After reading the entire book, the reader will come away with a thorough understanding of the rapid development of IoT-based systems and their impact on several scientific and engineering domains, including healthcare, smart homes, agriculture, robotics, industries, integration of leak detection in pipeline custody transfer of hydrocarbon products, and many others. Current trends and different architecture domains are explained systematically to motivate those in academia and industry to become familiar with the power of IoT. Also included are the heterogeneous used to enhance IoT security and an explanation of the framework of intelligent spaces needed for IoT-based optimized and secure ecosystem for the energy internet, handled by pervasive computing environment. The chapters thoroughly explain the transformation of business and ways of addressing its current needs, including how machine learning approaches play a greater role in achieving business intelligence in large commercial organizations, the role of big data analytics with the concept of automation, a roadmap for businesses to leverage big data analytics for creating business value implementing smartness in a smart environment where people are living and making an effort to develop it. Also, an analysis is conducted to examine how human and artificial intelligence might evolve together in future years and how it will impact humans with the help of business intelligence. Finally, this book portrays the difficulties experienced in business development consisting of a self-governing autonomous group setup with resources from both the IT and business advancement side of the organization.
On behalf of the entire editorial board, our heartfelt appreciation goes out to all the authors for considering us to publish their valuable work. Their overwhelming response has been a real factor in keeping us motivated and moving forward with this book, and therefore merits our sincere acknowledgement. The quality and diversity of their contributions have made the book more impactful, and their trust, patience and kind cooperation throughout the various stages of production played a vital role in its success. We also wish to thank the people at Scrivener Publishing for their guidance and support in bringing this edited collection to completion.
The Editors
November 2020
1
Applications of IIoT-Based Systems in Detection Leakage in Pipeline Custody Transfer of Hydrocarbon Products
Pragyadiya Das
National Institute of Technology, Trichy, India
Abstract
Custody transfer via pipelines has always been prone to losses due to leakages. Moreover, due to the large lengths of pipelines, leak detection is a tedious task. There are various methods that employ mechanical, mathematical and signal processing based approaches to detect leaks and their location. With the advent of Industrial Internet of Things and Machine Learning, the method of leak detection of pipelines using various machine learning methods have been analyzed and implemented in this chapter.
Keywords: Custody transfer, pipelines, leak detection, industrial Internet of Things, machine learning, ensemble learning
1.1 Introduction
The world of Oil and Gas has been moving at an alarming rate. The world is getting energy hungry [1] and the need for oil (or natural gas) is not going to go down anytime soon [2].
With this increased need for fuel, there is a constant need for the fuel (gasoline etc.) to be transported from one place to another. This gives rise to a need for a medium of transfer that would effectively transport the fuel in a secure and accountable manner.
Figure 1.1 Pipeline mapping—US.
As per data released in India’s official website (https://community.data.gov.in/), the total length of Natural Gas pipelines went from 10,246 to 17,753 km in the period of 31st March 2010 to 31st March 2017. That is a growth of 57% [3]. This is a significant growth and shows the need and increasing utility of pipeline in the energy scenario of an energy hungry nation like India [4].
Similarly, USA has a motor gasoline consumption of about 8,682 thousand barrels per day [4], and has a very significant crude and product pipeline, Figure 1.1 [5] explains this fact.
Looking at pipelines a major method of custody transfer of Gasoline, Diesel and other energy related fuels, it is important that there is proper monitoring of these structures to prevent any kind of adulteration, or more importantly, leakage causing financial losses.
With the advent of Wireless Sensor Networks (WSN) and Internet of Things (IoT), the monitoring of long distance pipelines have now become a task that can be achieved.
1.2 Industrial Internet of Things
The concept of “Internet of Things” has its core concepts set out as the interconnection of devices that have the capability to talk to themselves and “act” or take “decisions” based on each other’s statuses.
Usage of Industrial grade sensors to monitor industrial processes in real time and later achieving their interconnection is called Industrial Internet of Things.
Modern day refineries and pipelines consist of numerous numbers of sensors. The data that is generated is huge. This poses as a great opportunity to drive data analytics and Industrial Internet of things in this sector [6].
IIoT is the utilization of smart sensors (or actuators) to enhance manufacturing and industrial processes. The idea behind IIoT is intelligent machines are made better humans at capturing and analyzing data in real time, in addition they are also made better at prompting information that can be used to make decisions in lesser time and more effectively [7].
1.3 Pipeline Leaks
Pipelines are undoubtedly the most safe and reliable mode of custody transfer of fuel. Major reasons contributing to pipeline failure are depicted in Figure 1.2 [8].
The below detail explains the varied kind of reasons that are faced while analyzing pipeline leaks.
Figure 1.2 Percentage break-up of reasons causing pipeline leaks.
1.3.1 Various Techniques used to Detect Pipeline Leak
The methods that are usually used are divided into two classes,
1 Hardware techniques.
2 Software techniques.
We shall discuss in brief about three of the most common non-analytic and hardware-based technique used in pipeline leak detection. They are as follows:
1 a)