be provided. Attached Figure Description

[0015] Fig. 1 This is a flowchart of the method steps of the present invention.

[0016] Fig. 2 This is a schematic diagram of the system modules of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Refer to the instruction manual appendix Figs. 1-2 The method for predicting the full life-cycle performance of aluminum profiles for battery trays of the present invention includes: S1. Obtain edge IoT data, process identifier, acquisition time and profile identifier corresponding to the extrusion, aging, machining, welding and straightening of the aluminum profile of the battery tray. Write them in order according to the profile identifier, acquisition time and length direction position code to obtain the previous position record set. In this embodiment, the processing objective corresponding to S1 is to first merge the edge IoT acquisition data scattered across the preceding stages of extrusion, aging, machining, welding, and straightening, according to the same battery tray aluminum profile object, and then organize it according to the order of formation and the positional relationship along the length of the profile. This provides a unified, continuous, and directly callable data foundation for subsequent calculations of differences between stages under the same position code, calculations of differences between adjacent position codes under the same stage, and calculations of changes in values ​​between consecutive acquisition times under the same position code. Since the acquisition actions of each preceding stage are usually distributed across different workstations... Furthermore, the arrival time, writing time, and completion time of the collected messages are not entirely consistent. If the profile objects, process order, and position order are not uniformly organized first, the data of different processes on the same profile may be prone to cross-referencing, reverse writing, or position mismatch. Therefore, this implementation first extracts and corrects the basic fields in the edge IoT collected data, then forms a process time sequence queue, and then writes the data under the same profile identifier into the corresponding position record according to the length direction position code, so that each position record retains the collected content under different processes at that position, and also maintains the continuity of the adjacent relationship between positions. The implementation process includes the following steps: S1-1. After receiving edge IoT data uploaded from the extrusion, aging, machining, welding, and straightening stages, the edge gateway or workstation acquisition terminal first reads the profile identifier, stage identifier, and acquisition time from each data entry, and standardizes any missing or inconsistent fields. The profile identifier uniquely identifies the same battery tray aluminum profile within each preceding stage; the stage identifier identifies the process stage to which the current data belongs; and the acquisition time indicates the time the data was generated. If the stage name in the acquisition message uses a workstation code, the workstation code is first converted to the corresponding stage identifier in the extrusion, aging, machining, welding, or straightening stage using a pre-stored workstation code-stage name mapping table. If the acquisition time format is inconsistent (milliseconds, seconds, or string format), it is first uniformly converted to the same time format before sorting. After field extraction, the collected data is first grouped using the profile identifier as the key, so that all collected data corresponding to the same profile identifier enters the same profile data set. Then, within each profile data set, the data is sorted in ascending order by collection time from earliest to latest. If two or more data sets have the same collection time, they are arranged in a fixed order of extrusion, aging, machining, welding, and straightening, so that the data at the same collection time still have a definite sequential relationship. For example, if there are machining data and welding data with the time of 10:15:20 under the same profile identifier, the machining data will be placed before the welding data. After grouping and sorting, a sequential queue of process steps is formed under the same profile identifier. Each record in this sequential queue retains the corresponding profile identifier, process identifier, and collection time, and is used as direct input for subsequent position writing. S1-2. After obtaining the stage timing queue, continue to read the length direction position code from each record in the stage timing queue, and establish the corresponding position record writing area for the same profile identifier. The length direction position code is used to characterize the position of the collected value along the length direction of the profile. To ensure that subsequent stages can be compared around the same position, the extrusion, aging, machining, welding, and straightening stages all use the same length direction position code system. For multiple data entries with the same position code under the same profile identifier, write them sequentially into the same position record according to their order of appearance in the stage timing queue, and retain the corresponding stage identifier and collection time in each position record, so that the position record can reflect the collection process of the same position under different preceding stages. For multiple data entries with position codes arranged consecutively by value under the same profile identifier, write them sequentially into adjacent position records in ascending order of position code, so that adjacent position records are stored in the same way as the profile identifier. Adjacent physical positions along the length direction correspond one-to-one. The position codes mentioned here are arranged in numerical succession, meaning that the code difference between the next position code and the previous position code is consistent with the standard position code step size adopted by the system. For example, when the position codes along the length direction are generated at 10 mm intervals, position codes 100, 101, and 102 correspond to three adjacent position records respectively. If position codes 100 and 102 are missing 101, they are not directly regarded as adjacent position records, but are written into their respective position records with the middle position left empty. Through the above writing method, a set of preceding position records composed of multiple position records in sequence can be formed under the same profile identifier. Each position record contains at least the profile identifier, position code, and the data writing items of each link corresponding to that position. Adjacent position records are arranged in the order of position codes, thus providing a unified data organization basis for extracting adjacent link acquisition values ​​according to the same position code and extracting adjacent position code acquisition values ​​according to the same link. Through the above processing, the edge IoT data collected in the previous stages, which were originally scattered across different workstations and at different times, are organized into a data structure under the same profile identifier. This structure is first continuously expanded according to the sequence of stages and then continuously arranged according to the position along the length direction. This allows subsequent steps to directly obtain the corresponding data from the records that have already been merged, sorted by time, and written by position when reading changes between stages at the same location, changes between positions within the same stage, and changes in time at the same location. This avoids problems such as confusion of profile objects, reversal of stage sequence, and mismatch of positions. In practical applications: For example, during the extrusion process, an aluminum profile for a battery tray forms a profile identifier A001, and location codes 001 to 120 are generated along its length. Subsequently, during the aging, machining, welding, and straightening processes, the profile continues to upload corresponding data using the profile identifier A001. The system first categorizes all collected data into the data set corresponding to A001, and then arranges them according to the acquisition time. When the acquisition times for machining and welding data are the same, they are entered into the process sequence queue in the order of machining first, followed by welding. Then, the system reads the location code of each data entry and assigns it to the corresponding location code. The three data entries with code 050, belonging to extrusion, aging, and machining, are sequentially written to the same location record. The three consecutive data entries with location codes 050, 051, and 052 are written to adjacent location records respectively, ultimately forming the preceding location record set corresponding to A001. Subsequent steps can directly extract the difference between the collected values ​​of extrusion and aging from the location record corresponding to location code 050, or extract the position difference from the adjacent location records corresponding to location codes 050 and 051 in the machining stage, thus ensuring that the entire preceding data processing process can directly support the subsequent deviation calculation process.

