Skip to Main Content
The process industries produce a huge amount of data which needs to be compressed in order to reduce the cost of transmission and/or permanent storage. The motivation of any data compression technique is to reduce a given data set into a smaller one while preserving the important features in the data. However, data compression is found to have detrimental effects on the reliability/validity of control loop performance measures. When the controller performance indices estimated using the compressed data are used for performance assessment, we are inclined to make the errors in assessing the control loop performance. Consequently, we may end up taking corrective action which may not be necessary or fail to detect when corrective action is required. Hence, it is important to select a data compression technique which has little impact on performance measures. Comparison of different compression techniques is made using three statistical measures: the normalized Harris index, the root mean square error and the percentage difference in mean values. Simulation studies show that wavelet compression has less detrimental effect on control loop performance measures than alternative techniques.