|
Click
here to register for a FREE
trial copy of PlantWorkz
An efficient generation dispatch
solution can save a mid size power plant millions of dollars
per year in fuel cost savings alone. Historically unit
commitment & dispatch of power plants has been a difficult
problem. This is because unit commitment algorithms normally
use complex dynamic programming or mixed integer programming
techniques that must, in effect, search through a huge
space of possible schedules in order to find the best.
Incorporation of complex operating constraints (startup
costs, heat rate, maintenance scheduling, outages etc.),
fluctuations in fuel and power prices, and loads in dispatch
decisions stumble even the best programming techniques.
Let’s look at some of the common approaches in solving
the generation dispatch problem:
“Spread Sheet”
models: Fast but inaccurate
Many operators have developed in house “spread sheet”
models to optimize the operation of their power plants.
A typical “spread sheet” model may try to
model the difference between the price of electricity
and the price of fuel (normalized for a particular efficiency
of conversion from fuel to electricity), and make dispatch
decisions– known as “spark spread” models.
At a basic level, a generator resembles a spark-spread
model. When the price of power is above the generator’s
cost of production the generator can run profitably. Otherwise,
it can shut down and receive nothing. So, it is natural
to model a generator as a spark spread option where the
strike price is determined by the generator’s efficiency
in converting fuel to electricity.
Unfortunately, a spark-spread model ignores some important
aspects of typical generators, like the time-dependent
constraints on ramping and cycling, and time-dependent
costs associated with startup and cooling. For instance,
a generator can’t usually start up or shutdown as
often or as quickly as a spark spread option model would
imply. It can also be difficult to use a spark spread
option model to characterize the allocation of capacity
to ancillary services like regulation and spinning reserve.
(Some analysts embellish the spark spread option model
to account for some of these things, but in general it
is necessary to make compromises.)
Another criticism of spark spread option models is that
they are “reactive” – in essence, they
assume that a generator simply turns on or off in response
to the current price. In reality, the optimal schedule
for a generator must anticipate price changes, perhaps
incurring a loss in some periods in order to position
the generator to capture higher expected profits later
on. Again, this is related to the time-dependent constraints
and costs associated with typical generators.
Traditional unit commitment
models: Accurate but slow
In a traditional unit commitment model, a computer algorithm
is given a forecast of power and fuel prices, and it determines
a schedule for the generator that will maximize profitability
while adhering to all physical operating constraints,
and accounting for startup and other operating costs.
In addition, unit commitment and dispatch algorithms can
also be used to evaluate the financial properties of generation.
When unit commitment models are used in financial analysis-
e.g. what returns can I expect from adding a new duct-firing
unit to the current setup- would require inordinate amounts
of computing time for multi-year or multi-scenario studies.
Many of the solutions on the market today would require
more than an hour just to simulate a single five-year
scenario with hourly granularity – PlantWorkz would
take a second or less on a typical desktop. PlantWorkz’s
lightning speed allows a user to run multiple runs over
multiple years to simulate different scenarios without
investing in expensive hardware. More sophisticated financial
analysis, risk management, and portfolio optimization
(including multiple asset types such as service contracts,
financial derivatives, emissions credits and curtailment)
are also possible when PlantWorkz is used in an overall
EnWorkz software framework, known as EnWorkz PowerUpTM.
PlantWorkzTM : Accurate and
fast
PlantWorkz overcomes the accuracy/speed tradeoff by using
a new type of unit commitment algorithm. Developed by
a former computer science professor, the PlantWorkz unit
commitment algorithm exploits powerful heuristics and
algorithmic improvements that have been overlooked by
other researchers in the field.
The PlantWorkz generator model exhibits the accuracy normally
expected from a unit commitment model, but it runs about
1000 times faster than conventional unit commitment algorithms.
(See Table 1.)
The PlantWorkz generator model achieves high speed without
sacrificing completeness. It accounts for:
- ramp limits
- warmth-dependent startup times and costs
- uptime/downtime constraints
- multi-segment heat curves
- other common generator parameters.
|
Single-processor runtimes |
Table 1. PlantWorkz runtimes on a single 2.4GHz Pentium
4 processor for 5 generators.
PlantWorkzTM : Other advantages
- Rapid
Deployment – PlantWorkz is distributed
as a standard windows install file. It can be installed
to run on any regular Windows desktop PC.
- Supports
multiple Databases – Our object model
and persistence layer provides support for multiple
commercial databases, e.g. Oracle, SQL, Informix,
Sybase etc., and advanced features such as automated
database upgrades, data versioning, and data auditing.
For users without a commercial database, PlantWorkz
includes a license free database for immediate use.
This database requires zero administration.
- Rich
Look and Feel – PlantWorkz user interface
is straightforward and intuitive, it enables a productive
user experience with support for a rich look and
feel for desktop applications.
- Blinding
Operation Speed – PlantWorkz achieves
fast performance by using heavily optimized advanced
algorithms, by avoiding unnecessary layers, and
by object model caching. These techniques are recently
developed, and rely on an in-depth understanding
of chip performance design, and software interoperability.
- Cut and
Paste Data Between PlantWorkz and Excel Spreadsheet – For users who need to input data from an
Excel spreadsheet, or display results in an Excel
spreadsheet, PlantWorkz supports cut-and-paste of
data between the two applications. PlantWorkz also
supports import and export in the Excel XLS file
format.
- Scalable – For users who are considering using PlantWorkz
for financial analysis, portfolio optimization and
risk management, there is a clear migration path
to EnWorkz PowerUp.
|
Bottom line
Generation planning and scheduling is a difficult problem,
it is no wonder many vendors are asking for hundred of
thousands of dollars (if not more) for a deployed solution.
We at EnWorkz have developed a breakthrough unit commitment & dispatch technology in PlantWorkz that is part of
our overall software framework – EnWorkz PowerUp.
We offer EnWorkz PowerUp as an overall solution for your
financial analysis, portfolio optimization and risk management
needs; but we would like you to consider PlantWorkz as
a superior alternative to your existing “spread
sheet” or in house unit commitment & dispatch
solutions at an extremely affordable price. We believe
that this is a great opportunity considering the potential
millions you can add to your bottom line from fuel costs
savings alone! You don’t have to take our word for
it, just download it for a 30 days free trial, and you
will see the difference.
We are doing this for two reasons. One, we truly want
to help the industry become more profitable- this is a
good thing considering the current state of our industry.
Two, if you like what you see, you will have a reason
to look at the rest of our product suite.
Click here to register for a FREE
trial copy of PlantWorkz
Back
|
