Grant: Discovery Grants Program - Individual

Smart grid and renewable energy are promising solutions for the next generation of energy systems. This is motivated by environmental changes. Load demand and renewables forecasting, and fault detection are playing key roles in this process to improve the grid energy efficiency by implementing generation management. The accuracy of the forecasting models is very important to deal with new energy generation technology and consumption. Tidal currents, wind, solar and wave energies are the most challenging resources that require new paradigms in forecasting. The coordination between the distributed generation and the conventional centralized generating systems is considered as a very important element to reduce the overall energy cost and fossil fuel emissions. Decentralized and centralized control for the overall system is needed to enhance the reliability and fulfill smart grid requirements. Due to the fluctuation of the distributed generated power, conventional filters will not be sufficient for the desired power quality. As a result, there is a need for adaptive filters that depend on accurate forecasting models to improve the overall system quality. Due to the growth in residential energy demand, accurate load management and some management policies are required for optimal operation of the system. This research program proposes three intertwined and complementary themes to optimize renewables integration performance. The first theme deals with the development of advanced hybrid tools for renewables forecasting and fault classification and detection using novel blends of feature selection, clustering, and deep learning models for coupled data. The second theme deals with the development of decentralized and centralized control and filtering systems to improve renewables integration. The third theme aims to develop a predictive management framework for renewables using novel models, innovative algorithms, and decision support tools to assist renewables integration due to its fluctuation, nonlinearity and intermittency and minimize CO2 emissions and the overall cost. The development and testing of our models and machine learning algorithms will be based on large datasets from our industrial partners. In summary, cutting edge optimization models and techniques for renewables integration and management will be developed based on multiagent technology. A real time integration of micro-grid along with the main grid (simulated system) will be tested. An optimal filtering system will be used for improving the system quality based on an accurate forecasting model along with the real time data. Overall centralized and decentralized system control is monitoring and checking the multiagent system response. Study faulted cases situations will be tested to check the system stability under unexpected conditions. This proposal is associated and aligned with the Provincial and Federal strategies and policies towards green energy to reduce CO2 emissions.