Improving turbomachinery design process management

It is widely acknowledged that a company's ability to aquire market share, and hence its profitability, is very closely linked to the speed with which it can produce a new design. Indeed, a study by the U.K. Department of Trade and Industry has shown that the critical factor which determines profitability is the timely delivery of the new product. Late entry to market or high production costs dramatically reduce profits whilst an overrun on development cost has little significant effect. This paper describes a method which aims to assist the designer in producing higher performance turbomachinery designs more quickly by accelerating the process by which they are produced. The adopted approach combines an enhanced version of the 'Signposting' design process management methodology with industry-standard analysis codes and Computational Fluid Dynamics (CFD). It has been specifically configured to enable process-wide iteration, near instantaneous generation of guidance data for the designer and fully automatic data handling. A successful laboratory experiment based on the design of a large High Pressure Steam Turbine is described and this leads on to current work which incorporates the extension of the proven concept to a full industrial application for the design of Aeroengine Compressors with Rolls-Royce plc.

Uncertainty management for on-line optimisation of a POMDP-based large-scale spoken dialogue system

The optimization of dialogue policies using reinforcement learning (RL) is now an accepted part of the state of the art in spoken dialogue systems (SDS). Yet, it is still the case that the commonly used training algorithms for SDS require a large number of dialogues and hence most systems still rely on artificial data generated by a user simulator. Optimization is therefore performed off-line before releasing the system to real users. Gaussian Processes (GP) for RL have recently been applied to dialogue systems. One advantage of GP is that they compute an explicit measure of uncertainty in the value function estimates computed during learning. In this paper, a class of novel learning strategies is described which use uncertainty to control exploration on-line. Comparisons between several exploration schemes show that significant improvements to learning speed can be obtained and that rapid and safe online optimisation is possible, even on a complex task. Copyright © 2011 ISCA.