Fast and interpretable electricity consumption scenario generation for individual consumers

Abstract

To enable the transition from fossil fuels towards renewable energy, the low-voltage grid needs to be reinforced at a faster pace and on a larger scale than was historically the case. To efficiently plan reinforcements, one needs to estimate the currents and voltages throughout the grid, which are unknown but can be calculated from the grid layout and the electricity consumption time series of each consumer. However, for many consumers, these time series are unknown and have to be estimated from the available consumer information. We refer to this task as scenario generation. The state-of-the-art approach that generates electricity consumption scenarios is complex, resulting in a computationally expensive procedure with only limited interpretability. To alleviate these drawbacks, we propose a fast and interpretable scenario generation technique based on predictive clustering trees (PCTs) that does not compromise accuracy. In our experiments on three datasets from different locations, we found that our proposed approach generates time series that are at least as accurate as the state-of-the-art while being at least 7 times faster in training and prediction. Moreover, the interpretability of the PCT allows domain experts to gain insight into their data while simultaneously building trust in the predictions of the model.

Figures (6)

• Figure 1: The proposed predictive clustering technique performs on par with the existing data-driven approach, and, on the London dataset, slightly outperforms it. The bars show the average ES over all folds, the gray lines shows the standard deviation of ES over the folds.
• Figure 2: Evolution of the energy score as a function of training set size for every scenario generation method on the Flanders dataset.
• Figure 3: Training and prediction times as a function of training set size for every method on the Flanders dataset. For the data-driven and expert-based method, the relative execution time compared to predictive clustering is shown as well (e.g. for 250 years in the training set, data-driven needs 29.7 times more time to train a model than the proposed approach).
• Figure 4: The tree learned with predictive clustering on the Flanders dataset
• Figure 5: Visualization of the influence of the split on sunHour for consumers with a yearly consumption $\leq 4285$ kWh and PV power $> 4$ kVA for the Flanders dataset. Quantiles of all time series in the cluster are shown.