Natural gas pipeline leak detection based on data mining

This chapter is interested in discussing how to use data mining techniques to assist in achieving an acceptable level of quality of service of telecommunication systems. The quality of service is defined as the metrics which are predicated by using the data mining techniques, decision tree, association rules and neural networks. Routing algorithms can use this metric for optimal path selection which in turn will affect positively on the system performance. Also, in this chapter management axis using data mining techniques were handled, i.e., check the status of the telecommunication networks, role of data mining in obtaining optimal configuration, how to use data mining technique to assure high level of security for the telecommunication. The popularity of data mining in the telecommunications industry can be viewed as an extension of the use of expert systems in the telecommunications industry. These systems were developed to address the complexity associated with maintaining a huge network infrastructure and the need to maximize network reliability while minimizing labor costs (Liebowitz, J. 1988). The problem with these expert systems is that they are expensive to develop because it is both difficult and time consuming to elicit the requisite domain knowledge from experts.

Leakage diffusion characteristics and harmful boundary analysis of buried natural gas pipeline under multiple working conditions

Detecting leaks in natural gas pipelines using acoustic signals typically requires extensive prior knowledge and complex parameter designs, making it challenging to handle background noise and data distribution disparities simultaneously. This article proposes a dual-feature drift framework utilizing a nonparametric design approach for acoustic signal-based leak detection. This framework consists of two core technologies: first, feature backward normalization. Low-dimensional drift factors are designed based on transformed acoustic signals to exponentially normalize the time-periodic features of the signal feature matrix, thereby eliminating strong background noise. Second, constructing the feature drift layer within a one-dimensional convolutional neural network. Weighted parameters constrain a high-dimensional feature matrix, developing drift factors that perform exponential drift on each feature, thus enhancing gradient constraints and eliminating data distribution differences during model training. This framework achieves a fault identification accuracy of 95.46% for natural gas pipeline leaks, outperforming competing methods and representing a novel approach to intelligent pipeline leak detection.

The blending of hydrogen gas into natural gas pipelines is an effective way of achieving the goal of carbon neutrality. Due to the large differences in the calorific values of natural gas from different sources, the calorific value of natural gas after mixing with hydrogen may not meet the quality requirements of natural gas, and the quality of natural gas entering long-distance natural gas and urban gas pipelines also has different requirements. Therefore, it is necessary to study the effect of multiple gas sources and different pipe network types on the differences in the calorific values of natural gas following hydrogen admixing. In this regard, this study aimed to determine the quality requirements and proportions of hydrogen-mixed gas in natural gas pipelines at home and abroad, and systematically determined the quality requirements for natural gas entering both long-distance natural gas and urban gas pipelines in combination with national standards. Taking the real calorific values of the gas supply cycle of seven atmospheric sources as an example, the calorific and Wobbe Index values for different hydrogen admixture ratios in a one-year cycle were calculated. The results showed that under the requirement of natural gas interchangeability, there were great differences in the proportions of natural gas mixed with hydrogen from different gas sources. When determining the proportion of hydrogen mixed with natural gas, both the factors of different gas sources and the factors of the gas supply cycle should be considered.

The leakage of high-pressure natural gas pipeline can lead to serious accidents and pose a significant threat to people and buildings in the vicinity of the failure location. This paper examines the harm effect distances of different consequences based on harmful criteria. And possible consequences caused by pipeline leakage are assumed and their subsequent harm effect distances are computed using the software EFFECTS. Harm effect distances evaluation was performed considering jet fire and vapor cloud explosion (VCE) as the worst-case scenarios. Results show that for pipeline leakage, the jet fire produces longer harm effect distances, so the jet fire scenario should be considered as the limiter for harm effect distances evaluation. Based on the analysis of this paper, harm effect distances caused by the leakage of a high-pressure natural gas pipeline can be evaluated, which will not only be useful to prevent and control natural gas pipeline leakage accidents, but also provide a good reference to design of new constructed pipelines.

Pipelines are the safest and most economic transport systems for natural gas. In the long-term life of the pipeline, there will inevitably be a rupture, leak or damage caused by manmade events or by natural disasters. Leaks may generate dangerous clouds of gas which escape from high-pressure pipes and produce serious fires and explosions, endangering the lives and property of people in and around the area. Monitoring natural gas pipeline leaks will, in time, discover and locate these dangerous situations and reduce loss. The core contents of leak monitoring are the accurate location of leaks as well as rapid identification of different signal sources, thus reducing the occurrence of false alarms. There are a variety of approaches currently used by pipeline operators to detect leaks in natural gas pipelines. This paper describes a developed leak detection approach based on Real Time Transient Model (RTTM) for leak detecting in natural gas pipelines which has been used in Central Asia/China Gas Pipeline. The advantage of this developed approach is that a leak which occurs in all flow conditions (including operational transients) can be detected. This RTTM model provides a sensitive alarm threshold and takes a very short time to detect leaks in gas pipelines. It can identify and locate small leaks in seconds.

Alliance Pipeline operates an integrated Canadian and U.S. high-pressure, rich natural gas transmission pipeline system. Rich natural gas pipelines are unique in that the product transported in these pipelines contains greater amounts of higher molecular weight hydrocarbons than would be transported in a dry natural gas pipeline. The specifications for gas quality however are very similar and require the product to contain less than sixty five mg/m3 water, no free liquids and/or objectionable materials such as bacteria, ashphaltene, gum, etc. The acid gases, carbon dioxide and hydrogen sulphide, are also required to be below certain values (see Table 1). Corrosion is not expected to occur under these conditions due to the lack of free water available for the development of an electrochemical corrosion cell. However, there are instances where the gas quality may vary and this gas enters facility piping for short periods of time. A method has been developed by Pipeline Research Council International (PRCI) to determine the internal corrosion susceptibility for dry gas natural gas pipelines but there are currently no industry accepted models which determine the internal corrosion susceptibility for high energy natural gas (HENG) pipeline systems. Accordingly, it is important for operators of pipelines with high energy natural gas (HENG) to collect and analyze these off specification events and develop a method to determine the relative impact on internal corrosion susceptibility. It is perhaps more important for operators to use this method to develop a strategy to prioritize facility piping for inspection and confirm the absence of internal corrosion. An Internal Corrosion Susceptibility Assessment (ICSA) method has been developed for HENG which considers off specification water, carbon dioxide, and hydrogen sulphide contents in the HENG. The analysis has been enhanced to also consider low temperature operation and hydrocarbon dew-point variations. The model has been effectively trialed over the last number of years to prioritize inspections and has been further tested against PRCI research and models developed for dry gas internal corrosion susceptibility. All internal corrosion models need to identify free water as prime contributor to susceptibility, thus the subject model is considered adaptable to other gas pipeline systems. This paper discusses the methods used to develop the model, the challenges encountered and results of the field inspections conducted.

Data mining has been emerging recently as a viable computational tool for autonomous decision making especially in the field of medical applications. It has provided diagnostic solutions for skin and breast cancer detection, brain tumor detection, and also for other classification problems. In this chapter, we explore two data mining techniques, namely, association mining and decision tree mining, for predicting the life span of the kidney failure patients who have undergone routine dialysis. The total parameters used for this study were 28 attributes. The optimal prioritized parameters that decide the survival rate are reported and it can be concluded from the experimental results that the decision tree approach yields promising results.

Cardiovascular disease has increased rapidly in the past few decades. It has become a leading cause of death globally. Heart disease has affected the global heterogeneous population irrespective of age and gender. According to World Health Organization, an estimated 17.3 million people died from cardiovascular diseases in 2008, representing 30% of all global deaths. The accurate and timely prediction of these diseases has become a challenge for medical organizations. A mere assumption of absence or presence of disease is an approach used by many hospitals to give prediction results. The predictions of the heart disease are dependent mainly on the prominent factors involved and their effect weightage. Finding out the patterns and extracting knowledge from those patterns is the major task at hand. Data mining techniques have proven to be a good means for this knowledge discovery. This study makes use of the prominent features of two data mining techniques, namely, K-Means Clustering and Decision Tree. These methods, one being unsupervised learning and the other supervised learning, use very different approaches to predict the results. The positive factors of both the techniques have been used to build a Hybrid Classifier. The aim is to provide an algorithm which gives the best accuracy and performance for the Heart disease identification system.

Decision tree is an important method for both induction research and data mining, which is mainly used for model classification and prediction. ID3 algorithm is the most widely used algorithm in the decision tree so far. Through illustrating on the basic ideas of decision tree in data mining, in this paper, the shortcoming of ID3's inclining to choose attributes with many values is discussed, and then a new decision tree algorithm combining ID3 and Association Function(AF) is presented. The experiment results show that the proposed algorithm can overcome ID3's shortcoming effectively and get more reasonable and effective rules

The failure and leakage of the oilfield natural gas pipeline may lead to many serious sequences, such as the severe casualties and property losses. This paper analyzes the characteristics of the leakage of natural gas pipeline and the subsequent explosion accident of vapor cloud, adopts the TNT equivalent method to calculate the results of vapor cloud explosion of certain natural gas pipeline in Daqing Oilfield, and gets the relationship between the area of harm zone and leakage time as well as the harm radius corresponding to the different leakage time and area of discharge opening. In addition, it discusses the calculation results and gives the suggestions to reduce the loss resulting from the pipeline disaster.

