Vehicle calibration method, device, equipment, medium and program based on digital twinning

Abstract

The invention relates to the technical field of vehicle calibration, and discloses a digital twinning-based vehicle calibration method, device, equipment, medium and program, and the method comprises the steps: constructing a digital twinning model of a vehicle according to the design data of the vehicle and the multi-source data during operation; adjusting to-be-calibrated parameters in the digital twin model, and simulating operation processes of the vehicle under different working conditions to obtain simulation operation results corresponding to different to-be-calibrated parameters; determining target calibration parameters of the vehicle under different working conditions based on the simulation operation result; and carrying out vehicle calibration based on the target calibration parameter. According to the method, the digital twin model corresponding to the real vehicle is constructed in the virtual environment, the to-be-calibrated parameters of the digital twin model are adjusted, the operation process of the vehicle under different working conditions is simulated, and vehicle calibration is performed according to the obtained simulated operation result, so that the calibration efficiency and accuracy are improved, and the vehicle calibration accuracy is improved. Efficient development and optimization of a vehicle power system are achieved, and vehicle calibration work is efficiently and accurately carried out in a virtual environment.

Description

Technical Field

[0001] The present invention relates to the field of vehicle calibration technology, and in particular to a vehicle calibration method, device, equipment, medium and program based on digital twins. Background Art

[0002] With the rapid development of the automotive industry and the increasing demand of consumers for automobile performance, calibration is crucial to ensure vehicle performance during automobile design and manufacturing.

[0003] Traditional vehicle calibration methods require a large number of tests on actual vehicles, which is not only time-consuming and costly, but also difficult to cover various extreme working conditions. In addition, when the existing development and evaluation is aimed at complex vehicle power systems with highly coupled mechanical, electrical, magnetic and thermal multi-physics fields, there are problems such as long development cycles, high R&D costs, and difficulty in achieving optimal system performance. Summary of the invention

[0004] In view of this, the present invention provides a vehicle calibration method, device, equipment, medium and program based on digital twins to solve the problems of long time and high cost when performing calibration tests on actual vehicles.

[0005] In a first aspect, the present invention provides a vehicle calibration method based on digital twins, the method comprising:

[0006] Build a digital twin model of the vehicle based on the vehicle’s design data and multi-source data during operation;

[0007] Adjust the parameters to be calibrated in the digital twin model, simulate the operation process of the vehicle under different working conditions, and obtain the simulation operation results corresponding to different parameters to be calibrated;

[0008] Based on the simulation results, determine the target calibration parameters of the vehicle under different working conditions;

[0009] Based on the target calibration parameters, the vehicle is calibrated.

[0010] Beneficial effects: Through the vehicle's design data and multi-source data, a digital twin model corresponding to the real vehicle is constructed in a virtual environment. By adjusting the parameters to be calibrated of the digital twin model, the operation process of the vehicle under different working conditions is simulated. The vehicle is calibrated according to the simulated operation results to improve the calibration efficiency and accuracy, achieve efficient development and optimization of the vehicle power system, and perform vehicle calibration work efficiently and accurately in a virtual environment.

[0011] In an optional implementation, a digital twin model of a vehicle is constructed based on the design data of the vehicle and multi-source data during operation, including:

[0012] Extract implicit features from multi-source data;

[0013] Construct performance models and multi-physics coupling models corresponding to each vehicle component / subsystem based on design data, multi-source data and implicit features;

[0014] The performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated and optimized to obtain a digital twin model of the vehicle.

[0015] Beneficial effects: The performance models and multi-physics field coupling models of various vehicle components / component subsystems are constructed through the vehicle's design data, multi-source data and their implicit features, and a digital twin model of the vehicle is obtained to facilitate the understanding of the vehicle's performance during simulated operation. The multi-physics field coupling model is used to improve the accuracy of simulations of various working conditions and further improve the calibration accuracy.

[0016] In an optional implementation, based on the design data, multi-source data and implicit features, a performance model corresponding to each component / subsystem of the vehicle is constructed, including:

[0017] Based on the design data, determine the interaction mechanism and relationship map between each component / component subsystem, and determine the mechanism model and data-driven model corresponding to each component / component subsystem;

[0018] Based on multi-source data and implicit features, determine whether the mechanism model and data-driven model corresponding to each component / component subsystem meet the preset performance requirements, and determine the corresponding modeling method of the current component / component subsystem based on the judgment results;

[0019] According to the interaction mechanism, relationship map, multi-source data and implicit features, and in accordance with the modeling method of each component / component subsystem, the performance model corresponding to each component / component subsystem is constructed.

[0020] Beneficial effects: By determining the interaction mechanism and relationship map between each component / component subsystem, the interaction and influence mechanism between the various components of the vehicle are clarified, and by judging whether the mechanism model and data-driven model meet the preset performance requirements, the appropriate modeling method is selected to model the performance model corresponding to each component / component subsystem, so as to accurately evaluate the vehicle performance.

[0021] In an optional implementation, the preset performance requirement includes a preset accuracy requirement and a preset time constraint; and determining a corresponding modeling method of the current component / component subsystem according to the judgment result includes:

[0022] If the judgment result is that the mechanism model meets the preset accuracy requirements and the preset time constraints, the modeling method is to model according to the mechanism model and multi-source data;

[0023] Or, if the judgment result is that the mechanism model does not meet the preset accuracy requirement and the mechanism model meets the preset time constraint, then the modeling method is to model according to the data-driven model and multi-source data corresponding to the preset time constraint;

[0024] Alternatively, if the judgment result is that both the mechanism model and the data-driven model do not meet the preset accuracy requirements and the preset time constraints, the modeling method is to construct a mechanism-driven hybrid model based on the mechanism model and the data-driven model, and perform modeling based on the mechanism-driven hybrid model and multi-source data.

[0025] Beneficial effects: It is possible to determine whether the mechanism model and data-driven model meet the preset accuracy requirements and preset time constraints respectively, so as to select the appropriate modeling method to model the corresponding components / component subsystems and improve the accuracy of the performance model during the simulation process.

[0026] In an optional implementation, a multi-physics field coupling model corresponding to each component / subsystem of a vehicle is constructed based on design data, multi-source data and implicit features, including:

[0027] Based on the design data, determine the multi-physical field coupling relationship corresponding to each component / component subsystem, and determine the mechanism model, data-driven model and mechanism-driven hybrid model corresponding to each component / component subsystem; wherein the physical field includes at least two of the mechanical field, electric field, magnetic field, flow field and thermal field;

[0028] Perform multi-scale feature decomposition and reconstruction on latent features to obtain unit-scale latent features;

[0029] Based on the implicit features of the unit scale, multi-source data and multi-physical field coupling relationships, and in accordance with the mechanism model, data-driven model or mechanism-driven hybrid model of each component / component subsystem, a multi-physical field coupling model corresponding to each component / component subsystem of the vehicle is constructed.

