É«ºÐÖ±²¥

Enabling computers to predict the taste of new beers is the aim of mathematical models being developed at Swansea University for the brewing industry.

Intended to substantially reduce the cost of developing new flavours and blends, the five-year project co-ordinated by Bob Lovitt of the department of chemical engineering is being funded with Pounds 30,000 from the Brewing Research Foundation.

"Before embarking on the actual modelling, we had to investigate the effects of differences in ingredients, and changes in the brewing process, on the flavour of the finished product," he explains.

He began by chemically analysing the various amino acids and sugars that occurred in the differing samples of beer's basic ingredient, malt. He then charted the effects of different types of yeasts, and finally accessed the influence of the actual fermenting process. The type of fermenting used, for example, proved significant with long thin fermenters producing better quality beer than shallow ones.

After feeding all these variables into the models he hopes brewers will be able to turn to a computer to calculate how new beer blends and flavours will taste.

Dr Lovitt is convinced that computer flavoured prediction has wide potential in the food industry. After completing the brewing project he plans to adapt the models to suit the needs of the dairy industry, as he believes that they could be equally successful in predicting the flavours of yoghurt.