Location
Grand Wailea, Hawaii
Event Website
https://hicss.hawaii.edu/
Start Date
7-1-2020 12:00 AM
End Date
10-1-2020 12:00 AM
Description
An accurate representation of the voltage-dependent, time-varying energy consumption of end-use electric loads is essential for the operation of modern distribution automation (DA) schemes. Volt-var optimization (VVO), a DA scheme which can decrease energy consumption and peak demand, often leverages electric network models and power flow results to inform control decisions, making it sensitive to errors in load models. End-use load modeling can be improved with additional measurements from advanced metering infrastructure (AMI). This paper presents two novel machine learning algorithms for creating data-driven, time-varying load models for use with DA technologies such as VVO. The first algorithm uses AMI data, k-means clustering, and least-squares optimization to create predictive load models for individual electric customers. The second algorithm uses deep learning (via a convolution-based recurrent neural network) to incorporate additional data and increase model accuracy. The improved accuracy of the load models for both algorithms is validated through simulation.
Improving End-Use Load Modeling Using Machine Learning and Smart Meter Data
Grand Wailea, Hawaii
An accurate representation of the voltage-dependent, time-varying energy consumption of end-use electric loads is essential for the operation of modern distribution automation (DA) schemes. Volt-var optimization (VVO), a DA scheme which can decrease energy consumption and peak demand, often leverages electric network models and power flow results to inform control decisions, making it sensitive to errors in load models. End-use load modeling can be improved with additional measurements from advanced metering infrastructure (AMI). This paper presents two novel machine learning algorithms for creating data-driven, time-varying load models for use with DA technologies such as VVO. The first algorithm uses AMI data, k-means clustering, and least-squares optimization to create predictive load models for individual electric customers. The second algorithm uses deep learning (via a convolution-based recurrent neural network) to incorporate additional data and increase model accuracy. The improved accuracy of the load models for both algorithms is validated through simulation.
https://aisel.aisnet.org/hicss-53/es/monitoring/6