A wavelet theory-based adaptive trend analysis system for process monitoring and diagnosis

We discuss the development of a wavelet theory-based adaptive system for trend analysis (W-ASTRA). W-ASTRA performs process-monitoring and diagnosis. The main contributions of this paper are two fold. A wavelet theory based nonlinear adaptive algorithm has been developed for identification of trends from sensor data. In order to perform diagnosis using the identified trends, a knowledge base is required. Our second contribution is the development of an automated framework for knowledge base development. W-ASTRA uses the adaptive algorithm for identification of sensor trends and the knowledge base generated by the automated framework for diagnosing fault origins from the identified trends. The application of W-ASTRA is demonstrated on the Amoco Model IV FCCU.