Prop-blade section design optimization using weight/dynamic characteristic surrogate model with skin/spar design variable

Abstract

For prop-blade section design optimization, an analysis procedure was established to generate a weight/dy-namic characteristic surrogate model based on the general blade section design procedure. The general blade section design procedure consists of requirement analysis, cross-section design, section properties analysis, dynamic characteristics/load analysis and structure analysis to check all requirements are satisfied. Based on this procedure, the database was created by performing section design variable input value generation, section shape and finite element data generation, section properties analysis and dynamic characteristic analysis se-quentially. And an artificial neural network algorithm was applied to create a surrogate model of estimated weight and dynamic characteristics with a control prop-blade section design variable. The prop-blade section configuration has consisted of skin, C-spar and form. Section design variable was decided to a thickness of skin(Skin-t), a thickness of spar(Spar-t) and chordwise length of spar(Spar-l) from the leading edge. A data-base of design variable input values for the creation of a surrogate model was generated using the space-filling method, which is a kind of design of experience and design input variables were generated within a limited range by setting the ranges and constraints of each variable in consideration of the actual prop blade manufacturing characteristics. The weight/dynamic characteristics surrogate model derived through this pro-cess was applied to the genetic algorithm and the optimal cross-sectional shape that satisfies the dynamic characteristic requirements and weight minimization was derived.