[0030] Beneficial effects: Through the multi-physical field coupling relationship corresponding to each component / component subsystem, accurate mapping of data characteristics and physical phenomena is achieved to simulate various working conditions and states of the vehicle in actual operation, providing a simulation environment for the optimization of calibration parameters.

[0031] In an optional implementation, the performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated and optimized to obtain a digital twin model of the vehicle, including:

[0032] Based on the relationship graph, the performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated to obtain an initial system model; wherein the initial system model includes an environment setting model, a driver decision setting model, a calibration parameter setting model, a controller model and a drive mechanism model;

[0033] Perform feature prediction based on multi-source data to obtain predicted feature parameters; the predicted feature parameters include parameters to be calibrated;

[0034] Based on the predicted characteristic parameters, the initial system model is updated and optimized to obtain the digital twin model of the vehicle.

[0035] Beneficial effects: The performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated, and a prediction, update and optimization mechanism for characteristic parameters is established, so that the digital twin model can simulate the operation of the vehicle under different working conditions and can be adaptively updated and optimized.

[0036] In an optional implementation, based on the predicted characteristic parameters, the initial system model is updated and optimized to obtain a digital twin model of the vehicle, including:

[0037] Based on the predicted characteristic parameters, the data convergence rate, data forgetting rate and data update rate of the initial system model are adjusted, and the initial characteristic parameters of the initial system model are optimized to obtain a digital twin model.

[0038] Beneficial effects: By formulating an automatic identification and update strategy for characteristic parameters including parameter convergence control, forgetting factor and dynamic adjustment of update rate, the adaptive adjustment capability of the digital twin model is improved.

[0039] In an optional implementation, extracting implicit features of multi-source data includes:

[0040] Cluster multi-source data to obtain grouped data of different categories;

[0041] Extract association rule features and time-frequency domain features from grouped data to obtain implicit features of multi-source data.

[0042] Beneficial effects: Clustering multi-source data and extracting association rule features and time-frequency domain features, mining the implicit features of multi-source data, and facilitating the construction of a digital twin model of the vehicle.

[0043] In an optional implementation, before extracting implicit features of multi-source data, the method further includes:

[0044] Perform dimensionality reduction, sparse reconstruction, filtering, data correction and multi-scale feature fusion on multi-source data.

[0045] Beneficial effects: By performing dimensionality reduction and sparse reconstruction on multi-source data, data redundancy is eliminated, data sparsification and structuring are achieved, multi-source data noise and anomalies are eliminated through filtering processing, the expression ability of multi-source data is improved through data correction, and multi-source data information fusion of different batches and types is achieved through multi-scale feature fusion.

[0046] In an optional implementation, based on the simulation operation results, determining target calibration parameters of the vehicle under different operating conditions includes:

[0047] For each operating condition, determine the target simulation operation results that meet the preset calibration requirements of the current operating condition;

[0048] For each operating condition, the target calibration parameters of the vehicle under the current operating condition are obtained according to the parameters to be calibrated corresponding to the target simulation operation results.

[0049] Beneficial effects: By analyzing and evaluating the simulation operation results, the target calibration parameters of the vehicle under different working conditions are determined, the calibration time is reduced, and the vehicle calibration efficiency is improved.

[0050] In a second aspect, the present invention provides a vehicle calibration device based on digital twins, the device comprising:

[0051] A first processing module is used to build a digital twin model of the vehicle based on the design data of the vehicle and multi-source data during operation;

[0052] The second processing module is used to adjust the parameters to be calibrated in the digital twin model, simulate the operation process of the vehicle under different working conditions, and obtain simulation operation results corresponding to different parameters to be calibrated;

[0053] A third processing module is used to determine target calibration parameters of the vehicle under different working conditions based on the simulation operation results;

[0054] The fourth processing module is used to perform vehicle calibration based on the target calibration parameters.

[0055] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, computer instructions are stored in the memory, and the processor executes the digital twin-based vehicle calibration method of the above-mentioned first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0056] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the digital twin-based vehicle calibration method of the above-mentioned first aspect or any corresponding embodiment thereof.

[0057] In a fifth aspect, the present invention provides a computer program product, comprising computer instructions, which are used to enable a computer to execute the digital twin-based vehicle calibration method of the above-mentioned first aspect or any corresponding embodiment thereof.

[0058] The beneficial effects of the present invention are:

[0059] Through the vehicle's design data and multi-source data, a digital twin model corresponding to the real vehicle is constructed in a virtual environment. By adjusting the parameters to be calibrated of the digital twin model, the operation process of the vehicle under different working conditions is simulated. The vehicle is calibrated according to the simulated operation results to improve the calibration efficiency and accuracy, achieve efficient development and optimization of the vehicle power system, and perform vehicle calibration work efficiently and accurately in a virtual environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0061] Figure 1 is a flow chart of a vehicle calibration method based on digital twins according to an embodiment of the present invention;

[0062] Figure 2 is a flow chart of another vehicle calibration method based on digital twins according to an embodiment of the present invention;

[0063] Figure 3 is a schematic diagram of a process of multi-source data preprocessing and fusion according to an embodiment of the present invention;

[0064] Figure 4 is a schematic diagram of a process for building a performance model according to an embodiment of the present invention;

[0065] Figure 5 is a schematic diagram of a process for constructing a multi-physics field coupling model according to an embodiment of the present invention;

[0066] Figure 6 is a structural schematic diagram of a digital twin model according to an embodiment of the present invention;

[0067] Figure 7 is a schematic diagram of a process of adaptive optimization of a digital twin model according to an embodiment of the present invention;

[0068] Figure 8 is a schematic diagram of a vehicle calibration method according to an embodiment of the present invention;

[0069] Fig. 9 is a schematic diagram of a process of building a digital twin model according to an embodiment of the present invention;

[0070] Fig.10 is a structural block diagram of a vehicle calibration device based on digital twins according to an embodiment of the present invention;

[0071] Fig.11 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0073] With the rapid development of the automotive industry and the increasing demand of consumers for automotive performance, calibration is crucial to ensure vehicle performance during the design and manufacturing of automobiles. Existing development and evaluation methods for vehicle power systems with highly coupled mechanical, electrical, magnetic and thermal multi-physics fields have problems such as long development cycles, high R&D costs, and difficulty in achieving optimal system performance. Traditional vehicle calibration methods require a large number of tests on actual vehicles, which is not only time-consuming and costly, but also difficult to cover various extreme working conditions.

