• Lucian Fogoros, Co-Founder & Managing Director, lloT World
  ×

  Lucian Fogoros

  Co-Founder & Managing Director

  lloT World

  

What are effective cyber risk mitigation strategies that can be deployed for Operational Technology? The threat landscape is ever-changing in addition to the operational environments that we are trying to protect.

• James Goosby, I&C Support General Manager, Southern Company
  ×

  James Goosby

  I&C Support General Manager

  Southern Company

  
• Shahrul A. Rashid, Petronas
  ×

  Shahrul Rashid

  Petronas

  
• Kenneth Ferguson, Vice President, Nuclear Technology, Global Nuclear Associates
  ×

  Kenneth Ferguson

  Vice President, Nuclear Technology

  Global Nuclear Associates

  

  Kenneth Ferguson has been involved in high profile initiatives for over forty years as an engineer, senior manger, and consultant .His current affiliations include being involved as a technology Vice President as well as providing strategic planning services .Involvements have included the power generation, information technology ,petrochemical industries including the private sector and government agencies. A recurrent responsibility has involved program development, change management, and transforming organizations to respond to key situations ,establishing effective cyber culture is a relevant, high priority involvement. Dr. Ferguson has physics and engineering degrees from the University of Michigan(Ann Arbor) and Carnegie-Mellon University. Past affiliations have included as employee of Westinghouse Electric, and then Price-Waterhouse Coopers. He has operated his own consultancy for over ten years.
• David Foose, Security Program Manager, Emerson Power & Water Solutions
  ×

  David Foose

  Security Program Manager

  Emerson Power & Water Solutions

  
• Dennis Fairchild, Technical Sales Associate and Founding Member, IOT Integration Services
  How to talk Cyber security needs with your IT department, more that reciting NERC CIP Standards

  How to talk Cyber security needs with your IT department, more that reciting NERC CIP Standards

  ×

  Over the past 20 years, cyber security has become a major concern for the Power Generation Industry. Even though cyber security has become a critical risk to manage, many power generation personnel find the subject matter daunting. Power Generation personnel are confronted with ever changing government regulations, bombarded by social media reports of various security breaches, and forced to seek technical support from personnel who do not understand the critical functions of their business. Today, generation personnel need simple solutions to help secure their control system networks. Approach: To help power generation personnel build a working relationship with their organizations IT personnel and develop a simple framework for their control systems cyber security requirements. Findings: Developing a viable cyber security plan for any organization requires more than just learning the NERC CIP standards and wildly spending the organizations capital on the latest “Access Control Devices”. A methodical solution needs to be employed that is robust and scalable to the organizations current and future needs. Conclusions: Information Technology and Operations Technology professionals all want to protect their organizations equipment from cyber attacks. The keys to developing a viable cyber strategy is one that takes into account the NERC CIP requirements, does not present unnecessary burdens to site operations, and fosters a positive relationship with Information Technology professionals. Extras: Defense in Depth Strategies Network Trip / Isolation Device Unidirectional Gateway
  ×

  Dennis Fairchild

  Technical Sales Associate and Founding Member

  IOT Integration Services

  

What approach do you take to implement an OT network that supports your digitalization plan? This session offers approaches that can be utilized to "connect" your plant and realize effective and secure systems integration.

• Lauhael S. Godinho, Director, Digital IT: Logistics, Controls, PCN & Operational Systems, Phillips 66
  ×

  Lauhael Godinho

  Director, Digital IT: Logistics, Controls, PCN & Operational Systems

  Phillips 66

  
• Bradley G. Houk, PE, Senior Engineering Advisor ExxonMobil Open Process Automation Project Manager, ExxonMobil Corporation
  ×

  Bradley Houk, PE

  Senior Engineering Advisor ExxonMobil Open Process Automation Project Manager

  ExxonMobil Corporation

  
• Juan Panama, *Invited, Global Business Development Manager, Emerson
  ×

  Juan Panama

  Global Business Development Manager

  Emerson

  

  Juan Panama is the Global Business Development Leader for the Power Industry and has been with Emerson for six years. He is focused on helping utilities increase availability, reliability, safety and workforce effectiveness using wireless technology. He has an undergraduate degree in Mechanical Engineering as well as a Master’s in Business Administration from the University of Minnesota in Minneapolis Minnesota.
• Michael Dessauer, *Invited, Analytics Lead Architecture Specialist at The Dow Chemical Company Digital Operations Center, The Dow Chemical Company
  ×

