NanogridsLawrence Berkeley National Laboratory1 Cyclotron RoadBerkeleyCA94720USABNordman@LBL.govUniversity of South Florida4202 East Fowler Avenue, ENB 118TampaFL33620USAchristen@csee.usf.eduA nanogrid is a very small electricity domain that is distinct
from any other grid it is connected to in voltage, reliability, quality,
or price.
Nanogrids could form the basis of a future electricity system
built on a bottom-up, decentralized, and distributed network model rather than the
top-down centralized grid we have today in most parts of the world.
This document introduces the idea of a nanogrid to the IETF community
for two purposes -- to inform the work on energy management presently
underway in the EMAN working group, and to describe how future
communications within and between grids could be accomplished with
protocols that are the product of the IETF.
There appears to be no fundamental conflict between the nanogrid concept
and the current drafts in the EMAN working group.
A nanogrid is a very small electricity domain that is distinct from any
other grid it is connected to in voltage, reliability, quality, or price
[CIGRE] (also [NG-2009]).
Nanogrids could form the basis of a future electricity system built on
a bottom-up, decentralized, and distributed network model rather than the
top-down centralized grid we have today in most parts of the world.
Central to nanogrids is the ability
to communicate electricity price and availability
to enable matching demand with varying supply of electricity.
For the remainder of this document, we use "nanogrid" to refer
to those which use price to manage supply and demand.
Nanogrids bring an Internet approach and architecture to our
electricity system.
A nanogrid must have at least one load or sink of power (which could be electricity storage)
and at least one gateway to the outside.
Electricity storage may or may not be present.
Electricity sources are not part of the nanogrid, but often a source will be connected
only to a single nanogrid.
Interfaces to other power entities are through gateways within the nanogrid controller.
Nanogrids implement power distribution only and not any functional aspects of the devices
(or loads) that connect to the nanogrid.
Thus, the components of a nanogrid are a controller, loads, storage (optional), and gateways.
is a schematic of a nanogrid.
A nanogrid manages the power distributed to its loads.
All power flows are accompanied by communications and all communications are bi-directional.
Communication - either wired or wireless - is used to mediate local electricity supply
and demand using price, both within the nanogrid and in exchanges across the gateways.
The nanogrid controller receives requests for power, grants or revokes them,
measures or estimates power, and sets the local price.
Loads take the price into account in deciding how to operate.
Controllers negotiate with each other across gateways to buy or sell power.
Battery storage is optional - batteries can increase the reliability and stability of a nanogrid.
Controllers may resemble existing Power over Ethernet (PoE) switches, however unlike PoE
they need not be limited to one device per port.
To set the local price, the controller takes into account the price of any utility grid
electricity it has access to, as well as the quantity and price of any local power sources.
A nanogrid can exchange power with other nanogrids or with microgrids whenever mutually
beneficial (as indicated by relative price).
This enables optimal allocation of scarce and/or expensive power among loads and among
local grids.
A price will typically be a current price and non-binding forecast of future prices,
up to one day in advance.
Devices that connect to a nanogrid will ship with default price preference functions
that make sense given typical grid prices.
When a nanogrid is connected to the grid, the grid price will be a strong influence on
the local price, though local generation and storage can dramatically change that dependency.
When not grid-connected, the local price will reflect the local supply/demand condition,
the estimated replacement cost for battery power (which may be future grid power), and
an assessment of battery capacity.
Nanogrid policies establish the local price and load policies establish the price a
given load is willing to pay.
A core principle is to separate power distribution technologies from functional control technology.
Power distribution is envisioned to have three layers: layer 1 is power; layer 2 is power
coordination; and layer 3 is device functionality.
Nanogrids implement layer 2 to improve the efficiency and flexibility of power distribution and use
(layer 1), and isolate power distribution from device functionality (layer 3).
Separating power coordination from functionality has several purposes.
In future usage, devices that are in the same room or otherwise need to coordinate
functionally will often be powered differently, and devices that share a power
infrastructure may not have functional relationships.
Separating power distribution into different functional layers allows each function to
evolve separately, greatly easing and simplifying the development of new technologies and
deploying them alongside existing products.
To develop useful nanogrid technology we need
standards for communication internal to nanogrids, and for communication
between them via gateways.
Nanogrids use price to mediate their internal supply and demand with
attached loads, and to determine how power is acquired from external grids and
exchanged between nanogrids.
They require energy price information, common communications
protocols and interfaces, and standardized semantics.
Nanogrids could offer many benefits, broadly including:
Local RenewablesStorage and Reliability Security, Privacy, and ReliabilitySystem ReliabilityDemand ResponseSmart GridNew Electricity UsersDisaster ReliefMilitary ApplicationsReduced Capital CostsReduced Energy UseMobile and Off-GridNanogrids could provide smart grid benefits at the small (local)
scale, a capability we lack today; smart grid efforts only address grid
connected and large scale contexts.
Nascent nanogrids are common today in digitally managed forms (technologies
including USB and Power over Ethernet (PoE)), and unmanaged ones (vehicles,
emergency circuits, etc.).
However, they all lack the ability to use price
as the core prioritization mechanism and lack the ability to exchange power
with each other; a fully functioning "managed" nanogrid can do both.