[0019] S2. Extract the collected values ​​of adjacent links under the same position code in the preceding position record set, the collected values ​​of adjacent position codes in the same link, and the collected values ​​of consecutive collection times of the same position code, calculate the corresponding differences and changes, and obtain the preceding deviation record set. In this embodiment, the processing objective of S2 is to further transform the differences in collected values ​​formed by the same battery tray aluminum profile under different preceding stages, different length positions, and different collection times into deviation data that can be directly used in subsequent mapping and cumulative calculations, based on the already formed preceding position record set. Specifically, the preceding position record set has been organized according to profile identification, stage sequence, collection time, and position code, but this data form is still the original record form and cannot be directly used to determine whether a certain position has shifted between preceding and subsequent stages, whether there are local differences between adjacent positions under a certain stage, or whether the same position has changed between consecutive collection times. Therefore, this embodiment extracts two directly adjacent collected values ​​from the stage dimension, position dimension, and time dimension, respectively, based on the position records under the same profile identification, and then calculates the corresponding difference according to a unified subtraction direction, so that deviation information from different sources enters a unified data expression method. Subsequently, various differences are written back to the corresponding positions according to a unified field and summarized under the same position code to form a preceding deviation record set, so that subsequent steps can directly read the differences between stages, the differences between positions, and the time change values ​​based on the same position code. The implementation process includes the following steps: S2-1. After obtaining the preceding position record set, the preceding position record set is first grouped according to the profile identifier, so that all position records under the same profile identifier enter the same processing range; then, within the processing range corresponding to each profile identifier, three types of data extraction and difference calculation are performed respectively; the first type is the inter-stage difference calculation, that is, under the same position code, two adjacent data collection values ​​are extracted according to the stage sequence; the stage sequence here is consistent with the previous steps, arranged in the order of extrusion, aging, machining, welding and straightening. If the data of one stage is missing under a certain position code, the two actual adjacent data collection values ​​are used. The first category is the calculation of process data as a group of computational objects. For example, if there are three collected values ​​for extrusion, machining, and welding under the same location code, but time-related data is missing, then extrusion and machining are extracted as one group first, and machining and welding are extracted as another group. The second category is the calculation of differences between locations, that is, under the same process, two adjacent collected values ​​are extracted according to the order of location codes. Here, adjacent means that the collected values ​​corresponding to two adjacent location codes appearing in the same process record after sorting by location codes from smallest to largest. If there is no record for a certain location code in the middle, only the two actual adjacent location records are calculated, and no missing position is filled in. The third category is the value of change over time. The calculation involves extracting two consecutively collected values ​​in chronological order under the same location code. Here, consecutive order refers to two records that are adjacent to each other after being sorted from earliest to latest, under the same profile identifier, same stage identifier, and same location code. If only one collection record exists at the same location, that location is not included in the calculation of the time change value. After extracting the above three types of collected values, a unified calculation direction is used for each pair of consecutively collected values, i.e., subtracting the next collected value from the previous one to obtain the corresponding difference. Specifically, the difference between stages is calculated by subtracting the next collected value from the previous stage's collected value, and the difference between locations is calculated by subtracting the previous collected value from the next collected value. The difference between the values ​​collected at different locations is calculated by subtracting the values ​​collected at the next location from the values ​​collected at the previous location. For changes in time, the difference is calculated by subtracting the values ​​collected at the next location from the values ​​collected at the previous location. To ensure consistency in the calculation results, both collected values ​​used in the same difference calculation are taken from the same type of value. For example, when the collected values ​​are wall thickness measurements at the same location, the difference is obtained by directly subtracting the wall thickness; when the collected values ​​are hardness measurements at the same location, the difference is obtained by directly subtracting the hardness. Values ​​of different types are not cross-subtracted. Through the above extraction and calculation process, the deviation calculation results corresponding to the relationships between different locations, stages, and collection times under the same profile identification are obtained. S2-2. After obtaining the deviation calculation results, the inter-stage difference, inter-position difference, and time change value are written into the corresponding record positions according to a unified field standard, and further form a previous deviation record set. Specifically, the profile identifier, stage identifier, acquisition time, and location code corresponding to each deviation calculation result are read first, and this set of fields is used as the basis for writing the location. Specifically, the inter-stage difference is written into the record position corresponding to the next stage under the corresponding location code, the inter-position difference is written into the record position corresponding to the next location code, and the time change value is written into the record position corresponding to the next acquisition time, so that each deviation value can establish a correspondence with its actual current position status. Then, using the location code as the aggregation center, all inter-stage differences, inter-position differences, and time change values ​​corresponding to the same location code under the same profile identifier are merged and written, forming a single deviation record under the same location code. In this deviation record... In addition to retaining the profile identification and location code, the process also retains the process identification and acquisition time corresponding to the deviation value. This allows subsequent steps to retrieve various deviation values ​​and trace the process and time from which the deviation value originated when calling the deviation record. For cases where a certain type of difference value is not formed under a certain location code, such as a location only having a single acquisition without a time change value, the corresponding deviation record is written with a zero value to ensure that the field structure of the same deviation record remains consistent and to avoid interruption of subsequent mapping and accumulation steps due to missing fields. After writing and summarizing all deviation calculation results, the preceding deviation record set is obtained. Each deviation record in this preceding deviation record set uses the location code as the main index and contains the inter-process difference, inter-location difference, and time change value corresponding to that location, thus providing a unified input for subsequent assembly location mapping, establishment of location transfer relationships, and deviation accumulation calculation. Through the above processing, the original acquisition values ​​that were originally scattered in different stages, locations and acquisition times in the preceding position record set are further converted into deviation data with a clear calculation direction, clear source location and clear field structure. This makes it possible for subsequent steps to directly perform mapping, transmission and accumulation operations around the preceding deviation record set instead of backtracking the original acquisition records to recalculate item by item. In practical applications: For example, for a battery tray aluminum profile identified as A001, at location code 050, the collected values ​​are sequentially recorded as follows: extrusion stage value 120, aging stage value 116, and machining stage value 113. At location code 051, the machining stage value is 110. Simultaneously, at location code 050 in the machining stage, there are two corresponding collected values, 113 and 108. The system first extracts the two sets of collected values ​​(120 and 116, 116 and 113) at location code 050 according to the stage sequence, calculating 4 and 3 respectively. Then, it extracts the values ​​corresponding to location codes 050 and 051 in the machining stage. Values ​​113 and 110 are used to calculate 3; then, under the same position code 050 and the same machining process, the two consecutive acquisition times corresponding to the acquisition values ​​113 and 108 are extracted, and 5 is calculated; then, the inter-process difference 4 and 3, the inter-position difference 3, and the time change value 5 are written into the corresponding records according to the profile identifier A001, position code 050 and the corresponding process identifier, and acquisition time, and these values ​​are summarized into the same deviation record corresponding to position code 050, finally forming the previous deviation record set corresponding to A001; subsequent steps can directly read the various deviation values ​​in this deviation record to continue to complete the assembly position mapping and deviation accumulation processing.