  Michael Dessauer

  Analytics Lead Architecture Specialist at The Dow Chemical Company Digital Operations Center

  The Dow Chemical Company

  

  Michael P Dessauer is a data scientist in the analytics platform at the Digital Operation Center. Michael is currently focusing on manufacturing analytics areas such as predictive maintenance, geospatial productivity, and text analytics. Prior to joining the Dow Digital Operation Center, Michael was a data scientist with Dow’s Advanced Analytics group for six years. Michael has a PhD in Computational Analysis and Modeling from Louisiana Tech where he developed machine learning models for image classification and tracking applications.

Turn Arounds and Outages are high risk, critical moments in the life of any plant, where literally time is money. This session will look at how Digital tools can help better detect work required, schedule and execute safely during a turn around to complete on time and to budget.

• Matthew Peters, Manufacturing Engineer, 3M
  ×

  Matthew Peters

  Manufacturing Engineer

  3M

  
• James Del Rossi, Mobile Hardware Architect, Dow Water & Process Solutions
  ×

  James Del Rossi

  Mobile Hardware Architect

  Dow Water & Process Solutions

  
• David Harper, Chief Manager - Digital Sales Transformation, GE Power
  ×

  David Harper

  Chief Manager - Digital Sales Transformation

  GE Power

  

• Jon Towslee, Advanced Services Leader, Americas Industrial Automation, Energy Industries (IAEN), ABB Inc.
  ×

  Jon Towslee

  Advanced Services Leader, Americas Industrial Automation, Energy Industries (IAEN)

  ABB Inc.

  

  Jon Towslee is the Advanced Services Leader, Americas for ABB Inc. based in Cleveland, OH. Jon has 29+ years of energy industry experience. He has spent the last 19+ years leading, growing, and managing energy related businesses. Throughout his career, Jon has worked on both sides of the electrical industry from utility transmission, distribution, and generation on the supply side to commercial and industrial energy management, power distribution, and advanced energy analytics on the demand side. In his current capacity, Jon is responsible for leading ABB’s advanced services and digital sales for its Industrial Automation Energy Industries business unit. Jon joined ABB in 2016. Throughout his career, Jon has held numerous business development, management, and leadership positions. Prior to joining ABB, Jon was the Sr. Vice President, Industrial Segment for an energy analytics company, starting in 2011. He was also the Vice President, Strategic Accounts for a commercial energy management company which he joined in 2006. Prior to that, Jon was the Business Manager for an industrial / utility energy system integration business, which he joined in 1994 and held positions of increasing responsibility including: system engineer, project engineer, project manager, program manager, and business manager. Prior to that, Jon was the Engineering Manager and Supervisor of Electrical Engineering for an electrical equipment manufacturer, responsible for the design and application of metal-clad switchgear, capacitor banks, and harmonic filters. Jon earned his Bachelor of Science in Electrical Engineering from The University of Akron and his MBA from John Carroll University. He is a Registered Professional Engineer in the State of Ohio.
• Joe Michienzi, Vice President, Asset Management, Competitive Power Ventures, Inc. [CPV]
  Combing Technologies to Enhance Flexability

  Combing Technologies to Enhance Flexability

  ×

  According to Gartner Group, in 2018 more than one third of utilities began to bring digital technologies into their organizations. As this change happens across the industry, utilities are not only introducing brand new digital technologies, but also implementing new solutions into their workflow to maximize the systems that have kept the lights on for their customers for centuries. Together, the speakers will address how new technology solutions are helping create better outcomes for utilities and their customers by reducing asset maintenance costs and increasing profitability with novel dispatch and production planning solutions for a fleet of power plants. They will provide specific examples as to how a combination of modular edge-to-cloud solutions, combined with domain expertise and Artificial Intelligence and Machine Learning, create additional value in daily operations with intelligent asset management strategies, reduce manual processes, increase operational flexibility and more accurately generate day-ahead forecasts. They will also share how the adoption of the new technologies has impacted output and performance.
  ×

  Joe Michienzi

  Vice President, Asset Management

  Competitive Power Ventures, Inc. [CPV]

