Sizing and Life Cycle Assessment of Small-Scale Power Backup Solutions: A Statistical Approach

The increasing frequency of natural disasters, such as floods and wildfires, necessitates enhanced emergency management plans. Reliable backup power for critical infrastructure along evacuation routes is urgent. Using a case study approach, this study investigated small scale power backup solutions for a traffic light. We deployed a statistical approach to study the sizing problem for four alternative solutions: gasoline generator, battery pack, fuel cell system, and photovoltaic (PV) panels with a battery. Leveraging data from the State of California, we constructed probability distribution profiles to model power outage occurrence and duration and considered two representative high and low Global Horizontal Irradiance (GHI) locations. A Monte Carlo simulation, generating 10,000 power outage scenarios, was adopted to size the systems and evaluate their lifetime cost and Global Warming Potential (GWP). Following this approach allowed us to compare different systems and sizes based on the percentage of power outages that they can support, which is superior to sizing and analysis based on a single scenario. Our results show that the PV-battery system is the optimal choice, especially for areas with high GHI. The battery only option is a suitable choice only for short power outage scenarios, but increasing the battery size to cover prolonged outages will be both expensive and ungreen. Fuel cell systems, on the other hand, can be a better choice for the longest events; however, results show that, unlike the conventional belief, cheap gasoline generators can be comparable to fuel cell systems in terms of GWP due to the standby nature of power backup systems and the high CO2 emissions produced during the manufacturing process for fuel cell engine and hydrogen tanks. This study also explored additional factors for gasoline generators and fuel cells, such as PV panel shading, refueling cost and frequency, and the sensitivity of outcomes to component cost assumptions and variations.