自适应模糊神经网络推理系统在电缆故障定点检测中的应用

Filtering Technique Based on Adaptive Neural-fuzzy Inference System Applied to Cable Fault Location

电缆故障定点检测中存在着噪音复杂多变、信噪比较低的问题。自适应模糊神经网络推理系统（ANFIS），具有非线性映射和自学习能力，能够用于噪音信号的非线性建模，可以很好地解决故障定点检测中噪音的不确定性问题。建立了基于ANFIS的自适应滤波系统模型，讨论了ANFIS中隶属函数类型、数目的寻优以及参考函数的确定方法，设计了以ANFIS对消器为模型的滤波系统的硬件和软件。用该滤波系统对实际电缆故障信号进行了处理，结果表明了滤波系统的有效性。

There are some difficulties caused by complexity and variety of noise and lower signal-to-noise ratio in the signal processing of cable fault location. The adaptive neural-fuzzy inference system （ANFIS） has the ability of nonlinear mapping and self-learning property and can be used to achieve the nonlinear model of the noise. ANFIS can also solve the problem of uncertainty of noise. This article builds the model of adaptive filter based on ANFIS, and discusses the selection of the type, the number of membership functions and the reference functions in the ANFIS. The hardware and software of filtering system are also designed. At last we use the filtering system to access a factual cable fault signal. The result shows that the filtering system is reliable.