Unlike most competitors, PESC does not rely on opaque and rigid commercial simulation softwares for its motor design activities. PESC leverages PESTO, a software developed in-house, compiling more than 15+ years of experience in motor design and manufacturing in a single tool. PESTO is a finite element based motor design software, incorporating advanced models to evaluate performance (torque curves, efficiency maps, demagnetization, fault conditions…). PESTO has also been developed with computational efficiency in mind, in order to deliver the most robust sizing process by rapidly generating high fidelity databases of motor.

a numerical motor production line at your fingertips!

PESTO's high fidelity results are the fruit of the combination of the well known open source FEMM software with layers of semi-analytical models. PESTO delivers accurate information about torque, losses, fault conditions, demagnetization, etc. More complex phenomenon like 3D effects and high frequencies (due to PWM switching) are also considered thanks to advanced state of art models.

PESTO's ability to generate databases mostly rely on two aspects 

• smart motor design software architecture exploiting physical scaling laws and invariants

• connection to big data infrastructure on the cloud 

PESTO can handle a range of motor technologies, so far the focus has been on radial flux synchronous motors with concentrated windings. More motor types will get integrated on demand. Future features will also include thermal models. 

Typical PESTO outputs : geometry, torque-speed curves, efficiency and current control information

Numerical models are truly awesome but a model is only as good as its assumptions. So making sure predictions fit reality should be the #1 priority.  In the slides below we show how PESTO compares against experimental data. Most of the benchmark is coming from scientific literature and is focused around FSCW SPM machines (used in light EV, robotics, drones...). 

The goal is not to have perfect predictions per motor but to be accurate enough across a wide range of motors (relevant in the context of motor databases). So PESTO’s model has been calibrated at a high level and there is no case by case fudging to improve accuracy at an individual motor level. Although it’s not the point, it would be easy to tweak e.g. flux linkage/leakages and loss factors to reduce discrepancies if required.

Nevertheless, PESTO's accuracy is very acceptable. For papers [1-3] in the slides, PESTO predictions of power curves and losses are within 5-10% of experimental data.