Transcript MECE E4210 - Grids and Renewables

Grid and Renewables
SOME JARGON
Lots of issues of integrating
renewables
• Adequacy- there is simply not enough
generation to back-up variable generation
• Ramping- there is too rapid a drop
• Lots of solutions (demand response, real-time
optimal power flow, power electronics, smart
metering, storage etc etc)
• This lecture on where to locate renewables
Where to locate renewables
• BALANCE OF THESE FACTORS
• If one can locate say local wind resources
close to local demand AND if that resource
diurnal and seasonal profile is well matched to
the profile of demand that lowers cost
• Higher capacity factors lowers cost
• If one can use existing transmission capacity
then that lowers cost
• Current assignment tries to “optimize” these.
What we will not optimize for
• Reactive power
• What is it and how is it addressed by grid
• Some history
Why AC power
•
•
•
•
Transforming voltages was easier
Early generation was AC
Motor loads were AC
Three-phase: more thru same wire, easier to start
motors
• Why HVDC? Lower losses, long distances, but
expensive to tap power at intermediate locations,
legacy T&D systems is also AC
• Less issues of optimal power flow, DC backbones?
Optimal power flow- come to this later
but some intro
• Need to solve Kirchhoff’s laws for all nodes
• Multiple generators, loads…
• Multiple possible solutions, and to ensure
voltages within some band need to
inject/remove real/reactive power at different
locations.
• “optimal power flow” at any instant of time
• But loads are varying and new generation is
variable, multiple possible generators
• Each generator may have it’s own price
• Each load may have some price for it’s
demand response, so kind of like a generator
• Orchestrating which unit needs to come
on/off, which load comes on/off, and calling in
balancing, spinning reserves etc, doing this at
minimum cost and maint reliability role of ISO
Some background slides
DC
Pa same as Q in figure on right
• Reactive power flow is needed in an alternating-current transmission
system to support the transfer of real power over the network. In
alternating current circuits, energy is stored temporarily in inductive
and capacitive elements. AC connected devices that store energy in
the form of a magnetic field include inductors (a large coil of wire).
When a voltage is initially placed across the coil, a magnetic field
builds up, and it takes a period of time for the current to reach full
value. This causes the current to lag behind the voltage in phase;
hence, these devices are said to be sources of lagging reactive power.
• A capacitor is an AC device that stores energy in the form of an
electric field. When current is driven through the capacitor, it takes a
period of time for a charge to build up to produce the full voltage
difference. On an AC network, the voltage across a capacitor is
constantly changing – the capacitor will oppose this change, causing
the voltage to lag behind the current. In other words, the current
leads the voltage in phase; hence, these devices are said to be
sources of leading reactive power.
• Reactive power causes losses because it heats
up the lines….
• Instant power: real component when
integrated over one cycle is non-zero but
integral of the reactive component is zero
• Both the act of transmitting and running
motors etc leads to inductive loads
• Direct effect on system voltages
Steady-state
• Easier to deal with
transients
• Computational tools, do not scale with the size of
the system today
• So one resorts to simulations and hopes that the
system is either in one of stable trajectories or if
not, one can do something about it
• Deviations in sinusoidal behaviors and in voltages
• Solar and wind could introduce rapid transients
much faster than loads
Reactive Power Limitations
• Reactive power does not travel very far
• Usually necessary to produce it close to the
location where it is needed
• A supplier/source close to the location of the
need is in a much better position to provide
reactive power
Y versus one that is located far from the
location of the need
• Reactive power supplies are closely tied to the
ability to deliver real or active power
2
7
How Reactive Power Control
Implemented
• Regulate to control voltage to a desired
nominal value
• Often, reactive power injections regulate
voltage at the location of the injection
• Control effects tend to be localized
• Some reactive power supply mechanisms:
Y
Y
Y
Y
Shunt capacitors (fixed and switchable)
Synchronous condensers
Synchronous generators
Static VAR compensators
2
8
How Management of Reactive Power
Has Changed
• Under regulated environment, most utilities
owned/controlled G&T&D in its own control area
Y Provided reactive power just as it had to provide
sufficient generation and voltage
• Restructuring has changed this sometimes
causing problems dealing with reactive power
Y Merchant (non-utility) generation and related
financial incentives
Y Transmitting power over longer distances with
multiple transactions
2
9
What has Lead to Problems
• Regulated, electric rates based on kWh and kVA giving
incentive for pf correction
• Restructuring, separation of G&T&D businesses
Y Generation: More likely kW based from nonregulated generation removing incentive for pf
correction
Y Distribution: may not have significant incentive and
strict budget for installation of capacitors
Y Transmission: who will own and operate and thus no
incentive for improvement
• Electricity is transmitted between control areas
Y Has to be communication to properly operate the
system, including adjustments to reactive power.
