Forecasting the electricity consumption is one of the most challenging tasks for energy domain stakeholders. Having reliable electricity consumption forecasts can help minimizing the cost of electricity and also enable a better control on the electricity tariff. This paper presents a study regarding the forecast of electricity consumption using a methodology based on Hybrid neural Fuzzy Inference System (HyFIS). The proposed approach considers two distinct strategies, namely one strategy using only the electricity consumption as the input of the method, and the second strategy uses a combination of the electricity consumption and the environmental temperature as the input. A case study considering the forecasting of the consumption of an office building using the proposed methodologies is also presented. Results show that the second strategy is able to achieve better results, hence concluding that HyFIS is an appropriate approach to incorporate different sources of information. In this way, the environmental temperature can help the HyFIS method to achieve a more reliable forecast of the electricity consumption.

Reliable consumption forecasts are crucial in several aspects of power and energy systems, e.g. to take advantage of the full potential of flexibility from consumers and to support the management from operators. With this need, several methodologies for electricity forecasting have emerged. However, the study of correlated external variables, such as temperature or luminosity, is still far from adequate. This paper presents the application of the Wang and Mendel's Fuzzy Rule Learning Method (WM) to forecast electricity consumption. The proposed approach includes two distinct strategies, the first one uses only the electricity consumption as the input of the method, and the second strategy considers a combination of the electricity consumption and the environmental temperature as the input, in order to extract value from the correlation between the two variables. A case study that considers the forecast of the energy consumption of a real office building is also presented. Results show that the WM method using the combination of energy consumption data and environmental temperature is able to provide more reliable forecasts for the energy consumption than several other methods experimented before, namely based on artificial neural networks and support vector machines. Additionally, the WM approach that considers the combination of input values achieves better results than the strategy that considers only the consumption history, hence concluding that WM is appropriate to incorporate different information sources.

This paper proposes a Building Energy Management System (BEMS) as part of an organization-based Multi-Agent system that models the Smart Home Electricity System (MASHES). The proposed BEMS consists of an Energy Management System (EMS) and a Prediction Engine (PE). The considered Smart Home Electricity System (SHES) consists of different agents, each with different tasks in the system. In this context, smart homes are able to connect to the power grid to sell/buy electrical energy to/from the Local Electricity Market (LEM), and manage electrical energy inside of the smart home. Moreover, a Modified Stochastic Predicted Bands (MSPB) interval optimization method is used to model the uncertainty in the Building Energy Management (BEM) problem. A demand response program (DRP) based on time of use (TOU) rate is also used. The performance of the proposed BEMS is evaluated using a JADE implementation of the proposed organization-based MASHES.

Energy consumption forecasting is crucial in current and future power and energy systems. With the increasing penetration of renewable energy sources, with high associated uncertainty due to the dependence on natural conditions (such as wind speed or solar intensity), the need to balance the fluctuation of generation with the flexibility from the consumer side increases considerably. In this way, significant work has been done on the development of energy consumption forecasting methods, able to deal with different forecasting circumstances, e.g., the prediction time horizon, the available data, the frequency of data, or even the quality of data measurements. The main conclusion is that different methods are more suitable for different prediction circumstances, and no method can outperform all others in all situations (no-free-lunch theorem). This paper proposes a novel application, developed in the scope of the SIMOCE project (ANI|P2020 17690), which brings together several of the most relevant forecasting methods in this domain, namely artificial neural networks, support vector machines, and several methods based on fuzzy rule-based systems, with the objective of providing decision support for energy consumption forecasting, regardless of the prediction conditions. For this, the application also includes several data management strategies that enable training of the forecasting methods depending on the available data. Results show that by this application, users are endowed with the means to automatically refine and train different forecasting methods for energy consumption prediction. These methods show different performance levels depending on the prediction conditions, hence, using the proposed approach, users always have access to the most adequate methods in each situation.

This paper presents a Support Vector Machine (SVM) based approach for energy consumption forecasting. The proposed approach includes the combination of both the historic log of past consumption data and the history of contextual information. By combining variables that influence the electrical energy consumption, such as the temperature, luminosity, seasonality, with the log of consumption data, it is possible for the proposed method by find patterns and correlations between the different sources of data and therefore improves the forecasting performance. A case study based on real data from a pilot microgrid located at the GECAD campus in the Polytechnic of Porto is presented. Data from the pilot buildings are used, and the results are compared to those achieved by several states of the art forecasting approaches. Results show that the proposed method can reach lower forecasting errors than the other considered methods.