The paper considers two machine learning methods for predicting the growth of watercut on an oil field with the calculation of basic convergence metrics.

Keywords: watercut, statistics, machine learning, time series, oil production

A comprehensive methodology for calculating assessment of wells reliability and condition in terms of submersible equipment operation, operating time and additional negative factors of operation through Boolean and fuzzy logic has been developed. The methodology includes retrospective analysis, calculation for the current moment and calculation of future possible failures with the preparation of various option of wells repair schedule and their distribution into risk zones.

Keywords: boolean logic, fuzzy logic, time to failure, bath function, weighting factor, membership function, screening method