[0074] Patent 202110332720.8 discloses a method for virtually calibrating an engine model and its related device, providing a method for virtually calibrating an engine model to improve the accuracy of the engine model calibration value, thereby reducing the workload of the calibration work. However, the operation of the vehicle power system is often more complicated, and the operation of a certain component of the vehicle is inseparable from the participation of other components or component subsystems. Therefore, only establishing a digital twin model of a certain component cannot accurately simulate the vehicle operation process, and accurate calibration cannot be achieved.

[0075] Therefore, the embodiment of the present invention provides a vehicle calibration method based on digital twins, which simulates the vehicle operation and calibration process by building a digital twin model corresponding to the real vehicle in a virtual environment to improve the calibration efficiency and accuracy, achieve efficient development and optimization of the vehicle power system, and perform vehicle calibration work efficiently and accurately in a virtual environment. The present invention builds a highly accurate virtual simulation environment by integrating physical mechanisms and data-driven methods, and realizes rapid calibration and optimization of vehicle performance through the digital twin model of the vehicle.

[0076] According to an embodiment of the present invention, an embodiment of a vehicle calibration method based on digital twins is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0077] In this embodiment, a vehicle calibration method based on digital twins is provided, which can be used for a computer device or electronic device for vehicle calibration, such as a computer, a computer, etc. Figure 1 is a flow chart of a vehicle calibration method based on digital twins according to an embodiment of the present invention, such as Figure 1 As shown, the process includes the following steps:

[0078] Step S101, constructing a digital twin model of the vehicle based on the vehicle's design data and multi-source data during operation.

[0079] Specifically, the design data of the real vehicle and multi-source data from sensors, test benches, and historical databases can be collected to build a digital twin model corresponding to the real vehicle. The digital twin model simulates the vehicle operation process by adjusting parameters such as characteristic parameters, driver decision parameters, and environmental parameters.

[0080] Step S102, adjust the parameters to be calibrated in the digital twin model, simulate the operation process of the vehicle under different working conditions, and obtain simulation operation results corresponding to different parameters to be calibrated.

[0081] Specifically, the corresponding environment setting module in the digital twin model can be used to create a virtual environment similar to the real environment in the virtual environment, and set different road conditions, weather and other conditions to achieve simulation of different working conditions. In addition, by adjusting the characteristic parameters of the digital twin model (including parameters to be calibrated), the changes in the vehicle operation process or the driver's control strategy can be simulated to achieve the simulation of the vehicle's operation process under different working conditions and the calibration process (such as engine calibration, chassis calibration, body calibration, etc.), and obtain the simulation operation results corresponding to different parameters to be calibrated.

[0082] Step S103, based on the simulation operation results, determining the target calibration parameters of the vehicle under different working conditions.

[0083] Specifically, data analysis is performed on the simulation operation results output by the digital twin model, key parameters and performance indicators are extracted, and the target calibration parameters of the vehicle under different working conditions are obtained.

[0084] Step S104: perform vehicle calibration based on the target calibration parameters.

[0085] Specifically, according to the obtained target calibration parameters, the setting parameters of each component / subsystem of each component of the vehicle when operating under different working conditions are set.

[0086] The digital twin-based vehicle calibration method provided in this embodiment constructs a digital twin model corresponding to the real vehicle in a virtual environment through the vehicle's design data and multi-source data, simulates the vehicle's operation process under different working conditions by adjusting the parameters to be calibrated of the digital twin model, and performs vehicle calibration based on the obtained simulation operation results to improve calibration efficiency and accuracy, achieve efficient development and optimization of the vehicle's power system, and perform vehicle calibration work efficiently and accurately in a virtual environment.

[0087] In this embodiment, a vehicle calibration method based on digital twins is provided, which can be used for a computer device or electronic device for vehicle calibration, such as a computer, a computer, etc. Figure 2 is a flow chart of a vehicle calibration method based on digital twins according to an embodiment of the present invention, such as Figure 2 As shown, the process includes the following steps:

[0088] Step S201, constructing a digital twin model of the vehicle based on the design data of the vehicle and multi-source data during operation.

[0089] Specifically, the above step S201 includes:

[0090] S2011, Extracting implicit features from multi-source data.

[0091] Specifically, before extracting the implicit features of multi-source data, it is necessary to pre-process and fuse the multi-source data, for example, perform dimensionality reduction, sparse reconstruction, filtering, data correction, and multi-scale feature fusion on the multi-source data to eliminate data noise anomalies and correct and fuse the data. Furthermore, cluster the multi-source data to obtain grouped data of different categories, extract association rule features and time-frequency domain features in the grouped data, and obtain the implicit features of the multi-source data.

[0092] In some optional implementations, first, sparse information extraction and reconstruction of multi-source data is performed. Figure 3 As shown in the figure, dimensionality reduction methods such as principal component analysis and independent component analysis are used to map high-dimensional multi-source data into a low-dimensional subspace to eliminate the redundant characteristics of the data; sparse representation and dictionary learning methods are used to represent multi-source data as a linear combination of a set of basic elements (atoms) to achieve signal sparsification and structuring.

[0093] Next, we implement adaptive detection and repair of multi-source data. Figure 3, using methods such as adaptive filtering and Kalman filtering, extracting and utilizing the spatiotemporal correlation and dynamic characteristics among multi-source data, realizing data quality monitoring and adaptive adjustment, thereby eliminating multi-source data noise and anomalies; using methods such as interpolation and regression analysis, utilizing the statistical characteristics and functional relationships among multi-source data, realizing effective filling and correction of missing and abnormal data, so as to further enhance the expression capability of multi-source data.

[0094] Then, information fusion is performed on multi-source data. Figure 3 , feature selection and feature transformation methods are used to extract representative and discriminative features from multi-source data, and convert them into a unified format and scale to fuse multi-source data of different batches and types; then, support vector machine and random forest methods are used to utilize the classification and regression relationship between multi-source data to optimize and evaluate the fused data and improve the multi-source data fusion effect.