Such future nanogrids could be connected in arbitrary and dynamic networks to
each other, to microgrids, and to the utility grid.
Nanogrids are a new mechanism for managing power at the local level,
useful in a wide variety of applications.
They particularly enable more and better use of local generation (including
intermittent renewables) and local storage, as well as facilitate "Direct DC" -
powering loads with local renewable power without converting to and from AC.
Recent studies have estimated 5-13% electricity savings from Direct DC in
residences [DIRECTDC], and local renewables also avoid transmission system losses.
Many people value local renewable energy more than grid power and value
the reliability and certainty of local storage and off-grid capability.
Nanogrids offer the possibility of moving to a less reliable large-scale grid,
providing increased quality and reliability locally, and saving capital and
energy in a distributed, bottom-up manner.
While the smart grid will better match supply and demand at the large scale,
we lack mechanisms to do this at small scales.
Nanogrids fill this gap.
Microgrids are important and necessary, but lack
near-term potential for dramatic scale-up of deployment, lack standards-based
plug-and-play technologies, lack comprehensive visibility into individual loads,
and lack pervasive use of price.
Nanogrids build on standard semiconductor and
communications technologies already produced at mass-scale, and can be deployed
incrementally and at low capital and installation cost.
This will enable them to spread rapidly and quickly become a standard fixture
in buildings.
While existing nanogrid technologies enable only relatively small
loads, there is no power limit to nanogrid loads or controllers.
While nanogrids work best with communicating loads, for legacy devices, with
one device per port, the controller can implement the load control function
itself for on/off loads, as well as variable loads like lights and motors.
By being directly and correctly responsive to the most local conditions of
energy supply, storage, and demand, nanogrids can provide price and other
control abilities not possible with other technologies which treat electricity
distribution at a more aggregated and abstracted level.
Nanogrids are also inherently more flexible and should be less capital-intensive
than alternatives, and provide a more nimble infrastructure for local generation
and storage.
The concept of power interface in the current EMAN drafts is consistent
with the interfaces that nanogrids have, both those from controllers to
loads, and those at gateways between nanogrids.
A load could report via EMAN protocols directly, or a controller could
report information about loads on their behalf; these are both basic
EMAN functions.
The role that batteries play in nanogrids is consistent with EMAN's treatment of them.
Nanogrids enable bi-directional exchange of power between grids;
recent versions of EMAN documents acknowledge this as a possibility and support it
(of course, the power flows in only one direction at any given time).
Two existing power distribution technologies, UPAMD and HDBaseT, support
bi-directional power flows.
Nanogrids have two characteristics that could be challenging for EMAN
to handle and deserve further consideration.
The first is that grids can be arranged in any topology and may
lack a single "root" as the utility grid generally provides.
The second is that connections among grids and connections to loads
may be intermittent and dynamic.
Accommodating these does not seem contrary to the goals of EMAN, but
EMAN semantics could be defined in a way which makes doing so
difficult or impossible.
Communication internal to a nanogrid will be specific to the particular
physical layer technology.
USB, for example, could add nanogrid capability by simply extending the
existing protocols it provides for coordinating power distribution on
USB links.
For PoE, it would be possible to do this with LLDP, or with some higher-layer
protocol.
Communication between nanogrids will require standards for gateways between
them that cover both electrical and communications aspects.
IEEE is a likely choice for at least most of this.
Some of these may benefit from using IETF protocols, though core to the
concept of local power distribution is that it only requires communication
between immediately adjacent (electrically-connected) grids - just one hop.
Whether or not the IETF is involved in power distribution protocols,
most of the devices in future that are on nanogrids, and the controllers
themselves, will likely also implement IETF protocols, so that semantic
consistency between the two domains would be extremely beneficial.
Just as EMAN provides visibility into device power (measurement and control)
at the network level, the IETF may want to in future support management
protocols for small (microgrid or smaller) grids (that is, not intruding
into the utility grid space where other standards organizations are active).
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119.This mechanism introduces no information security vulnerabilities.
A security advantage of nanogrids is that they only need to communicate
with other grids (or power sources) to which they are directly electrically
connected. This requirement for physical connection greatly reduces their
vulnerability, and is in sharp contrast to many grid architectures which
require communication across many network links.Nanogrid gateways need only communicate information about the price
and quantity of electricity, not about their internal structure or
electricity-consuming loads.
This makes them exceptionally protective of privacy.Key words for use in RFCs to Indicate
Requirement LevelsHarvard University1350 Mass. Ave.CambridgeMA 02138- +1 617 495 3864sob@harvard.edu
General
keywordIn many standards track documents several words are used to
signify the requirements in the specification. These words are
often capitalized. This document defines these words as they
should be interpreted in IETF documents. Authors who follow these
guidelines should incorporate this phrase near the beginning of
their document: The key words "MUST", "MUST NOT", "REQUIRED", "SHALL",
"SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and
"OPTIONAL" in this document are to be interpreted as described
in RFC 2119.Note that the force of these words is modified by the
requirement level of the document in which they are used.Nanogrids: Evolving our Electricity Systems from the Bottom UpLBNLFuture Roles of Milli-, Micro-, and Nano- GridsLBNLLBNLLBNLOptimizing Energy Savings from Direct-DC in U.S. Residential Buildings