| Steel is the one of the major materials widely used in most of the industry fields for its high strength,low cost and good processability.However,there inevitably appear some corrosion pits,cracks or grooves in the steel structure,which would lead to a potential engineering accident.Thus,the non-destructive testing,as the last line of the engineering safety,is required to detect various defects in different cases.For the case of large-scale of steel plate and long-distance pipeline,the defect detection technique shall own high efficiency,reliability,and automation,which prompts the MFL(Magnetic Flux Leakage)testing to be the main detection technique in this occasion.This thesis first focuses on the noise problem of the MFL signals,analyzing the noise sources and classifying the noise as background noise,distension noise and vibration noise.To address the noise issue,a new detection strategy by measuring the MFL change rate is proposed.The strategy is demonstrated by the MFL change rate model established on the basis of magnetic charge theory.Then through a series of simulations on the detection process,the deduction results of the model are verified,including the detection feasibility and the noise suppression.Finally,the practical detection is carried on and the results agree well with the simulation data,manifesting the feasibility and advantage of the new detection strategy.Then the thesis concentrates on the low defection sensitivity of the axial crack in the steel pipe.Based on the ACFM(Alternating Current Field Measurement)principle,the traditional PIG(Pipeline Inspection Gauge)is modified to induce an extra eddy current flowing around the pipe wall,enhancing the axial crack detection ability.The optimal parameters of the new PIG are determined by a set of tests before crack detection experiment.By analyzing and imaging the composite detection signal,it is found that the static MFL field and the secondary induced magnetic field are significantly impacted by the circumferential crack and the axial crack respectively.This well verifies the modified detection method and shows good performance of the new PIG. |