Article
Computer Science, Interdisciplinary Applications
Xiaobing Shang, Zhi Zhang, Hai Fang, Lichao Jiang, Lipeng Wang
Summary: Global sensitivity analysis (GSA) is commonly used to explore the contribution of input variables to the model output and identify the most important variables. However, it often requires a large number of model evaluations, which can be computationally expensive. This paper proposes an efficient Sobol index estimator called PCEGPR, which combines polynomial chaos expansion (PCE) and Gaussian process regression (GPR). The proposed method incorporates PCE into the GPR surrogate model and derives analytical expressions for the main and total sensitivity indices. The effectiveness of the proposed estimator is demonstrated through numerical examples.
ENGINEERING WITH COMPUTERS
(2023)
Article
Chemistry, Multidisciplinary
Xuejun Liu, Hailong Tang, Xin Zhang, Min Chen
Summary: A fast, accurate, and robust uncertainty quantification method is proposed in this paper to investigate the impact of component performance uncertainty on the performance of a classical turboshaft engine. The method utilizes Gaussian process model and Latin hypercube sampling to accurately approximate the relationships between inputs and outputs of the engine performance simulation model. The study explores two main scenarios where uncertain parameters are considered to be mutually independent and partially correlated, respectively, providing new insights into engine performance uncertainty and important component performance parameters.
APPLIED SCIENCES-BASEL
(2021)
Article
Management
Wei Xie, Russell R. Barton, Barry L. Nelson, Keqi Wang
Summary: Motivated by challenges in biopharmaceutical manufacturing, this paper proposes a metamodel-assisted stochastic simulation uncertainty analysis framework to accelerate the development of flexible production processes. The framework considers both simulation and model uncertainties and produces a confidence interval using a metamodel-assisted bootstrapping approach. The relative contributions of model uncertainty and simulation uncertainty are estimated through variance decomposition, allowing for improved system performance estimation.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2023)
Article
Astronomy & Astrophysics
Marc Tunnell, Nathaniel Bowman, Erin Carrier
Summary: The NASA Ames Mars Global Climate Model (MGCM) is a computer model that simulates the weather on Mars. The MGCM is used by NASA to help understand weather data collected from satellites and other sources. To address the computational challenge of sensitivity studies, a surrogate model using Gaussian processes (GP) has been developed. This surrogate model can accurately and quickly approximate the output of the MGCM with a relatively small amount of training data.
EARTH AND SPACE SCIENCE
(2023)
Article
Engineering, Geological
Hu Zhao, Florian Amann, Julia Kowalski
Summary: The integration of Gaussian process emulation into landslide run-out modeling was shown to be feasible and efficient based on the 2017 Bondo landslide event. First-order effects were consistent with common one-at-a-time sensitivity analyses, while the approach also allows for a rigorous investigation of interactions.
Article
Engineering, Electrical & Electronic
Ketian Ye, Junbo Zhao, Fei Ding, Rui Yang, Xiao Chen, George W. Dobbins
Summary: This paper proposes a data-driven GSA method for large-scale distribution systems, using deep Gaussian process to identify the mapping relationship between uncertain power injections and voltages, with much better scalability.
IEEE TRANSACTIONS ON POWER SYSTEMS
(2021)
Article
Green & Sustainable Science & Technology
Qi Liu, Jing Yang, Lei Gao, Yucheng Dong, Zhaoxia Guo, Enayat A. Moallemi, Sibel Eker, Michael Obersteiner
Summary: This study conducted robust sensitivity analyses of scenario parameters in a complex socio-ecological system model and identified sensitive parameters that remain sensitive under five representative scenarios. The findings make it easier to improve and apply the model by focusing on key parameters and understanding important sector linkages in socio-ecological systems.
JOURNAL OF CLEANER PRODUCTION
(2023)
Article
Environmental Sciences
A. Trucchia, L. Frunzo
Summary: In this study, Global Sensitivity Analysis (GSA) and Uncertainty Quantification (UQ) were conducted on an ADM1-based Anaerobic Digestion Model focused on municipal solid waste digestion. The model introduced a surface-based kinetic approach to better model the disintegration step of complex organic matter. GSA and UQ were essential for further improvements and a deeper understanding of the main processes and input factors.
JOURNAL OF ENVIRONMENTAL MANAGEMENT
(2021)
Article
Mathematics, Applied
Yisi Liu, Xiaojun Wang, Yunlong Li
Summary: This paper introduces an improved Bayesian collocation method for steady-state response analysis of structural dynamic systems, which uses a bidirectional global optimization process and Gaussian process surrogate model to improve efficiency and accuracy of nonlinear interval analysis.
APPLIED MATHEMATICS AND COMPUTATION
(2021)
Article
Engineering, Electrical & Electronic
Xiaobing Liao, Min Zhang, Jian Le, Lina Zhang, Zicheng Li
Summary: This paper studies a global sensitivity analysis method of static voltage stability based on an extended affine model. The analysis of simulation results shows that this method can effectively suppress the effect of interval expansion and accurately determine the importance of input interval variables.
