Performance Evaluation of Deep Recurrent Neural Networks Architectures: Application to PV Power Forecasting

Smart grid systems require an accurate energy prediction from renewable sources to ensure high sustainability and power quality. For PV plants, a precise estimation of the generated PV power is crucial for the reduction of the production/demand unbalance. This essential need comes from the high variability of weather parameters during the PV electricity generation. Long Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) recurrent neural networks proved their high efficiency in forecasting applications. Thus, this paper proposes a comprehensive evaluation of the LSTM and GRU techniques for PV power estimation in the medium/long horizon. The evaluation is based on a fair assessment of the aforementioned architectures for one week and more than three months (98 days) periods.

[1]  Emad M. Ahmed,et al.  PV power forecasting using different Artificial Neural Networks strategies , 2014, 2014 First International Conference on Green Energy ICGE 2014.

[2]  R. Falkner.  The Paris Agreement and the new logic of international climate politics , 2016 .

[3]  Francesco Grimaccia,et al.  Hybrid model for hourly forecast of photovoltaic and wind power , 2013, 2013 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE).

[4]  Patrice Simardy,et al.  Learning Long-Term Dependencies with , 2007 .

[5]  Francesco Grimaccia,et al.  A Physical Hybrid Artificial Neural Network for Short Term Forecasting of PV Plant Power Output , 2015 .

[6]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[7]  Arvind Chel,et al.  Renewable energy technologies for sustainable development of energy efficient building , 2017, Alexandria Engineering Journal.

[8]  R. Betts,et al.  Climate change in cities due to global warming and urban effects , 2010 .

[9]  Salvatore Carlucci,et al.  A Review of Systems and Technologies for Smart Homes and Smart Grids , 2016 .

[10]  Tao Hong,et al.  Long Term Probabilistic Load Forecasting and Normalization With Hourly Information , 2014, IEEE Transactions on Smart Grid.

[11]  Gerald Steinmaurer,et al.  Solar Irradiance Forecasting, Benchmarking of Different Techniques and Applications of Energy Meteorology , 2010 .

[12]  Hu Du,et al.  Understanding the reliability of localized near future weather data for building performance prediction in the UK , 2016, 2016 IEEE International Smart Cities Conference (ISC2).

[13]  T. Ari,et al.  Life history predicts advancement of avian spring migration in response to climate change , 2009 .

[14]  Robert G. Watts,et al.  Global Warming and the Future of the Earth , 2007, Global Warming and the Future of the Earth.

[15]  Yoshua Bengio,et al.  Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling , 2014, ArXiv.

[16]  S. Elangovan.  Recent trends in sustainable development of renewable energy , 2017, 2017 International Conference on Advances in Electrical Technology for Green Energy (ICAETGT).

[17]  N. Lior,et al.  Vulnerability to climate change impacts of present renewable energy systems designed for achieving net-zero energy buildings , 2016 .

[18]  Bismark Singh,et al.  A Guide to Solar Power Forecasting using ARMA Models , 2018, 2019 IEEE PES Innovative Smart Grid Technologies Europe (ISGT-Europe).

[19]  M. Diagne,et al.  Review of solar irradiance forecasting methods and a proposition for small-scale insular grids , 2013 .