Artificial Intelligence-based Optimization System for Thermal Power Generating Units
Research project at JD Intelligent Cities Research
2018-2020: Funded by JD Intelligent Cities Research.
Research project manager for developing a deep reinforcement learning model and system to optimize power generation efficiency for the thermal power generating unit (TPGU) in power plants. This research has already finished software productization in JD Technology and been deployed in 6 power plants and 9 TPGUs in China!
Papers Published
- AAAI 2022: DeepThermal: Combustion Optimization for Thermal Power Generating Units Using Offline Reinforcement Learning
- AAAI 2022: Constraints Penalized Q-Learning for Safe Offline Reinforcement Learning
- IJCAI 2022 : Model-Based Offline Planning with Trajectory Pruning
Patents Granted
- CN112130538B: 磨煤机的控制优化及模型训练的方法、装置、设备和介质。2022.2.1。
- CN111389575B: 磨煤机组合运行控制方法、装置、设备及存储介质。2021.11.2。
- CN110533192B: 强化学习方法、装置、计算机可读介质及电子设备。2021.5.04。
- CN110717600B: 样本池构建方法和装置、以及算法训练方法和装置。2021.1.26。
- CN110673485B: 用于燃烧控制的模型训练方法、装置、电子设备以及介质。2020.11.24。
- CN110673478B: 磨煤机的控制方法、装置、系统及存储介质。2020.09.29。
- CN110490132B: 数据处理方法和装置。2020.09.29。
Media Coverages
- South China Morning Post: “JD builds AI control system that can save China’s thermal power plants billions and reduce pollution”.
- 人民日报：”“用先进科技为产业赋能 - “Alpha电”让电厂变“聪明””.
- 经济日报：”“京东数字科技CEO陈生强：数字科技能解决火电行业三大痛点”.
- 巴伦周刊：”“京东的技术“再造””.
- 中国能源报：”“AI+大数据”助火电行业降本增效”.
- 新浪科技：”京东对发电站下手了 一年可为中国节省70亿”.
- 极客网：”京东数科用AI优化火力发电 比Alpha Go围棋问题还复杂”.