Erosion Testing System And Achievement Of The Data Method Based On Multisensor For Oil Transportation Pipelines In Seabed

The paper summarized the resent researching actuality, developing direction and problems in seabed and inland about the inservice oil pipelines nondestructive testing (NDT) techniques. The preponderances about the testing precision, reliability and facticity of multisensor method were analyzed theoretically, and the necessary and advantages of this method were also reviewed., the characteristics and lacks of ultrasonic and magnetic-flux testing methods were comparatively analyzed by the thesis, we consider the range may be widened, the probability of missing and error may be reduced, the reliability of the testing results may be increased if the testing pig for oil transportation pipelines was designed with the method combined ultrasonic and magnetic-flux. With 8 ultrasonic and 4 magnetic-flux detections, a detection flan was designed. The design of body and driver devices of the pig is part content of the paper, the pig drove by batteries schlepped by it may have much strong independence, even exactitude and adjustable movement speed. The essential parameters of ultrasonic and magnetic-flux detections was calculated and designed, the concernful parts such as batteries and motors were selected in this thesis too. About the design, the emphasis is the design of high rate sampling and saving of ultrasonic signals, in order to realize it, one 100MHz rate sampling A/D card was used in our system. We also used a PCI bus to control sampling and saving of ultrasonic, magnetic-flux and other assistant signals, as well as send of ultrasonic signals, operate and shut down of the pig and detection plan, regulate velocity and so on. On the other hand, a database was built to save original data in order to lookup and relate expediently when the data fusion was processing. In the end of the system design, we simply evaluated the pig, include precision, utility range and types of damnifications which can be tested out, the result is the pig which jointly used two sensors was excelled comparing with the pig only used one sensor in any aspect. The other emphasis of this thesis was the design of a new multisensor data fusion arithmetic that was a pivotal technique, can ensure system of the multisensor successfully used. After the data fusion technique based on the multisensor was reviewed in this paper, the fuzzy regression analysis based on the D-S evidence theory and support integration analysis was made use of two grades data fusion. Through narrow mathematic reasoning, the paper gave out calculation formulas of the subjection degree and interval estimates, farther achieved these formulas about different sensors. For farther described degree of the damnification,the fuzzy characterization measurement analysis based on the fuzzy regression analysis was used in our thesis, obtained the upper and lower limit calculation formulas of the fuzzy characterization measurement under given reliability and subjection degree. These formulas can depict any testing data to characterization measurement, which is much close to the true flaw. Based on the fuzzy regression analysis, the support integration analysis was used as the second grade data fusion. The purpose is confirming the integration information put up by a set of flaw data——support integration flaw size. It can integrate some of scattered flaw sizes to get the integration flaw's geometric shape, describe the flaw degree from characteristic level. If we combined this arithmetic with expert database and experience characteristic database, we can report much important information include the type and size of the flaw and if or not need to maintain. In order to testify the validity of the arithmetic, selecting actual data to be fused by it. From fusion results of the sixth chapter we can find the effect is consistent with the expectation. Fusion data were much close to the true flaw data after were fused. The result which regarding superscale flaws as un-superscale flaws may be avoided through adding safety factor, so the economic closing and safety problems produced by it can also be avoided effectively. We introduced VB language to write programs in the progress of calculating these simples. All programs were wrote as functions, which can be embedded and modified expediently when they were using.