SYNTHESIS OF NEURO-FUZZY NETWORK FOR STABILIZING TEMPERATURE IN PROCESS OF CONTINUOUS STERILIZATION

Abstract

To synthesize a neuro-fuzzy network of stabilizing temperature there is a view on the development of a base of rules for a fuzzy controller taking into account the optimal object management and training of a hybrid neural network. The optimal trajectory was accepted as the performance-optimal (detected by the maximum principle) management for a closed loop system for automatic management. There has been a transition made from the temporary area where the optimal management has been found, to the phase plane of the system, which allowed a direct use of the solution to develop a rule base for a fuzzy controller. There has been a neuro-fuzzy network (Adaptive-Network-Based Fuzzy Inference System - ANFIS) employed to develop a temperature stabilization system for sterilization with two management influences.

Keywords

• нейро-нечеткая система
• стабилизация температуры
• оптимальная траекто-рия
• база правил
• обучение нейронной сети
• моделирование
• neuro-fuzzy network
• temperature stabilization
• optimal trajectory
• rule base
• neural network training
• simulation