Abstract: Uncertainties analysis has a significant influence on the whole design optimization process of the engineering system in multidisciplinary coupled environment. The variance calculation of collaborative uncertainty analysis (CUA), one of the methods for uncertainty analysis in multidisciplinary system, has disadvantages of poor accuracy and low efficiency. On the basis of analyzing the reasons leading to the above–mentioned disadvantages, a modified collaborative uncertainty analysis (MCUA) is presented. The MCUA is an approach that utilizes parallel and distributed optimization system of CUA, Taylor approximation and sensitivity analysis to evaluate the mean and variance of coupled variables and system outputs, and obtains the robust optimum solution through second optimization. The computational results of the example indicate that the variance of system outputs calculated by the MCUA has higher accuracy than the variance calculated by the original method, and it is an effective method to analyze the uncertainty propagation in multidisciplinary system.

Key words: Collaborative uncertainty analysis, Modified collaborative uncertainty analysis, Multidisciplinary system, Sensitivity analysis, Taylor approximation