Browsing by Author "Yin, Jianhua"

Now showing 1 - 7 of 7
  • Thumbnail Image
    Item
    Active learning with generalized sliced inverse regression for high-dimensional reliability analysis
    (Elsevier, 2022-01) Yin, Jianhua; Du, Xiaoping; Mechanical and Energy Engineering, School of Engineering and Technology
    It is computationally expensive to predict reliability using physical models at the design stage if many random input variables exist. This work introduces a dimension reduction technique based on generalized sliced inverse regression (GSIR) to mitigate the curse of dimensionality. The proposed high dimensional reliability method enables active learning to integrate GSIR, Gaussian Process (GP) modeling, and Importance Sampling (IS), resulting in an accurate reliability prediction at a reduced computational cost. The new method consists of three core steps, 1) identification of the importance sampling region, 2) dimension reduction by GSIR to produce a sufficient predictor, and 3) construction of a GP model for the true response with respect to the sufficient predictor in the reduced-dimension space. High accuracy and efficiency are achieved with active learning that is iteratively executed with the above three steps by adding new training points one by one in the region with a high chance of failure.
  • Thumbnail Image
    Item
    Distributed Nonlinear Model Predictive Control for Heterogeneous Vehicle Platoons Under Uncertainty
    (IEEE Xplore, 2021-09) Shen, Dan; Yin, Jianhua; Du, Xiaoping; Li, Lingxi; Electrical and Computer Engineering, School of Engineering and Technology
    This paper presents a novel distributed nonlinear model predictive control (DNMPC) for minimizing velocity tracking and spacing errors in heterogeneous vehicle platoon under uncertainty. The vehicle longitudinal dynamics and information flow in the platoon are established and analyzed. The algorithm of DNMPC with robustness and reliability considerations at each vehicle (or node) is developed based on the leading vehicle and reference information from nodes in its neighboring set. Together with the physical constraints on the control input, the nonlinear constraints on vehicle longitudinal dynamics, the terminal constraints on states, and the reliability constraints on both input and output, the objective function is defined to optimize the control accuracy and efficiency by penalizing the tracking errors between the predicted outputs and desirable outputs of the same node and neighboring nodes, respectively. Meanwhile, the robust design optimization model also minimizes the expected quality loss which consists of the mean and standard deviation of node inputs and outputs. The simulation results also demonstrate the accuracy and effectiveness of the proposed approach under two different traffic scenarios.
  • Thumbnail Image
    Item
    Distributed Stochastic Model Predictive Control With Taguchi’s Robustness for Vehicle Platooning
    (IEEE, 2022-02-03) Yin, Jianhua; Shen, Dan; Du, Xiaoping; Li, Lingxi; Mechanical and Energy Engineering, School of Engineering and Technology
    Vehicle platooning for highway driving has many benefits, such as lowering fuel consumption, improving traffic safety, and reducing traffic congestion. However, its performance could be undermined due to uncertainty. This work proposes a new control method that combines distributed stochastic model predictive control with Taguchi’s robustness (TR-DSMPC) for vehicle platooning. The proposed method inherits the advantages of both Taguchi’s robustness (maximizing the mean performance and minimizing the performance variation due to uncertainty) and stochastic model predictive control (ensuring a specific reliability level). Taguchi’s robustness is achieved by introducing a variation term in the control objective to bring a trade-off between mean performance and its variation. TR-DSMPC propagates uncertainty via an approximation method: First-Order Second Moment, which is far more efficient than Monte Carlo-based methods. The uncertainty is considered from two perspectives, time-independent uncertainty by random variables and time-dependent uncertainty by stochastic processes. We compare the proposed method with two other MPC-based methods in terms of safety (spacing error) and efficiency (relative velocity). The results indicate that our proposed method can effectively reduce the performance variation and maintain the mean performance.
