Vladimir Pechenkin and Biplay Chapagain
DV8 Energy
Modern controllers are required to estimate various parameters from field data to provide effective diagnostics and control of sucker rod pumping installations. In some cases, however, the data are not only corrupted by noise but also contain outliers that are in gross disagreement with the postulated model. If included, outliers can distort the fitting process so dramatically that the fitted parameters become arbitrary.
In such circumstances, the deployment of robust estimation methods is essential. This paper discusses the application of one of such estimators to rod pump systems. The approach is capable of identifying the outliers even when they constitute up to 50% of data. The problem that motivated this research is the estimation of the plunger leakage from the travelling valve check. Several other aspects of the system that can benefit from this method are also considered. The results are demonstrated using real data from the field.