Application of Machine Learning Methods in Vancouver Island Supply Capability Determination

This project is a feasibility study of applying Machine Learning to Vancouver Island Load Supply Capability determination. The intern is expected to apply neural networks and other machine learning methods to train a transmission operation decision making model to determine the load supply capability of Vancouver Island. An operation scenario database covering equipment status and load information will be generated as the training data to the decision making model. The proposed project will improve the transmission operation of BC Hydro by avoiding over-control, such as unnecessary generator shedding and transmission line tripping. In addition, instead of the current practice, no equations are needed anymore in the operating orders, thus it will save huge amount of human efforts for developing these equations. Furthermore, the proposed project will make it easy for engineers and operators to do sensitivity check, better understanding of system operation will be gained.

Faculty Supervisor:

Jose Marti

Student:

Zemeng Wang

Partner:

BC Hydro

Discipline:

Engineering - computer / electrical

Sector:

Energy

University:

University of British Columbia