Creating the Grid OS: A
Computing Systems Approach to Energy Problems
CS294-49, Fall 2009
Professors David Culler and Randy Katz
M/W 4-5:30, 320 Soda Hall, 3 Units
Limited Enrollment By Instructors Consent
The world's electric grids are an engineering wonder of last century's physical age, each with a
vast geographic reach, epitomized by a highly centralized, synchronized, and reliable distribution
tree that allows electric power to be consumed without concern for its source. But rapidly changing
energy demands, incorporation of non-dispatchable renewable sources, and the need to proactively
manage load, have pushed this aging marvel to its limit. As the rise in greenhouse gases threatens
civilization, it is time to examine how pervasive information can fundamentally change the nature
of energy production, distribution and use. Taking guidance from the design principles of the
dominant triumph of the cyber age, the Internet, how can we design an essentially
more scalable, flexible and resilient electric power infrastructure - one that encourages efficient
use, integrates local generation, and manages demand through omnipresent awareness of energy
availability and use over time. The crucial insight is to integrate information exchange everywhere
that power is transferred. 21st Century Energy Infrastructure is in reality a computing systems
The purpose of this course will be to
develop the design of a new kind of Energy Network, an information overlay
on the energy distribution system in its various physical manifestations, e.g., machine rooms, buildings, neighborhoods, isolated generation islands
and regional grids. Pervasive information exchange will enable a more efficient scalable energy system
with improved resilience and quality of delivered power. This information overlay brings together
(1) pervasive information about energy availability and use,
(2) interactive load/supply negotiation protocols,
(3) controllable loads and sources, and
(4) logically "packetized" energy, buffered and forwarded over a physical energy network.
Together these yield a system for agile, distributed, and integrated management
of energy that can buffer energy on the path to reduce peak-to-average energy consumption, moderate
infrastructure provisioning, and encourage power-limited design and operation.
Analogous to a router, one can imagine as a building block
an intelligent power switch that logically connect sources to loads by bundling information (bits) with
energy (electrons) flows.
The first third of the course will be dedicated to getting students rapidly up to speed on modern energy systems architectures, primarily from a computing and information systems perspective.
We make NO assumptions about the students' knowledge of energy systems--you need never have taken a
course in electrical engineering!
The middle third of the course will focus on a group design of the information overlay.
The final third will shift to a project intensive mode,
with the goal of developing prototype implementations for major components
of the envisioned energy information and network architecture.
This is a ground floor opportunity to get involved in a new and exciting research project that combines superb
technical opportunities with the possibility of materially affecting people's lives - for the good!
Week 0. (W 8/26) Introduction, The Computer Science in the Energy Problem
- Randy Katz, "A Computer Scientist Looks at the Energy Problem."
- David Culler, "Energy Issues in Buildings"
- Discussion about Course Organization
Luiz Andre Barroso, Urs Hoelzle, "The Datacenter as a Computer: An Introduction to the Design of Warehouse-Scale Machines", 2009. Chapters 1, 4-6.
The Climate2020 Group, "Smart2020: Enabling the Low Carbon Economy in the Information Age," 2008.
Week 1: (M 8/31, W 9/2) The Physical Layer
- Randy Katz:
- Power systems
- David Culler:
- Alexandra von Meier, "Power Delivery Systems Tutorial," 2006.
- Recommended: Alexandra von Meier, Electric Power Systems: A Conceptual Introduction, Wiley, 2006.
- PHY notes [ppt, pdf]
- Electric Power Transmission article from Answer.com [html]
- HVAC notes [ppt, pdf]
Fundamentals of HVAC Controls from www.PDHcenter.com
Week 2: (W 9/9, M 9/14) The Device Layer
- Randy Katz:
- Loads continued (datacenters, microgrids)
- David Culler:
- Devices and Device interfaces (RS485, modbus, SCADA)
- Measurement and Performance (EMS)
- DEV notes [ppt, pdf]
- LBNL Datacenter Energy Roadmap [pdf]
- CERTS Microgrid [pdf]
Week 3: (W 9/16, M 9/28) Information Flows and Protocols
- David Culler:
- Gridwise, SEP, Intelligrid
- Power proportionality
Week 4: (M 9/21, W 9/23) Resource Allocation and Control
- Randy Katz:
- Auotmatic Demand Response, load shifting and shedding
- IOC, Wholesale markets, deregulation, models
Week 5 & 6: (W 9/30, M 10/5, W 10/7, M 10/12) System Architecture: Putting it all together
- David Culler, Randy Katz
- Lifecycle, Carbon Tax, Policy, Regulation, Title 24
- RAS, Security, Vulnerability
- Occupant and stakeholder interactive feedback and visualization
- Energy in emerging regions
Week 7-9: Architectural Design: Project meetings, Formulation, Thrust
Week 10-15: Prototype Implementation and Evaluation
Related Courses (in progress)
- Evaluate--Design--Implement Cycle applied to energy system problems
- Tools, Workloads, Benchmarks, Models, Layered Architectures, Interfaces and
- Energy Equivalent of Workload Characterization:
Analysis of CalISO pricing/regional demand signals (historical energy prices),
California Grid topology/distribution bottleneck analysis and implications for pricing,
Historical weather data (e.g., data from Weather Underground web site),
other real time feeds
- Analysis: Economic/Architectural Analysis of Storage in Grid; Wholesale vs. Retail Analysis of Energy supply and consumption
- Grid Model Formulation: Extraction of demand from observation of energy supply and consumption
(sensing and measurement, data mining, compressed sensing),
energy market formation and matching supply and load, peak shifting and saving, load adaptation;
How to model/validate aggregated load and supply? Predict load and supply. Analogy
with statistical workload generation. How much room for moving load around for peak shifting/shaving--how much load or supply can you actually shift or match in time?
- Building Model Formulation: Autodesk/EcoTect BIM (building information model) building description, e.g., create one for Soda Hall; Load slack aspect of Demand-Response
- Building behavioral models, building population statistical analysis to understand distributions--DOE Buildings Databook
- Tools: market simulators, load model simulators (e.g., GridSpice)
- Node OS prototypes: Zero-idle Motherboard, Power proportional data center services and energy-application performance tradeoffs, Zero-time restart after power shutdown
- Bldg OS prototypes: control systems of a variety of building-scale energy systems, e.g., machine room components, lighting, HVAC; Architecture of a building wide monitoring system
- Grid OS prototypes: control systems of a variety of grid-scale energy systems, e.g., (simulated) power plants, energy markets, aggregated loads, IPS decision control algorithms
- Architectural Implication: System design impact on the penetration of renewables, role of storage
- Retreat Demos as projects
Last Updated: 09 Sep 2009