JuliaSim HVAC: Thermofluid Design & Control | JuliaHub

JuliaHub is thrilled to announce the launch of Dyad (Formerly JuliaSim) HVAC, a suite of components for the modeling and simulation of complex thermofluid systems, revolutionizing the design, optimization and control of heating, ventilation, and air-conditioning (HVAC) systems.

HVAC systems are indispensable to maintaining comfort, safety, and efficiency in buildings, automobiles, airplanes, industrial facilities, and various applications. However, designing and optimizing these complex systems presents numerous challenges requiring interdisciplinary collaboration, including numerical complexity, model calibration, design optimization for energy efficiency, control synthesis, etc.

Dyad (Formerly JuliaSim) HVAC comes with a pre-built library of ready-to-use HVAC components and refrigerant models that integrate with advanced solvers, optimized for HVAC systems and compatible with the Dyad (Formerly JuliaSim) Scientific Machine Learning (SciML) ecosystem.

Challenges in HVAC System Design and Optimization

"Standard HVAC design tools are fragmented and lack the flexibility to handle the complexities of modern systems," said Dr. Avinash Subramanian, Dyad (Formerly JuliaSim) Simulation and Control Engineer. "Dyad (Formerly JuliaSim) HVAC shifts the paradigm by integrating physics-based modeling, advanced solvers, machine learning capabilities, and control design into a single, unified environment."

“HVAC & Refrigeration (HVAC&R) systems are inherently complex, involving multiple phases, fluid dynamics, thermodynamics, and control interactions. Current design workflows use disparate tools for each step such as modeling and simulation, calibration, design optimization, control design and machine learning,” says Dr. Chris Rackauckas, JuliaHub VP of Modeling and Simulation. “This siloed approach is not only computationally expensive and resource intensive, it sometimes fails to capture the full complexity of the system accurately. The Dyad (Formerly JuliaSim) HVAC platform enables all of these activities to be completed in one environment.”

The Dyad (Formerly JuliaSim) HVAC library integrates machine learning tools that can be used for automated model calibration to plant data, surrogate modeling for accelerated simulation, the discovery of unknown physics, automatic differentiation for efficient sensitivity analysis and model-based control to enable the design and operation of the next generation of HVAC systems. Case studies have shown a 60 - 570x speed-up compared with alternative tools attained by combining numerical techniques with scientific machine learning.

What’s the solution?

• Advanced Solvers: Robust solvers capable of handling stiff, nonlinear, and discontinuous dynamics
• Automated Model Calibration: Fine-tune simulation parameters to match experimental data using automatic differentiation with Dyad (Formerly JuliaSim) Model Optimizer
• Model Acceleration: Leverage Dyad (Formerly JuliaSim) Surrogates for accelerating models and smoothing out irrelevant or stiff dynamics with the help of neural surrogates
• External Model Integration: Integrate external Functional Mock-up Units (FMUs) and pre-trained surrogate models
• Export to FMUs: generate Model Exchange and Co Simulation FMUs for use in other simulation platforms
• Connect to Controls: Integrate with Dyad (Formerly JuliaSim) Control for PID, Linear and Nonlinear Model Predictive Control (MPC)
• Machine Learning Integration: Integration of Machine Learning workflows for enhanced simulation
• Specialized DAE Initialization: Specialized initialization routines for large-scale models with Differential-Algebraic Equations (DAE) provide the robustness needed to solve the non-linearities unique to multi-phase fluid systems.
• Thermodynamic Property Models: Spline-based thermodynamic property models for refrigerants (R32, R1234YF, R290, R152a, R134a, R410A, R717), dry air, and moist air
• Industry-Grade Pre-built Components: Pre-built components for Tube-Fin Heat Exchangers, Compressors, Valves, Fans, Conditioned spaces, and Pipes, with configurable options
• GUI for Rapid Prototyping: Intuitive drag-and-drop GUI functionality for rapid prototyping of system configurations

Dyad (Formerly JuliaSim) HVAC is now available for professionals and researchers looking to accelerate their approach to HVAC system design, operation, and optimization. For more information and to request a demo, contact the Dyad (Formerly JuliaSim) HVAC team at sales@juliahub.com.