Line leakage of oil and natural gas systems can result in significant financial and environmental consequences. Often, small leaks lead to ruptures in the pipeline that result in product escaping into the surrounding soil. Oil product can taint water supplies by migrating into aquifers, while natural gas leaks are susceptible to catastrophic explosions. A joint academic-industry research initiative funded by the U.S. Department of Transportation's Pipeline and Hazardous Materials Safety Administration (PHMSA) has lead to the development and refinement of a free-swimming device capable of detecting small leaks in oil product and natural gas pipelines. The SmartBall swims through a pipeline being assessed and produces results at significantly reduced cost to the end user compared to current leak detection methods. Geographic information systems (GIS) based above ground loggers that are global positioning systems (GPS) synchronized capture low frequency acoustic signatures and digitally log the passage of the SmartBall through a pipeline. This paper presents the development, laboratory and field validation testing of the SmartBall for both oil and natural gas applications with discussion of demonstrated case studies.

With the introduction of various carbon reduction policies around the world, hydrogen energy, as a kind of clean energy with zero carbon emission, has attracted much attention. The safe and economical transportation of hydrogen is of great significance to the development of hydrogen energy industries. Utilizing natural gas pipelines to transport hydrogen is considered to be an efficient and economical way. However, hydrogen has a higher risk of leakage due to its strong diffusion capacity and lower explosive limit than conventional natural gas. Therefore, it is of great significance to study the leakage and diffusion law of hydrogen-enriched natural gas (HENG) pipelines for the safe transportation of hydrogen energy. In this study, the leakage and diffusion characteristics of urban buried HENG pipelines are investigated numerically, and the dangerous degree of leakage is analyzed based on the time and area when the gas concentration reaches the lower explosive limit. The influences of hydrogen blending ratio (HBR), operating pressure, leakage hole size and direction, as well as soil type on the leakage and diffusion law of HENG are analyzed. Results show that the hydrogen mixing is not the key factor in increasing the degree of risk after gas leakage for urban buried HENG pipelines. When the HBR is 5%, 10%, 15% and 20%, the corresponding first dangerous time is 1053, 1041, 1019 and 998 s, respectively. This work is expected to provide a valuable reference for the safe operation and risk prevention of HENG pipelines in the future.

CO2 corrosion is one of the most common corrosion forms in the process of oil and gas field development and oil and gas gathering and transportation. In the process of natural gas gathering and transportation, the problem of pipeline corrosion caused by CO2 is widespread, which often leads to serious local thinning, even perforation and fracture of the pipe body. Once it occurs, it will cause great economic losses and disastrous consequences to the actual production. Therefore, it is of great practical significance to study the mechanism of CO2 corrosion and analyze the influencing factors of CO2 corrosion in order to prevent the safety accidents and improve the efficiency of natural gas transportation. In this paper, a natural gas pipeline wall perforation failure analysis as an example, the sample from the natural gas pipeline corrosion product morphology and composition of micro analysis, and the CO2 corrosion mechanism, influencing factors and other aspects are described in detail.

A vehicle-deployed parts-per-billion in volume (ppbv)-level off-axis integrated cavity output spectroscopic (OA-ICOS) CH4/C2H6 sensor system was experimentally presented for mobile inspection of natural gas leakage in urban areas. For the time-division-multiplexing-based dual-gas sensor system, an antivibration 35-cm-long optical cavity with an effective path length of ∼2510 m was fabricated with a high-stability temperature and pressure control design. An Allan deviation analysis yielded a minimum detection limit of 0.2 ppbv for CH4 detection and 10 ppbv for C2H6 detection for a 1 s averaging time. A natural gas leakage source location algorithm was proposed using an improved hybrid Nelder-Mead simplex search method and a particle swarm optimization (NM-PSO) algorithm. For field industrial application, the accuracy of the sensor system and leakage source location algorithm was confirmed through a CH4/C2H6 cylinder leakage experiment on the campus. Furthermore, through natural gas pipeline network inspection measurements in urban areas, three types of leakage sources, including natural gas, biogas, and possible leakage source were respectively located and confirmed using the global positioning system and wind speed and direction measurement system, verifying the reliability and potential application of the vehicle-deployed inspection system for future natural gas pipeline leakage monitoring.