This paper will present calculation techniques to predict the cross flow rate resulting from a casing leak in a water injection well. The techniques do not require the conventional use of downhole flowmeter (spinner) to obtain the flow rate. Rather, continuous surface injection data prior to leak development and shut-in well data are used to estimate the casing leak cross flow rate. Well performance modeling and nodal analysis techniques were deployed to carry out the computations associated with the proposed method.
The case study of this paper is a tubingless water injector. The well was identified with an upward cross flow by a sudden drop of wellhead pressure to zero psi. To quantify the leak cross flow rate, the following calculation methodology was applied:
Generating a well performance model using pre-leak injection data: The case study's physical dimensions, most recent static bottomhole pressure, and injection pressures and rates are all used to build the inflow performance model (IPR). The calculated IPR is then calibrated using the wellhead (surface) injection pressures and rates data.
Generating an imaginary production well model: A production well mimicking the flow characteristics and properties of the case study is envisioned to simulate leak cross flow at shut-in conditions. The imaginary production well has the same reservoir pressure of the injection well and the productivity index of the imaginary well is assumed to be equal to the injectivity index of the injector. The imaginary production well's system node is selected to be at the leak point and performance curves are generated at different system's node pressures.
Calculating the system's node pressure of the imaginary production well: The pressure across the leak point is needed as an input to the production well model (which simulate the cross flow rate). The leak zone and water level depths were identified using a gauge ring survey. Hence, the system's node pressure was calculated using Bernoulli energy balance equation.
The results of this numerical methodology was verified by running an e-line deployed flowmeter (spinner survey). The results of the numerical methodology was 1.37 % over the actual spinner rate measurement.
Number of Pages
Looking for more?
Some of the OnePetro partner societies have developed subject- specific wikis that may help.