PDO is managing some 850 Electrical Submersible Pump (ESP) systems scattered across North & South fields, which is continue to grow in the next five years business plan. All ESP wells have real time down-hole sensors that measures intake and discharge pressures, intake and motor temperatures, vibration and current leakage. The oil producing fields are equipped with real time data transmission system where several data measurements; down hole (such as pump intake and discharge pressures and temperatures) and surface (such as volts, amps and frequency) are transmitted directly from the well site to the gathering stations, central control rooms and even to the engineers' desktop.
At present, PDO is deploying an integrated smart tool which will monitor, control, and optimize oil production and ESP performance to the various disciplines involved in oil production and optimization like Reservoir and Petroleum Engineers, Programmers, and Field Operation Teams. However, in order to enable these modern well surveillance systems, which often produce an overwhelming quantity of information but the data is often misleading or difficult to interpret, establishing the Pattern recognition of the trended real time data is key to make the software intelligent enough to be effective to the work places.
This paper will demonstrate how precise ESP, well and reservoir performance can be predicted from simple physical relationships and how these relate to the trends of surface and downhole data. A number of real field examples of data trends will be shown to illustrate how a proper understanding of these patterns will allow prompt ESP troubleshooting and ensure the correct actions are taken. The results are correlated with equipment pull and inspection reports to validate the diagnosis.
Pattern recognition trends and analysis will be presented for common problems such as hole in tubing, shut in at surface, ESP wear, blockage at pump intake, debris in pump, broken shaft, change in reservoir pressure, blockage at perforations, etc. A proper understanding of these trends will allow the correct settings of alarm and trips and assist in the implementation of semi-automated well surveillance and diagnostic system which being currently deployed in the Company. A pattern recognition analysis check sheet will be included in the paper to allow users to quickly interpret data trends and diagnose well, ESP and reservoir performance problems.