COMPRESSION OF CHEMICAL PROCESS DATA BY FUNCTIONAL APPROXIMATION AND FEATURE-EXTRACTION

Effective utilization of measured process data requires efficient tech niques for their compact storage and retrieval as well as for extracti ng information on the process operation. Techniques for the on-line co mpression of process data were developed based on their contribution i n time and in frequency wing the theory of wavelets. Existing techniqu es for compression via wavelets and wavelet packets are inconvenient f or on-line compression and are best suited for stationary signals. The se methods were extended to the on-line decomposition. and compression of nonstationary signals via time-varying wavelet packets. Various cr iteria for the selection of the best time-varying wavelet packet coeff icients ave derived. Explicit relationships among the compression unti e, local and global errors of approximation, and features in the signa l were derived and used for efficient compression. Extensive case stud ies on industrial data demonstrate the superior performance of wavelet -based techniques as compared to existing piecewise linear techniques.