In many photovoltaic projects, performance assessment cannot rely on a single deterministic estimate. Instead, results are expressed in terms of exceedance probabilities that quantify the likelihood of achieving or exceeding a given level of performance.
Indicators such as P50 or P90 are commonly used in energy yield assessments and financial evaluations to characterise risk. These indicators are central to bankability analyses, contractual discussions, and investment decisions.
LuSim supports exceedance probability analysis by explicitly quantifying and propagating uncertainties along the photovoltaic modelling chain, rather than relying on generic margins or fixed uncertainty percentages.
Exceedance analysis in LuSim is strictly separated from design optimisation. It is performed for a defined reference configuration, corresponding to a specific system layout, geometry, and set of operating assumptions.
The objective is not to explore alternative designs, but to quantify the residual uncertainty associated with the selected design. This uncertainty reflects limitations in data, modelling assumptions, environmental variability, and system behaviour that cannot be fully resolved at the design stage.
This distinction ensures that exceedance probabilities represent risk and confidence levels, not deliberate design variability.
LuSim follows a system level view of photovoltaic energy yield, where the conversion of solar resource into delivered electricity is represented as a sequence of gains and losses. Each stage of this modelling chain introduces both a deterministic contribution and an associated uncertainty.
Uncertainty sources may include solar resource assessment, irradiance transposition, shading and mismatch, soiling, module and inverter performance, system availability, and long term degradation. The relative importance and maturity of these sources vary depending on technology, site conditions, and project phase.
By making uncertainty sources explicit and traceable, LuSim avoids treating uncertainty as a single aggregated correction and instead links confidence bounds to identifiable physical and modelling assumptions.
Exceedance probabilities require the propagation of uncertainties through the full PV energy yield and financial modelling chain.
LuSim supports several uncertainty propagation strategies depending on the context and available information. These include Gaussian approximations for small and well behaved uncertainties, convolution based approaches for asymmetric or bounded distributions, and Monte Carlo sampling when correlations or nonlinearities are significant.
The choice of propagation method is part of the modelling trade off, balancing accuracy, transparency, and computational effort. Importantly, the resulting distributions are always anchored to the same reference configuration.
LuSim supports the derivation of exceedance probabilities for energy related indicators such as annual energy yield or delivered energy.
By analysing the resulting output distributions, values such as P50 or P90 can be extracted directly. These values reflect uncertainty driven by site specific geometry, shading complexity, and modelling assumptions rather than generic industry averages.
This approach improves the credibility and interpretability of energy based exceedance metrics in complex PV projects.
Exceedance probabilities can also be applied to financial performance indicators derived from the reference configuration.
LuSim supports exceedance analysis for quantities such as revenue, net present value, internal rate of return, or levelised cost of energy. These indicators are computed consistently for each simulated outcome using the same financial assumptions.
This allows financial risk to be evaluated as a direct consequence of technical and modelling uncertainty, rather than as a separate or purely financial adjustment.
Uncertainty levels are not static across the lifetime of a PV project. In the pre construction phase, exceedance probabilities reflect limited information and conservative assumptions. During operation, monitoring data and diagnostics can reduce some uncertainty components while revealing others.
LuSim is designed to support this evolution by allowing uncertainty assumptions to be refined as better data becomes available. This makes it possible to revisit exceedance metrics over time and to distinguish between reducible and irreducible sources of uncertainty.
Exceedance probabilities produced with LuSim are intended to support informed risk assessment and transparent decision making.
By grounding uncertainty quantification in detailed physical modelling and traceable assumptions, LuSim provides exceedance metrics that are consistent with the actual complexity of the photovoltaic system being assessed.
This approach is particularly relevant for projects where geometry driven effects, innovative configurations, or atypical environments make simplified uncertainty treatments insufficient.
English English 
French
Spanish
Portuguese
