نوع مقاله : مقاله پژوهشی
کلیدواژهها
موضوعات
عنوان مقاله English
نویسندگان English
Drilling in gas hydrate formations in deep water poses significant risks and challenges. This research aims to determine the temperature window for drilling fluid in deep water drilling, after examining the relations of predicting the temperature of gas hydrate formation, using two machine learning algorithms, "polynomial regression in Python and multilayer perceptron in MATLAB". The predictions are used to establish the required temperature range for the drilling fluid. Among the examined methods for predicting gas hydrate formation temperatures, the Saharkhizan relationship, with an average relative error of 0.49, demonstrated the highest accuracy. Furthermore, for six gas compounds with varying specific gravities, the gas compound with a specific gravity of 0.61 yielded an R² value of 0.9743 using the polynomial regression algorithm. Similarly, the gas compound with a specific gravity of 0.65 showed the highest accuracy when using the multilayer perceptron algorithm.
کلیدواژهها English