LIFE2019 General Session Keynotes
Don’t forget to share your feedback by filling out the surveys in your mobile app.
Eric Carre, EVP Global Business Lines and Chief HSE Officer, Halliburton – Watch Now »
Anders Opedal, EVP Development & Production, Equinor – Watch Now »
Tom Keane, Corporate VP, Microsoft – Watch Now »
Tommy Sigmundstad, SVP, Drilling and Wells, AkerBP – Watch Now »
Mik Kersten, CEO and Co-Founder, Tasktop Technologies – Watch Now »
Nagaraj Srinivasan, SVP, Landmark and Halliburton Digital Solutions – Watch Now »
Qasem Al Kayoumi, SVP of the Technical Center, ADNOC – Watch Now »
Roberta L. Schwartz, EVP and Chief Innovation Officer of Houston Methodist Hospital – Watch Now »
Johan Krebbers, GM Digital Emerging Technologies / VP IT Innovation, Shell – Watch Now »
Panel Discussion – Integrating Data Science Into the Exploration and Production Lifecycle – Watch Now »
LIFE2019 Technical Presentations
Geological Context – The Missing Element in The Interpretation Domain
Development of Geofizyka Torun’s Innovations in Seismic Data Processing Using SeisSpace® Software
Velocities Are Geology – Interpretation-Based Velocity Analysis
Reservoir Characterization Improvement Through Integrated Approach to Capillary Pressure Curves Handling. Case Field: Llanos Orientales, Colombia
Global Patterns in the Geographic and Stratigraphic Distribution of Hydrocarbons
Based on our studies, along with updates in velocity model building with improved technologies and advancements in seismic survey design for better acquisition and enhanced imaging techniques, challenges and risks can be mitigated for successful exploration of the Triassic and Paleozoic reservoirs. Data was gathered from various sources (including well data, petrophysical logs, and seismic) to understand reservoir complexity and heterogeneity. Greater depth of occurrence and poor imaging of the Paleozoic sequences in seismic are main deterrents in pursuing potential leads. Recently, 3D seismic data were analyzed by integrating the data with drilled well data, using a seismic sequence stratigraphic approach, identifying Kuwait’s paleobasin architecture. Evaluation of 3D seismic data in West Kuwait showed the possible presence of carbonate reefs, talus, and grain flow deposits within the Permian Khuff formation, identifying potential features such as buried structures, alluvial fans, and channel-levy systems. Well data from deep wells drilled in Kuwait indicate that the Permian reservoirs may have gaseous hydrocarbon potential. Well logs from nine deep exploratory and development wells have been used to study petrophysical characteristics and their effect on the reservoir quality of the Paleozoic Khuff and Unayzah formations. Petrophysical log data have been calibrated with core analysis available at some intervals. Data indicate that the Triassic and Paleozoic formations show potential for gas reserves, and work is being done to mitigate challenges and unlock the formations’ potential. Improved acquisition, imaging, and velocity-model building are underway, and some of these involve using improved DecisionSpace® tools. Acquisition of advanced suite of PP logs suitable for high-pressure/high-temperature (HP/HT) wells are being planned. Improved drilling and completion are also being designed to enable successful drilling, completion, and testing in the future.
Landmark Earth® Engineered Appliance – On-Premise Private Cloud Platform for Geoscience Application Delivery
Seismic Multi-Attribute Realistic Co-Visualization Techniques – Multidimensional Analysis and Delineation of Complex Structure Tools
AVO Overview and Seismic Screening for a Quick Sweet Spot Evaluation – An Interpreter’s Perspective
Generating Complex Fault in a Strike-Slip Area and Its Impact On Prospect Evaluation, Kutei Basin, Indonesia
Integrated Geosciences Analysis In Mature Field to Improve Reservoir Management and Production
Well Engineering on the GO – Realtime
Well Design Concepts Accelerated with the Power Engine from Aker BP: Digital Well Program™
Drilling Software Evaluation and Directional Survey Simulation Project to Determine Accurate Survey Tool Code
A total of 2,426 wellbores has been entered in the EDM database, comprising 1,897 onshore wellbores and 529 offshore wellbores. In addition, a Survey Tool Code Bridging study (ISCWA and OWSG standards) between the T-DESK and COMPASS systems has been conducted and documented.