[0095] Finally, we can mine implicit knowledge from multi-source data. Figure 3 , clustering analysis methods such as K-means algorithm and hierarchical clustering algorithm are used to group multi-source data according to indicators such as distance, density, and connectivity; association rule mining methods such as Apriori algorithm and FP-Growth algorithm are used to extract strong association rules from multi-source data according to indicators such as support, confidence, and lift; sequence pattern mining methods are used to mine time-frequency domain features of multi-source data according to indicators such as length, frequency, and compactness.

[0096] In the embodiment of the present application, by performing dimensionality reduction and sparse reconstruction on multi-source data, data redundancy is eliminated, data sparsification and structuring are achieved, multi-source data noise and anomalies are eliminated through filtering, multi-source data expression capabilities are improved through data correction, and multi-scale feature fusion is used to achieve information fusion of multi-source data of different batches and types. Subsequently, the multi-source data is clustered and association rule features and time-frequency domain features are extracted to mine the implicit features of the multi-source data, so as to facilitate the construction of a digital twin model of the vehicle.

[0097] S2012, based on design data, multi-source data and implicit features, build performance models and multi-physics field coupling models corresponding to each vehicle component / component subsystem.

[0098] In some optional implementations, the above step S2021 includes:

[0099] Step a1, based on the design data, determine the interaction mechanism and relationship map between each component / component subsystem, and determine the mechanism model and data-driven model corresponding to each component / component subsystem.

[0100] Specifically, Figure 4As shown, the multi-level complex interactions and influence mechanisms between the components / subsystems of the power system are determined, such as the interaction mechanism in which the engine transmits power to the transmission, the transmission controls the power, and they work together to efficiently transmit the power to the wheels.

[0101] Furthermore, combined with the prior knowledge of the physical / chemical theoretical system followed by the system operation (such as engine combustion cycle efficiency, conversion of electrical energy into mechanical energy, vehicle dynamics calculation, transmission ratio calculation, etc.) and the implicit knowledge between multi-source data, a model structure of the vehicle power system is constructed (such as the mechanism model of each component / component subsystem, data-driven model, mechanism-data-driven hybrid model), the specific function of each sub-model is clarified, the key model parameters in each sub-model are sorted out, the state connection and data interface between the models are constructed, and the model library and relationship map construction at the power system component level and component subsystem level are realized.

[0102] It should be noted that the components may include components of the vehicle power system such as the engine, motor, battery, clutch, transmission, drive shaft, etc., but the present invention is not limited thereto.

[0103] Step a2, based on multi-source data and implicit features, judge whether the mechanism model and data-driven model corresponding to each component / component subsystem meet the preset performance requirements, and determine the corresponding modeling method of the current component / component subsystem according to the judgment result.

[0104] Specifically, based on the model library and relationship map, guided by the model application scenario, combined with the bench test data in the multi-source data, see again Figure 4 , explore the comprehensive performance of the models of each component / component subsystem in terms of accuracy, real-time performance, etc., and determine the modeling method that characterizes the nonlinear / time-varying / uncertainty characteristics of the system based on this. For example, the nonlinear modeling method can be a nonlinear differential equation, etc., the time-varying modeling method can be a multidimensional Taylor network, etc., and the uncertainty modeling method can be additive / multiplicative uncertainty modeling, etc., but the present invention is not limited thereto.

[0105] Specifically, the preset performance requirements include preset accuracy requirements and preset time constraints. If the judgment result is that the mechanism model meets the preset accuracy requirements and the preset time constraints, the modeling method is to model based on the mechanism model and multi-source data. Figure 4 For each component or component subsystem with clear physical laws, clear action mechanisms and mechanism models that meet the real-time and accuracy requirements, we carry out mechanism model construction based on prior knowledge of physical / chemical theoretical systems and model verification and correction based on measured data.

[0106] In addition, if the judgment result is that the mechanism model does not meet the preset accuracy requirement and the mechanism model meets the preset time constraint, the modeling method is to model according to the data-driven model and multi-source data corresponding to the preset time constraint. Figure 4 For each component / component subsystem with unclear physical laws and unclear action mechanisms, consider the complexity, real-time, accuracy, generalization and other performance factors of the data-driven model, select a suitable basic mathematical model architecture to build a data-driven model that ensures coordination and consistency with the performance level of the above-mentioned mechanism model. And use existing measured data or design experiments to prepare new measured data, improve the training and verification data sets, and carry out training and verification of data-driven models.

[0107] In addition, if the judgment result is that both the mechanism model and the data-driven model do not meet the preset accuracy requirements and the preset time constraints, the modeling method is to build a mechanism-driven hybrid model based on the mechanism model and the data-driven model, and modeling is performed based on the mechanism-driven hybrid model and multi-source data. Figure 4 In order to solve the problem that it is difficult to fully capture the dynamic characteristics of some power system components / component subsystems with high nonlinearity, time-varying and uncertainty by relying solely on mechanism modeling or data-driven modeling, multi-source, multi-scale and multi-modal mechanism-data-driven hybrid modeling is carried out, and physical information neural networks, physical-guided neural networks, physical-information machine learning and other methods are used to achieve high-precision hybrid modeling of the complex dynamic characteristics of the system.

[0108] In an embodiment of the present application, it is determined whether the mechanism model and the data-driven model meet the preset accuracy requirements and the preset time constraints, so as to select a suitable modeling method to model the corresponding components / component subsystems and improve the accuracy of the performance model during the simulation process.

[0109] Step a3, constructing the performance model corresponding to each component / each component subsystem according to the interaction mechanism, relationship map, multi-source data and implicit features and the modeling method of each component / each component subsystem.

[0110] Specifically, see again Figure 4 , it is also possible to integrate and construct multi-level models at the vehicle subsystem level, component level, and power system level based on the relationship graph.

[0111] Furthermore, identify or design experiments to obtain measured data under complex working conditions, extreme environments, human interference and other usage scenarios, analyze the influencing factors and evaluation indicators of the generalization ability of the models of various components of the vehicle power system under complex working conditions, extreme environments, human interference and other usage scenarios, and clarify the key systems that affect the generalization ability of the vehicle power system model under the above scenarios. For key influencing factors, combined with test data sets, physical / empirical knowledge guidance and constraints, considering the complexity, reliability, real-time performance and other factors of the adaptive decision / adjustment method, carry out adaptive decision / adjustment methods based on optimization algorithms such as gradient descent, genetic algorithm, particle swarm algorithm, etc., to achieve adaptive adjustment of the type / structure / parameters of each model in different scenarios.

