Distributed photovoltaic bus-level power prediction method

The invention relates to the technical field of power systems, and discloses a distributed photovoltaic bus-level power prediction method, which comprises the following steps of: 1, calculating starting time and deadline of daily photovoltaic output; 2, classifying the users; 3, constructing a prediction model based on a long short-term memory network; and step 4, distributed roof photovoltaic day-ahead power prediction: using sunrise and sunset time to approximately replace starting time and ending time of photovoltaic output, carrying out hierarchical clustering on distributed photovoltaic users based on shape distance, constructing a prediction model based on a long-short-term memory network for each type of distributed photovoltaic users, and carrying out distributed roof photovoltaic day-ahead power prediction. Model parameters are trained, short-term power of each type of users is predicted, and prediction results of each type of distributed photovoltaic users are added. A user classification method of