APPARATUS AND METHOD FOR ONLINE SIGNAL DATA VALIDATION VIA AUTO ASSOCIATIVE BILATERAL KERNEL REGRESSION

摘要

The present invention relates to a technique for monitoring an industrial process using a weighted distance self-associative bilateral kernel regression (AABKR), the signal data to be measured by an on-line signal data verification device installed in an industrial facility The range kernel weight of the signal data is calculated based on the range change of, and the query time index related to the temporal change of the signal data is determined, and the time kernel weight of the signal data is calculated based on the determined query time index. And calculates bidirectional kernel weights of the signal data using the calculated range kernel weights and the calculated time kernel weights, and verifies process defects of the industrial equipment using the signal data and the calculated bidirectional kernel weights. It is about technology.