RTOC Transformation with iEnergy® Tenant, Using SmartDigital™ Co-innovation Service
EDM™ Insights – Improving User Proficiency and Performance in Real Time
Living on the Edge – Digital Challenges for Extreme Environments
The Performinator Project – A Rig Upgrade with Robotic Pipe Handling, Open Source, Autonomous Operations, Edge, Simulators, and Digital Twins
Well-Architecture-Driven Transverse Shear Stresses – Engineering Challenges in Horizontal Extended-Reach Wells
Applying DecisionSpace® Gun Barrel View for Easy Visualization and Interpretation of Complex Unconventional Well Planning Scenarios
Cuttings Image Classification in Real Time, Using Machine Learning Applications
Empowering Data-Driven Upstream on Azure
A 4D Small Data Solution in a Deepwater Gulf of Mexico Seismic-Driven History Matching Workflow
History Matching of Multiple Wells with Actual Downhole ICD Configuration in a Five-Spot Pattern Reservoir Development
Handling of Static and Dynamic Uncertainty in a Coupled Reservoir-Surface Model for the First Deepwater Field Development in Mexico
Multiple-Cloud, Multidisciplinary Workflows are Possible with Standards-Based Dataset Transfer Packaging
Rethinking Earth and Reservoir Modeling: Is the Path Forward White, Black, or Gray?
Managing Large Amounts of Capillary Pressure Data, Using DecisionSpace® Petrophysics Software to Generate a Comprehensive Saturation Height Model
Revitalization of Oil Production in Nezzazat Reservoir, Using Low-Cost Water Dump-Flood in Offshore Environment: Case Study from Gulf of Suez Petroleum
DecisionSpace® 365 Full-Scale Reservoir Simulation
Field Gas Lift Distributed Optimization with Automated Control for Holonic Systems
Large-Scale Subsurface and Surface Integrated Asset Modeling – Lessons Learned
Simplified Physics-Based Reservoir Models for Country-Wide Integrated Capacity Planning
Initiate Digital Oilfield Application at Mature Offshore Oil Field in South East Sumatera, Indonesia
Reaching Asset Potential – A Multiple Time Horizon Optimization Problem
Production Surveillance Digital Solution for Supporting a Mature Field Redeveloping Business Model in Eastern Ecuador Basins: IGAPO Consortium Case
Reducing the Time to Completion and Increasing Production, Utilizing Oracle Information Management Technologies
A Collaborative Solution for Integrated Water Management and Quality Surveillance in Production and Injection Operations – Ecopetrol Case
Intelligent Water-Alternating-Gas Process Using Downhole Control Valve (WAG-CV): Concepts, Tools and Simulations
Reservoir Simulators Performance Evaluation in the Gulf of Mexico
Hybrid Metamorphosis: New Data, Science, and Capabilities in Existing Systems Blur Technology Lines Between Fossils, Renewables, Cloud, and the Edge
The Cloud Experience at Eni, Using Geoscience Workflows
Petrobras Implements Web Applications to Meet Regulatory Changes in Well Integrity
Holistic Data Management Strategy for Large NOC’s
How Gyrodata’s Guide Center Has Increased Efficiency, Productivity, and Deliverables for Well Engineering Services
Fueling Digital Innovation with AWS Cloud-Native Services
Agile Database Approach to Non-Operated Assets
Digital Solution for Drilling and Well Services – An Example of How to Take Advantage of Your Data
BHP’s Next Generation Work Station Ready (WSR) Well Logs
Best Practices in Drilling Data Quality Management and Automation
Insights Into Hydraulic Fracture Geometries, Using Fracture Modeling Honoring Field Data Measurements and Post-Fracture Production
The Modern Geophysicist: Adding Value and Influencing Business Decisions
Reducing Uncertainty in Unconventional Reservoir Characterization
Unconventional Drilling in the New Mexico Delaware Basin
Applying DecisionSpace® Gun Barrel View for Easy Visualization and Interpretation of Complex Unconventional Well Planning Scenarios
Elastic Mechanical Properties from SEM and X-ray CT Imaging in Tight Formations
Integrating a Drilling Events Database, DecisionSpace® Geosciences Software, and Seismic Imaging to Predict Subsurface Drilling Events
Getting the Most Out of Your COMPASS™ Software
Pore Pressure Estimation from Velocity Model in Tight-Gas Sandstones in Argentina’s Neuquén Basin
Automated Unconventional Production Forecasting Based on SPEE Monograph 4 Guidelines
Misguided Approaches to Artificial Intelligence and Machine Learning Result in Suboptimal Value
Rapid Basin-Scale Geologic Reconnaissance Using Unsupervised Learning
Understanding the combined effects of multiple wireline logs, along with depositional environments and lithologies, are key steps in developing a comprehensive understanding of unconventional reservoirs. However, to characterize reservoirs on a basin scale, it is necessary to efficiently analyze and integrate large wireline log data sets. Current software packages offer very limited basic unsupervised learning algorithms that can be used to quickly understand the natural separation of data sets. This time-consuming challenge is compounded by significant heterogeneities that exist within a basin, along with issues associated with data integrity. This presentation will highlight the possibility of analyzing tens of thousands of well logs, using an automated workflow that involves quality control (QC), statistical analyses, and unsupervised learning analyses, with the intention of rapidly identifying log-scale rock types and core-scale lithofacies. To further capture the varying degrees of heterogeneity observed within each of the target formations, clustering algorithms for each formation are chosen based on key performance indicators (KPIs). This integrated workflow allows us to rapidly perform high-resolution reconnaissance of the geology on a basin scale, and to assist in geological mapping and modeling, and petrophysical analyses. The results of this automated workflow highlight laterally continuous clusters that reflect the different paleoenvironments of the Delaware Basin – from the margins of the carbonate platform to the proximal part of the basin. Contrary to existing workflows using unsupervised learning, results can be quantified using KPIs (i.e., gap statistics). Using a similar workflow on the core data, a lithofacies scheme for the Permian Basin was rapidly established for each of the key formations. Benchmarking results against historical regional studies indicated that this workflow was able to quickly establish high-quality reservoir characteristics on a basin scale.
ESP Failure Prediction
Smarter Well Engineering Concepts Aid in Reducing Planning Time and Increasing ROP
Modern well engineers struggle with digital confusion; they either have too much data, or not enough, and the quality of the data is often questionable. This presentation will illustrate how the rate of penetration (ROP) can be optimized in any given field with an automated and time-saving process for designing wells, using machine learning (ML) techniques. We will discuss applying ML algorithms to predict optimized ROP for a prospect well. This model was trained with historical datasets from offset wells for two primary fields in Wintershall DEA North Sea operations, and the detailed workflow included data analysis for quality and completeness, statistical regression, and optimization. Results of the ML computations to an actual dataset of a well drilled were compared, and a high correlation was observed. The optimized results showed a 20% to 30% improvement in ROP from the benchmark well. This concept can be further developed to recommend optimum drilling parameters while drilling in real time in the absence of drilling engineers’ interpretations. A real-time drill-off test was validated on these datasets for optimizing the ROP based on weight on bit (WOB), flow rate, and total downhole RPM. Using a simple ML model and real-time feedback loop, drillers can continuously optimize based on real-time and historical data feed without the need to stop and perform a drill-off test. By prescribing optimized ROPs through automated ML of offset well attributes, engineers can push technical limits. Automated analysis, regression, and visualization of high-volume data can significantly reduce planning time and help establish optimized operational parameters to reduce drilling time and costs. The next step is to build a real-time downhole advisory system to help achieve the predicted ROPs by predicting and prescribing drilling parameters ahead of the bit. With this process, drilling engineers are able to do more value-added engineering and less data mining and management, and to ultimately drive efficiencies throughout the well construction process.
