IoT for Smart Operations in the Oil and Gas Industry: From Upstream to Downstream

Front Cover
IoT for Smart Operations in the Oil and Gas Industry
Copyright
Contents
Biography
Razin Farhan Hussain
Ali Mokhtari
Ali Ghalambor (Ph.D., P.E.)
Mohsen Amini Salehi (Ph.D.)
Preface
1 Introduction to smart O&G industry
1.1 Challenges of the O&G industry
1.2 Objectives of the smart O&G industry
1.3 Smart O&G: computing and middleware aspects
1.3.1 Landscape of computing infrastructure for O&G industry
1.3.2 Edge-to-Cloud continuum and O&G industry
1.3.2.1 Cloud computing
1.3.2.2 Edge-fog-cloud computing
1.4 Smart O&G: data and software aspects
1.4.1 Big data in the O&G industry
1.4.1.1 Big data in the O&G upstream
1.4.1.2 Big data in the O&G midstream
1.4.1.3 Big data in the O&G downstream
1.4.2 AI-based software systems in O&G
1.4.3 Digital twin: another data-driven applications in O&G
1.4.4 Edge-to-Cloud for AI and other data-driven applications in smart O&G
1.5 Roadmap of the book
2 Smart upstream sector
2.1 Introduction and overview
2.2 O&G exploration
2.2.1 Survey studies for potential reservoir
2.2.1.1 Magnetometry survey
2.2.1.2 Gravimetric survey
2.2.1.3 Seismic survey
2.2.1.4 Aerial photogrammetric survey
2.2.1.5 Well logging
2.2.1.6 Core sampling
2.2.1.7 Well testing
2.2.2 Smart O&G exploration: how can computing help?
2.2.2.1 Machine learning in geoscience for oil and gas exploration:
2.2.2.2 Using cloud data centers to archive exploration data
2.2.2.3 Automation in exploration data acquisition
2.2.2.4 Virtual reality in seismic imaging
2.3 Well production
2.3.1 Drilling operations
2.3.2 Well completion
2.3.3 Smart solutions in well production
2.3.3.1 Machine learning to predict drilling rate of penetration (ROP)
2.3.3.2 Predicting drilling fluid density
2.3.3.3 Drilling fluid optimal pressure estimation
2.3.3.4 Intelligent decisions to mitigate lost circulation
2.4 Smartness in the upstream sector
2.4.1 Overview of dataflow in upstream
2.4.2 Smart sensor data acquisition
2.4.2.1 Data acquisition in survey operations
2.4.3 Smart data preprocessing on the edge platforms
2.4.3.1 Data cleaning
2.4.3.2 Data transformation
2.4.3.3 Data conversion
2.4.4 Data analytics across Edge-to-Cloud continuum
2.4.4.1 Descriptive data analytics in the O&G industry
2.4.4.2 Predictive data analytics in the O&G industry
2.4.5 Synergy between High Performance Computing (HPC) and Cloud computing to provide real-time insights from simulations
2.4.6 Upstream as a cyber-physical system: detection to action framework
2.4.7 IoT-based monitoring and control solutions
2.4.7.1 Wired technology
2.4.7.2 Wireless technology
2.4.7.3 Supervisory control and data acquisition (SCADA) system
2.4.7.4 IoT in upstream of a smart O&G industry
2.4.7.5 Smart IoT applications in the O&G industry
2.4.7.6 Computing industry developing IoT solutions for the O&G
2.5 Summary
3 Smart midstream of O&G industry
3.1 Introduction and overview
3.1.1 The use of computation in midstream
3.2 Transportation of O&G
3.2.1 Various types of pipeline
3.2.2 Distance issue of transportation
3.2.2.1 Short distance transportation
3.2.2.2 Medium distance transportation
3.2.2.3 Long distance transportation
3.2.3 Challenges of transportation systems
3.2.4 Leakage detection systems in O&G transportation
3.3 Smart transportation
3.3.1 Smart vehicular transportation
3.3.1.1 Platoon coordination system
3.3.1.2 Vehicle tracking
3.3.1.3 Optimal routing
3.3.2 Smart pipelines for transportation
3.3.2.1 Pipeline design
3.3.2.2 Predictive maintenance
3.3.2.3 Smart pipeline monitoring
3.3.2.4 Equipment failure and leakage detection in pipeline system
3.3.2.5 Smart management system for long distance pipeline
3.3.2.6 Equipment failure detection
3.3.2.7 Case study: pump failure cause investigation
