Design of Experiments with Dynamic Responses

动态响应实验设计

• 批准号: 1726445
• 负责人: Rong Pan
• 金额: $ 31.46万
• 依托单位: Arizona State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-01 至 2021-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1726445&HistoricalAwards=false
• 关键词: Design; Experiments; Dynamic; Responses

Many engineered systems, such as those in the automotive and chemical industries, have dynamic performance or system outputs. The overall performance of such systems cannot be captured by a single number, but is instead represented by a response profile over time. Conducting experiments on these systems is costly and time-consuming; as a result, improving system performance through experimentation is difficult. This project will overcome these challenges through novel experimental design methodologies for engineered systems with dynamic outputs. The project is expected to impact a broad range of engineering applications. The academic research team will work closely with industrial partners to ensure applicability to real life settings.The research objective of this project is to improve experimental design methods for systems with dynamic response. The project will make advances in (1) dynamic response modeling, (2) optimal experimental designs, and (3) computational algorithm development. A novel hierarchical mix effect modeling approach will capture both the dynamic behavior of response variables and the effects of experimental factors. At the experimental design stage, the optimal sampling locations of the spectrum variables and design points in the experimental factor space will be investigated. Finally, fast computer algorithms that take advantage of the properties of B-spline basis function and the patterns in design matrices will be developed.

许多工程系统，如汽车和化学工业中的系统，具有动态性能或系统输出。这种系统的整体性能不能用一个数字来表示，而是用一段时间内的响应曲线来表示。在这些系统上进行实验是昂贵且耗时的;因此，通过实验来改善系统性能是困难的。该项目将通过新的实验设计方法来克服这些挑战，为工程系统的动态输出。该项目预计将影响广泛的工程应用。学术研究团队将与工业合作伙伴密切合作，以确保适用于真实的生活环境。本项目的研究目标是改进具有动态响应的系统的实验设计方法。该项目将在（1）动态响应建模，（2）最佳实验设计和（3）计算算法开发方面取得进展。一种新的分层混合效应建模方法将捕获响应变量的动态行为和实验因素的影响。在实验设计阶段，将研究光谱变量和设计点在实验因子空间中的最佳采样位置。最后，利用B样条基函数的性质和设计矩阵中的模式，将开发快速计算机算法。

项目成果

Boost-R: Gradient boosted trees for recurrence data

Boost-R：用于重复数据的梯度提升树

• DOI: 10.1080/00224065.2021.1948373
• 发表时间: 2021
• 期刊: Journal of Quality Technology
• 影响因子: 2.5
• 作者: Liu, Xiao;Pan, Rong
• 通讯作者: Pan, Rong

Design optimal sampling plans for functional regression models

为函数回归模型设计最佳抽样计划

• DOI: 10.1016/j.csda.2020.106925
• 发表时间: 2020
• 期刊: Computational Statistics & Data Analysis
• 影响因子: 1.8
• 作者: Rha, Hyungmin;Kao, Ming-Hung;Pan, Rong
• 通讯作者: Pan, Rong

Copula-based reliability analysis of degrading systems with dependent failures

• DOI: 10.1016/j.ress.2019.106618
• 发表时间: 2020-01-01
• 期刊: RELIABILITY ENGINEERING & SYSTEM SAFETY
• 影响因子: 8.1
• 作者: Fang, Guanqi;Pan, Rong;Hong, Yili
• 通讯作者: Hong, Yili

On designing experiments for a dynamic response modeled by regression splines

关于设计由回归样条建模的动态响应的实验

• DOI: 10.1002/asmb.2490
• 发表时间: 2020
• 期刊: Applied Stochastic Models in Business and Industry
• 影响因子: 1.4
• 作者: Pan, Rong;Saleh, Moein
• 通讯作者: Saleh, Moein

Bagging-enhanced sampling schedule for functional quadratic regression

用于函数二次回归的套袋增强采样方案

• 发表时间: 2021
• 期刊: Journal of statistical theory and practice
• 影响因子: 0.6
• 作者: Rha, H;Kao, M.-H.;Pan, R.
• 通讯作者: Pan, R.