Grant: Discovery Grants Program - Individual
École Polytechnique de Montréal
$31,000.00
Digital Twinning Approach for Autonomous Asset Performance Management
EF 2023-24
Résumé
Energy efficiency (EE) is a prime and major concern for Canadian energy-intensive industries. Improving EE is the most cost-effective way to secure energy supplies while reducing not only energy costs but also the environmental footprint. The large final emitters (e.g., pulp & paper mills and the oil & gas refineries) are the facilities that are contributing most to greenhouse gas (GHG) emissions. The main reason for the consumption of excess energy in these industries is the inefficient operation of their processes. This leads to a high frequency of abnormal events, non-accurate controller tuning, poor adaptability to system disturbances, and excessive maintenance. The success of artificial intelligence (AI) offers appropriate candidates to exploit the available data and build accurate models that help understand how various components in the system interact and influence each other. The proposed research aims at developing a digital twining approach made of hybrid causal models using predictive, descriptive, and prescriptive AI techniques combined with physical models and simulation-based optimization methods. The objective is to optimize the performance of complex assets in heavy-process industries and achieve autonomous operations. These hybrid models will be built from different industrial data sources and represent the building blocks of what is called "Active Digital Twins" that can provide plant operators with accurate and updated actionable insights. These digital twins will serve several purposes such as abnormal event management, supervisory control, and real-time optimization in complex equipment. The challenges and contributions of the proposed research fall under the following three development activities that involve the training of highly qualified persons. ?Development of new data and decision fusion method that preprocesses different types of data by utilizing an ensemble of different machine learning techniques. This method will support building accurate models based on interpretable and accurate information. ?Development of accurate and representative data-driven dynamic causal models that construct an efficient causality analysis tool by combining explainable AI and process mining methods with human expertise. This tool will automatically diagnose interdependence between several abnormal events and describe evolving mechanisms in monitored equipment. ?Development of a robust supervisory control method that uses event-data logs to better understand the behavior of local controllers and to extract the necessary knowledge to better supervise them. The method will also be used to avoid abnormal events and maintain the desired energy-efficient process operation. This research will contribute to reducing GHG emissions from the industrial sector. This reduction is an important step to help Canada achieve its commitment to move to Net-Zero emissions by 2050.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RGPIN-2023-05502
Name nom
Ragab, Ahmed
Department département
Department of Mathematics and industrial Engineering
Organization id
64
Institution établissement
École Polytechnique de Montréal
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2023
Award amount
31000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
306
Area of application group en
Energy resources
Area of application group fr
Ressources d'énergie
Area of application en
Energy efficiency
Area of application fr
Efficacité énergétique
Research subject code
1606
Research subject group en
Industrial engineering
Research subject group fr
Operations management
Research subject en
Génie industriel
Research subject fr
Gestion de la production
Application title
Digital Twinning Approach for Autonomous Asset Performance Management
Keyword
Artificial Intelligence; Causality Analysis; Condition Monitoring; Deep Learning; Digital Twins; Energy Efficiency; Fault Diagnosis and Prognosis; Greenhouse Gas Emission; Reinforcement Learning; Supervisory Control
Application summary
Energy efficiency (EE) is a prime and major concern for Canadian energy-intensive industries. Improving EE is the most cost-effective way to secure energy supplies while reducing not only energy costs but also the environmental footprint. The large final emitters (e.g., pulp & paper mills and the oil & gas refineries) are the facilities that are contributing most to greenhouse gas (GHG) emissions. The main reason for the consumption of excess energy in these industries is the inefficient operation of their processes. This leads to a high frequency of abnormal events, non-accurate controller tuning, poor adaptability to system disturbances, and excessive maintenance. The success of artificial intelligence (AI) offers appropriate candidates to exploit the available data and build accurate models that help understand how various components in the system interact and influence each other. The proposed research aims at developing a digital twining approach made of hybrid causal models using predictive, descriptive, and prescriptive AI techniques combined with physical models and simulation-based optimization methods. The objective is to optimize the performance of complex assets in heavy-process industries and achieve autonomous operations. These hybrid models will be built from different industrial data sources and represent the building blocks of what is called "Active Digital Twins" that can provide plant operators with accurate and updated actionable insights. These digital twins will serve several purposes such as abnormal event management, supervisory control, and real-time optimization in complex equipment. The challenges and contributions of the proposed research fall under the following three development activities that involve the training of highly qualified persons. ?Development of new data and decision fusion method that preprocesses different types of data by utilizing an ensemble of different machine learning techniques. This method will support building accurate models based on interpretable and accurate information. ?Development of accurate and representative data-driven dynamic causal models that construct an efficient causality analysis tool by combining explainable AI and process mining methods with human expertise. This tool will automatically diagnose interdependence between several abnormal events and describe evolving mechanisms in monitored equipment. ?Development of a robust supervisory control method that uses event-data logs to better understand the behavior of local controllers and to extract the necessary knowledge to better supervise them. The method will also be used to avoid abnormal events and maintain the desired energy-efficient process operation. This research will contribute to reducing GHG emissions from the industrial sector. This reduction is an important step to help Canada achieve its commitment to move to Net-Zero emissions by 2050.