中国机械工程

• 机械基础工程 • 上一篇    下一篇

面向响应准确度的参数不确定性模型确认方法

项涌涌;潘柏松;罗路平;谢少军   

  1. 浙江工业大学机械工程学院,杭州,310014
  • 出版日期:2019-04-10 发布日期:2019-04-04
  • 基金资助:
    国家自然科学基金资助项目(51475425,51075365)

Validation Method of Parameter Uncertainty Models for Response Accuracy

XIANG Yongyong;PAN Baisong;LUO Luping;XIE Shaojun   

  1. College of Mechanical Engineering,Zhejiang University of Technology,Hangzhou,310014
  • Online:2019-04-10 Published:2019-04-04

摘要: 针对参数不确定性模型中概率与区间混合不确定性情况的模型确认问题,提出了面向响应准确度的模型确认方法,并阐明了确认过程的具体实施步骤。根据响应数值结果的非精确累积分布函数和响应试验结果的经验分布函数求解模型确认准则参数,并评估模型确认结果。对于不满足评估要求的模型,建立以待修正区间不确定性参数总置信水平最大为优化目标,以模型确认评估标准和模型参数初始值为约束条件的优化模型,运用遗传算法求解优化问题,并得到模型参数修正值。将模型确认方法应用在圣地亚热传导问题中,结果表明:所提方法可明显提高预测精度和响应准确度,结果真实可信。

关键词: 参数不确定性模型, 概率与区间混合不确定性, 模型确认, 圣地亚热传导问题

Abstract: Aiming at the problems of model validation for mixed uncertainty cases of probability and interval in parameter uncertainty models, a model validation method was proposed for response accuracy, and the specific implementation steps were described for the validation processes.Model validation criterion parameters were obtained by using the imprecise cumulative distribution function of response numerical results and empirical distribution function of response test results, then the model validation results were assessed. For the model that didn't satisfy the evaluation requirements, an optimization model was established, the maximum total confidence level of interval uncertainty parameters to be corrected was taken as the optimization object, the assessment standard of model validation and initial values of model parameters were used as constraint conditions, finally the optimization model was solved by the genetic algorithm and the corrected values of model parameters were obtained.The model validation method was applied to the Sandia thermal challenge problem, the results show that the proposed method may significantly improve the forecast accuracy and response accuracy, the results are credible.

Key words: parameter uncertainty model, mixed uncertainty of probability and interval, model validation, Sandia thermal challenge problem

中图分类号: