Abstract for: Forecasting Photovoltaic Decommissioning: Integrating Survival Analysis and System Dynamics for Long-Term PV Waste Projections

Rapid growth in photovoltaic (PV) installations across Europe raises concerns about future PV waste. Decommissioning is driven by a complex mix of technical, economic, and environmental factors, whereas empirical data on failures and deinstallations remain sparse. This uncertainty challenges planning of recycling capacity and circular‑economy strategies. We develop a System Dynamics model for France that integrates survival‑analysis lifetime distributions using conveyor‑belt ageing chains and Weibull‑based hazard functions. This structure avoids cohort blending and enables explicit modelling of early, constant, and wear‑out failures. The model incorporates installation scenarios, evolving technologies, and material compositions and was tested through behavioral validation and calibration using collected waste data. Simulations reveal wide but plausible ranges of future waste volumes. Annual PV waste is projected to reach 154,000–300,000 t by 2050 and 238,000–434,000 t by 2070, depending on lifetime assumptions and installation goals. Cumulative waste is expected to reach 1.95–3.85 Mt by 2050 and 6.0–11.4 Mt by 2070. Waste streams are dominated by young modules due to rapidly increasing installation rates and early failures. Integrating survival analysis methods into System Dynamics broadens the achievable uncertainty range and enables more transparent modelling of PV decommissioning. The findings highlight the strong influence of lifetime variance on long‑term forecasts and show that recyclers must prepare for heterogeneous waste streams, including different technologies. While uncertainty remains high, even conservative scenarios indicate substantial growth in PV waste, underscoring the need for improved monitoring and recycling infrastructure. AI has been used for text refinments and to summarise content.