[0020] S3. Obtain the assembly position data, connection position data and load input position data of the battery tray aluminum profile in the assembly process, establish the mapping relationship between the previous deviation record set and the assembly position data, and determine the transfer order between each assembly position to obtain the position transfer relationship set. In this embodiment, the processing objective of S3 is to establish a one-to-one correspondence between the positional deviation information in the preceding deviation record set and the actual assembly position in the assembly stage. Then, by combining the connection position data and the load input position data, a path relationship that can reflect the continued transmission of deviation along the assembly structure is formed. The position coding in the preceding deviation record set still adopts the position coding system of the original length direction of the profile. The start position coding and end position coding in the assembly position data also adopt the same position coding system. Therefore, the initial mapping of the preceding deviation record to the assembly position can be completed first using the position coding interval. Then, invalid records occurring after assembly are deleted by using the sequential relationship between the acquisition time and the assembly time. Subsequently, a continuous transmission chain starting from the load position is formed based on the connection start point, connection end point, and connection sequence. Since the same preceding deviation record may fall into multiple assembly position intervals at the same time, and the same assembly position may be located in multiple transmission chains at the same time, it is necessary to further perform single mapping convergence and single chain convergence to ensure that subsequent deviation accumulation and risk positioning are based on a unique assembly position and a unique transmission chain. The implementation process includes the following steps: S3-1. After obtaining the preceding deviation record set, first read the profile identifier, position code, and acquisition time from each preceding deviation record. Then, read the profile identifier, assembly position identifier, start position code, end position code, and assembly sequence from each assembly position data. The assembly position data stores not only the assembly sequence but also the corresponding assembly time. The assembly sequence is used for subsequent chain sorting, and the assembly time is used for time-series filtering. Subsequently, using the profile identifier as the first matching condition, match the preceding deviation records and assembly position data within the same profile range. If the profile identifiers match, then determine whether the position code in the preceding deviation record is at the start or end position of the corresponding assembly position. Within the coding interval, if the position code is greater than or equal to the starting position code and less than or equal to the ending position code, the preceding deviation record is written under the assembly position identifier to form a mapping record. If the same position code is in multiple assembly position intervals, all of them are written under the corresponding assembly position identifier and reserved as subsequent candidate mapping states, and are not deleted in advance in this step. After completing the interval matching of all preceding deviation records and assembly position data, a position mapping table is obtained. Each record in the position mapping table includes at least the profile identifier, position code, acquisition time, assembly position identifier, starting position code, ending position code, and assembly sequence, thereby fixing the candidate correspondence between preceding deviation records and assembly positions. S3-2. After obtaining the position mapping table, read the acquisition time and the assembly time corresponding to the assembly position identifier in each record, and perform a time sequence check on each mapping record. Specifically, records with acquisition times earlier than assembly times are determined to be valid mapping records and retained, while records with acquisition times equal to or later than assembly times are determined to be invalid mapping records and deleted. The reason for adopting this processing method is that the preceding deviation records should come from the processing deviations formed before assembly. If the acquisition time is not earlier than the assembly time, the record cannot be used as a pre-assembly deviation to enter the subsequent assembly position transfer analysis. After completing the time sequence filtering of all mapping records, output the valid position mapping table. Each record in the valid position mapping table simultaneously satisfies three conditions: consistent profile identifier, position code falling within the assembly range, and acquisition time earlier than the assembly time, thus providing valid input for writing the assembly position into the transfer chain. S3-3. After obtaining the effective position mapping table, read the connection start point, connection end point, and connection sequence from the connection position data one by one, and read the load position and load direction from the load input position data one by one. The load direction is used to determine the reading direction from the load position into the connection relationship; that is, using the connection start point corresponding to the load position as the starting side of the transfer chain, read backwards along the connection sequence. Then, first filter out the connection records where the connection start point equals the load position, and write them as the first connection record at the beginning of the transfer chain. Next, using the connection end point of the first connection record as the current end point, search for the next connection record one by one according to the connection sequence. If the connection start point of the next connection record equals the current end point, then write the next connection record sequentially into the same transfer chain, and set the next... The endpoint of a connection record is updated to the new current endpoint, and subsequent searches continue. If the same current endpoint corresponds to multiple subsequent connection records with the same starting point, they are all written to different candidate transfer chains and retained as objects for subsequent single-chain convergence processing. Through the above end-to-end search, one or more candidate transfer chains can be formed, starting from the load position and extending continuously along the connection sequence. The system writes the starting point, endpoint, and position in each candidate transfer chain into the position transfer list in sequence. The position in the chain is defined as the internal order. The internal order of the first connection record is 1, and the internal order is incremented by 1 for each subsequent end-to-end connection record. After completing the search and writing of all connection records, the position transfer list is output. S3-4. After obtaining the valid position mapping table and the position transfer list, read the assembly position identifiers in the valid position mapping table one by one, and then read the connection start point, connection end point, and intra-chain order in the position transfer list one by one, mapping the assembly position identifiers to the connection segments in the transfer chain; in specific execution, first determine whether a certain assembly position identifier is located within a connection segment formed by a certain connection start point and connection end point; if the assembly position interval corresponding to the assembly position identifier overlaps with the connection segment, then write the assembly position identifier to the intra-chain order position corresponding to the connection segment; complete the assembly position mapping table. After the identifier is written, the preceding and following transfer positions are generated sequentially for the assembly position identifiers arranged in the same transfer chain according to the chain order. That is, for two assembly position identifiers that are adjacent in the chain order, the assembly position identifier that is earlier in the chain order is recorded as the preceding transfer position, and the assembly position identifier that is later in the chain order is recorded as the following transfer position. This correspondence is written into the initial position transfer table one by one. Through this process, the preceding deviation record no longer stays at the original position encoding level, but is transformed into the preceding and following transfer relationship in the assembly structure, providing the basis for the propagation order of the assembly stage for the subsequent deviation accumulation. S3-5. After obtaining the initial position transfer table, read the multiple assembly position identifiers corresponding to the same preceding deviation record one by one, and perform single mapping convergence. Specifically, first read the position code corresponding to the preceding deviation record, and then read the start position code and end position code corresponding to each candidate assembly position identifier. For each candidate assembly position identifier, calculate the absolute value of the difference between the position code and the start position code, and the absolute value of the difference between the position code and the end position code, and then add these two absolute values ​​to obtain the boundary difference sum corresponding to the candidate assembly position identifier. Then compare the boundary difference sums of each candidate assembly position identifier corresponding to the same preceding deviation record, and retain the boundary difference sums. The assembly position identifiers whose boundary difference sum is the lower limit are selected, and the mapping records corresponding to the remaining assembly position identifiers are deleted. The reason for adopting this method is that the closer the position code is to the two boundaries of the assembly interval, the more concentrated the affiliation of the preceding deviation record is within the assembly position interval. If the boundary difference sums of two candidate assembly position identifiers are the same, the absolute value of the difference between their starting position code and the position code is compared, and the assembly position identifier with the smaller absolute value of the difference is retained. If they are still the same, the assembly position identifier with the smaller assembly sequence value is retained. After the single mapping convergence of all preceding deviation records is completed, a single mapping position table is output. This single mapping position table ensures that each preceding deviation record corresponds to only one assembly position identifier. S3-6. After obtaining the single-mapped location table, read each of the multiple location transfer chains corresponding to the same assembly location identifier, and perform single-chain convergence. Specifically, for each candidate location transfer chain, first count the number of preceding and following connection records for that assembly location identifier within the chain. The number of preceding connection records refers to the number of consecutive connection records preceding the assembly location identifier within the chain, and the number of following connection records refers to the number of consecutive connection records following the assembly location identifier within the chain. Then, calculate the sum of the number of preceding and following connection records for each candidate location transfer chain, and maintain... Retain the position transfer chains whose sum of quantities is the upper limit, and delete the remaining candidate position transfer chains; if the sum of quantities corresponding to two or more candidate position transfer chains is the same, continue to compare the connection order value of the first connection record of each candidate position transfer chain, and retain the position transfer chain with the smaller connection order value of the first connection record; if they are still the same, compare the connection order value of the last connection record, and retain the position transfer chain with the smaller connection order value of the last connection record; after completing the convergence of the single chains corresponding to all assembly position identifiers, output the position transfer relationship set; each assembly position identifier in this position transfer relationship set is located in only one unique transfer chain, and this transfer chain maintains the structure of starting from the load position and continuously expanding along the connection records connected end to end; Through the above processing, the positional deviation information in the preceding deviation record set is gradually converted into effective mapping relationships and unique transfer relationships in the assembly stage. This allows subsequent steps to directly perform deviation accumulation around the assembly position without needing to return to the original position code and original connection record for repeated judgment. Specifically, S3-1 and S3-2 complete the candidate mapping and temporal filtering from the original position code to the assembly position identifier; S3-3 and S3-4 complete the establishment of a continuous transfer chain starting from the load position and the generation of the preceding and following relationships of the assembly position in the transfer chain; S3-5 and S3-6 further eliminate the uncertainty caused by multiple position assignments of the same preceding deviation record and multiple chain assignments of the same assembly position, thus ensuring that the position transfer relationship set remains unique in terms of object scope, time scope, and path scope. In this way, during subsequent deviation accumulation, each preceding deviation record has a unique assembly position, each assembly position has a unique preceding and following transfer relationship, the calculation chain is continuous, the field source is clear, and deletion and retention rules can be directly executed. In practical applications: For example, in the preceding deviation records of profile identification A001, if the record with position code 050 and acquisition time of 08:15 falls simultaneously into the interval from start position code 045 to end position code 055 corresponding to assembly position P3, and the interval from start position code 050 to end position code 060 corresponding to assembly position P4, the system first writes this record into the candidate mappings corresponding to P3 and P4 simultaneously. If the assembly time corresponding to P3 is 09:10 and the assembly time corresponding to P4 is 09:20, then both candidate mappings are retained. Subsequently, the sum of the absolute values ​​of the differences between position code 050 and the boundaries 045 and 055 of P3 is 10, and the sum of the absolute values ​​of the differences between position code 050 and the boundaries 050 and 060 of P4 is also 10, then the system continues to compare position code 050 with the two... The absolute value of the difference between the starting position codes is 5 for P3 and 0 for P4, so P4 is retained and P3 is deleted. For example, if the load position is L1, and there are two subsequent branches in the connection position data: L1 to C1, C1 to C2, C2 to C3, and C2 to C4, then the system first forms two candidate transfer chains: L1-C1-C2-C3 and L1-C1-C2-C4. If the assembly position P4 falls in both chains, the number of preceding and following connection records of P4 on each chain is counted, and the transfer chain with the larger sum is retained. If the sums are the same, the transfer chain with the smaller connection order value of the first connection record is retained. Thus, the unique assembly position and unique position transfer relationship set corresponding to A001 can be obtained for subsequent steps to continue deviation accumulation and risk location.

[0021] S4. For each assembly position, read the previous deviation record of the mapping, and accumulate the difference, change value and connection crossing number under the corresponding position code one by one along the position transmission relationship to obtain the deviation accumulation result; In this embodiment, the processing purpose of S4 is to further transform the assembly positions that have completed unique mapping and unique transmission chain convergence into deviation accumulation results that can be directly used for risk positioning. The preceding deviation record centrally stores the inter-link difference, inter-position difference, and time change value corresponding to the position code, and the position transmission relationship centrally stores the forward and backward transmission relationship between assembly positions. Therefore, it is necessary to first write the deviation value corresponding to the position code into the corresponding assembly position, and then accumulate it sequentially along the position transmission relationship to obtain the cumulative deviation status of each assembly position in the assembly transmission path. In order to ensure that the subsequent target risk position extraction can be directly obtained, in this embodiment, the inter-link difference and inter-position difference are uniformly treated as difference items, the time change value is treated as a change value item, and the number of connection traversals is generated synchronously as an independent counting field. The implementation process includes the following steps: S4-1. After obtaining the position transfer relationship set, read the position code, previous transfer position, and subsequent transfer position corresponding to each assembly position one by one. The position code is the position code corresponding to the preceding deviation record mapped to that assembly position. The preceding and subsequent transfer positions represent the previous and next assembly positions of that assembly position in the position transfer relationship set, respectively. Then, according to the position code, read the corresponding inter-stage difference, inter-position difference, and time change value from the preceding deviation record set, and write the read inter-stage difference and inter-position difference into the difference item under the assembly position identifier. The time change value is then... The change value is written to the change value item under the assembly position identifier; if the same assembly position corresponds to multiple preceding deviation records, the inter-stage difference, inter-position difference, and time change value in each preceding deviation record are read separately, and summarized and written according to the same assembly position identifier; after writing, a position deviation table is formed; each record in the position deviation table includes at least the assembly position identifier, the preceding transfer position, the following transfer position, the difference item, and the change value item, where the difference item consists of the inter-stage difference and the inter-position difference, and the change value item is the time change value, thereby uniformly transferring the position deviation in the preceding deviation records to the assembly position level; S4-2. After obtaining the position deviation table, read the difference and change values ​​corresponding to each assembly position identifier line by line, and perform accumulation according to the position transfer relationship set from the previous transfer position to the next transfer position. Specifically, first, take the difference and change values ​​at the starting assembly position of each transfer chain as the initial difference accumulation result and initial change value accumulation result of that transfer chain, respectively, and set the initial value of the number of connection traversals to zero. Then, traverse sequentially along the next transfer position of the transfer chain. Each time a new position transfer relationship is entered, add the difference value of the current assembly position to the difference accumulation result corresponding to the previous assembly position, and add the change value of the current assembly position to the result corresponding to the previous assembly position. In the cumulative result of the change values, the number of connection traversals is incremented by one. Following this processing method, the cumulative difference result always represents the cumulative difference between links and positions on the current assembly position's transfer chain up to the current position; the cumulative change value result always represents the cumulative change value of the current assembly position on its transfer chain up to the current position; and the number of connection traversals always represents the number of positional transfer relationships traversed from the starting assembly position to the current assembly position. After traversing and accumulating all transfer chains, a position accumulation table is output. Each record in the position accumulation table includes at least the assembly position identifier, the cumulative difference result, the cumulative change value result, and the number of connection traversals. S4-3. After obtaining the location accumulation table, read the difference accumulation result, change value accumulation result, and connection span count corresponding to each assembly location identifier, and generate the deviation accumulation result. Specifically, first add the difference accumulation result and the change value accumulation result to obtain the deviation accumulation value corresponding to the assembly location. Then, write the deviation accumulation value, difference accumulation result, change value accumulation result, and connection span count together under the corresponding assembly location identifier to form a deviation accumulation record corresponding to the assembly location. After writing all assembly locations, output the deviation accumulation result. Each record in the deviation accumulation result includes at least the assembly location identifier, deviation accumulation value, difference accumulation result, change value accumulation result, and connection span count, so that subsequent steps can directly read the deviation accumulation value and connection span count according to the assembly location and continue to execute the continuous increase judgment and target risk location extraction. Through the above processing, the inter-stage differences, inter-position differences, and time-varying values ​​in the preceding deviation records are uniformly transferred to the assembly position level and accumulated one by one along the unique position transmission relationship to form a deviation accumulation result oriented towards the assembly position. In this way, subsequent steps do not need to backtrack to the preceding deviation records and position transmission relationship set for repeated calculations, but can directly read the deviation accumulation value and connection scrambling number around the assembly position. Furthermore, this embodiment fixes the source, accumulation caliber, and output caliber of the difference item, change value item, and connection scrambling number. The deviation accumulation value is obtained by adding the difference accumulation result and the change value accumulation result. The connection scrambling number is stored independently and is not included in the deviation accumulation value, thereby ensuring that the field source is clear and the calculation chain is continuous when locating risks later. In practical applications: for example, a transmission chain includes assembly positions P1, P2, and P3 in sequence. The inter-stage difference value corresponding to P1 is 3, the inter-position difference value is 2, and the time-varying value is 1. The inter-stage difference value corresponding to P2 is 2, and the inter-position difference value is 1. The time change value is 2. The inter-link difference value for P3 is 1, the inter-position difference value is 3, and the time change value is 1. The system first records the difference item for P1 as 5 and the change value item as 1 in the position deviation table; the difference item for P2 as 3 and the change value item as 2; and the difference item for P3 as 4 and the change value item as 1. Then, it performs accumulation in the order from P1 to P2 to P3, resulting in a difference accumulation result of 5 for P1, a change value accumulation result of 1, and a connection scramble count of 0. The difference accumulation for P2... The results are 8, the cumulative change value is 3, and the number of connections is 1. The cumulative difference of P3 is 12, the cumulative change value is 4, and the number of connections is 2. Then, the cumulative difference of P1, P2, and P3 is added to the cumulative change value, respectively, to obtain the cumulative deviation values ​​of 6, 11, and 16, and these, along with their respective number of connections, are written into the cumulative deviation result. Subsequent steps can then directly use the cumulative deviation values ​​and the number of connections of P1, P2, and P3 to continue extracting the target risk location.

[0022] S5. Extract the assembly locations in the cumulative deviation results where the cumulative deviation value continues to increase, the number of connection crossings is greater than one, and the assembly locations are located on the transmission path between the load input location and the connection location. These locations are then identified as target risk locations, and a set of target risk locations is obtained. In this embodiment, the processing objective of S5 is to screen out the assembly positions that truly need to enter the subsequent performance prediction output from the deviation accumulation results. The previous step has already obtained the deviation accumulation value and the number of connection crossings corresponding to each assembly position. However, not all assembly positions with large deviation accumulation values ​​constitute target risk positions. It is necessary to further determine whether the assembly position is located in a continuously growing segment on the same transmission path, whether it has experienced sufficient connection crossings, and whether it is indeed on a continuous transmission path formed from the load position. Therefore, this embodiment first compares the deviation accumulation value segment by segment along the position transmission relationship set, screens out the assembly positions that are continuously growing on the same path and whose number of connection crossings meets the conditions, then checks back the corresponding connection path starting from the load position, deletes the assembly positions that do not hit the continuous transmission path, and finally summarizes the remaining assembly positions to form a target risk position set. The implementation process includes the following steps: S5-1. After obtaining the cumulative deviation results, first read the cumulative deviation value and connection span number corresponding to each assembly position one by one. Then, extract three or more consecutive assembly positions according to the transmission order of the same transmission path in the position transmission relationship set. Here, three or more consecutive assembly positions refer to three or more assembly positions that are consecutively adjacent in the same position transmission relationship chain. This is used to determine whether the cumulative deviation value forms a continuous increase along the transmission path, rather than just a local increase between two adjacent positions. Subsequently, for each group of extracted consecutive assembly positions, calculate the difference between the cumulative deviation value corresponding to the next assembly position and the cumulative deviation value corresponding to the previous assembly position segment by segment. If the difference values ​​of each segment in the same group of consecutive assembly positions are all greater than zero... If the difference between any two consecutive assembly positions is less than or equal to zero, the group of consecutive assembly positions is determined to satisfy continuous growth on the transfer path. If the difference between any two positions is less than or equal to zero, the group of consecutive assembly positions is determined not to satisfy continuous growth, and the corresponding next assembly position in the group of consecutive assembly positions is deleted. For assembly positions that satisfy continuous growth, their connection span count is further read, and only assembly positions with a connection span count greater than one are retained. Assembly positions with a connection span count equal to zero or equal to one are directly deleted. After completing the segment-by-segment comparison and span count screening of all transfer paths, the retained assembly positions, along with their corresponding cumulative deviation value, connection span count, and transfer path identifier, are written into the candidate risk position table, thereby obtaining candidate assembly positions that simultaneously satisfy the conditions of continuous growth and connection span count. S5-2. After obtaining the candidate risk location table, read the assembly locations one by one, then read the load locations in the load input location data one by one, and read the connection start and connection end points in the connection location data one by one. Perform a path lookup for each candidate assembly location, starting from the load location. Specifically, first, filter out connection records where the connection start point equals the load location, and use this connection record as the first connection record for the path search of that candidate assembly location. Then, using the connection end point of the current connection record as the connection start point for the next search, continue searching the connection location data for the next connection record with the same connection start point and connection end point, and write the retrieved connection records sequentially into the path record corresponding to the candidate assembly location according to the connection order. If writing continuously... If the endpoint of the last connection record after entry falls into the connection segment corresponding to the candidate assembly position, then the candidate assembly position is determined to be on a continuous transmission path formed with the load position as the starting point, and is retained. If, during the retrieval process, the endpoint of the connection record is different from the starting point of the next connection record, or the endpoint of the continuous connection record does not reach the connection segment corresponding to the candidate assembly position, then the candidate assembly position is determined not to have hit the continuous transmission path and is deleted. After completing the path back lookup for all candidate assembly positions, the retained assembly positions and their corresponding path records are written into the path hit position table. Through this step, the assembly positions in the candidate risk position table are further limited to assembly positions on the continuous transmission path between the load position and the connection position. S5-3. After obtaining the path hit location table, read the assembly locations one by one and perform summarization according to the transmission path. Specifically, first, use the transmission path identifier as the grouping key to group assembly locations belonging to the same transmission path into the same summary group. Then, write the assembly locations in each summary group into the target risk location set in the order of their transmission in the transmission path. If the same assembly location repeatedly enters the path hit location table due to different continuous growth segments, deduplication is performed by assembly location before writing to the target risk location set, keeping only one record. After completing the grouping, summarization, and deduplication of all transmission paths, output the target risk location set. Each record in the target risk location set includes at least the assembly location, transmission path identifier, cumulative deviation value, and number of connection crossings, thus providing direct input for subsequent steps to generate full life cycle performance prediction results according to the target risk location. Through the above processing, all assembly positions in the cumulative deviation results are further compressed into assembly positions that simultaneously meet three conditions: continuous growth along the same transfer path, more than one connection span, and located on a continuous transfer path formed from the load position. Thus, subsequent steps no longer require re-checking the cumulative deviation results, position transfer relationship set, and connection position data; instead, they can directly generate sorting results and performance prediction results around the target risk position set. Furthermore, this implementation separates the three actions of continuous growth determination, path hit determination, and target position aggregation sequentially, while maintaining consistent field definitions. Continuous growth is determined based on the cumulative deviation value with segment differences greater than zero; path hit is determined based on the connection endpoint being the same as the connection start point of the next connection record; and the target risk position set is formed by grouping and deduplicating based on the transfer path. Therefore, the entire screening process has a clear source of values, a fixed calculation direction, and direct connections between steps. In practical applications: for example, a position transfer path includes assembly positions P1, P2, P3, and P4 sequentially, whose deviations... The cumulative values ​​are 6, 11, 16, and 18, and the number of connection spans are 0, 1, 2, and 3, respectively. The system first extracts the continuous assembly positions from P1 to P4. The difference between P2 and P1 is 5, the difference between P3 and P2 is 5, and the difference between P4 and P3 is 2. Since all differences are greater than zero, this group of assembly positions satisfies continuous growth. Then, it filters by the number of connection spans, retaining only P3 and P4, where the number of connection spans is greater than one. Subsequently, if the load position is L1, and the connection position data includes L1 to C1, C1 to C... 2. For continuous connection records from C2 to C3, and where P3 and P4 are located within the corresponding connection segments from C2 to C3 and their subsequent connection segments respectively, then P3 and P4 are both written into the path hit location table. If a certain assembly location satisfies continuous growth but cannot be continuously retrieved from the connection location data by L1, then that assembly location is deleted. Finally, the system writes P3 and P4 retained under the same path into the target risk location set for subsequent steps to continue generating the full life cycle performance prediction results of the battery tray aluminum profile.

[0023] S6. Generate the full life cycle performance prediction results of the aluminum profile for the battery tray according to the magnitude of the cumulative deviation value, the order of the growth of the cumulative deviation value, and the order of the location transfer in the target risk location set. In this embodiment, the processing purpose of S6 is to uniformly sort the identified target risk locations and generate a full life cycle performance prediction result that can be directly output. The previous step has obtained a set of target risk locations, but each target risk location has only completed risk screening and has not yet formed a clear order. In order to enable the subsequent output results to directly represent the risk level and deterioration order of each target risk location, this embodiment first reads the cumulative deviation value, the order of growth of the cumulative deviation value, and the order of location transmission corresponding to each target risk location, and then sorts them level by level according to a fixed sorting rule to form a prediction sorting table. Subsequently, the prediction fields are written in sequence according to the sorting results, and a deterioration order number is generated. The implementation process includes the following steps: S6-1. After obtaining the target risk location set, read the cumulative deviation value, cumulative deviation value growth order, and location transmission order for each target risk location. The cumulative deviation value characterizes the degree of cumulative deviation at that target risk location; the cumulative deviation value growth order characterizes the order in which the target risk location enters a continuously growing state in the corresponding transmission path; and the location transmission order characterizes the transmission order of the target risk location in the location transmission relationship. Then, perform a three-level sorting of all target risk locations: first, compare the cumulative deviation values ​​and arrange them from largest to smallest; if two or more target risk locations have the same cumulative deviation value, continue comparing the cumulative deviation value growth order and arrange them from front to back; if both the cumulative deviation value and the cumulative deviation value growth order are the same, continue comparing the location transmission order and arrange them from front to back. After completing the three-level sorting of all target risk locations, write the sorting results into the prediction sorting table in order of priority. Each record in the prediction sorting table includes at least the target risk location, cumulative deviation value, cumulative deviation value growth order, location transmission order, and sorting position. S6-2. After obtaining the prediction ranking table, read each target risk location from front to back according to the ranking position, and write them into the full life cycle performance prediction results in the order of reading. In specific execution, for each ranking record, write the target risk location, cumulative deviation value, cumulative deviation value growth order, and location transmission order in sequence. At the same time, use the reading order of the record in the prediction ranking table as the degradation sequence number and write it into the corresponding position. The degradation sequence number corresponding to the first record in the prediction ranking table is 1, and the degradation sequence number of the subsequent record is 1 based on the previous record. After writing all ranking records, output the full life cycle performance prediction results. Each record in the full life cycle performance prediction results includes at least the target risk location, cumulative deviation value, cumulative deviation value growth order, location transmission order, and degradation sequence number, so that the output results have both risk level ranking and degradation order ranking information. Through the above processing, each target risk location in the target risk location set is further organized into a full life cycle performance prediction result with a unified sorting caliber. In this way, subsequent users no longer need to repeatedly compare the original fields in the target risk location set. They can directly determine which locations have higher accumulated deviations, which locations enter the growth state earlier, and which locations are earlier in the transmission path based on the output results. Furthermore, this implementation method fixes the sorting process into a three-level sorting of accumulated deviation value, accumulated deviation value growth order, and location transmission order. The generated degradation sequence number is directly derived from the writing order after sorting. Therefore, the field source is clear, the sorting direction is fixed, and the output results can be directly called. In practical applications: For example, the target risk location set includes assembly locations P3, P5, and P7. P3 has a cumulative deviation value of 16, a cumulative deviation value growth order of 2, and a position transfer order of 4. P5 has a cumulative deviation value of 16, a cumulative deviation value growth order of 1, and a position transfer order of 5. P7 has a cumulative deviation value of 12, a cumulative deviation value growth order of 1, and a position transfer order of 3. The system first arranges these locations in descending order of cumulative deviation value, with P3 and P5 preceding P7. Then, it compares the cumulative deviation value growth order of P3 and P5. Since P5's cumulative deviation value growth order is earlier, P5 is ranked before P3. The resulting prediction ranking table is P5, P3, and P7. Subsequently, the system writes the full lifecycle performance prediction results in this order and assigns them degradation sequence numbers 1, 2, and 3 respectively, thus generating a prediction result that can be directly output.

