Facial reconstruction

Search LJMU Research Online

Browse Repository | Browse E-Theses

Development of an inverse FE modelling method for material parameters identification based on indentation tests

Li, B (2009) Development of an inverse FE modelling method for material parameters identification based on indentation tests. Doctoral thesis, Liverpool John Moores University.

[img] Text
505926.pdf - Published Version

Download (37MB)

Abstract

In this work, an inverse FE modelling program based on the Kalman filter technique has been developed and used to study three material models (linear elastic, hyperelastic and hyper foam). Two error treatment methods have been developed and implemented in the program and their feasibilities for different material systems were systematically investigated. FE models simulating the indentation process of three typical material behaviours have been developed and some important factors including mesh sensitivity, frictional conditions and material properties have been systematically studied to validate the FE models. The use of single indenter and dual indenters have been comparatively studied in terms of accuracy, convergence and robustness of the inverse program, which are important for materials characterisation. The program was evaluated with blind tests using numerical experimental data of known material properties. The validated method was then successfully used to study the properties of EVA foams for midsole of sport shoes and human heel pad using a newly develop continuous indentation testing system. Blind tests have been successfully used to establish the validity, efficiency and robustness of the program with different material models, error treatment and selection of indenters. The results showed that the double indenters method is better than the single indenter method, which is initial value dependent for some materials. The results also showed that the new random error treatment method is applicable to all the three material models while the converged results based on the program with constant error distribution was initial values dependent. The blind test results showed that Kalman filter is a feasible method and the random error treatment is more practical approach for characterisation material and can be universally applied to different materials models. Sensitivity test with perturbation in the indentation force demonstrated that the program is robust against potential experimental noise/errors. The framework established has been successfully used to characterise the properties of EVA foams in comparison with conventional compression, and compression-shear methods. The prediction from indentation tests showed comparable accuracy to the standard combined compression-shear tests, while pure compression could not predict the parameters accurately describing the material at complex situations. The parameters inversely predicted can be directly used in the product design and simulation process. A new in vivo test machine has been developed and performed on human subject with good accuracy and repeatability. The inverse method has been used to predict the elastic and nonlinear parameters of the heel pad. The predict elastic and hyperelastic properties showed good correlation for all the subjects tested. However the nonlinear model is more accurate, in describing the stiffening effect of the heel pad. This method would provide a practical way for detecting the property change of the heel pad with different conditions.

Item Type: Thesis (Doctoral)
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
T Technology > TS Manufactures
Divisions: General Engineering Research Institute
Date Deposited: 14 Mar 2017 10:48
Last Modified: 14 Mar 2017 10:48
URI: http://researchonline.ljmu.ac.uk/id/eprint/5923

Actions (login required)

View Item View Item