花_格管线腐蚀穿孔失效模型分析.pdf

2000年7月第15卷第4期西安石油学院学报(自然科学版)Journal of Xian Petroleum Institute(N atural Science Edition)Jul.2000Vol.15 No.4 收稿日期:2000202218 作者简介:樊玉光(19642),男,山西翼城人,副教授,主要从事机械工程方面的研究工作.文章编号:100125361(2000)0420070203花-格管线腐蚀穿孔失效模型分析Buried Pipeline Corrosion Punching Fa ilureM odel樊玉光1,王金刚1,陈朝达1,王全喜2(11 西安石油学院机械系化机研究所,陕西 西安710065;21 青海石油管理局格尔木管道输油处,青海 格尔木816000)摘要:为了寻找花-格管线腐蚀穿孔的原因,对西起柴达木盆地西部的油沙山,东至其东部格尔木,全长43517km的花-格管线运行以来的腐蚀穿孔次数进行统计分析和相关性计算,得出腐蚀穿孔累积数对数和运行时间的线性关系、腐蚀穿孔累积数和运行时间的指数关系.模型表明腐蚀穿孔失效呈加速态势,其腐蚀失效机理是管道防腐层下电屏蔽的电化学自催化腐蚀和管道保温层中干湿交界处氧浓差腐蚀.关键词:埋地管道;腐蚀;穿孔;数学模型中图分类号:TE973.91文献标识码:A花-格管线西起柴达木盆地西部的花土沟油沙山,东至盆地东部格尔木,全长435.7km,管径273mm,采用加热输送,出站温度7080,进站3040.沿线大部分地段为地下水位较低的戈壁荒漠,部分地段为地下水位较高的草原和沼泽,同时该管线的东部还穿越那陵格勒河、大灶火河和格尔木河,穿越河流的总长度近4.78km.沿线所经区域降雨量少、昼夜温差大、气侯干燥、风沙多,属典型的大陆高原性气侯.花-格管线采用的是“管中管”夹克防腐保温技术1,在管线外部有50mm厚的硬质聚氨酯泡沫作保温层,其外是厚4mm的黑色高密度聚乙烯作防水防腐蚀层,全线采用电溶套补口方式.但是花-格管线建成投入使用后,由管道腐蚀穿孔产生的原油泄漏事故逐年增加,不仅因维修影响输油生产,造成了严重的经济损失(维修费和漏油),而且还污染了环境.通过对花-格管线使用以来的腐蚀和维修资料的分析,得到了腐蚀穿孔统计数据,并由此建立腐蚀穿孔数学模型,对该管线腐蚀穿孔机理和影响因素进行了分析.1花-格管线腐蚀穿孔数学模型青海石油局的花-格管线于1990年建成投入运行,其历年的腐蚀穿孔统计数据见表1.表1花格管线腐蚀穿孔数统计表年份1991 1992 1993 1994 1995 1996 1997 1998运行时间xi?a12345678当年度穿孔数0213251518腐蚀穿孔是一个多因素的结果,直接采用每年度的穿孔数进行分析,数据波动大,随机误差大,在此采用统计分析的累积数法2,将年度穿孔数累加为该年度前整个运行时间的累积穿孔数,见表2.这样减少了数据波动,可建立更准确的数学模型.图1为花格管线腐蚀穿孔累积数yi对数和运行时间xi的关系曲线,可以看出二者符合线性关系.为得到两者间的统计关系,以最小二乘法建立腐蚀穿孔数量和运行时间之间的数学模型,并进行相关性和误差分析.设yi和xi数据样本的线性关系为:yi=axi+b令 Q=ni=1yi-(axi+b)2达到最小值(最小二乘法),yi和xi线性回归系数为:a=SxySxxb=y-ax(1)式中x=1nni=1xiy=1nni=1yiSxx=ni=1(xi-x)2Sxy=ni=1(xi-x)(yi-y)表2花格管线腐蚀穿孔分析表年份1992199319941995199619971998累积数平均值运行时间xi?a2345678355xi-x-3-2-10123(xi-x)2941014928累积穿孔数N2368132846累积穿孔数对数yi=1nN0.6931.0991.7922.0792.5653.3323.82915.3892.198yi-y-1.505-1.099-0.406-0.1190.3671.1341.631(yi-y)22.2651.2080.1650.0140.1351.2862.6607.733(xi-x)(yi-y)4.5152.1980.40600.3672.2684.89314.647注:1991年由于全线凝管事故、热水解堵,该年度不作为统计分析样本.由表2得=(xi-x)(yi-y)(xi-x)2=14.647?28=0.523b=y-ax=2.198-0.5235=-0.417运行时间xi与腐蚀穿孔累积数N的对数yi=1nN的线性关系为yi=axi+b=0.523xi-0.417(2)现场实际样本xi,yi的相关系数为rxy=SxySxxSyy=14.647287.733=0.995(3)式中 Syy=ni=1(yi-y)2=7.733此样本的rxy接近于1,故两个变量(腐蚀穿孔累积数N的对数yi、运行时间xi)符合线性关系.所回归的线性关系和实际现场样本的相对误差为QSyy=1-r2xy=0.92%(4)所回归的线性关系误差很小,可很好地描述花格管线现场腐蚀穿孔规律.将腐蚀穿孔累积数N代入方程(2),可得腐蚀穿孔累积数N和运行时间xi的数学模型:N=ebeax=0.659e0.523x(5)运行时间xi和腐蚀穿孔累积数N的关系,可用图1、图2表示.图1腐蚀穿孔累积数的对数和运行时间关系图2腐蚀穿孔累积数和运行时间的关系由图1、图2和式(2)、(5)知:花格管线腐蚀穿孔数学模型是指数模型,累积穿孔数有较快的增加,输油处可根据其预测以后的腐蚀穿孔数量制订相应的维修计划,更好地保证输油生产、降低维修成本.2花-格管线腐蚀穿孔原因和机理从花格管线的运行和维修情况知,该管线在施工时质量控制不严,夹克层机械损伤较多、补口质量差、聚氨脂保温层发泡有缺陷,加上花-格管线全线凝管修复3,这些均造成了防腐层进水点较多.根据历年维修、开挖检测记录,管外防腐保温保护层进水的管段总长度118.1km,占管线总长度435km的27.1%.保温层内进水或充水使金属与水直接接触,阴极保护措施等一度运行不是很正常,这些造成了花格管线腐蚀严重.由花格管线的腐蚀事故记录知,己产生的46处腐蚀穿孔事故中,发生在管外保温防腐保护层进水的管段有23处,占全部腐蚀穿孔数的一半.这说明17樊玉光等:花-格管线腐蚀穿孔失效模型分析腐蚀穿孔主要是由于管外壁金属与水接触后产生电化学腐蚀造成的,而未进水管段的土壤腐蚀穿孔发生率相对较低.