[0024] Furthermore, it also includes a full life-cycle performance prediction system for battery tray aluminum profiles, including: The location recording module is used to acquire edge IoT data, process identifiers, acquisition times and profile identifiers corresponding to the extrusion, aging, machining, welding and straightening processes of the battery tray aluminum profile. The data is written sequentially according to the profile identifier, acquisition time and length direction position code to obtain the previous position record set. The deviation extraction module is used to extract the collected values ​​of adjacent links under the same position code in the preceding position record set, the collected values ​​of adjacent position codes of the same link, and the collected values ​​of consecutive collection times of the same position code, calculate the corresponding difference and change value, and obtain the preceding deviation record set. The relationship mapping module is used to obtain the assembly position data, connection position data and load input position data of the battery tray aluminum profile in the assembly process, establish the mapping relationship between the previous deviation record set and the assembly position data, and determine the transfer order between each assembly position to obtain the position transfer relationship set. The deviation accumulation module is used to read the previous deviation record of the mapping for each assembly position, and accumulate the difference, change value and connection span number under the corresponding position code one by one along the position transmission relationship to obtain the deviation accumulation result; The risk location module is used to extract assembly locations in the deviation accumulation results that have continuously increasing cumulative deviation values, more than one connection span, and are located on the transmission path between the load input position and the connection position, and determine them as target risk locations to obtain a set of target risk locations; The results generation module is used to generate full life-cycle performance prediction results for battery tray aluminum profiles based on the magnitude of the cumulative deviation value, the order of the cumulative deviation value growth, and the order of location transmission within the target risk location set. Working Principle: This solution first collects edge IoT data generated during the extrusion, aging, machining, welding, and straightening processes of the battery tray aluminum profile. This data is organized by profile type, time sequence, and length position to form a preceding position record. Then, it calculates the preceding deviation record based on the changes at the same position between preceding and following processes, between adjacent positions within the same process, and between different time points. Subsequently, these preceding deviation records are mapped to the actual assembly positions after assembly, and a transmission path for deviations in the assembly structure is established by combining connection relationships and load input positions. The difference, change value, and number of connection crossings corresponding to each assembly position are then gradually accumulated along the transmission path to obtain the cumulative deviation result. Finally, assembly positions that continuously increase along the same path and are located on the load transmission path are selected as target risk positions, and the full life-cycle performance prediction results are output according to the degree of deviation accumulation and the order of occurrence. In other words, this solution does not only consider the single inspection result of a particular process, but connects the preceding processing deviations, subsequent assembly relationships, and the load transmission process into a unified whole. The complete calculation chain determines which deviations will continue to propagate and amplify in subsequent structures. For example, after the aluminum profile of the battery tray is welded and shaped, dimensional deviations, local deformation values, or detection values ​​at certain lengths are collected at the edge nodes on site. This data is first organized by profile and position to calculate the preceding deviations. When the profile enters the tray assembly, the system determines which assembly position in the assembly falls on the corresponding section of the profile, the connection position, and the load input position, and continues to propagate it along the actual connection path. If the cumulative deviation value of several subsequent assembly positions on a certain transmission path increases continuously, and these positions are on the continuous transmission path after the load is input, the system will identify these positions as target risk positions and provide predictions of which position will deteriorate earlier and which position will have a higher cumulative deviation. In this way, in actual production and assembly scenarios, workers do not need to wait until the tray is in service to see where the problem is, but can know in advance which positions are more likely to form performance risks during the assembly stage.

[0025] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the full life-cycle performance of aluminum profiles for battery trays, characterized in that, include: S1. Obtain edge IoT data, process identifier, acquisition time and profile identifier corresponding to the extrusion, aging, machining, welding and straightening of the aluminum profile of the battery tray. Write them in order according to the profile identifier, acquisition time and length direction position code to obtain the previous position record set. S2. Extract the collected values ​​of adjacent links under the same position code in the preceding position record set, the collected values ​​of adjacent position codes in the same link, and the collected values ​​of consecutive collection times of the same position code, calculate the corresponding differences and changes, and obtain the preceding deviation record set. S3. Obtain the assembly position data, connection position data and load input position data of the battery tray aluminum profile in the assembly process, establish the mapping relationship between the previous deviation record set and the assembly position data, and determine the transfer order between each assembly position to obtain the position transfer relationship set. S4. For each assembly position, read the previous deviation record of the mapping, and accumulate the difference, change value and connection crossing number under the corresponding position code one by one along the position transmission relationship to obtain the deviation accumulation result; S5. Extract the assembly locations in the cumulative deviation results where the cumulative deviation value continues to increase, the number of connection crossings is greater than one, and the assembly locations are located on the transmission path between the load input location and the connection location. These locations are then identified as target risk locations, and a set of target risk locations is obtained. S6. Generate the full life cycle performance prediction results of the aluminum profile for the battery tray according to the magnitude of the cumulative deviation value, the order of the growth of the cumulative deviation value, and the order of the location transmission of the target risk location set.

2. The method for predicting the full life cycle performance of battery tray aluminum profiles according to claim 1, characterized in that: S1 includes: S1-1. Extract the profile identifier, process identifier, and acquisition time from the edge IoT acquisition data. First, group them according to the profile identifier, then sort them in ascending order according to the acquisition time. If the acquisition time is the same, sort them in the process order of extrusion, aging, machining, welding, and straightening to obtain the process time sequence queue. S1-2. Extract the length direction position code corresponding to each edge IoT acquisition data in the time sequence queue of the extraction process. Write the data with the same position code under the same profile identifier into the same position record in sequence. Write the data with the position code arranged in consecutive numerical order into adjacent position records in sequence to obtain the previous position record set.

3. The method for predicting the full life cycle performance of battery tray aluminum profiles according to claim 2, characterized in that: S2 includes: S2-1. Group the previous position record set by profile identification, extract two adjacent collection values ​​arranged in the link sequence under the same position code, two adjacent collection values ​​arranged in the position code sequence under the same link, and two consecutive collection values ​​arranged in the collection time sequence under the same position code. Subtract the next collection value from the previous collection value to obtain the corresponding difference value, and form the deviation calculation result. S2-2. Write the inter-stage difference, inter-position difference, and time change value in the deviation calculation results into the corresponding positions according to the profile identifier, stage identifier, collection time, and position code. Then, summarize the inter-stage difference, inter-position difference, and time change value corresponding to the same position code into the same deviation record to obtain the previous deviation record set.

