Join Christopher Laughman. Senior Principal Research Staff, Mitsubishi Electric Research Laboratories(MERL), and Avinash Subramanian, Software Engineer - Simulation, Control and Optimization at JuliaHub for a webinar on Insights from Mitsubishi Electric: Building High-Fidelity HVAC Models.
Accurately estimating refrigerant mass in vapor-compression cycles (air conditioners, heat pumps) is crucial for optimal performance and environmental impact assessment. However, traditional methods are either invasive or lack precision.
In this webinar, Christopher Laughman and Avinash Subramanian will share insights from the Mitsubishi Electric Research Laboratories case study where they used a novel approach using ModelingToolkit and machine learning.
In particular, we will focus on the Mitsubishi Electric Research Laboratories’s use case of estimating refrigerant quantities in air conditioners and heat pumps—critical for both system performance and environmental impact assessments. Where traditional methods for determining refrigerant mass are invasive and disruptive, Mitsubishi Electric’s novel state estimation technique offers a non-invasive alternative that uses temperature and pressure measurements to predict unmeasurable quantities like refrigerant mass.
Highlights:
- Limitations of traditional diagnostic methods for HVAC systems.
- How JuliaSim empowers high-fidelity HVAC modeling.
- Leveraging machine learning for state estimation and accurate refrigerant mass determination.
- Real-world case study: Achieving <2% error in refrigerant mass estimation using JuliaSim.
- Potential of this approach for improved diagnostics and control strategies in HVAC systems.
