Use this URL to cite or link to this record in EThOS:
Title: Modelling damage in turbine blades and their coatings
Author: Ostovari, Farnaz
ISNI:       0000 0004 8504 6306
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
In the process of modelling damage in turbine blades and their coatings, four studies were carried out in this thesis. The first two chapters focus on modelling the damage in thermal barrier coatings or TBCs which are commonly used for Ni-based turbine blades. The second two chapters focus on development of meshfree methods for modelling micro-cracks in woven SiC_f/SiC_m composites which are a potential future material to be used for turbine blades. A brief summary of the main achievements from each chapter is given here as an insight into what is expected from each work. Swelling as the Main Source of Rumpling in TBC, chapter 2: Previously it was believed that the main source of rumpling growth in TBC systems is from phenomenon such as phase transformation and thermal mismatch that occur during the heating and cooling processes of a thermal cycle. However, the findings from an experimental work by Tolpygo & Clarke could not be explained with the previously suggested theories, where no difference in the rumpling amplitude was observed as the lower temperature of the thermal cycles was changed, except for the isothermal case. This behaviour was puzzling because it mitigated the effects of phase transformation and thermal mismatch. In this work, the existing analytical model of rumpling by Balint et al. was modified to include a relatively new phenomenon known as swelling, and used to reproduce and explain the experimental results. The analysis of the data from the developed model revealed that most of rumpling occurs during the dwell which is caused by swelling; its effects are also apparent during heating and cooling processes. Therefore, swelling proves to be the main source of rumpling growth. Lateral Growth in the Bond Coat and Inter-diffusion Layers, chapter 3: In the process of understanding the puzzling outcome from the experimental work of Chen et al. on measuring the lateral growth of the bond coat/inter-diffusion layers of TBC system after 50 hours of isothermal heat treatment at 1150 C, two finite element models of the system were produced; one with the coating modelled as two layers and another with the coating modelled as four layers. Chen et al. experimental results showed a large lateral deformation for the bond coat and almost none for the inter-diffusion layer, which was surprising because swelling effect which is a volumetric phenomenon, was observed in both layers, hence, it should have led to lateral swelling for the inter-diffusion layer as well. In this work, it was shown that because of the non-uniform nature of the Ni/Al inter-diffusion the two-layered model is not detailed enough to capture the real behaviour of the system, hence, the four-layered model is introduced which more closely matches the experimental results. This outcome indicated that modelling this system with two layers can create implications when modelling rumpling, therefore, a multi-layered coating system, such as the four-layered model shown in this work, is needed for modelling rumpling more accurately. MQ-RPIM Optimisation for Engineering Single Body Problems, chapter 4: Multi-quadrics radial point interpolation meshfree (MQ-RPIM) method is one of the common mehsfree methods currently used. However, the shape parameters involved in the generalised multi-quadric method have a strong influence on the accuracy of the solutions. In addition to the shape parameters, there are variables related to the integrations involved in the MQ-RPIM method that affect its accuracy. In this work a novel systematic algorithm was introduced which produces the best values for the variables involved in the MQ-RPIM method, including the integration and MQ shape parameters, for any engineering problem. For demonstration, this method is applied to three solid mechanic problems in both two- and three-dimensional forms. MQ-RPIM Model of Plain Woven Composite with Frictionless Contact, chapter 5: In the process of developing an explicit model for the micro-cracks at the yarn-matrix interface of a woven SiC_f/SiC_m composite using meshfree methods, the first three-dimensional MQ-RPIM frictionless contact code is developed from scratch and successfully applied to the preliminary model of a plain woven composite unit cell for two limiting conditions; i) full-stick (0% delamination) and ii) full-slip (100% delamination). As part of the development for this contact model, a two-dimensional frictionless contact model is also produced, where both two- and three-dimensional forms of the contact code are verified against analytical and finite element results for two Hertzian contact problems. The MQ-RPIM results for the Hertzian examples made a use of optimisation algorithm introduced in chapter 4, confirming the use of this algorithm and flexibility of the MQ-RPIM method compared to the FEM for models with non-uniform distribution of nodes, particularly at contact regions.
Supervisor: Balint, Daniel S. ; Aliabadi, M. H. Ferri ; Haynes, Peter Sponsor: Engineering and Physical Sciences Research Council
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral