Multiple stages of model optimization

MapleSim uses powerful symbolic techniques to generate model equations, which allows for multiple stages of model optimization. Complex models are simplified and streamlined, resulting in a very compact, numerically efficient model that can run significantly faster than other simulation systems when dealing with large and/or complex models.

Click to enlarge

Symbolic preprocessing provides optimal formulation
Since model equations are generated symbolically, MapleSim can simplify your equations using symbolic techniques before solving them numerically, thus increasing computational efficiency and simulation performance without losing fidelity of results.

Simplification algorithms are designed for efficient simulation. Simplification techniques include:

• Symbolic index reduction of differential algebraic equations (DAE) and differential elimination to simplify system equations into equivalent but more tractable forms
• Automatic substitution and algebraic simplification
• Elimination of redundancies such as operations that are equivalent to multiplication by 0 and 1

Built-in compilation speeds up numeric computation
After the simplification phase, you have the option of using compilation to speed up the simulation. Maple procedures generated by the simulation engine are translated to C code, which is compiled automatically. Built-in model compilation of the simplified model can dramatically speed up the simulations, especially for complex models.

Model caching delivers high-speed simulation for multiple runs
Unlike purely numeric solvers, which need to iterate to a solution at each time step in the simulation, MapleSim only needs to create the system equations once, at the beginning of the simulation. The same equations are then used for every time step of the simulation. Once formulated, MapleSim stores these equations so that they can be reused for multiple simulation runs with different parameter values, making subsequent runs significantly faster than the initial run.