GridPACK Application Concept

From GridPACK
Revision as of 20:58, 10 December 2012 by Dchassin (talk) (Created page with 'This document describes the GridPACK application concept for the various solution methods available __TOC__ [[image:GridPACK_Application_Concept_Figure_1.png|thumb|Figure 1 - P...')

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

This document describes the GridPACK application concept for the various solution methods available

Figure 1 - Protocol and data model layers of GridPACK

GridPACK provide a family of libraries that are necessary to implement power flow applications. In general, these libraries are packaged use a set of protocols and data models that transform the problem from a concrete power model to an abstract numerical solution, as shown in Figure 1.


Steady State Powerflow

Steady state powerflow analysis is method of solving a general power system problem where the currents and voltages on the busses (nodes) and branches (lines) of an electric network are computed based on the impedances, current and power injections at nodes (if any), the boundary voltages (if any) and topology of the network.

Power system applications provide this information to a solver in the form of a matrix of bus admittance (called Y), bus injections (called S) and boundary conditions (e.g., V).

Because the problem is essentially non-linear, various iterative solution methods are employed to obtain the solution to this flow problem. The include

Gauss-Seidel (GS) 
This method is the simplest to describe and implement, is guaranteed to converge for any network for which a solution exists, but converges quite slowly. It is sometimes used to initialize other faster solution methods that sometimes can converge reliably (such as Newton-Raphson). The method is implicitly parallel.
Newton-Raphson (NR) 
This method requires the a Jacobian be computed and maintained for the current system. It is not guaranteed to converge but when it does converge it is quite fast. The method is not readily parallelized.
Forward-Backsweep (FBS) 
This method works only on radial flow models but is extremely fast and can be readily parallelized for more networks.
Conjugate-Gradient (CG) 
Yousu to summarize this method.

Dynamic Simulation

State Estimation