Catalytic Converter Design, Development and Manufacturing

Computer aided engineering is used to design, develop, optimize and manufacture catalytic converter. Heatcad, a transient heat transfer analysis is used to simulate the temperature response in the exhaust system to locate the catalytic converter to achieve maximum performance. Heatcad analysis provides the easy way to identify thermal management issues and to design and optimize the runner lengths and material thicknesses of the manifold, and downpipes. P-Cat is used to estimate back pressure due to substrates, end cones, and inlet/outlet pipes. Catheat, a one dimentional heat transfer tool is used to identify the converter insulation to maintain the required external skin temperature. Computational Fluid Dynamics (CFD) analysis, a powerful means of simulating complex fluid flow situations in the exhaust system, is used to optimize the converter inlet and outlet cones and the downpipes to obtain uniform exhaust gas flow to achieve maximum converter performance and reduce mat erosion. The uniformity index, velocity index (eccentricity) and pressure drop index predicted are used to optimize the geometry and orientation of the converter components. WAVE simulation is used to predict the exhaust system back pressure from the engine headface to tail pipe to estimate engine performance. Finite Element Analysis (FEA) is used to predict structural mechanics and structural dynamics of the full exhaust system to give insight about the thermal fatigueness of the converter assembly. Heat transfer analysis performed with thermal, mechanical and road load conditions is used to predict the static and vibrational stresses of the converter components. Monte-Carlo statistical simulation is used to study the effect of material tolerances on Gap Bulk Density of the mounting mat and to control the converter manufacturing process. The measured GBD of the assmbled converter selected from the manufacturing lot on a randam selection is plotted along with the statistical simulation results to validate the modeling capability.