Study Of Magnetic Leakage Field Characteristics Defect Quantitative Method Under Multi-magnetized Units Testing

With the development of natural gas and petrochemical industry,domestic pipeline industry also rapidly develops in recent years.In service,the pipeline is often under the attack of internal and external environment.For the causes of environment,medium,effusion,pressure influctuation,etc.the pipeline can be easily corroded and its lift is shortened.Now,the managers and inspectors mainly concern themselves with several aspects,including the detecting means on pipeline corrosion state,targeted repairment,life extension techinques,and so on.At present,there are two method for pipeline testing,namely,inspection in pipe and detection at outside.For the pipelines in petrochemical industry,most of them are made up with small diameter pipes and a large number of elbows,it is difficult to detect them with internal testing methods.To ensure their safe operation and do not disturbing normal production,a new applicable testing method carried out at outside is urgently needed.In the paper,a external inspection method based on magnetic flux leakage is studied after the investigation on characters of petrochemical pipelines above the ground.According to the requirement of pipe testing,the 3d static magnetic flux leakage testing model is established with finite element method.Then,the spatial distribution of magnetic leakage field is obtained uner different magnetized units number and different magnetized unit distances.At last,the influence rules beween magnetized units and cumulative rules of defect magnetic leakage field are searched out.In order to uniformly distribute excitation magnetic field,we make effort to determine reasonable magnet thickness and boots structure.Of course,the key is the design of adjustable magnetized structure with three magnetized units which can be used for pipeline whose diameter is 159mm or bigger.Aiming at the question that for a defect,different amplitudes will be obtained by different sensors,a MFL signals pretreatment model is established.After that,amplitude correct factor is advanced and the relation between signal amplitudes of three magnetic units is set up with equations.Then consistent processing of each detecting channel amplitude is realized.After the pretreatment of pipeline detecting data,the modern signal process methods are used to decompose and denoise the MFLsignals.Then the time and frequency domain are extracted as samples.With them,corrosion defect quantitative BP neural network is established for pipeline testing at outside under three magnetized units.In laboratory,MFL detection system is set up with different diameter pipes.After the inspection for artificial corrosion defects,the result shows that adjustable three magnetized units structure is feasible for pipeline inspection.With the raised analysis method and BP neural network,defect quantization can be realized.In field,the MFL testing result is matched with ultrasonic testing result on four pipelines including a Φ159 × 6mm hydrocorbon pipeline.