<p>Small modular reactors (SMRs) are expected as a suitable candidate to fulfill energy needs in the future. The regulation of the emergency planning zone (EPZ) has been a controversial issue. The possibility of smaller EPZs because of their small core size and passive safety functions is still under discussion. The major emergency responses to radiological incidents in the early phase are evacuation from the area and shelter-in-place within a building. Comparison between the dose incurred during evacuation and that with shelter-in-place is necessary to consider the proper protective actions. The effect of shelter-in-place from small modular reactor hypothetical accident was studied. The source term came from a long-term station blackout (LTSBO) and loss of cooling accident (LOCA), and the time change of air concentration and the ground deposition data through the atmospheric spread around the plant was calculated with Radiological Assessment System for Consequence Analysis (RASCAL), a software developed by United States Nuclear Regulatory Commission (NRC) to provide dose projection around the plant. Then general one-story and two-story houses were set up, and 6 wall materials were selected for calculating indoor doses. Cloudshine and groundshine were calculated with Monte Carlo methods. In addition, the conservation of mass, air flow model was established to evaluate the inhalation for sheltered cases. The shielding function of each house for each pathway was evaluated by comparing the indoor dose with outdoor dose. The projected dose for sheltered cases was much smaller than that for unsheltered cases. Even though the projected dose will not completely perish, it was quite effective to reduce radiation exposure and can be superior to evacuation. The result will be a basis for calculating the radiological dose for sheltered cases in case of nuclear emergency for SMRs, which will be valuable to have a more effective emergency planning.</p>

En les darreres dècades el nombre d'accidents ocurreguts en la industria química i durant el transport de mercaderies perilloses ha augmentat substancialment, registrant-se en la seva majoria en zones densament poblades. Els núvols de gasos tòxics solen originar-se en aquests accidents i malgrat que són menys probables que altres tipus d'accidents, poden afectar grans extensions i contaminar zones poblades, provocant greus conseqüències. Això comporta un gran repte per a les autoritats civils, que han d'avaluar i decidir l'àrea que cal evacuar i l'àrea en la que s'ha d'implementar el confinament com a mesura de protecció. L'avaluació de l'efectivitat del confinament comprèn tres etapes fonamentals: el càlcul de la dispersió exterior, el càlcul de la concentració interior en funció de la concentració exterior i l'avaluació dels efectes adversos per a la salut. Aquesta tesi s'enfoca principalment en l'estudi de la segona etapa, la qual és funció de la taxa d'infiltració d'aire en les edificacions.Inicialment es va realitzar una extensa revisió bibliogràfica sobre les tres etapes, fent èmfasi en la cerca de models pel càlcul de la concentració interior, la taxa d'infiltració y l'hermeticitat de les vivendes. Posteriorment, a través d'una anàlisi de sensibilitat es trobà que la taxa de renovació d'aire té una gran influencia sobre l'efectivitat del confinament i, a més, atès que aquesta varia per cada edificació, el coneixement de la seva distribució en una població és necessari per a una avaluació adequada de l'efectivitat del confinament, ja que suposar-la constant per a totes les edificacions pot comportar sobreestimacions o subestimacions del radi d'evacuació. Per tant, amb la finalitat d'obtenir una aproximació de la distribució de l'hermeticitat, es va aplicar el model desenvolupat pel Lawrence Berkeley National Laboratory (LBNL), que prové de dades de vivendes nord-americanes, a les vivendes catalanes. De tota manera, els resultats obtinguts es trobaven esbiaixats a les zones climàtiques, essent les prediccions per a vivendes ubicades en zones seques més hermètiques que en zones humides. En el cas de Catalunya, on les tècniques constructives no varien significativament d'una zona a una altra i la majoria de vivendes estan construïdes a base de materials pesats, no és d'esperar una diferència tan marcada com la predita pel model del LBNL. Per tant, es va decidir desenvolupar un model per a les vivendes catalanes utilitzant la base de dades de taxes d'infiltració de vivendes unifamiliars del CETE de Lyon, ja que aquestes vivendes tenen més similitud amb les vivendes catalanes que no pas les nord-americanes.El model desenvolupat, denominat UPC-CETE, permet estimar l'hermeticitat de les vivendes unifamiliars en funció de l'àrea, el número de pisos, l'edat i el tipus d'estructura constructiva: lleugera