Message-ID: <22232469.1075856566564.JavaMail.evans@thyme>
Date: Mon, 31 Jul 2000 05:09:00 -0700 (PDT)
From: tom.halliburton@enron.com
To: jobrien@snowyhydro.com.au
Subject: SDDP Model
Cc: vince.kaminski@enron.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Bcc: vince.kaminski@enron.com
X-From: Tom Halliburton
X-To: jobrien@snowyhydro.com.au
X-cc: Vince J Kaminski
X-bcc: 
X-Folder: \Vincent_Kaminski_Jun2001_5\Notes Folders\Australia
X-Origin: Kaminski-V
X-FileName: vkamins.nsf

John,
Attached is a copy of the SDDP Methodology manual, written by the 
developers.  The model was developed by Mario Pereira, who is the proprietor 
of PSRI (Power Systems Research Inc) in Rio de Janeiro.   Their web site is: 
www.psr-inc.com where you can find more information about the company and 
contact details.  Mario and most of the staff speak English, so phone calls 
are possible.  

Development of SDDP was originally done on a World Bank contract to compare 
the benefits of new generation and transmission system interconnections in 
Central America.  It has since found much wider application.  The code runs 
on a PC, and is licensed to a number of consultants, power companies, etc.  
While I was with ECNZ in Wellington, New Zealand, we purchased a copy around 
1994.

The model works with weekly or monthly time steps, with a time horizon of 
from 1 to 10 years, for mid to long range planning of hydro thermal power 
systems.  It represents loads in each period as from one to five load blocks, 
so it is not a chronological model.  This type of model can not handle 
thermal plant ramp rates and similar chronological system constraints.  I'm 
not sure how important that would be to you.  The optimization phase of the 
model takes into account hydrological uncertainty using a form of stochastic 
dynamic programming with sampling.  This phase sets up a function for water 
value in each large reservoir in the system.  The simulation phase uses a 
number of hydrological outcomes to collect statistics on how the system would 
operate.  The water values can be used with the short term optimization mode 
of SDDP or in some other more detailed short term optimization model.

Most constraints found in hydro systems can be modelled.  Many transmission 
constraints can be modelled.  A DC load flow option can be used to determine 
bus nodal prices.

A Visual Basic user interface is provided which reads and writes text files.

If you'd like to know more about SDDP, you could call me here in Houston. 
(tel 001 713 345 8539)

yours

Tom Halliburton