Vehicle simulation calculation data post-processing method and device, equipment and storage medium

Abstract

The method, apparatus, equipment, and storage medium for post-processing vehicle simulation calculation data identify discrete simulation result files through global scanning, encapsulate them into atomic tasks, construct a globally unified queue, and silently distribute and execute them using a multi-core parallel resource pool. Based on preset semantic anchors, dynamic addressing and semantic mapping are completed, adapting to heterogeneous files with non-deterministic suffixes. At the data extraction source, structural-level dimensionality reduction and purification are performed to eliminate redundant data, ultimately outputting standardized structured data. This significantly improves post-processing efficiency, reduces system resource consumption, enhances system robustness, and allows the output data to be directly connected to modern data analysis tools, ensuring the consistency and accuracy of data processing in long-cycle projects.

Description

Technical Field

[0001] This application relates to the field of automotive simulation, specifically to a method, apparatus, device, and storage medium for post-processing vehicle simulation calculation data. Background Technology

[0002] With the rapid development of automotive digital R&D technology, computer-aided virtual verification has become a core means of evaluating vehicle powertrain configuration, control strategies, and overall vehicle performance. To obtain the optimal design solution, the R&D process requires multiple rounds of large-scale simulation calculations under various operating conditions and parameter combinations. This generates massive amounts of multi-dimensional raw calculation results data, leading the industry to place increasingly higher demands on the automation level, processing efficiency, data purity, and ecosystem compatibility of simulation data post-processing.

[0003] In related technologies, manual data extraction is usually performed using the graphical user interface built into engineering software, or by writing single-session scripts that rely on hard-coded parameters to complete the processing operations.

[0004] However, it cannot achieve concurrent batch processing of massive discrete simulation result files. The post-processing stage relies on manual interaction triggering, which is inefficient and seriously restricts the operation speed of the digital R&D pipeline. The hard-coded static addressing method cannot adapt to the heterogeneous files with non-deterministic suffixes output by the computing engine, which can easily lead to processing interruption and poor system robustness. Summary of the Invention

[0005] This application provides a method, apparatus, device, and storage medium for post-processing vehicle simulation calculation data, which can solve the technical problems existing in related technologies, such as low automation of simulation data post-processing, inability to achieve concurrent batch processing of massive discrete result files, difficulty in adapting static addressing to heterogeneous files with non-deterministic suffixes, easy processing interruption, and poor system robustness.

[0006] In a first aspect, embodiments of this application provide a method for post-processing vehicle simulation calculation data, which includes the following steps: Read the preset system configuration information and scan the vehicle simulation data directory to be processed. Based on the scan results, construct a global atomic task queue and map each discrete simulation result file to an atomic task. Based on a multi-core parallel computing resource pool, tasks in the global atomic task queue are silently distributed to multiple worker processes, and each worker process concurrently executes the parsing and processing of the corresponding atomic tasks. Each process performs dynamic addressing and semantic mapping of the corresponding simulation result files within the topology directory space based on preset semantic anchor points, and performs structural-level dimensionality reduction and purification of the parsed simulation data channels based on the naming ontology features of the industrial solver. The purified effective time-series data channels output by each working process are flattened and reorganized into a two-dimensional table structure. Storage paths are generated according to the semantically mapped working condition names, and the reorganized structured data is serialized into a preset columnar storage format.

[0007] In conjunction with the first aspect, in one implementation, the step of reading preset system configuration information and scanning the catalog of vehicle simulation data to be processed, and constructing a global atomic task queue based on the scan results, includes: Read and parse the system configuration file, which includes the source data path, target output path, list of tasks to be processed, and list of component semantic identifiers; Recursively scan all simulation result files and associated metadata files under the source data path according to the preset directory topology hierarchy; Based on the list of tasks to be processed, the scanned files are filtered to select the target simulation result files that need to be processed. Extract the file identifier, storage path, and associated metadata information of each target simulation result file, and encapsulate them into indivisible atomic task units; All atomic tasks are sorted according to a preset priority rule to generate a globally unified atomic task queue.

[0008] In conjunction with the first aspect, in one implementation, the step of silently distributing tasks in the global atomic task queue to multiple worker processes based on a multi-core parallel computing resource pool, with each worker process concurrently executing the parsing and processing of the corresponding atomic task, includes: Create a corresponding number of headless worker processes based on the current number of processor cores in the device, and build a parallel computing resource pool. A load balancing strategy is adopted to distribute tasks in the global atomic task queue to idle worker processes in sequence. All worker processes silently execute data parsing tasks in the background, skipping the graphical rendering stage.

