Performance prediction and fault diagnosis in photovoltaic systems for optimal energy management

In this project, we will develop software-based models to monitor and to predict the performance of Perovskites-based
PVBlindsTM for optimal energy management and optimal integration into buildings, as well as to diagnose faults of PV cells for
safe, efficient, and reliable operation. The PVBlindsTM are developed by Solaires Inc. and will be deployed at various locations
within the Greater Vancouver area, and in various types of buildings. To develop the software-based models, machine learning
approaches will be studied and implemented. In addition, we will identify the best sets of model parameters to optimize the
accuracy of prediction and fault diagnosis, and to ensure a balance between electricity production and building energy
consumption.

Faculty Supervisor:

Ahmad Al-Dabbagh

Student:

Marzieh Kooshbaghi

Partner:

Solar-Ventures

Discipline:

Engineering

Sector:

Energy

University:

University of British Columbia Okanagan