井下多元测试系统状态监测及其容错技术研究
详细信息    本馆镜像全文|  推荐本文 |  |   获取CNKI官网全文
摘要
井下多元测试系统,是石油开采中的重要配套设备,用于测量井下温度、压力、泵的出口流量、振动和泄漏电流等环境、工作参数,以实现采油装备运行情况的监测。井下多元测试系统能够完成其使命的前提,是其自身要具有较高的可靠性。然而,由于测试仪工作环境恶劣,其某些薄弱环节,尤其是传感器等,相比工作于常温、常压环境下的系统更容易损坏,且难以修复或更换。因此,对测试仪状态监测和容错的研究具有重要的理论意义和使用价值。
     作者设计了基于Windows的系统上位数据接收处理软件,概述了其具体的功能特点,并对软件各部分的程序设计作了详尽的说明。采用故障树分析的方法,并引入三角模糊数和中位数的概念,对系统可靠性进行了分析,实现了对难以获得精确的基本事件发生概率的故障树的量化分析和比较,找出了系统理论上的薄弱环节。
     为了提高系统可靠性,针对和温度参数密切相关的压力传感器,采用数据融合的方法,将温度参数融合进了压力值的计算公式当中。实验验证了该方法可明显提高压力传感器的温度适应性和测量精度。针对最重要的温度传感器,提出了基于双传感器备份冗余,并且结合了基于RBF神经网络时间序列预测理论的传感器状态监测模型,改进了K-均值聚类算法,使之在保证原有精度的前提下,减少了迭代次数,提高了运算速度,有效实现了井下测试传感器的状态监测和容错。针对传感器的故障类型识别,提出了一种根据故障传感器输出特性参数的变化,对故障类型进行分类的新方法。
     除此之外,本文还针对实际试验当中发现的影响系统可靠性的其余几个重要方面,如调制电阻电容的容错、抗干扰滤波、软件的容错等问题作了分析,并提出了作者的设想和建议。
Multi-sensor Down-hole Tool (MDT) is the important equipment used in oil extraction, which works underground and provides parameters such as temperature, pressure, discharge flux, vibration and current leakage, according to which the operator could monitor the oil extraction equipment. But the precondition of all above is that the MDT should be very reliable. However, the problem is that the MDT works in very poor condition, some weak parts of the system, especially the sensors etc. are more easily to be broken-down compare with those working in condition with normal temperature and pressure, and it is hard to repair or replace them. So, it's very significant and worthful to make research of the condition monitoring and fault tolerance of the MDT.
     The MDT software based on Windows system is designed, its function and characteristics are summarized and its program design is recited in detail. In order to analyze the reliability of the whole system and find out the weak parts in theory, the method based on Fault Tree Analysis (FTA) is used and the conception of Triangular Fuzzy Number (TFN) and the median are inducted.
     To enhance the reliability, for the pressure sensors which are related to the temperature parameter, the data fusion processing technology to calculate the pressure is applied. The experiment validates the ability of the method on enhancing its temperature adaptability and precision. For the temperature sensors, a model is presented, which is based on prepared redundancy of double sensors, combining the time series predicting theory based on the RBFNN. The K-means algorithm is improved, the alternate time is reduced and the calculate speed is enhanced while reserving the inhere precision of the K-means algorithm. It actualizes the condition monitoring and fault tolerance for the MDT. For the fault recognition of the sensors, a new way is presented to identify the fault byanalyzing the change of the sensors' characteristic parameters in the paper.
     Besides, this paper elucidates some other aspects discovered among the test that will affect the reliability of the MDT, such as fault tolerance of the resistances and capacitance for modulation, wave sieve and fault tolerance of the software etc. and presents the author's consideration and suggestion.
引文
[1] 王慧勋等.潜油电泵的原理及使用.北京:石油工业出版社,1994:1-64页
    [2] 葛光涛.油井参数遥测系统接收分系统设计.哈尔滨工程大学硕士学位论文.2005:1-3,8-12,67页
    [3] 郑青玉,刘刚,梁福平.电力载波在油田井下压力测量中的应用.北京机械工业学院学报.1997,12(1):56—61页
    [4] Moffatt T, Craig D. Innovative real-time well-monitoring systems (ESP Applications).Speprod Facil, 2001,16(3):151-155P
    [5] Anon. ESP monitoring and diagnosis. Journal of Petroleum Technology 2005,57(1):18—22P
    [6] Wenrong WU. New Nonlinear Algorithms for Estimating & Suppressing Narrowband Interference in DS Spread Spectrum Systems. IEEE Trans on Commun, 1996, Com-44(4):508-515P
    [7] 王仲生.智能故障诊断与容错控制.西安:西北工业大学出版社,2005
    [8] 孙瑞琛.自主式水下机器人传感器状态监测技术研究.哈尔滨工程大学硕士学位论文.2005:19—26页
    [9] 张铭钧,孙瑞琛,王玉甲.基于RBF神经网络的水下机器人传感器状态监测方法研究.哈尔滨工程大学学报.2005,26(6):726—731页
    [10] 房方,魏乐.传感器故障的神经网络信息融合诊断方法.传感技术学报.2000(4):272—276页
    [11] 李金伴,李捷辉.基于RBF神经网络的机器人故障诊断研究.自动化博览.1999,1:22—24页
    [12] 何佳洲等.基于新型神经网络分类器的故障诊断模型.计算机研究与发展.2001,38(1):93—97页
    [13] 沈毅,刘宜平,刘志言.模糊神经网络在多传感器故障监测与诊断中的应用.中国机械工程.2001,12(10):1176—1179页
    [14] 聂北刚,李初琴.传感器故障诊断实时监测系统设计的讨论.机械强度.2001,23(3):273—276页
    [15] 谷立臣,张优云,丘大谋.基于神经网络的传感器故障监测与诊断方法研究.西安交通大学学报.2002,36(9):959—962页
    [16] 王仲生.智能容错技术及应用.北京:国防工业出版社,2002:1-15页
    [17] 闻新,张洪钺,周露.控制系统的故障诊断和容错控制.北京:机械工业出版社,1998
    [18] Koren I, Pradhan D K. Yield and Performance Enhancement Through Redundancy in VLSI and WSI Multiprocessor Systems. Proc. IEEE, Vol. 74, No. 5, May 1986:699-711P
    [19] 张翰英.必须加速发展实用性的容错系统.控制工程.1986,4
    [20] 周东华等.以故障检测为中心的非线性动态系统的容错控制.控制理论与应用第四届学术年会论文集.1990:189-194页
    [21] 程一.线性多变量系统控制器和传感器的动态冗余性分析.信息与控制.1990,19(1):12—17页
    [22] 葛建华等.传感器失效故障系统的容错控制研究.浙江大学学报.1990,24(3):481-483页
    [23] Patton R J. Fault Tolerant Control: the 1997 Situation. Proc. Of IFAC/IMAC Symposium on Fault Detection, Supervision and Safety for Technical Process , Hull, England, 1997:1033-1055P
    [24] 葛建华,孙优贤.容错控制系统的分析与综合.杭州:浙江大学出版社,1994
    [25] 周东华,叶银忠.现代故障诊断与容错控制.北京:清华大学出版社,2000
    [26] Mehra R K, Peshon J. An Innovations Approach to Fault Detection and Diagnosis in Dynamic Systems. Automation. 1971:637-640P
    [27] Reiter R.A theory of diagnosis from first principles. Artificial Intelligence, 1987,32(1):57-96P
    [28] Poggio T, Girosi F. Network for Approximation and Learning, Proc. Of IEEE, 1990,78(9):1481-1497P
    [29] 龚建伟,熊光明.Visual C++/Turbo C串口通信编程实践.北京:电子工业出版社,2004
    [30] Peter W.Gofton著.王仲文,薛荣华,巴雪静译.精通串行通信.北京:电子工业出版社,1995
    [31] 隋鹏程,陈宝智.安全原理与事故预测.北京:冶金工业出版社,1988
    [32] 廖炯生.机器人的可靠性、维修性、安全性.北京:科学出版社,1994
    [33] 金伟娅,张康达.可靠性工程.北京:化学工业出版社,2005
    [34] 曾声奎等.系统可靠性设计分析教程.北京:北京航空航天大学出版社,2001
    [35] Ernst G. Frankel, Systems Reliability and Risk Analysis, Martinus Nijhoff Pub. 1984
    [36] James N. Siddall, Probilistic Engineering Design, Principles and Applications, Marcel Dekker Inc. 1983
    [37] 张慧慧,谢礼忠,李大寨.角膜移植显微手术机器人系统的安全分析.北京工业大学学报.2004,30(2):151-155页
    [38] 高峻蛲,龚建伟.基于逆向FTF的履带式移动机器人安全性分析.测控技术.2003,22(6):49~51页
    [39] Carreras C, Walker ID. Interval methods for fault-tree analysis in robotics [J] . IEEE Transactions on Reliability. 2001, 50(1):3-8P
    [40] 李锡江等.基于模糊故障树法的清洗机器人安全性研究.北京航空航天大学学报.2004,30(4):344—348页
    [41] 王永传,郁文贤,庄钊文.一种故障树模糊重要度分析的新方法.国防科技大学学报.1999,21(3):63—66页
    [42] 李青,陆廷金.模糊重要度分析方法的研究.模糊系统与数学.2000,14(1):89—93页
    [43] V.M.Stuchebnikov.钛-硅蓝宝石式压力传感器温度误差的整定.http://www.ru-cn.net/download/MIDA技术讨论/钛-硅蓝宝石式压力传感器温度误差的整定.doc 2006.10.18
    [44] Lawrence A. Klein. Sensor and Data Fusion Concepts and Applications. SPIE Press, 1993
    [45] 常炳国,张声良,刘君华.采用数据融合处理技术提高传感器的可靠性. 西安交通大学学报.1998,32(12):5—7,28页
    [46] 李利平等.传感器输出特性校正方法与多参数测量的研究.仪器仪表学报.1999,20(5):537—540页
    [47] 阳春华等.深海底集矿机传感器故障容错方法研究.计算机测量与控制.2005.13(9):1006-1009页
    [48] 翁桂荣,叶萍.基于RBF的传感器在线故障诊断和信号恢复.传感器技术.2003,22(10):50-53页
    [49] 钮永胜,赵新民,孙金玮.采用基于神经网络的时间序列预测器的传感器故障诊断新方法.仪器仪表学报.1998,19(40):383—388页.
    [50] 王让丁,梁正峰,陈华辉.基于联合神经网络的传感器故障诊断与重构.传感技术学报.2005,18(2):230-234页
    [51] 高隽.人工神经网络原理及仿真实例.北京:机械工业出版社,2003:55-75页
    [52] Simon Haykin著.神经网络原理.叶世伟,史忠植译.第二版.北京:机械工业出版社,2004:213—218页
    [53] 党开放,赵弘,林廷圻.电液伺服系统的模糊径向基函数网络监督控制.系统仿真学报.2004,16(3):593-596页
    [54] S. Chen. Nonlinear time series modeling and prediction using Gaussion RBF networks with enhanced clustering and RLS learning. ELECTRONICS LETTERS. 1995,31(2):117-I18P
    [55] 朱明星,张德龙.RBF网络基函数中心选取算法的研究.安徽大学学报.2000,24(1):72—78页
    [56] 李杰星,章云,符曦.一种改进的最近邻聚类学习算法.控制理论与应用.2000,17(5):735—738页
    [57] 孙延风,梁艳春,孟庆福.改进的神经网络最近邻聚类学习算法及其应用.吉林大学学报.2002,20(1):63—66页
    [58] 刘光斌,刘冬,姚志成.单片机系统实用抗干扰技术.北京:人民邮电出版社,2003:154—159,329—358页