CN-121980805-A - Processing method of virtual calibration model input boundary of post-processing system
Abstract
The invention discloses a processing method of an input boundary of a virtual calibration model of an aftertreatment system, relates to the technical field of simulation and calibration of the aftertreatment system of an engine, and solves the technical problems of insufficient test data integrity, poor consistency and prediction precision and application range of a model containing abnormal values in the existing virtual calibration practice of the aftertreatment system of the engine. The processing method comprises the steps of S1, performing formal test according to a preset target test working condition, S2, performing quality check on original test data to judge whether the quality requirement is met, S3, performing data preprocessing on the original test data meeting the quality requirement to generate preprocessed data, returning to S1 to perform formal test again when the original test data does not meet the quality requirement, and S4, introducing the preprocessed data into a post-processing virtual calibration model to perform preliminary simulation operation, and verifying the validity of the preprocessed data according to a simulation operation result. The invention improves the integrity, consistency and physical rationality of the input data.
Inventors
- LIU XU
- LI ZHIHUA
- HE GUANZHANG
- LIN TIEJIAN
- Huang Guicong
- WEI PEILI
- CAI RUIFANG
Assignees
- 广西玉柴机器股份有限公司
Dates
- Publication Date
- 20260505
- Application Date
- 20260130
Claims (6)
- 1. A processing method of virtual calibration model input boundary of post-processing system is characterized in that S1, setting target test working condition, and performing formal test according to the target test working condition; s2, acquiring original test data after a formal test, and performing quality inspection on the original test data to judge whether the original test data meets quality requirements; s3, carrying out data preprocessing on the original test data meeting the quality requirement to generate preprocessed data, and returning to S1 to carry out a formal test again when the original test data does not meet the quality requirement; And S4, importing the preprocessing data into a post-processing virtual calibration model to perform preliminary simulation operation and outputting a simulation operation result.
- 2. A method for processing virtual calibration model input boundaries in a post-processing system according to claim 1, wherein in S3, the method for preprocessing the raw test data satisfying the quality requirement comprises the steps of, S31, formulating a unified variable naming mapping rule, and unifying the names of the original test data by applying the variable naming mapping rule; s32, acquiring variable characteristics and engineering knowledge, setting reasonable thresholds according to the variable characteristics and the engineering knowledge, and performing amplitude limiting or eliminating treatment on the original test data with unified names according to the reasonable thresholds; s33, performing variable secondary calculation processing on the original test data subjected to the amplitude limiting or eliminating processing to generate variables meeting the input requirements of the virtual calibration model; and S34, according to the output requirement of the virtual calibration model, aligning and formatting the variables meeting the input requirement of the virtual calibration model to generate a model input data set, and taking the model input data set as preprocessing data.
- 3. The method for processing virtual calibration model input boundaries in a post-processing system according to claim 2, wherein in S33, the method for performing variable secondary calculation processing on the raw test data after clipping or eliminating processing is, S331, converting units in the original test data subjected to amplitude limiting or eliminating processing into measurement units required by a model; S332, calibrating or compensating the original test data after the conversion unit; S333, calculating the input quantity which cannot be directly acquired by the virtual calibration model according to the calibrated or compensated original data, and taking the input quantity which cannot be directly acquired by the calibrated or compensated original data and the virtual calibration model as a variable which meets the input requirement of the virtual calibration model.
- 4. A method for processing virtual calibration model input boundaries for a post-processing system as defined in claim 1, wherein in S2, the method for quality checking the raw test data comprises, Checking whether all the measurement and control system variables, the sensor data, the ECU data and the emission equipment data in the original test data are recorded and are free from missing, and judging that the original test data do not meet the quality requirement when the measurement and control system variables, the sensor data, the ECU data and the emission equipment data are not recorded or are missing; Checking whether the recording frequency of the original test data meets the frequency of a preset model simulation requirement, and judging that the original test data does not meet the quality requirement when the recording frequency of the original test data does not meet the frequency of the model simulation requirement; Checking whether the original test data has the phenomena of data interruption, abnormal jump or long-time stagnation, and judging that the original test data does not meet the quality requirement when the phenomena of data interruption, abnormal jump or long-time stagnation exist.
