Accelerated multi-physical evaluation for air bag design

Abstract

A method for modeling an air bag design includes training a reduced order parametric model (104) of an air bag system to evaluate different air bag designs. Training the model includes applying a higher order multi-physics model (102) to the air bladder system in the design space to produce a performance snapshot (103) that varies over time. The reduced order model (104) constructs a reduced basis (105) from the higher order snapshot (103) and projects a set of modal coefficients (107). The regression model (108) is fitted to produce a parametric model (109). The parameter model is established and trained in an offline process, and the evaluation of the air bag design is completed online. The reduced order model may be created using an eigenorthogonal decomposition greedy algorithm, and the regression model may be implemented in a machine learning model or Gaussian process. The airbag design is evaluated using parameters of the parametric model that describe the airbag or the environment in which the airbag is operating.

Description

Technical Field

[0001] The present application relates to the design of manufactured objects. Background Art

[0002] In the design of some objects, such as motor vehicle air bags, reliable simulation of air bag inflation requires the combination of multiple physics, such as large displacement structural dynamics, fluid dynamics, and multibody dynamics. Numerical methods exist to simulate each of these physics. Finite element methods are widely used for structures, including large deformation scenarios required for air bag inflation. Fluid dynamics problems are usually solved by computational fluid dynamics (CFD), which has been successfully verified in a variety of applications including container gas filling. In the field of multibody simulation, time advancement schemes (such as modified Euler and Runge-Kutta) are usually used to evaluate solutions. However, when the interaction between air bags and dynamic physical environments is taken into account, the complexity is greatly increased, and the simulation time for analyzing a given air bag inflation in a given motor vehicle and a given passenger is very long due to the required number of physics and their coupling. As non-limiting examples, possible dynamic physical environments include, but are not limited to, the interior of a motor vehicle, a passenger wearing a seat belt, a pedestrian on the road, or a cyclist with a protective system. Summary of the invention

[0003] According to the embodiments described in the present disclosure, a method for modeling an air bag design includes: for a given air bag design, applying a high-order multi-physics model to generate a snapshot of the air bag performance over time. Then, in a reduced-order model, an air bag basis is constructed from the high-order snapshot, and the reduced-order output and the snapshot are projected to generate a set of modal coefficients for the air bag basis. The fitted regression model is evaluated to generate a parametric model of the air bag. The parametric model is then used to evaluate the air bag design, thereby saving time and allowing more designs to be evaluated. The parametric model of the air bag is built and trained in an offline process, while the evaluation of the air bag design in the parametric model is completed in an online process. In the air bag design process, design decisions are made based on the evaluation results. As non-limiting examples, the reduced-order model can be created using proper orthogonal decomposition, greedy algorithms, or nonlinear manifold learning. As non-limiting examples, the regression model can be implemented in a machine learning model or a Gaussian process. Evaluating the air bag design can be performed by inputting parameters describing one of the multiple air bags into the parametric model. According to some embodiments, evaluating the air bag design can be performed by inputting parameters describing the environment in which the air bag operates into the parametric model. The air bag design for training the parametric model includes selecting an air bag system from a target design space. Evaluating the air bag system using the parametric model includes generating a full order response in the parameter space using back projection.

[0004] A system for modeling an air bag design, comprising: a computer processor; and a non-transitory computer memory in communication with the computer processor, the non-transitory computer memory storing instructions that, when executed by the computer processor, cause the computer processor to perform the following steps: for a given air bag design, applying a high-order multi-physics model to generate a snapshot of the air bag performance over time; in a reduced-order model, constructing an air bag basis from the high-order snapshot; projecting the snapshot onto a reduced basis to generate a set of modal coefficients for the air bag basis; performing a fitted regression model to generate a parametric model of the reduced-order coefficients of the air bag; and evaluating the air bag design using the parametric model and the reduced-order basis. The generation of the parametric model of the air bag can be performed offline. The evaluation of the air bag design can be done in an online process. Based on the evaluation results using the parametric model, a design alternative can be selected. Reduced-order modeling can be performed using proper orthogonal decomposition, by using a greedy algorithm or nonlinear manifold learning. By way of example, regression modeling can include a machine learning model or a Gaussian process. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] The above and other aspects of the present invention can be best understood from the following detailed description when read in conjunction with the accompanying drawings. For the purpose of illustrating the present invention, presently preferred embodiments are shown in the accompanying drawings, however, it should be understood that the present invention is not limited to the specific means disclosed. The following figures are included in the accompanying drawings:

[0006] Figure 1 is a block diagram of an offline training process for creating a parametric model from a reduced order model according to aspects of an embodiment of the present disclosure.

