José Ricardo Andrade

Abstract
In the power system decarbonization roadmap, novel grid management tools and market mechanisms are fundamental to solving technical problems concerning renewable energy forecast uncertainty. This work proposes a predictive algorithm for procurement of grid flexibility by the system operator (SO), which combines the SO flexible assets with active and reactive power short-term flexibility markets. The goal is to reduce the cognitive load of the human operator when analyzing multiple flexibility options and trajectories for the forecasted load/RES and create a human-in-the-loop approach for balancing risk, stakes, and cost. This work also formulates the decision problem into several steps where the operator must decide to book flexibility now or wait for the next forecast update (time-to-decide method), considering that flexibility (availability) price may increase with a lower notification time. Numerical results obtained for a public MV grid (Oberrhein) show that the time-to-decide method improves up to 22% a performance indicator related to a cost-loss matrix, compared to the option of booking the flexibility now at a lower price and without waiting for a forecast update.

Abstract
This paper explores the application of Deep Learning techniques to forecast electricity market prices. Three Deep Learning (DL) techniques are tested: Dense Neural Networks (DNN), Long Short-Term Memory Networks (LSTM) and Convolutional Neural Networks (CNN); and two non-DL techniques: Multiple Linear Regression and Gradient Boosting (GB). First, this work compares the forecast skill of all techniques for electricity price forecasting. The results analysis showed that CNN consistently remained among the best performers when predicting the most unusual periods such as the Covid19 pandemic one. The second study evaluates the potential application of CNN for automatic feature extraction over a dataset composed by multiple explanatory variables of different types, overcoming part of the feature selection challenges. The results showed that CNNs can be used to reduce the need for a variable selection phase.

Abstract
Network human operators' decision-making during grid outages requires significant attention and the ability to perceive real-time feedback from multiple information sources to minimize the number of control actions required to restore service, while maintaining the system and people safety. Data-driven event and alarm management have the potential to reduce human operator cognitive burden. However, the high complexity of events, the data semantics, and the large variety of equipment and technologies are key barriers for the application of Artificial Intelligence (AI) to raw SCADA data. In this context, this paper proposes a methodology to convert a large volume of alarm events into data mining terminology, creating the conditions for the application of modern AI techniques to alarm data. Moreover, this work also proposes two novel data-driven applications based on SCADA data: (i) identification of anomalous behaviors regarding the performance of the protection relays of primary substations, during circuit breaker tripping alarms in High Voltage (HV) and Medium Voltage (MV) lines; (ii) unsupervised learning to cluster similar events in HV line panels, classify new event logs based on the obtained clusters and membership grade with a control parameter that helps to identify rare events. Important aspects associated with data handling and pre-processing are also covered. The results for real data from a Distribution System Operator (DSO) showed: (i) that the proposed method can detect unexpected relay pickup events, e.g., one substation with nearly 41% of the circuit breaker alarms had an 'atypical' event in their context (revealed an overlooked problem on the electrification of a protection relay); (ii) capability to automatically detect and group issues into specific clusters, e.g., SF6 low-pressure alarms and blocks with abnormal profiles caused by event time-delay problems.

Abstract

Abstract
Forecasts of residual demand curves (RDCs) are valuable information for price-maker market agents since it enables an assessment of their bidding strategy in the market-clearing price. This paper describes the application of deep learning techniques, namely long short-term memory (LSTM) network that combines past RDCs and exogenous variables (e.g., renewable energy forecasts). The main contribution is to build up on the idea of transforming the temporal sequence of RDCs into a sequence of images, avoiding any feature reduction and exploiting the capability of LSTM in handling image data. The proposed method was tested with data from the Iberian day-ahead electricity market and outperformed machine learning models with an improvement of above 35% in both root mean square error and Frechet distance.