Artificial neural networks for automation technology
Industrial process control using neural networks

Artificial neural networks are tools for representing the complex functional relationships required by modern control systems. Recorded information is sent to the neurons in a network and stored using weighting factors. Networks are trained on the basis of real-life examples, online if possible.

Since these neural networks can be trained to acquire the desired behavior, they are universal modules designed for use wherever it is necessary to describe complex behavior by example and illustrate it with a functional representation.

Networks are therefore used to develop models describing the transfer characteristics of non-linear systems, which are very difficult to describe. They are used for signal prediction, adaptive control, adaptive filtering, classification, monitoring and system diagnosis.