Y ISOs (i.e., MISO) has not yet defined any system
rules concerning reactive power
3
0
•
•
•
•
•
•
•
•
•
The heart of economically efficient + reliable ISO power markets is AC optimal
power flow (ACOPF) problem. This problem is complex economically, electrically
and computationally.
Generators offer LBMP price/quantity pairs
System operator collects these offers and constructs cost functions at each node
Solves the optimal power flow, while balancing the loads, respecting transmission,
voltage, reliability constraints
After solving you get prices at EACH node, (through shadow prices of the
optimization prob) you pay uniform pricing at that node (can be at/above bid price)
Economically: efficient market equilibrium requires multi‐part nonlinear pricing.
Electrically, the power flow is AC, with nonlinearities.
Computationally, the optimization is a MI NLP, including both binary variables and
non-convex continuous functions, which makes the problem difficult to solve.
must be able to handle loss of any generator or transmission element, and the
system operator must make binary decisions to start up and shut down generation
and transmission assets in response to system events.
• For investment planning purposes, the problem needs binary investment
variables and a multiple year horizon.
• Where you put what and how much to invest
• Braess’s paradox
• Non-intuitive things can happen, need to careful how you expand
• Objective Function. They include minimizing generation
costs, maximizing market surplus, minimizing losses,
minimizing generation (equivalent to minimizing losses)….
optimal transfers.
• full ACOPF: model all constraints/controls with an objective
function of minimize cost would meet the objectives of
minimize (generator fuel costs, generation output, losses,
load shedding, control actions).
• objective functions and constraints are not algebraic or
differentiable, and that multiple solutions are likely to exist,
in particular when there are many reactive power controls
(such as switched capacitors, FACTS devices, or generators)
in network loops.
• Bus‐type. In P, Q, |V|, θ space, there are four quantities at each
bus: voltage magnitude (V), voltage angle (θ), real power (P), and
reactive power (Q). In a power flow solution without optimization,
buses are classified into three bus types: PQ, PV and slack. PQ
buses generally correspond to loads and PV buses to generators.
Generator buses are called PV buses because power and voltage
magnitude are fixed; load buses are known as PQ buses because
real and reactive power are fixed, that is, Pmin = Pmax and Qmin =
Qmax; slack or reference buses have a fixed voltage magnitude and
voltage angle.
• For a power flow to be solved, the slack bus needs to have
sufficient real and reactive power to make up for system losses and
hold the slack bus voltage magnitude at 1; for this reason, a bus
with a large generator is commonly chosen as a slack bus.
• HVDC systems based on the Voltage-Source
Converter (VSC) technology, which started to be
deployed during the last decade, exhibit
significant flexibility as they can control
independently the active and reactive power.
Such systems can prove helpful in the
maintenance of power system security. Due to
their fast response, they are able to undertake
corrective control actions in order to relieve
overloads and prevent voltage drops in case of
contingency
Security-Constrained Optimal Power Flow including Post-Contingency
Control of VSC-HVDC lines
Spyros CHATZIVASILEIADIS, Thilo KRAUSE, Göran ANDERSSON
• The EU Project IRENE-40 (www.irene-40.eu) aims to identify the
appropriate transmission expansion measures in order to achieve a
more secure, sustainable, and economically competitive European
power system. Within this context, simulations based on different
future generation scenarios are carried out. With respect to
security, the objective is to identify the appropriate reinforcements
in order to minimize the ‘cost of security’. Hourly generation and
load data for the years 2010, 2020, 2030, 2040, and 2050 are
provided for each EU-27 member state, Norway, and Switzerland
• Here, we study a single-hour snapshot of the year 2050. The
generation and load data are taken from the scenario RES, which
projects a high share of renewable generation in Europe (~80%) by
the year 2050.
Large transformers
Where to place wind
• Example New York
• Complex problem as you can see
• Even if we do not take into account issues of
reactive power and control, the first order
problem is that of optimizing locations so that
one can ensure supply is closer to demand
(both in time and space)
• That may mean even a lower capacity factor
wind farm may be cost-effective