输送管道泄漏智能诊断

对于管道运输的经济安全运行、防范损失以及环境保护,流体输送管道安全监控是一项极为重要的任务。在管道系统的整体监控当中,泄漏检测与定位尤其占据着重要的地位。因而深入研究流体输送管道泄漏的智能诊断与定位方法,不但具有重要的理论意义,而且也具有巨大的实际应用价值。目前,管道泄漏检测定位技术的研究已引起了广泛关注,已经成为管道安全监控领域的热门研究课题之一。 本文系统地研究了流体输送管道的泄漏智能诊断与定位方法,主要研究工作如下: 1 针对具有先验知识的一般非线性系统,提出了一种新型的基于ANFIS非线性观测器的连续小波变换故障检测方法。首先利用系统输入—输出数据设计非线性系统的ANFIS输出观测器,然后根据实际系统与观测器输出设计出了残差产生器。利用小波变换分析残差,判断被监测系统是否发生故障。所设计的非线性观测器具有通用逼近性能,同时克服了现有方法中对噪声干扰信号突变特征点不易辨识的难点。该方法的适用对象范围相对较大。 2 为满足实时性需要,研究了基于负压波法的管道泄漏智能诊断与定位方法。提出了一种基于管道运行工况模式识别方法的泄漏实时检测策略。保证管道运行工况平稳,给出了执行器故障时容错控制器的设计方法。为提取真实信号,采用小波变换方法来消除采样噪声。为实现快速定位,采用负压波方法进行泄漏点定位。这一方法既获得了检测的实时性,又保证了定位的准确性,该方法同时也具有较高的灵敏度和稳健性。 3 如何提高泄漏定位精度一直是管道安全监控领域中热门的研究课题。为此,提出了一种基于双传感器压力梯度的管道泄漏定位方法,该方法克服了假设管道泄漏前后流体的摩擦阻力系数为常数的局限。对于难以建立精确动态数学模型的输油管道系统,这一方法仍然可以提高泄漏点的定位精度。通过与相关参考文献的比较,表明所提出的方法具有更好的优越性。 4 当前关于流体输送管道泄漏智能诊断与定位的研究大多使用单一方法,而采用多方法融合的策略进行研究却鲜有提及。受到多方法结合思路的启迪,我们提出了一种能够迅速、准确实现检测与定位的方法,即模糊决策管道泄漏诊断与定位方法。该方法利用模糊自适应推理,对于输油管道的各种不同程度的泄漏
电加热器
Monitoring of fluid transporting pipeline is an important task for economical and safe operation, loss prevention, and environmental protection. Leak detection and localization of fluid transporting pipeline play a key role, especially in the overall integrity monitoring for a pipeline system. Hence, the intensive study on intelligent diagnosis and localization approach of leak for fluid transporting pipeline is important on theory and application. At present, it has become a popular research focus, and has been one of the main topics in safe monitoring domain of pipeline.The problem of intelligent diagnosis and localization of leak for oil transporting pipeline has been studied systematically in this dissertation. The main research work is as follows:1 For the common nonlinear system with prior knowledge, a novel fault detection method based on ANFIS nonlinear observer using continuous wavelet transform is proposed. First we design the output observer of the plant by input-output data of the system. Then we design a residual generator. The residual, that is difference of system output and ANFIS observer output, is analyzed by the method of continuous wavelet transform. The result of analysis is used to alarm when a fault happen in the supervised system. The designed nonlinear observer has the performance of universal approximation. The method overcomes the difficulty of identification for sudden change of disturbed signals, and can be used in a great deal of practical systems.2 To satisfy the need for real time, the method based on negative pressure wave is studied. A kind of real time detection strategy, which is based on method of pattern identification of pipeline running status, is proposed. When actuator is in trouble, we give a design method of fault tolerant controller for pipeline’s normally running. To extract the true signal from its corrupted signal, the sample noise is eliminated by wavelet transformation. To locate leak quickly, negative pressure wave method is used. The characteristics of both real time detection and localization precision are all obtained. The proposed method is sensitive and robust.3 How to improve location precision is always the research hotspot in the domain of pipeline safe supervision at all times. A novel localization method based on pressure gradient using double sensors is proposed. The proposed method can overcome the major shortcoming of usual method, which assumes that the fluid