The application of formal optimizing methods to the design of steam turbine reaction bladepaths

Due to the increasingly competitive nature of the world market-place, steam turbine manufacturers must routinely produce designs which are extensively optimised whilst adhering to demanding tender and contract lead-times. The current research is concerned with the application of formal optimising strategies to the mean-line and throughflow optimisation of reaction turbine bladepaths using a rule-base, numerical optimisation, and genetic algorithms. A generic software framework is described, which manages the iterative running of design codes and also the interchange of data between them. It is demonstrated that this method produces entirely viable designs with a high level of optimisation. Significantly, design times are reduced by an order of magnitude, performance gains are predicted, and new design rules can be deduced.