Thermo-Mechanical Fatigue and Press-Fit Loss Analysis Of Valve Seat Insert

Valve seat inserts (VSI) are installed in cylinder heads to provide a seating surface for poppet valves. Insert material is more heat and wear resistant than the base cylinder head material and hence it makes them better suited for valve seating and improved engine durability. Also using inserts permits easier repair or rebuild of cylinder heads as only the wear surfaces need to be replaced. Desirable performance characteristics are appropriate sealing, heat-transfer and minimizing valve to VSI wear and undesired outputs include valve seat dropping and cracking. With downsizing trend of diesel engines, it leads to increasing power density and therefore higher cylinder pressure and temperatures. Hence the engine components are getting exposed to more severe loadings and hence to failure modes, which were not heretofore experienced based on the warranty data. Possible identified failure modes are insert’s yielding and corresponding press-fit loss leading to either it’s cracking or drop-off. During the scope of this work, initially thermo-mechanical finite element analysis is carried out to check for the insert’s yielding and further calculating press-fit loss and fatigue life. FEA predicted temperatures results are calibrated considering the thermal survey data available. First using the temperature dependent elastic material properties, appropriate yielding factor is considered to evaluate the insert’s yielding. Further non-linear finite element analysis is carried out to evaluate press-fit loss and low cycle fatigue life of the insert. Required material properties towards fatigue analysis are derived using Seeger’s approximation method. Stabilized stress-strain data is considered towards fatigue analysis and press-fit loss calculations. Commercial software’s like ANSYS and Fe-Safe are used for finite element analysis and fatigue analysis respectively. Very good co-relation is observed against the actual crack location and fatigue life. Further with alternative material option, analysis tasks are repeated and predicted analysis results also correlate well with experimental findings.