  
• Zoran Veljkovic, Principal Digital Product Manager, GE Power Digital
  Digital Worker and Remote Power Plant Operations for safer and cost effective Plant O&M

  Digital Worker and Remote Power Plant Operations for safer and cost effective Plant O&M

  ×

  Remote Operations is a suite of software and hardware solutions designed to enable power generation operators to further optimize and reduce the power plant staffing required to operate and monitor a power plant. This on-premises and off-premises solution will allow an operator to operate and monitor a power plant away from the control room: • Location somewhere in the plant (mobile fashion). Mobile solutions within the plant are provided to enable full plant control, electronically take plant logs and rounds, access CMMS, create Permit to Work , and Lock Out Tag Out – (Optional). • Location outside of the plant boundaries- fixed location. allows for the consolidation of operators across several peaking plants to a single, centralized location. The solution is applicable for Single cycle and combine cycle gas power plants As it is today, Fossil fueled power plant O&M cost is becoming too High. There is no Plant centralized operations capability in the market as well as not fully automated / autonomous Plant operation. The work force is changing very fast and Cyber Security and power plant vulnerability are in customer concern focus. Regional / Governments Regulations and Standards are very much diversified and not aligned. Insurability cost and Plant total installed cost stays high as well. With proven technology backed by extensive industry expertise GE provides remote plant operations as a solution with the highest security and reliability, and best value for worldwide power producers on plant and/or subsystems level to enable profitable power generation.
  ×

  Zoran Veljkovic

  Principal Digital Product Manager

  GE Power Digital

  
• Ed Dupre, Regional Sales Manager, ABB Inc.
  Leveraging Digital Solutions to increase Operator Effectiveness & Plant Performa

  Leveraging Digital Solutions to increase Operator Effectiveness & Plant Performa

  ×

  Operator effectiveness is fundamental to power plant performance. With daily decisions of control room personnel directly impacting safety, production, process availability and quality, it is important to keep the operator in focus. This session will cover an end user case study and best practices of Operator HMI upgrades across their large fleet. It will cover the integration approach to High Performance Graphics, the migration strategy, including benchmarking and Key Performance Indicators. The discussion will also include how a step-by-step migration approach made the process more manageable for the Operators as well as some possible future next steps.
  ×

  Ed Dupre

  Regional Sales Manager

  ABB Inc.

  

  Ed Dupre is a Strategic Account Manager working with Wheelabrator and a select few other of ABB’s most important customers. Ed works with customers that need consultation at the corporate level with such things as long term budget optimization, life cycle management, standardization across the fleet, multi-plant long term service agreements, and many other initiatives. Ed has been with ABB for 23 years, holding front line, management and executive positions in Sales and Marketing, with both domestic and international responsibilities. Ed started his career with AEP at their Muskingum River Power Plant and then moved on to Bailey Controls Company, which was subsequently acquired by ABB.

A technical presentation showing how Nova Scotia Power has leveraged a common equipment hierarchy across multiple software platforms to create Operational Intelligence for their fleet of assets. Nova Scotia Power are currently in the middle of a pilot project that uses the Element Analytics PI Asset Framework Accelerator to construct a PI AF hierarchy that matches the master hierarchy in the MMS system and in GE APM. This PI AF framework is used to contextualize equipment assets with their associated tags and to construct asset templates for the major asset classes that can be cloned across the fleet.

• Mike Greene, Operational Technology Lead, Nova Scotia Power
  ×

  Mike Greene

  Operational Technology Lead

  Nova Scotia Power

  

  Mike has over 25 years of experience with Databases, Information Systems and Plant Automation. He is currently the Operational Technology Lead for the Enterprise Asset Management Office of Nova Scotia Power, the utility for the province of Nova Scotia. He is responsible for the successful integration of NSPs many Operational information systems into a platform that provides end users with a dynamic asset based Operational Risk across the NSP Thermal fleet.
• Jacqueline Grunwald, CIO, AdvanSix, Inc.
  End to end Digital Engagement/ From Value Estimation to Adoption to Value Realization

  End to end Digital Engagement/ From Value Estimation to Adoption to Value Realization

  ×

  "The debate over ‘why’ companies need to embrace digitization is over. Despite this shift, the actual results and value derived from digital investments remain mixed. As we enter 2020, digital investments will continue to rise, so too will the expectations for value – from shareholders, executives (CEO, CFO, CIO), employees, and yes, most importantly, customers. Current research indicates that the odds are against those companies embarking on Digital Transformations with several reports suggesting that over 2/3rds fail to meet expected results. In most cases, technology is not the culprit of these failures. This two forces are shifting leaders and decision makers to focus on ‘how’. While we have sat through several presentations that touch on this topic, many remain at the strategic level leaving lots of gaps on the details of how companies realized the value from a cultural and change perspective. Our presentation will take the audience inside a real on-going digitization initiative sharing and discussing the frameworks and best practices that are making this transformation successful as well as some of the obstacles and learnings along the way. This is a use case that would be presented to the audience while highlighting the frameworks, best practices, and key learnings experienced from over 16 months of the engagement. We would periodically pause for audience questions and lively discussions. Example speaker outline: End to End Engagement Cycle: From Value Estimation to Adoption to Evaluation 1. Identifying and Estimating Potential Value: Top Down and Bottoms Up Approach 2. Outcome Mapping – from Corporate KPIs to Operational Milestones 3. Data Mapping – creating a data collection strategy and visualization for Operational and Executive Decision Making 4. Change Management Plan – Cross Functional Ownership (CIO/Business) who owns what? 5. Culture & Behavior Change – Sub Use Cases that Delivered quick wins and drove adoption (e.g. mobility) 6. Communications Strategy – How do you effectively communicate with Maintenance & Reliability Teams to drive adoption? 7. Measuring, Evaluating, and Sustaining Value
  ×

  Jacqueline Grunwald

  CIO

  AdvanSix, Inc.

  
• Michael Magerstädt, Manager, Global Industry Division Chemical, KROHNE AG
  Chemical Plant Instrumentation IOT and Industry 4.0 – the last 12 months and the future

  Chemical Plant Instrumentation IOT and Industry 4.0 – the last 12 months and the future

  ×

  Last year, we reported on developments on digitalization of chemical plant automatization and particularly on process instrumentation. Core automatization of chemical processes has assured a very high level of safety as well as ever-improving efficiency in this industry over decades. Feedback loops that are designed to maintain chemical reactions within the safe and most efficient range consist of a process control system (PCS), instrumentation to monitor process conditions, and actuators to control the process according to the PCS’s instructions. The advent of large data processing capabilities allows acquisition and utilization of large amounts of data that can make chemical plants self-reliant. These data have neither been acquired nor utilized in the past. In addition, over the last years, almost an entire generation of technicians and engineers is retiring. Up to two decades of reducing headcount in chemical plants combined with a worldwide focus of education systems away from science over a large part of this time led to a generation gap. Since the renewed interest of young people in science and technology has yielded graduates, highly qualified young technicians and engineers began entering the industry. Still, a generation gap exists that needs to be filled. Hence, the driving force to combine the data processing capabilities of today and the data available from chemical plant instrumentation to „industry 4.0“ is the need to make chemical plants more self-sufficient. Intelligent systems are required that assess large sensor data volumes and that derive information by sophisticated pattern recognition algorithms. These systems are able to emulate the knowledge about a process which the human brain can acquire by subconscious connection of seemingly unconnected sensory input once a certain level of experience has been reached. In last year’s conference paper, we showed how chemical operators and instrumentation manufacturers are approaching the challenges of industry 4.0 and are utilizing the possibilities that an internet of things can provide. Examples for additional data transmission channels, wireless data transmission, and the modular plant concept MTP (Module Type Package) were presented. In this presentation, progress that both the global chemical players as well as the process instrumentation industry have made over the last twelve months, will be reviewed. A specific look will be taken on the actual need as well as the requirements for so-called M&O data, i.e., sensor data that are not part of the data acquired for the core instrumentation of a plant, but that are providing additional data for monitoring and optimization. M & O information can be gathered by installing additional sensors that are transmitting information relevant to maintenance. A very simple example would be measurement of environmental temperature or humidity that may have a long-term influence on the lifetime of chemical plant components. More efficient and economical than adding new sensors is the utilization of data that process instrumentation can provide beyond the data the instrument generates by fulfilling its original purpose. This presentation will describe the options to do this from an instrument manufacturer’s perspective.
  ×

