PROBABILISTIC FINITE ELEMENT ANALYSIS OF A HEAVY DUTY RADIATOR UNDER INTERNAL PRESSURE LOADING

• Main Authors: ROBIN ROY P., V. HARIRAM, M. SUBRAMANIAN
• Format: Article
• Language: English
• Published: Taylor's University 2017-09-01
• Series: Journal of Engineering Science and Technology
• Subjects: Heavy duty radiator; Nominal finite element analysis; Probabilistic finite element analysis; Monte Carlo method; Robust design; Reliability.
• Online Access: http://jestec.taylors.edu.my/Vol%2012%20issue%209%20September%202017/12_9_11.pdf

Engine cooling is vital in keeping the engine at most efficient temperature for the different vehicle speed and operating road conditions. Radiator is one of the key components in the heavy duty engine cooling system. Heavy duty radiator is subjected to various kinds of loading such as pressure, thermal, vibration, internal erosion, external corrosion, creep. Pressure cycle durability is one of the most important characteristic in the design of heavy duty radiator. Current design methodologies involve design of heavy duty radiator using the nominal finite element approach which does not take into account of the variations occurring in the geometry, material and boundary condition, leading to over conservative and uneconomical designs of radiator system. A new approach is presented in the paper to integrate traditional linear finite element method and probabilistic approach to design a heavy duty radiator by including the uncertainty in the computational model. As a first step, nominal run is performed with input design variables and desired responses are extracted. A probabilistic finite elementanalysis is performed to identify the robust designs and validated for reliability. Probabilistic finite element includes the uncertainty of the material thickness, dimensional and geometrical variation. Gaussian distribution is employed to define the random variation and uncertainty. Monte Carlo method is used to generate the random design points.Output response distributions of the random design points are post-processed using different statistical and probability technique to find the robust design. The above approach of systematic virtual modelling and analysis of the data helps to find efficient and reliable robust design.