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

Autor: Hassan Hamie, Anis Hoayek, Bassam El-Ghoul, Mahmoud Khalifeh
Přispěvatelé: Vienna University of Technology (TU Wien), École des Mines de Saint-Étienne (Mines Saint-Étienne MSE), Institut Mines-Télécom [Paris] (IMT), Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes (LIMOS), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Centre National de la Recherche Scientifique (CNRS)-Université Clermont Auvergne (UCA)-Institut national polytechnique Clermont Auvergne (INP Clermont Auvergne), Université Clermont Auvergne (UCA)-Université Clermont Auvergne (UCA), Institut Henri Fayol (FAYOL-ENSMSE), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT), Département Génie mathématique et industriel (FAYOL-ENSMSE), Ecole Nationale Supérieure des Mines de St Etienne (ENSM ST-ETIENNE)-Institut Henri Fayol, Phoenicia University, University of Stavanger
Rok vydání: 2022
Předmět:
Zdroj: Journal of Petroleum Science and Engineering
Journal of Petroleum Science and Engineering, 2022, 212, pp.110333. ⟨10.1016/j.petrol.2022.110333⟩
ISSN: 0920-4105
1873-4715
DOI: 10.1016/j.petrol.2022.110333
Popis: International audience; As a potential alternative to Portland cement, geopolymers are getting wider acceptance in the scientific world. On a laboratory scale, the latter is being experimented repeatedly to extract valuable and valid results. To complement the experimental work and to make use of the data that resulted from previous experiments, statistical and mathematical models are developed. This article aims to anticipate test results, extract statistical relationships from measured properties, and therefore minimize the time and trials needed to run experiments in laboratories. Five independent input parameters are measured that cover the SiO2/K2O ratio, temperature, time, liquid to solid ratio and the total water content. For each set of these input variables, the consistency of geopolymers was obtained.Two statistical models have been developed in this regard, the Decision Tree, which is a heuristic machine learning model, and the Logistic Regression which is a probabilistic model that calculates and estimates the probability for a certain mixture, at different time, temperature, and other independent variables, to reach a certain consistency threshold.Both model results indicate sufficient performance, and the modelers can use such methods to predict the consistency (pumping time) trends of an untested geopolymer mixture. The results of our models are further validated by additional statistical tests, such as the receiver operating characteristic curve.
Databáze: OpenAIRE
Externí odkaz: