A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Electric Vehicle Identification in Low-Sampling Non-Intrusive Load Monitoring Systems Using Machine Learning
The prediction is that by 2030, electric vehicles (EVs) will make up 40% of total passenger car sales in the USA. This surge in EV adoption poses a significant challenge to the electric grid due to the increased demand for charging. To proactively address this demand, power agencies must be able to identify EV charging activities. Previous studies have shown the potential of machine learning (ML) in High-Sampling Non-Intrusive Load Monitoring (NILM) systems. However, it remains uncertain whether these ML approaches are equally effective in Low-Sampling NILM systems, which are more representative of real-world conditions. Therefore, this paper investigates the operation of ML in Low-Sampling NILM systems. We propose a novel feature extraction that improves the accuracy by up to 25%. Additionally, we conduct experiments to gain insights into the relationship between system sampling rate (SR), training data size, and identification accuracy. By investigating these factors in realistic settings, this research takes a significant step towards the practical implementation of EV charging identification methods in NILM systems.
Electric Vehicle Identification in Low-Sampling Non-Intrusive Load Monitoring Systems Using Machine Learning
The prediction is that by 2030, electric vehicles (EVs) will make up 40% of total passenger car sales in the USA. This surge in EV adoption poses a significant challenge to the electric grid due to the increased demand for charging. To proactively address this demand, power agencies must be able to identify EV charging activities. Previous studies have shown the potential of machine learning (ML) in High-Sampling Non-Intrusive Load Monitoring (NILM) systems. However, it remains uncertain whether these ML approaches are equally effective in Low-Sampling NILM systems, which are more representative of real-world conditions. Therefore, this paper investigates the operation of ML in Low-Sampling NILM systems. We propose a novel feature extraction that improves the accuracy by up to 25%. Additionally, we conduct experiments to gain insights into the relationship between system sampling rate (SR), training data size, and identification accuracy. By investigating these factors in realistic settings, this research takes a significant step towards the practical implementation of EV charging identification methods in NILM systems.
Electric Vehicle Identification in Low-Sampling Non-Intrusive Load Monitoring Systems Using Machine Learning
Khaleghian, Seyedmehdi (author) / Tran, Toan (author) / Cho, Jin (author) / Harris, Austin (author) / Sartipi, Mina (author)
2023-09-24
800201 byte
Conference paper
Electronic Resource
English
Comparing four machine learning algorithms for household non-intrusive load monitoring
| DOAJ | 2024