[0009] In conjunction with the first aspect, in one implementation, the step of silently distributing tasks in the global atomic task queue to multiple worker processes, with each worker process concurrently executing the parsing and processing of the corresponding atomic task, further includes: Real-time monitoring of the running status of all working processes and the execution progress of each atomic task; If an atomic task fails to execute and the cumulative number of failures has not reached the preset retry threshold, the atomic task will be added back to the tail of the global atomic task queue to wait for re-distribution and execution. If the cumulative number of failures reaches the preset retry threshold, the atomic task is marked as a permanently failed task, an exception alarm record is generated and written to the global processing log.

[0010] In conjunction with the first aspect, in one implementation, each working process, based on semantic anchors in a preset component semantic identifier list, performs dynamic addressing and semantic mapping of the corresponding simulation result files within the topological directory space, and performs structural-level dimensionality reduction and purification of the parsed simulation data channels based on the naming ontology features of the industrial solver, including: Each working process obtains the file information corresponding to the assigned atomic task, uses the semantic anchor point identifier fragment in the component semantic identifier list to perform fuzzy matching on the corresponding simulation result file with non-deterministic suffix, and establishes a dynamic mapping relationship between the logical processing component and the physical simulation data entity. Parse the metadata configuration file corresponding to the atomic task and reverse map the mechanized task directory code to the real working condition name with physical meaning; The simulation result file obtained by parsing is traversed through all data channels, and redundant data channels with system-level metadata pre-reservation characters and non-temporal static attribute post-labels are identified and eliminated, retaining only the effective temporal physical quantity data channels with dynamic evolution processes.

[0011] In conjunction with the first aspect, in one implementation, the step of flattening and recombining the purified effective time-series data channels output by each working process into a standard two-dimensional table structure, generating a hierarchical storage path according to the semantically mapped working condition name and the preset target output path, and serializing the recombined structured data into a preset columnar storage format includes: Summarize all valid timing data channels output by each working process, perform flattening processing, and construct a standard two-dimensional table structure; By combining the target output path with the actual working condition name corresponding to each atomic task, a hierarchical target storage path is automatically generated. Serialize time-series data with a standard two-dimensional table structure into Parquet or CSV format; The serialized data file is stored in the corresponding target storage path, and a processing log file including task execution status, processing time and exception information is generated.

[0012] In conjunction with the first aspect, in one implementation, after performing structural-level dimensionality reduction and purification on the parsed simulation data channels based on the industrial solver naming ontology features, the method further includes: Data integrity verification and physical quantity rationality verification are performed on each effective time-series physical quantity data channel after purification; If the verification passes, proceed with the subsequent steps of flattening and reassembling the valid time-series data channels. If the verification fails, the corresponding atomic task is marked as an abnormal task, the abnormality type and corresponding file information are recorded, and the subsequent processing flow of the task is skipped.

[0013] Secondly, embodiments of this application provide a vehicle simulation calculation data post-processing device, the vehicle simulation calculation data post-processing device comprising: The task preparation module is used to read the preset system configuration information and scan the directory of vehicle simulation data to be processed. Based on the scan results, a global atomic task queue is constructed, and each discrete simulation result file is mapped to an atomic task. The parallel scheduling module is used for the multi-core parallel computing resource pool. It silently distributes tasks in the global atomic task queue to multiple worker processes, and each worker process concurrently executes the parsing and processing of the corresponding atomic task. The parsing and purification module is used by each working process to perform dynamic addressing and semantic mapping of the corresponding simulation result files in the topology directory space based on preset semantic anchors, and to perform structural dimensionality reduction and purification of the parsed simulation data channels based on the naming ontology features of the industrial solver. The output module is used to flatten and reorganize the purified effective time-series data channels output by each working process into a two-dimensional table structure, generate storage paths according to the semantically mapped working condition names, and serialize the reorganized structured data into a preset columnar storage format.

[0014] Thirdly, embodiments of this application provide a vehicle simulation calculation data post-processing device, which includes a processor, a memory, and a vehicle simulation calculation data post-processing program stored in the memory and executable by the processor. When the vehicle simulation calculation data post-processing program is executed by the processor, it implements the steps of the vehicle simulation calculation data post-processing method as described in some of the above embodiments.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium storing a vehicle simulation calculation data post-processing program, wherein when the vehicle simulation calculation data post-processing program is executed by a processor, it implements the steps of the vehicle simulation calculation data post-processing method as described in some of the above embodiments.

