测井曲线形态的自动识别方法研究

为了进行测并沉积学和测并层序地层学的研究，需要对测井曲线形态进行识别。首先进行测井曲线一、二阶导数的求解，当二阶导数呈现符号转换时的深度点即曲线的拐点，亦即测井曲线的地层分界面，然后利用拐点限定范围内测井曲线的变化趋势进行曲线的形态识别，即当测井曲线数值呈现减小的变化特点称曲线的形态为钟形，数值呈现增大的曲线形态为漏斗形，数值变化幅度较少时则称曲线形态为箱形，从而完成测井曲线形态的自动识别。按照以上方法进行了某取心井测井曲线的数据处理和形态的自动识别。从实例分析知，通过测并曲线二阶导数的求取、拐点的识别以及拐点限定范围内测井曲线的变化趋势分析，能较好完成测井曲线形态的自动识别。

Study on well logging curve shape automatic identification method

Well logging curve shape identification is needed in study of logging sedimentology and logging sequence stratigraphy. Firstly, the first and second derivatives of logging curve are solved. The depth point where the second derivative shows symbol conversion is the inflection point of the curve, that is, the formation interface of logging curve. Next, curve shape is identified according to the variance trend of logging curve in the range defined by inflection point. Curve shape is called campaniform when the logging curve values show decline feature, while curve shape is called infundibular when the values show ascending feature. If value change amplitude is small, curve shape is called boxshaped. Thus the automatic identification of logging curve shape is achieved. According to the above method, data processing and automatic shape identification have been implemented to certain coring well logging curve. Analysis of this example shows that automatic identification of logging curve shape can be successfully achieved through logging curve second derivative solving, inflection point identification and logging curve variance trend analysis within inflection point defined range.