CN118094163B - Digital physical model management system and method based on multi-feature association - Google Patents
Digital physical model management system and method based on multi-feature association Download PDFInfo
- Publication number
- CN118094163B CN118094163B CN202410494673.0A CN202410494673A CN118094163B CN 118094163 B CN118094163 B CN 118094163B CN 202410494673 A CN202410494673 A CN 202410494673A CN 118094163 B CN118094163 B CN 118094163B
- Authority
- CN
- China
- Prior art keywords
- feature
- data
- physical model
- model
- association
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 230000008859 change Effects 0.000 claims abstract description 47
- 238000007726 management method Methods 0.000 claims abstract description 26
- 238000011156 evaluation Methods 0.000 claims abstract description 22
- 238000002474 experimental method Methods 0.000 claims abstract description 22
- 238000000605 extraction Methods 0.000 claims abstract description 17
- 238000010276 construction Methods 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims description 17
- 230000008569 process Effects 0.000 claims description 16
- 238000012360 testing method Methods 0.000 claims description 13
- JXASPPWQHFOWPL-UHFFFAOYSA-N Tamarixin Natural products C1=C(O)C(OC)=CC=C1C1=C(OC2C(C(O)C(O)C(CO)O2)O)C(=O)C2=C(O)C=C(O)C=C2O1 JXASPPWQHFOWPL-UHFFFAOYSA-N 0.000 claims description 6
- 238000011835 investigation Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 230000001172 regenerating effect Effects 0.000 claims description 4
- 241000764238 Isis Species 0.000 claims description 3
- 239000012141 concentrate Substances 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- WLRMANUAADYWEA-NWASOUNVSA-N (S)-timolol maleate Chemical compound OC(=O)\C=C/C(O)=O.CC(C)(C)NC[C@H](O)COC1=NSN=C1N1CCOCC1 WLRMANUAADYWEA-NWASOUNVSA-N 0.000 claims description 2
- 239000000463 material Substances 0.000 description 15
- 238000004364 calculation method Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012067 mathematical method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to the technical field of model construction, in particular to a digital physical model management system and method based on multi-feature association, wherein the management method comprises the following steps: acquiring output data of each physical model of the history, performing error data checking to obtain an effective data set, and performing feature extraction to obtain a feature set; analyzing the change conditions of all the features in the feature sets of different time periods, and establishing a dynamic feature association model; in the vehicle performance evaluation experiment, generating a real-time feature set; selecting data from the two physical models according to the quantity ratio of the same features in the two real-time feature sets and the predicted association degree in the dynamic feature association model, and constructing a digital physical model; and according to the actual association degree between each physical model feature set, analyzing the difference between the actual association degree and the predicted association degree change condition, and adjusting the dynamic feature association model and the digital physical model.
Description
Technical Field
The invention relates to the technical field of model construction, in particular to a digital physical model management system and method based on multi-feature association.
Background
The digital physical model refers to a process of modeling, simulating and analyzing a physical system by using a computer technology and a mathematical method; the method is an extension and development of the traditional physical model in the digital era, has higher flexibility, expandability and visualization capability, and can better simulate a complex physical system and conduct intensive research and analysis.
In the conventional construction of the digital physical model, the situation of mismatch between data and the model exists, for example, the model assumption is inconsistent with the actual data distribution, the model parameters are inconsistent with the data characteristics, and the like, so that the adaptability and the prediction capability of the model are influenced; if the digital physical model is built only according to the existing data, the digital physical model is influenced by other factors, so that the obtained result is greatly different from the expected result when the originally built model is subsequently put in new data for output.
Disclosure of Invention
The invention aims to provide a digital physical model management system and method based on multi-feature association, which are used for solving the problems in the background technology.
In order to solve the technical problems, the invention provides the following technical scheme: a digital physical model management method based on multi-feature association, the management method comprising the steps of:
step S100: acquiring output data of each physical model in any vehicle performance evaluation experiment process of history, performing error data investigation on the output data to obtain an effective data set corresponding to the physical model, and performing feature extraction on the effective data set to obtain a feature set; among the physical models that may be involved in the vehicle performance evaluation experiment are an engine performance evaluation model, a collision detection model, a vehicle speed test model, and the like;
Step S200: acquiring a feature set of each physical model, and analyzing the change condition of each feature in the feature set of different time periods; comparing the association degree between any two feature sets, establishing a dynamic feature association model according to the association degree change condition between the feature sets in different time periods, and predicting the association degree change between the feature sets;
Step S300: in the vehicle performance evaluation experiment, input data and output data of each physical model are obtained in real time, and a real-time feature set is generated; acquiring the feature quantity of the same features in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity proportion of the same features in the two real-time feature sets and the predicted association degree in the dynamic feature association model, and constructing a digital physical model for simulating the vehicle performance change process; in the vehicle performance evaluation experiment, different physical models focus on performance detection in different directions, so that data in different physical models are integrated by constructing a digital physical model, and the vehicle performance change is evaluated in multiple dimensions.
