Ziyang Wang

Ziyang Wang

Dr. Ziyang Wang is a Project Researcher in EEIS at the University of Tokyo.

Publications

Novel Custom Loss Functions and Metrics for Reinforced Forecasting of High and Low Day-ahead Electricity Prices Using CNN-LSTM and Ensemble Learning

Published in Energies, 2024

This paper presents an innovative framework for day-ahead electricity price forecasting (DAEPF) using a Convolutional Neural Network–Long Short-Term Memory (CNN–LSTM) model to predict prices in Kyushu, Japan. It addresses the challenge of zero prices during high RES periods by incorporating a natural logarithm transformation and a novel “policy-versus-policy” strategy. The CNN–LSTM model outperforms a traditional LSTM in both accuracy and computation time. By integrating multimodal features, the model achieves high prediction accuracy, highlighting the benefits of a comprehensive approach to DAEPF.

Recommended citation: Wang Z, Mae M, Yamane T, Ajisaka M, Nakata T, Matsuhashi R, Novel Custom Loss Functions and Metrics for Reinforced Forecasting of High and Low Day-ahead Electricity Prices Using CNN-LSTM and Ensemble Learning. Energies 2024, 17, 2687. https://doi.org/10.3390/en17112687 http://ziyangwangacademics.github.io/files/Energies3.pdf

Enhanced Day-Ahead Electricity Price Forecasting Using a Convolutional Neural Network–Long Short-Term Memory Ensemble Learning Approach with Multimodal Data Integration

Published in Energies, 2024

This paper presents an innovative framework for day-ahead electricity price forecasting (DAEPF) using a Convolutional Neural Network–Long Short-Term Memory (CNN–LSTM) model to predict prices in Kyushu, Japan. It addresses the challenge of zero prices during high RES periods by incorporating a natural logarithm transformation and a novel “policy-versus-policy” strategy. The CNN–LSTM model outperforms a traditional LSTM in both accuracy and computation time. By integrating multimodal features, the model achieves high prediction accuracy, highlighting the benefits of a comprehensive approach to DAEPF.

Recommended citation: Wang Z, Mae M, Yamane T, Ajisaka M, Nakata T, Matsuhashi R, Enhanced Day-Ahead Electricity Price Forecasting Using a Convolutional Neural Network–Long Short-Term Memory Ensemble Learning Approach with Multimodal Data Integration. Energies 2024, 17, 2687. https://doi.org/10.3390/en17112687 http://ziyangwangacademics.github.io/files/Energies2.pdf

Vehicular fuel consumption and CO₂ emission estimation model integrating novel driving behavior data using machine learning

Published in Energies, 2024

This paper introduces a novel approach to estimating vehicular fuel consumption and CO₂ emissions by incorporating driving behavior data such as speeding, sudden accelerating, and braking, along with driving time and distances on various road types. Using random forest regression, the study demonstrates the relevance of these behaviors for monthly CO₂ emission estimation and proposes a generalizable model with high prediction accuracy. This research offers insights into CO₂ emission reduction and energy conservation in road transportation.

Recommended citation: Wang Z, Mae M, Nishimura S, Matsuhashi R. Vehicular fuel consumption and CO₂ emission estimation model integrating novel driving behavior data using machine learning. Energies 2024, 17, 1410. https://doi.org/10.3390/en17061410 http://ziyangwangacademics.github.io/files/Energies1.pdf

Estimating vehicular fuel consumption and CO₂ emissions by machine learning using only speed and acceleration

Published in Journal of Japan Society of Energy and Resources, 2023

This paper introduces a method for instantaneous vehicular fuel consumption assessment by analyzing driving data, specifically speed and acceleration, using random forest (RF) regression.

Recommended citation: Maroju R, Nishimura S, Wang Z, Matsuhashi R. Estimating Vehicular Fuel Consumption and CO₂ Emissions by Machine Learning Using Only Speed and Acceleration. Journal of Japan Society 2023. https://doi.org/10.24778/jjser.44.1_30 http://ziyangwangacademics.github.io/files/JSER1.pdf

Intrusive and non-intrusive early warning systems for thermal discomfort by analysis of body surface temperature

Published in Applied Energy, 2023

This paper delves further into the concept of Relative Thermal Sensation (RTS) and explores both its intrusive and non-intrusive assessment methods by analysis of body surface temperature using machine learning. Additionally, this paper proposes a novel composite thermal comfort model that can predict an individual’s current thermal comfort (hot, cold, or cozy) and the current thermal sensation transitioning trend (hotter, colder, or no change). Preliminary work on this topic can be found in our earlier paper titled “Proposal of relative thermal sensation: Another dimension of thermal comfort and its investigation.”

Recommended citation: Wang Z, Matsuhashi R, Onodera H. Intrusive and non-intrusive early warning systems for thermal discomfort by analysis of body surface temperature. Appl Energy 2023;329:120283. https://doi.org/10.1016/j.apenergy.2022.120283. http://ziyangwangacademics.github.io/files/AppliedEnergy1.pdf

Towards wearable thermal comfort assessment framework by analysis of heart rate variability

Published in Building and Environment, 2022

This paper introduces a real-time thermal comfort assessment method based on the analysis of heart rate variablity (HRV). The heartbeat data were measured using a portable, non-intrusive, and low-cost wearable heartbeat sensor.

Recommended citation: Wang Z, Matsuhashi R, Onodera H. Towards wearable thermal comfort assessment framework by analysis of heart rate variability. Build Environ 2022:109504. https://doi.org/10.1016/j.buildenv.2022.109504. http://ziyangwangacademics.github.io/files/BuildingandEnvironment1.pdf

Proposal of relative thermal sensation: Another dimension of thermal comfort and its investigation

Published in IEEE Access, 2021

This paper introduces the concept of Relative Thermal Sensation (RTS) and presents a preliminary investigation of its assessment method using an intrusive approach. Further discussions on this topic can be found in our subsequent paper titled “Intrusive and non-intrusive early warning systems for thermal discomfort by analysis of body surface temperature.”

Recommended citation: Wang Z, Onodera H, Matsuhashi R. Proposal of relative thermal sensation: Another dimension of thermal comfort and its investigation. IEEE Access 2021;9:36266–81. https://doi.org/10.1109/ACCESS.2021.3062393. http://ziyangwangacademics.github.io/files/IEEEAccess1.pdf

Research on Thermal Comfort by Analyzing LF/HF Value and Heat Flow Rate

Published in エネルギー・資源学会論文誌 2019, 2019

Preliminary work involved using the heart rate variability (HRV) index, specifically the LF/HF value, for real-time thermal comfort assessment.

Recommended citation: Wang Z, Matsuhashi R. Research on thermal comfort by analyzing LF/HF value and heat flow rate. エネルギー・資源学会論文誌 2019;40:154–9. https://doi.org/10.24778/jjser.40.5_154. http://ziyangwangacademics.github.io/files/JSER2019.pdf