Вибір нейронної мережі для розпізнавання стану нафтової свердловини

Abstract

The paper discusses a way of determining the oil well state with the help of a neural network and methods of image recognition for time-current curves. Various types of neural networks were tested for efficiency of image recognition. A string of experiments studied the accuracy of recognition and speed of the circuit. The conclusions were made regarding the optimal structure of the neuron neural for the set tasks.