The digitalization of the energy system will computers take over? Michael Weinhold CTO Siemens Energy Management

Transcription

1 The digitalization of the energy system will computers take over? Michael Weinhold CTO Siemens Energy Management Unrestricted Siemens AG Österreich 2017 siemens.at/future-of-energy

2 Agenda Digitalization of the Energy System Artificial Intelligence (AI) AI Use Cases 4 Outlook Page 2

3 The Energy Revolution: Big Picture From centralized power and unidirectional grid Distributed Energy Systems to Decentral and Distributed Energy Systems and bidirectional balancing Transmission Distribution and Consumption 1 Changing generation mix 2 Generation capacity additions Distance from source to load Decentralization (public/private) Refurbishment/ upgrades Page 3

4 Germany: more renewable generation capacity than peak load 5/2017: ca. 105 GW Peak Load ( Spitzenlast ) ca. 82 GW 12/2015: ca. 93 GW More installed Wind- and PV-capacity than Peak Load Base Load ( Grundlast ) ca. 40 GW >65GW of Wind and PV-Power Plants connected to LV- and MV-systems Trend towards an Electronic Grid Resulting Challenges for grid operators: Changing system dynamics Frequency and Voltage Stability Short Circuit Power Source: Page 4

5 Germany: Power Generation and Consumption in April 2017 Renewable Power Generation: ca. 55 GW ca. 10 GW Source: Page 5

6 Generation mix in 2030, example European Union (forecast to be further verified) EU28 generation capacity until 2030 (TW) Impact on Grid business +1.4% Integration of Renewables % 7.8% Grid extensions 9% 12% 3% 12% 21% 12% 3% 9% 5.1% 0.5% 2.6% -0.7% Stability challenges (less inertia, towards solid state grid ) Power quality and reliability of supply Cyber Security Automated operation and situational awareness 50% 33% -1.3% New business models, solutions and customers 2015 Retirements Additions Solar PV Hydro Wind other Renewables Source: Siemens 2030 Nuclear x% CAGR Fossil (incl. Engines, CHP) Regulatory uncertainty and public acceptance Disruptive potential from cheap storage Page 6

7 Energy storage applications and sector couplings Application cases by location of storage Central Large Utilities Distributed Small utilities, municipalities, industry prosumer Pumped storage H2/Chemicals Battery Thermal Electricity Electricity H 2 / H 2 Fuel Methane for car Electricity Heating, Cooling (gas grid) Grid balancing and stability Power to gas Power-to-chemicals Grid stability, self-supply, electro-mobility Power-to-heating and -cooling Page 7

8 Possible advantages of the blockchain technology No separate financial management process necessary any more Cost reduction: (real-time) peer-to-peer transactions with integrated payment / currency Security: Immutable records & Trust of source No verification costs to prove the origin e.g. of renewable energy, performance validation Blockchain (P2P) energy trading (incl. e. g. regulation power), grid charging, nodal pricing, EV-Charging New Business Models: Multi-stakeholder collaboration in different use cases Optimization: Data available in one distributed / shared database analytics across shared data (e. g. price data, market data, load profiles, grid utilization) Page 8

9 Partnership Siemens LO3 Energy Innovative Microgrid solution supporting New York s Reforming the Energy Vision (REV) program (currently under development) Page 9

10 Digitalization drives dramatic change in energy systems Affordability Availability Environment Renewables and conventional Forecasting Distributed energy Generation control HVDC / FACTS Energy Storage Electrification Agility in energy Automation Digitalization Grid stability Security Power to X MDM e-mobility Customer engagement HVDC/FACTS = High Voltage Direct Current/Flexible AC Transmission Systems MDM = Meter Data Management Page 10

11 Challenges Energy System Stakeholder challenges Complexity and Uncertainty (Technology, Regulation) System Dynamics (Stability) Vulnerability (e. g. Physical and Cyber Attacks) Competitive advantage through: Adaptability, flexibility, speed Forecasting accuracy Decision Quality Data Analytics turns data into knowledge Page 11

12 Application of Artificial Intelligence (including Machine Learning) Act Analyze Prescriptive Analytics Inform Predictive Analytics Diagnostic Analytics Descriptive Analytics What happened? Why did it happen? What will / should happen? What shall we do? Business Applications Service statistics Sales reports Root cause identification Fault analysis Condition monitoring Fault prediction Power optimization Load balancing Data Mining Machine Learning Page 12

13 Agenda Digitalization of the Energy System Artificial Intelligence (AI) AI Use Cases 4 Outlook Page 13

14 Definition and Evolution of AI Definition of AI Creating machines that perform functions that require intelligence when performed by people (Kurzweil, 1990) Before : Expected by : Zuse s Z3, first programmable electronic computer 1997: IBM Deep Blue defeats world s chess champion Kasparov 2005: Honda s humanoid robot Asimo comes to life 2011: Watson wins Jeopardy! against most successful contestants 2014: Alexa, Amazon s intelligent assistant debuts 2016: AlphaGo beats Lee Sedol in a Go match ~2020: All-over virtual personal assistants as interface for consumers 202x: Fully autonomously driving cars become market-ready 20xx: Robots may build robot children on their own Major breakthroughs Algorithmic advances in deep learning Increasing computing power Usage of huge datasets leverage full potential of AI Open platforms and data bases Page 14

15 Other existing AI Use Cases Smart Home Thermostats Energy Management Smart Sales Individualized Advertisement Purchase Description Online Customer Support Customer behavior prediction Smart Security Fraud Detection Security surveillance Smart Entertainment Movie and Music Recommendations Video Games Healthcare Diagnostic tools Personalized medicine Smart Information Information Retrieval News Generation Fake news identification Smart Mobility Dynamic Routing Traffic jam prescription Page 15

16 AI framework mirrors the human brain activation of different brain areas for different types of intelligence AI framework as mirror of human brain Sensor processing Image processing Speech recognition Text processing Reasoning Learning Creativity Key & driving aspect Potential supporting aspect Aspect not in focus Decision Subsets of AI aspects (not exhaustive) System Intelligence Physical Intelligence Visual Intelligence The ability to gather (e.g., sensor streaming) data, recognize and infer patterns in it and generate recommendations or feedback loops The ability to plan and implement movement, as well as react on unexpected events in a human-like way (e.g., autonomous driving) The ability to perceive and interpret visual content Example: Visual intelligence Knowledge Intelligence The ability to understand the semantics of natural language (text or speech), generate hypotheses and even converse with a human Page 16

17 Agenda Digitalization of the Energy System Artificial Intelligence (AI) AI Use Cases 4 Outlook Page 17

18 Use Case: Next Generation Root Cause Analysis Sensor processing Reasoning Image processing Speech recognition Text processing Learning Creativity Decision High Level View Detailed AI Work Flow Efficiency & competitiveness Page 18

19 Use Cases in Power Generation: Gas Turbine Emission Reduction Wind Turbine Control Benefit for Gas Turbines: Reduced costs from emissions Benefit for Wind Turbines: Increased Energy yield Sensor processing Image processing Speech recognition Text processing Reasoning Learning Creativity Decision Page 19

20 Use Cases in Electricity Grids: Online Decision Support for Power Grids (EM DG PROducts) Growing share of renewable energy Growing and share distributed of renewable power generation call for enhanced capabilities of intelligent devices. Wide-Area Disturbance Classification Wide area monitoring combined with decision support Disturbance identification and compensation Fault Localization and Classification Increase quality of supporting information in case of faults Localization of faults even in difficult cases Operation Center: Disturbance Classification Infield: Fault location using neural networks Stream Data Recognize Contingencies Counter Measures Model Training Model Generation Model Deployment Interpreter TIME contingency Source code MLP, CNN, Embedded Analytics Framework (LEAF) Page 20

21 Agenda Digitalization of the Energy System Artificial Intelligence (AI) AI Use Cases 4 Outlook Page 21

22 Outlook More Wind- and PV, Electrification, Distributed Energy Systems Sector-couplings and Energy Storage increasingly relevant Digitalization is key key enabler (simulation, operation, market integration) Emerging Sharing Economy concepts for Prosumers Artificial Intelligence gaining momentum Page 22

23 Contact siemens.at/future-of-energy Page 23