基于模糊模式识别的抽油井示功图判别算法研究

Pumping Well Dynamometer Discriminant Algorithm Based on FPR

针对有杆抽油井示功图诊断分析问题,研究特定故障情况下示功图特征值算法。把示功图数据通过傅里叶变换由空域转换成频域形式,仅保留较低频率上的系数a0、a1、b1,并将其作为特征值。根据所得到的特征值结合模式识别算法能够实现有杆抽油井示功图快速准确的识别和分类。实际应用表明,利用基准示功图28种类型的特征值的选取,计算出每种示功图的特征矢量,并保存到特征库中。对给定示功图特征值进行模式识别,即可得出该图隶属于哪种模式。

For rod pumping wells diagnosis of the indicator diagram analysis to study the eigenvalue diagram shown in the specific fault conditions. The indicator diagram data converts the airspace to the form of a frequency domain through the Fourier transform, retaining only the lower frequencies on the coefficients a0, a1, b2, which is as a characteristic value. Eigenvalue pattern recogni- tion algorithm achieves that a rod pumping wells show fast and accurate identification and classification of the diagram. The prac- tical application shows that using the selection of the benchmark midicator diagram 28 types of eigenvalue diagram, calculates the feature vectors for each indicator diagram, and saves it to the feature library. To make mode recognition the given diagrams of ei- genvalues, it can be derived the diagram belongs to which mode.