OLUWAMUYIWA OLATUNDE OLADELE

PhD Graduate

PhD program:: XXXV


supervisor: Prof. Paola Russo

Thesis title: Quantitative Risk Assessment of Hydrogen Fuel Cell Electric Vehicles in Road Infrastructures

ABSTRACT A new methodology has been developed (The Sapienza university of Rome (URS) in collaboration with the Denmark Technical University (DTU) to calculate probability of fatality, probability of structural failure, individual risk and hazard distance associated with an accident involving a hydrogen powered vehicle in a road tunnel. Also, an event tree was also developed to show all the possible scenarios that can occur in the event an HFCEV is involved in an accident. Furthermore, a Bayesian network model was developed to give a probabilistic analysis of the risk scenarios of a HFCEV involved in a road tunnel. Bayesian network model is a dynamic tool and also, it is able to give a relationship between cause to effect and also interrelationship between variables in the model. Finally, QRA was carried out for hydrogen refuelling station , a compressed natural gas refuelling station a nd a liquefied petroleum gas station. From the methodology developed, in the case of tank rupture, the individual risk is in the range of 6.8 ×10^(-4) to 1.0 × 10^(-5) fatality per year up to a distance of 160 m from the centre of the tunnel for tank type IV of 62.4L SoC =99%, and of 100 m for SoC=40%. Assuming a risk acceptance criterion of 10^(-6) fatality per year, a fatality hazard distance of 375 m and 240 m is evaluated for SoC=99% and SoC=40%, respectively. On the other hand, using as risk acceptance criterion of 10^(-5) fatality per year, hazard distance is reduced to 160 m and 100 m for SoC=99% and SoC=40%, respectively. The Bayesian network model followed the principles of fire science and engineering which was obvious /evident upon updating with new evidence and thus will be a valuable tool for prediction , diagnosis , causal relationship and intercausal relationship has new data on HFCEV become available. The QRA for the hydrogen refuelling station is 3.192E-006 Fatalities/year, for the CNG is 2.929 E-005 Fatalities/year and the LPG 1.560E-007 Fatalities/year, which are all within the acceptable limits of standard international organizations. . ABSTRACT A new methodology has been developed (The Sapienza university of Rome (URS) in collaboration with the Denmark Technical University (DTU) to calculate probability of fatality, probability of structural failure, individual risk and hazard distance associated with an accident involving a hydrogen powered vehicle in a road tunnel. Also, an event tree was also developed to show all the possible scenarios that can occur in the event an HFCEV is involved in an accident. Furthermore, a Bayesian network model was developed to give a probabilistic analysis of the risk scenarios of a HFCEV involved in a road tunnel. Bayesian network model is a dynamic tool and also, it is able to give a relationship between cause to effect and also interrelationship between variables in the model. Finally, QRA was carried out for hydrogen refuelling station , a compressed natural gas refuelling station a nd a liquefied petroleum gas station. From the methodology developed, in the case of tank rupture, the individual risk is in the range of 6.8 ×10^(-4) to 1.0 × 10^(-5) fatality per year up to a distance of 160 m from the centre of the tunnel for tank type IV of 62.4L SoC =99%, and of 100 m for SoC=40%. Assuming a risk acceptance criterion of 10^(-6) fatality per year, a fatality hazard distance of 375 m and 240 m is evaluated for SoC=99% and SoC=40%, respectively. On the other hand, using as risk acceptance criterion of 10^(-5) fatality per year, hazard distance is reduced to 160 m and 100 m for SoC=99% and SoC=40%, respectively. The Bayesian network model followed the principles of fire science and engineering which was obvious /evident upon updating with new evidence and thus will be a valuable tool for prediction , diagnosis , causal relationship and intercausal relationship has new data on HFCEV become available. The QRA for the hydrogen refuelling station is 3.192E-006 Fatalities/year, for the CNG is 2.929 E-005 Fatalities/year and the LPG 1.560E-007 Fatalities/year, which are all within the acceptable limits of standard international organizations. . ABSTRACT A new methodology has been developed (The Sapienza university of Rome (URS) in collaboration with the Denmark Technical University (DTU) to calculate probability of fatality, probability of structural failure, individual risk and hazard distance associated with an accident involving a hydrogen powered vehicle in a road tunnel. Also, an event tree was also developed to show all the possible scenarios that can occur in the event an HFCEV is involved in an accident. Furthermore, a Bayesian network model was developed to give a probabilistic analysis of the risk scenarios of a HFCEV involved in a road tunnel. Bayesian network model is a dynamic tool and also, it is able to give a relationship between cause to effect and also interrelationship between variables in the model. Finally, QRA was carried out for hydrogen refuelling station , a compressed natural gas refuelling station a nd a liquefied petroleum gas station. From the methodology developed, in the case of tank rupture, the individual risk is in the range of 6.8 ×10^(-4) to 1.0 × 10^(-5) fatality per year up to a distance of 160 m from the centre of the tunnel for tank type IV of 62.4L SoC =99%, and of 100 m for SoC=40%. Assuming a risk acceptance criterion of 10^(-6) fatality per year, a fatality hazard distance of 375 m and 240 m is evaluated for SoC=99% and SoC=40%, respectively. On the other hand, using as risk acceptance criterion of 10^(-5) fatality per year, hazard distance is reduced to 160 m and 100 m for SoC=99% and SoC=40%, respectively. The Bayesian network model followed the principles of fire science and engineering which was obvious /evident upon updating with new evidence and thus will be a valuable tool for prediction , diagnosis , causal relationship and intercausal relationship has new data on HFCEV become available. The QRA for the hydrogen refuelling station is 3.192E-006 Fatalities/year, for the CNG is 2.929 E-005 Fatalities/year and the LPG 1.560E-007 Fatalities/year, which are all within the acceptable limits of standard international organizations. .

© Università degli Studi di Roma "La Sapienza" - Piazzale Aldo Moro 5, 00185 Roma