Cerca nel sito per parola chiave

pubblicazioni - Memoria

Improving streamflow prediction for run-of-river hydropower operation by error correction: the Aniene River case study

pubblicazioni - Memoria

Improving streamflow prediction for run-of-river hydropower operation by error correction: the Aniene River case study

La previsione della produzione di energia elettrica da fonti rinnovabili è un mezzo importante per la gestione del sistema elettrico. In questo articolo si presenta un esempio di modello statistico in grado di migliorare la previsione della produzione idroelettrica prevedendo, nel punto di interesse, le fluttuazioni orarie di afflusso generate da altri impianti posti a monte.

The European policies’ focus on renewable energy expansion calls for a more efficient integration of distributed generation within the existing electric power systems. Growing large-scale renewable energy production will considerably challenge scheduling and dispatch decisions by increased variability and uncertainty in the supply patterns. Effective renewable energies forecasting has been identified as an essential tool to address the issue of efficiently operating power systems characterised by large renewable energies penetration. In this study, we developed and tested a streamflow hourly prediction system at a run-of-river power plant in the Aniene river basin, a medium-sized catchment in Central Italy The catchment is heavily exploited for energy production by different power companies which do not operate in coordination, hence upstream plant daily scheduling is mostly unknown to the plant manager at the study site. The model is composed of two modules: 1) a lumped, conceptual rainfall-runoff model, whose input consists of antecedent rainfall and temperature data, for estimating initial conditions, and precipitation and temperature forecasts provided by a Limited Area Meteorological model, to compute the streamflow prediction; 2) a data-driven error correction filter to correct the flow forecast provided by the rainfall-runoff model, taking into proper consideration the deterministic hourly signal related to energy generation upstream, which causes a significant degradation of the model predicting capability for hourly fluctuations in dry conditions. The correction module adopts a local modelling approach mapping different flow conditions into different model response mechanisms. Results show that the integrated model describes accurately all the flow conditions, either when the natural signal is prevailing in flood conditions, or the anthropic activity is predominant in dry conditions.

Progetti

Commenti