AbstractThe lifetime estimation of photovoltaic (PV) systems is an important consideration in their life cycle. This paper presents a multi-level methodology to estimate the availability and the lifetime of a PV system using stochastic Petri networks and taking into account the time distribution to failure and to repair. The following components – module, wires, and inverter – are modeled with a Petri network. The evolution of the MTBF, availability, and outages over a 30-year period is simulated for different series/parallel configurations of connected PV modules. Findings are in agreement with reported experimental results, namely the evolution of the availability which tends to be near 96% for long times. Numerical results show that the highest MTBF is 3.2 years and the lowest outage is of 11.84 days/year, which are both obtained for the string configuration. The approach is interesting for the dimensioning of a PV plant or an installation.