4. The method for predicting the full life cycle performance of battery tray aluminum profiles according to claim 3, characterized in that: S3 includes: S3-1. Read the profile identifier, position code and acquisition time from the previous deviation record set. Read the profile identifier, assembly position identifier, start position code, end position code and assembly sequence from the assembly position data. Write the previous deviation records with the same profile identifier and position code within the range of start position code and end position code under the corresponding assembly position identifier. Output the position mapping table. S3-2. Read the acquisition time and assembly order from the position mapping table, retain the records whose acquisition time is earlier than the corresponding time of the assembly order, delete the records whose acquisition time is not earlier than the corresponding time of the assembly order, and output the valid position mapping table.

5. The method for predicting the full life cycle performance of battery tray aluminum profiles according to claim 4, characterized in that: S3 further includes: S3-3. Read the connection start point, connection end point and connection sequence from the connection position data. Read the load position and load direction from the load input position data. Take the connection record where the connection start point is equal to the load position as the first connection record. Search for the connection records where the connection end point is the same as the connection start point of the next connection record in the connection sequence. Write the search results into the transfer chain in order and output the position transfer chain list. S3-4. Read the assembly position identifier in the valid position mapping table, read the connection start point, connection end point and intra-chain order in the position transfer chain, write the assembly position identifier into the intra-chain order position between the corresponding connection start point and connection end point, generate the previous transfer position and the next transfer position according to the intra-chain order, and output the initial position transfer table.

6. The method for predicting the full life cycle performance of aluminum profiles for battery trays according to claim 5, characterized in that: S3 further includes: S3-5. Read multiple assembly position identifiers corresponding to the same preceding deviation record in the initial position transfer table, calculate the absolute value of the difference between the position code and the starting position code corresponding to each assembly position, and the absolute value of the difference between the position code and the ending position code corresponding to each assembly position. Keep the assembly position identifiers whose sum of absolute differences is the lower limit and delete the rest of the assembly position identifiers. Output the single mapping position table. S3-6. Read multiple position transfer chains corresponding to the same assembly position identifier in the single mapping position table, count the number of preceding and following connection records that are connected to the assembly position identifier in each position transfer chain, retain the position transfer chains whose sum of preceding and following connection records is the upper limit, delete the remaining position transfer chains, and output the position transfer relationship set.

7. The method for predicting the full life cycle performance of aluminum profiles for battery trays according to claim 6, characterized in that: S4 includes: S4-1. Read the position code, previous transfer position and subsequent transfer position corresponding to each assembly position in the position transfer relationship set. Read the inter-link difference, inter-position difference and time change value under the corresponding position code in the previous deviation record set. Write the inter-link difference, inter-position difference and time change value corresponding to each assembly position under the corresponding assembly position identifier. Output the position deviation table. S4-2. Read the inter-link difference, inter-position difference, and time change value corresponding to each assembly position identifier in the position deviation table. Accumulate the inter-link difference, inter-position difference, and time change value one by one in the order from the previous transfer position to the next transfer position. Increment the connection crossing number by one each time a position transfer relationship is passed, and output the position accumulation table. S4-3. Read the difference accumulation result, change value accumulation result, and connection span count corresponding to each assembly position identifier in the position accumulation table. Write the difference accumulation result, change value accumulation result, and connection span count under the corresponding assembly position identifier, and output the deviation accumulation result.

8. The method for predicting the full life cycle performance of battery tray aluminum profiles according to claim 7, characterized in that: S5 includes: S5-1. Read the cumulative deviation value and connection span number of each assembly position in the cumulative deviation result. Extract three or more consecutive assembly positions according to the transmission order of the same transmission path in the position transmission relationship set. Calculate the difference between the cumulative deviation value of the next assembly position and the cumulative deviation value of the previous assembly position segment by segment. Keep the assembly positions where the difference of each segment is greater than zero and the connection span number is greater than one. Output the candidate risk position table. S5-2. Read the assembly position in the candidate risk position table, read the load position in the load input position data, read the connection start point and connection end point in the connection position data, retrieve the connection record with the load position as the start point and the assembly position in the candidate risk position table as the end point one by one, write the connection record with the connection end point and the connection start point of the next connection record into the corresponding assembly position in the transmission order, retain the assembly position on the continuous transmission path between the load position and the connection position, and output the path hit position table. S5-3. Read the assembly positions in the path hit position table, summarize each assembly position and write it into the target risk position set, and output the target risk position set.

9. The method for predicting the full life cycle performance of battery tray aluminum profiles according to claim 8, characterized in that: S6 includes: S6-1. Read the cumulative deviation value, the order of growth of the cumulative deviation value, and the order of location transmission for each target risk location in the target risk location set. Sort them from largest to smallest according to the cumulative deviation value. For target risk locations with the same cumulative deviation value, sort them from front to back according to the order of growth of the cumulative deviation value. For target risk locations with the same cumulative deviation value and the same order of growth of the cumulative deviation value, sort them from front to back according to the order of location transmission. Output the prediction sorting table. S6-2. Read each target risk location in the prediction sorting table, write the target risk location, cumulative deviation value, cumulative deviation value growth order and location transmission order in sequence according to the sorting order, generate the deterioration sequence number according to the writing order, and output the full life cycle performance prediction result.

10. A full life-cycle performance prediction system for battery tray aluminum profiles, characterized in that, include: The location recording module is used to acquire edge IoT data, process identifiers, acquisition times and profile identifiers corresponding to the extrusion, aging, machining, welding and straightening processes of the battery tray aluminum profile. The data is written sequentially according to the profile identifier, acquisition time and length direction position code to obtain the previous position record set. The deviation extraction module is used to extract the collected values ​​of adjacent links under the same position code in the preceding position record set, the collected values ​​of adjacent position codes of the same link, and the collected values ​​of consecutive collection times of the same position code, calculate the corresponding difference and change value, and obtain the preceding deviation record set. The relationship mapping module is used to obtain the assembly position data, connection position data and load input position data of the battery tray aluminum profile in the assembly process, establish the mapping relationship between the previous deviation record set and the assembly position data, and determine the transfer order between each assembly position to obtain the position transfer relationship set. The deviation accumulation module is used to read the previous deviation record of the mapping for each assembly position, and accumulate the difference, change value and connection span number under the corresponding position code one by one along the position transmission relationship to obtain the deviation accumulation result; The risk location module is used to extract assembly locations in the deviation accumulation results that have continuously increasing cumulative deviation values, more than one connection span, and are located on the transmission path between the load input position and the connection position, and determine them as target risk locations to obtain a set of target risk locations; The results generation module is used to generate full life cycle performance prediction results for battery tray aluminum profiles based on the magnitude of the cumulative deviation value, the order of the cumulative deviation value growth, and the order of location transfer in the target risk location set.

Images