Benchmarking – Clustering Wells in the Construction of Offshore Wells
The implementation of benchmarking, as an instrument of business management, aims to verify the processes adopted, comparing them to each other, in order to increase the efficiency of operations. Applying it to the oil industry, the main challenge is to properly group the wells, so that their similar characteristics can be compared properly. Usually, the groups are divided based on the experience of the technicians, but this is completely error prone due to manual intervention – thus, triggering the need for generating an algorithm that can create the grouping objectively, considering all the parameters available. Initially, according to the method developed jointly by Petrobras and Halliburton, the wells were divided into two large groups: pre-salt and post-salt. Then, within each of these groups, artificial intelligence techniques were used to group wells with greater similarity, using clustering algorithms on variables obtained from historical data, such as drilled interval, water depth, fluid weight, start and end true vertical depth (TVD), and pre-salt and post-salt, among other variables. Given this cluster, new wells can be allocated through the application of classification algorithms. Our experiments indicated a difference of up to 60 percent in the duration of wells belonging to different clusters. In this way, clustering can be applied directly to decision-making processes, providing support for monitoring all the operations involved in well construction. Also, automatic clustering and allocation of the wells can mitigate distortions generated in a prevalent process, since it reduces the group subjectivity. Finally, in the future, we plan to extend this approach to completion and workover interventions, covering the entire life cycle of a well construction.
Building Human Superpowers With Machine Intelligence & Software UX
Human actors working inside digital systems during both exploration and production (E&P) cycles require unique forms of situational awareness and contextual interaction across the spectrum of data, information, and insights to take on the appropriate amount of cognitive load for decision and action. When bridging physical and digital workflows across production operations, human intelligence and muscle memory still provide critical checks and balances at scale. Machine learning cannot (yet) replace human intuition and experience. The explosive proliferation of business data and its supporting technologies can easily swamp traditional methods of processing and analysis, resulting in enormous expenditures in time and effort with potentially little gain. Methods of artificial intelligence, machine learning, natural language processing, and computer vision have developed at pace with the data explosion, and offer an exceptional set of tools for extracting actionable intelligence and business value in a more efficient and automated manner. However, like any complex set of tools, their proper use and potential misuse can be highly nuanced. Embracing this new role of data and this new toolset as enablers for highly intelligent user experiences provides tremendous opportunities. Valuable feedback loops for software user experience are created in the product and service development life cycle when data is understood and managed as both an output and an input. Workflows become more contextualized and operations are made more efficient when human intelligence becomes empowered and augmented with artificial intelligence, rather than completely replaced by it in full-automation scenarios.
A Multidisciplinary Approach for Dynamic Rock Typing Characterization Using Artificial Neural Network Methods in Hamra Quartzites Reservoir
The previous geological model of the quartzites Hamra (QH) reservoir failed to accurately classify and estimate the reservoir’s petrophysical properties. The QH reservoir comprises five electrofacies (EFs): EF1, EF2, and EF3 are similar to clean sandstones with different rock qualities (porous to compact sandstone), while EF4 is shaly sandstone and EF5 is shale. In this work, we suggested a multidisciplinary workflow to recharacterize the QH formation, using artificial neural network (ANN) methods. The main propose was to delineate EFs present in the QH reservoir by using an unsupervised ANN self-organizing map (SOM) clustering algorithm, and also to determine the hydraulic flow unit (HFU) and permeability in uncored wells by using a supervised neural network (NN) algorithm and a flow zone indicator (FZI) approach. Our results concerning the HFU indicated complete agreement with the geological and sedimentary control. Also, the FZI approach suggested the presence of eight HFUs in the QH reservoir. The best ones (such as HFU1, HFU2, and HFU3) were mainly located in zones QH2 and QH4. In this approach, our results also indicated complete agreement with the production testing results. Finally, permeability determined by NN is correlated to core permeability with good correlation coefficient (0.6), which indicated the conformity of our proposed approach.
