Artificial Intelligence to Reduce GHG Emissions in Energy Production and Transport Applications

点击翻译按钮获取中文摘要

One of the most difficult challenges facing our society is to reduce GHG emissions in an attempt to mitigate climate changes and their effect on the planet. Most of GHG emissions in Canada are from the use of fossil fuels for transportation, heating and electricity production in remote areas with diesel generators. In this research program, we specifically address these issues and apply artificial intelligence techniques to increase the impact of energy efficiency solutions developed in our research group for electricity production, road, rail and maritime transport. For the last 10 years, our research team has contributed to the development of new technologies for renewable energy and energy efficiency applications. The most important contributions were the use of Compressed Air Energy Storage (CAES), Phase Change Materials Heat Storage (PCMHS) and Pneumatic Hybridization of Diesel Engines (PHDE). We thoroughly studied the application of these technologies for the optimization of hybrid wind-diesel systems (electricity production in remote areas) and transportation (road, rail and maritime). CAES, PCMHS and PHDE significantly improve the renewable energy penetration in hybrid Wind-Diesel systems with Compressed Air Storage (WDCAS). In a typical WDCAS application we have higher wind power penetration and the surplus of wind power during strong winds is used to compress and store air. During compression, the heat is recovered and stored in a PCMHS for future use. When the wind energy is insufficient to supply the charge, the stored compressed air is used to overcharge the diesel such as to operate at an optimal air-fuel ratio at every regime. Before entering the engine, the compressed air is heated using the PCMHS. The overall renewable energy percentage in the total consumption increases between 30% and 60% compared with a wind-diesel system without storage. The application in transportation consists mainly in energy recuperation during breaking, in form of CAES and PCMHS, and restitution of this energy through overcharge of the diesel engine such as to minimize fuel consumption. The theoretical analysis based on thermodynamic models shows up to 60% fuel reduction for an urban driving cycle (ARTEMIS). While we completed the thermodynamic and heat transfer analysis, established the theoretical models for the different solutions, made a preliminary performance assessment, there are significant challenges before these solutions can be applied to industry. Artificial Intelligence (AI) techniques will be used to choose the most significant operating parameters involving CAES, PCMHS and PHDE according to the specific application and build models that can optimize real time operation. The AI models, that accurately represents the diversity and complexity of the systems and phenomena involved in these solutions, should operate sufficiently fast to optimize in real time the control parameters over a wide span of operating conditions and constraints.