Step S400: obtaining actual association degree between feature sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic feature association model, and adjusting the dynamic feature association model; and adjusting the digital physical model according to the actual association degree between the feature sets.
Further, the step S100 includes the following steps:
Step S101: setting the first The output data set of the physical model isSequencing the output data sets according to the output time of the output data sets in the physical model to obtain the output data setsIs a profile of (2); checking the distribution of the data of the distribution by using the box diagram, if the data group is outputIf the numerical value of a plurality of data in the box body exceeds the upper boundary and the lower boundary of the box body, the plurality of data are used as error data, and the output data group is deletedError data in (a) is obtainedEffective data group of individual physical model；
Step S102: acquisition of the firstInput data of a physical model, wherein one physical model has three parts of input, processing and output, so that the input data can be directly obtained from the physical model; extracting data sources from input data as the firstOne feature of the physical model is derived from the active data setExtracting several features from the distribution map to generate the firstFeature set of individual physical models。
Further, the step S200 includes the following steps:
Step S201: acquisition of Feature sets of any two physical models at any moment, and setting one physical model as a first physical modelA physical model, the feature set isAnother physical model is the firstA physical model, the feature set is; Randomly selecting feature setsOne feature of (2)Acquiring featuresIn the first placeCorresponding data in a physical modelWherein, the dataIn the first placeThe data volume duty ratio in each physical model is；
Step S202: acquiring dataIn the first placeData sources in physical model, obtaining feature setData sources of the corresponding data for each feature; according to the formula:
Wherein, Is a feature setIs used for the number of features of (a),Judging data as a judging functionAnd feature setMiddle featureCorresponding dataWhether the data sources of (a) are the same, if so,Otherwise, the device can be used to determine whether the current,; Calculated to obtainTime of day featureAnd feature setIs related to the degree of association of (2)；
The similarity between the features is influenced by two factors, namely a data source and a data duty ratio, and the importance of the features in the physical model is reflected on the basis of the data duty ratio of the corresponding data of the features in the physical model, so that the relevance between the features can be influenced, and the same data source is directly connected; therefore, the accuracy can be ensured by performing association degree calculation through the data source and the data duty ratio, and the data acquisition method is simpler;
Step S203: acquiring feature sets Data and data sources corresponding to all features in the system are provided with feature setsTogether withThe individual features are according to the formula:
Calculated to obtain Time feature setAnd feature setIs related to the degree of association of (2)；
Step S204: every other unit timeReacquiring the firstFeature set of individual physical modelsAnd (d)Feature set of individual physical modelsCalculated to obtainTime feature setAnd feature setThe degree of association is; According to the acquisition time sequence of the feature set, for the firstPhysical model number oneSorting the association degree among the physical models; is provided to obtain togetherDegree of association before selectionThe association degree is used as a training set, and the restThe correlation degrees are used as test sets, and a dynamic characteristic correlation model is generated;
The degree of association between features changes because factors affecting the physical model change; when the influencing factors change, the input and output data of the model are influenced by the influencing factors, and the data sources and the duty ratio of different data in the model are changed; because the feature extraction is performed according to the data source of the input data, the data source can be changed due to the change of influencing factors; so as long as the feature sets under different time are obtained, when the feature sets are changed, the association degree among the features is changed;
step S205: will be trained to concentrate The degree of association is put into a dynamic feature association model for training, and when the degree of association is output from the dynamic feature association modelWhen the number is equal to the correlation in the test set, counting and outputting the number of the correlation equal to the correlation in the test setSetting an accuracy thresholdIf (if)Selecting a plurality of relevancy from the test set, putting the relevancy into the training set, and retraining the dynamic characteristic relevancy model until the relevancy meets the requirement。
Further, the step S300 includes the following steps:
Step S301: acquisition of the first in real time Input data and output data in each physical model are set inExtracting a plurality of characteristics at any time to generate the firstReal-time feature set of individual physical models；
Step S302: acquisition of the firstThe physical model isReal-time feature set of momentsStatistics of real-time feature setsAnd real-time feature setThe same feature quantity is that; If it isOr alternativelyWherein, the method comprises the steps of, wherein,Is the firstThe physical model isThe number of features of the time of day,Is the firstThe physical model isThe feature quantity of the moment is fromAcquiring data corresponding to the same characteristics from the physical models, wherein the data quantity corresponding to the same characteristics is as follows; Randomly extracting the data quantity from the data corresponding to the same characteristics asData of (1), and (b)Integrating data corresponding to the same characteristics in the physical models; if it isOr alternativelyThen from the firstThe total data quantity selected by the physical model isWherein, the method comprises the steps of, wherein,Is of the same character in the firstThe corresponding data amount in the physical model; extracting the first data except the data corresponding to the same featuresThe data quantity corresponding to other characteristics in the physical model is as follows；
Setting how much data amount is extracted from one physical model according to the feature quantity of the same feature between the two physical models by taking the data in the one physical model as a reference; if the number of the same features is large, the association degree of the two physical models is higher to a certain extent, and the data corresponding to the same features are extracted; if the number of the same features is small, the fact that the complete and accurate physical model can be generated only by extracting the data corresponding to the other features in the physical model is indicated, and the data corresponding to the other features is extracted;
Step S303: will be Time of day (time)Data corresponding to the same feature in the physical model and the firstData extracted from the physical models are integrated, and thenThe same feature at the moment is put into a dynamic feature association model to obtain a time of arbitrary unit timeThe degree of correlation after that, from the firstPhysical model number oneAnd extracting data from the physical models for integration to obtain the digital physical model.
