Impact-CO₂ project: 36 months final report

The Impact-CO₂ project enabled to improve knowledge to contribute to evaluation of possible exposure inside a building in case of CO₂ leakage from a geological site that would reach the near-surface. It uses an approach that combines experiments and modeling. Experiments at a representative scale were carried out on the PISCO₂ site in Spain. For the first time in the context of Carbon Capture and Storage (CCS), we performed an experiment that reproduced a near-surface leak including intrusion in a building prototype. During several-days experiments, we observed peaks of CO₂ concentrations in the prototype, likely correlated to atmospheric changes. This result is of prior importance in the context of CCS risk analysis to assess acute risks of exposure during short periods. Assessing this kind of risk requires indeed estimating maximal concentrations and not only average values. Experiments were also performed at the laboratory scale, with the objective to progress on the understanding of transport mechanisms' theoretical aspects. These experiments included unidimensional experiments with a glassy column filled with sand, with CO₂ injection at the bottom in the form of pulse, as well as experiments on a two-bulb apparatus. These experiments, combined with modeling developments, enabled significant progress in the understanding of influence on transport from gravity; gas composition, flux and humidity. The most important result with regard to risk assessment is the evolution of the dispersion coefficient with the flux intensity. This coefficient is generally assumed to be constant, while it is in fact highly variable. Based on a state of the art in several fields, we defined leakage risk scenarios. We focused on human indoor exposure. We considered an a priori conservative configuration where the leak occurs right below a building, in the unsaturated zone. Four compartments are involved in the analysis: the unsaturated zone, the slab, the air inside the building and the atmosphere. Scenarios were modelled with TOUGH2, coupled with the code SIREN developed by CSTB to integrate more precisely the building's aspects. The results were confronted with regulatory exposure thresholds. Model limitations, notably those identified thanks to laboratory experiments, were investigated and discussed with regard to risk assessment. Steady state simulations show a complex influence of the soil, foundations and building parameters on the resulting CO₂ concentration. They highlight a high variability of the flow rate entering the building, which may lead to indoor CO₂ concentration levels exceeding the short-term exposure thresholds, while most results remain below the long-term exposure threshold. Taking into account wind and outdoor temperature variations leads to significant variations of the resulting CO₂ concentration, more pronounced for a building with only natural ventilation than with mechanical ventilation. Atmospheric pressure variations induce transient effects that might lead to very high CO₂ concentration levels in indoor air. These phenomena may, for the range of values investigated, drive to concentrations exceeding the short-term exposure thresholds. The quantitative results have to be cautiously considered, notably because the model is very rough. Nevertheless these simulations show the significant influence of the meteorological conditions on CO₂ exposure. In a risk management perspective, where acute exposure must be assessed, these variations should not be neglected. In terms of risk assessment, the simple model considered in our study considers several safety margins. As a consequence it can be very conservative, and not conclusive. Two situations can indeed occur with such a model: 1) either the exposure is acceptable, and it is possible to conclude directly that the risk is acceptable; 2) or the exposure overpasses thresholds, and in such a case it is not possible to discriminate between over-conservative safety margins or true risk. In this latter case it would become necessary to realize simulations with more realistic safety margins. Our results plead for the use of an exposure model at the scale of the studied territory. If a need for monitoring CO₂ exposure in buildings is established, installing CO₂ sensors in the concerned buildings appears the most efficient solution. In case of leakage, a monitoring campaign should not only focus on identifying its cause, but also how it behaves in the near-surface. Regarding remediation, the first measure should consist in informing the populations and inviting them to carefully ventilate the building. If the leakage situation is expected to last, CO₂ sensors equipped with alarms and automatic ventilation systems controlled by the level of the CO₂ indoor concentration might be installed. Despite the severe limitations on the use of the model developed in the Impact-CO₂ project, we compiled an IMPACTCO₂-Toolbox from the results of the numerical simulations run in the project. This spreadsheet provides rough indicative values for estimating the intrusion flow rate in a building following a leakage, and allows converting it into a concentration that can be compared to exposure thresholds. A CO₂ geological storage site shall be selected, designed and operated in a way that prevents leakage out of the storage complex. Nevertheless the ability to understand what could happen and what could be the consequences is important in a risk management perspective. Additional studies should be conducted to further the effort performed in the Impact-CO₂ project, in particular regarding the influence of meteorological conditions on CO₂ migration in the near-surface, CO₂ atmospheric dispersion models, or the design of coupled underground-atmosphere CO₂ migration models to be able to assess the risk at the scale of a territory