  Michael Magerstädt

  Manager, Global Industry Division Chemical

  KROHNE AG

  

  Michael Magerstädt received his Ph.D. degree in Chemistry from Berlin Technical University. Later, he completed a postgraduate study in business administration, focus marketing. Michael holds the position of Manager, Global Industry Division Chemical at KROHNE group a leading global supplier of innovative products, Solutions, and services for the process industry with close to 4000 employees and manufacturing plants on 4 continents. After graduating, Michael spent 4 years conducting research in the United States; first at the University of Georgia, Athens, GA and later at the National Cancer Institute, NIH, Bethesda, MD. Then, he spent 13 years at Hoechst AG (later on Hoechst-Celanese, Frankfurt, Germany), formerly one of the three largest German chemical and pharmaceutical companies. After his time at Hoechst, he spent 19 years as CTO and CEO of privately owned manufacturing and technology companies in in Germany and Switzerland before joining KROHNE, the most recent of which was ROSEN, world leader in pipeline inspection and a major player in Asset Integrity Management. Michael is author of one scientific book as well as numerous publications and patents. He was and is active in building and developing technology startup companies. Michael Magerstädt is member of a number of professional organizations, including Gesellschaft Deutscher Chemiker, American Chemical Society, VDI (German Engineering Association), SME (Society of Mining Engineers), AWWA (American Water Works Association), and NACE International (National Association of Corrosion Engineers), where he co-chairs a Technology Exchange Group on “Corrosion in Mining/Slurry Pipelines”. Michael is married, and has two adult children; one son and one daughter. His non-professional interests besides his family include flying single-engine aircraft, skiing, and his international network of friends.

Industrial companies are sitting on a tremendous amount of data - a treasure that grows every single day. With the advent of AI (Artificial Intelligence) and ML (Machine Learning) technology mining this data for value is quickly becoming reality. This session will focus on providing an understanding of some of the building blocks needed to deploy AI and ML, as well as provide a few examples and learning journeys for specific applications of AI and ML for predictions, creating insights and taking further steps towards automation (predictive maintenance, corrosion detection and plant operation, to name a few).

• Mr. Pete Davies, VP of Digital Engineering Solutions, Uniper Enerlytics
  ×

  Pete Davies

  VP of Digital Engineering Solutions

  Uniper Enerlytics

  
• Dr. Yves Gorat Stommel, Director Business Development and Innovation, Evonik Corporation
  ×

  Yves Gorat Stommel

  Director Business Development and Innovation

  Evonik Corporation

  

  Yves Gorat Stommel is currently working as the director of the department for business development and innovation in the technological/methodological area, with a focus on implementing best practices (processes, methodologies and technology) in production / operations. In his current function, Yves’ motivation stems from continuously improving the processes that lead to an increasing contribution of the operations to the business success, mainly through identifying, benchmarking and implementing best practices, coupled with targeted communication, linking knowledge and an emphasis on global instead of local mindsets and approaches.
• Gabriel Prado, Product Marketing Manager, SparkCognition
  Turning Data into Intelligent Action for Industrial Consumers

  Turning Data into Intelligent Action for Industrial Consumers

  ×

  Industrial consumers have more access to data than ever before, but how do you turn that data into intelligent action? Advancements in technology have provided ample opportunities for intelligent automation from machine learning to artificial intelligence. The potential value is immense, and we are constantly being presented with the potential business benefits of leveraging these technologies. To truly benefit from these technologies however, a foundation of contextualized data needs to be readily available. This thought-provoking discussion will introduce execution strategies to build a robust foundation that will ultimately achieve technologically driven intelligence into industrial operations. Our discussion will focus on industrial water treatment applications for both power and chemical processing industries, with the ability to apply these concepts to any industrial consumer application looking to leverage the latest technology to improve their business and operations.
  ×

  Gabriel Prado

  Product Marketing Manager

  SparkCognition

  
• Prakhar Behar, Mechanical Engineer, Oceaneering International, Inc.
  Automatic Rust Detection and Classification for Asset Integrity using Machine Learning

  Automatic Rust Detection and Classification for Asset Integrity using Machine Learning

