Differential Transform and Its Application to Engineering Problems

博士 === 國立成功大學 === 航空太空工程學系 === 86 === Differential transform and its application to engineering problems, include initial-value problems, boundary-value problems, and parameter identification problems are studied in this theses. Since the operation of dif...

Full description

Bibliographic Details
Main Authors: Liu, Yung-Chin, 劉永欽
Other Authors: Chieh-Li Chen
Format: Others
Language:zh-TW
Published: 1998
Online Access:http://ndltd.ncl.edu.tw/handle/90538602444789359380
Description
Summary:博士 === 國立成功大學 === 航空太空工程學系 === 86 === Differential transform and its application to engineering problems, include initial-value problems, boundary-value problems, and parameter identification problems are studied in this theses. Since the operation of differential transform can be developed from one dimension to multi-dimension, both ODE and PDE can be solved by this technique. In some sense, differential transform is more universal and practical than integral transform. Differential transform method provides an adaptive technique, which involved more concise adjustment policy and raises the efficiency of numerical computation of initial-value problems. Stiff equations can also be solved by this method with specific stability and accuracy. Both linear and nonlinear boundary-value problems can be solved as initial-value problems after taking differential transform. From the problem considered, it shows that the proposed technique can obtain reasonably accurate approximate solutions and find all possible solutions of the problem, which is distinct from the existing approaches. Differential transform method is also applied to parameter identification problems. The system model and the criterion function can be obtained via the unknown parameter and the initial value of state variable. The proposed method provides direct search algorithm to find the maximum likelihood estimate. Both linear and nonlinear problems can be solved in the same process and the problem of singularity and sensitivity in solving traditional inverse problems can be avoided.