Voltage Estimation in Low-Voltage Distribution Grids with Distributed Energy Resources

TL;DR

This paper introduces a machine learning-based method using cable TV sensors to estimate voltage in low-voltage distribution grids, enhancing observability and supporting increased integration of distributed energy resources.

Contribution

It proposes a novel voltage estimation approach leveraging widely available CATV sensors and spatio-temporal data, improving accuracy with minimal sensor deployment.

Findings

High accuracy voltage estimates achieved in simulations

Effective even with very few observable nodes

Demonstrated on large 1572-bus feeder data set

Abstract

The present distribution grids generally have limited sensing capabilities and are therefore characterized by low observability. Improved observability is a prerequisite for increasing the hosting capacity of distributed energy resources such as solar photovoltaics (PV) in distribution grids. In this context, this paper presents learning-aided low-voltage estimation using untapped but readily available and widely distributed sensors from cable television (CATV) networks. The CATV sensors offer timely local voltage magnitude sensing with 5-minute resolution and can provide an order of magnitude more data on the state of a distribution system than currently deployed utility sensors. The proposed solution incorporates voltage readings from neighboring CATV sensors, taking into account spatio-temporal aspects of the observations, and estimates single-phase voltage magnitudes at all non-monitored buses using random forest. The effectiveness of the proposed approach was demonstrated using a 1572-bus feeder from the SMART-DS data set for two case studies - passive distribution feeder (without PV) and active distribution feeder (with PV). The analysis was conducted on simulated data, and the results show voltage estimates with a high degree of accuracy, even at extremely low percentages of observable nodes.