Differential pipe sticking in oil wells parameters using neural algorithms - fireflies in one of the fields in southtern Iran

The phenomenon of stuck pipe in the drilling industry has been one of the major problems that increase costs by increasing the drilling of a well is drilling. Generally, two types of pipe sticking hydrocarbon reservoirs occur during drilling include: consuming mechanical tube and differential pipe sticking. In this paper data pipe sticking out of the 12 wells from existing wells in Iran s southwestern oil fields have been used for this purpose. The optimal amount is equal to 60 are listed in the study. In general, it was found that a neural network with three hidden layers has the best performance in detecting pipe is stuck. Also, statistical analysis results obtained by fireflies predictive dialer algorithm neural tube using parameters TPR, SPC, ROC and TCA is detected The neural network developed able to predict with 72% accuracy related to stuck pipe, with an accuracy of 93.9% able to predict with accuracy of 89.4 of pipe sticking able to predict both phenomena are correct.

Keywords:

differential pipe sticking , sensitivity analysis , neural networks , Firefly Algorithm