管外腐蚀有几种情况,一是均匀腐蚀,它由均匀土壤腐蚀或保温层充水后的水腐蚀产生,这种腐蚀基本上为氧含量控制的电化学腐蚀,腐蚀速度不高,是大面积腐蚀,不会在短时间内产生腐蚀穿孔,不是产生腐蚀穿孔事故的主要原因,埋地管线的阴极保护可降低该类腐蚀的速度;二是由于腐蚀介质或钢管不均匀性产生的腐蚀,如干湿交替区域的氧浓差腐蚀,或材料及安装制造产生的不均匀性;三是由于防腐蚀层的电化学屏蔽作用产生的封闭体系自催化局部腐蚀4.后两种情况将会产生快速腐蚀,花格管线产生腐蚀穿孔的主要因素是:1)夹克进水不充满或水位变化使夹克保温层充水,存在干湿交界,在干湿交界处由于氧浓差会产生很强的局部腐蚀,当干湿交界固定时这种腐蚀就可能形成局部集中的腐蚀穿孔;2)防腐聚乙烯夹克层的小缺陷(如机械损伤等)导致较长保温层进水,由于防腐聚乙烯夹克层有很好的电绝缘特性,从而屏蔽了阴极保护电流进入金属腐蚀电极区,此时保温层内腐蚀产物和土壤水不溶物自然堆积将使防腐聚乙烯夹克层的缺陷(水、氧交换通道)堵塞,从而形成闭塞腐蚀区.闭塞腐蚀区初期的耗氧腐蚀产生的氢氧化铁液的电中性平衡使Cl2大量进入闭塞腐蚀区,产生高Cl2低pH值的酸性腐蚀区,从而会使管道快速腐蚀穿孔,这一点己有现场检修时保温层内的黑色酸性溶液得到证实.这两类腐蚀情况均难以通过阴极保护的改善而改善,1995年以后,花格管线阴极电化学保护得到了很好的运行,但是腐蚀穿孔数量仍然以指数规律大幅度增加也说明了这一点.3结论(1)花格管线腐蚀穿孔累积数对数和运行时间符合线性关系,腐蚀穿孔累积数N和运行时间xi是指数关系,穿孔速率随时间快速增长.(2)腐蚀穿孔的主要原因有两个:一是夹克进水不充满或水位变化使夹克保温层充水有干湿交界,二是由防腐蚀层的电化学屏敝作用产生的封闭体系自催化腐蚀区.(3)为预防此类腐蚀穿孔产生,必须进行定期检测,并可参考本文数学模型对将要产生的腐蚀穿孔数量进行预测,做好维修计划,保证输油生产.参考文献:1张俊义.埋地管道腐蚀规律初探J.油气储运,1991,10(3):30.2田中玄一.实验设计法M.魏锡禄,王世芳译.北京:机械工业出版社,1987.3冯洪臣.花格线腐蚀原因分析及处理J.油气储运,1994,13(5):52.4樊玉光,王金刚,高护生.管线涂层局部损伤时腐蚀及电化学保护A.石油科技理论与应用新进展C,西安:陕西科技出版社,1996.5汤兵勇.油田管线腐蚀更新数学模型的探讨J.石油学报,1993,14(3):118.编辑:田美娥(上接第69页)2M ichalew icz Z.Genetic algorithm s+data structures=evolutionprogram s J.IEEET rans on N euralN etworks,1994,5(1):96298.3Rudolph G.Convergence analysis of canonical geneticalgorithm s J.IEEE T rans on N eural N etworks,1994,5(1):962101.4 M ukhopadhyay B K,M alik O P.Opti mal control ofsynchronous machine excitation by quasi2linearisationJ.Proc.IEE 1972,119(1):91298.编辑:张新宝27西安石油学院学报(自然科学版)alsodecrease w iththeincrease ofshear rate.The watersolution displays the characteristics ofpseudoplastic fluid.Key words:ethylene oxide,propylene oxide,copolymer,viscosity,tri2isobutyl alum ium,lanthanumacetylacetoneL IU X iang,L I Q ian2d ing and YU H ong2J iang(Xian Petroleum Institute,Xian,Shaanxi 710065,China)JXA P I 2000 V.15 N.4p.5658Regression Analysis of the Correlation Between the Power Factor of the M otor and the Frequency of theFrequency convertor in an O il Transportation SystemAbstract:U surally,the speed of the motor used in an oil transportation system is adjustable,thefrequency of frequency convertor is also variable,and the power factor of the motor varies w ith thefrequency.So it is important to know the interrelationship between the power factor and the frequcncy.Inthis paper,the regressive curve of the interrelationship was obtained based on the data of the power factorfrom fivemotorsworking at five different frequencies separately.W hen the frequency is known,the powerfactor can be gained fromthe relationship.The power factor can be controlled under meeting therequirements of oil transportation system.Key words:variable2frequency adjustable2speed oil transportation system,frequency,power factor,regressive equation,prediction and controlCA O Yu2quan,LZuo2liang and L IA N G Ya2dong(Daqing Petroleum Institute,A nda,Heilongjiang151400,China)JXA