Non-intrusive load monitoring (NILM) is the main method used to monitor energy footprint of a residential building and disaggregate total electrical usage into appliance-related signals. The most common disaggregation algorithms are based on Hidden Markov Model, while solutions deep neural networks have recently caught attention researchers. In this work we address problem through recognition state activation appliances using fully convolutional network, borrowing some techniques in semantic...

Indoor heating and cooling systems largely influence the power demand of residential buildings can play a significant role in Demand Side Management for energy communities. We propose novel method probabilistic forecasting total load community its base thermal components, combining conformalized quantile regression causal machine learning techniques, using only aggregate consumption environmental conditions data. applied proposed methods to dataset Germany, which includes separate...

The inclusion of photo-voltaic generation in the distribution grid poses technical difficulties related to variability solar source and determines need for Probabilistic Forecasting procedures (PF). This work describes a new approach PF based on quantile regression using Gradient-Boosted Regression Trees (GBRT) method fed by numerical weather forecasts European Centre Medium Range Weather Forecast (ECMWF) Integrated System (IFS) Ensemble Prediction (EPS). proposed methodology is compared...

In this article, a nowcasting technique for meteorological radar images based on generative neural network is presented. This technique’s performance compared with state-of-the-art optical flow procedures. Both methods have been validated using public domain data set of images, covering an area about 104 km2 over Japan, and period five years sampling frequency minutes. The the network, trained three data, forecasts time horizon up to one hour, evaluated year proved be significantly better...

Estimating household energy use patterns and user consumption habits is a fundamental requirement for management control techniques of demand response programs, leading to growing interest in non-intrusive load disaggregation methods. In this work we propose new methodology disaggregating the electrical from low-frequency measurements obtained smart meter contextual environmental information. The method proposed allows, with an unsupervised approach, separate loads into two components...

Rooftop photovoltaic solar systems can be an essential tool to support the energy transition of Europe. The assessment power generation potential in urban areas, necessary for smart grid planning, requires processing data different types, such as building cadastral information, a detailed description available roof and irradiation data. We introduce algorithm fast calculation building’s shadows procedure integration time. therefore develop methodology that allows evaluation with minimal...

The ability to predict consumption is an essential tool for the management of a power distribution network. availability advanced metering infrastructure through smart meters makes it possible produce forecasts down level individual user and introduce intelligence control at every grid. While aggregate load forecasting mature technology, single more difficult problem address due multiple factors affecting consumption, which are not always easily predictable. This work presents hybrid machine...

The balance between production and consumption in a smart grid with high penetration of renewable sources the presence energy storage systems benefits from an accurate load prediction. A general approach to forecasting is not possible because additional complication due increasing distributed usually unmeasured photovoltaic production. Various methods are proposed literature that can be classified into two classes: those predict by separating portion habits part local weather conditions,...

We conduct a comparative study of deterministic-to-probabilistic (D2P) and probabilistic-to-probabilistic (P2P) forecasting methods for photovoltaic (PV) power generation. In this analysis, we go beyond traditional statistical metrics to introduce novel metric in the field PV forecasting. This evaluates economic value production across all possible cost–loss ratios, offering comprehensive view forecast's utility at different probability thresholds. The study, based on real-world data from...

The implementation of the energy transition and building communities are driving forward exploitation potential for rooftop photovoltaic power generation. Estimating PV generation requires processing different types data, such as cadastral information buildings, a detailed description available areas, solar irradiance data. High-resolution estimation based on GIS data is normally limited to small survey areas. Instead, by using an algorithm efficient calculation shadows over rooftops,...

An accurate forecast of the power generated by a wind turbine is paramount importance for its optimal exploitation. Several forecasting methods have been proposed either based on physical modeling or using statistical approach. All them rely availability high quality measures local speed, corresponding and numerical weather forecasts. In this paper, simple effective technique, probability distribution mapping speed observed data, presented it applied to two turbines located island Borkum...

Abstract A detailed knowledge of the energy consumption and activation status electrical appliances in a house is beneficial for both user supplier, improving awareness allowing implementation management policies through demand response techniques. Monitoring individual certainly expensive difficult to implement technically on large scale, so non-intrusive monitoring techniques have been developed that allow be derived from sole measurement aggregate house. However, these methodologies often...

The High Concentrator Photovoltaic (HCPV) technology, due to its high efficiency, is considered one of the most promising solutions for exploitation sun-irradiation-based Renewable Energy Sources (RES). Nevertheless, HCPV production strictly connected Direct Normal Irradiation (DNI) making this photovoltaic technology more sensible cloudiness than traditional ones. In order mitigate power intermittence and improve programmability, integration between Storage Systems (ESSs) HCPV, resorting...

The application of load disaggregation techniques based on neural networks is often limited to users included in the training dataset. A methodology typical semantic segmentation images has been proposed for this task, which allows obtain a high accuracy and good generalization. We introduce here novel data augmentation technique improving forecasts unmonitored houses that does not require any interaction with user, nor further measurements consumption household appliances.

The Spectral Element Method (SEM), widely used in engineering sciences but almost unknown Oil&Gas exploration, is a numerical technique for large scale and multi-physics simulations. SEM an h-p approximation of the problem solution that allows quasi-optimal distribution computational points on unstructured grids so as to reduce both memory usage processing time. Thus, when accuracy mandatory complex 3D geometries, becomes excellent alternative Finite Elements (FE) Differences (FD). Its...