o pesada. Els valors d'hermeticitat predits amb aquest model foren menors que els obtinguts amb el model del LNBL, tal com s'esperava. Finalment, per tal de validar i millorar el model desenvolupat, es van realitzar mesures de la taxa de renovació d'aire en diverses vivendes de Catalunya i també en habitacions prèviament condicionades per ser utilitzades com a refugi, per tal d'avaluar la reducció guanyada sobre la taxa de renovació de tota la vivenda. Com a mitjana, s'obtingueren reduccions d'un 35% i es trobà que les reduccions més grans tenien lloc en vivendes antigues, amb àrees petites d'una o dues plantes. El model UPC-CETE millorat a partir dels resultats obtinguts en les proves experimentals, s'incorporà a la metodologia per avaluar l'efectivitat del confinament en l'etapa d'estimació de la taxa de renovació d'aire, evitant l'ús d'un valor constant per a totes les vivendes i promovent així l'ús d'una distribució d'aquest paràmetre per secció censal afectada dins la població. / During the last decades the number of accidents in chemical industries and during transportation of hazardous substances has significantly increased, with most of them occurring in highly populated areas. One of the possible accidents is a toxic gas cloud, which although less common than other major hazards could affect larger areas reaching populated zones and producing more severe consequences. This implies then, a great challenge to emergency managers who must plan and decide the areas where protection measures should be implemented: shelter in place and/or evacuation. The assessment of the effectiveness of shelter in place is subjected to three main stages: the calculation of the outdoor gas dispersion, the estimation of indoor concentration from outdoor concentration and the evaluation of human vulnerability. This thesis is mainly focused on the study of the second stage which is primarily a function of buildings leakage.Initially we performed a bibliographic survey with special interest on the models to estimate indoor concentration from outdoor concentration, airtightness of dwellings and ventilation models. Then, through a sensitivity analysis, we found that the air exchange rate has a great influence on the effectiveness of shelter in place. Moreover, since this parameter is different for each building, the knowledge of the distribution of this variable in the affected population would lead to a more accurate assessment of the effectiveness of shelter in place, because if we assume it as a fix value, constant for all buildings, over or underestimations of the evacuation radius may occur. Therefore, with the aim of making an estimation of the airtightness distribution in Catalunya, we applied the model developed by the Lawrence Berkeley National Laboratory (LBNL), a model based on data from North American dwellings, to Catalan dwellings. The results obtained were influenced by climate zones, due to the coefficients of the model, being more airtight the predictions for dwellings located in dry climates than for dwellings in humid zones. In the case of Catalunya, where constructions techniques do not differ significantly from one zone to another and most of the dwellings consist of a heavy structure, a difference such as that predicted by the model of the LBNL is not expected. Consequently, we decided to develop a model for Catalan dwellings using the air leakage database from the CETE de Lyon, since French dwellings are more likely to Catalan dwellings than US dwellings. The model developed, named the UPC-CETE model, predicts the airtightness of single-family dwellings as a function of the floor area, the age, the number of stories and the structure type: light or heavy. The airtihgtness values predicted with this model were smaller than those predicted with the model of the LBNL, as was expected. Finally, in order to validate and improve the model developed we carried out a series of trials to measure the air exchange rate in some Catalan dwellings. Measurements in sealed rooms were also performed with the aim of assessing the reduction gained on the air exchange rate with regards to the air exchange rate of the whole dwelling. On average, we obtained reductions of 35% and found that larger reductions belonged to old dwellings with small floor areas and 1 or 2 stories. The improved model was incorporated on the methodology to assess shelter in place effectiveness on the stage concerning the estimation of the air exchange rate of the dwellings located on the affected zone; therefore, the assumption of a constant value is avoided. These measurements and the model constitute therefore the first proposal for estimating the airtightness distribution of single-family dwellings that could be used by Catalan authorities for emergency response planning.