Moving your model from desktop simulation to real-time simulation is an iterative process that can require extensive model reconfiguration. The real-time model preparation workflow shows how to analyze and decrease the computational cost of your model on a development machine. After completing the workflow, you can determine optimal fixed-cost solver settings for simulation on a real-time target machine.
Determine if your Simscape™ model is accurate enough to generate expected simulation results and fast enough to run on your real-time target machine without generating an overrun.
Real-Time Model Preparation Workflow
Follow the real-time model preparation workflow to make your Simscape model real-time capable.
Simulating a Simscape model in real time requires a balance of speed and accuracy that you can attain by reducing computational costs, optimizing solver configurations, or increasing processing power.
Determine the maximum step size to use for fixed-step simulation by analyzing the results from a variable-step simulation of your Simscape model.
Increase Simulation Speed Using the Partitioning Solver
Improve performance by using the Simscape Partitioning solver to convert a large system of equations into several smaller systems of equations that are easier to solve.
Make your Simscape model real-time capable by identifying and eliminating unnecessary, computationally costly processes such as redundant monitoring and data logging.
To make your Simscape model real-time capable, decrease computational cost by reducing fast dynamics that you identify using frequency-response and pole-speed analyses.
Eliminate components that cause rapid changes to reduce the computation cost of simulation and to make your Simscape model real-time capable.
Eliminate components that cause zero crossings to increase the minimal step-size for fixed-step simulation and to make your Simscape model real-time capable.
Partition a Simscape model for parallel processing on real-time processors.
Select model variants for dynamic systems using variant subsystems.