UBC Theses and Dissertations
Real time paper machine data wavelet analysis and cross directional actuator response identification Kan, Ka Keung
The development of an online automated diagnostics tool for analyzing paper machine data is useful to mill operators when monitoring the performance of a cross directional (CD) control system. The use of Object linking and embedding for Process Control (OPC) interface as a medium for client-server based data acquisition provides an efficient, robust and accurate data transfer between a CD control system and a monitoring workstation. Therefore, a real time diagnostics toolbox has been developed to help paper mill companies reduce costs and save time on troubleshooting. The Real Time Wavelet Toolbox (known as Waveplot and developed over several years at UBC, ) provides color maps of raw, wavelet and residue profiles for both basis weight and moisture. Moreover, machine directional (MD) and cross directional (CD) control performance indices as well as the standard deviations for controllable and non-controllable CD components are calculated and displayed. Other important plots such as power spectral density and two-sigma plots are also included in the toolbox for further analysis. Waveplot also helps mill operators identify the high-resolution CD actuator response shape and CD mapping through bump tests. The identification is achieved by using a continuous wavelet transform with the chosen CD actuator response shape model as the mother wavelet. The control performance index reflects recent views on CD performance index in terms of CD controllability given the knowledge of the identified CD actuator response shape. Furthermore, the identified actuator response centers could help mill operators investigate any problems due to CD mapping misalignment and shrinkage, particularly at the paper edges. The displays and identification results are validated and demonstrated by collecting paper profiles through the OPC server from a multi-grade, operating paper machine.
Item Citations and Data