Products + Technology

Product Monday: Generative-Designed Heat Sinks Using Topology Optimization

Diabatix nv custom designed heat sink.

A Belgian thermal management company named Diabatix nv is offering custom-designed heat sinks that use artificial intelligence/machine learning (AI/ML) generative design to create computer-generated shapes that can be optimized to achieve different engineering goals.

Generative design is a process that uses algorithms and machine learning techniques to create and optimize designs based on specified design goals and constraints. It is a form of computer-aided design (CAD) that allows designers and engineers to explore many possible design variations and alternatives quickly and efficiently. In generative design, the designer inputs a set of design parameters such as dimensions, material properties, and performance requirements, and the software generates multiple design options based on these inputs.

The generated designs are then evaluated using simulations and other analysis tools to determine their suitability. The resulting designs can be highly optimized for specific performance criteria, such as weight, strength, or cost. Generative design can be applied in a variety of fields, including architecture, industrial design, product development, and engineering. It can help designers create innovative, optimized designs faster and more efficiently and has the potential to revolutionize the design process.

Starting from any geometrical shape and with minimal input, generative design utilizing topology optimization can be used to determine the best shape for a manufacturer’s needs. Product developers set the objectives and impose the constraints most suitable to the design problem.

More information is available here.

Want to read our other stories about heat sinks? Click here


Leave a Reply


LightFair 2023
IALD Enlighten Europe
Lightovation – Dallas Market Center
2023 IES Annual Conference
Click For More


Coming Soon - to list your position here, please contact Suelynn at








%d bloggers like this: