Analysis of control valves stiction quantification tool

Abstract

Control valve stiction is considered as one of the main sources of control loops nonlinearities which impacts plants profitability. In turn, this phenomenon hinders the plant from being operated at optimal conditions. Therefore, an efficient and accurate stiction quantification algorithm is required for accurate stiction compensation and timely scheduling of control valve maintenance. This research investigates the robustness and recommends improvements to the previously developed stiction quantification approach by Zabiri et al. The approach was tested under several operating conditions which were simulated in five case studies by using MATLAB software. The case studies investigated the impact of a wide range of stiction values, controller tuning, disturbance, time delay and noise on the quantification approach. The algorithm was found to be robust since it quantified the correct values of stiction regardless of the operating conditions. It was found that the accuracy of the quantification results depends on the process model accuracy, number of data samples and the search resolution. A number of improvements were recommended and validated by simulation in order to further enhance the current quantification approach. As conclusion, the algorithm can be applied on any type of process due to its robustness.