Long-term generation forecasting at DER level
19.1. DER forecasting in long(seasonal/year) term
19.1 Rationale & Link to BEYOND Apps
Long-term DER forecasting can play a pivotal role in the maintenance and sizing of the electrical infrastructure. Typically, there is no such thing as long-term DER forecasting. Literature of long-term predictions is focused in the prediction of a single renewable resource such as PV or wind turbines or in the long term prediction of influencing factors such as wind characteristics or solar irradiation. Literature surrounding DER influence in the grid exists but is taken as an influencing factor in long term net load prediction[1].
Since one of the goals of Beyond and in particular task T5.2 is to optimize network asset planning and sizing, we need long term hourly predictions to assess events like the possibility of network congestion (without excluding their reusability in other applications that may need to enrich their feature list in the future).
19.2 Overview of relevant implementations
Forecasting in general is usually seen as a timeseries problem and so is in the case of PV and wind turbine generation forecasting. Forecasting is usually done by using either a statistical or machine learning approach. Examples in literature of long term forecast use ML (machine learning) methods such as gated recurrent units [2] or fully recurrent neural networks [3].
19.3 Implementation in BEYOND
In beyond we intend to explore a two-tiered model mixing ML or statistical techniques to forecast generation with a model that tries to predict DER penetration into the grid. We will use data coming from the Cuerva pilot site as well as typical meteorological year data coming from the PVGIS platform [4]. We also want to explore the addition of other data which might help us predict DER penetration.
19.3.1. Analytics Libraries Employed
We intend to research python libraries such as Pandas, Keras, TensorFlow, Numpy and Scikit-learn but actual usage will depend on the final implementation
References
[1] Zhao, M. a. (2014). Medium and long term load forecasting method for distribution network with high penetration DG. 2014 China International Conference on Electricity Distribution (CICED), 442-444.
[2] Aslam, M. a.-M. (2019). Long-term Solar Radiation Forecasting using a Deep Learning Approach-GRUs. 2019 IEEE 8th International Conference on Advanced Power System Automation and Protection (APAP), 917-920.
[3] Sopian, H. A.-W. (2022). Long-term power forecasting using FRNN and PCA models for calculating output parameters in solar photovoltaic generation. Heliyon.
[4] PVGIS. (s.f.). Obtenido de https://joint-research-centre.ec.europa.eu/pvgis-photovoltaic-geographical-information-system/pvgis-tools/tmy-generator_en
[4] Godinho, Xavier, et al. "Forecasting Heating and Cooling Energy Demand in an Office Building using Machine Learning Methods." 2020 International Young Engineers Forum (YEF-ECE). IEEE, 2020.
Back to BEYOND_Baseline_Analytics