dc.description.abstract |
Generally, Ordinary Portland Cement (OPC) is used as a well cement during the CO2
sequestration process; however, it shows adverse failures in a CO2-rich environment and loses
its isolation properties in a short time. Based on the previous findings on OPC-based gas well
cement, its uncertainty in providing effective well integrity is revealed. Therefore, studying a
novel well cement is one of the main requirements to conduct a sustainable CO2 sequestration
process. Among them, fly ash (FA)-based geopolymer has a higher prominence due to the ability
to reduce the gigantic amounts of fly ash piled up due to coal-fired power plant operations. The
compressive strength and CO2 permeability of well cement play major roles in downhole
conditions to maintain the wellbore integrity at different temperature and pressure variations.
This study was carried out to develop predictive models for compressive strength and
permeability of FA-based geopolymer cement using different independent variables. For this
purpose, databases were developed to collect data from many laboratory studies available in the
literature. Two models were developed for predicting 7 days of compressive strength of well
cement using linear and nonlinear multivariable regression (MVR) analyses and Artificial
Neural Network (ANN), and they were validated using the experimental data. One of the models
developed using Si/Al ratio and curing temperature as independent variables have shown a good
prediction accuracy with R2 values of 0.9332 for training data and 0.9761 for validating data.
In the case of developing prediction models for CO2 permeability, five equations were
developed under selected confining pressures using injection pressure and the curing
temperature as independent variables. Coefficient of determination values (R2) of 0.880, 0.955,
0.959, 0.964, and 0.980 were obtained for each trained data in categorised subgroups under
confining pressure values of 12, 16, 20, 25, and 35 MPa respectively for these developed
equations. |
en_US |