3.3.3 Shallow and deep learning methods for pump cavitation detection
3.3.3.1 Leakage detection in pipeline
3.3.3.2 Various computing solutions for pipeline leakage detection
3.4 Storage facilities for O&G
3.4.1 Types of atmospheric storage tanks
3.4.1.1 Open-top tanks (OTT)
3.4.1.2 Fixed-roof tanks
3.4.1.3 Floating-roof tanks
3.4.2 Smart O&G storage system
3.4.2.1 Overview
3.4.2.2 Atmospheric storage tank features
3.4.2.3 Fire protection
3.4.2.4 Smart fire fighting solution with fog computing and 5G
3.4.2.5 Pressure control system of AST
3.4.2.6 Remote monitoring to control pressure of AST
3.4.2.7 Inert gas blanket
3.4.2.8 Water drainage in AST
3.4.2.9 Drainage monitoring applications
3.4.3 Smart solution—automatic water drainage system
3.4.3.1 Hardware equipment for smart solution
3.4.3.2 Working principle of the smart solution
3.4.3.3 Advantages of using smart solution
3.4.3.4 Lightning protection of AST
3.4.3.5 A brief use case scenario of cloud computing in oil refinery
3.4.3.6 Smart solutions in oil storage tank maintenance system
3.5 Safety in midstream sector
3.5.1 Overview
3.5.2 Workers safety in midstream
3.5.2.1 The impact of toxic gas on workers health
3.5.2.2 Sources of toxic gas emission
3.5.2.3 Edge-to-cloud solutions for toxic gas detection
3.5.2.4 Motor vehicle accidents
3.5.2.5 Fire hazards
3.5.3 Environmental impacts of midstream
3.5.3.1 Methane emission
3.5.3.2 Carbon emission
3.5.3.3 Oil spill
3.6 Security aspects of midstream sector
3.6.1 Cyber-attacks on O&G midstream
3.6.1.1 A brief recent history of cyber-attacks against O&G industry
3.6.1.2 Cyber-attacks in various sectors of O&G
3.6.1.3 Cyber-attack target area—SCADA
3.6.1.4 Prevention from cyber-attacks
3.6.2 Physical threats of O&G
3.7 Summary
4 Smart downstream sector of O&G industry
4.1 Introduction and overview
4.1.1 Downstream taxonomy
4.2 Refining & processing of crude oil
4.2.1 Definition of O&G refinery
4.2.2 Different configurations of petroleum refineries
4.2.2.1 Topping and hydroskimming refineries
4.2.2.2 Catalytic conversion refineries
4.3 Petroleum refining processes
4.3.1 Separation
4.3.2 Conversion
4.3.3 Treating
4.4 Smartness in refinery units
4.4.1 Distillation unit
4.4.2 Smart distillation
4.4.2.1 Artificial intelligence method for smart distillation
4.4.2.2 Computing solutions for smart distillation
4.4.3 Cooling towers
4.4.3.1 Smart cooling tower systems
4.4.3.2 Smart solutions for cooling tower issues
4.4.3.3 Artificial intelligence & IoT in cooling towers
4.4.4 Smart boilers
4.4.4.1 IoT enabled boilers for refinery systems
4.4.4.2 Case study: Optimal design of heat exchanger using a combination of machine learning and computational fluid dynamics (CFD) methods
4.4.5 Smartness in distillate hydrotreater, fluid catalytic cracker and alkylation
4.4.5.1 Distillate hydrotreater
4.4.6 Fluid catalytic cracker (FCC)
4.4.7 Alkylation
4.4.8 Smart IoT based predictive maintenance
4.5 Distribution of end products
4.5.1 Blockchain based supply chain management
4.6 Safety of refinery facilities
4.6.1 Refinery process safety
4.6.2 Refinery equipment safety
4.6.3 Smart predictive maintenance for refinery equipment safety
4.6.4 Refinery workers safety
4.6.4.1 Chips in smart tech helmets
4.6.5 Environmental impacts of oil refineries
4.6.5.1 Air pollution
4.6.5.2 Water pollution
4.6.5.3 Soil fertility
4.7 Summary
5 Threats and side-effects of smart solutions in O&G industry
5.1 Introduction and overview
5.2 Taxonomy of cyber-threats and side-effects in the smart O&G industry
5.3 Vulnerabilities caused by the interplay of informational and operational technologies
5.4 Cyber threats in smart O&G industry
5.4.1 Vulnerabilities of sensitive data
5.4.2 Vulnerabilities of smart systems
5.4.3 Malware and vulnerability of information technology (IT)
