Smart Buildings

The buildings of the 21st century are being design and built in a new context – one that clearly calls for development of new cutting-edge information and communication technologies providing decision support tools for efficient energy management in buildings. As a result, the design of smart buildings is not limited to installing various sensors: it is part of a broader context where distributed energy resources and the active participation of consumers are integrated into the effective management of demand throughout the building. Given advances in smart metre infrastructure, digital strategy has become vital for leveraging the data collected. Artificial intelligence (AI) seems essential for big data processing, increasing the performance of short-term energy demand forecasting models, gaining a better understanding of the volatility in each consumption profile and generating learning mechanisms adapted to demand management in several types of buildings.

Moreover, recent advances in the Internet of Things offer interesting opportunities for buildings to communicate, allowing the pooling of difference resources at the neighbourhood level as well as the possibility of energy trading to reduce peak demand. Along with the big data revolution brought about by widespread deployment of metres, sensors and smart technologies, there is a need to better explore the potential of predictive models based on deep learning to improve the accuracy and efficiency of forecasting in the energy context.

The team of Hanane Dagdougui, professor in the Department of Mathematics and Engineering and a member of GERAD, is particularly interested in the development of mathematical models and the application of machine learning techniques to energy management problems in buildings. Hanane Dagdougui is working on development of distributed algorithms and new approaches based on machine learning as well as implementation of applications derived from them in smart building networks. These management algorithms will make use of demand response strategies that will increase the flexibility of the building and the network. Hanane Dagdougui is currently developing several large-scale projects with major partners such as CanmetÉNERGIE, the Hydro-Québec Research Institute, Innovée, Hitachi ABB, Fusion Énergie, VadimUS. She works in collaboration with Charles Audet, Sébastien Le Digabel and Antoine Lesage-Landry, professors at Polytechnique Montréal and GERAD members. When demand side management of a significant number of buildings is precisely controlled by aggregators, this can play an increasing role in the wholesale electricity market. In this case, demand-side management can help the power system operator better manage peak demand while exploiting the potential for flexibility and enabling consumers to benefit from rewards or lower energy bills.

TIMES modelling

Canada is aiming for zero net greenhouse gas (GHG) emissions by 2050. But how can we achieve this ambitious carbon neutrality objective? Mainly be taking action on energy production and consumption, both of which can be assessed using so-called “techno-economic” energy models. These models detail the entire energy sector with its various forms of energy (oil, bioenergy, electricity, eat.) and associated technologies, in order to identify strategies that would avoid or sequester GHG emissions.

Over time, GERAD members have developed several versions of such models, following particularly the TIMES approach as developed by the International Energy Agency. TIMES is a large mathematical program – consisting of millions of variables and equations – which can be used to create a model which can then be used to identify the most economically efficient GHG reduction scenarios and the optimal timeframe for implementing them.

TIMES is a mathematical program that identifies the most efficient GHG reduction scenarios.

GERAD designed an initial TIMES model adapted to the Canadian context (TIMES Canada). ESMIA Consultants – a firm founded by Kathleen Vaillancourt, an entrepreneur trained at GERAD – then transformed TIMES Canada into a North American version called NATEM. ESMIA uses this model to advise companies and government; for example, in collaboration with the Trottier Energy Institute and Olivier Bahn, professor in the Department of Decision Science at HEC Montréal and Director of GERAD, it is using the model to develop a Canadian energy outlook. In this regard, the latest report, published in 2018, highlights the importance of electrification and bioenergy deployment in achieving Canada’s ambitious GHG reduction targets.

ESMIA and Olivier Bahn also use NATEM in the university setting, for example to assess the economic and ecological relevance of a new construction material developed at McGill University with a view to replacing cement.

Optimizing Hydroelectric Production and Maintenance

Hydroelectricity is a clean and renewable energy source that uses the power of water to produce energy. In Canada, more than 60% of electricity is generated by hydroelectricity, while in Quebec this figure rises to 97%. A hydroelectric system comprises power plants and reservoirs. Optimization models are used to efficiently manage these complex systems due to the spatiotemporal dependencies between facilities and uncertain water supplies at the planning stage. Indeed, these systems are subject to the vagaries of weather and hydrological forecasting: they also rely on turbines with different efficiencies.

The field of hydroelectric optimization is interested in development of mathematical models to solve these systems. Sara Séguin, professor in the Department of Computer Science and Mathematics at the Université du Québec à Chicoutimi and a GERAD member, works particularly on short-term models which aim to make decisions on short-term horizons (in hours or days) over periods of one to a few weeks. Sara Séguin has been working with Rio Tinto since 2013 to refine and develop new models. Projects are also underway with Dominique Orban and Miguel F. Anjos, respectively professors at Polytechnique Montréal and the University of Edinburgh. The research aims at refining models in order to produce more energy using the same facilities. Innovative modelling of production functions has made it possible to model and solve the problem over a very short computation time. It is particularly important to model production functions, considering they are essentially non-linear and non-convex. Solving these models makes it possible to avoid floods by keeping the reservoir levels at safe values, and allows for boating on lakes over the summer months by keeping the level high enough. In addition, producing more energy with the same amount of water ends up reducing costs for consumers.

Real-time decision-making in industrial mining complexes

An industrial mining complex or mineral value chain is an engineering system managing the extraction of materials from a group of mines and the treatment of the extracted materials through different interconnected processing facilities and waste/tailings locations, all leading to sellable products delivered to various customers and/or the spot market. The inherent uncertainty in the supply of materials from the mineral deposits mined, the market demand, as well as the various related operational aspects is the core element critically affecting decision-making. These sources of uncertainty are the central facets of this integrated business and the related planning, production scheduling, performance forecasting and product delivery.

Research by GERAD members Roussos Dimitrakopoulos and Michel Gamache contributes new advances towards an all-inclusive supply-meets-demand simultaneous stochastic optimization framework. This framework includes the development of new stochastic, intelligent and self-learning approaches, enhancing the capability of a mining complex optimization to respond to new unveiling information in operating environments in real-time. It also supports related operational decision-making in addition to adapting both strategic and operational policies, including environmental aspects. The so-termed “self-learning mining complex” is viewed as a future research outcome incorporating in real-time new unveiling production data, including sensor data, and market information, to dynamically adapt operational, production and scheduling decisions. At present, this is approached via self-play reinforcement learning including a multiple agent setting that integrates uncertainties and manages related risk. New developments are tested and evaluated in actual case studies.

Examples of related publications:

• Kumar, A., Dimitrakopoulos, R., Maulen, M., Adaptive self-learning mechanisms for updating short-term production decisions in an industrial mining complex. Journal of Intelligent Manufacturing, 31(7), 1795–1811 2020.
• Rimélé, A., Dimitrakopoulos, R., Gamache, M., A dynamic stochastic programming approach for open-pit mine planning with geological and commodity price uncertainty, Resources Policy, 65, 101570, 2020.
• de Carvalho, J.P., Dimitrakopoulos, R., Integrating production planning with truck-dispatching decisions through reinforcement learning while managing uncertainty, Minerals, 11(6), 587, 2021.