[0007] Figure 2 is an illustration of online evaluation of a reduced-order parameter model according to aspects of an embodiment of the present disclosure.

[0008] Figure 3 is a block diagram of a computer system that can be used to implement a reduced-order parameter model for a design in accordance with aspects of an embodiment of the present disclosure. DETAILED DESCRIPTION

[0009] The complexity involved in multi-physics air bag inflation simulations causes the air bag design process to be very time consuming. For example, given the design requirements, many air bag variations must be analyzed and simulated to arrive at the most appropriate air bag. Similarly, automotive design involves the selection and placement of air bags, which is also a time consuming process for similar reasons. The computational demands of such simulations limit the design space that can be explored, resulting in a lower number and more robust designs than would otherwise be possible within the time allowed.

[0010] Current computationally expensive approaches to solving air bag inflation hinder product quality and extend time to market for several automotive and / or aerospace stakeholders, such as air bag designers, mobile original equipment manufacturer (OEM) designers, and testers. The embodiments described in this disclosure are intended to provide air bag simulation technology that reduces time to market and improves product quality and robustness for air bag and mobile OEM designers.

[0011] The current state-of-the-art approach to simulating air bag inflation is to couple the solvers of the involved multiphysics. The physics are simulated separately over a certain time span and coupled together at a given time to share information and boundary conditions so that they converge and remain consistent. Depending on how often the multiphysics are coupled, the solvers are called loosely coupled or tightly coupled. In the former, the solvers are coupled and communicate with each other less frequently than in the latter.

[0012] A traditional solution for airbag design allows designers to model and simulate restraint systems such as airbags, seatbelts, passengers, and other vehicle interior components relevant to vehicle crashes. This solution incorporates multibody, finite element analysis (FEA), and computational fluid dynamics (CFD) into the solver to balance solver speed and modeling detail. However, simulation time still hinders airbag and vehicle time to market because many simulations need to be run when designers evaluate the robustness of design changes or optimizations. Despite the state-of-the-art solution balance between speed and detail, this still leads to long design times. Based on multibody methods, these solvers are much faster than most FEA solutions alone. However, depending on the airbag design case, and when considering finite element (FE) airbags with CFD interaction, the computation time slows down significantly, resulting in a reduction or loss of time advantage. Therefore, alternative methods such as deep neural networks (DNNs) that can provide sufficient accuracy and robustness while maintaining or increasing the time advantage are critical to maintaining a competitive advantage.

[0013] Active research focuses on solver coupling acceleration, which is invasive (requiring source code access and even source code modification). These solutions have limited simulation time reductions since each related physics still uses a high-fidelity and high-order solver. Additionally, they are not easy to use because these multiphysics codes are complex and often convoluted.

[0014] Embodiments of the present disclosure propose a non-intrusive workflow that combines model order reduction and machine learning (ML) to create a parametric model that accelerates the simulation of dynamic multi-physics airbag inflation for dynamic physical environments (e.g., passengers, vehicle / train / airplane interiors, seat belts, cyclists, etc.). The parametric nature of the model allows the ML model to evaluate multiple airbags and / or interactive changes with humans and the physical environment, enabling rapid simulation of the entire design space defined by these model parameters.

[0015] The proposed workflow consists of two parts: an offline process, in which the model Figure 1 and an online process in which the trained model is used to evaluate and design air bags and / or inflation for multiple objects or agents embedded in a physical environment, such as Figure 2 As shown. During the training phase, a state-of-the-art multi-physics estimator 102 is used to evaluate multiple air bag systems in a desired parameter space 101. The results or snapshots 103 are used to estimate system modes or bases 105 via any relevant order reduction method 104, including but not limited to proper orthogonal decomposition (POD), greedy algorithms, or manifold learning. Once the modes 105 are available, the modal coefficients 107 of the training snapshots are obtained by projecting 106 the snapshots onto the modes 105. Finally, a regressor 108 is constructed that maps the air bag system parameters and time to the modal coefficients. Potential regressors 108 include but are not limited to neural networks, Gaussian processes, or response surfaces.

[0016] During the online phase, the designer or optimization engineer evaluates the airbag system performance in the parametric model by evaluating the reduced order model 109 and subsequently obtaining the full order response through a computationally inexpensive back-projection step. Due to their computational simplicity, mode expansion and regression reasoning, these online evaluations are much faster than state-of-the-art multiphysics-based simulations.

