【摘要】 地震地下流体观测仪器传感器故障频发，如何快速、准确的检测故障点是国内外一直以来研究的课题。传统的硬件冗余法经济、人力成本太高；而传统的傅里叶变换算法又具有局限性。由于系统的复杂性，可靠故障检测与隔离(FDI)技术中非线性过程的方案往往耗时且难以实现。神经网络和模糊模型能够逼近非线性动态函数，为解决上述问题提供了有力的工具。提出一种基于小波变换的非线性过程的局部线性模型算法，并结合湖北台站地下流体数据对此算法进行仿真。结果表明上述算法在信号去噪、故障检测中效果显著，具有很广泛的应用前景，为台网仪器维护人员判断仪器故障提供了一个全新的研究方向。更多还原

【Abstract】 The sensor faults of seismic underground fluid observation instruments occur frequently. How to detect the fault point quickly and accurately has been a research topic at home and abroad. The traditional hardware redundancy method is economic and the labor cost is too high; while the traditional Fourier transform algorithm has limitations. Due to the complexity of the system, the solution of nonlinear processes in reliable fault detection and isolation(FDI) technology is often time-consuming and difficult to implement. Neural networks and fuzzy models can approximate nonlinear dynamic functions, providing a powerful tool for solving this problem. This paper proposed a local linear model algorithm based on the nonlinear process of wavelet transform, and combined the underground fluid data of Hubei station to simulate the algorithm. The results show that the algorithm is effective in signal denoising and fault detection, and has a wide range of application prospects. It provides a new research direction for station network instrument maintenance personnel to judge instrument faults.更多还原