Stochastic Modeling and Analysis of a Heavy Duty Radiator

Abstract

Governmental agencies across the globe are constantly evolving with stringent emission laws to tackle the problem of CO2 and NOx/SOx emissions. New emission standards force the Truck OEM’s to redesign the engine. The paper is aimed to measure the header tube joint stress of the radiator subjected to random variations in geometry, shape and material properties. Linear analysis will not consider the uncertainty and randomness due to tolerance, process changes, part-part variation etc. Stochastic finite element analysis (FEA) is carried out to account the uncertainty in the system. The finite element model of radiator system is built and baseline linear simulation is performed to obtain the baseline deformation and baseline stress responses. Then the uncertainty and random variation due to the geometry, material and shape variable is defined by a normal distribution function. Random designs are generated by defining the upper and lower bound limit values for the input design variable. Random designs are populated using Monte-Carlo simulation technique. 250 random design points are created for each design variables. Then stochastic simulation is performed to evaluate the responses at random design points. Statistical and probabilistic tools are used to post process the simulation results. The paper showcases application of stochastic simulation method which aids in indentifying the robust design with minimum variations. This also enables engineers and designers to understand the relationship and significance between different design variables in designing energy efficient systems.