[0112] Finally, it is also possible to integrate the data-driven models of the above-mentioned key components / component subsystems with adaptive decision-making / adjustment methods to achieve automatic switching of the type / structure / parameters of the data-driven models of each component in different usage scenarios, and to realize data-driven modeling of the performance of vehicle power system components and adaptive adjustment of model structure and parameters.

[0113] In the embodiments of the present application, by determining the interaction mechanism and relationship map between each component / component subsystem, the interaction and influence mechanism between the various components of the vehicle are clarified, and by judging whether the mechanism model and the data-driven model meet the preset performance requirements, a suitable modeling method is selected to model the performance model corresponding to each component / component subsystem, so as to accurately evaluate the vehicle performance.

[0114] In some optional implementations, the above step S2021 further includes:

[0115] Step b1, based on the design data, determine the multi-physical field coupling relationship corresponding to each component / each component subsystem, and determine the mechanism model, data-driven model and mechanism-driven hybrid model corresponding to each component / each component subsystem; wherein the physical field includes at least two of the mechanical field, electric field, magnetic field, flow field and thermal field.

[0116] For example, Figure 5 As shown, according to the working principles and physical characteristics of the engine / motor / battery / clutch / transmission / drive shaft, combined with the test data, the coupling mechanisms of the key physical fields such as the mechanical field, electric field, magnetic field, flow field, and thermal field inside the above components are analyzed, and the coupling relationship between the key state variables in the multi-physical fields of each component is determined, such as the coupling between the thermal field of the engine and the mechanical field of the transmission, the coupling between the flow field and the thermal field in the oil supply pipeline, and the coupling between the magnetic field and the electric field in the motor.

[0117] Step b2, perform multi-scale feature decomposition and reconstruction on the latent features to obtain unit-scale latent features.

[0118] For example, see again Figure 5 Based on the structural parameters, operating mechanisms and test data of different time scales of key components, the basic granularity units are defined, decomposed and reconstructed for implicit features such as the time domain dynamic characteristics of each key component, such as the cross-scale decomposition and reconstruction of data features of fuel injection quantity and SOC.

[0119] Step b3, based on the implicit features of the unit scale, multi-source data and multi-physical field coupling relationship, and according to the mechanism model or data-driven model or mechanism-driven hybrid model of each component / component subsystem, construct a multi-physical field coupling model corresponding to each component / component subsystem of the vehicle.

[0120] For example, see again Figure 6 , combining data-driven modeling and mechanism modeling theory, establish a method to express the dynamic response characteristics of key components with different resolutions on the unit-level time scale, and realize multi-physics field modeling at the component level based on mechanism modeling, data-driven modeling, and mechanism-data-driven hybrid modeling methods. First, make reasonable assumptions based on the application scenarios of the modeling object and the purpose of the model, and use simple mechanism analysis, expert experience, artificial intelligence, and some statistical theory methods (such as correlation analysis, principal component analysis, etc.) to determine the correlation structure between input variables and output variables; then, according to the inherent mechanism of the system, use neural networks, Bayesian networks, partial least squares, support vector machines and other tools to train and learn to obtain the approximate mathematical function relationship between input variables and output variables, and verify its rationality; then solve the established mathematical model and explain the practical significance of the mathematical solution; finally, compare the model analysis results with the actual results to verify the accuracy and generalization of the model. By analyzing the energy flow interaction mechanism between components, clarify the data interface of the multi-physics field model of each component, and realize the integration of multi-physics field models at the system level.

[0121] In addition, random forest and deep learning algorithms are used to establish a data-driven model that reflects the aging mechanism, service characteristics and interaction characteristics of the battery / clutch / shift actuator. Based on a multi-objective optimization algorithm, adaptive decision-making and adjustment of the data-driven model type, structure and parameters are achieved for different service states.

[0122] In the embodiments of the present application, accurate mapping of data features and physical phenomena is achieved through the multi-physical field coupling relationship corresponding to each component / component subsystem, so as to simulate various operating conditions and states of the vehicle in actual operation and provide a simulation environment for the optimization of calibration parameters.

[0123] S2013: Integrate and optimize the performance models and multi-physics field coupling models corresponding to each component / component subsystem to obtain a digital twin model of the vehicle.

[0124] In some optional implementations, the above step S2013 includes:

[0125] Step c1, based on the relationship graph, integrate the performance model and multi-physics field coupling model corresponding to each component / component subsystem to obtain an initial system model; wherein the initial system model includes an environment setting model, a driver decision setting model, a calibration parameter setting model, a controller model and a drive mechanism model.

[0126] For example, Figure 6 A schematic diagram of the structure of a digital twin model provided by an embodiment of the present invention is shown in FIG. Figure 6 As shown in the figure, the virtual scene can be constructed through the environment setting model, and the data of the virtual operating environment such as altitude, temperature, slope, wind speed, etc. can be set; the driver decision setting model can be used to simulate the behavior and decision-making process of the real driver, including the setting of signals such as target vehicle speed, throttle opening, brake opening, etc.; the calibration parameter setting module can set and simulate the operation and calibration process of the vehicle in the virtual environment, including engine calibration, chassis calibration, body calibration, etc.

[0127] See again Figure 6 The controller model and drive mechanism model can simulate the process of vehicle operation and physical action execution. The controller model contains the actual strategies and control algorithms of each system in the vehicle (such as ECU control strategy, MCU control strategy, BMS control strategy, etc.), and the drive mechanism model contains the engine, transmission, drive system, brake system, etc. on the vehicle power transmission route. The power and economy parameter output window can display the virtual calibration results, that is, the simulation operation results, in the time domain.

[0128] Step c2, perform feature prediction based on multi-source data to obtain predicted feature parameters, wherein the predicted feature parameters include parameters to be calibrated.

[0129] Specifically, Figure 7 As shown, based on multi-source data analysis, operation mechanism and modeling method, a set of vehicle power system characteristic parameters that reflect the nonlinearity / time-varying and multi-input and multi-output coupling characteristics of the system is identified and extracted, where the characteristic parameters include parameters that need to be calibrated. For the multivariable time-varying nonlinear component / component subsystem model, considering the associated information under the system-level large model framework and the system accuracy requirements under different application scenarios, the extended Kalman filter method is used to achieve online continuous estimation of characteristic parameters based on the latest multi-source data to obtain predicted characteristic parameters.

[0130] Step c3: based on the predicted characteristic parameters, the initial system model is updated and optimized to obtain a digital twin model of the vehicle.

