Hybrid short-term forecasting of the electric demand of supply fans using machine learning

Runge, Jason; Zmeureanu, Radu; Cam, Mathieu Le

doi:10.1016/j.jobe.2019.101144

Cited by 23 publications

(10 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This category refers to publications, which have applied DL-based models to submeter and/or components within the building. Typically, models applied at these levels find it more challenging to achieve accurate forecasts due to: (i) a lack of available data, (ii) a larger volatility and uncertainty in their profiles [81], and (iii) component and/or systems may be more sensitive to changes in operation (e.g., HVAC [93] and [94]). Applications of DL-based techniques include: (i) the electric demand of a ground source heat pump [37];…”

Section: Summary Of Applications At the Sub-meter And Component Levelmentioning

confidence: 99%

A Review of Deep Learning Techniques for Forecasting Energy Use in Buildings

Runge

Zmeureanu

2021

Energies

Self Cite

View full text Add to dashboard Cite

Buildings account for a significant portion of our overall energy usage and associated greenhouse gas emissions. With the increasing concerns regarding climate change, there are growing needs for energy reduction and increasing our energy efficiency. Forecasting energy use plays a fundamental role in building energy planning, management and optimization. The most common approaches for building energy forecasting include physics and data-driven models. Among the data-driven models, deep learning techniques have begun to emerge in recent years due to their: improved abilities in handling large amounts of data, feature extraction characteristics, and improved abilities in modelling nonlinear phenomena. This paper provides an extensive review of deep learning-based techniques applied to forecasting the energy use in buildings to explore its effectiveness and application potential. First, we present a summary of published literature reviews followed by an overview of deep learning-based definitions and techniques. Next, we present a breakdown of current trends identified in published research along with a discussion of how deep learning-based models have been applied for feature extraction and forecasting. Finally, the review concludes with current challenges faced and some potential future research directions.

show abstract

Section: Summary Of Applications At the Sub-meter And Component Levelmentioning

confidence: 99%

A Review of Deep Learning Techniques for Forecasting Energy Use in Buildings

Runge

Zmeureanu

2021

Energies

Self Cite

View full text Add to dashboard Cite

show abstract

“…It uses the resistor-capacitor and black-box models based on generalized LR, SVM, ANN, RF, and GB. Reference [39] combines a grey-box model and ANN to predict the electric demand of fans. Reference [40] joins the physical model of photovoltaics and ANN into a grey-box model.…”

Section: Introductionmentioning

confidence: 99%

Classification of retrofit measures for residential buildings according to the global cost

2021

View full text Add to dashboard Cite

Data-driven black-box surrogate models are widely used in research related to buildings energy efficiency. They are based on machine learning techniques, learn from available data, and act as a replacement for or an addition to complex and computationally intensive knowledge-based models. Surrogate models can predict energy demand, indoor air temperature, or occupants behavior; explore search space in optimization problems; learn control rules; etc. This paper analyzes surrogate models that classify building retrofit measures directly according to the global cost. In addition, they quantify the importance of each variable for the classification process. The models are based on random forest classifiers, which are fast and powerful ensemble learners. They can be applied to effectively reduce search spaces when optimizing energy renovation measures or to rapidly identify projects that deserve financial support. This approach is applied to two residential buildings and three scenarios of price development. The training process uses a small share of retrofit options assessed with standard calculations of the heating and cooling demands, as well as the global cost. The results show very high classification performance, even when the models are trained with small and imbalanced training sets. The obtained recall, precision, and F-score values are mostly above 95%, except for extremely small training sets.

show abstract

“…In the same vein, in a previous work of the same authors (Shiel & West, 2015), the research is focused on analyse their results by providing a huge table with the outputs obtained by the whole process in which the task of drawing any conclusion is not easy at first sight because of the high information depicted. An interesting recent study is carried out in (Runge, Zmeureanu & Le Cam, 2020) that proposes an hybrid short-term predictive solution to model electric demand, unlike the rest of the works mentioned so far, in this work the authors employ a carpet plot that can be very illustrative and useful to extract information, however, in this case, they also choose a non-dynamic solution. Probabilistic Algorithms and Model predictive control we used in (Gómez-Romero, et al, 2019) reporting energy savings around 35% in the intermediate winter season in an office building in Helsinki.…”

Section: Introductionmentioning

confidence: 99%

A case study on understanding energy consumption through prediction and visualization (VIMOEN)

Ruiz

Pegalajar

Molina-Solana

et al. 2020

Journal of Building Engineering

View full text Add to dashboard Cite

Energy efficiency has emerged as an overarching concern due to the high pollution and cost associated with operating heating, ventilation and air-conditioning systems in buildings, which are an essential part of our day to day life. Besides, energy monitoring becomes one of the most important research topics nowadays as it enables us the possibility of understanding the consumption of the facilities. This, along with energy forecasting, represents a very decisive task for energy efficiency. The goal of this study is divided into two parts. First to provide a methodology to predict energy usage every hour. To do so, several Machine Learning technologies were analysed: Trees, Support Vector Machines and Neural Networks. Besides, as the University of Granada lacks a tool to properly monitoring those data, a second aim is to propose an intelligent system to visualize and to use those models in order to predict energy consumption in real-time. To this end, we designed VIMOEN (VIsual MOnitoring of ENergy), a web-based application to provide not only visual information about the energy consumption of a set of geographically-distributed buildings but also expected expenditures in the near future. The system has been designed to be easy-to-use and intuitive for non-expert users. Our system was validated on data coming from buildings of the UGR and the experiments show that the Elman Neural Networks proved to be the most accurate and stable model and since the 5 th hour the results maintain accuracy.

show abstract

Hybrid short-term forecasting of the electric demand of supply fans using machine learning

Cited by 23 publications

References 43 publications

A Review of Deep Learning Techniques for Forecasting Energy Use in Buildings

A Review of Deep Learning Techniques for Forecasting Energy Use in Buildings

Classification of retrofit measures for residential buildings according to the global cost

A case study on understanding energy consumption through prediction and visualization (VIMOEN)

Contact Info

Product

Resources

About