掘进履带行驶系统作业于煤矿巷道恶劣地形和复杂环境中，其关键承载部件驱动轮长期承受不均匀载荷，导致轮齿损伤甚至断裂，影响掘进装备的正常生产作业和行驶平稳性。为了及时检测驱动轮轮齿损伤状态，避免故障扩大和降低维修成本，在阶次跟踪算法、包络谱分析、线性峭度算法和滤波算法的基础上，提出了一种适用于变转速机械的阶域线性峭度算法。通过对只含基频、包含基频和谐频的合成仿真信号进行包络谱分析及特征分布统计，表明线性峭度相较于峭度更具优势。通过使用变转速和多噪声振动信号进行驱动轮齿损伤状态识别，并与快速谱峭度、阶域线性峭度等算法进行对比分析，表明阶域线性峭度算法具有适用性和鲁棒性。该方法可有效提高故障检测的准确性和效率，为履带行驶系统的安全健康运行提供了有力保障。

The crawler maneuvering system used for tunneling operates in the harsh terrain and complex environment of coal mine tunnels. The driving wheel, as a key load-bearing component, is subjected to uneven loads for extended periods, leading to damage or even fracture of the wheel teeth. This can impact the normal production operation and smooth travel of tunneling equipment. In order to detect the damage state of the driving wheel gear teeth in time, avoid the expansion of the fault, and red
uce maintenance costs, an order-domain linear kurtosis algorithm for variable speed machinery is proposed. This algorithm is based on the order tracking algorithm, envelope spectral analysis, linear kurtosis algorithm, and filtering algorithm. Through envelope spectrum analysis and feature distribution statistics of synthetic simulation signals containing only fundamental frequency, fundamental frequency, and harmonic frequency, it is shown that linear kurtosis has more advantages than kurtosis. By using variable speed and multi-noise vibration signals to detect the damage state of driving gear teeth, and comparing it with algorithms such as fast spectral kurtosis and order domain linear kurtosis, it is shown that the order domain linear kurtosis algorithm has applicability and robustness. This method can effectively improve the accuracy and efficiency of fault detection, providing a strong guarantee for the safe and healthy operation of the crawler maneuvering system.