Machine learning techniques for developing virtual sensors of water flow rates in building cooling systems

The scope of this project is the development of virtual sensors that use mathematical models along with measurements data from the Building Automation System, and can be installed in mechanical cooling systems pf large commercial and institutional buildings, instead of electromagnetic or ultrasonic water flow meters. The proposed virtual sensors would provide a low-cost, practical, and non-intrusive method to monitor the chilled and condenser water mass flow rates, and help for the fault detection and diagnosis of chillers, and the monitoring of chillers performance.
The partner, Mariner Partners Inc. DBA SHIFT Energy, from Saint John, New Brunswick, Canada, can use the thermodynamic mathematical models generated in this project to develop computer applications that can be integrated in the Building Automation Systems of HVAC systems. The computer applications developed by the partner can be used across Canada in large commercial and institutional buildings, for example for more efficient energy use, and detection and diagnostic of faults. The improved energy efficiency in chilled water plants represents a large opportunity for demand and GHG reductions – a priority for Canada going forward. Virtual flow metering will expand the addressable market for energy efficiency projects in central cooling plants.

Faculty Supervisor:

Radu Zmeureanu

Student:

Partner:

Mariner Partners Inc.

Discipline:

Engineering

Sector:

Professional, scientific and technical services

University:

Concordia University

Program: