Large Vessel Nozzle Penetration Geometric Optimisation Study in 3D Incorporating Design of Experiments Techniques

Abstract

This paper describes the results of an optimisation study into the effects of changing geometrical variables local to the nozzle entry on a large pressure vessel. The purpose of the study was to quantify the effects of altering the geometry, and thereby provide trade-offs between key responses such as primary strength and shakedown performance. A 3D Finite Element model was built of a 90 degree sector of the pressure vessel which was sufficiently detailed to allow the thermal and structural effects of the vessel remote from the nozzle to be included. Specifically, this included the thermal and structural influence of the closure head and bolting assembly. The model was then parameterised for 5 independent variables, including the extent of the nozzle reinforcement, crotch corner fillet radius and nozzle thickness. The parameterised model was then subjected to a number of thermal and structural transient analyses during Level A operation, as well as a representative strength loadcase. A full factorial design study was undertaken, comprising of separate 243 analysis runs covering the 5 independent variables at 3 levels. A number of output metrics were monitored, and the effects on the output metrics resulting from changes to the inputs were quantified. Due to the full factorial nature of the experimental design, interactions between variables could also be investigated. The response for each output metric was then fitted to a response surface, which allows a polynomial (meta-model) of each metric to be calculated. These responses were input to a simple Excel chart which allows the designers to perform rapid what-if design scenarios, and see the resulting effects of their changes on the responses. This allows the trade-offs between responses, for example shakedown and strength trade-off for shell thickness, to be easily seen and quantified.