5.4.3.1 Ransomware attack incidents
5.4.4 Vulnerabilities in operational technology (OT) protocols
5.4.5 Improving the security of the OT platforms
5.4.6 Data siphoning: vulnerabilities in IoT data transmission
5.4.7 Vulnerabilities of IoT devices
5.5 Incompatible IoT devices
5.5.1 Hardware-level incompatibility
5.5.2 Software-level incompatibility
5.5.3 Data pipeline incompatibility
5.6 Blockchain to overcome cyber-threats in smart O&G
5.6.1 Blockchain-based control systems (SCADA)
5.6.1.1 Consensus mechanism
5.6.1.2 Mining node selection
5.6.2 Blockchain to enable trust across industrial IoT
5.6.3 Blockchain for result verification in compute offloading
5.6.4 Aligning IT and OT to fill the security gaps
5.7 Risks of smart solutions in industrial IoT
5.7.1 Human-machine interaction issues
5.7.2 Machine-to-machine interaction issues
5.8 Bias in smart O&G industry
5.8.1 Biases caused by the artificial intelligence (AI) solutions
5.8.2 Automation bias
5.8.3 Other forms of (human-related) biases
5.8.3.1 Gender bias
5.8.3.2 Cognitive bias
5.9 Summary
6 Designing a disaster management system for smart oil fields
6.1 Introduction and overview
6.1.1 Smart oil fields
6.1.2 Challenges of the current smart oil field solutions
6.1.3 Contributions of current research works
6.2 System model for IoT- and edge-based task allocation
6.3 Robust resource allocation using federation of edge computing systems in remote smart oil fields
6.4 Performance evaluation of resource allocation method
6.4.1 Experimental setup
6.4.2 Baseline task assignment heuristics for load balancer
6.4.3 Experimental results
6.4.3.1 Analyzing the impact of oversubscription
6.4.3.2 Analyzing the impact of urgent tasks ratio
6.4.3.3 Analyzing communication overhead of edge federation
6.4.3.4 Analyzing average makespan of tasks
6.5 Summary
7 Case study I: Analysis of oil spill detection using deep neural networks
7.1 Introduction
7.2 Data acquisition
7.2.1 SAR images
7.3 Dataset overview
7.3.1 Challenges of oil spill datasets
7.4 Machine learning models
7.4.1 Semantic segmentation
7.4.2 Oil spill detection DNN model overview
7.4.3 Oil spill detection DNN model architecture
7.4.3.1 Encoder
7.4.3.2 Decoder
7.5 Performance evaluation of oil spill detection
7.5.1 Experimental setup
7.5.2 Preprocessing dataset for FCN-8s
7.5.2.1 Original dataset
7.5.2.2 Identify and labeling classes in dataset
7.5.2.3 Data preprocessing
7.5.3 Evaluation criteria
7.5.3.1 IoU
7.5.3.2 Pixel accuracy
7.5.4 Experimental results
7.5.5 Experiments with optimizers
7.5.5.1 SGD optimizer
7.5.5.2 Adadelta optimizer
7.5.5.3 Adamax optimizer
7.6 Summary
8 Case study II: Evaluating DNN applications in smart O&G industry
8.1 Introduction
8.2 DNN-based applications in O&G Industry 4.0
8.2.1 Fire detection (abbreviated as Fire)
8.2.2 Human activity recognition (abbreviated as HAR)
8.2.3 Oil spill detection (abbreviated as Oil)
8.2.4 Acoustic impedance estimation (abbreviated as AIE)
8.3 Inference datasets for DNN-based applications
8.3.1 Fire detection inference dataset
8.3.2 HAR inference dataset
8.3.3 Oil inference dataset
8.3.4 AIE inference dataset
8.4 Cloud computing platforms for Industry 4.0
8.4.1 Amazon cloud
8.4.2 Chameleon cloud
8.5 Performance modeling of inference time
8.5.1 Application-centric analysis of inference time
8.5.1.1 Overview
8.5.1.2 Statistical distribution of inference execution time
8.5.1.3 Shapiro-Wilk test to verify normality of the sampled data
8.5.1.4 Kolmogorov-Smirnov goodness of fit test
8.5.1.5 Analysis of central tendency and dispersion measures
8.5.2 Resource-centric analysis of inference time
8.5.2.1 Estimating confidence interval using Jackknife method
8.5.2.2 Estimating confidence interval using Bootstrap method
8.6 Summary and discussion
Bibliography
Index
Back Cover