关键词: 齿轮箱, 加速寿命试验, 预测与健康管理, 全寿命周期, 多模态信号

Abstract: Prognostics and health management(PHM) is a pivotal technology to industrial intelligence and digital transformation, effectively reducing economic losses caused by unexpected downtime and extending equipment lifespan. However, PHM technology faces severe challenges of data scarcity, manifested in the fact that existing public failure datasets predominantly consist of artificially induced fault conditions, with limited availability of full-life cycle datasets spanning from healthy to complete failure. In addition, issues such as idealized test conditions, limited modal diversity in data, and insufficient data for multiple gearbox types also severely restrict the development of this field. To address the challenges, Professor Li Yongbo's research team selected parallel-shaft gearboxes, planetary gearboxes, and bevel gearboxes. Accelerated life test of gearboxes was conducted, and the degradation data has been publicly released to the scholarly community as the “NWPU Multi-Type Gearbox Accelerated Life Test Datasets”. This dataset encompasses full-life signals for 3 gearbox types and 7 operating conditions. It includes detailed data such as multi-point vibration and acoustic signals, clear fault types, failure start and end times, effectively creating a detailed “multidimensional portrait” for the gearbox. The release of this dataset provides robust empirical support for in-depth research into the full-life cycle failure evolution patterns of gearboxes. It cultivates fertile ground for the development of next-generation intelligent operating and maintenance algorithms, bringing industrial intelligent maintenance technologies to fruition.

Key words: gearbox, accelerated life test, prognostics and health management, full-life cycle, multimodal signals