[0131] Specifically, based on the predicted characteristic parameters, the data convergence rate, data forgetting rate and data update rate of the initial system model are adjusted, and the initial characteristic parameters of the initial system model are optimized to obtain a digital twin model. Figure 7 , comprehensively considering the stability, time-varying and adaptability of the model, the unknown parameters are estimated and compensated online in real time through the radial basis function (RBF) neural network, so as to control the convergence of parameters. Then, the forgetting factor is used to control the degree of forgetting of historical data by the current data. By adjusting the value of the forgetting factor, the adaptation and optimization of the dynamic characteristics of the system are achieved. Finally, the dynamic adjustment technology of the update rate is introduced, allowing the system to automatically adjust the data refresh rate according to the current environmental conditions and optimize the system performance.

[0132] Specifically, see again Figure 7 Based on multi-source data, we identify the impact of complex working conditions and environmental interference on the accuracy of key state variables of the vehicle power system, and identify the logic of the role of characteristic parameter sets in the accuracy of key state variables. Based on the measured multi-source data under complex working conditions and environmental interference and the simulation data of the data-driven model, we use the particle swarm optimization algorithm to achieve automatic optimization of characteristic parameters and ensure the accuracy and robustness of the model under complex conditions.

[0133] This embodiment improves the adaptive adjustment capability of the digital twin model by formulating an automatic identification and update strategy for characteristic parameters including parameter convergence control, forgetting factor, and dynamic adjustment of update rate.

[0134] The embodiment of the present application uses the vehicle's design data, multi-source data and its implicit features to construct performance models and multi-physics field coupling models of each vehicle component / component subsystem to obtain a digital twin model of the vehicle, so as to understand the performance of the vehicle during simulated operation, and through the multi-physics field coupling model, to improve the accuracy of simulation of various working conditions and further improve the calibration accuracy.

[0135] Step S202, adjust the parameters to be calibrated in the digital twin model, simulate the operation process of the vehicle under different working conditions, and obtain simulation operation results corresponding to different parameters to be calibrated.

[0136] Step S203, based on the simulation operation results, determine the target calibration parameters of the vehicle under different working conditions.

[0137] Specifically, the above step S203 includes:

[0138] Step S2031, for each working condition, determine the target simulation operation result that meets the preset calibration requirements of the current working condition.

[0139] Specifically, the simulation operation results may include economic parameters and dynamic parameters. By evaluating the performance of economic parameters and dynamic parameters under different working conditions, the target simulation operation results corresponding to different working conditions are determined. For example, when the SOC is low, the simulation operation results with higher economic parameters should be given priority.

[0140] Step S2032, for each working condition, according to the parameters to be calibrated corresponding to the target simulation operation result, obtain the target calibration parameters of the vehicle under the current working condition.

[0141] Specifically, the parameters to be calibrated corresponding to the target simulation operation results are used as target calibration parameters of the vehicle in the current working condition.

[0142] This embodiment analyzes and evaluates the simulation operation results to determine the target calibration parameters of the vehicle under different working conditions, thereby reducing the calibration time and improving the vehicle calibration efficiency.

[0143] Step S204: perform vehicle calibration based on the target calibration parameters. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0144] The present invention first collects multi-source data from sensors, test benches, historical databases, etc., and uses sparse information extraction, adaptive detection and repair, information fusion and other technologies to clean, integrate and standardize the data to achieve multi-source data fusion and preprocessing; through data mining technology, the vehicle performance characteristics and failure modes implicit in the data are extracted to provide data support for subsequent calibration and optimization.

[0145] Next, we integrate multidisciplinary knowledge such as vehicle dynamics, thermodynamics, and electromagnetism, and use advanced modeling technology to build an integrated simulation model covering all key components and systems of the vehicle, which reflects the vehicle's performance under various working conditions in real time, including power, economy, emissions, thermal stability, etc.

[0146] Then, based on the multi-physics field data-driven modeling method, the preprocessed data is deeply integrated with the vehicle physical model to achieve accurate mapping of data features and physical phenomena, build a multi-physics field coupling model, simulate the actual various working conditions and states of the vehicle, and provide a simulation environment for the optimization of calibration parameters.

[0147] Furthermore, intelligent optimization algorithms (such as genetic algorithms, particle swarm optimization algorithms, etc.) are used to automatically identify and optimize key parameters in the model by combining the simulation model with actual test data. Through the iterative optimization process, the model parameters are gradually adjusted to make the performance of the simulation model closer to that of the actual vehicle, thereby achieving rapid calibration and optimization of vehicle performance.

[0148] Finally, the virtual calibration results are verified and feedback is given, and the optimized simulation model is compared with the real vehicle to verify the accuracy and reliability of the virtual calibration. According to the real vehicle test results, the simulation model is corrected and feedback is given to continuously improve the accuracy and scope of application of the model.

[0149] The vehicle calibration method based on digital twin of the present invention is described in detail below with reference to a specific application example. Figure 8 As shown, the application example includes the following steps:

[0150] Step S801, collect vehicle design data, sensor data, etc., and build a vehicle digital twin model.

[0151] Specifically, the digital twin model can be constructed using professional modeling software such as MATLAB / Simulink, ADAMS, etc., and a high-precision integrated simulation model can be constructed based on the vehicle's structure and performance requirements. Necessary interfaces and parameters are set in the model to facilitate subsequent data input and parameter adjustment. Fig. 9 As shown, the details are as follows:

[0152] Step S901, multi-source data preprocessing and fusion. Through sensors, test benches and other equipment, various data (such as speed, acceleration, temperature, pressure, etc.) of the vehicle during operation are collected in real time. The collected data is cleaned, integrated and standardized to eliminate noise and outliers and improve data quality. Using data mining technology, the vehicle performance characteristics and failure modes hidden in the data are extracted.

[0153] Step S902: Data-driven modeling of vehicle powertrain component performance. Using a large amount of actual data, through mathematical models such as statistics and machine learning, the model has the characteristics of strong nonlinear mapping, and better captures complex system relationships and trends, nonlinearity and time-varying characteristics.

[0154] Step S903, cross-scale multi-physics fusion modeling. According to the actual working conditions and performance requirements of the vehicle, a multi-physics coupling model is constructed to simulate the operating status of the vehicle under various working conditions. The pre-processed data is fused with the multi-physics coupling model to achieve accurate mapping of data features and physical phenomena. According to the mapping results, the relevant parameters and constraints in the simulation model are adjusted to make the model more in line with the actual situation.

