# Long-term demand forecasting (building/ portfolio) / electricity and heating

**6. Load forecasting in medium to long (7-30 days) and long term (seasonal/year) term**

**6.1 Rationale & Link to BEYOND Apps**

Medium to long- and long-term load forecasting can play a pivotal role in the maintenance and sizing of the electrical infrastructure. Typically, long term load forecasting is done with quite a big granularity (i.e. only total energy consumption of the day is estimated) There is a growing interest in long term load forecasting with an at least hourly granularity[1],[2]. 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. This forecast will also be necessary for other WP5 task like T5.1. (without excluding their reusability in other applications that may need to enrich their feature list in the future).

**6.2 Overview of relevant implementations**

Forecasting in general is usually seen as a timeseries problem and so is in the case of load forecasting. Forecasting is usually done by using either a statistical or machine learning approach. Statistical approaches use techniques like ARIMA or exponential smoothing [3] while ML techniques use various flavours of neural networks such as Long-short term memory recurrent neural networks (LSTM-RNN) (Agrawal, 2018) or ensemble methods like swarm particle optimization [4] among others

**6.3 Implementation in BEYOND**

For beyond we intend to explore both statistical and ML methods in order to find the one that best suits our needs attending not only accuracy measurements like the normalised RMSE (nRMSE), the NMAE and MAPE metrics but also the ability to explain the obtained results. The model will be based on data coming from pilots and specifically from Cuerva but we also intend to experiment with introducing other external variables such as the plans for urban development in the area which might deeply affect long-term increase of powe demand.

**6.3.1. 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] Agrawal, R. K. (2018). Long term load forecasting with hourly predictions based on long-short-term-memory networks. 2018 IEEE Texas Power and Energy Conference (TPEC), 1-6

[2] Basaran Filik, U. a. (2009). ourly Forecasting of Long Term Electric Energy Demand Using a Novel Modeling Approach. 2009 Fourth International Conference on Innovative Computing, Information and Control , 115-118.

[3] Rahman, M. Z.-E.-H. (2016). Forecasting the long term energy demand of Bangladesh using SPSS from 2011–2040. 016 3rd International Conference on Electrical Engineering and Information Communication Technology (ICEEICT), 1-5.

[4] Hafez, A. A. (2016). Particle swarm optimization for long-term Demand Forecasting. 2016 Eighteenth International Middle East Power Systems Conference (MEPCON), 179-183.

Back to BEYOND_Baseline_Analytics