[0016] The beneficial effects of the technical solutions provided in this application include: By globally scanning and identifying discrete simulation result files, encapsulating them into atomic tasks and constructing a globally unified queue, and silently distributing and executing them using a multi-core parallel resource pool, dynamic addressing and semantic mapping are completed based on preset semantic anchors, adapting to heterogeneous files with non-deterministic suffixes. At the data extraction source, structural-level dimensionality reduction and purification are performed to remove redundant data, ultimately outputting standardized structured data, significantly improving post-processing efficiency, reducing system resource consumption, enhancing system robustness, and allowing the output data to be directly connected to modern data analysis tools, ensuring the consistency and accuracy of data processing in long-cycle projects. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating an embodiment of the vehicle simulation calculation data post-processing method of this application; Figure 2 This is a flowchart illustrating another embodiment of the vehicle simulation calculation data post-processing method of this application; Figure 3 This is a schematic diagram of the hardware structure of the vehicle simulation calculation data post-processing device involved in the embodiments of this application. Detailed Implementation

[0018] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0019] First, some of the technical terms used in this application will be explained to help those skilled in the art understand this application.

[0020] Semantic Anchor: In this application, it specifically refers to a predefined file or directory identifier fragment with stable semantic features, used for fuzzy matching and positioning in heterogeneous file systems with non-deterministic suffixes (such as timestamps and version numbers), and is the core basis for realizing dynamic mapping between logical components and physical data entities.

[0021] Atomic task queue: refers to a global task list that maps each discrete simulation result file to be processed into an indivisible smallest execution unit (atomic task) and arranges them according to a unified rule. It is the basis of the parallel scheduling mechanism.

[0022] Silent Batch Processing / Headless Execution: refers to a running mode that automatically completes all batch data processing tasks in the background without requiring manual interaction or monitoring through a graphical user interface (GUI), skipping all graphics rendering stages.

[0023] Naming ontology features: These refer to the inherent semantic rules followed in the naming system of output variables of industrial simulation solvers, including pre-reservation characters of system-level metadata (such as double underscores) and post-labels of non-temporal static attributes (such as _Property), which are the core judgment criteria of the data purification algorithm in this application.

[0024] Structural-level dimensionality reduction refers to a data processing method that directly identifies and removes non-temporally redundant dimensions (such as metadata and static attributes) based on the semantic features of variable naming at the data extraction source, retaining only the effective temporally and physically quantity channels. This is different from the traditional mode of full loading followed by filtering.

[0025] Channel purification: This refers to the operation of verifying all data channels in the simulation data one by one through a preset judgment function, retaining only the valid timing physical quantity channels with the status of Keep, and eliminating all redundant channels.

[0026] Semantic inversion: refers to the process of reversibly mapping the machined task directory code (such as task.083) generated by the underlying computing engine into a real working condition name (such as WLTC, CHTC-TT) with business physical meaning by parsing the metadata configuration file.

[0027] Columnar storage format: refers to a storage format that organizes data by columns rather than rows (such as Parquet). It features high compression ratio, efficient column lookup and low I / O overhead, making it suitable for storing and analyzing large-scale structured data.

[0028] Analysis-Ready Data: refers to standardized data that has undergone automated cleaning, dimensionality reduction, and structure reorganization, and can be directly input into modern data science toolchains for advanced analysis without additional preprocessing.

[0029] CSV format: Short for Comma-Separated Values, it's a common plain text structured data storage format. It stores tabular data in plain text, with each row representing a data record. Different fields within each record are separated by commas (or other preset delimiters). This format is characterized by its simple structure, high readability, and excellent cross-platform compatibility; almost all mainstream data analysis tools, spreadsheet software, and database systems natively support its read and write operations.

[0030] Parquet format: Apache Parquet is designed for efficient storage and querying of large-scale structured data. Its core feature is that it organizes data by column rather than by row, supports nested data structures, various compression algorithms (such as Snappy and Gzip), and predicate pushdown optimization, and has extremely high compression ratios and extremely low disk I / O overhead, which can significantly improve the efficiency of reading, filtering, and aggregation analysis of massive amounts of data.

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0032] In a first aspect, embodiments of this application provide a method for post-processing vehicle simulation calculation data.

