Effect of Al2O3 nanoparticle, Silica Powder microparticle, and Fibers on the newly formulated geopolymer: Experimental and Modelling Studies.
Master thesis
Permanent lenke
https://hdl.handle.net/11250/3089275Utgivelsesdato
2023Metadata
Vis full innførselSamlinger
- Studentoppgaver (TN-IEP) [341]
Sammendrag
When the productivity of a reservoir becomes uneconomical, the final fate of the well is to be plugged and abandoned. NORSOK D-10 defines the well integrity criteria and sets criteria for cement properties to reduce the risk level of undesired leakage from the reservoir to the surface. However, well integrity survey data from the Norwegian Continental Shelf (NCS) shows that out of the seventy-five production and injection wells, cement recorded about an 11% integrity issue rate. Surveys from different countries around the world show that cement failure is one of the major reasons for well integrity failure. This shows that cement does not satisfy the NORSOK D-10 standard’s requirement. Currently, the oil and gas industry is searching for new alternative materials that can be used for plug and abandonment operations. Among others, geopolymer is one of the alternative candidate materials. Several studies have shown that the addition of nanoparticles and fibers improved the strength and quality of geopolymers. In this thesis work, a total of four experimental test designs have been formulated to investigate the effect of Silica powder, Al2O3 nanoparticle solution, carbon fiber, and human hair on the properties of 12M NaOH and 14M NaOH-based neat geopolymer. In addition, the properties of the newly formulated neat geopolymer were compared with Portland G class cement. The geopolymer plugs were cured for 3 and 10 days at 62oC and atmospheric pressure. Results showed that silica powder increased the strength of the geopolymer with an increase in concentration. The optimum concentrations of Al2O3 nanoparticles and fibers have also shown positive impacts on the neat geopolymer. Furthermore, an empirical UCS wave velocity model was developed for the measured data, which showed a good prediction of the test dataset.