[0155] Step S904, self-identification and self-optimization of system calibration parameters. Select a suitable intelligent optimization algorithm (such as genetic algorithm, particle swarm optimization algorithm, etc.), set the parameters and constraints of the algorithm. Compare the simulation model with the actual test data, and calculate the difference and error between the model and the actual. Use the optimization algorithm to automatically identify and optimize the key parameters in the model, and gradually reduce the difference and error between the model and the actual.

[0156] Step S802, creating a virtual environment similar to the real environment, and setting different road conditions, weather conditions, etc.

[0157] Step S803: simulate the operation and calibration process of the vehicle on the digital twin model, including adjusting parameters, setting control strategies, etc.

[0158] Step S804: Display the results of the virtual calibration through the result display and analysis module, and perform data analysis to extract key parameters and performance indicators.

[0159] The embodiments of the present application have the following beneficial effects: high efficiency, calibration is performed through a virtual environment, and there is no need to actually manufacture the vehicle, which greatly shortens the calibration cycle; flexibility, the parameters and settings of the digital twin model can be adjusted according to actual needs to simulate various extreme working conditions; accuracy, the digital twin model based on real data has high accuracy, and the virtual calibration results can provide a reliable reference for the design and manufacture of actual vehicles; visualization, by presenting the calibration process and results in an intuitive form, it is easy to understand and analyze. According to the results of the virtual calibration, the design or manufacturing process of the real vehicle is optimized or adjusted.

[0160] The present invention can be widely used in the fields of vehicle design, manufacturing, testing, etc., providing strong support for the development of the automotive industry. At the same time, with the continuous development and improvement of vehicle-cloud collaborative technology, the application prospects of the present invention will be broader.

[0161] In this embodiment, a vehicle calibration device based on digital twins is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and will not be repeated here. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0162] This embodiment provides a vehicle calibration device based on digital twins, such as Fig.10 As shown, including:

[0163] The first processing module 1001 is used to build a digital twin model of the vehicle according to the design data of the vehicle and multi-source data during operation;

[0164] The second processing module 1002 is used to adjust the parameters to be calibrated in the digital twin model, simulate the operation process of the vehicle under different working conditions, and obtain simulation operation results corresponding to different parameters to be calibrated;

[0165] The third processing module 1003 is used to determine the target calibration parameters of the vehicle under different working conditions based on the simulation operation results;

[0166] The fourth processing module 1004 is used to perform vehicle calibration based on the target calibration parameters.

[0167] In some optional implementations, the first processing module 1001 is further used for:

[0168] Extract implicit features from multi-source data;

[0169] Construct performance models and multi-physics coupling models corresponding to each vehicle component / subsystem based on design data, multi-source data and implicit features;

[0170] The performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated and optimized to obtain a digital twin model of the vehicle.

[0171] In some optional implementations, before extracting implicit features of multi-source data, the first processing module 1001 is further used to:

[0172] Perform dimensionality reduction, sparse reconstruction, filtering, data correction and multi-scale feature fusion on multi-source data.

[0173] In some optional implementations, the first processing module 1001 is further used for:

[0174] Cluster multi-source data to obtain grouped data of different categories;

[0175] Extract association rule features and time-frequency domain features from grouped data to obtain implicit features of multi-source data.

[0176] In some optional implementations, the first processing module 1001 is further used for:

[0177] Based on the design data, determine the interaction mechanism and relationship map between each component / component subsystem, and determine the mechanism model and data-driven model corresponding to each component / component subsystem;

[0178] Based on multi-source data and implicit features, determine whether the mechanism model and data-driven model corresponding to each component / component subsystem meet the preset performance requirements, and determine the corresponding modeling method of the current component / component subsystem based on the judgment results;

[0179] According to the interaction mechanism, relationship map, multi-source data and implicit features, and in accordance with the modeling method of each component / component subsystem, the performance model corresponding to each component / component subsystem is constructed.

[0180] In some optional implementations, the preset performance requirement includes a preset accuracy requirement and a preset time constraint; the first processing module 1001 is further used to:

[0181] If the judgment result is that the mechanism model meets the preset accuracy requirements and the preset time constraints, the modeling method is to model according to the mechanism model and multi-source data;

[0182] Or, if the judgment result is that the mechanism model does not meet the preset accuracy requirement and the mechanism model meets the preset time constraint, then the modeling method is to model according to the data-driven model and multi-source data corresponding to the preset time constraint;

[0183] Alternatively, if the judgment result is that both the mechanism model and the data-driven model do not meet the preset accuracy requirements and the preset time constraints, the modeling method is to construct a mechanism-driven hybrid model based on the mechanism model and the data-driven model, and perform modeling based on the mechanism-driven hybrid model and multi-source data.

[0184] In some optional implementations, the first processing module 1001 is further used for:

[0185] Based on the design data, determine the multi-physical field coupling relationship corresponding to each component / component subsystem, and determine the mechanism model, data-driven model and mechanism-driven hybrid model corresponding to each component / component subsystem; wherein the physical field includes at least two of the mechanical field, electric field, magnetic field, flow field and thermal field;

[0186] Perform multi-scale feature decomposition and reconstruction on latent features to obtain unit-scale latent features;

[0187] Based on the implicit features of the unit scale, multi-source data and multi-physical field coupling relationships, and in accordance with the mechanism model, data-driven model or mechanism-driven hybrid model of each component / component subsystem, a multi-physical field coupling model corresponding to each component / component subsystem of the vehicle is constructed.

[0188] In some optional implementations, the first processing module 1001 is further used for:

[0189] Based on the relationship graph, the performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated to obtain an initial system model; wherein the initial system model includes an environment setting model, a driver decision setting model, a calibration parameter setting model, a controller model and a drive mechanism model;

[0190] Feature prediction is performed based on multi-source data to obtain predicted feature parameters; the predicted feature parameters include parameters to be calibrated.

[0191] Based on the predicted characteristic parameters, the initial system model is updated and optimized to obtain the digital twin model of the vehicle.

[0192] In some optional implementations, the first processing module 1001 is further used for:

[0193] Based on the predicted characteristic parameters, the data convergence rate, data forgetting rate and data update rate of the initial system model are adjusted, and the initial characteristic parameters of the initial system model are optimized to obtain a digital twin model.

[0194] In some optional implementations, the third processing module 1003 is further configured to:

[0195] For each operating condition, determine the target simulation operation results that meet the preset calibration requirements of the current operating condition;

[0196] For each operating condition, the target calibration parameters of the vehicle under the current operating condition are obtained according to the parameters to be calibrated corresponding to the target simulation operation results.

