Abstract
The problem discussed in this article relates to the integration of knowledge about the design process of products from ADI. Specialised domain knowledge, often resulting from costly experiments, published in the world in a large number of magazines, is an invaluable source of information for its users and researchers. The authors draw attention to the need to develop a system that will facilitate selective access to specific passages of this knowledge, allowing for its automatic processing. The cost of developing such a system is certainly lower than the cost of multiple repetition of the experiments. Proposed under the system, aggregation and centralisation of the results of studies already carried out may be the basis for planning and execution of subsequent experiments, covering areas hitherto unexplored. Another outcome may relate to the creation of new knowledge through the discovery of relationships and dependencies that are not visible in individual, single, experiments, but emerge when the results of a large number of different tests are compared. This paper proposes the use of artificial neural networks to explore the relations between the properties of ADI and selected heat treatment parameters based on a set of integrated experimental data from various publications. In the future, this form of knowledge representation may be used in intelligent computer systems in the knowledge acquisition module on the manufacturing process of ADI.