Further, the step S400 includes the following steps:
step S401: when a unit time passes Respectively obtain the firstPhysical model number oneOutput data of the physical model is obtainedActual feature set of individual physical modelsAnd (d)Actual feature set of individual physical modelsCalculating to obtain the firstPhysical model number oneThe physical model isActual degree of relevance of time of day；
Step S402: set in dynamic feature correlation model, the firstPhysical model number oneThe physical model isThe degree of correlation of the time isIf (if)Wherein, the method comprises the steps of, wherein,Is the error threshold value, thenActual degree of relevance of time of dayTraining the dynamic feature association model as a training set in the dynamic feature association model;
Step S403: if it is Continue from the firstThe data quantity extracted from each physical model isIf (if)Then at the firstThe data quantity extracted from each physical model isThe number of randomly removed data in the data of (a) isIs used for regenerating the digital physical model.
In order to better realize the method, a digital physical model management system is also provided, wherein the management system comprises a feature extraction module, a relevance prediction module, a model construction module and an adjustment module;
The feature extraction module is used for acquiring output data of each physical model in the history vehicle performance evaluation experiment process at any time, performing error data investigation on the output data to obtain an effective data set corresponding to the physical model, and performing feature extraction on the effective data set to obtain a feature set;
The association degree prediction module is used for acquiring a feature set of each physical model and analyzing the change condition of each feature in the feature set of different time periods; comparing the association degree between any two feature sets, establishing a dynamic feature association model according to the association degree change condition between the feature sets in different time periods, and predicting the association degree change between the feature sets;
The model construction module is used for acquiring input data and output data of each physical model in real time in a vehicle performance evaluation experiment to generate a real-time feature set; acquiring the feature quantity of the same features in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity proportion of the same features in the two real-time feature sets and the predicted association degree in the dynamic feature association model, constructing a digital physical model, putting the data into the digital physical model to be converted into a digital form, and carrying out simulation analysis on the vehicle performance;
The adjustment module is used for obtaining the actual association degree between the feature sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic feature association model, and adjusting the dynamic feature association model; and adjusting the digital physical model according to the actual association degree between the feature sets.
Further, the feature extraction module comprises a data checking unit and a feature set generating unit;
The data checking unit is used for acquiring output data of each physical model in the history vehicle performance evaluation experiment process at any time, and performing error data checking on the output data to obtain an effective data set corresponding to the physical model; the feature set generating unit is used for extracting features of the effective data set to obtain a feature set.
Further, the association degree prediction module comprises an association degree comparison unit and a prediction unit;
The association degree comparison unit is used for acquiring the feature set of each physical model, analyzing the change condition of each feature in the feature set of different time periods and comparing the association degree between any two feature sets; the prediction unit is used for establishing a dynamic feature association model according to the association degree change condition among feature sets in different time periods and predicting the association degree change among the feature sets.
Further, the model building module comprises a real-time data acquisition unit and a model building unit;
The real-time data acquisition unit is used for acquiring input data and output data of each physical model in real time in a vehicle performance evaluation experiment to generate a real-time feature set; the model building unit is used for obtaining the feature quantity of the same feature in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity ratio of the same feature in the two real-time feature sets and the predicted association degree in the dynamic feature association model, and building a digital physical model.
Further, the adjustment module comprises a characteristic association model adjustment unit and a physical model adjustment unit;
The characteristic association model adjusting unit is used for obtaining the actual association degree between the characteristic sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic characteristic association model and adjusting the dynamic characteristic association model; the physical model adjusting unit is used for adjusting the digital physical model according to the actual association degree among the feature sets.