  • Thumbnail Image
    Item
    High-Dimensional Reliability Method Accounting for Important and Unimportant Input Variables
    (ASME, 2022-04) Yin, Jianhua; Du, Xiaoping; Mechanical and Energy Engineering, School of Engineering and Technology
    Reliability analysis is a core element in engineering design and can be performed with physical models (limit-state functions). Reliability analysis becomes computationally expensive when the dimensionality of input random variables is high. This work develops a high-dimensional reliability analysis method through a new dimension reduction strategy so that the contributions of unimportant input variables are also accommodated after dimension reduction. Dimension reduction is performed with the first iteration of the first-order reliability method (FORM), which identifies important and unimportant input variables. Then a higher order reliability analysis is performed in the reduced space of only important input variables. The reliability obtained in the reduced space is then integrated with the contributions of unimportant input variables, resulting in the final reliability prediction that accounts for both types of input variables. Consequently, the new reliability method is more accurate than the traditional method which fixes unimportant input variables at their means. The accuracy is demonstrated by three examples.
  • Thumbnail Image
    Item
    Label Free Uncertainty Quantification
    (ARC, 2022-01) Li, Huiru; Yin, Jianhua; Du, Xiaoping; Mechanical Engineering, School of Engineering and Technology
    View Video Presentation: https://doi.org/10.2514/6.2022-1097.vid Uncertainty quantification (UQ) is essential in scientific computation since it can provide the estimate of the uncertainty in the model prediction. Intensive computation is required for UQ as it calls the deterministic simulation repeatedly. This study discusses a physics-based label-free deep learning UQ method that does not need predictions at training points or labels. It satisfies the physical equations from which labels could be generated without solving the equations during the training process. Then inexpensive surrogate models are built with respect to model inputs. The surrogate models are used for UQ with a much lower computational cost. Two examples demonstrate that the label-free method can efficiently produce probability distributions of model outputs for given distributions of random input variables.
  • Thumbnail Image
    Item
    A Safety Factor Method for Reliability-Based Component Design
    (American Society of Mechanical Engineers, 2021-09) Yin, Jianhua; Du, Xiaoping; Mechanical and Energy Engineering, School of Engineering and Technology
    Reliability-based design (RBD) employs optimization to identify design variables that satisfy the reliability requirement. For many routine component design jobs that do not need optimization, however, RBD may not be applicable, especially for those design jobs which are performed manually or with a spreadsheet. This work develops a modified RBD approach to component design so that the reliability target can be achieved by conducting traditional component design repeatedly using a deterministic safety factor. The new component design is based on the first-order reliability method (FORM), which iteratively assigns the safety factor during the design process until the reliability requirement is satisfied. In addition to several iterations of deterministic component design, the other additional work is the calculation of the derivatives of the design margin with respect to the random input variables. The proposed method can be used for a wide range of component design applications. For example, if a deterministic component design is performed manually or with a spreadsheet, so is the reliability-based component design. Three examples are used to demonstrate the practicality of the new design method.
  • Thumbnail Image
    Item
    Uncertainty Quantification by Convolutional Neural Network Gaussian Process Regression with Image and Numerical Data
    (AIAA, 2022-01) Yin, Jianhua; Du, Xiaoping; Mechanical and Energy Engineering, School of Engineering and Technology
    Uncertainty Quantification (UQ) plays a critical role in engineering analysis and design. Regression is commonly employed to construct surrogate models to replace expensive simulation models for UQ. Classical regression methods suffer from the curse of dimensionality, especially when image data and numerical data coexist, which makes UQ computationally unaffordable. In this work, we propose a Convolutional Neural Network (CNN) based framework, which accommodates both image and numerical data. We first transform numerical data into images and then combine them with existing image data. The combined images are fed to CNN for regression. To obtain the model uncertainty, we integrate CNN with Gaussian Process (GP), which results in the mixed network CNN-GP. The simulation results show that CNN-GP can build accurate surrogate models for UQ with mixed data and that CNN-GP can also provide the uncertainty associated with the model prediction.