Study On Oil And Gas Pipeline Leak Detection Methods And Their Applications

The pipelines leak detection technique is a kind of important tool at the aspect of ensuring safe running of pipeline. This technique has typically gone through two stages of growth over the past twenty years. In the first stage, hardware based methods are studied and applied. During the second stage, software based methods begin to be studied. With the development of information technique and modern control theory, the leak detection methods based on the SCADA are widely used because of its advantages such as high performance-to-price ratio, easy to update and good expansibility. Especially in our country, the above technique has become leading measure which is used to detect leakage continuously. The research on pipeline leakage detection and localization is carried out in this dissertation. The main efforts accomplished can be summed up as the following aspects:1. The cross-correlation technique combined with the wave propagation model is analyzed. From this, the relationship between the correlation coefficient and leaks location is acquired. And this paper develops a residual correlation analytical model based the state estimate theory for detecting the leaks in oil pipeline. The field experimentations show that the effect is satisfactory.2. The multi-scale system theory is studied and the multi-scale transformation arithmetic for decomposing pipeline dynamic system and process technique is proposed. By use of multi-scale correlation analytical method, more leak information is preserved and high frequency noise is removed effectively.3. The principle of the hypothesis testing is studied and the difference between the likelihood ratio test and sequential probability ratio test is analyzed. The pipe signals are pre-processed base on normality test and stationary test before the improved sequential probability ratio test is used to detect the leakage. The shortcoming of the traditional localization method based on sequential probability ratio test is investigated. The improved localization method is put forward for the first time, which is based on the time diversity between pressure sensors in the beginning and end of the pipeline. Spot running has proved that this leak detection method based on SPRT can extremely diminish the occurrence of the false alarm.4. The transient pipeline flows model is studied. The characteristic method which is used to acquire the numeric solution of the model and ascertainment ofboundary conditions are analyzed. The transient model method base on the state estimate theory is investigated and its state transfer matrix and the dynamic noise covariance matrix are derived in detail. The computer simulated results show the validity of this method.5. The Hilbert-Huang Transform theory is studied and the signals decomposition method based on this theory is presented. According this method, signals are decomposed multi-component, called intrinsic mode functions. The signals decomposition effects is analyzed from three aspects which include the mean of the envelopes, stopping criterions, boundary conditions.6. The fuzzy cluster analysis method is used to classify the pipeline signals. The best-classified matrix and clusters center are obtained by the fuzzy C-mean dynamic cluster arithmetic. The cluster algorithm is used in the signals by the HHT transform. Three kinds of standard sample are constructed and new data are tested. The results prove the validity of the proposed method.