Research On Integrity Evaluation Method Of Gathering And Transportation Pipelines Based On Multi-source Data Fusion

Oil and gas gathering and transportation pipelines in China are large in scale,wide in coverage,highly corrosive in conveying media,diverse in pipeline specifications,complex in operating conditions,and there are many old pipelines that have been in service for more than 10 years,it is easy to cause pipeline accidents,causing serious economic losses and environmental pollution.The current pipeline integrity evaluation technology mainly relies on mechanical parameters such as internal pressure,and does not consider the influence of various factors such as medium and environmental complexity on the pipeline,and the gathering pipeline is not easy to meet the internal detection conditions,so the current pipeline integrity evaluation method cannot be directly applied in the gathering pipeline.Therefore,the research on the integrity evaluation technology of gathering and transportation pipelines is of great significance to the safe operation and maintenance of gathering and transportation pipelines.This paper analyzes the risk factors affecting the safety of gathering and transportation pipelines from the aspects of corrosion,man-made,design,operation,and natural factors,and forms a four-level big data characteristic index system that characterizes the integrity of gathering and transportation pipelines.The applicability of internal and external inspection methods for pipelines to gathering and transportation pipelines is analyzed,and the application scope and advantages and disadvantages of three integrity evaluation standards of ASME B31 G,API 579 and DNV RP-F101 are compared.The combination of non-contact magnetic stress detection is suitable for the evaluation method of corrosion defects in gathering and transportation pipelines.The research on the collection and processing methods of multi-source data was carried out,the data collection range and preprocessing method were proposed,the BWP index was used to evaluate the clustering effect and the optimal number of clusters was calculated,and the K-means algorithm was used to perform attribute domain analysis of continuous variables.Discretization converts multi-source heterogeneous data of gathering and transportation pipelines into data that can be analyzed and calculated.26 data indicators including pipeline ontology,inspection,operation and environment were collected,and a correlation analysis model based on mutual information was established with the estimated defect maintenance coefficient as the decision variable,and the key risk factors were extracted and the correlation degree of each indicator was ranked,the most relevant factors are defect size related factors such as defect length and defect depth,and the more relevant factors are the pipeline body and operating factors such as defect mileage,slope,and operating temperature.A method of jointly determining the weight of key risk factors by correlation analysis and analytic hierarchy process is proposed,and a risk scoring system is established to correct the safety factor of important parameters for corrosion defect evaluation in ASME B31G;the key risk factors are used as the input of the neural network,and the safety factor is Neural network output,establish a safety factor calculation model based on radial basis neural network,and form a 7-15-1 network structure.After training and testing,the average error of the model is 1.25%,and the maximum error is 7.81%,indicating that the model The prediction effect is better.The revised safety factor is combined with the ASME B31 G evaluation method to evaluate two gathering pipelines with different detection methods.For gathering pipeline 1,the ASME B31 G evaluation method considers that all detected defects are in a safe state,and the new method considers that there are 1 defect requires maintenance and repair as soon as possible.For gathering pipeline 2,MTM inspection and evaluation believes that there are 6 defect points that need to be repaired as soon as possible,and 19 defect points need to be repaired as soon as possible.The new method believes that there are 2defects that need to be repaired as soon as possible,and other defects are in a safe state.Based on the analysis of source information,the new method considers the influence of various factors on defects,and the evaluation results can better reflect the safety status of the defect points.