倒谱法检测信号异常的灵敏度分析

The sensitivity analysis of signal abnormality detection based on cepstrum

信号异常检测的主要目的是寻找对信号反应灵敏的特征参数.本文从倒谱法的算法、性质等阐释了倒谱系数对信号异常具有较好的灵敏度,并通过实际计算,验证了倒谱系数对信号的振幅、时间延迟和相位参数微小变化的灵敏度反应较好.对所设计的指数衰减正弦波信号,倒谱法能够较容易地检测出振幅、时间延迟和相位参数相对变化为百分之一的信号,而时间延迟和相位参数相对变化仅为千分之一也能较容易地检测出来.并利用该方法对某气田的含气性检测进行实验性处理,取得了较好的效果,表明倒谱法检测由油气引起的地震信号异常也是可行性.

The main purpose of signal abnormality detection is to find the sensitivity of the signal characteristics. In fact, using seismic signals to detect Hydrocarbon is similar to the signal detection. It finds that oil or gas, which only accounts for a small part in reservoirs caused by seismic signal anomalies. In this paper, we introduce method of cepstrum and explain the method of cepstrum and the property of cepstrum, and also expound cepstral coefficient is good for the sensitivity of the signal abnormality detection. Through practical calculation, it verifies that the cepstral coefficient is good sensitive to be response to the signal amplitude, time delay and phase parameters, which changed a little. For the design of the exponential decay of the sine wave signal, the method of cepstrum can easily detect the relatively change one percent of the signal amplitude, time delay and phase parameters, the time delay and phase parameters of the relatively change is only one in a thousand also can be easily detected. Finally, we use the method of cepstrum to experimental processing for oil and gas detection for a gas field T3 X2 in Sichuan, and have achieved good results. This indicates that the method of cepstrum detected seismic signals, which caused by the oil and gas anomalies is also feasible.