To address the limitations of the electric drive system in terms of performance, this paper proposes a performance detection method based on key parameters, which introduces machine learning algorithm based on Gradient Boosted Decision Tree (GBDT) to construct and verify the performance quality detection model, including quality sample data collection, feature engineering analysis, model construction, model fusion and model self-training mechanism, etc., to determine product performance inspection quality based on key production and processing parameters of electric drive system. Combined with production practice, the paper proposes the idea of application scenario construction and its deepening direction. The results show that electric drive system quality inspection can be realized based on key parameters.

Artificial intelligence  /  Performance inspection  /  Electric drive systems  /  Quality modeling