Thank u very much sir,i tried it by reducing step size, it brings close to nose
point.
thanks
From: "Abhyankar, Shrirang G." <abhy...@anl.gov>
Sent: Thu, 20 Aug 2015 21:30:43
To: MATPOWER discussion forum <matpowe...@list.cornell.edu>
Subject: Re: Jacobian at nose point
Sorry for the delay in replying.
Just to clarify, the continuation power does not stop actly⠡t the nose point,
but pretty close it. The condition for detecting nose point in runcpf is "lam
< lamprv" (line 303 in runcpf), i.e., lambda at the current continuation
step is less than
lambda at previous step. So, runcpf actually gives a close estimate of the
nose point which can be improved by reducing the continuation step size,
mpoption(ᣰf.step⬦lt;stepsize>). However, I do not think you can calculate an
exact zero determinant in MATLAB
for the power flow partly because of the way MATLAB calculates the determinant
(pivoted LU factorization) and the ill-conditioning of the Jacobian near the
nose point.
Shri
From: nilesh patel <nk2...@rediffmail.com>
Reply-To: MATPOWER discussion forum <matpowe...@list.cornell.edu>
Date: Tuesday, August 18, 2015 at 12:29 PM
To: matpower-l <matpowe...@list.cornell.edu>, MATPOWER-L
<MATPOWER-L@cornell.edu>
Subject: Jacobian at nose point
Sir,
At nose point in P-V curve using cpf, Jacobian matrix becomes singular.
i.e. determinant of Jacobian to be zero.
I am getting very high negative value of determinant of jacobian at nose
point. I have used following code:
define_constants;
mpopt = mpoption('out.all',0,'verbose',2,'out.bus',1);
mpopt = mpoption(mpopt,'cpf.stop_at','nose','cpf.step',0.03);
mpopt = mpoption(mpopt,'cpf.plot.bus',7,'cpf.plot.level',2);
mpcb = loadcase('case39'); % load base case
mpct = mpcb; % set up target case with
mpct.gen(:,[PG QG]) = mpcb.gen(:,[PG QG])*1.35
mpct.bus(:,[PD QD]) = mpcb.bus(:,[PD QD])*1.35
results = runcpf(mpcb, mpct, mpopt);
J=makeJac(results)
%%% determinant of J
det(J)
ans =
-1.0471e+126
Could you please suggest me what necessary steps I should follow to get
determinant zero?
Thanking you.
Nilesh Patel
