Leak detection in liquid subsea flowlines with no recorded feed rate

This paper presents a statistical method to detect leaks in subsea liquid flowlines when the inlet flow rate measurements are unavailable and conventional mass balance techniques cannot be used. Presently the only method used in these situations is to "predict" the inlet rate from a steady-state reservoir/wellbore model for use in a transient pipeline simulator. The purpose of this work was to examine the utility of a pressure loss model which requires only routine production data. This technique is derived from the statistical theory of estimation and testing, and treats fluctuations due to turbulence, transients, and measurement errors as statistical noise. This method compares two mean values of the pipeline resistance coefficient, R; the new updated value representing the potential leak and the old value being R without leak. Precision of the new method was demonstrated using the 9460-ft long, (3.64 in, i.d.) flowloop at LSU, with flowrates up to 10,000 bbl/day of water and pressures up to 2400 psi, The leak was simulated by an orifice which discharged the water from the flowline into a shut-in, well. An environmental (wellhead) pressure up to 2200 psi was applied to simulate hydrostatic pressure in the deep-sea environment. The results presented in this paper show how detection of small leaks relates to the number of R measurements, leak position, and statistical noise resulting from flow rate-pressure fluctuations, transients, and measurement uncertainties.