[0017] An important improvement of the described embodiments is the time for system evaluation, which is much shorter than conventional techniques. Due to the reduced complexity of the ML-enhanced model, the airbag system evaluation is greatly shortened and the evaluation speed can be up to four orders of magnitude faster. For example, compared to the embodiments described in this invention, performing an airbag simulation using a multiphysics solver may take 2500 seconds, while a simulation using a reduced-order parameter model only takes 0.2 seconds, both using the same computing hardware.

[0018] Additionally, the parametric nature of the solution makes the ML-enhanced model effective over a wide design space of airbag systems. That is, the same ML-enhanced model can be used to evaluate multiple airbag systems within the design space defined by the training snapshots.

[0019] Combining the time advantage with the parametric nature of the model enables faster design, optimization, and robustness checking. Thus, time to market is reduced, and better, more robust system performance is achieved for airbag and vehicle designers. Furthermore, since the estimation of modes and their coefficients is done through model reduction and regression, the ML-enhanced solution is non-intrusive. It should be noted that neither technique requires changes to the underlying airbag inflation simulation code.

[0020] Figure 2 An online process for evaluating an air bag and its environment in a reduced order parametric model is shown in accordance with an embodiment of the present disclosure. The machine learning enhanced parametric model 109 is fed by the parameters 202 of the air bag being evaluated to produce snapshot base coefficients which, together with the air bag base 105, produce an air bag solution 203. The air bag parameters may represent properties of the air bag itself, or may include values ​​representing objects in the environment in which the air bag operates. The machine learning enhanced parametric model 109 performs the evaluation 201 together with the air bag base 105 to produce an air bag solution 203. Due to the computational simplicity of the parametric model 109, designers can perform more evaluations within a given time budget than traditional solutions that rely on physics solvers. This allows for a more comprehensive exploration of the design space to ensure better solutions in less time, thereby improving results while reducing time to market for higher quality products.

[0021] Figure 3 An exemplary computing environment 300 is shown within which embodiments of the present invention may be implemented. Computers and computing environments, such as computer system 310 and computing environment 300, are known to those skilled in the art and, therefore, are briefly described herein.

[0022] like Figure 3 As shown, computer system 310 may include a communication mechanism such as a system bus 321 or other communication mechanism for transferring information within computer system 310. Computer system 310 also includes one or more processors 320 coupled to system bus 321 for processing information.

[0023] The processor 320 may include one or more central processing units (CPUs), graphics processing units (GPUs), or any other processors known in the art. More generally, a processor as used in the present invention is a device for executing machine-readable instructions stored on a computer-readable medium for performing tasks, and may include any one of hardware and firmware or a combination of hardware and firmware. The processor may also include a memory storing executable machine-readable instructions for performing tasks. The processor acts on information by manipulating, analyzing, modifying, converting, or sending information for use by an executable program or information device, and / or by routing information to an output device. The processor may use or include the capabilities of, for example, a computer, a controller, or a microprocessor, and may be adjusted using executable instructions to perform special functions that are not performed by a general-purpose computer. The processor may be coupled to any other processor (electrically and / or include executable components) to enable interaction and / or communication therebetween. A user interface processor or generator is a known element including an electronic circuit system or software or a combination of both for generating a display image or part thereof. The user interface includes one or more display images to enable interaction between a user and a processor or other device.

[0024] Continue to refer Figure 3 , the computer system 310 also includes a system memory 330 coupled to the system bus 321 for storing information and instructions to be executed by the processor 320. The system memory 330 may include computer-readable storage media in the form of volatile and / or non-volatile memory such as read-only memory (ROM) 331 and / or random access memory (RAM) 332. RAM 332 may include other dynamic storage devices (e.g., dynamic RAM, static RAM, and synchronous DRAM). ROM 331 may include other static storage devices (e.g., programmable ROM, erasable PROM, and electrically erasable PROM). In addition, the system memory 330 may be used to store temporary variables or other intermediate information during the execution of instructions by the processor 320. A basic input / output system 333 (BIOS) containing basic routines that help transfer information between elements within the computer system 310 (such as during startup) may be stored in ROM 331. RAM 332 may contain data and / or program modules that are immediately accessible and / or currently being operated by the processor 320. System memory 330 may additionally include, for example, operating system 334 , application programs 335 , other program modules 336 , and program data 337 .

