2025.6 – 2025.11
LaST-BO: Charging Load Forecasting
Charging-station load forecasting under realistic scenarios, paired with publication acceptance and software copyright.
ForecastingOptimizationWeb PlatformPower Systems
A LaST-BO charging-load forecasting study for complex real-world EV-charging scenarios. The work combined decoupled representation learning with Bayesian optimization, led to an acceptance from Zhejiang Electric Power in 2025, and produced a registered software copyright.
01
Background
To meet the demand for EV charging-station load forecasting under complex realistic scenarios, this project built a LaST-BO model that combines decoupled representation learning and Bayesian optimization.
02
Core Work
- Developed the forecasting-results display platform, including frontend design and implementation
- Integrated data interaction, charging-station geographic cloud maps, and an AI assistant module
03
Methods
- Participated in parameter design and experimental validation
- Built the modeling analysis on real charging-station data from Wuhan
04
Outcomes
- Paper accepted by Zhejiang Electric Power in 2025: "LaST-BO Charging Load Forecasting Model Improved by Bayesian Optimization"
- Co-second author
- Registered software copyright: "Power Load Forecasting System Based on the LaST-BO Model V1.0"
05
Tech Stack
JavaScriptPythonBayesian Optimization
06
Figures