- 5. The method for processing virtual calibration model input boundaries of a post-processing system according to claim 1, wherein a pre-processing test is performed according to a preset test specification to bring the post-processing system to an initial state at the time of the formal test before S1 is performed.
- 6. A method of processing virtual calibration model input boundaries for an aftertreatment system according to claim 5, wherein temperature and pressure sensors are arranged on the aftertreatment system according to preset test requirements prior to performing the pretreatment test.
Description
Processing method of virtual calibration model input boundary of post-processing system Technical Field The invention relates to the technical field of simulation and calibration of an engine aftertreatment system, in particular to a processing method of an input boundary of a virtual calibration model of the aftertreatment system. Background The modern engine aftertreatment system converts harmful components in the tail gas into harmless substances through technologies such as catalytic conversion, particle filtration and the like, and is a core component meeting strict emission regulations. Compared with the traditional entity rack calibration, the post-processing system calibration is carried out in the virtual simulation platform, so that the development efficiency can be remarkably improved, the test cost can be reduced, more comprehensive system performance parameters can be obtained, and important support is provided for optimizing design and control strategies. However, in the current virtual calibration practice, the test data collected from the physical bench often has the problems of insufficient integrity, poor consistency, abnormal value inclusion and the like, so that the input boundary of the model is inaccurate or incomplete, and the prediction precision and the application range of the virtual model are further restricted. Disclosure of Invention The invention aims to solve the technical problems of insufficient test data integrity, poor consistency and prediction precision and application range of a model containing abnormal value restriction in the existing virtual calibration practice of an engine aftertreatment system. The invention relates to a processing method of an input boundary of a virtual calibration model of a post-processing system, which comprises the following steps of S1, setting a target test working condition, and performing a formal test according to the target test working condition; s2, acquiring original test data after a formal test, and performing quality inspection on the original test data to judge whether the original test data meets quality requirements; s3, carrying out data preprocessing on the original test data meeting the quality requirement to generate preprocessed data, and returning to S1 to carry out a formal test again when the original test data does not meet the quality requirement; And S4, importing the preprocessing data into a post-processing virtual calibration model to perform preliminary simulation operation and outputting a simulation operation result. Further improved, in S3, the method for preprocessing the data of the original test data meeting the quality requirement is that, S31, formulating a unified variable naming mapping rule, and unifying the names of the original test data by applying the variable naming mapping rule; s32, acquiring variable characteristics and engineering knowledge, setting reasonable thresholds according to the variable characteristics and the engineering knowledge, and performing amplitude limiting or eliminating treatment on the original test data with unified names according to the reasonable thresholds; s33, performing variable secondary calculation processing on the original test data subjected to the amplitude limiting or eliminating processing to generate variables meeting the input requirements of the virtual calibration model; and S34, according to the output requirement of the virtual calibration model, aligning and formatting the variables meeting the input requirement of the virtual calibration model to generate a model input data set, and taking the model input data set as preprocessing data. Further, in S33, the method for performing variable secondary calculation processing on the original test data after the clipping or eliminating processing is that, S331, converting units in the original test data subjected to amplitude limiting or eliminating processing into measurement units required by a model; S332, calibrating or compensating the original test data after the conversion unit; S333, calculating the input quantity which cannot be directly acquired by the virtual calibration model according to the calibrated or compensated original data, and taking the input quantity which cannot be directly acquired by the calibrated or compensated original data and the virtual calibration model as a variable which meets the input requirement of the virtual calibration model. Further, in S2, the method for quality inspection of the raw test data includes, Checking whether all the measurement and control system variables, the sensor data, the ECU data and the emission equipment data in the original test data are recorded and are free from missing, and judging that the original test data do not meet the quality requirement when the measurement and control system variables, the sensor data, the ECU data and the emission equipment data are not recorded or are missing; Checking whether the recording frequency of the