Predictive Engines For Smart Grid
1 like370 views
The document discusses smart grid technology and predictive engines for operational control. It introduces Metro Power Company as an energy services consultancy focused on real-time energy management. It outlines the smart grid concept and maturity model roadmap for implementation. With increasing data from sensors, artificial intelligence like the Hybrid Artificially Intelligent Statistical Engine (HAISE) will be needed to analyze data and allow for autonomous control elements in smart grid architecture. The convergence of technologies will make smart grids a reality enabled by predictive engines providing autonomous operational control through situational awareness.
Predictive Engines For Smart Grid
- 1. Presentation to 85th National Electric Energy Society of Australia Conference Predictive Engines for Smart Grid Operational Control Timothy Edwards Managing Director August 09
- 2. What is Metro Power Company? Advisory & services consultancy Also known as an ESCO in Europe Energy Services Company Also known as a CleanTech with Proprietary analytical systems Focused on real-time energy & demand management for grid control Energy Conservation | Complexity Management | Smart Grid
- 3. Smart Grid Concept Courtesy of EPRI (2009) Best definition in the American Energy Independence & Security Act
- 5. The Road Map ~ Maturity Model IBM-underwritten Maturity Model is a proven process to create a Road Map Collaboration effort between IBM, APQC, and the Global Intelligent Utility Collaboration (GUIC) Country Energy part of GUIC 5 stage plan, across 8 focus areas
- 8. Wide Area Situational Awareness Courtesy of EPRI (2009)
- 9. Data, data everywhere. Interoperability & Standards soon addressed More data to/from AMI, EMS, SCADA, DMS, OMS, GIS. Complexity theory to adapt to complex problems 2007 Yahoo records 12 TB daily 2015 SKA needs to record 6400 TB hourly Moores Law (CPU Hz x2 per 2 yrs) Kryders Law (Storage x2 per 2 yrs) x 0.5 cost per 2 yrs
- 10. Artificial Intelligence and Data Mining True Data Mining est. 1990s Started as a mix of Statistical Analysis, AI, machine learning and very large data sets. Three paths since 2004: 1. Decision trees 2. Artificial Neural Networks 3. Machine Learning Data Mining uses patterns in data to build models & adjust Neither A.N.N or Regression Statistics will do for autonomous control elements.
- 11. Hybrid Artificially Intelligent Statistical Engine Regression vs Neural Nets AUTONOMOUS MANAGEMENT both inadequate Change Request ANALYSE PLAN HAISE invented 2007 Virtualized Hardware and Change Software environments enable Symptom KNOWLEDGE Plan simultaneous analytical applications to run MONITOR EXECUTE HAISE provides the brain power within autonomous RESOURCE computing architecture HAISE Control Loop for Autonomic Computing Element
- 12. Autonomous Grid Architecture with HAISE SYSTEM OPERATIONS MARKET OPERATIONS Orchestrating Autonomic Managers LMS WMS OMS EMS FCS LOAD WORK OUTAGE ENERGY FINANCIAL MANAGEMENT MANAGEMENT MANAGEMENT MANAGEMENT CONTROL SYSTEM SYSTEM SYSTEM SYSTEM SYSTEM Self-Configuring Autonomic Managers Layered Architecture of a Smart Grid Autonomic Computer System
- 13. Summary Convergence of technologies, roadmaps and interoperability will make smart grid a reality soon With data, data everywhereartificial intelligence will be required to turn that data into useful information Predictive engines such as the HAISE will provide the brains behind autonomous self-configuring, self-healing grid operations through Wide Area Situational Awareness Live Outage Visualisation - VERDE Weather Impact Model Visualisation
- 14. Thank You Timothy Edwards Managing Director E: timothy.edwards@metropower.com.au P: 08 9485 1121 F:08 9485 1119 PO Box 774. West Perth. WA 6872 Energy Conservation | Complexity Management | Smart Grid