Water Coning From Simulation to Neural Network – Comprehensive Study of Coning Prediction and Factors Affecting It
Producing undesirable phases, like water and free gas in oil wells, is a challenging problem in the oil industry. The main major reason for that problem is water coning, which is a rate-sensitive phenomenon generally associated with high producing rates. Strictly, a near-wellbore phenomenon, it develops when pressure forces draw fluids toward the wellbore to overcome the natural buoyancy forces that segregate gas and water from oil. This work uses Nexus® simulation software to build different mechanistic models with different parameters that affect water coning formation in oil reservoirs. Simulating water coning is a very challenging problem due to the instabilities of solvers regarding the severe saturation change around the wellbore, unless small time steps and small grid sizes were used. Local grid refinement (LGR) is used to accurately follow up water coning formation and to minimize solver convergent errors. Simulations were used to quantify the effect of every parameter on water progress to form coning around a wellbore. In this regard, a neural network was built using input and output parameters from the simulation runs to devise a simple approach for evaluating the critical rate of production and how the uncertainty in every parameter would affect the recovery and coning formation.
Digital FPSO – Water Injection System Digital Twin Using Machine Learning
This presentation will focus on how digital solutions can be orchestrated to improve the operational efficiency of the water injection system of an offshore production unit operating in the Brazilian pre-salt area. Problems related to the seawater injection pumps and sulfate removal units (SRUs) have been impacted by the water injection system availability on a daily basis, which, in the long run, can have a major impact on the reservoir material balance, the field’s life cycle, and cash flow. Based on machine learning techniques, a digital representation of the water injection system was modeled using approximately 180 real-time process variables with 2 seconds of frequency, as well as flat files of daily operation reports, chemical injections, and injection losses for two years of operation. The solution combines machine learning models running in an orchestrated way to provide alerts and warnings for any underperformance of the SRU and main injection pumps. SRUs analyze real-time operational parameters, predicting the next stoppage due to high saturation of the SRU membranes, and specifying the causation of the performance decrease. Seawater injection pumps analyze real-time operational pump parameters and predict different vibration behaviors, as well as horsepower levels for the coming hours. Based on real-time data or forecasted vibration parameters, a second model can predict if the pumps are operating normally or if they are in a pre-fail or failure zone. Our results illustrated: 1) earlier prediction of pump failures by the combination of vibration forecast and an operational condition classification algorithm, making it possible to predict failures within 15 days in advance on average, and resulting in an estimated yearly return of USD 650,000; 2) the data-driven model provided early identification of pump performance degradation, allowing for the pump’s condition to be evaluated and for maintenance to be scheduled; and the SRU model predicted the next stoppage (due to high saturation of the SRU membranes) with 82 percent accuracy.
Improving Machine Learning Workflow and Business Value for Oil and Gas Applications
Machine learning (supervised and unsupervised) and deep-learning techniques have been widely used to solve a variety of oil and gas problems. However, a few characteristics limit its success, such as data availability and quality, the heavy requirement of data preprocessing and feature engineering, the lack of correct labels, and the complicated business value not in pre-existed mathematical optimization. So, we applied two deep-learning approaches, including supervised and unsupervised, to oil and gas business cases. The first approach is a hybrid approach combining a fully convolutional network and a long short-term memory network (LSTM) with an attention mechanism. Temporal convolutions capture the local variations in the data, whereas LSTMs capture the long-term variations in the data. Fine-tuning of the model involves developing customized metrics based on business need. The second approach is utilizing an autoencoder, a deep-learning unsupervised method, for anomaly detection. The key idea is to train a model that encodes high-dimensional data to a low-dimensional space (latent-space representation) and then decodes them back to the original input. In comparison to a supervised machine learning technique for anomaly detection that can only detect a fixed set of features, an autoencoder is unsupervised and useful properties can be learned in latent-space representation. The model parameters are optimized to reduce reconstruction error. If the features within the input data are independent of each other, the decoding could be difficult, leading to reconstruction error. However, if there is a norm within the input data, this norm could be learned automatically by an autoencoder and leveraged when decoding the data back to the original input. When this approach was tested on equipment, the result was positive, presenting a difference in reconstruction error between data from the beginning and end of the equipment’s life span. Methods above were applied to different projects, resulting in substantially reduced turnaround time and improved business value.
Nature (Sometimes) Knows Best
The natural world has served as a rich source of design inspiration throughout history. Design aesthetics and functionality have both been enhanced by imitating nature – a practice referred to as biomimicry. The appeal of biomimicry is not difficult to appreciate. Some proponents maintain that nature serves as an excellent laboratory and incubator where designs are tested and refined over many years. The durability and commercial success of products as diverse as wind turbine blades and Velcro bolster this argument. Moreover, the promise of biomimicry is not limited to the materials domain. Scientists, engineers, programmers, economists, and more are becoming increasingly interested in how human processes/systems can be improved by imitating nature. Given the complexity of some of the system-level challenges that we as a species are facing, the optimism underwriting this trend is understandable. However, natural processes have often evolved as the result of different pressures than we humans are facing, and, consequently, they serve different purposes. Imitating nature without understanding contextually significant details risks not only failing to solve the problem at hand, but also exacerbating it. In this talk, two examples of process biomimicry are discussed with an eye toward identifying general design heuristics that can be used to increase the likelihood of success. It is appropriate to continue to look to nature for design inspiration, so long as this is done from an informed, critical perspective.
Artificial Intelligence (AI) or Intelligence Augmentation (IA)? How to Produce Not Only Better Engineering Designs, But Also Better Designers
Artificial intelligence (AI) systems – and, in particular, machine learning (ML) algorithms – seem to be taking roles we thought were reserved for specialists, spanning medicine to engineering. Maybe Warren Bennis, the famed University of Southern California Marshall School of Business professor was right about the “factory of the future.” He claimed, “The factory of the future will have only two employees, a man and a dog. The man will be there to feed the dog. The dog will be there to keep the man from touching the equipment.” However, now there is a push in the other direction – namely, that AI and ML systems do not replace the need for specialists, but rather augment their skills, especially when dealing with uncertainty. Researchers, analysts, and decision makers are not only interested in understanding their data, but also in understanding the uncertainty present in the data. Quantification, communication, and interpretation of uncertainty are all necessary for the understanding and control of the impact of variability. These three things – quantification, communication, and interpretation of uncertainty – help add both understanding and robustness to the design process. In this talk, we will present recent work done at the University of Utah on the topic of “materials by design” – our attempts to engineer for specific purposes and operating conditions. We will focus on how AI/ML combined with visualization can aid in not only better designs, but also in producing better designers.
Enabling Creativity and Agility in the Workplace
Organizations in every industry are scrambling to respond to the rapid technological advancements and changing consumer expectations that are challenging their relevance and ability to deliver innovative solutions. From startups reinventing or creating new markets seemly overnight, to increasing expectations to deliver responsible and relevant products and experiences, enterprises are looking both inward and outward for a path forward. It is essential across industries to unlock new ways of working so that organizations can bring innovative products and services to market and stay relevant with their customers and employees. We have helped many of our clients reinvent the way they work by embracing creativity and collaboration as the foundation of their future. This takes a commitment to an aligned purpose, and to understanding your staff and empowering them to work in the way they know best. It means giving them the right tools and the right space. It also means a shift in language, behavior, and the way we measure success. We must think holistically and design for a future that we are creating. We will share the work with companies that are leaders in their industries – in particular, by putting the focus on creativity, not innovation. Participants will come away with the perspective necessary to champion and lead these efforts within their own organizations.
