·         Coupled magneto-elastostatic analysis using Implicit Boundary Finite Element Method

 Implicit boundary finite element method (IBFEM) uses solution structures constructed using step functions to enforce boundary and interface conditions so that a structured grid can be used to perform the analysis. A structured grid, which consists of regular shaped elements, is much easier to generate than conforming mesh thus eliminating the difficulties associated with mesh generation for complex assembles.  In this study, IBFEM is extended to solve 2D and 3D magnetostatics, compute magnetic forces and to solve coupled magneto-elastostatic problems that typically involve an assembly of parts made of several different materials. The geometry is accurately modeled using equations from CAD models and a separate structured mesh is used for each part in an assembly. Specially constructed solution structures are used to represent test and trial functions such that boundary and interface conditions are enforced. Several magnetostatic problems with known solutions are modeled to validate the method. The magnetostatic problems are classified as unbound problems so that sometimes a very large analysis domain should be modeled to get more accurate results. In order to reduce the analysis domain, two open boundary techniques are developed for IBFEM: asymptotic boundary conditions and decay function infinite element. In addition, a magnetostatic problem with permanent magnets is solved using IBFEM.

·         Long-term reliability prediction of LED packages using numerical simulation

 Solid-state lightings (SSL) rapidly penetrate the global illumination market because of the energy efficiency and the reliability. The energy efficiency can be easily evaluated but the reliability is not convenient to be estimated. SSL is based on semiconductor and brings new manufacturing process and new materials, which introduces a series of new and unknown failure modes. Using several reliability tests, the reliabilities of LED packages are evaluated. However, the general experiments are mostly time-consuming and expensive. In this chapter, two failure modes caused by the corresponding reliability tests are focused, which are wire bond breakage for the thermal shock cycle test and silicone degradation for high temperature operating life test (HTOL). As these failure modes belong to long-term reliability issues for LED packages, solving these issues are crucial in term of time to market (TTM). Two failure modes were quantified by using finite element analysis. For the wire bond failure, a wire bond lifetime model was developed to predict the number of cycles to failure during thermal shock tests. For the silicone degradation, a lumen depreciation model was suggested to estimate the lumen depreciation for HTOL. Using two models, long-term reliability of LED packages could be predicted using numerical simulation without time-consuming experiments.