New online load forecasting system for the Spanish Transport System Operator

López, María Dolores Ruiz; Verdú, S.; Rodriguez, Ana; Veiras, Iago; Senabre, C.

doi:10.1016/j.epsr.2017.09.003

Cited by 25 publications

(40 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Both forecasting models under analysis are extracted from the STLF system currently working at Red Eléctrica de España (REE), the Transport System Operator in Spain. They have been thorughly described in [39], and have been running on the REE headquarters for over two years now. Both forecasting engines use the same data filtering system to discard outliers, usually caused by malfunctioning of the data acquisition systems.…”

Section: Forecasting Modelsmentioning

confidence: 99%

“…This non-linear function is, in our case, a feedforward neural network. Further description of this model can be found in [39]. Figure 1 shows a visualization of this type of networks working online.…”

Section: Neural Networkmentioning

confidence: 99%

“…where T med,d is the average temperature of day d, TH hot and TH cold are the thresholds for hot and cold days and CDD d and HDD d are the values of each series for day d. This technique requires the thresholds to be properly tuned to each location's effect on the load. This optimization process is described in [39] and the optimal threshold for each zone has been calculated. However, the robustness of each model against the variation of these values has been tested by introducing variations of up to 12 degrees on each threshold.…”

Section: Neural Network Size Redundancy and Computational Burdenmentioning

confidence: 99%

“…The days are classified into 34 exclusive categories some of which include several days under a general rule: Mondays, Wednesdays, national holidays, Mondays before a holiday… and others for specific and unique days: 1 May, 25 December, 1 January… In addition to the exclusive categories (each day can only be assigned to one of these), there are also 18 modifying categories that may be simultaneously active with an exclusive one. These include regional holidays, days affected by DST… The complete classification can be found in [39].…”

Section: Calendarmentioning

confidence: 99%

“…To sum up, it is also relevant to mention examples of real world applications [37][38][39]. The publishing of models that are validated through actual use by the industry instead of through lab conditions is especially important for the advancement of the field [2].…”

mentioning

confidence: 99%

See 4 more Smart Citations

Empirical Comparison of Neural Network and Auto-Regressive Models in Short-Term Load Forecasting

et al. 2018

Self Cite

View full text Add to dashboard Cite

Artificial Intelligence (AI) has been widely used in Short-Term Load Forecasting (STLF) in the last 20 years and it has partly displaced older time-series and statistical methods to a second row. However, the STLF problem is very particular and specific to each case and, while there are many papers about AI applications, there is little research determining which features of an STLF system is better suited for a specific data set. In many occasions both classical and modern methods coexist, providing combined forecasts that outperform the individual ones. This paper presents a thorough empirical comparison between Neural Networks (NN) and Autoregressive (AR) models as forecasting engines. The objective of this paper is to determine the circumstances under which each model shows a better performance. It analyzes one of the models currently in use at the National Transport System Operator in Spain, Red Eléctrica de España (REE), which combines both techniques. The parameters that are tested are the availability of historical data, the treatment of exogenous variables, the training frequency and the configuration of the model. The performance of each model is measured as RMSE over a one-year period and analyzed under several factors like special days or extreme temperatures. The AR model has 0.13% lower error than the NN under ideal conditions. However, the NN model performs more accurately under certain stress situations.

show abstract

Section: Forecasting Modelsmentioning

confidence: 99%

Section: Neural Networkmentioning

confidence: 99%

Section: Neural Network Size Redundancy and Computational Burdenmentioning

confidence: 99%

Section: Calendarmentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations