A Supervised Learning-Based Min/Max Voltage Estimation Model for All Nodes in Low-Voltage Networks
- Resource Type
- Conference
- Authors
- Park, Woan-Ho; Hwang, Jin Sol; Hwang, Joonbyeok; Park, Yohan; Namkoong, Won; Kim, Yun-Su
- Source
- 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE) Innovative Smart Grid Technologies Europe (ISGT EUROPE), 2023 IEEE PES. :1-5 Oct, 2023
- Subject
- Components, Circuits, Devices and Systems
Computing and Processing
Power, Energy and Industry Applications
Low voltage
Renewable energy sources
Error analysis
Estimation
Europe
Real-time systems
Voltage control
Voltage estimation
Low-voltage networks
Supervised learning
Multi-layer perceptron
Feeder Remote Terminal Unit
Advanced Metering Infrastructure
- Language
Voltage regulation is performed primarily to control low-voltage networks, which can be challenging for large networks. This study presents the development of a voltage estimation technique for low-voltage networks based on supervised learning. The proposed technology utilizes a feeder remote terminal unit (FRTU) and advanced metering infrastructure (AMI) measurement data to estimate the voltage range of a low-voltage system by deploying smart metering in a distribution network A multi-layer perceptron technique is implemented to perform near real-time voltage estimation using measured data from part of FRTU and AMI. The voltage range estimation model is simulated based on radial networks of the Korea Electric Power Corporation network system. The test results indicate that the proposed model can accurately estimate the maximum and minimum voltage range in near real-time tests and in photovoltaic integration scenarios with low error rates across various cases.