A first-order meteorological Markov chain-based method for generating a long-term dataset of photovoltaic outputs

The paper sets out a method for generating a medium- to long-term PV output dataset based on a firstorder meteorological Markov chain. The method generates shortterm PV data through a generator and employs a filter to ensure that the data meets the upper and lower limits of power generation, thereby stitching together the medium- and long-term PV output datasets. In comparison with conventional methodologies, the Markov chain-based generator demonstrates superior capabilities in modelling shifts in meteorological states and in reflecting the impact of meteorological factors on PV output. By incorporating considerations of cloud transmittance and seasonal variations, the PV data generated by this method exhibits superior performance in terms of daily and monthly screening pass rates when compared with the Latin hypercube sampling and Monte Carlo methods. The experimental findings demonstrate that the dataset produced by this method exhibits enhanced data quality and broader coverage of scenarios, while the simulation time is reduced, rendering it well-suited for the simulation of medium- to long-term PV power generation datasets.