A vast quantity of data in the form of correlations is available for use in design and analysis. Even using hitherto-available databases there are significant difficulties in obtaining the best correlation for a particular purpose. The limitations of current databases in this respect are discussed and the implementation of a deductive database to overcome these limitations is described.
A database has been developed which makes available a wide variety of correlation information; the data includes arithmetic expressions, tables, graphs and procedural methods on the topic of jet impingement heat transfer. Logic programming is used to provide access to the relational database and to declare rules which act on thi~ data: the resulting knowledge-based system is able to confine its selection of information to relevant and valid correlations. In addition, facilities created using knowledge-based programming techniques assist the engineer in the task of providing a full description of the heat transfer system and in evaluating the correlations, returning values for heat transfer coefficients and flux for the system described. The knowledge-based methods described provide improved access to the correlation data and are believed to be applicable to other fields where large quantities of data are presented in the form of correlations.