Rahul Upreti

Graduation Semester and Year




Document Type


Degree Name

Master of Science in Mechanical Engineering


Mechanical and Aerospace Engineering

First Advisor

Brian Dennis


This study provides an insight into the factors which have a major impact on the probability of failure of a Ball Grid Array package. A stochastic finite element analysis was done on the package using First-order reliability method (FORM). The effects of uncertainty in material thermal conductivities, heat source, heat transfer coefficient and ambient temperature were studied. The average temperature at the die-junction interface was selected as failure criterion because the excess temperature at the junction is often a cause of failure. The performance of Finite Difference Method(FDM) and Semi-Analytic Complex Variable Method(SACVM) in computing sensitivities are compared. Results show that for some parameters, the probability of failure is more sensitive to uncertainties than other parameters, which provides crucial information to select the manufacturing tolerance of material properties and tolerance for parameters which are controlled experimentally. For sensitive parameters, a tighter manufacturing tolerance will decrease the randomness and ultimately increase the reliability. The parameters for which probability of failure show low sensitivity with respect to uncertainties, large tolerance for manufacturing could be selected which will drop down the cost.


Uncertainty analysis, First order reliability method, Stochastic finite element analysis, BGA package


Aerospace Engineering | Engineering | Mechanical Engineering


Degree granted by The University of Texas at Arlington