Robots as a Service: See How Automated Drones Can Monitor Operations in Real Time
The emerging technology of the Robots as a Service business model has the potential to revolutionize the way companies operate. Airobotics has selected Houston to launch this world-first initiative with its “robot” (the Optimus drone), Airbase docking station, and artificial intelligence (AI)-enabled analytics. Unlike traditional drone programs, which rely on drone pilots, Airobotics’ fully automated system has the unique capability of swapping its own batteries and payloads, using a robotic arm. By combining the automated nature of aerial data collection and in-house data processing algorithms, Airobotics’ fully automated drone system enables the delivery of customer insights within hours. It no longer takes days or weeks between data collection processing and final delivery of insights. Insights derived from an automated platform can offer ways to optimize operations and maintenance of physical assets, systems, and processes of the site’s day-to-day operations – from inspecting key assets and infrastructure to increasing safety, security, and compliance through real-time video data. One key example of this has been with a major operator in Australia, where our solution has become an integral part of the daily mining process. In fact, an entire Mining Playbook has been created that outlines key data to be captured and monitored, as well as safety and training exercises that are monitored through live video feed from the drone. Efficiencies, safety protocols, and improved processes are resulting from this being incorporated into the “business as usual” model for Area C. For all the talk of automation and robots replacing human jobs, that is hardly the case with Airobotics’ technology. Rather, the new uses of an automated, pilotless system can remove dangerous, time-consuming tasks to move workers upstream. As a result, companies can cut costs while dramatically reducing risk, increasing efficiency, and even saving lives.
Counting Alligators and Catching Crooks: Stories from a Life Misspent in the Swamp of Data
Data surrounds us. We count, we measure, and we create databases to help describe, analyze, and interpret our world, and to predict what we can’t see: the future, and things hidden by design or obscured by nature. Though data science tools are amenable to automated recipes, many data science problems still require the spice of creativity and imagination. This talk looks at the magic of data science through four very different projects from the files of Mo Srivastava, a geostatistician who has spent a career pondering data, mostly from earth sciences, but sometimes from things found in a corner grocery store. In 2011, Mo was written up in Wired for breaking an instant scratch lottery. Pattern recognition and a data transformation common in geostatistics led him to a trick for separating winners from losers without scratching anything off. After teaching the trick to his young daughter, Mo reported the problem to the lottery corporation, which pulled the game off the market. In the 1990s, Canada’s nuclear waste program recognized the need for earth modeling tools that supported studies of flow and contaminant transport in a way that was not only technically sound but that would also inspire confidence and comfort in public hearings. Mo developed a new method for modeling fracture networks that honored all available data with a high degree of visual realism. When the State of Florida wanted consulting advice on its census of Everglades alligators, Mo signed on – partly because he had developed methods for leveraging secondary information to model things that are hard to see (like healthy alligators) … but also because he wanted to blast around in a giant fan boat. Mo’s final example is based on the work of Kim Rossmo, the first beat cop in Vancouver to earn a PhD in criminology. Rossmo’s “geographic profiling” toolkit assists in investigations of serial crimes. Mo adapted a data declustering method from geostatistics to enhance a tool that aims to identify something that serial criminals want to keep secret: where they live.
How Accenture ‘Wise Pivoted’ to the Future
The oil and gas industry has survived the largest downturn in history and faces an uncertain future going forward. There are lessons from other industries and companies that can provide a roadmap for energy companies as they plan for the future. Accenture faced a series of existential threats in the first decade of the millennium. We live in an age that demands companies stay in a permanent state of change. The digital age calls for a new approach to organizational change that enables companies to make a wise pivot successfully. This approach requires companies to:
- TRANSFORM THE CORE BUSINESS … to drive up investment capacity
- GROW THE CORE BUSINESS … to sustain the fuel for growth
- SCALE NEW BUSINESS … to identify and scale new growth areas at pace
This session will focus on Accenture’s “wise pivot” transformation, some of the key initiatives and imperatives to support the transformation, and what others can learn from our own experience.
