Design Automation of Heat exchangers using Deep reinforcement learning

点击翻译按钮获取中文摘要

The objective of this project is to further develop a software with an AI engine for design automation of thermal devices. This phase of the project is essential for the product to become commercially viable and competitive. Heat exchangers play a vital role for the removal of unwanted heat generated in a device and, as such, have wide variety of applications in cooling electronics, HVAC, data centers, and electric vehicles. Heat exchanger design automation software is capable of automatic optimization of heat exchanger geometry based on parameters such as available space and operating conditions. Design Automation of Heat exchangers using Deep reinforcement learning provides significant improvement in design performance of up to 60 percent and enhances productivity of engineering design teams through automation. This project aims to implement new features into the existing design automation software to include high volume manufacturing processes and compressible flow conditions. Another goal of this project is the case-study and manufacturing of prototypes with industrial applications to demonstrate the manufacturability of the designed heat exchangers. Successful implementation of these features along with cost per design analysis will greatly improve the commercial viability of heat exchanger design automation software. It also provides industry leaders with proof of concept of the capabilities of our software.