Project 37: Learning-Driven Power Maps

Contact Information:

Assoc. Prof. Ce Shang

Email: shangce@sjtu.edu.cn

Project Description and Objectives:

In this project, machine learning tools will be used to depict a nation-wide map of power and energy supply and demand. The power system that connects the power supply and demand, with the uncountable devices it is made of, has become the biggest and most complex artificial system of data of all time. The well-functioning of the power system, with a balanced power supply and demand, determines the well-being of all aspects in modern society. Mapping the power supply and demand benefits the system operation in the short term and planning in the long run. Extracting useful information drives the application of machine learning tools to power systems, especially when the power system is being developed towards the ubiquitous Internet of Things and the data capacity consequently explodes. This program studies the application of machine learning tools for drawing power maps, which is aimed to assist power system operation and planning.

Eligibility Requirements:

Power System Analysis; Linear Algebra, Probability and Statistics; Computer Programming (with C or Python language).

This program involves the interdisciplinary study of machine learning and power engineering. Background knowledge and interest in both fields is required as well as the willingness to do hands-on programming work. An in-term oral presentation and a final written report are required for a mark to be given.

Main Tasks:

Learning the background knowledge of the power system: its history, the cutting-edge research topics with a focus on power system operation and planning.

Reviewing machine learning tools: different algorithms, the basic math they require, and typical problems they can solve.

Implementing a machine learning tool, which can either be during supervised or unsupervised learning; targeting a specific group data of the power system, which can either be operational or planning data.

Tuning the learning tool with acquired power system data; mapping power system supply demand either in the short term of operation or in the long term of planning.

Website:

Lab: http://www.ssgc.sjtu.edu.cn/ 

School: www.seiee.sjtu.edu.cn/