IMPROVING THE ACCURACY OF IDENTIFYING ANOMALIES CRITICAL TO AN OBJECT USING A NEURAL NETWORK MODEL

Eremin Igor V.

The article proposes a two-stage method for accurately detecting anomalies in industrial process data using the CascadeForwardNet neural network. The first stage involves prediction, while the second focuses on error correction and architecture modification. This approach improves anomaly identification accuracy, aiding in the analysis of equipment failure causes.

Keywords:APCS, industrial automation, chemical-technological process, neural networks, models, systems approach, anomalies