Analysis of multi-factor influence on measurement of water content in crude oil and its prediction model

The measurement of water content in crude oil based on method of dielectric coefficient is affected by multi-factor, including temperature, salinity content, flow states of oil/water mixture and output characteristic of measuring-sensor, is regarded as a multi-input and single-output nonlinear system with time-variance, strong uncertainty and randomicity. In this paper, a measuring system based on multi-sensor is designed for experiment of oil/water two-phase flow, the influence relationship and complexity of multi-factor in the parameter detection of oil/water mixture is developed by numerical simulation combined with theoretical computation. Moreover, a multi-sensor data fusion method based on artificial neural network is presented to establish a prediction model for water content in crude oil, the simulation result shows this method is effective to deal with the influence of nonlinear characteristic of measuring-sensor, temperature and salinity on measurement of water content in crude oil, evidently improving the measuring accuracy of water content.