Inverter semiconductor Temperature Prediction AI Modeling Challenge

The Inverter Semiconductor Temperature Prediction AI Modeling Challenge is a new challenge created for the first time this year, requiring players to solve specific application problems by mining digital value: Based on the massive data from real industrial scenarios provided by ABB, we comprehensively utilize advanced technologies such as machine learning and big data prediction to analyze the relationship between equipment data, and independently develop and establish the temperature estimation model of insulated gate bipolar transistor (IGBT), the core component of inverter, in the Python environment. The temperature fluctuation of IGBTs under different load power is monitored with high precision and fast response. The accuracy and innovation of the model is the key to determining the success of the player.

86 teams and 194 people from 51 universities across the country participated in the competition. In the process of the competition, the judges evaluated and scored the contestants’ works from multiple dimensions such as model accuracy, innovation, feature engineering, etc. A total of 5 teams were selected for the final and participated in the on-site final defense. After a rigorous review by a number of experts from the Service and Transmission Products business unit of ABB China Motion Control Business Unit, Wu Shunyu and Gan Ziyi from Shanghai Jiao Tong University won the first prize, 4 groups of players won the second and third prizes, and 13 groups of teams won the excellence award.

V0701596-001102

Mr. Qi Luping, head of ABB Motion Control Business Division in China, spoke highly of the organization of the competition and the performance of the participating students, saying: “We are very happy to see so many college students actively participate in the competition, and we also hope that through this competition to provide a platform for college students to help them apply the methods and theoretical knowledge learned in school to production practice, experience how to use and analyze data in real industrial scenarios, and find ways to find solutions in non-ideal or missing data states.” Solve real production problems. Participants in this process will get a very good experience, but also expand the understanding of frequency converter technology and industrial knowledge. At the same time, their innovative ideas can also provide us with some new ideas and inspiration. We also hope to use this platform to provide excellent college students with the opportunity to show their talents, and cultivate interdisciplinary talents with excellent innovation and practical ability for society.”