
In this study, the damage process of Thermal Barrier Coatings (TBC), which were exposed at 1273K for 500hours in advance, under tensile loading was observed continuously, and the influence of exposure temperature on crack initiation strength and interface strength was clarified quantitatively based upon interfacial fracture mechanics. TBC/IN738LC specimen, which is consisted of yttria-stabilized zirconia as top coat (TC) and CoNiCrAlY as bond coat (BC) in TBC, was prepared and the high-temperature exposure treatment was then conducted before tensile test. Continuous observation of damage process was revealed that (i) interfacial crack propagated along the interface between TC and BC in exposed TBC tested at room temperature however the crack propagated along BC/substrate interface in as-sprayed TBC, and (ii) interfacial crack propagated along TC/BC interface in the exposed TBC specimen as well as as-sprayed one. It was considered that the crack propagation path was changed by strong effect of a high-temperature exposure treatment. Quantitative evaluation also revealed that critical strain up to crack initiation in TC was increased by the exposed treatment and number of cracks per length was decreased in contrast with the critical strain. The exposure treatment stiffened top coat layer, which means that cohesive strength between splats became stronger. The interfacial fracture energy of the exposed TBC specimen was lower than that of as-sprayed one. Therefore, in order to assess precisely remaining life of main hot-parts of gas turbine, data for as-sprayed TBC is needed as initial information.
In a growth prediction for fitness-for-service assessment, mechanical interaction between multiple cracks was considered by assuming a coalescence of interacting cracks. The current coalescence rule for crack growth prediction is prescribed for interacting two cracks and the effect of size difference is not taken into account. No rule is provided for cracks of more than three and it was pointed out that the current rule is overly conservative. In this study, the coalescence criterion (N-criterion), which has been developed for two cracks of different size, was applied to the multiple cracks. The S-version finite element method was employed to simulate the crack growth behavior under interaction. It was shown that the growth of three cracks could be predicted conservatively by applying the N-criterion. For applying the N-criterion to three cracks, however, the small crack has to be treated properly when the crack was judged to stop growing. Furthermore, it was pointed out that the relative position between multiple cracks should be modeled with care when cracks were coalesced. The detailed procedure for the growth prediction was shown and its validity was shown by predicting the growth of six cracks, which was observed in an actual plant. The proposed procedure could predict the growth of interacting multiple cracks conservatively.
In this paper, a parameter-free optimization method is proposed for designing the optimal free-formed shell with an arbitrary surface. The compliance minimization problem of a linear elastic shell is dealt with here as an application example of this method. The optimum design problem is formulated as a distributed-parameter shape optimization problem under the assumptions that a shell is varied in the normal direction to the surface. The shape gradient function and the optimality conditions for this problem are theoretically derived by using the material derivative method and the Lagrange multiplier method. The shape gradient function is applied to the shell surface as a pseudo distributed force under a Robin condition for varying the surface, for minimizing the objective functional and for regularizing the mesh. With this approach, an optimal global and local curvature distribution of shell surface can be determined without any shape design parameterization. The validity and practical utility of this method are verified through several design problems. It is also confirmed that a membrane stress state is obtained by this method.
Sandwich panels such as roll core sandwich panel (hereafter, RSP for brevity) and honeycomb core sandwich panel (hereafter, HSP for brevity) are superior in impact absorption under whole surface compression, because of the buffer effect of core. On the other hand, impact properties of sandwich panels under local compression such as drop weight impact are affected by mechanical properties of the plate as the face sheet in addition to the core layer. This paper describes drop weight impact properties of double layer RSP which consists of two core layers and three plates. In order to compare RSP with HSP and build up the experimental database for design of double layer sandwich panel, a series of drop weight tests was carried out by using double layer RSPs and HSPs. From test results, the follows were summarized. The RSP and the HSP have common characteristics: 1) Three peaks of the load appeared in the load-displacement curve. The peak loads were larger in sequence of their appearances. 2) In the load-displacement curve until the last peak of the load, there were load increase regions by deep drawing of the plates and load equilibrium regions similar to plateau region under whole surface compression. Then drop weight impact properties of the RSP are higher than the HSP in terms of 1) the deformation resistances of the load increase regions, 2) the peak loads, 3) the loads of the load equilibrium regions, and 4) the absorption energy until the 2nd peak of the load.
In this paper, the mechanism of stress concentration due to defects in honeycomb sandwich structure is studied by using finite element method. It is found that the stress concentrantion can be expressed as a function of Ectc=Eptp, where tc and tp are the thickness of core and plate, respectively, and Ec and Ep are the Young's modulus of core and plate, respectively. Also, it is found that the stress concentrantion can be evaluated by an approximation of an elliptical inclusion. The validly of this technique is verified through comparison of the evaluation values with the numerical results obtained by FEM.