Dear Ray,

When I do an OPF with variable voltages I get different results while for
the runmarket I get the same results.  runmarket is done underlying an
OPF. What is the reason?

Best Wishes

Silvio Miceli


On Sun, Mar 25, 2012 at 03:28, Silvio Miceli <[email protected]>wrote:

> Dear Prof. Zimmerman,
>
> Thank you very much.
>
> Best Wishes
>
> Silvio Miceli
>
>
>
> On Sat, Mar 24, 2012 at 01:44, Ray Zimmerman <[email protected]> wrote:
>
>> As I mentioned before, in MATPOWER's OPF the voltages *are* optimization
>> variables. In other words, they are *output* values that depend on all of
>> the input parameters. If you want to treat the voltages at generator buses
>> as inputs (that you can set randomly according to some distribution), then
>> you will need to set the VMIN and VMAX values at that corresponding bus to
>> the desired value before calling the OPF.
>>
>> Something like this, where randomvoltage() is a function that returns a
>> voltage for each generator bus drawn from some specified distribution and
>> ig is the vector of indexes of generator buses ...
>>
>> mpc = loadcase('mybasecase');
>> for k = 1:10000
>>     V = randomvoltage();   % generate a random draw of voltages
>>     mpc.bus(ig, VMIN) = V;
>>     mpc.bus(ig, VMAX) = V;
>>     r = runopf(mpc);
>>     % save whatever outputs you want to collect
>> end
>>
>>  --
>> Ray Zimmerman
>> Senior Research Associate
>> 419A Warren Hall, Cornell University, Ithaca, NY 14853
>> phone: (607) 255-9645
>>
>>
>>
>>
>> On Mar 23, 2012, at 7:13 PM, Silvio Miceli wrote:
>>
>> Dear Johnny
>>
>> I want to do a probabilistic OPF by using Monte Carlo. So, I have to load
>> a distribution of voltage at the buses. I think MATPOWER is not able to
>> load variables as optimization variable. If you have an idea please let me
>> know.
>>
>> Best Wishes
>>
>> Silvio Miceli
>>
>>
>> On Fri, Mar 23, 2012 at 23:39, Johnny <[email protected]> wrote:
>>
>>> Silvio Miceli <[email protected]> writes:
>>>
>>> > Dear Shiyang,
>>> >
>>> > I also want to do it by Monte Carlo, but not loads. I want to have a
>>> > distribution of voltage as input. When I load loadcase it gives error.
>>> My mean
>>> > is this:
>>> > I want load variable voltage values in bus matrix. Is it possible?
>>> >
>>>
>>> Silvio,
>>>
>>> I believe the best way to achieve this is to generate the normal
>>> distibution in a script which calls the OPF solver and collect the
>>> results to display as a distribution. So, in the Monte Carlo, you use
>>> stochastic input on a Black Box model and collect stochastic
>>> output. This normally takes some time, but what do you look for really?
>>>
>>>
>>>
>>> > Best Wishes
>>> >
>>> > Silvio Miceli
>>> >
>>> > On Fri, Mar 23, 2012 at 23:19, 李诗旸 <[email protected]> wrote:
>>> >
>>> >     HI Silvio Miceli,
>>> >
>>> >     I think Mr. Zimmerman’s suggestion is the way like Monte Carlo
>>> Method.
>>> >     Normally, you may have the distributions of some parameters
>>> (inputs) such
>>> >     as loads and active generations, then you could use that way to
>>> get the
>>> >     distributions of the optimized voltages (output) numerically.
>>> >
>>> >
>>> >
>>> >     Sincerely
>>> >
>>> >     Shiyang Li
>>> >
>>> >
>>> >
>>> >     发件人: [email protected] [mailto:
>>> >     [email protected]] 代表 Silvio Miceli
>>> >     发送时间: 2012年3月24日 4:31
>>> >     收件人: MATPOWER discussion forum; Ray Zimmerman
>>> >     主题: Re: probabilistic voltage
>>> >
>>> >
>>> >
>>> >     I want to a have a distribution of voltage at buses that it is an
>>> >     optimization variable in MATPOWER? These distributions are the
>>> inputs for
>>> >     the OPF.
>>> >
>>> >     Best Wishes
>>> >
>>> >
>>> >
>>> >     Silvio Miceli
>>> >
>>> >     On Fri, Mar 23, 2012 at 21:25, Ray Zimmerman <[email protected]>
>>> wrote:
>>> >
>>> >     I'm afraid I don't know what *you* mean. The voltages are
>>> *outputs* of
>>> >     MATPOWER's OPF. So what are the input parameters that you would
>>> like to be
>>> >     stochastic? Voltage limits?
>>> >
>>> >
>>> >
>>> >     --
>>> >
>>> >     Ray Zimmerman
>>> >
>>> >     Senior Research Associate
>>> >
>>> >     419A Warren Hall, Cornell University, Ithaca, NY 14853
>>> >
>>> >     phone: (607) 255-9645
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >     On Mar 23, 2012, at 2:40 PM, Silvio Miceli wrote:
>>> >
>>> >     My problem is this I cannot do loadcase. How can I do many, many
>>> >     deterministic optimization? If I introduce voltage as x and do it
>>> with a
>>> >     FOR?
>>> >
>>> >
>>> >
>>> >     Could you please explain better? I didn't understand. I am sorry.
>>> >
>>> >     Best Wishes
>>> >
>>> >
>>> >
>>> >     Silvio Miceli
>>> >
>>> >     On Fri, Mar 23, 2012 at 19:22, Ray Zimmerman <[email protected]>
>>> wrote:
>>> >
>>> >     MATPOWER does not implement any stochastic optimization. All of
>>> it's
>>> >     variables and parameters are assumed to be deterministic. However,
>>> one
>>> >     approach to handling probabilistic parameters is to simply do
>>> many, many
>>> >     deterministic optimizations, where each one involves, for each
>>> >     probabilistic parameter, drawing a value from the corresponding
>>> >     distribution. Such an approach could use MATPOWER's OPF as a
>>> subroutine to
>>> >     solve the deterministic sub-problem.
>>> >
>>> >
>>> >
>>> >     --
>>> >
>>> >     Ray Zimmerman
>>> >
>>> >     Senior Research Associate
>>> >
>>> >     419A Warren Hall, Cornell University, Ithaca, NY 14853
>>> >
>>> >     phone: (607) 255-9645
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >     On Mar 23, 2012, at 12:42 PM, Silvio Miceli wrote:
>>> >
>>> >     Dear Dr. Zimmerman,
>>> >
>>> >
>>> >
>>> >     I want to have probabilistic voltage or reactive power in an OPF.
>>> In other
>>> >     words, I want to do OPF with probabilistic variables,e.g. voltage
>>> should be
>>> >     a normal distribution of voltage. How can insert this kind of
>>> variables to
>>> >     case file of MATPOWER? Is there this possibility in MATPOWER?
>>> >
>>> >
>>> >
>>> >     Best Wishes
>>> >
>>> >
>>> >
>>> >     Silvio Miceli
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >
>>>
>>> --
>>> Johnny
>>>
>>>
>>>
>>
>>
>

Reply via email to