ELECTRIC POWER SYSTEMS RESEARCH
(2022)
Article
Engineering, Chemical
Aaron S. Yeardley, Stefan Bellinghausen, Robert A. Milton, James D. Litster, Solomon F. Brown
Summary: Model-driven design requires efficient workflows for model calibration by identifying critical process parameters and impactful modelling parameters, followed by targeted experimental campaigns. Conducting a global sensitivity analysis is essential to identify these parameters and reduce computational effort. The use of Gaussian Process surrogate method in this study significantly decreases the input space and improves the ability to determine impactful parameter values.
CHEMICAL ENGINEERING SCIENCE
(2021)
Article
Environmental Sciences
Stephanie Aparicio, Rebecca Serna-Garcia, Aurora Seco, Jose Ferrer, Luis Borras-Falomir, Angel Robles
Summary: The study assessed a microalgae model applied to a Membrane Photobioreactor (MPBR) pilot plant through a global sensitivity and uncertainty analysis. The influential factors of the model were identified, calibrated offline or online, and the model's uncertainty was evaluated. The results indicated a need for offline calibration methods to improve model performance.
SCIENCE OF THE TOTAL ENVIRONMENT
(2022)
Article
Energy & Fuels
Sinan Xiao, Timothy Praditia, Sergey Oladyshkin, Wolfgang Nowak
Summary: Global sensitivity analysis is conducted to identify the impact of uncertain parameters on the outputs of a thermochemical energy storage model, aiming to better understand predictive uncertainties and streamline uncertainty quantification efforts.
Article
Optics
Huanhuan Qin, Jingyuan Liang, Xizheng Ke
Summary: This paper proposes a noise analysis method based on Allan variance to investigate the noise components in Vehicular visible light communications (VVLC) systems. The results show the presence of white noise and random walk noise in VVLC systems, and the system performance degrades significantly in the presence of Gaussian mixture (GM) noise.
Article
Engineering, Civil
Baixi Chen, Luming Shen, Hao Zhang
Summary: In this paper, a data-driven stochastic constitutive model based on Gaussian process regression (GPR) is proposed, which captures the constitutive relationship of materials along with their uncertainties. The GPR model shows higher accuracy compared to other data-driven approaches, especially when dealing with small datasets.
ASCE-ASME JOURNAL OF RISK AND UNCERTAINTY IN ENGINEERING SYSTEMS PART A-CIVIL ENGINEERING
(2021)
Article
Engineering, Civil
Lei Su, Hua-Ping Wan, You Dong, Dan M. Frangopol, Xian-Zhang Ling
Summary: The scenario-based seismic assessment approach is demonstrated in a large-scale pile-supported wharf structure, presenting a novel and efficient method to enhance traditional simulation methods by utilizing a Gaussian Process surrogate model. This model replaces the time-consuming FE model of the wharf structure, making the quantification of uncertainty in seismic response more computationally-efficient. The feasibility of the approach is verified through Monte Carlo simulation in assessing seismic performance of the wharf structure under a given seismic scenario.
JOURNAL OF EARTHQUAKE ENGINEERING
(2021)
Article
Construction & Building Technology
Yang Li, Yaozhi Luo, Hua-Ping Wan, Chung-Bang Yun, Yanbin Shen
STRUCTURAL CONTROL & HEALTH MONITORING
(2020)
Article
Engineering, Mechanical
Hua-Ping Wan, Yi-Qing Ni
JOURNAL OF ENGINEERING MECHANICS
(2020)
Article
Construction & Building Technology
Hua-Ping Wan, Yanfeng Zheng, Yaozhi Luo, Chao Yang, Xian Xu
Summary: The rotational stability of the central detector at the Jiangmen Underground Neutrino Observatory, influenced by uncertain model parameters, was studied using polynomial chaos expansion surrogate model for global sensitivity analysis. The results effectively revealed the sensitivity of the structure to uncertain parameters.
ADVANCES IN STRUCTURAL ENGINEERING
(2021)
Article
Engineering, Multidisciplinary
Yi-Ming Zhang, Hao Wang, Hua-Ping Wan, Jian-Xiao Mao, Yi-Chao Xu
Summary: The study presents an online approach for detecting anomalies in structural health monitoring data based on Bayesian dynamic linear model, utilizing EM algorithm and Kalman smoother, along with subspace identification method to overcome initialization issues, showing good accuracy and efficiency in both simulation and real-world data applications.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2021)
Article
Engineering, Civil
Hua-Ping Wan, Lei Su, Dan M. Frangopol, Zhiwang Chang, Wei-Xin Ren, Xianzhang Ling
Summary: This study focuses on assessing the effects of acceleration-pulse ground motions on the seismic response of a pile-supported bridge at a canyon site. The results show that acceleration-pulse ground motions had a more significant impact on the seismic response of the ground-bridge system compared with non-acceleration-pulse ground motions. Special care should be taken on the acceleration-pulse effect of ground motion, as bridges are more vulnerable to damage when subjected to acceleration-pulse ground motions.
