关键词: 示功图, 模式识别, 傅里叶变换, 特征值算法, 离散傅里叶

Abstract: 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 a0, a1, b1, which is as a characteristic value. Eigenvalue pattern recognition algorithm achieves that a rod pumping wells show fast and accurate identification and classification of the diagram. The practical 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 eigenvalues, it can be derived the diagram belongs to which mode.


Key words: indicator diagram, pattern recognition, Fourier transform, eigenvalue algorithm, DFT