METHOD AND SYSTEM FOR ESTIMATION OF ERROR OR OMITTED MEASURED VALUES USING ARTIFICIAL NEURAL NETWORK

摘要

The present invention relates to an error measurement value estimation method using an artificial neural network and a system thereof. The present invention includes: an S1 step where a training set organizing unit organizes a self-reference training set including a normal measurement value into an input value and an output value by using the normal measurement value from a sensor system; an S2 step where a neural network organizing unit organizes an artificial neural network to calculate an estimated measurement value from the normal measurement value using the self-reference training set organized in the S1 step; an S3 step where an estimated data calculating unit inputs the normal measurement value after a measurement failure on a building structure through the sensor system among one or more measurement values and calculates the estimated measurement value by including an error measurement value after the measurement failure; an S4 step where an estimated data classifies the estimated measurement values obtained in the S3 step into sections based on the sensor number given to one or more sensors and corrects the estimated measurement values to make the estimated measurement value and the normal measurement value identical by applying a correction function by section; and an S5 step where an error measurement value extracting unit extracts a final value by estimating the error measurement value from the estimated measurement values obtained in the S4 step. [Reference numerals] (S1) Step of organizing a self-reference training set including a normal measurement value into an input value and an output value; (S2) Step of organizing an artificial neural network to calculate an estimated measurement value using the self-reference training set; (S3) Step of inputting the normal measurement value into the artificial neural network and calculating the estimated measurement value including an error measurement value; (S4) Step of correcting the estimated measurement value to make the estimated measurement value and the normal measurement value identical by changing the estimated measurement value; (S5) Step of extracting a final value by estimating the error measurement value from the estimated measurement values