Deep and Efficient Impact Models for Edge Characterization and Control of Energy Events

Network control in microgrids is an active research area driven by a steady increase in energy demand, the necessity to minimize the environmental footprint, yet achieve socioeconomic benefits and ensure sustainability. Reducing deviation of the predicted energy consumption from the actual one, softening peaks in demand and filling in the troughs, especially at times when power is more affordable and clean, present challenges for the demand-side response. In this paper, we present a hierarchical energy system architecture with embedded control. This architecture pushes prediction models to edge devices and executes local control loops to address the challenge of managing demand-side response locally. We employ a two-step approach: At an upper level of hierarchy, we adopt a conventional machine learning pipeline to build load prediction models using automated domain-specific feature extraction and selection. Given historical data, these models are then used to label prediction failure events that force the operator to use backup energy sources to stabilize the network. On a lower level of hierarchy, computed labels are used to train impact models realized by LSTM networks running on edge devices to infer the probability that the power consumption of the player contributes to the upper level prediction failure event. The system is evaluated on clustered and aggregated energy traces from a public data set of academic buildings. The results show the benefits of the proposed hierarchical energy system architecture in terms of impact prediction with 55% accuracy. This allows minimizing the number of prediction failure events by 11.69 % by executing targeted local control.

[1]  Francesco Borrelli,et al.  Model Predictive Control of thermal energy storage in building cooling systems , 2009, Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference.

[2]  Jui-Sheng Chou,et al.  Real-time detection of anomalous power consumption , 2014 .

[3]  Phuong H. Nguyen,et al.  Collaborative learning for classification and prediction of building energy flexibility , 2019, 2019 IEEE PES Innovative Smart Grid Technologies Europe (ISGT-Europe).

[4]  Clayton Miller,et al.  What's in the box?! Towards explainable machine learning applied to non-residential building smart meter classification , 2019, Energy and Buildings.

[5]  Sanjeevikumar Padmanaban,et al.  Study and Analysis of an Intelligent Microgrid Energy Management Solution with Distributed Energy Sources , 2017 .

[6]  Albert Y. Zomaya,et al.  From Insight to Impact: Building a Sustainable Edge Computing Platform for Smart Homes , 2018, 2018 IEEE 24th International Conference on Parallel and Distributed Systems (ICPADS).

[7]  Jui-Sheng Chou,et al.  Forecasting energy consumption time series using machine learning techniques based on usage patterns of residential householders , 2018, Energy.

[8]  Jian LUO,et al.  Real-time anomaly detection for very short-term load forecasting , 2018 .

[9]  Margarida Silveira,et al.  Unsupervised Anomaly Detection in Energy Time Series Data Using Variational Recurrent Autoencoders with Attention , 2018, 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA).

[10]  Alex Sherstinsky,et al.  Fundamentals of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) Network , 2018, Physica D: Nonlinear Phenomena.

[11]  Yanghui Wu,et al.  A XGBoost Model with Weather Similarity Analysis and Feature Engineering for Short-Term Wind Power Forecasting , 2019, Applied Sciences.

[12]  Xiandong Xu,et al.  Hierarchical microgrid energy management in an office building , 2017 .

[13]  A. Stephen McGough,et al.  Segmenting Residential Smart Meter Data for Short-Term Load Forecasting , 2018, e-Energy.

[14]  Pushpendra Singh,et al.  Rimor: towards identifying anomalous appliances in buildings , 2018, BuildSys@SenSys.

[15]  S. Beck,et al.  Using regression analysis to predict the future energy consumption of a supermarket in the UK , 2014 .

[16]  Florian Ziel,et al.  Forecasting Electricity Spot Prices Using Lasso: On Capturing the Autoregressive Intraday Structure , 2015, IEEE Transactions on Power Systems.

[17]  Antonio F. Gómez-Skarmeta,et al.  A methodology for energy multivariate time series forecasting in smart buildings based on feature selection , 2019, Energy and Buildings.

[18]  Hao Wang,et al.  A New Anomaly Detection System for School Electricity Consumption Data , 2017, Inf..

[19]  Yang Zhao,et al.  Deep learning-based feature engineering methods for improved building energy prediction , 2019, Applied Energy.

[20]  Andreas W. Kempa-Liehr,et al.  Time Series FeatuRe Extraction on basis of Scalable Hypothesis tests (tsfresh - A Python package) , 2018, Neurocomputing.

[21]  Nadeem Javaid,et al.  Electricity Price and Load Forecasting using Enhanced Convolutional Neural Network and Enhanced Support Vector Regression in Smart Grids , 2019, Electronics.

[22]  Iulia Stamatescu,et al.  Decision Support System for a Low Voltage Renewable Energy System , 2017 .

[23]  Yoshua Bengio,et al.  Deep Sparse Rectifier Neural Networks , 2011, AISTATS.

[24]  Grigore Stamatescu,et al.  Evaluation of Sequence-Learning Models for Large-Commercial-Building Load Forecasting , 2019, Inf..

[25]  Kevin M. Smith,et al.  Forecasting energy consumption of multi-family residential buildings using support vector regression: Investigating the impact of temporal and spatial monitoring granularity on performance accuracy , 2014 .

[26]  Aven Satre-Meloy,et al.  Investigating structural and occupant drivers of annual residential electricity consumption using regularization in regression models , 2019, Energy.

[27]  Clayton Miller,et al.  The Building Data Genome Project: An open, public data set from non-residential building electrical meters , 2017 .