Prediction of CO2 storage site integrity with rough set-based machine learning
CO2 capture and storage (CCS) and negative emissions technologies (NETs) are considered to be essential carbon management strategies to safely stabilize climate. CCS entails capture of CO2 from combustion products from industrial plants and subsequent storage of this CO2 in geological formations or reservoirs. Some NETs, such as bioenergy with CCS and direct air capture, also require such CO2 sinks. For these technologies to work, it is essential to identify and use only secure geological reservoirs with minimal risk of leakage over a timescale of multiple centuries. Prediction of storage integrity is thus a difficult but critical task. Natural analogues or naturally occurring deposits of CO2, can provide some information on which geological features (e.g., depth, temperature, and pressure) are predictive of secure or insecure storage. In this work, a rough set-based machine learning (RSML) technique is used to analyze data from more than 70 secure and insecure natural CO2 reservoirs. RSML is then used to generate empirical rule-based predictive models for selection of suitable CO2 storage sites. These models are compared with previously published site selection rules that were based on expert knowledge.
KeywordsCO2 storage CO2 removal (CDR) Artificial intelligence (AI) Classification
This work was supported via the Philippine Higher Education Research Network (PHERNet) Sustainability Studies Program Granted to De La Salle University by the Commission on Higher Education of the Republic of the Philippines.
- Breiman L (1984) Classification and regression trees, 1st edn. Routledge, New YorkGoogle Scholar
- CASSEM (2011) CO2 aquifer storage site evaluation and monitoring. Heriot-Watt University, EdinburghGoogle Scholar
- Chadwick RA, Arts R, Bernstone C, May F, Thibeau S, Zweigel P (2008) Best practice for the storage of CO2 in saline aquifers. British Geological Survey Occasional Publication, NottinghamGoogle Scholar
- Cortes C, Vapnik V, Saitta L (1995) Support-vector networks. Mach Learn 20:273–297Google Scholar
- Duan X (2018) Application of artificial intelligence in evaluation and management of SEC oil and gas reserves. Chem Eng Trans 71:925–930Google Scholar
- IEA GHG (2009) CCS site characterisation criteria. IEA Greenhouse Gas R&D Programme, ReadingGoogle Scholar
- IPCC (2018) Summary for policymakers. In: Masson-Delmotte V, Zhai P, Pörtner HO, Roberts D, Skea J, Shukla PR, Pirani A, Moufouma-Okia W, Péan C, Pidcock R, Connors S, Matthews JBR, Chen Y, Zhou X, Gomis MI, Lonnoy E, Maycock T, Tignor M, Waterfield T (eds) Global warming of 1.5 °C. An IPCC special report on the impacts of global warming of 1.5 °C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and efforts to eradicate poverty. World Meteorological Organization, GenevaGoogle Scholar
- Mahajan P, Kandwal R, Vijay R (2012) Rough set approach in machine learning: a review. Int J Comput Appl 56:1–13Google Scholar
- Minx JC, Lamb WF, Callaghan MW, Fuss S, Hilaire J, Creutzig F, Amann T, Beringer T, De Oliveira Garcia W, Hartmann J, Khanna T, Lenzi D, Luderer G, Nemet GF, Rogelj J, Smith P, Vicente Vicente JL, Wilcox J, Del Mar Zamora Dominguez M (2018) Negative emissions—part 1: research landscape and synthesis. Environ Res Lett 13:063001CrossRefGoogle Scholar
- ProSoft (1999) User’s guide ROSE 2 rough set data explorer. idss.cs.put.poznan.pl/site/fileadmin/projects-images/rose_manual.pdf. Accessed 18 March 2019