In this paper, effects preceding a latch-up fault in insulated gate bipolar transistors (IGBTs) are studied. Primary failure modes associated with IGBT latch-up faults are reviewed. Precursors to latch-up, primarily an increase in turn-off time as a consequence of elevated junction temperature, are examined for an IGBT. The relationship between junction temperature and turn-off time is explained by modeling the parasitic properties of an IGBT. A metric is derived from the model to standardize the relative estimates in junction temperature from measurements of turn-off time. To evaluate the effects preceding latch-up in-situ, seeded fault testing is conducted on a three-phase power inverter using aged transistors induced with a fault located in the die-attach solder layer. Experimental results demonstrated the feasibility of using the proposed metric as a precursor to transistor latch-up.

No abstract

No abstract

No abstract

In this paper, a branch and bound multi-objective based method is proposed for reaching the non-dominated set. Two types of nodes are considered in the tree-search. The first type characterises the non-integer solutions found which are transformed to integer solutions by applying a branching procedure. The second type of nodes contains an integer solution and in this case efficient cuts are established in order either to remove dominated integer vectors or to fathom them. The method is compared advantageously with two exact methods of the literature tailored for the general case and also analysed computationally on benchmarks of MCDM library.Keywords: branch and bound; multi-objective programming; nondominated solution.Reference to this paper should be made as follows: Abbas, M., Chergui, M.E-A. and Mehdi, M.A. (2012) 'Efficient cuts for generating the non-dominated vectors for Multiple Objective Integer Linear Programming', Int. Efficient cuts for generating the non-dominated vectors for MOILP 303of Mathematics and Laboratory LAID3, USTHB, his research interests include mathematical technics and tools for combinatorial optimisation problems. He works on modelling and solving real-world problems as an expert and develops exact and metaheuristic methods for multiobjective discrete optimisation problems. His recent works are focused on Multiobjective Combinatorial Optimisation (MOCO) in linear, hyperbolic and quadratic cases.Meriem Ait Mehdi received her Master's Degree in Operation Research from the