[0197] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0198] The digital twin-based vehicle calibration device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above-mentioned functions.

[0199] The embodiment of the present invention also provides a computer device having the above Fig.10 The digital twin-based vehicle calibration device is shown.

[0200] See also Fig.11 , Fig.11 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Fig.11 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig.11 A processor 10 is taken as an example.

[0201] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0202] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0203] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0204] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0205] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Fig.11 The example of connecting through bus is taken in the following.

[0206] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0207] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0208] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0209] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A vehicle calibration method based on digital twins, characterized in that: The method comprises: Build a digital twin model of the vehicle based on the vehicle’s design data and multi-source data during operation; Adjusting the parameters to be calibrated in the digital twin model, simulating the operation process of the vehicle under different working conditions, and obtaining simulation operation results corresponding to different parameters to be calibrated; Based on the simulation operation results, determining target calibration parameters of the vehicle under different working conditions; Vehicle calibration is performed based on the target calibration parameters.

2. The method according to claim 1, characterized in that: The digital twin model of the vehicle is constructed based on the design data of the vehicle and the multi-source data during operation, including: Extracting implicit features of the multi-source data; Constructing performance models and multi-physics coupling models corresponding to various components / subsystems of the vehicle based on the design data, the multi-source data and the implicit features; The performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated and optimized to obtain a digital twin model of the vehicle.

3. The method according to claim 2, characterized in that The constructing the performance model corresponding to each component / subsystem of each component of the vehicle based on the design data, the multi-source data and the implicit features includes: Determine the interaction mechanism and relationship map between the components / component subsystems according to the design data, and determine the mechanism model and data-driven model corresponding to the components / component subsystems; Based on the multi-source data and the implicit features, respectively judging whether the mechanism model and the data-driven model corresponding to each component / component subsystem meet the preset performance requirements, and determining the corresponding modeling method of the current component / component subsystem according to the judgment result; According to the interaction mechanism, the relationship map, the multi-source data and the implicit features, and in accordance with the modeling method of each component / each component subsystem, a performance model corresponding to each component / each component subsystem is constructed.

4. The method according to claim 3, characterized in that The preset performance requirement includes a preset accuracy requirement and a preset time constraint; and the corresponding modeling method of the current component / component subsystem is determined according to the judgment result, including: If the judgment result is that the mechanism model meets the preset accuracy requirement and the preset time constraint, the modeling method is to perform modeling according to the mechanism model and the multi-source data; Or, if the judgment result is that the mechanism model does not meet the preset accuracy requirement and the mechanism model meets the preset time constraint, the modeling method is to perform modeling according to the data-driven model corresponding to the preset time constraint and the multi-source data; Alternatively, if the judgment result is that both the mechanism model and the data-driven model do not meet the preset accuracy requirements and the preset time constraints, the modeling method is to construct a mechanism-driven hybrid model based on the mechanism model and the data-driven model, and perform modeling based on the mechanism-driven hybrid model and the multi-source data.

5. The method according to claim 4, characterized in that The constructing of a multi-physics field coupling model corresponding to each component / subsystem of each component of the vehicle based on the design data, the multi-source data and the implicit features includes: Based on the design data, determine the multi-physical field coupling relationship corresponding to each component / component subsystem, and determine the mechanism model, data-driven model and mechanism-driven hybrid model corresponding to each component / component subsystem; wherein the physical field includes at least two of the mechanical field, electric field, magnetic field, flow field and thermal field; Performing multi-scale feature decomposition and reconstruction on the latent features to obtain unit-scale latent features; Based on the implicit features of the unit scale, the multi-source data and the multi-physical field coupling relationship, and in accordance with the mechanism model or data-driven model or mechanism-driven hybrid model of each component / component subsystem, a multi-physical field coupling model corresponding to each component / component subsystem of the vehicle is constructed.

6. The method according to claim 3, characterized in that The performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated and optimized to obtain a digital twin model of the vehicle, including: Based on the relationship graph, the performance models and multi-physics field coupling models corresponding to each component / component subsystem are integrated to obtain an initial system model; wherein the initial system model includes an environment setting model, a driver decision setting model, a calibration parameter setting model, a controller model and a drive mechanism model; Perform feature prediction based on the multi-source data to obtain predicted feature parameters; the predicted feature parameters include parameters to be calibrated; Based on the predicted characteristic parameters, the initial system model is updated and optimized to obtain a digital twin model of the vehicle.

7. The method according to claim 6, characterized in that The updating and optimizing of the initial system model based on the predicted characteristic parameters to obtain a digital twin model of the vehicle includes: Based on the predicted characteristic parameters, the data convergence rate, data forgetting rate and data update rate of the initial system model are adjusted, and the initial characteristic parameters of the initial system model are optimized to obtain a digital twin model.

8. The method according to claim 2, characterized in that: The extracting implicit features of the multi-source data includes: Clustering the multi-source data to obtain grouped data of different categories; The association rule features and time-frequency domain features in the grouped data are extracted to obtain implicit features of the multi-source data.

9. The method according to claim 8, characterized in that Before extracting the implicit features of the multi-source data, the method further includes: The multi-source data are subjected to dimensionality reduction, sparse reconstruction, filtering, data correction and multi-scale feature fusion.

10. The method according to claim 1, characterized in that Determining target calibration parameters of the vehicle under different working conditions based on the simulation operation results includes: For each operating condition, determine the target simulation operation results that meet the preset calibration requirements of the current operating condition; For each operating condition, target calibration parameters of the vehicle under the current operating condition are obtained according to the parameters to be calibrated corresponding to the target simulation operation result.

11. A vehicle calibration device based on digital twins, characterized in that: The device comprises: A first processing module is used to build a digital twin model of the vehicle based on the design data of the vehicle and multi-source data during operation; The second processing module is used to adjust the parameters to be calibrated in the digital twin model, simulate the operation process of the vehicle under different working conditions, and obtain simulation operation results corresponding to different parameters to be calibrated; A third processing module, used for determining target calibration parameters of the vehicle under different working conditions based on the simulation operation results; The fourth processing module is used to perform vehicle calibration based on the target calibration parameters.

12. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the vehicle calibration method based on digital twins as described in any one of claims 1 to 10 by executing the computer instructions.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the vehicle calibration method based on digital twins according to any one of claims 1 to 10.

14. A computer program product, characterized in that It comprises computer instructions, which are used to enable a computer to execute the vehicle calibration method based on digital twins as described in any one of claims 1 to 10.

Images