Compared with the prior art, the invention has the following beneficial effects: (1) According to the invention, the digital physical model is adjusted by analyzing the association degree change among all the features among the physical models, so that the finally obtained digital physical model can be more accurate; (2) According to the invention, a dynamic feature association model is established through the association degree between the features extracted by the historical data, the subsequent feature association degree is predicted, and then the model is retrained according to the acquisition of the actual association degree, so that the association degree prediction can be more accurate, the finally generated digital physical model can be gradually separated from manual adjustment, and an accurate digital physical model is automatically generated; (3) According to the method for extracting the characteristics of the data, the information source and the information duty ratio of the data are used as the characteristics, and in the subsequent association degree change calculation, whether the association degree is changed can be judged only by acquiring whether the information source and the duty ratio are changed, so that the judgment is simpler and certain accuracy can be ensured.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
FIG. 1 is a schematic diagram of the steps of a digital physical model management method based on multi-feature correlation;
FIG. 2 is a schematic diagram of a digital physical model management system based on multi-feature correlation.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1 and 2, the present invention provides the following technical solutions: a digital physical model management method based on multi-feature association, the management method comprising the steps of:
Step S100: acquiring output data of each physical model in any vehicle performance evaluation experiment process of history, performing error data investigation on the output data to obtain an effective data set corresponding to the physical model, and performing feature extraction on the effective data set to obtain a feature set;
Wherein, the step S100 includes the following steps:
Step S101: setting the first The output data set of the physical model isSequencing the output data sets according to the output time of the output data sets in the physical model to obtain the output data setsIs a profile of (2); checking the distribution of the data of the distribution by using the box diagram, if the data group is outputIf the numerical value of a plurality of data in the box body exceeds the upper boundary and the lower boundary of the box body, the plurality of data are used as error data, and the output data group is deletedError data in (a) is obtainedEffective data group of individual physical model；
Step S102: acquisition of the firstInput data of a physical model, wherein one physical model has three parts of input, processing and output, so that the input data can be directly obtained from the physical model; extracting data sources from input data as the firstOne feature of the physical model is derived from the active data setExtracting several features from the distribution map to generate the firstFeature set of individual physical models。
Step S200: acquiring a feature set of each physical model, and analyzing the change condition of each feature in the feature set of different time periods; comparing the association degree between any two feature sets, establishing a dynamic feature association model according to the association degree change condition between the feature sets in different time periods, and predicting the association degree change between the feature sets;
wherein, the step S200 includes the following steps:
Step S201: acquisition of Feature sets of any two physical models at any moment, and setting one physical model as a first physical modelA physical model, the feature set isAnother physical model is the firstA physical model, the feature set is; Randomly selecting feature setsOne feature of (2)Acquiring featuresIn the first placeCorresponding data in a physical modelWherein, the dataIn the first placeThe data volume duty ratio in each physical model is；
Step S202: acquiring dataIn the first placeData sources in physical model, obtaining feature setData sources of the corresponding data for each feature; according to the formula:
Wherein, Is a feature setIs used for the number of features of (a),Judging data as a judging functionAnd feature setMiddle featureCorresponding dataWhether the data sources of (a) are the same, if so,Otherwise, the device can be used to determine whether the current,; Calculated to obtainTime of day featureAnd feature setIs related to the degree of association of (2)；
Step S203: acquiring feature setsData and data sources corresponding to all features in the system are provided with feature setsTogether withThe individual features are according to the formula:
Calculated to obtain Time feature setAnd feature setIs related to the degree of association of (2)；
Step S204: every other unit timeReacquiring the firstFeature set of individual physical modelsAnd (d)Feature set of individual physical modelsCalculated to obtainTime feature setAnd feature setThe degree of association is; According to the acquisition time sequence of the feature set, for the firstPhysical model number oneSorting the association degree among the physical models; is provided to obtain togetherDegree of association before selectionThe association degree is used as a training set, and the restThe correlation degrees are used as test sets, and a dynamic characteristic correlation model is generated;
step S205: will be trained to concentrate The degree of association is put into a dynamic feature association model for training, and when the degree of association is output from the dynamic feature association modelWhen the number is equal to the correlation in the test set, counting and outputting the number of the correlation equal to the correlation in the test setSetting an accuracy thresholdIf (if)Selecting a plurality of relevancy from the test set, putting the relevancy into the training set, and retraining the dynamic characteristic relevancy model until the relevancy meets the requirement。
Step S300: in the vehicle performance evaluation experiment, input data and output data of each physical model are obtained in real time, and a real-time feature set is generated; acquiring the feature quantity of the same features in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity proportion of the same features in the two real-time feature sets and the predicted association degree in the dynamic feature association model, and constructing a digital physical model for simulating the vehicle performance change process;
wherein, the step S300 includes the following steps:
Step S301: acquisition of the first in real time Input data and output data in each physical model are set inExtracting a plurality of characteristics at any time to generate the firstReal-time feature set of individual physical models；
Step S302: acquisition of the firstThe physical model isReal-time feature set of momentsStatistics of real-time feature setsAnd real-time feature setThe same feature quantity is that; If it isOr alternativelyWherein, the method comprises the steps of, wherein,Is the firstThe physical model isThe number of features of the time of day,Is the firstThe physical model isThe feature quantity of the moment is fromAcquiring data corresponding to the same characteristics from the physical models, wherein the data quantity corresponding to the same characteristics is as follows; Randomly extracting the data quantity from the data corresponding to the same characteristics asData of (1), and (b)Integrating data corresponding to the same characteristics in the physical models; if it isOr alternativelyThen from the firstThe total data quantity selected by the physical model isWherein, the method comprises the steps of, wherein,Is of the same character in the firstThe corresponding data amount in the physical model; extracting the first data except the data corresponding to the same featuresThe data quantity corresponding to other characteristics in the physical model is as follows；
Step S303: will beTime of day (time)Data corresponding to the same feature in the physical model and the firstData extracted from the physical models are integrated, and thenThe same feature at the moment is put into a dynamic feature association model to obtain a time of arbitrary unit timeThe degree of correlation after that, from the firstPhysical model number oneAnd extracting data from the physical models for integration to obtain the digital physical model.
Step S400: obtaining actual association degree between feature sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic feature association model, and adjusting the dynamic feature association model; adjusting the digital physical model according to the actual association degree between the feature sets;
Wherein, the step S400 includes the following steps:
step S401: when a unit time passes Respectively obtain the firstPhysical model number oneOutput data of the physical model is obtainedActual feature set of individual physical modelsAnd (d)Actual feature set of individual physical modelsCalculating to obtain the firstPhysical model number oneThe physical model isActual degree of relevance of time of day；
Step S402: set in dynamic feature correlation model, the firstPhysical model number oneThe physical model isThe degree of correlation of the time isIf (if)Wherein, the method comprises the steps of, wherein,Is the error threshold value, thenActual degree of relevance of time of dayTraining the dynamic feature association model as a training set in the dynamic feature association model;
Step S403: if it is Continue from the firstThe data quantity extracted from each physical model isIf (if)Then at the firstThe data quantity extracted from each physical model isThe number of randomly removed data in the data of (a) isIs used for regenerating the digital physical model.
A digital physical model management system based on multi-feature association comprises a feature extraction module, an association degree prediction module, a model construction module and an adjustment module;
The feature extraction module is used for acquiring output data of each physical model in the history vehicle performance evaluation experiment process at any time, performing error data investigation on the output data to obtain an effective data set corresponding to the physical model, and performing feature extraction on the effective data set to obtain a feature set;
The association degree prediction module is used for acquiring a feature set of each physical model and analyzing the change condition of each feature in the feature set of different time periods; comparing the association degree between any two feature sets, establishing a dynamic feature association model according to the association degree change condition between the feature sets in different time periods, and predicting the association degree change between the feature sets;
The model construction module is used for acquiring input data and output data of each physical model in real time in a vehicle performance evaluation experiment to generate a real-time feature set; acquiring the feature quantity of the same features in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity proportion of the same features in the two real-time feature sets and the predicted association degree in the dynamic feature association model, constructing a digital physical model, putting the data into the digital physical model to be converted into a digital form, and carrying out simulation analysis on the vehicle performance;
The adjustment module is used for obtaining the actual association degree between the feature sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic feature association model, and adjusting the dynamic feature association model; and adjusting the digital physical model according to the actual association degree between the feature sets.
The feature extraction module comprises a data checking unit and a feature set generating unit;
The data checking unit is used for acquiring output data of each physical model in the history vehicle performance evaluation experiment process at any time, and performing error data checking on the output data to obtain an effective data set corresponding to the physical model; the feature set generating unit is used for extracting features of the effective data set to obtain a feature set.
The relevance prediction module comprises a relevance comparison unit and a prediction unit;
The association degree comparison unit is used for acquiring the feature set of each physical model, analyzing the change condition of each feature in the feature set of different time periods and comparing the association degree between any two feature sets; the prediction unit is used for establishing a dynamic feature association model according to the association degree change condition among feature sets in different time periods and predicting the association degree change among the feature sets.
The model building module comprises a real-time data acquisition unit and a model building unit;
The real-time data acquisition unit is used for acquiring input data and output data of each physical model in real time in a vehicle performance evaluation experiment to generate a real-time feature set; the model building unit is used for obtaining the feature quantity of the same feature in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity ratio of the same feature in the two real-time feature sets and the predicted association degree in the dynamic feature association model, and building a digital physical model.
The adjusting module comprises a characteristic association model adjusting unit and a physical model adjusting unit;
The characteristic association model adjusting unit is used for obtaining the actual association degree between the characteristic sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic characteristic association model and adjusting the dynamic characteristic association model; the physical model adjusting unit is used for adjusting the digital physical model according to the actual association degree among the feature sets.