[0033] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the vehicle simulation calculation data post-processing method of this application. Figure 1 As shown, the post-processing method for vehicle simulation calculation data includes: S100: Reads the preset system configuration information and scans the directory of vehicle simulation data to be processed. Based on the scan results, it constructs a global atomic task queue and maps each discrete simulation result file to an atomic task. S200: Based on a multi-core parallel computing resource pool, tasks in the global atomic task queue are silently distributed to multiple worker processes, and each worker process concurrently executes the parsing and processing of the corresponding atomic tasks. S300: Each working process completes the dynamic addressing and semantic mapping of the corresponding simulation result file in the topology directory space based on the preset semantic anchor point, and performs structural-level dimensionality reduction and purification of the parsed simulation data channel based on the naming ontology features of the industrial solver. S400: Flatten and reorganize the purified effective time-series data channels output by each working process into a two-dimensional table structure, generate storage paths according to the semantically mapped working condition names, and serialize the reorganized structured data into a preset columnar storage format.

[0034] In this embodiment, addressing the issue of heterogeneous multidimensional matrix data output by the vehicle simulation underlying computing engine being scattered across multiple nested "condition-task" topologies and unable to be processed in batches, a global recursive directory scan is used to uniformly identify all discrete simulation result files. Each file is then encapsulated as an independent atomic task and a globally unified queue is constructed. Finally, these tasks are incorporated into a parallel scheduling system for concurrent processing, completely eliminating the process bottlenecks caused by manual file location, interactive triggering, and graphical interface rendering in traditional post-processing modes. Semantic anchors are used to achieve loosely coupled mapping between logical components and physical data entities, which can adaptively resist file naming variations caused by upstream model version updates or different computing nodes. By performing structural dimensionality reduction and purification at the data extraction source, a large amount of high-frequency non-temporally redundant data is intercepted, avoiding memory loading and disk I / O consumption of invalid data. The final output standardized structured data removes the restrictions on data flow imposed by the underlying computing engine's proprietary format and can be directly connected to modern data science toolchains. Using the method in this embodiment, the post-processing time can be drastically reduced from days / weeks in the traditional manual mode to minutes, peak memory usage can typically be reduced by more than 90%, and disk storage space requirements for output results can be reduced by 50%-80%, while ensuring absolute consistency and accuracy of the post-processing data cleaning pipeline in long-cycle projects.

[0035] Furthermore, in one embodiment, S100 includes the following steps: S101: Read and parse the system configuration file, which includes the source data path, target output path, list of tasks to be processed, and list of component semantic identifiers; S102: Recursively scan all simulation result files and associated metadata files under the source data path according to the preset directory topology hierarchy; S103: Filter the scanned files based on the task list to be processed, and select the target simulation result files that need to be processed; S104: Extract the file identifier, storage path, and associated metadata information of each target simulation result file, and encapsulate them into indivisible atomic task units; S105: Sort all atomic tasks according to preset priority rules to generate a globally unified atomic task queue.

[0036] In this embodiment, the entire system configuration is read and parsed in one go, avoiding disk I / O overhead caused by repeatedly reading the configuration in multiple steps. It also ensures that all working processes share the same set of processing rules, avoiding inconsistencies in data processing standards. The full scan results are filtered through a list of tasks to be processed, supporting flexible specification of certain tasks for processing without the need to parse all files under the source data path. This adapts to project needs of different scales and scenarios where urgent tasks are prioritized. Each target simulation result file is encapsulated as an independent atomic task unit, decoupling the task from the processing logic and facilitating subsequent parallel distribution, status tracking, and anomaly management.

[0037] Furthermore, in one embodiment, step S200 includes the following steps: S201: Create a corresponding number of headless worker processes based on the current number of processor cores in the device to build a parallel computing resource pool; S202: A load balancing strategy is adopted to distribute tasks in the global atomic task queue to idle worker processes in sequence. All worker processes silently execute data parsing tasks in the background, skipping the graphical rendering stage.

[0038] In this embodiment, a corresponding number of headless worker processes are dynamically created according to the device hardware configuration, which maximizes the utilization of the computing power of the multi-core processor and avoids resource idleness. A load balancing strategy is adopted to realize the dynamic distribution of tasks, ensuring that the task load of each worker process is relatively balanced, preventing some processes from being overloaded while others are idle. All worker processes are executed silently in the background in headless mode, skipping all graphical rendering processes and concentrating all computing power on data parsing operations, further improving the processing speed of a single task and the throughput of the overall system.

[0039] Furthermore, in one embodiment, step S200 further includes the following step: S203: Real-time monitoring of the running status of all working processes and the execution progress of each atomic task; S204: If an atomic task fails to execute and the cumulative number of failures has not reached the preset retry threshold, then the atomic task is added back to the tail of the global atomic task queue to wait for re-distribution and execution. S205: If the cumulative number of failures reaches the preset retry threshold, mark the atomic task as a permanently failed task, generate an exception alarm record and write it to the global processing log.

[0040] In this embodiment, by monitoring the working process and the running status of atomic tasks in real time, the overall progress of batch processing tasks can be fully grasped; the introduction of a graded retry mechanism can effectively solve the problem of task failure caused by occasional factors such as network fluctuations and temporary file locking, and improve the fault tolerance of the system; atomic tasks that have accumulated a threshold of failures are marked as permanent failures and an abnormal alarm is generated to avoid invalid retries occupying computing resources, and at the same time facilitate subsequent manual investigation and processing of abnormal files, ensuring the stable operation of large-scale concurrent processing.

[0041] Furthermore, in one embodiment, step S300 includes the following steps: S301: Each working process obtains the file information corresponding to the assigned atomic task, uses the semantic anchor point identifier fragment in the component semantic identifier list to perform fuzzy matching on the corresponding simulation result file with non-deterministic suffix, and establishes a dynamic mapping relationship between the logical processing component and the physical simulation data entity. S302: Parse the metadata configuration file corresponding to the atomic task and reverse map the mechanized task directory code to the real working condition name with physical meaning; S303: Traverse all data channels of the simulation result file obtained by parsing, identify and remove redundant data channels with system-level metadata pre-reservation characters and non-temporal static attribute post-labels, and retain only valid temporal physical quantity data channels with dynamic evolution processes.

[0042] In this embodiment, fuzzy matching is performed using preset semantic anchor identifier fragments, which can automatically adapt to heterogeneous files with non-deterministic suffixes such as timestamps and version numbers output by the computing engine. This eliminates the need to modify processing code due to file naming changes, significantly reducing system maintenance costs. Semantic inversion is achieved by parsing metadata configuration files, reversing the machine-generated task directory code (e.g., task.083) into a physically meaningful real-world operating condition name (e.g., CHTC-TT), making the output data business-readable. When traversing all data channels, the following judgment rules are applied to filter the channels: in This represents the data channel identifier parsed from the underlying layer. This represents a predefined set of solver system-level metadata identifiers. This represents a pre-defined set of non-temporally static physical attribute identifiers, retaining only those in the state of... An effective time-series physical quantity channel enables dimensionality reduction and purification of engineering data at the data extraction source.

[0043] Furthermore, in one embodiment, after S300, the following steps are included: S304: Perform data integrity verification and physical quantity rationality verification on each effective time-series physical quantity data channel after purification; S305: If the verification passes, proceed with the subsequent steps of flattening and reassembling the valid time-series data channels; S306: If the verification fails, mark the corresponding atomic task as an abnormal task, record the abnormal type and corresponding file information, and skip the subsequent processing flow of the task.

[0044] In this embodiment, data integrity checks are performed to check for missing or truncated channel data, and physical quantity rationality checks are performed to verify whether the data conforms to the value range and variation law of the corresponding physical quantity. The dual verification mechanism can effectively eliminate invalid data caused by simulation calculation anomalies or file parsing errors. Atomic tasks that fail the verification are marked and skipped from subsequent processing to avoid invalid data polluting the final output result set and ensure the accuracy and reliability of the output data.

[0045] Furthermore, in one embodiment, step S400 includes the following steps: S401: Summarize all valid timing data channels output by each working process, perform flattening processing, and build a standard two-dimensional table structure; S402: Combine the target output path with the actual working condition name corresponding to each atomic task to automatically generate a hierarchical target storage path; S403: Serialize time-series data with a standard two-dimensional table structure into Parquet or CSV format; S404: Store the serialized data file to the corresponding target storage path and generate a processing log file that includes task execution status, processing time, and exception information.

[0046] In this embodiment, the dispersed effective time-series data channels are aggregated and flattened into a standard two-dimensional table structure, giving the data a unified organizational form for easy viewing and analysis. A hierarchical storage path is generated by combining the target output path and the actual operating condition name, achieving standardized management of output files and facilitating subsequent retrieval and archiving. It supports output in Apache Parquet columnar storage format and Comma-Separated Values ​​(CSV) general text format. The Parquet format offers high compression ratios and efficient column query capabilities, significantly reducing storage costs and data reading overhead. The CSV format boasts excellent cross-platform compatibility and can be directly imported into spreadsheet software and conventional data analysis tools for processing. The generated processing log file fully records key information throughout the entire post-processing process, facilitating subsequent process traceability and problem auditing.

[0047] Furthermore, in another embodiment, referring to Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the vehicle simulation calculation data post-processing method of this application. Figure 2 As shown, the post-processing method for vehicle simulation calculation data includes: S1: Read the configuration and scan the data directory; S2: Construct a global atomic task queue; S3: Initialize the multi-core parallel computing resource pool; check if the global atomic task queue is empty. If the queue is empty, end the entire post-processing process. If the queue is not empty, take the task and enter the parallel computing unit processing process. S4: Dynamically locate target files based on feature prefixes; perform inverse parsing of task configuration to complete the mapping from mechanized task identifiers to business condition names; S5: Read the file's internal key list; iterate through all data channel keys within the file. Once the iteration is complete, execute S6. During the iteration, perform double filtering checks sequentially. First check: Check if the key starts with a system-level metadata reserved character. If so, discard the key and continue iterating to the next key. Otherwise, execute the second check. Second check: Check if the key ends with a non-temporal static attribute tag. If so, discard the key and continue iterating to the next key. Otherwise, retain the valid temporal channel and continue iterating to the next key. S6: Data Dimensionality Reduction and Two-Dimensional Table Restructuring; S7: Serialization output in columnar storage format; after a single task is completed, return to the global atomic task queue for an empty check step, and repeat until all tasks are processed.

[0048] In this embodiment, a system configuration file, including the source data path, target output path, list of tasks to be processed, and list of component semantic identifiers, is read and parsed in one go. All simulation result files and associated metadata files under the source data path are recursively scanned according to a preset directory topology hierarchy. The scan results are filtered based on the list of tasks to be processed, and the identifiers, paths, and metadata information of the target files are extracted and encapsulated into indivisible atomic task units. These units are then sorted according to a preset priority to generate a globally unified atomic task queue. A parallel computing resource pool is constructed by creating a corresponding number of headless worker processes based on the number of processor cores in the current device. All worker processes skip all graphical rendering stages and silently execute data parsing tasks in the background.

[0049] Through a closed-loop architecture of queue empty check - task acquisition - single task full-process processing - return to queue, fully unattended automated batch processing is achieved without manual intervention to trigger any intermediate steps. The parallel computing unit adopts a modular decomposition design, dividing the single task processing flow into six independent stages: location, inversion, reading, filtering, reorganization, and output. Each stage has clear responsibilities and can be optimized independently, improving the system's maintainability and scalability.

[0050] In this embodiment, the feature prefix is ​​a preset semantic anchor identifier fragment. The semantic anchor is used to perform fuzzy matching on simulation result files with non-deterministic suffixes such as timestamps and version numbers to establish a dynamic mapping relationship between logical processing components and physical data entities. By parsing the metadata configuration file corresponding to the task, the machined task directory code (such as task.083) is reverse-mapped to a real working condition name with physical meaning (such as CHTC-TT), so that the output data directly has business semantics.

[0051] When traversing the internal data channels of a file, the following judgment rules are applied to perform double filtering: ;in This represents the data channel identifier parsed from the underlying layer. This represents a predefined set of solver system-level metadata identifiers. This represents a pre-defined set of non-temporally sequential static physical attribute identifiers. The filtering logic, which combines key-by-key traversal with dual conditional checks, can accurately remove all non-temporally redundant data at the data extraction source, while avoiding the accidental deletion of valid temporally sequential physical quantities.

[0052] Furthermore, in this embodiment, after performing the dual filtering judgment in S5, the method further includes: performing data integrity verification and physical quantity rationality verification on the retained valid time-series physical quantity data channels; if the verification passes, then performing the data dimensionality reduction and two-dimensional table reconstruction steps in S6; if the verification fails, then marking the corresponding atomic task as an abnormal task, recording the abnormality type and corresponding file information, skipping the subsequent processing flow of the task and returning to the queue empty judgment step.

[0053] Furthermore, in this embodiment, during the entire process of task acquisition and execution, the running status of all working processes and the execution progress of each atomic task are monitored in real time. If an atomic task is detected to have failed and the cumulative number of failures has not reached the preset retry threshold, the atomic task is added back to the tail of the global atomic task queue to wait for re-distribution and execution. If the cumulative number of failures reaches the preset retry threshold, the atomic task is marked as a permanently failed task, an abnormal alarm record is generated and written to the global processing log.

[0054] In this embodiment, the purified effective time-series data channels are flattened and a standard two-dimensional table structure is constructed. Combined with the preset target output path and the mapped real-world working condition names, a hierarchical target storage path is automatically generated. The time-series data in the standard two-dimensional table structure is serialized into Apache Parquet or Comma-SeparatedValues ​​format and stored in the corresponding path. A global processing log file containing task execution status, processing time, and exception information is generated. Using this method, the post-processing time can be drastically reduced from days / weeks in traditional manual methods to minutes. Peak memory usage can typically be reduced by more than 90%, and the disk storage space requirement for the output results can be reduced by 50%-80%, while ensuring absolute consistency and accuracy of the post-processing data cleaning pipeline in long-term projects.