[0025] The computer system 310 also includes a disk controller 340 coupled to the system bus 321 to control one or more storage devices for storing information and instructions, such as a magnetic hard disk 341 and a removable media drive 342 (e.g., a floppy disk drive, an optical drive, a tape drive, and / or a solid state drive). Storage devices can be added to the computer system 310 using an appropriate device interface (e.g., small computer system interface (SCSI), integrated device electronics (IDE), universal serial bus (USB), or FireWire).

[0026] The computer system 310 may also include a display controller 365 coupled to the system bus 321 to control a display or monitor 366, such as a cathode ray tube (CRT) or a liquid crystal display (LCD), for displaying information to a computer user. The computer system includes an input interface 360 ​​and one or more input devices, such as a keyboard 362 and a pointing device 361, for interacting with a computer user and providing information to the processor 320. For example, the pointing device 361 may be a mouse, a light pen, a trackball, or a pointing stick for communicating directional information and command selections to the processor 320 and for controlling cursor movement on the display 366. The display 366 may provide a touch screen interface that allows input to supplement or replace the communication of directional information and command selections by the pointing device 361. In some embodiments, an augmented reality device 367 that a user can wear may provide input / output functionality that allows the user to interact with both the physical world and the virtual world. The augmented reality device 367 communicates with the display controller 365 and the user input interface 360, allowing the user to interact with virtual items generated by the display controller 365 in the augmented reality device 367. The user can also provide gestures, which are detected by the augmented reality device 367 and sent to the user input interface 360 ​​as input signals.

[0027] The computer system 310 may perform some or all of the processing steps of the processing steps of an embodiment of the present invention in response to the processor 320 executing one or more sequences of one or more instructions contained in a memory, such as a system memory 330. Such instructions may be read into the system memory 330 from another computer-readable medium, such as a magnetic hard disk 341 or a removable media drive 342. The magnetic hard disk 341 may contain one or more data repositories and data files used by embodiments of the present invention. The data repository contents and data files may be encrypted to increase security. The processor 320 may also be employed in a multi-processing arrangement to execute one or more sequences of instructions contained in the system memory 330. In alternative embodiments, hardwired circuitry may be used in place of or in combination with software instructions. Therefore, the embodiments are not limited to any particular combination of hardware circuitry and software.

[0028] As described above, the computer system 310 may include at least one computer-readable medium or memory for storing instructions programmed according to embodiments of the present invention and for containing data structures, tables, records, or other data described herein. As used herein, the term "computer-readable medium" refers to any medium that participates in providing instructions to the processor 320 for execution. Computer-readable media can take many forms, including but not limited to non-transient media, non-volatile media, volatile media, and transmission media. Non-limiting examples of non-volatile media include optical disks, solid-state drives, magnetic disks, and magneto-optical disks, such as magnetic hard disks 341 or removable media drives 342. Non-limiting examples of volatile media include dynamic memory, such as system memory 330. Non-limiting examples of transmission media include coaxial cables, copper wires, and optical fibers, including wires that constitute the system bus 321. Transmission media can also take the form of sound waves or light waves, such as those generated during radio wave and infrared data communications.

[0029] The computing environment 300 may also include a computer system 310 that operates in a networked environment using a logical connection to one or more remote computers (such as a remote computing device 380). The remote computing device 380 may be a personal computer (laptop or desktop), a mobile device, a server, a router, a network PC, a peer device, or other common network node, and typically includes many or all of the elements described above for the computer system 310. When used in a networked environment, the computer system 310 may include a modem 372 for establishing communications over a network 371 (such as the Internet). The modem 372 may be connected to the system bus 321 via a user network interface 370 or via another appropriate mechanism.

[0030] The network 371 can be any network or system generally known in the art, including the Internet, an intranet, a local area network (LAN), a wide area network (WAN), a metropolitan area network (MAN), a direct connection or a series of connections, a cellular telephone network, or any other network or medium that can facilitate communication between the computer system 310 and other computers (e.g., a remote computing device 380). The network 371 can be wired, wireless, or a combination thereof. The wired connection can be implemented using Ethernet, a universal serial bus (USB), RJ-6, or any other wired connection generally known in the art. The wireless connection can be implemented using Wi-Fi, WiMAX, and Bluetooth, infrared, cellular networks, satellites, or any other wireless connection methodology generally known in the art. In addition, several networks can work alone or communicate with each other to facilitate communication in the network 371.

[0031] An executable application as used in the present invention includes code or machine-readable instructions for adjusting a processor to implement predetermined functions (such as those of an operating system, a contextual data acquisition system, or other information processing system), for example, in response to user commands or inputs. An executable program is a fragment of code or machine-readable instructions for performing one or more specific processes, a subroutine, or other different sections of code or a part of an executable application. These processes may include receiving input data and / or parameters, performing operations on received input data, and / or performing functions in response to received input parameters, and providing resulting output data and / or parameters.

[0032] A graphical user interface (GUI) as used in the present invention includes one or more display images generated by a display processor and realizing the interaction between a user and a processor or other device and associated data acquisition and processing functions. The GUI also includes an executable program or an executable application. The executable program or executable application adjusts the display processor to generate signals representing GUI display images. These signals are provided to a display device, which displays images for viewing by the user. Under the control of the executable program or executable application, the processor manipulates the GUI display image in response to signals received from an input device. In this way, the user can use the input device to interact with the display image, thereby realizing the interaction between the user and the processor or other device.

[0033] The functions and process steps of the present invention may be automatically executed or executed in whole or in part in response to user commands. Automatically executed activities (including steps) are executed in response to one or more executable instructions or device operations without direct initiation of the activity by the user.

[0034] The systems and processes shown in the figure are not exclusive. Other systems, processes and menus can be derived according to the principles of the present invention to achieve the same goal. Although the present invention has been described with reference to specific embodiments, it will be understood that the embodiments and variations shown and described in the present invention are only for illustrative purposes. Those skilled in the art can realize the modification of the current design without departing from the scope of the present invention. As described in the present invention, various systems, subsystems, agents, managers and processes can be implemented using hardware components, software components and / or their combinations.

Claims

1. A multi-physics modeling approach for air bag design, include: For a given air bag design and the environment in which said air bag operates, a high-order multiphysics model is applied to generate a snapshot of the air bag performance over time; In the reduced-order model, an air pocket reduction basis is constructed from the high-order multiphysics snapshot; projecting the snapshot onto the air bag reduced basis to produce a set of modal coefficients for an air bag basis; performing a fitting regression model to generate a parameter model of the air bag; and The parametric model was used to evaluate air bag designs.

2. The method according to claim 1, in, The generation of the parametric model of the air bag is performed off-line.

3. The method according to claim 1, in, Evaluation of the air bag design is done in an online process.

4. The method according to claim 1, further comprising: include: Design decisions are performed based on results of the evaluation using the parametric model.

5. The method according to claim 1, in, The reduced order model is performed using proper orthogonal decomposition.

6. The method according to claim 1, in, The reduced order model is performed using a greedy algorithm.

7. The method according to claim 1, in, The reduced order model is performed using nonlinear manifold learning.

8. The method according to claim 1, in, The regression model includes a machine learning model.

9. The method according to claim 1, in, The regression model is a Gaussian process.

10. The method according to claim 1, in, Evaluating the air bag design includes inputting parameters describing an air bag of the plurality of air bags into the parametric model.

11. The method according to claim 1, in, Evaluating an air bag design includes inputting into the parametric model parameters describing the environment in which the air bag operates.

12. The method according to claim 1, in, The air bag design includes selecting an air bag system from a target design space.

13. The method according to claim 1, further comprising: include: Use backprojection to generate full-order responses in the parameterization.

14. A system for modeling air bag design, include: Computer processors; as well as a non-transitory computer memory in communication with the computer processor, the non-transitory computer memory storing instructions that, when executed by the computer processor, cause the computer processor to perform the following steps: For each airbag design in the training set and the environment in which said airbag operates, applying a high-order multiphysics model to generate a snapshot of the airbag performance over time; In the reduced-order model, an air pocket reduction basis is constructed from high-order snapshots; projecting the snapshot onto the air bag reduced basis to produce a set of modal coefficients for an air bag basis; performing a fitting regression model to generate a parameter model of the air bag; as well as The parametric model is used to evaluate air bag designs for unseen air bag designs or environments in which the air bag operates (that are not part of the training set).

15. The system according to claim 14, in, The generation of the parametric model of the air bag is performed off-line.

16. The system according to claim 14, in, Evaluation of the air bag design is done in an online process.

17. The system of claim 14, the non-transitory computer memory further storing instructions that, when executed by the computer processor, cause the computer processor to perform the steps of: performing design decisions based on results of the evaluation using the parametric model.

18. The system according to claim 14, in, The reduced order model is performed using proper orthogonal decomposition.

19. The system according to claim 14, in, The reduced order model is performed using a greedy algorithm.

20. The system according to claim 14, in, The reduced order model is performed using nonlinear manifold learning.

21. The system according to claim 14, in, The regression model includes a machine learning model.

22. The system according to claim 14, in, The regression model is a Gaussian process.

Images