Application of non-parametric statistical methods to predict pumpability of geopolymers for well cementing

Prediction of geopolymer pumpability and setting time for well zonal isolation - Using machine learning and statistical based models

“…Furthermore, Hamie et al [ 7 ] constructed a DT module to predict the pumpability of geopolymers developed for well cementing applications. An additional discrete type model, logistic regression (LR) was applied in order to estimate the probability that a certain geopolymer mixture, subjected to different external properties (time, temperature), will attain a certain desired level of consistency.…”

“…

Fig. 1 Consistency data of the slurries, for a variety of geopolymer mixtures, at ambient pressure and two different bottomhole circulation temperatures [ 7 , 27 , 31 ].

…”

“…These prediction methods include regression, fuzzy logic, and Artificial Neural Network (ANN) [ [4] , [5] , [6] ]. While most of these publications investigate the prediction performance of geopolymer properties using a variety of mathematical models to be used in the field of construction and building materials, few scholars focused on geopolymer application in well cementing for Oil and Gas industry [ 7 ]. Despite this, computational modeling methods are widely used throughout the oil and gas value chain, with some products being commercialized, particularly in the field of reservoir engineering.…”