| R. S. Bain, "NNES User's Manual", (1993). |
....submatrices [20] The block Newton iteration scheme can be combined with the chord method for a further reduction in computational complexity. However these methods can only be employed for almost linear circuits, since they present poor convergence properties. The globally convergent methods [13] are the line search methods and the trust region methods. These methods have better convergence properties when the initial guess is not close to the solution. The line search method expands on the strict Newton method. The point generated by equation 2.10 is not necessarily accepted as a new ....
....Fourier transform, and it is simpler to implement [17] One of the advantages is that the analysis remains essentially the same as with single tone excitation. CHAPTER 3. IMPROVED HARMONIC BALANCE 29 3.2. 3 Solution of the nonlinear system The nonlinear system is solved using the library NNES [13]. It is written in Fortran and it provides Newton and quasi Newton methods with many options, such as the use of analytic Jacobian or forward, backwards or central di#erences to approximate it, di#erent quasi Newton Jacobian updates, two globally convergent methods, etc. After formulating all the ....
R. S. Bain, "NNES User's Manual", (1993).
....in Equation (5.3) can be solved using di#erent numerical methods. Zhou et al. 34, 37] used a fixed point scheme. We have chosen to use Newton and quasi Newton methods because of the higher convergence rate and the global convergence properties of some variants such as the line search method [59]. There are two main issues in solving the nonlinear system. Obtaining a good initial guess and reducing the number of unknowns. If the circuit being simulated has a periodic excitation, the time window size can be chosen equal to the period. Then the solution for a given time window can be used ....
....CIRCUIT SIMULATOR 74 recently developed formal OO techniques and C features. The design intent was to to combine the advantages of previous OO circuit simulators with these new developments as well as expanding capability. Transim uses C libraries [72, 73] and several written in C or Fortran [52, 59, 74]. In the following the specific OO programming construction used in the current work is described. Then an example is presented integrating electromagnetic and circuit analysis in modeling a microwave CPW active antenna. 7.2 The Network Package The network package is the core of the simulator. ....
[Article contains additional citation context not shown here]
R. S. Bain, NNES user's manual, 1993.
....using some of the recently developed formal OO techniques and C features. The design intent was to to combine the advantages of previous OO circuit simulators with these new developments as well as expanding capability. Transim uses C libraries [13, 14] and several written in C or Fortran [15, 16, 17]. In the following OO circuit simulators are first reviewed and the specific OO programming construction used 2 in the current work is described. Then an example is presented integrating electromagnetic and circuit analysis in modeling a microwave CPW active antenna. 2 Background APLAC 2 [18, ....
....Numerical Toolkit 8 are available but the GNU gcc compiler version 2.8.1 and earlier used at the time to develop Transim are not capable of compiling them. Transim is now being developed using GNU gcc 2.95. 4. 3 Solution of nonlinear systems Nonlinear systems are solved using the NNES 9 [17] library. This package is written in Fortran and provides Newton and quasi Newton methods with many options including the use of analytic Jacobian or forward, backwards or central di#erences to approximate it, di#erent quasi Newton Jacobian updates, or two globally convergent methods, etc. This ....
R. S. Bain, NNES user's manual, 1993.
....they were implemented. 3.3.1 Packages used in the implementation In developing the frequency model, the Sparse 1.2 package [30] was used. This is a package that solves linear equations which have sparse matrices. The nonlinear error equation formulated in (3. 8) is solved using the NNES package [31]. At each time point the error function is calculated using the state variable vector that has an initial value. This initial value is set to zero at zero time, but at other times di#erent than zero it is set to the previous time solution to improve the convergence and to decrease the number of ....
R. S. Bain, NNES User's Manual, 1993.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC