Research Projects

• Project Physically-Constrained Price-Quantity Bidding for Thermal Price-Takers in Multi-Settlement Electricity Markets in IMSE7011 Data-driven Optimization Taught by Prof. Jiayang Li@HKU

  This study investigates how conventional thermal power plants, operating as price-takers in a hybrid market framework comprising a uniform pricing-based day-ahead market and a pay-as-bid real-time balancing market, can maximize their expected revenue through coordinated optimization of bidding strategies across both temporal domains while adhering to operational physical constraints such as ramp rate limitations and minimum stable generation requirements.

  Code is available in GitHub：Price-Taker Offering Strategy in Electricity Pay-as-Bid Markets

  

• Project Energy Hub Planning in ELEC7011 Energy Internet Taught by Prof. Yi Wang@HKU

  In this project, we aim to design a multi-energy system with a set of given devices to meet the electricity, heat and cooling demands of a building at the possible lowest total cost.

  Code is available in GitHub：Energy Hub Planning

  

• Research 《Online Transfer Learning for Load Forecasting under the Lack of Historical Data》

  With the liberalization of the electricity market, it has become increasingly common for residential customers to switch retail electricity providers in pursuit of more favorable tariffs. Due to privacy‐protection regulations and other constraints, these retailers are often faced with a lack of historical load data for newly contracted customers. Considering such cases, an online transfer learning-based load forecasting method is proposed in this paper.

• Research Cooperated with Prof. Kangping Li@SJTU 《Meteorological-implicit Graph Learning-based Distributed PV Data Cleaning Framework》 (Submitted)

  Distributed PV installations suffer serious data degradation and lack of site‑specific meteorological references which is usually necessary for PV data cleaning. We propose MiGL, a meteorological‑implicit graph‑learning framework that jointly captures spatial and temporal dependencies via a data-driven learned adjacency matrix—eliminating explicit priors—and validate its superior cleaning performance on real data from 69 PV sites.

• 全国大学生智能汽车竞赛讯飞组

  项目内容：在搭载激光雷达等传感器的无人驾驶车上实现机器人定位、自动驾驶、自动避障、目标检测、智能语音交互等功能，并完成赛项规定任务

  负责工作：完成整体任务思路构建，协调团队内部工作。实现机器人SLAM环境构建算法、自动驾驶与导航算法、机器人自主定位算法、目标检测算法与车辆运动控制算法的构建与优化

  项目成果：国家级竞赛一等奖两项（其中一项全国第四名）同时于Github开源本队伍参赛代码框架Code is available in GitHub：CUMTB-2022-Opensource

  目前收获27个star，6个fork