It is difficult to detect leakage for oil pipelines in complicated conditions. For solving the difficulty a method based on SVM (Support Vector Machine) is proposed and the diagnosis model was established. Model train can be completed in few samples to distinguish different conditions of pipelines. The experimental result demonstrates it was effective in the classification with a few samples and the correct rate increased more greatly compared with traditional BP method. Moreover, in hot pipelines pressure velocity is affected by oil and pipeline axial temperature drop. Location usually has obvious error. For solving this problem axial temperature drop was analyzed and pressure velocity was revised. By means of Newton-Cotes integration method location formula was improved. The field experiments show that the improved located formula made location accuracy increased from 2.5% to 1.0%.
[1]
Pijush Samui,et al.
Support vector machine applied to settlement of shallow foundations on cohesionless soils
,
2008
.
[2]
Wen Xingpu.
Application of Numerical Analysis in VAR
,
2010
.
[3]
Patrick Haffner,et al.
Support vector machines for histogram-based image classification
,
1999,
IEEE Trans. Neural Networks.
[4]
Peng Shan-bi.
Leakage detection for oil pipeline based on negative pressure wave theory
,
2009
.
[5]
Luis F. G. Pires,et al.
TRANSIENT FLOW ANALYSIS OF FAST VALVE CLOSURE IN SHORT PIPELINES
,
2004
.
[6]
R. T. Walczak,et al.
Using Support Vector Machines to Develop Pedotransfer Functions for Water Retention of Soils in Poland
,
2008
.