In recent years, the safety and failure of industrial pipelines have received extensive attention. Corrosion of industrial piping is a common but inevitable problem. In reality, in order to prevent sudden failure of industrial pipelines, it is necessary for the staff to periodically obtain data to calculate the corrosion rate of the pipeline. However, due to the unreliability that may occur during manual inspection, it is urgent to establish a model to help workers intelligently predict the corrosion rate of pipelines.

At the same time, more and more industrial fields are beginning to use machine learning to process huge amounts of data and exploiting the potentially immeasurable value behind industrial data. How to make in-depth analysis of complex and diverse data based on machine learning, and more efficient use of information has become the main direction of machine learning research in the current big data environment. Common algorithms for machine learning include decision tree algorithm, naive Bayes algorithm, artificial neural network algorithm, support vector machine algorithm and random forest algorithm. Among them, the support vector machine algorithm is based on a single kernel function, which maps the indivisible data to the high-dimensional space for data classification, so as to classify the data. Since the mapping methods of different kernel functions are different and the experience is only based on experience, there will be some contingency in the results.

In this paper, a predictive pipeline corrosion rate model based on multi-core learning is proposed. The corrosion data of Changqing Oilfield Gas Well pipeline collected by actual process is tested to prove the reliability of the model. First, the standard normalized data set, select the test data label (corrosion rate high, low and medium) and test data characteristics (daily gas, H2S concentration, pressure, CO2 content, etc.), and then practice testing, select 11 times the polynomial kernel function training kernel function selects the optimal classification output of the model when the Lam values ​​are 0, 0.01, 0.1, 0.2, 0.9, 1. The results show that the prediction accuracy of the corrosion rate for the Changqing Oilfield Gas Well pipeline corrosion data is over 95%. It has guiding significance for industrial prevention of pipeline safety.


Prediction of Pipeline Corrosion Rate Model Based on Multi-Core Learning

Division / Sponsors: 
Pressure Vessels & Piping (PVP)


Paper Proceedings: https://asmedigitalcollection.asme.org/PVP/PVP2020/volume/83815

PVP2020 - Virtual Conference

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.