Energy forecasting is a technique to predict future energy needs to achieve demand and supply equilibrium. In this paper we aim to assess the performance of a forecasting model which is a weather-free model created using a database containing relevant information about past produced power data and data mining techniques. The idea of using a weather-free data-driven model is first to alleviate the dependence on weather data which, in some scenarios is difficult to obtain and second to reduce the computational effort. In this work, we aim first to evaluate the interplay between anomaly detection techniques and forecasting model accuracy. Secondly we will determine out of the three defined performance metrics, which one is the best for this particular application.

This paper addresses the estimation of household communities' overall energy usage and solar energy production, considering different prediction horizons. Forecasting the electricity demand and energy generation of communities can help enrich the information available to energy grid operators to better plan their short-term supply. Moreover, households will increasingly need to know more about their usage and generation patterns to make wiser decisions on their appliance usage and energy-trading programs. The main issues to address here are the volatility of load consumption induced by the consumption behaviour and variability in solar output influenced by solar cells specifications, several meteorological variables, and contextual factors such as time and calendar information. To address these issues, we propose a predicting approach that first considers the highly influential factors and, second, benefits from an ensemble learning method where one Gradient Boosted Regression Tree algorithm is combined with several Sequenceto-Sequence LSTM networks. We conducted experiments on a public dataset provided by the Ausgrid Australian electricity distributor collected over three years. The proposed model's prediction performance was compared to those by contributing learners and by conventional ensembles. The obtained results have demonstrated the potential of the proposed predictor to improve short-term multi-step forecasting by providing more stable forecasts and more accurate estimations under different day types and meteorological conditions.

No abstract

This paper presents a co-simulation model using MATLAB ® toolboxes to illustrate an interaction between the communication system and the energy grid, coherent with the concept of smart grid that employs IEC 61850 communication standard. The MMS (Manufacturing Message Specification) protocol supported by IEC 61850, based on TCP/IP, is used for the vertical communication between the Supervisory and Data Acquisition (SCADA) system and Intelligent Electronic Devices (IEDs) embedding the local control of different parts of the smart grid. In this paper an IED supporting the power control of a photovoltaic (PV) plant connected to a low-voltage (LV) utility grid is considered. Communication system consisting of the transport layer and a router placed on the network layer is modeled as an event-driven system using SimEvents ® toolbox and energy grid is modeled as a time-driven system using SimPowerSystems ® toolbox. Co-simulation results are obtained by combining different communication scenarios and time-varying irradiance scenarios for the PV plant when this one is required to provide a certain power in response to a power reference received from SCADA over the communication network. The analysis aims at illustrating the impact that stochastic behavior and delays due to network communication have on the global system behavior.