A Diagnostically Lossless Compression Method and Fpga Implementation for Pipeline Magnetic Flux Leakage Inspection Data

In order to enhance the precision of the pipeline MFL (Magnetic Flux Leakage) detecting system measurement, detailed flux information obtained by decreasing time and space intervals for sampling is needed. The huge amount of data produced by this will exceed the processing ability of the system. Data compression is an effective way of solving this problem. In this paper we present a diagnostically lossless compression method of MFL inspection data. By analyzing the character of the MFL inspection data, we use axial differencing threshold and dynamic range threshold of data-block to determine the diagnostically importance of the block, and then use different Huffman coding method to compress data of different diagnostic data blocks. We also introduce the FPGA (Field Programmable Gate Array) implementation of the compression method presented above. Background Pipelines are recognized as the safest method for oil/gas long-distance transmission. There are more than 30,000km oil/gas pipelines in use in China, and nearly 10,000km pipelines abuilding. The basic requirement of oil/gas pipelines transportation is safety and efficiency. However 50% of the pipelines in China have been used for several decades. Pipeline leakage sometimes occurs due to defects caused by erosion, abrasion and so on. If not detecting and repairing these defects in time, huge natural resources will be wasted and great damages will be done to the environment and people once the pipeline fails [1]. The integrity of pipeline is a worldwide problem. Many countries have paid more attention to the detection of metal loss of the long pipelines and spent much money on developing pipeline-inspecting devices. Among techniques for in line inspection (ILI) such as: MFL, ultrasonic and eddy current, MFL is the most reliable and established one, holding around 80% of the market for metal loss inspection, because it doesn’t require coupling liquid and very ‘clean’ pipeline, and is capable of providing high resolution data on the sizes and geometry of defects in most pipeline environments [2]. MFL uses the magnetic intensity leaked out from the surface of magnetized ferromagnetic material to estimate the size of the defect. Figure 1 is the illustration of MFL test. If there are no defects on pipe wall, magnetic flux tends to travel through steel rather than through air or oil/gas, as figure.1a shows. If a defect exists, as fig.1b shows, it will increase the magnetic resistance and make the magnetic field aberrance at and near the defect. Some flux will leave the surface of pipe wall and pass the defect. So it is called leakage flux. Hall effect sensors can measure the leakage flux. The signal amplitude is closely related to the parameters of the defect such as size, depth and shape. The working process of MFL inspection is: the device carrying electrical source moves with the medium transmitted by the pipe, magnetizes the pipe wall by the magnets on the device, and uses sensors to measure the flux leakage intensity. Measured data are stored in storage device after being compressed. We can import the data into the computer, and then provide the information of the defect size and location by tables and images after inspection. Pipeline operator can thus use this information to ensure pipeline integrity. 1 The project is supported by the China National Natural Science Fund. No. 60272015 Key Engineering Materials Vols. 270-273 (2004) pp 651-656 online at http://www.scientific.net © (2004) Trans Tech Publications, Switzerland Online available since 2004/Aug/15 All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 130.203.133.34-16/04/08,14:29:55) Title of Publication (to be inserted by the publisher) (a) (b) Fig. 1. The principle of MFL test Fig.2 shows the basic structure of MFL PIG. It is made up of four parts: driving section, measuring section, computer section and battery section. Scrapers installed in driving section to generate pressure so as to drive the PIG and clean pipes. Measuring section contains magnets and sensors. Computer section is the kernel part of the device. It controls the whole in-line inspection process. Battery section offers enough power for the device. Fig. 2. The structure of MFL PIG Since 1997, the industry measure and control center of Shenyang University of Technology has designed and built Φ377mm, Φ325mm, Φ273mm, Φ426mm oil pipe MFL PIGs and supplied inspection services for nearly 2000km pipes in several oil fields in China. Inspection results have been validated by field digging. Now we are designing MFL PIGs for gas pipelines inspection. Objective Because inspection devices run fast in gas pipelines, the decrease of the time interval is needed in order to get more leakage flux information. High resolution will produce huge amount of data: in our gas pipe inspection device, the amount of data exceeds 70Gbyte. Because of the limitation of device size, we cannot use normal computer system. The measured data must be compressed in order to be stored in real time. The compression ratio is low when using lossless methods to compress the MFL data: we now use Huffman coding to compress the differencing value of closest data in every axial channel and get compression performances of 3:1. But we hope the ratio can exceed 10:1, then we must use lossy methods. However the losses of data should not affect the accuracy of inspection and the important data associated with the diagnosis must be preserved. The compression is diagnostically lossless. The diagnostically lossless compression use lossless methods to preserve the diagnostically crucial information and use lossy methods to compress the unimportant data. By this, we can get more compression while ensuring the inspection accuracy. Various diagnostically compression methods were introduced [3,4], but these methods are so complex that it is difficult to be implemented by hardware and it can not fit the real time requirement by software realization. In this paper, we 652 Advances in Nondestructive Evaluation