PowerBrain is an intelligent AI-based power converter design automation tool. It is a software design aided tool which integrates novel AI algorithms and power electronics design simulations for realizing a fast and optimal power converter design. It consists of two major parts: one is the intelligent datasheet interpreter, which can automatically extract dynamic data from datasheets, providing adequate data for component selection. Another part is the magnet core loss dataset, which provide adequate data for magnetic design optimization.
Power Electronics design, which has been working as an important enabler technology in electrical engineering, has encountered new challenges these days. On the one hand, new semiconductor technologies have emerged, providing more choices for power converter designers with different performances on efficiency, frequency etc. On the other hand, new applications of power converters have been developed, such as the electrification of vehicles and aircraft, solar and wind power conversion. Those applications normally require power converters having higher efficiency and high power density. This is a challenge for power converter designers: how to design a power converter with good performance in a shorter time.
PowerBrain is an intelligent AI-based power converter design automation tool which integrates AI algorithms, power electronics design simulations for realizing a fast and optimal power converter design. It can break through the limitation of traditional empirical design methods and realize the optimal design in a faster way.
PowerBrain consists of two major parts: the intelligent datasheet interpreter and magnet core loss dataset, which targets the two pain points of semiconductor component choice and magnetic component design optimization. The datasheet interpreter is an online program that can extract dynamic data from semiconductor datasheets by recognizing the necessary figure elements, then extracting and exporting the data automatically in a machine-readable format. It facilitates the effective construction of semiconductor datasets, hence enhancing power electronics design automation. The magnet core loss dataset provides an experimental database for core loss characterization, in which the core losses in different materials with various sizes and shapes can be analysed accurately. An artificial neural network model for core loss predictions will be displayed.
The tool is under continuous development. To learn more, visit the PowerBrain website.