1. Yes.
2. You have correctly calculated the sensitivity of total system real power 
loss to simultaneous and equal changes in PD and QD. Whether that is a 
reasonable sensitivity on which to base your solar placement decisions, you 
will have to decide/justify. It seems a bit strange to me, since your solar 
panels are not likely to be generating a VAr of reactive power for every W of 
real power they produce.
3. I don’t know where your formulas come from or what the notation means. The 
code computes real power loss sensitivities w.r.t real power injections and 
w.r.t. reactive power injections. It does not compute any reactive power loss 
sensitivities.

You need to decide precisely which sensitivity you are referring to when you 
say you want to use LSF to determine the order of placement. You can calculate 
almost any sensitivity numerically, and if it is a sensitivity of real power 
loss to real power (or reactive power) injections that you want, then 
t/t_get_losses.m includes code for computing them analytically (LSF) and 
numerically (numLSF). I would compute them analytically since the results will 
be more exact and the computation much more efficient. This is the relevant 
portion of code …

ri = ext2int(r);            %% results with internal indexing
[loss, fchg, tchg, dloss_dV] = get_losses(ri);
nb = size(ri.bus, 1);
nl = size(ri.branch, 1);
[ref, pv, pq] = bustypes(ri.bus, ri.gen);
J = makeJac(ri);
dL = real([dloss_dV.a(:, [pv;pq]) dloss_dV.m(:, pq)]) / ri.baseMVA;
LSFi = zeros(nl, 2*nb);
LSFi(:, [pv; pq; nb+pq]) = dL / J;  %% loss sensitivity factors in internal 
indexing

%% convert to external indexing
nb = size(r.bus, 1);
nl = size(r.branch, 1);
LSF = zeros(nl, 2*nb);
LSF(ri.order.branch.status.on, [ri.order.bus.status.on; 
nb+ri.order.bus.status.on]) = LSFi;

Hope this helps,

-- 
Ray Zimmerman
Senior Research Associate
B30 Warren Hall, Cornell University, Ithaca, NY 14853  USA
phone: (607) 255-9645

> On Feb 5, 2015, at 5:34 AM, nivedita arunachalam 
> <[email protected]> wrote:
> 
> <New Microsoft Office Word Document.docx>

Reply via email to