Design optimization of rotor-bearing systems for industrial turbomachinery applications

The design of large scale rotor-bearing systems for industrial turbomachinery is a multifaceted task involving many engineering disciplines. The accurate analytical prediction of the behavior of such systems prior to their manufacturing and commissioning to the user is largely responsible for their reliability and conformity to the application specifications. As the complexity of the design of a rotor-bearing system increases, as in the instance of multi-stage and multi-shaft turbomachinery, the requirement of an automated computer aided design protocol becomes justified from both engineering and economical points of view. Since a systematic process, corroborated by the designer's experience, is customarily followed in the conceptual development and verification of rotating machinery, it appears feasible to automate such a process by devising numerical algorithms that, with the aid of a digital computer, would not only allow engineers to develop a workable machine design but also to obtain a near optimum design meeting the specifications of the given application. Inherently, design is an iterative process in which pertinent parameters are feasibly modified until established objectives are achieved.; Several automated optimization procedures have been developed within the scope of the dissertation to aid the industrial design of large scale rotor-bearing systems. Design objectives considered in this dissertation include: optimal configuration of hydrodynamic radial bearings, selection of bearings and their characteristic parameters to minimize rotor dynamic response while stability of motion is maintained or improved.; The rotating assembly has been modeled using the finite element method formulation employing a matrix condensation approach to significantly reduce solution time. Numerical optimization techniques utilizing gradient information of the problem functions coupled with multidiscipiinary optimization protocols involving problem decomposition and subtask solutions have been applied to assist the design of hydrodynamic journal bearings supporting industrial rotors. Selected example cases regarding bearings and complete rotating systems are presented to illustrate the feasibility and usefulness of automated optimal bearing design procedures.