GT-SUITE New 2026 Features

Export of Metamodels to FMU

Metamodels can now be exported to .fmu files for easy integration in other modeling frameworks.  Since FMUs streamline tool integration, metamodel FMUs function as simple plug-and-play models to ease the sharing of metamodels with other teams.

Faster Machine Learning Training

Accelerate your machine learning workflows with GT-SUITE’s new parallel metamodel training feature. Leveraging distributed computing, GT-SUITE now enables users to train multiple metamodels simultaneously, dramatically reducing computation time and unlocking faster insights.  Whether you’re innovating in design optimization or pushing the boundaries of simulation, this enhancement empowers engineers and analysts to scale their experimentation with efficiency and precision.

In addition, parallel training of up to 4 metamodels is automatically handled when using the Machine Learning Assistant on your local machine.

3D Flow

Optimized Cooling Module Simulation with GT-Auto-3DFlow: A prominent use case for Auto-3DFlow is the design of cooling modules with the tightly integrated 3D airflow solution, 2D heat exchangers and the 1D system. This workflow has been significantly enhanced to improve accuracy and reduce run time. This is achieved via optimized mesh and solver settings, improved initial values, and an enhanced fan modeling approach. It is also possible to reuse meshes from previously completed simulations, eliminating meshing time at the start of new runs and streamlining the setup process. The new ability to define mesh refinement zones enables added flexibility on mesh density, resulting in improved accuracy while retaining the fast run times that GT-Auto-3DFlow is known for.

Faster Steady-State Solutions with GT-CONVERGE: The latest solver significantly enhances numerical stability and convergence efficiency of steady-state problems thanks to a new under-relaxation-based approach. This results in simulation times that are 3x to 25x faster than previous versions. Additionally, the solver demonstrates improved robustness, especially when applied to large and complex models. Additionally, steady-state simulations now feature runtime monitors on the simulation dashboard, allowing users to track convergence progress in real time.

Fluid Properties Evaluator

Thermal fluid engineers occasionally need to evaluate fluid properties in early design and sizing studies, to validate test data, and more. The new fluid property calculator makes this information available instantly for any fluid, including refrigerants, liquids, gases, mixtures, and even humid gases. Points for each fluid state are overlayed with fluid property plots (pressure/temperature contours, psychrometric charts, P-h/T-s diagrams, etc.) to visualize the operating points.

Rapid model-building utilities for scroll and screw compressors

During compressor design studies, it is critical to evaluate the thermodynamic performance of different geometries using 1D simulation. Whether studying variable wall thickness (hybrid) scroll compressors or novel screw compressor rotors, engineers first need to preprocess the geometry before they can simulate the machine. With the latest features in GT-SUITE, it is easier than ever to build a detailed compressor model starting from 2D or 3D geometry.

Scroll compressors can now be modeled directly from 2D geometry, enabling faster design iterations. Additionally, hybrid scrolls with significant wall thickness variation are now supported so engineers can assess not only the compression characteristics but also potential force and moment imbalances. For screw machines, engineers can now automatically build a screw compressor model from 3D geometry using GT-SCORG.

Predictive PEM Fuel Cell Degradation

Minimizing fuel cell degradation is vital for extending the lifetime of the stack and reducing total cost of ownership. Physical testing requires hundreds of hours per iteration, and the result is a destroyed fuel cell. Predicting degradation via simulation saves time and avoids the expense of replacing damaged prototypes.

New, physical models for cathode catalyst layer degradation and membrane fluoride release are now available in GT-SUITE. The cathode catalyst layer models a distribution of platinum particle radii which evolve over time as various degradation reactions occur, such as platinum oxidation, carbon corrosion, dissolution, and Ostwald ripening.

This results in reduction of electrochemically active surface area (ECSA) as a function of operating conditions, such as temperature, humidity, and cell voltage. The membrane fluoride release model predictively captures how reactive radicals gradually thin the membrane over time, which leads to larger crossover rates that can reduce cell performance. These models enable engineers to avoid problematic operation and increase fuel cell reliability.

