Weatherford Technology Holdings, LLC

Endre Anderssen

Hamed Chok

Jess V Ford

Methods and systems are described for estimating the level of contamination of downhole fluid and underlying composition using physical property measurements, and mathematical modeling of contamination functions and fluid property mixing laws. The proposed approaches enable computation of estimates of the pumping time needed to achieve a certain contamination threshold level and the determination of whether or not sampling is appropriate at the current point in time based on the predicted compositional properties of the formation fluid.

Real-time fluid contamination prediction using bilinear programming

A method is disclosed of inverting three dimensional fluid property distribution. The method includes comparing a detected NMR signal with a plurality of modeled signal values derived from precomputed values of NMR signal contribution values at prechosen (T, T, D) value tuples; identifying one or more modeled signals satisfying domain constraints and in respect of which an objective function involving a respective detected NMR signal and a modeled signal is optimized; selecting one or more of the solutions resulting in optimized objective; and using each selected optimized solution to characterize the one or more properties of fluid in the formation. The method also includes processing the resulting solutions as e.g. graphical or tabular data. Also disclosed is apparatus for performing the method.

Method of and apparatus for inverting three-dimensional fluid property distribution over the (T1,T2,D)domain from NMR measurements

Systems and methods for predicting fluid composition in a reservoir formation are disclosed. To accurately and efficiently predict the composition of a sample fluid, an absorption spectrum is obtained over a prefixed set of wavelength channels and a set of basis spectra is received for all constituents of interest in the sample fluid. A set of a-priori constraints on the obtained absorption spectrum and constituent properties to be determined is then obtained, before calculating a joint probability distribution of all constituents of interest based on the obtained spectrum, basis spectra and a-priori constraints. In one embodiment, qualitative states are also computed for all constituents of interest based on the obtained spectrum, basis spectra and a-priori constraints.

Growing community of inventors

Houston, TX, United States of America

Endre Anderssen