P I 2000 V.15 N.4p.5962Singe-chip Computer Control System for Hydrometallurgical ExtractionAbstract:A method was put forward for controlling the experimental condition of hydrometallurgicalextraction.The signals of temperature,voltage,acidity and agitating speed are acquired by sensors,thenconverted into digital signals,and transferred to 8031 single2chip processor.A fter the digital signals arecompared w ith setting values,binary control signals are produced.This system can control four extractionconditions at the same time.The constitution and working principle of it are introduced in the paper.Key words:hydrometallurgy,single2chip computer,multiparameters control systemX ION G Guang2yu and H UO A i2qing(Xian PetroleumInstitute,Xian,Shaanxi 710065,China)JXA P I 2000 V.15 N.4p.6366Design of Neural Network D irect Inverse Controllers Based on Genetic Algorithm in Continous SpaceAbstract:A generalized design method of neural controllers was proposed for the acquisition oftraining data in the neural network direct inverse control.The genetic algorithm in continous space wasimproved in coding and genetic operators by using linear sequencing selection,arithmetic crossover andmultiplier mutation,which overcomes its premature convergence and make it posess high search precisionand efficiency.The control imputs of nonlinear dynam ic system s were optim ized by the modified geneticalgorithm.A neuro2controller was trained w ith the obtained desirable response trajectory.A synchronousmachine was used as a test2bed to demonstrate the effectiveness of the proposed design method,and thesimulation results were also given.Key words:neural network,genetic algorithm,coding,genetic operator,optimal controlSUNQ u,L I R en2hou and W A N G K ui2sheng(Xian Jiaotong U niversity,Xian,Shaanxi 710049,China)JXA P I 2000 V.15 N.4p.6769,72Buried Pipeline Corrosion Punching Fa ilureM odelAbstract:Corrosion punching hole is a main failure way of buried pipeline.A ccording to the statisticdata of the number of the corrosive holes in the buried pipeline from Youshashan to Geermu in ChaidamuBasin,it is known that the relationship is linear between the logarithm of the number and service time.The exponent relationship was also obtained between the number and the service time.It is held that thecorrosion punching hole is caused from the autocatalyzed electrochem ical corrosion of the electric shieldunder the anticorrosion layer of the pipeline,and the oxygen concentration corrosion of insulating layer atdry2wet boundary.The mechanism s of the two kinds of corrosion were analyzed.Key words:buried pipeline,corrosion,puncture,mathematicalmodelFA NYu2guang,W A N G J in2gang,CH ENZhao2da,et al(Xian PetroleumInstitute,Xian,Shaanxi 710065,China)JXA P I 2000 V.15 N.4p.7072