【摘要】 文章针对大系统模型降阶问题,提出了基于低阶组合模型的降阶方法。充分利用系统响应图形所表征大系统的宏观性和特殊性,把响应曲线模态分解为平滑模态和振荡模态,降阶模型结构为2种模态模型的线性组合。针对降阶模型模态提取关键问题,提出基于EMMD技术的提取方法。采用四阶复杂系统模型进行降阶验证,结果表明只需采用2个二阶模型的线性组合,即可达到相对误差为-0.0207的效果。该方法可以应用到采用纯数学方法降阶效果不理想的大系统降阶问题。更多还原

【Abstract】 Our brief review of existing literature at the start of our paper reveals the disadvantages of using pure mathematical methods as large scale system model reduction.In our opinion,system response curves provide much information so that they can be used for large scale system model reduction.To our knowledge,no paper in the open literature used this method.We now present our research results.First,system modes consist of smooth mode and oscillation mode.These two modes are modeled by using low order models.Large scale system reduction model is the linear combination of these two low order models.Secondly,the key issue of model reduction is mode extraction based on system response curve.Extremum Field Mean Mode Decomposition(EMMD) technology is used for mode extraction.Finally,a four-order complex large scale system model is used for verifying the feasibility of the proposed scheme.Results show that only using the linear combination of two two-order models has reached good results.Relative error is below-0.0207.This method can be applied to large scale system reduction problems when using the pure mathematical methods cannot achieve desired results.更多还原