Baseline methodologies for small scale residential demand response

Demand response (DR) programs are designed to reduce electricity load in periods of peak electricity demand, which helps avoiding expensive network upgrades. This reduction is measured by comparing the actual load with a baseline estimate. The baseline is a counter-factual load that could have been consumed in the absence of the DR program. Criteria for a good baseline methodology are accuracy, simplicity, and integrity. Five baseline methodologies are tested on actual smart meter data from 66 residential customers. These methodologies are High X of Y, Last Y Days, Regression, Neural Network, and Polynomial Interpolation. The performances of the selected methodologies are evaluated using bias and estimation error. The merits of each methodology are summarised in terms of simplicity, accuracy, and integrity. The results indicate that machine learning with neural network and polynomial extrapolation outperform the other methodologies.

[1]  I. Rowlands,et al.  A comparison of four methods to evaluate the effect of a utility residential air-conditioner load control program on peak electricity use , 2011 .

[2]  Koen Vanthournout,et al.  A norm behavior based deterministic methodology for demand response base lines , 2014, 2014 Power Systems Computation Conference.

[3]  Nicholas A. Engerer,et al.  Real-Time Simulations of 15,000+ Distributed PV Arrays At Sub Grid Level Using the Regional PV Simulation System (RPSS) , 2016 .

[4]  Ram Rajagopal,et al.  Data-Driven Targeting of Customers for Demand Response , 2016, IEEE Transactions on Smart Grid.

[5]  Omid Motlagh,et al.  Analysis of household electricity consumption behaviours: Impact of domestic electricity generation , 2015, Appl. Math. Comput..

[6]  Siobhán Clarke,et al.  A hybrid approach to very small scale electrical demand forecasting , 2014, ISGT 2014.

[7]  Janina M. Jolley,et al.  Research design explained: Instructor's edition, 7th ed. , 2010 .

[8]  Rui Xu,et al.  A Cluster-Based Method for Calculating Baselines for Residential Loads , 2016, IEEE Transactions on Smart Grid.

[9]  Karen Herter,et al.  Residential response to critical-peak pricing of electricity: California evidence , 2010 .

[10]  Karl Aberer,et al.  When Bias Matters: An Economic Assessment of Demand Response Baselines for Residential Customers , 2014, IEEE Transactions on Smart Grid.

[11]  Uwe Aickelin,et al.  An Approach for Assessing Clustering of Households by Electricity Usage , 2014, ArXiv.

[12]  Michael Y. Hu,et al.  Forecasting with artificial neural networks: The state of the art , 1997 .

[13]  Hamid Shaker,et al.  Short-term electricity load forecasting of buildings in microgrids , 2015 .

[14]  Siobhán Clarke,et al.  Residential electrical demand forecasting in very small scale: An evaluation of forecasting methods , 2013, 2013 2nd International Workshop on Software Engineering Challenges for the Smart Grid (SE4SG).

[15]  Yogesh L. Simmhan,et al.  Empirical Comparison of Prediction Methods for Electricity Consumption Forecasting , 2014 .

[16]  Ram Rajagopal,et al.  Short Term Electricity Load Forecasting on Varying Levels of Aggregation , 2014, ArXiv.

[17]  David G. Holmberg,et al.  Towards Demand Response Measurement and Verification Standards | NIST , 2013 .

[18]  Ram Rajagopal,et al.  Thermal profiling of residential energy use , 2015 .

[19]  Jaime Lloret,et al.  Artificial neural networks for short-term load forecasting in microgrids environment , 2014 .

[20]  Walid Saad,et al.  Game-Theoretic Methods for the Smart Grid: An Overview of Microgrid Systems, Demand-Side Management, and Smart Grid Communications , 2012, IEEE Signal Processing Magazine.

[21]  Clifford Grimm.  Evaluating Baselines for Demand Response Programs , 2008 .

[22]  Carmen Baskette Henrikson.  Designing a Successful Demand Response Program: It's Not Your Grandfather's Load Control Program , 2008 .

[23]  Hongseok Kim,et al.  Data-Driven Baseline Estimation of Residential Buildings for Demand Response , 2015 .