A kind of detection method of the water content of crude oil in the porous medium

Abstract

With the development of science and technology, the petroleum industry plays an increasingly important role in people's lives. The water content of crude oil can not only predict the oil layer location and water level of oil wells, but also further infer the output and value of the remaining recoverable crude oil in the oil field, and formulate corresponding development plans. To sum up, it is of indispensable significance to monitor the water content of crude oil. This paper mainly introduces a method for calculating the water content of crude oil based on the improved ELM algorithm. Compared with the traditional machine learning algorithm, the algorithm not only optimizes the activation function with the Gaussian kernel function but also adopts the method of segmental modeling to ensure that the oil-water two-phase mixed state can be sufficiently usable. The experimental results show that the method has achieved good results in the application of the water content of crude oil detection devices based on the conductometric method, which is superior to the general BP neural network, the ELM algorithm, and the ELM algorithm that is only piece-wised but not optimized.