3 Physical models are used in the vehicle performance evaluation experiment process at any time of the history setting, wherein the first model is a motor material model of the vehicle, the second model is a motor material state and motor rotating speed relation model, and the third model is a motor rotating speed and vehicle speed relation model; extracting output data of the three models and checking error data;
Acquiring input data of a first model as a material type, taking the material type as one characteristic of the first model, and outputting data as a material state, so that the material state is taken as one characteristic of the first model, and the characteristic set of the first model is the material type and the material state; and so on, obtaining a characteristic set of the second model as a material state and a motor rotating speed, and obtaining a characteristic set of the third model as the motor rotating speed and a vehicle speed;
analyzing the association degree of the features between the first model and the second model, wherein one common feature exists in the two models, namely the data sources corresponding to the feature are the same, the other feature is different, the data proportion of the material type in the first model is 50%, and the association degree of the material type feature and the feature set in the second model is obtained =0.5× (1+0) =0.5; The data of the material state is 50 percent, and the obtained association degree is=0.5× (1+0.5) =0.75; The association of the feature set of the first model with the feature set of the second model is thus 0.625;
Re-acquiring the characteristic set of the second model as the material state, the motor rotating speed and the temperature every other unit time; the characteristic set of the first model is the material type, material state and temperature, and as the temperature is increased in the characteristic sets of the two models, the association degree of the two models is changed, and the calculation is needed to be carried out again, so that a new association degree is obtained; training the dynamic characteristic association model by taking the obtained association degrees as a training set, extracting rules of the dynamic characteristic association model according to the association degree changes at different times, and predicting the subsequent association degree changes;
acquiring actual feature sets of the first model and the second model in real time, wherein two identical features exist in the two models, and the association degree of the feature sets of the first model and the second model in the dynamic feature association model is set to be 0.8; obtaining all data of the first model after calculation, obtaining partial data with different characteristics from the second model, and integrating to obtain a digital physical model, wherein 2/3 is less than 0.8;
When one unit time passes, the association degree between two physical models is 0.8, but the actual association degree is 0.75, and the set error threshold is 0.1, because 0.8-0.75=0.05 < 0.1, the dynamic characteristic association model does not need to be retrained, and 0.8 > 0.75, partial data needs to be randomly removed, the data volume is (0.8-0.75) ×100=5, and the data volume corresponding to the same characteristics as the first model in the second model is assumed to be 100; and regenerating the digital physical model.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Finally, it should be noted that: the foregoing description is only a preferred embodiment of the present invention, and the present invention is not limited thereto, but it is to be understood that modifications and equivalents of some of the technical features described in the foregoing embodiments may be made by those skilled in the art, although the present invention has been described in detail with reference to the foregoing embodiments. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (9)
1. A digital physical model management method based on multi-feature association is characterized in that: the management method comprises the following steps:
step S100: in the process of acquiring any one vehicle performance evaluation experiment of history, carrying out error data investigation on the output data by the input data and the output data of each physical model to obtain an effective data set corresponding to the physical model, and carrying out feature extraction on the input data and the effective data set to obtain a feature set;
Step S200: acquiring a feature set of each physical model, and analyzing the change condition of each feature in the feature set of different time periods; comparing the association degree between any two feature sets, establishing a dynamic feature association model according to the association degree change condition between the feature sets in different time periods, and predicting the association degree change between the feature sets;
Step S300: in the vehicle performance evaluation experiment, input data and output data of each physical model are obtained in real time, and a real-time feature set is generated; acquiring the feature quantity of the same features in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity proportion of the same features in the two real-time feature sets and the predicted association degree in the dynamic feature association model, and constructing a digital physical model for simulating the comprehensive change process of the vehicle performance;
step S400: obtaining actual association degree between feature sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic feature association model, and adjusting the dynamic feature association model; adjusting the digital physical model according to the actual association degree between the feature sets;
the step S300 includes the steps of:
Step S301: acquisition of the first in real time Input data and output data in each physical model are set inExtracting a plurality of characteristics at any time to generate the firstReal-time feature set of individual physical models；
Step S302: acquisition of the firstThe physical model isReal-time feature set of momentsStatistics of real-time feature setsAnd real-time feature setThe same feature quantity is thatSetting upTime feature setAnd feature setIs related to the degree of association of (2); If it isOr alternativelyWherein, the method comprises the steps of, wherein,Is the firstThe physical model isThe number of features of the time of day,Is the firstThe physical model isThe feature quantity of the moment is fromAcquiring data corresponding to the same characteristics from the physical models, wherein the data quantity corresponding to the same characteristics is as follows; Randomly extracting the data quantity from the data corresponding to the same characteristics asData of (1), and (b)Integrating data corresponding to the same characteristics in the physical models; if it isOr alternativelyThen from the firstThe total data quantity selected by the physical model isWherein, the method comprises the steps of, wherein,Is of the same character in the firstThe corresponding data amount in the physical model; extracting the first data except the data corresponding to the same featuresThe data quantity corresponding to other characteristics in the physical model is as follows；
Step S303: will beTime of day (time)Data corresponding to the same feature in the physical model and the firstData extracted from the physical models are integrated, and thenThe same feature at the moment is put into a dynamic feature association model to obtain a time of arbitrary unit timeThe degree of correlation after that, from the firstPhysical model number oneAnd extracting data from the physical models for integration to obtain the digital physical model.
