UBC Theses and Dissertations
Control loop performance Lynch, Christopher
Control loop monitoring provides the opportunity to maintain high process performance. In this thesis, the regulation performance of a process is monitored by examining the time delay, static input-output relation, and the minimum achievable output variance of the process. The time delay estimate is obtained from the Fixed Model Variable Regressor Estimator and the static input-output relation is estimated using the Adaptive Nonlinear Modeller. The minimum achievable output variance of the process is determined by modelling the closed loop noise transfer function by a Laguerre network and finding the Laguerre gains of the network using Recursive Extended Least Squares estimation. An extensive simulation study was performed to examine the operation, usefulness and limitations of the monitoring tools. These simulations confirmed the operation and benefit of the monitoring methods, and defined their limitations. The monitoring tools were applied to data generated by industrial processes. The successful application to industrial processes highlighted the benefits obtainable by control loop monitoring. Due to the ease of applying the monitoring tools and the valuable insight they provide, these monitoring tools should be applied to any process where regulation performance is a concern.
Item Citations and Data