关键词: 成组算法, 特征提取, 核元素, 非平稳信号, 广义夹角余弦, 刀具

Abstract: Traditional group algorithm couldn't guarantee core element as optimal. To deal with this problem, an improved group algorithm was proposed. By this algorithm, optimum core elements were obtained. As a result, each group had the most elements satisfying similarity threshold consequently the number of groups reached the fewest. Compared to fuzzy clustering algorithm, the proposed method had better convergence time and simple structure. Vibration signals acquired from a cutting tool degradation experiment was used to evaluate this method, and satisfactory results were obtained.

Key words: group algorithm, feature extraction, core elements, nonstationary signals, cosine of generalized angle, cutting tool