No abstract

Systems that rely on forecasts to make decisions, e.g. control or energy trading systems, require frequent updates of the forecasts. Usually, the forecasts are updated whenever new observations become available, hence in an online setting. We present the R package onlineforecast that provides a generalized setup of data and models for online forecasting. It has functionality for time-adaptive fitting of linear regression-based models. Furthermore, dynamical and non-linear effects can be easily included in the models. The setup is tailored to enable effective use of forecasts as model inputs, e.g. numerical weather forecast. Users can create new models for their particular system applications and run models in an operational online setting. The package also allows users to easily replace parts of the setup, e.g. use kernel or neural network methods for estimation. The package comes with comprehensive vignettes and examples of online forecasting applications in energy systems, but can easily be applied in all fields where online forecasting is used.

Conventional residential electricity consumers are becoming prosumers who not only consume electricity but also produce it. This shift is expected to occur over the next few decades at a large scale, and it presents numerous uncertainties and risks for the operation, planning, investment, and viable business models of the electricity grid. To prepare for this shift, researchers, utilities, policymakers, and emerging businesses require a comprehensive understanding of future prosumers’ electricity consumption. Unfortunately, there is a limited amount of data available due to privacy concerns and the slow adoption of new technologies such as battery electric vehicles and home automation. To address this issue, this paper introduces a synthetic dataset containing five types of residential prosumers’ imported and exported electricity data. The dataset was developed using real traditional consumers’ data from Denmark, PV generation data from the global solar energy estimator (GSEE) model, electric vehicle (EV) charging data generated using package, a residential energy storage system (ESS) operator and a generative adversarial network (GAN) based model to produce synthetic data. The quality of the dataset was assessed and validated through qualitative inspection and three methods: empirical statistics, metrics based on information theory, and evaluation metrics based on machine learning techniques.

Home energy management systems (HEMSs) are expected to become a crucial part of future smart grids. However, there is a limited number of studies that comprehensively assess the potential economic benefits of home energy management systems (HEMSs) for consumers under real market conditions and which take account of consumers' capabilities. In this study, a new optimization-based HEMS controller is presented to operate a photovoltaic and battery system. The HEMS controller considers the consumers' electrical load uncertainty by integrating multivariate probabilistic forecasting methods and a stochastic optimization in a rolling horizon. As a case study, a comprehensive simulation study is designed to emulate the operation of a real HEMS using real data from nine Danish homes over different seasons under real-time retail prices. The optimization-based control strategies are compared with a default (naive) control strategy that encourages self consumption. Simulation results show that seasonality in the consumers' load and electricity prices have a significant impact on the performance of the control strategies. A combination of optimization-based and naive control strategy presents the best overall results.