JOURNAL OF BRIDGE ENGINEERING
(2021)
Article
Engineering, Mechanical
Hua-Ping Wan, Guan-Sen Dong, Yaozhi Luo
Summary: This paper proposes a new method to construct a dedicated dictionary for wind speed signals using the time-shift strategy, improving the performance of compressive sensing (CS) methodology in wind monitoring. The results demonstrate that the improved CS methodology outperforms the traditional CS algorithm and explores the influences of critical parameters comprehensively.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Engineering, Civil
Ning-Bo Wang, Wei Shen, Chuanrui Guo, Hua-Ping Wan
Summary: This study proposes an influence line-based moving load test method for rapid evaluation of the load capacities of aged bridges. By extracting the influence line and defining capacity evaluation indices, the accuracy and effectiveness of the method are validated.
ENGINEERING STRUCTURES
(2022)
Article
Engineering, Civil
Yu Xue, Yaozhi Luo, Xian Xu, Hua-Ping Wan, Yanbin Shen
Summary: The paper proposes a robust approach for pre-stress adjustment of cable-strut structures based on sparse regression. By adjusting the lengths of structural members, the method significantly decreases pre-stress errors and is robust to noise. Experiments demonstrate its effectiveness in reducing errors by adjusting a small number of active elements.
ENGINEERING STRUCTURES
(2021)
Article
Engineering, Mechanical
Hua-Ping Wan, Guan-Sen Dong, Yaozhi Luo, Yi-Qing Ni
Summary: This paper explores an improved method for compressing and reconstructing structural health monitoring data using compressive sensing and multi-task learning techniques, aiming to enhance computational efficiency and accuracy by exploiting data sparsity in the frequency domain.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Engineering, Multidisciplinary
Zhi Ma, Yaozhi Luo, Chung-Bang Yun, Hua-Ping Wan, Yanbin Shen
Summary: This paper presents an improved approach using a mixture of probabilistic principal component analysis (MPPCA) for the anomaly detection of structures under multiple operational conditions with missing measurement data. The effectiveness of the MPPCA-based method was investigated by applying the method to the anomaly detection of a retractable roof structure with numerically simulated and real monitored stress data.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2023)
Article
Construction & Building Technology
Xuan Liu, Hua-Ping Wan, Yaozhi Luo, Chao Yang
Summary: The roof of a large-span structure is susceptible to wind-induced damage. Structural condition assessment using structural health monitoring data is effective in analyzing structural safety and detecting damage. This study proposes a data-driven combined deterministic-stochastic subspace identification method to identify the state of a wind-induced vibrating roof structure. A specific damage-sensitive feature is extracted using the system matrices from the identified state space model, and a statistics-based damage index is defined to detect structural state changes.
STRUCTURAL CONTROL & HEALTH MONITORING
(2022)
Article
Engineering, Civil
Wenwei Fu, Bochao Sun, HuaPing Wan, Yaozhi Luo, Weijian Zhao
Summary: This paper proposes a new damage detection approach based on temperature-induced strains. The method improves the modeling accuracy of damage detection by separating the temperature-induced component using independent component analysis, and establishes a damage index to determine the presence of structural damage. Results demonstrate the effectiveness of the proposed method in detecting damage and identifying the damage occurrence time.
ENGINEERING STRUCTURES
(2022)
Article
Engineering, Multidisciplinary
Hua-Ping Wan, Zi-Nan Zhang, Yaozhi Luo, Wei-Xin Ren, Michael D. Todd
Summary: This study proposes a multi-fidelity Gaussian process modeling (GPM) method for uncertainty quantification (UQ), which combines high-fidelity (HF) samples with computationally-cheaper low-fidelity (LF) ones to achieve a trade-off between computational cost and accuracy. A generalized co-Gaussian process model (GC-GPM) is developed to mix LF and HF samples, and an adaptive sampling strategy is introduced to reduce the number of training samples. The results show that the proposed adaptive GC-GPM method outperforms traditional GC-GPM in terms of computational accuracy and efficiency.
INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING
(2022)
Article
Construction & Building Technology
Xiao-Wei Ye, Yang Ding, Hua-Ping Wan
Summary: Wind speed forecasting is crucial for various purposes, and the use of Bayesian model is highlighted due to the random, nonlinear, and uncertain characteristics of wind speed. The selection of covariance function directly affects the modeling performance, and different covariance functions have different impacts on forecasting performance.
STRUCTURAL CONTROL & HEALTH MONITORING
(2021)