PREDICTING METHOD AND SYSTEM OF REAL-TIME RISK SIGN FOR ENERGY PLANTS USING INTELLIGENT RISK SIGN PATTERN MODEL

摘要

PURPOSE: A real time damage sign prediction method an energy plant using an intelligent damage sign pattern analyzing model and a system thereof are provided to predict a damage sign state based on extended spherical self organized map algorithm, thereby increasing safety and decreasing an accident of a related facility. CONSTITUTION: A data preprocessing unit obtains a data quantity for constructing an intelligent learning model by using real data transmitted from a sensor in real time and pseudo data generated from the real data. A data filter unit converts time series data into logical virtual transaction based on a length and a moving size of a time window. A real time damage sign predicting unit recognizes a damage sign state by using an intelligent damage sign pattern analyzing model and provides a result. [Reference numerals] (AA) Real Data; (BB,DD) Sensor 01-n; (CC) Generation Data; (EE) Data Viewer; (FF) Data Filter; (GG) Window Controller; (HH) Data Standard input; (II) Data Normalization; (JJ) Bayesian Network; (KK) SOM Learning; (LL) SOM cognition; (MM) Output Data; (NN) Dangerous sign 00%