Abstract
The article presents a new bio-inspired algorithm for industrial equipment diagnostics based on the joint use of a neural network and an immune network. The idea of integrating neural and immune networks was inspired by the human body, where these two mechanisms effectively interact to maintain immune regulation. The neural part of the algorithm is responsible for reducing the dimensionality of the data, which reduces the load on computing resources and improves the accuracy of predictions. The immune network is used to diagnose equipment operation. The feature of the proposed neuro-immune algorithm is the ability to work with industrial raw data, including sensor readings and the current state of the equipment. The predictive model created on the basis of this data is capable of detecting faults in advance. During the modeling, two industrial databases were used to evaluate the efficiency of the algorithm: "Motor" and "TSHO", which are time series of data from sensors. The mechanism for implementing the algorithm using functional blocks on the Modicon M340 series industrial controller and the EcoStruxture Control Expert platform is shown. The modeling results showed the prediction accuracy of 97 Artificial Intelligence • Prediction • Neuro-immune algorithm