[0055] Secondly, this application also provides a vehicle simulation calculation data post-processing device, which includes: a task preparation module, which is used to read preset system configuration information and scan the vehicle simulation data directory to be processed, construct a global atomic task queue based on the scanning results, and map each discrete simulation result file to an atomic task; a parallel scheduling module, which is used to silently distribute the tasks in the global atomic task queue to multiple working processes based on a multi-core parallel computing resource pool, and have each working process concurrently execute the parsing and processing of the corresponding atomic task; a parsing and purification module, which is used for each working process to complete the dynamic addressing and semantic mapping of the corresponding simulation result file in the topology directory space based on preset semantic anchors, and to perform structural dimensionality reduction and purification of the parsed simulation data channels based on the naming ontology features of the industrial solver; and an output module, which is used to flatten and reorganize the purified effective time-series data channels output by each working process into a two-dimensional table structure, generate storage paths according to the semantically mapped working condition names, and serialize the reorganized structured data into a preset columnar storage format.

[0056] The functions of each module in the above-mentioned vehicle simulation calculation data post-processing device correspond to the steps in the above-mentioned vehicle simulation calculation data post-processing method embodiment, and their functions and implementation processes will not be described in detail here.

[0057] Thirdly, embodiments of this application provide a vehicle simulation calculation data post-processing device, which can be a personal computer (PC), laptop computer, server, or other device with data processing capabilities.

[0058] Reference Figure 3 , Figure 3 This is a schematic diagram of the hardware structure of the vehicle simulation calculation data post-processing device involved in the embodiments of this application. In the embodiments of this application, the vehicle simulation calculation data post-processing device may include a processor, a memory, a communication interface, and a communication bus.

[0059] Fourthly, embodiments of this application also provide a readable storage medium.

[0060] The present application has a readable storage medium storing a vehicle simulation calculation data post-processing program, wherein when the vehicle simulation calculation data post-processing program is executed by a processor, it implements the steps of the vehicle simulation calculation data post-processing method as described above.

[0061] The method implemented when the vehicle simulation calculation data post-processing program is executed can be referred to in various embodiments of the vehicle simulation calculation data post-processing method of this application, and will not be repeated here.

[0062] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0063] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0064] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0065] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0066] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0067] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0068] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for post-processing vehicle simulation calculation data, characterized in that, It includes the following steps: Read the preset system configuration information and scan the vehicle simulation data directory to be processed. Based on the scan results, construct a global atomic task queue and map each discrete simulation result file to an atomic task. Based on a multi-core parallel computing resource pool, tasks in the global atomic task queue are silently distributed to multiple worker processes, and each worker process concurrently executes the parsing and processing of the corresponding atomic tasks. Each process performs dynamic addressing and semantic mapping of the corresponding simulation result files within the topology directory space based on preset semantic anchor points, and performs structural-level dimensionality reduction and purification of the parsed simulation data channels based on the naming ontology features of the industrial solver. The purified effective time-series data channels output by each working process are flattened and reorganized into a two-dimensional table structure. Storage paths are generated according to the semantically mapped working condition names, and the reorganized structured data is serialized into a preset columnar storage format.

2. The vehicle simulation calculation data post-processing method as described in claim 1, characterized in that, The process of reading preset system configuration information and scanning the catalog of vehicle simulation data to be processed, and constructing a global atomic task queue based on the scan results, includes: Read and parse the system configuration file, which includes the source data path, target output path, list of tasks to be processed, and list of component semantic identifiers; Recursively scan all simulation result files and associated metadata files under the source data path according to the preset directory topology hierarchy; Based on the list of tasks to be processed, the scanned files are filtered to select the target simulation result files that need to be processed. Extract the file identifier, storage path, and associated metadata information of each target simulation result file, and encapsulate them into indivisible atomic task units; All atomic tasks are sorted according to a preset priority rule to generate a globally unified atomic task queue.

3. The vehicle simulation calculation data post-processing method as described in claim 1, characterized in that, The multi-core parallel computing resource pool silently distributes tasks in the global atomic task queue to multiple worker processes, which then concurrently execute the parsing and processing of the corresponding atomic tasks, including: Create a corresponding number of headless worker processes based on the current number of processor cores in the device, and build a parallel computing resource pool. A load balancing strategy is adopted to distribute tasks in the global atomic task queue to idle worker processes in sequence. All worker processes silently execute data parsing tasks in the background, skipping the graphical rendering stage.

4. The vehicle simulation calculation data post-processing method as described in claim 3, characterized in that, The step of silently distributing tasks in the global atomic task queue to multiple worker processes, with each worker process concurrently executing the parsing and processing of the corresponding atomic task, further includes: Real-time monitoring of the running status of all working processes and the execution progress of each atomic task; If an atomic task fails to execute and the cumulative number of failures has not reached the preset retry threshold, the atomic task will be added back to the tail of the global atomic task queue to wait for re-distribution and execution. If the cumulative number of failures reaches the preset retry threshold, the atomic task is marked as a permanently failed task, an exception alarm record is generated and written to the global processing log.

5. The vehicle simulation calculation data post-processing method as described in claim 1, characterized in that, Each working process, based on semantic anchors in a preset component semantic identifier list, performs dynamic addressing and semantic mapping of corresponding simulation result files within the topology directory space. It also performs structural-level dimensionality reduction and purification of the parsed simulation data channels based on the industrial solver's naming ontology features, including: Each working process obtains the file information corresponding to the assigned atomic task, uses the semantic anchor point identifier fragment in the component semantic identifier list to perform fuzzy matching on the corresponding simulation result file with non-deterministic suffix, and establishes a dynamic mapping relationship between the logical processing component and the physical simulation data entity. Parse the metadata configuration file corresponding to the atomic task and reverse map the mechanized task directory code to the real working condition name with physical meaning; The simulation result file obtained by parsing is traversed through all data channels, and redundant data channels with system-level metadata pre-reservation characters and non-temporal static attribute post-labels are identified and eliminated, retaining only the effective temporal physical quantity data channels with dynamic evolution processes.

6. The vehicle simulation calculation data post-processing method as described in claim 1, characterized in that, The process of flattening and recombining the purified effective time-series data channels output by each working process into a standard two-dimensional table structure, generating a hierarchical storage path according to the semantically mapped working condition name and the preset target output path, and serializing the recombined structured data into a preset columnar storage format includes: Summarize all valid timing data channels output by each working process, perform flattening processing, and construct a standard two-dimensional table structure; By combining the target output path with the actual working condition name corresponding to each atomic task, a hierarchical target storage path is automatically generated. Serialize time-series data with a standard two-dimensional table structure into Parquet or CSV format; The serialized data file is stored in the corresponding target storage path, and a processing log file including task execution status, processing time and exception information is generated.

7. The vehicle simulation calculation data post-processing method as described in claim 1, characterized in that, After performing structural-level dimensionality reduction and purification on the parsed simulation data channels based on the naming ontology features of the industrial solver, the following is also included: Data integrity verification and physical quantity rationality verification are performed on each effective time-series physical quantity data channel after purification; If the verification passes, proceed with the subsequent steps of flattening and reassembling the valid time-series data channels. If the verification fails, the corresponding atomic task is marked as an abnormal task, the abnormality type and corresponding file information are recorded, and the subsequent processing flow of the task is skipped.

8. A post-processing device for vehicle simulation calculation data, characterized in that, The vehicle simulation calculation data post-processing device includes: The task preparation module is used to read the preset system configuration information and scan the directory of vehicle simulation data to be processed. Based on the scan results, a global atomic task queue is constructed, and each discrete simulation result file is mapped to an atomic task. The parallel scheduling module is used to silently distribute tasks in the global atomic task queue to multiple worker processes based on a multi-core parallel computing resource pool, and each worker process concurrently executes the parsing and processing of the corresponding atomic task. The parsing and purification module is used by each working process to perform dynamic addressing and semantic mapping of the corresponding simulation result files in the topology directory space based on preset semantic anchors, and to perform structural dimensionality reduction and purification of the parsed simulation data channels based on the naming ontology features of the industrial solver. The output module is used to flatten and reorganize the purified effective time-series data channels output by each working process into a two-dimensional table structure, generate storage paths according to the semantically mapped working condition names, and serialize the reorganized structured data into a preset columnar storage format.

9. A post-processing device for vehicle simulation calculation data, characterized in that, The vehicle simulation calculation data post-processing device includes a processor, a memory, and a vehicle simulation calculation data post-processing program stored in the memory and executable by the processor, wherein when the vehicle simulation calculation data post-processing program is executed by the processor, it implements the steps of the vehicle simulation calculation data post-processing method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a vehicle simulation calculation data post-processing program, wherein when the vehicle simulation calculation data post-processing program is executed by a processor, it implements the steps of the vehicle simulation calculation data post-processing method as described in any one of claims 1 to 7.

Images