Date of Award
12-2021
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Mechanical Engineering
Committee Chair/Advisor
Dr. John R Wagner
Committee Member
Dr. Cameron Turner
Committee Member
Dr. Gregory Mocko
Abstract
As the complexity of industrial equipment continues to increase, the management of the individual machines and integrated operations becomes difficult without computer tools. The availability of streaming data from manufacturing floors, plant operations, and deployed fleets can be overwhelming to analyze, although it provides opportunities to improve performance. The use of dedicated monitoring systems in the plant and field to troubleshoot machinery can be integrated within a product lifecycle management (PLM) architecture to offer greater features. PLM offers virtual processes and software tools for the design, analysis, monitoring, and support of engineering systems and products. Within this paradigm, a digital twin can estimate system behavior based on the assembled physical models and the operating data for preventive maintenance efforts. PLM software can store computer-aided-design, computer-aided-engineering, advanced manufacturing, and data in cloud form for remote access. Integrating physical and performance data into a single database provides flexibility and adaptability while allowing distant commanding and health monitoring of dynamic systems.
The recent attention on global warming, and the minimization of energy consumption can be partially addressed by examining those economic sectors that use large quantities of electric power. Across the United States, heating, ventilation, and air conditioning (HVAC) systems use a collective $14 Billion of resources to control the temperature of commercial and residential spaces. A typical commercial HVAC system consists of a chiller plant, water pumps for fluid circulation, multiple heat exchangers, and iii forced air blowers. In this research project, a digital twin is created for a single compressor chilled water-based HVAC system using a multi-disciplinary CAE software package. The system level models are assembled to describe a 1400 ton chiller located in the East-side chiller plant on the Clemson University (Clemson, SC) campus. The dynamic models that estimate the fluid pressures, temperatures, and flow rates, as well as the electrical and mechanical power consumption, are validated against the operating data streamed through the OptiCX System.
To demonstrate the capabilities of this digital twin tool in a preventive maintenance mode, various degradations are virtually investigated in the chiller plant's components. The mechanical pump efficiency, electric pump motor friction, pipe blockage, air flow rate sensor, and the expansion valve opening were degraded by 3% to 5%, which impacted component behavior and system performance. The analysis of these predicted plant signals helped to establish preventive maintenance thresholds on these components, which should promote improved plant reliability. A digital twin provides additional flexibility than stand-alone monitoring technologies due to the capability of simulating customized scenarios for analyzing failure-prone conditions and overall equipment effectiveness (OEE). The PLM-based digital twin offers a design and prognostic platform for HVAC systems.
Recommended Citation
Kale, Mihir, "Development and Application of a Digital Twin for Chiller Plant Performance Assessment" (2021). All Theses. 3683.
https://tigerprints.clemson.edu/all_theses/3683
Included in
Energy Systems Commons, Other Mechanical Engineering Commons, Systems Engineering Commons