SmartSuit for Extravehicular Activity: Current Challenges and Spacesuit Technology Development to Enable Future Planetary Exploration Missions
Extravehicular activity (EVA) is one of the most challenging activities that astronauts need to perform in space. NASA’s current gas-pressurized spacesuit, the Extravehicular Mobility Unit (EMU) is used in a microgravity environment, and it has not been designed to operate in different conditions such as planetary surfaces. Additionally, gas-pressurized spacesuits require astronauts to use their strength to move the suit, which can be fatiguing and can significantly affect the metabolic cost of human movement and locomotion in space. In particular, the current EMU spacesuit causes many astronauts musculoskeletal injuries and discomfort, which could lead to suboptimal EVA performance and could negatively impact mission success. In this talk, we will review the challenges of current spacesuits, and will introduce future concepts and technologies for upcoming planetary missions. Particularly, we will introduce the SmartSuit, a new spacesuit concept that our lab is investigating in collaboration with Prof. Robert Shepherd at Cornell University. The SmartSuit concept incorporates soft robotics technology in order to provide better mobility, and a soft and stretchable self-healing skin with optoelectronic sensors for enhanced safety and interaction with the astronauts’ surroundings. We expect this novel spacesuit intelligent architecture for EVA operations to increase human performance by an order of magnitude on several quantifiable fronts for exploration missions on Mars and other planetary environments.
Automated Artificial Intelligence – The Next Era of Data Science
The value of artificial intelligence (AI) is widely recognized, but most organizations struggle with finding data scientists and getting data science models deployed. DataRobot addresses these issues through automation. Automation allows organizations to improve their data science productivity. Automation also allows for a wider group of people to build data science models. To demonstrate these ideas, this presentation will cover how organizations can leverage automation to build models on a larger scale (thousands of models), while also achieving improved accuracy and rapid deployment. Automation also means less of a reliance on hand-coding, which opens analytics to more than just data scientists. This talk will share how DataRobot’s automation is enabling organizations like the Global Water Challenge to address needs such as clean water in Liberia and Sierra Leone.
From the Edge to the Cloud – A Single Data Platform
Cruising on the edge makes data management a challenge. In this talk, we will look at the challenges faced by the cruise industry with data collection in remote locations and synchronizing to the extreme distributed system. Learn how a unified data platform and Data as a Service (DaaS) strategy solve these challenges to enable digital transformation and drive business value.
An Introduction to the OpenEarth® Community and its Offerings
Learn Application Development to Consumption (Production Deployment) in 30 minutes
Geosciences – Easy Steps to Seismic Attribute Calculations Using OEC
Learn How Integration Server is Making Life Easy to Consume E&P Data (Includes Python Hands-on)
Building Web Applications Using Web Framework Available in OEC And Building Workflow Orchestration Using DecisionSpace® Business Process Management Engine available in OEC
Using Digital Well Program™ Solution to Customize the Well Design Process
DecisionSpace® 365 Enterprise Services Software – Modern Approach for Service Lifecycle
OpenEarth® Community – Enabling Development via DevOps and Collaboration
Integrated Development Platform for E&P
Assisted Lithology Interpretation – Turning A Machine Learning Scientific Algorithm into Product Using OEC and DSIS
DecisionSpace® 365 Seismic Engine Software – How to Build A Highly Scalable Seismic Processing Application
A Collaborative Analysis Solution for Integrated Data from Drilling Operations, Daily Reports, and Real Time
Creating Scalable Data Science Applications in OEC
Design Thinking for E&P Software Innovation
How OEC Can Help With Project Management and Development
Delivering Innovation at Scale – Getting Ideas to Value in the Oil and Gas Space
Optimized Production Solutions – DevOps Practices
Agile Data Science with Red Hat OpenShift Container Platform
New Science and Innovation Breakthroughs in IoT Prognostics
Building Open Platforms in the Kubernetes Renaissance
AI-Powered Geosteering-Based Drilling Optimization with Reinforcement Learning
Field Gas Lift Distributed Optimization with Automated Control for Holonic Systems
Intelligent Predictive Maintenance as a Service (IPMaaS)
ASTRA – Rapid Digital Transformation of Legacy to Cloud-Native Applications
The contents of this presentation are for informational purposes only. Halliburton** makes no representation or warranty about the accuracy or suitability of the information provided in this presentation and any related materials. Nothing in this presentation constitutes professional advice or consulting services. No contractual relationship with Halliburton is established by attending or viewing this presentation. No rights to intellectual property of Halliburton are granted through this presentation. The opinions expressed in this presentation are those of the author and do not necessarily represent the views of Halliburton. **Halliburton means Halliburton Energy Services, Inc., Landmark Graphics Corporation, and their affiliates.