Multi-angle valve seat machining: experimental analysis and numerical modelling

Modern automotive manufacturers operate in highly competitive markets, heavily influenced by Government regulation and ever more environmentally conscious consumers. Modern high-temperature, high-pressure engines that use high hardness multi-angle valve seats are an attractive environmental option, but one that manufacturers find requires more advanced materials and tighter geometric tolerances to maintain engine performance.

Tool manufacturers meet these increasingly tougher demands by using, higher hardness cutting materials such as polycrystalline cubic boron nitride (pcBN), that on paper, promise to wear at a lower rate, require less coolant and deliver tighter tolerances than their carbide counterparts.

The low brittle fracture toughness of pcBN makes tools that use it vulnerable to minute chipping. A review of literature for this work pointed to no clear answer to this problem, although suggestions range from manufacturing defects, dynamic and flexibility problems with the production line machinery and fixtures, and radial imbalances in the cutting loads.

This work set about experimentally investigating those potential explanations, coming to the conclusion that the high radial imbalance of the cutting loads is responsible for pcBN cutting insert failure during multi-angle valve seat machining, and that by simply relocating the cutting inserts around the multi angle cutting tool, the imbalance can be reduced, thus extending the life of the cutting inserts.

It is not always easy to predict the imbalance due to the multiple flexibilities in the system, and simulating such a system in 3D with all its associated cutting phenomena such as friction, thermal expansion, chip flow and shearing, would call upon extraordinary computational power and extremely precise experimental inputs to reduce cumulative error.

This thesis proves that such a 3D simulation can be made, that runs in exceptionally short durations compared to traditional methods, by making a number of simplifications.

MSC Marc was used to host the simulation, with a parametric script written in Python responsible for generating the model geometry and cutter layout. A Fortran program was developed that is called upon by Marc to calculate the required cutting load outputs and generate new workpiece meshes as material is removed.