Co-Electrolysis

The possibility to model co-electrolysis, which simulates the simultaneous conversion of H₂O and CO₂ into syngas (H₂ and CO) has been added to GT-SUITE. Operating at typical temperatures of 800°C, the model can accurately capture complex electrochemical reactions and the water-gas shift equilibrium that determines the H₂/CO ratio in the product gas. The effects of different reactant concentrations on system performance can be studied by controlling inlet gas compositions and temperatures to predict outlet species concentrations across various electrolysis current ranges.

The ElectrolyzerStack template integrates seamlessly with GT-SUITE’s thermal and electrical modeling capabilities. The thermal interface allows detailed analysis of temperature distribution within the stack. The electrical interface enables connection to various electrical systems, with the direction of electrical signal determining operation mode, electrolysis or fuel cell.

Predictive Combustion Modeling for Alternative Fuels

To meet increasingly stringent decarbonization targets, many engine manufacturers are considering use of non-carbon or net-zero carbon fuels such as hydrogen, methanol, ammonia, etc. to help reduce net carbon emissions.  In recent GT-SUITE versions, new laminar flame speed and knock models have been introduced to help support predictive combustion modeling for these new fuels.  This trend continues in V2026 as we are introducing new sub-models for hydrogen blends, including the following new models:

• V2025 Build 2 – Laminar flame speed model for natural gas + hydrogen blends
• V2026 – Laminar flame speed model for ammonia + hydrogen blends
• V2026 – Knock model for natural gas + hydrogen blends

It should also be noted that the two models for natural gas + hydrogen blends were developed using data that included 0% hydrogen, so they can also be used to improve combustion model accuracy for natural gas mixtures without hydrogen blending, especially for natural gas mixtures with lower methane numbers.

Fast 3D Transport Model to Predict Fuel Stratification for Gaseous Direct Injection

While initial development of hydrogen engines focused on use of port fuel injection, many engine manufacturers are looking to switch to direct injection to both improve performance and help mitigate abnormal combustion events such as pre-ignition and backfire.

Due to the late injection timings required to achieve full performance benefits associated with hydrogen direct injection, it is increasingly important to model fuel stratification within the cylinder and the corresponding impact on the combustion rate and emissions formation.  To support predictive combustion modeling of direct injected hydrogen engines, we have introduced several new stratification options for our SITurb combustion model, including a 3D charge motion and scalar transport model which is capable of predicting the fuel distribution along the flame surface at each time step during combustion, leading to improved combustion rate and NOx predictions for operating conditions where significant fuel stratification is present.

Scroll Compressor FSI Solution – Leakage Prediction

Over the past couple of versions, developments have been made to generate the chamber maps that allows accurate application of pressure on the scroll structure, to capture the elastic deformations on both sides of the thrust bearing providing more accurate oil film thickness, load sharing and tilt effects.

Coupling these with the flexible contact solution and the thermal deformation of the structure, the scroll compressor leakage model was implemented by first extracting deformation-informed clearances along the radial and flank direction. These evolving contact gaps between the scrolls were then looked up on the chamber maps to identify the combination of chambers with leakage.

Speed Controller for Shafts

Accurate modeling of shaft torsionals requires proper representation of speed oscillations, which traditional “Speed” mode in Crank/Cam/ShaftAnalysis fails to deliver by imposing constant instantaneous speed. This approach contradicts real-world behavior where torsional vibrations cause speed variations around an average value and artificially stiffens the shaft by preventing natural oscillations. The current workaround requires engineers to use “Load” mode, where they must implement and manually tune a PID controller to adjust external load torque for maintaining target average speed. This process demands sensing, actuation signal configuration, adding unnecessary complexity to what should be a straightforward modeling task.

We are introducing “ControllerShaftSpeed” in v2026, a sophisticated new template designed specifically for maintaining target average shaft speed over a cycle. The solution dramatically simplifies the modeling process by eliminating the need for manual PID controller implementation and tuning – engineers simply connect a single template to the shaft requiring speed control.

The system automatically senses shaft inertia, calculates initial torque estimates, and self-tunes by running with constant speed for initial cycles to determine optimal PI gains.