2. The digital physical model management method based on multi-feature association according to claim 1, wherein: the step S100 includes the steps of:
Step S101: setting the first The output data set of the physical model isSequencing the output data sets according to the output time of the output data sets in the physical model to obtain the output data setsIs a profile of (2); checking the distribution of the data of the distribution by using the box diagram, if the data group is outputIf the numerical value of a plurality of data in the box body exceeds the upper boundary and the lower boundary of the box body, the plurality of data are used as error data, and the output data group is deletedError data in (a) is obtainedEffective data group of individual physical model；
Step S102: acquisition of the firstInput data of the physical model, extracting data sources from the input data as the firstOne feature of the physical model is derived from the active data setExtracting several features from the distribution map to generate the firstFeature set of individual physical models。
3. A digital physical model management method based on multi-feature association according to claim 2, wherein: the step S200 includes the steps of:
Step S201: acquisition of Feature sets of any two physical models at any moment, and setting one physical model as a first physical modelA physical model, the feature set isAnother physical model is the firstA physical model, the feature set is; Randomly selecting feature setsOne feature of (2)Acquiring featuresIn the first placeCorresponding data in a physical modelWherein, the dataIn the first placeThe data volume duty ratio in each physical model is；
Step S202: acquiring dataIn the first placeData sources in physical model, obtaining feature setData sources of the corresponding data for each feature; according to the formula:
；
Wherein, Is a feature setIs used for the number of features of (a),Judging data as a judging functionAnd feature setMiddle featureCorresponding dataWhether the data sources of (a) are the same, if so,Otherwise, the device can be used to determine whether the current,; Calculated to obtainTime of day featureAnd feature setIs related to the degree of association of (2)；
Step S203: acquiring feature setsData and data sources corresponding to all features in the system are provided with feature setsTogether withThe individual features are according to the formula:
；
Calculated to obtain Time feature setAnd feature setIs related to the degree of association of (2)；
Step S204: every other unit timeReacquiring the firstFeature set of individual physical modelsAnd (d)Feature set of individual physical modelsCalculated to obtainTime feature setAnd feature setThe degree of association is; According to the acquisition time sequence of the feature set, for the firstPhysical model number oneSorting the association degree among the physical models; is provided to obtain togetherDegree of association before selectionThe association degree is used as a training set, and the restThe correlation degrees are used as test sets, and a dynamic characteristic correlation model is generated;
step S205: will be trained to concentrate The degree of association is put into a dynamic feature association model for training, and when the degree of association is output from the dynamic feature association modelWhen the number is equal to the correlation in the test set, counting and outputting the number of the correlation equal to the correlation in the test setSetting an accuracy thresholdIf (if)Selecting a plurality of relevancy from the test set, putting the relevancy into the training set, and retraining the dynamic characteristic relevancy model until the relevancy meets the requirement。
4. A digital physical model management method based on multi-feature association according to claim 3, wherein: the step S400 includes the steps of:
step S401: when a unit time passes Respectively obtain the firstPhysical model number oneOutput data of the physical model is obtainedActual feature set of individual physical modelsAnd (d)Actual feature set of individual physical modelsCalculating to obtain the firstPhysical model number oneThe physical model isActual degree of relevance of time of day；
Step S402: set in dynamic feature correlation model, the firstPhysical model number oneThe physical model isThe degree of correlation of the time isIf (if)Wherein, the method comprises the steps of, wherein,Is the error threshold value, thenActual degree of relevance of time of dayTraining the dynamic feature association model as a training set in the dynamic feature association model;
Step S403: if it is Continue from the firstThe data quantity extracted from each physical model isIf (if)Then at the firstThe data quantity extracted from each physical model isThe number of randomly removed data in the data of (a) isIs used for regenerating the digital physical model.
5. A digital physical model management system applied to a digital physical model management method based on multi-feature association as claimed in any one of claims 1 to 4, characterized in that: the management system comprises a feature extraction module, a relevance prediction module, a model construction module and an adjustment module;
The feature extraction module is used for acquiring output data of each physical model in the history vehicle performance evaluation experiment process at any time, performing error data investigation on the output data to obtain an effective data set corresponding to the physical model, and performing feature extraction on the effective data set to obtain a feature set;
The association degree prediction module is used for acquiring a feature set of each physical model and analyzing the change condition of each feature in the feature set of different time periods; comparing the association degree between any two feature sets, establishing a dynamic feature association model according to the association degree change condition between the feature sets in different time periods, and predicting the association degree change between the feature sets;
The model construction module is used for acquiring input data and output data of each physical model in real time in a vehicle performance evaluation experiment to generate a real-time feature set; acquiring the feature quantity of the same features in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity proportion of the same features in the two real-time feature sets and the predicted association degree in the dynamic feature association model, constructing a digital physical model, putting the data into the digital physical model to be converted into a digital form, and carrying out simulation analysis on the vehicle performance;
The adjustment module is used for obtaining the actual association degree between the feature sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic feature association model, and adjusting the dynamic feature association model; and adjusting the digital physical model according to the actual association degree between the feature sets.
6. The digital physical model management system of claim 5, wherein: the feature extraction module comprises a data checking unit and a feature set generating unit;
The data checking unit is used for acquiring output data of each physical model in the history vehicle performance evaluation experiment process at any time, and performing error data checking on the output data to obtain an effective data set corresponding to the physical model; the feature set generating unit is used for extracting features of the effective data set to obtain a feature set.
7. The digital physical model management system of claim 5, wherein: the association degree prediction module comprises an association degree comparison unit and a prediction unit;
The association degree comparison unit is used for acquiring the feature set of each physical model, analyzing the change condition of each feature in the feature set of different time periods and comparing the association degree between any two feature sets; the prediction unit is used for establishing a dynamic feature association model according to the association degree change condition among feature sets in different time periods and predicting the association degree change among the feature sets.
8. The digital physical model management system of claim 5, wherein: the model building module comprises a real-time data acquisition unit and a model building unit;
The real-time data acquisition unit is used for acquiring input data and output data of each physical model in real time in a vehicle performance evaluation experiment to generate a real-time feature set; the model building unit is used for obtaining the feature quantity of the same feature in the real-time feature sets of any two physical models, selecting data from the two physical models according to the quantity ratio of the same feature in the two real-time feature sets and the predicted association degree in the dynamic feature association model, and building a digital physical model.
9. The digital physical model management system of claim 5, wherein: the adjusting module comprises a characteristic association model adjusting unit and a physical model adjusting unit;
The characteristic association model adjusting unit is used for obtaining the actual association degree between the characteristic sets according to the output data of each physical model, analyzing the difference between the actual association degree and the association degree change condition predicted by the dynamic characteristic association model and adjusting the dynamic characteristic association model; the physical model adjusting unit is used for adjusting the digital physical model according to the actual association degree among the feature sets.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410494673.0A CN118094163B (en) | 2024-04-24 | 2024-04-24 | Digital physical model management system and method based on multi-feature association |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410494673.0A CN118094163B (en) | 2024-04-24 | 2024-04-24 | Digital physical model management system and method based on multi-feature association |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118094163A CN118094163A (en) | 2024-05-28 |
| CN118094163B true CN118094163B (en) | 2024-07-02 |
Family
ID=91150246
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410494673.0A Active CN118094163B (en) | 2024-04-24 | 2024-04-24 | Digital physical model management system and method based on multi-feature association |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118094163B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113642177A (en) * | 2021-08-16 | 2021-11-12 | 清华大学 | Digital twin virtual-real multi-vehicle mixed-driving simulation method and device |
| CN117113682A (en) * | 2023-08-24 | 2023-11-24 | 北京航空航天大学 | Digital twin digital-analog linkage system and method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117476235A (en) * | 2023-12-11 | 2024-01-30 | 北京大学第三医院（北京大学第三临床医学院） | Method for predicting pathological features of diseases by artificial intelligence technology |
| CN117725437B (en) * | 2024-02-18 | 2024-05-24 | 南京汇卓大数据科技有限公司 | Machine learning-based data accurate matching analysis method |
-
2024
- 2024-04-24 CN CN202410494673.0A patent/CN118094163B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113642177A (en) * | 2021-08-16 | 2021-11-12 | 清华大学 | Digital twin virtual-real multi-vehicle mixed-driving simulation method and device |
| CN117113682A (en) * | 2023-08-24 | 2023-11-24 | 北京航空航天大学 | Digital twin digital-analog linkage system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118094163A (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111459700B (en) | Equipment fault diagnosis method, diagnosis device, diagnosis equipment and storage medium | |
| CN108229588B (en) | Machine learning identification method based on deep learning | |
| CN112541532B (en) | Target detection method based on dense connection structure | |
| CN115907436B (en) | Quality coupling prediction-based water resource water environment regulation and control method and system | |
| CN113608916A (en) | Fault diagnosis method and device, electronic equipment and storage medium | |
| CN115408925A (en) | Rock mass parameter prediction method and device for tunnel construction | |
| CN111783843A (en) | Feature selection method and device and computer system | |
| CN115081515A (en) | Energy efficiency evaluation model construction method and device, terminal and storage medium | |
| CN111177500A (en) | Data object classification method and device, computer equipment and storage medium | |
| CN118094163B (en) | Digital physical model management system and method based on multi-feature association | |
| CN113420506A (en) | Method for establishing prediction model of tunneling speed, prediction method and device | |
| CN115907178A (en) | Clean ecosystem CO 2 Method for predicting exchange amount | |
| CN115080386A (en) | Scene effectiveness analysis method and device based on automatic driving function requirement | |
| CN114529096A (en) | Social network link prediction method and system based on ternary closure graph embedding | |
| CN113435519A (en) | Sample data enhancement method, device, equipment and medium based on antagonistic interpolation | |
| CN114049764A (en) | Traffic simulation method and system based on convolution long-time and short-time memory neural network | |
| CN109308565B (en) | Crowd performance grade identification method and device, storage medium and computer equipment | |
| CN111625525A (en) | Environmental data repairing/filling method and system | |
| CN115146596B (en) | Recall text generation method and device, electronic equipment and storage medium | |
| CN112434839B (en) | Distribution transformer heavy overload risk prediction method and electronic equipment | |
| CN116541252B (en) | Computer room fault log data processing method and device | |
| CN117076293B (en) | Software performance visual evaluation method based on lean sample hierarchy credible clustering thermodynamic diagram | |
| CN116415423B (en) | Computer simulation data processing system and method based on big data analysis | |
| CN114266953B (en) | Target detection flow self-optimization method based on deep learning | |
| CN117113484A (en) | Method and terminal for determining tunnel support design scheme |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |