基于极值分析的钻井参数刺峰噪点数据识别研究

Study on identification of spike noise data of drilling parameters based on extreme value analysis

在石油勘探开发钻井施工过程中,工程监测中产生的离群刺峰噪点数据严重影响智能化诊断报警的准确度。为了准确识别噪点数据,提出了一种基于极值分析的钻井参数刺峰噪点数据识别方法,该方法以噪点数据明显偏离趋势线特征为标准依据,以期精准识别并剔除噪点数据,提升工程数据分析准确度。为此,首先介绍了样本数据极值点的筛选算法,再对极值点刺峰噪点识别算法进行了详细论述,并阐述了刺峰噪点附近数据的噪点识别判断算法,进而完成了对样本数据全部刺峰噪点的识别。将该算法应用于实际钻井现场30口井5种钻井参数的噪点数据识别,试验后识别的噪点数据与作业现场的实际情况吻合度达82%以上,经专业技术人员评估后,证实该方法可应用于实际作业现场。

During drilling construction of oil exploration and development process,the outlier spike noise data generated in engineering monitoring seriously affects the accuracy of intelligent diagnosis and alarm.To accurately identify the noise data,a drilling parameter spike noise data identification method based on extreme value analysis is proposed.The method is based on the characteristic of noise data that obviously deviate from the trend line,in order to accurately identify and eliminate noise data and improve the accuracy of engineering data analysis.Therefore,the screening algorithm for extreme points of sample data was first introduced,followed by a detailed discussion on the algorithm for identifying extreme point spike noise.The noise identification and judgment algorithm for the data near the spike noise points was also expounded,thereby completing the identification of all spike noise points in the sample data.The algorithm was applied to the noise data identification of 5drilling parameters in 30 wells at the actual drilling sites.The noise data identified after the test were consistent with the actual situation at the working sites by more than 82%.After evaluation by professional technicians,it has been confirmed that the method can be applied to actual working sites.