  ×

  "Background: The total annual cost of corrosion is $2.5 trillion. This is nearly 3.4% of the world’s gross domestic product (GDP). This study by NACE also predicts that implementing strategies to prevent corrosion can result in savings of about $375-$875 billion. Corrosion is a huge problem in both the power generation and chemical processing fields, with the costs in the power generation industry alone costing the U.S. over $23 billion annually. In the Oil & Gas industry, corrosion continues to be a significant problem. The UK’s Energy Institute ranks corrosion as the second most frequent cause in initiating loss of hydrocarbon containment, and highlights it as a major threat to asset integrity and plant efficiency. The cost of corrosion can be broken down into surface pipeline and facility costs, downhole tubing expenses and in capital expenditures related to corrosion. In both, power generation & chemical processing field, the challenge of corrosion is further exacerbated by an aging infrastructure utilizing obsolete technologies that increase the vulnerability to theft and cyber attacks. These could lead to outages and health, safety and environmental damage costs. Hence, there is an urgent need for a robust corrosion inspection system with state-of-the-art technological capabilities to keep costs and downtime to a minimum. Currently, inspection to detect corrosion is performed manually. Inspectors watch hours of video footage and analyze tens of thousands of images for damages caused by corrosion. Few companies have, thus far, implemented machine learning with the aim of automating this laborious and inefficient process. Our Approach & Results: We considered several machine learning models with the goal of predicting the area and severity of rust within images. The labelled data was split into a training set (75%) and a test set (25%). The selected machine learning algorithm learns the features from the training images independently and attempts to find similar patterns in the test images to predict the levels of rust. This algorithm had an average accuracy of about 60%. This project has shown that there is a definite use for machine learning in the Oil and Gas industry, especially for asset integrity. We have concluded that deep learning models should be employed where the format of the data is an image. This will enable semantic segmentation, where each pixel in the image is categorized into a class. According to state-of-the-art research, this has been noted to outperform machine learning models in the problems involving segmentation. Our aim now is to make available more high-quality data for our algorithm to improve further. Next steps: Our vision for the future entails integrating the finished rust detection software into a drone for automatic detection and classification. The drone will also have the capability of alerting relevant maintenance teams so that immediate remedial action is taken. Furthermore, by using predictive analysis to pinpoint the time to failure, it will allow maintenance to be scheduled well before that."
  ×

  Prakhar Behar

  Mechanical Engineer

  Oceaneering International, Inc.

  

  "I am working as a Mechanical Engineer in the product development team at Oceaneering in Houston TX. During my time here, I have had the fortune to collaborate and work on multiple projects along side experienced engineers and SMEs from various departments dispersed throughout the globe and learn a great deal from them. This has helped me to expand my knowledge base in technologies employed in both offshore and onshore spaces. My expertise includes CAD modeling, Finite element analysis, GD&T analysis, and Lean Six Sigma with high proficiency in all the statistical process control tools I have graduated with a Masters in Mechanical Engineering from University of Houston, and currently working towards PE license in the state of Texas, after acquiring E.I.T. (Engineer in Training) certification in July 2018."
• Gabriel Prado, Product Marketing Manager, SparkCognition
  Why breakthroughs in scaled asset-centric modeling are critical for the connected plant

  Why breakthroughs in scaled asset-centric modeling are critical for the connected plant

  ×

  "Plant-wide predictive maintenance programs are only as good as the accuracy and deployability of their predictive models, but traditionally, these models have required significant overhead to develop and maintain. This has resulted in disparate systems and processes that are deployed only on select assets, maybe delivering below-average results, and may have resulted in other business trade-offs. However, with recent advancements in both science and automation, machine learning models offer a clear path to improved accuracy, standardization, and scale based on asset-centric modeling. This session will explore the techniques and fundamentals for building asset-centric machine learning models quickly without sacrificing accuracy and will discuss best practices around successful deployments at plant-wide scale. These concepts will then be mapped to a recent case study for an offshore oil and gas production platform where SparkCognition was able to reduce the deployment time for 40 asset-centric models by 50% to two months while delivering KPI-exceeding results. After this session, operations executives, managers, and leads will: 1) Understand the differences and performance trade-offs between generalized and asset-centric models and link it to their business problem 2) Be able to identify the requirements, technology, and process for developing and deploying these asset-centric models 3) Have learned best practices from a live, scaled deployment in a complex production environment
  ×

  Gabriel Prado

  Product Marketing Manager

  SparkCognition

  
• Dr. Berk Birand, CEO, Fero Labs
  Leveraging explainable machine learning to drive plant optimization

  Leveraging explainable machine learning to drive plant optimization

  ×

  Power and chemical plants produce and store significant amounts of data. Yet, the machine learning (ML) and artificial intelligence (AI) techniques that have revolutionized the technology sector have not been widely adopted amongst operations improvement groups. A major shortcoming of the popular ML methodologies in use today is their black-box nature. Google may identify a specific image with high precision, but neither the developers nor the consumers know why the algorithms think those pixels define the target image. As long as most predictions are accurate, they are happy. At a plant setting having mostly reliable prediction is not enough; a single false negative can lead to the costly shutdown of critical operations. Plant engineers need to go beyond accurate predictions and understand the whys and hows and corroborate these insights with their subject matter expertise before they can actually trust the software. Lack of explainability of popular ML techniques is why engineers still rely on methods such as Six Sigma. Yet Six Sigma doesn’t scale with data and engineers can conduct only so many single shot design of experiments. Fero, with its Explainable ML technology based on Bayesian machine learning, addresses this problem. Explainable ML continuously learns from the entire plant’s data. The models reveal root-causes of production problems, map input/output relationships for emission reduction, and enable what-if simulations for planning optimal asset performance. By explaining the reasons behind the predictions, Explainable ML builds trust with the user. The easy-to-use software comes with automatic data cleaning and ML modeling features. As a result, engineers without data science training can easily conduct advanced data analytics projects. By simplifying ML modeling and making models transparent, Fero builds trust between ML technology and human users and makes these powerful new techniques accessible to subject matter experts of the plant. Publicly available information: Uniper Energy: https://www.littlegatepublishing.com/2019/05/endeavour-magazine-june-2019/ (page 48-49) AkzoNobel: https://www.chemicalprocessing.com/articles/2018/imagine-chemistry-challenge-awards-support-to-startups/ Session Takeaways: - Truly explainable machine learning techniques do exist and pave the way for a wide range of applications not possible with black-box methods - Democratizing ML and empowering subject matter experts with powerful, modern ML tools is the only scalable approach - Using ML purely for predictions is a loss of opportunity; these new techniques can be used for various process and quality improvement projects
  ×

  Berk Birand

  CEO

  Fero Labs

  

  Berk holds a Ph.D. in electrical engineering and computer science from Columbia University. His academic research focused on optimizing wireless and optical networks with efficient cross-layer algorithms. He developed scheduling algorithms for LTE-A and fiber-optic networks. He holds a patent in IoT systems for resilient fiber-optic networks.

• Sydni Credle, Ph.D., Project Manager and General Engineer Enabling Technologies & Partnerships, US Dept of Energy/National Energy Technology Lab
  ×

  Sydni Credle, Ph.D.

  Project Manager and General Engineer Enabling Technologies & Partnerships

  US Dept of Energy/National Energy Technology Lab

  
• Mr Chris Welch, Rotational Development, Tennessee Valley Authority (TVA)
  ×

  Chris Welch

  Rotational Development

  Tennessee Valley Authority (TVA)

  

  I've been involved in some fantastic projects over the years. They tend to be bleeding edge technologies, which is a blast! Wired networks, wireless networks, servers, and everything in between are my modus operandi. My goals in life include solving problems with new technologies, and using technology to make a difference in the lives of those around me. Specialties: My specialties include engineering solutions, and translating between those who speak technology and those who need it. I love to build strong/passionate teams that can reach the moon and beyond. Come to us with your problems, and we will solve them. Recent Projects: Industrial sensors (IIoT), non-ip based wireless communications, machine learning, AI, robotics, hybrid manufacturing, and changing the world bits at a time. Putting the techy stuff aside for a bit, my other passion is developing leadership and making connections between amazing people. Taking advantage of my experiences from a broad range of fields, I have been able to leverage my time around some very different thinkers. I like to share those thinking styles to others over discussions and sessions while we develop ideas and solutions. I cannot think of anything that is more fun than that at work.
• Hannah Davis, Engineer/Scientist II, EPRI Electric Power Research Institute
  The Growth of BlockChain

  The Growth of BlockChain

  ×
  ×

  Hannah Davis

  Engineer/Scientist II

  EPRI Electric Power Research Institute