O slideshow foi denunciado.
Utilizamos seu perfil e dados de atividades no LinkedIn para personalizar e exibir anúncios mais relevantes. Altere suas preferências de anúncios quando desejar.

AI Solutions in Manufacturing

160 visualizações

Publicada em

This presentation was made on June 11, 2020.

Recording from the presentation can be viewed here: https://youtu.be/02Gb062U_M4

The manufacturing industry is adopting artificial intelligence (AI) at a fast rate. This century-old industry is complex but has seen constant transformation across all of its facets.

Led by big data analytics, miniaturization of sensors enabling the Internet of Things (IoT), and, now, AI machine learning (ML), manufacturers everywhere have embarked on an AI transformation that is opening up potential new revenue streams as well taking costs and time out of existing processes.

This talk will walk through a use case for enterprise AI solutions within the manufacturing sector. We will discuss the challenges, motivation, and tool selection process, then cover the solution development in detail.

Speaker Bio:

eRic is armed with the technical know-how of Data Science, Machines Learning, and Big Data Analytics. He. is equipped with skill-sets to value-add businesses exploring into areas of Artificial Intelligence (AI) with an AI consultation approach. Translating BDA, Machine Learning, and AI into Business Values.
eRic CHOO had spent the last 8 years in the IT industry from integration of Infrastructure (Storage and Back-up) solutions to Advance Analytics Software specializing in BDA, Machines Learning, and AI. Before joining the IT industry, he had vast experience in the Semiconductor industry, thus a deep understanding in advance manufacturing processes.

SIONG Jong Hang works as a Solutions Engineer/Data Scientist at H2O.ai based in Singapore where he helps business, government, academia, and non-profit organizations in their transformation into AI. Prior to H2O.ai, he has worked at the Quant Group at Bank of America Merrill Lynch in Hong Kong and Teradata in Singapore as a data scientist. He has completed data science projects for various verticals in Europe and Asia. After hours, he’s an avid learner and has attained 100 MOOC certificates in various fields such as AI, science, engineering, and maths. He has also authored articles to instill interest in science, technology as well as AI.

Publicada em: Tecnologia
  • Seja o primeiro a comentar

  • Seja a primeira pessoa a gostar disto

AI Solutions in Manufacturing

  1. 1. H2O.ai Empowers an AI Transformation Presented by: eRic Choo, MBA Technical Business Development Manager
  2. 2. Confidential2 H2O.ai is the open source leader in AI and Machine Learning Democratize AI for Everyone Make Your Company an AI Company 2 Confidentia l
  3. 3. Confidential3 Confidential3 Industry 4.0 describes the trend where automation powered by data may be able to perform everyday tasks. It is blurring the lines between the physical, digital and biological worlds. It also refers to the blending of two or more technologies such as genetic engineering, materials science, Artificial Intelligence (AI), advanced manufacturing and the Internet of Things (IoT) to result in innovations and solutions. Industry 4.0 has ushered in the digital economy where technology is ubiquitous, driving transformative change in the way we live and radically disrupting every business sector. SGINNOVATE, (2020), “Industry 4.0: Singapore’s Strategy”, 20 February 2020, [online]. Available at https://www.sginnovate.com/blog/industry-40-singapore’s-strategy Industry 4.0: Singapore’s Strategy Spectral Engines Oy, (2018), “Industry 4.0 and how smart sensors make the difference”, 26 February 2018, [online]. Available at https://www.spectralengines.com/articles/industry-4-0-and-how-smart-sensors-make-the-difference
  4. 4. Confidential4 Closer to i4.0: How is AI improving Manufacturing? • Machinery Maintenance and Quality are the leading AI transformation projects in manufacturing operations today. • Caterpillar's Marine Division is saving $400K per ship per year after machine learning analysed data on how often hulls should be cleaned for maximum efficiency. • The BMW Group uses AI to evaluate component images in ongoing production lines to spot deviations from the standard in real-time. • Improving demand forecast accuracy is showing solid results across multiple industries with Consumer-Packaged Goods manufacturers leading all others. – Danone Group Columbus, L., (2020), “10 Ways AI Is Improving Manufacturing In 2020”, 18 May 2020, [online]. Available at https://www.forbes.com/sites/louiscolumbus/2020/05/18/10-ways-ai-is-improving-manufacturing-in-2020/#66cc30891e85
  5. 5. Confidential5 AI in Manufacturing: In a nutshell $ Introducing AI into different aspects of a manufacturing process can potentially reduce running cost in terms of maintaining asset machineries and material usual. AI on the other hand can maximise yield performance of a manufactured product in many ways.
  6. 6. Confidential6 Data + AI + People Transformations
  7. 7. Confidential7 CPO Refinery Process (producing better quality of CPO)Incomingandunloading 0.5% moisture oven weighing Unloading area Storage tank Transfertoprocess Refinery Process Filtration (5μm) Not more than 265℃ PLC with SCADA for monitoring RTD sensors to measure final temperature Passthroughdifferent filtrationprocess Furtherremove moisture Leaf Filter Process Dryer Tank Pressure Sensor (Outlet vs Inlet) Hourly Delta measurement CPO
  8. 8. Confidential8 Biodiesel (FAME) and Glycerin Process (FAME process optimisation) Incoming Chemical Refining Trans-esterification Washing Drying Catalyst (fats & Methanol) Biodiesel (FAME) Glycerin (Detergent)
  9. 9. Confidential9 Robotic Welding Arm Welding process Photodiode Camera / Image Processing Visible Light Emission Plume and Molten Pool Feature Keyhole Feature Time Domain Frequency Domain Feature Vector Machine Learning Algorithm Welding Status Prediction Training Yetkin, C., (2019), “Application of Machine Learning and AI in Industrial Welding”, 28 January 2019, [online]. Available at https://datapao.com/application-of-machine-learning-and-ai-in-industrial-welding/ Monitor & raise alarms -online- Compute model -offline- ABB Robotic, (2013), “ABB Robotics - Welding Scooter & Motorcycle Frames”, [youtube-online]. Available at https://www.youtube.com/watch?start=21&feature=oembed&v=tRuqb6wquig
  10. 10. Confidential10 Our AI and ML Platforms. The leading and visionary platforms today.
  11. 11. Confidential11 Model Development Model Deployment H2O MOJO Model Object Smartphone Watch Factory T-Shirt Engine Predictive Maintenance Direct plugin to Edge & IoT Production Cases Learning Feedback Loop From Model to Production Deployment on Edge Load Data Run AutoML Winning Model Generated
  12. 12. Confidential12 Driverless AI Solves Talent, Time, Trust Challenges Explainable AITop 10 Experts Data science experts and Kaggle Grandmasters built the tips and techniques within the platform. Months to Hours Driverless AI has the technology that accelerates the data science workflow. Technology that provides interpretability, reason codes, visualization, and easy to read documentation.
  13. 13. Confidential13 Goal: Focus and Ensure Customer Success that results in overall happiness leading to expansion of product usage. • Resolution of technical errors via the Support Portal, e-mail, or phone • Installation support • Assignment of an H2O Program Manager • Regular cadence with H2O Team • Access to H2O Data Scientist and Engineers for guidance and support on model implementation and production Roles and Responsibilities • Ensure Success • Monitor Delivery • Know the Customer • Manage Documentation • Engage Stakeholders • Secure Resources • Establish a Cadence Key Performance Indicators • Models in Production or business outcomes affected • Customer Success/Happiness Index (Net Promoter Score) • Expansion of Product Value and Usage • Referenceable Customer • Visibility to the highest Ranking Executive Champion (Closest to CEO or • CXO) Where We Engage? • On-Boarding • Touch Points • Support Tickets • Customer Training & Activities • Expansion and Renewals H2O.ai Customer Success Team
  14. 14. Confidential14 H2O.ai: AI and ML Platforms – Open Source • 100% open source – Apache V2 Licensed • Enterprise support subscriptions • Interface using R, Python on H2O Flow H2O Enterprise Support delivers the training, support and optimization services you need to go faster and deliver results in enterprise production environments for H2O, H2O Sparkling Water and H2O4GPU. Enterprise Steam (available w/Enterprise Support) Secure, Self-Service Artificial Intelligence Environments with Comprehensive IT Control In-memory, distributed machine learning algorithms with H2O Flow GUI H2O AI open source engine integration with Spark H2O4GPU is an open source, GPU-accelerated machine learning package with APIs in Python and R that allows anyone to take advantage of GPUs to build advanced machine learning models. web-based interface of H2O allowing users to explore all available features and algorithms within H2O 3
  15. 15. Confidential15 We Partner to Make Your Company an AI Company We will guide you on your AI journey Your AI Journey Access our expert data scientist, software developers and Kaggle Grandmasters Experts We reference learnings from our customers, community and hundreds of successful use cases and lend our experience to you Empathy We work with your domain experts and data scientists to become AI experts We offer world class technology and support Teach World Class Technology We will work with you on innovative AI and new business models (JV, rev share) Your Future
  16. 16. CONFIDENTIA Thank you