Digital Solution of Power-flow Problems by Newton's Method of Using a Hybrid Matrix

Document Type

Thesis - Open Access

1972

Degree Name

Master of Science (MS)

Department / School

Electrical Engineering

Abstract

The last decade and a half has witnessed dramatic developments in the application of digital computers for solving power-flow problems. Previously these problems were analyzed on the direct analog computers called a-c calculating boards. With the enormous growth of the interconnected power systems during this period of time, digital computers established a distinct advantage over the analog computers for such reasons as: (a) Their ability to analyze large-size systems (with such features as automatic tap setting, automatic area interchange control, and control of reactive constraints of generators). (b) Elimination of human error in reading data and recording information on the system diagram. (c) Accessibility and economy in making only a few changes from the base case. (d) Availability of additional information such as the total transmission loss by easy extension of the power-flow program. The power-flow problem can be solved by both direct and iterative methods. In fact, all the methods are iterative in the sense that the load flow problem involves the solution of a system of nonlinear equations. However, the so-called direct methods employ the direct solution of a related linear system in the iterative algorithm, whereas the iterative methods use a scheme of successive displacements such as Gauss-Seidel. Newton's method has an advantage over an iterative method because of its much faster (quadratic) convergence to a solution, thus saving computer time. The usual approach has been to use the bus admittance matrix for the network-defining equations. The purpose of this investigation has been to apply. Newton 's method for the solution of power-flow problems employing a hybrid matrix for the network-defining equations in order to confirm the possibility of affecting further saving in computer time. A sample 6-bus problem was solved on an. IBM 360 Model 40 computer with 1 28 K core memory with single precision programming for the precision indices of 1 x 10-3 and 1 x 10-S for real and reactive power mismatches at the busses. A double precision program was written for the precision index of 5 x l0-7. The hybrid matrix was formed by considering generator busses (1 and 2) as voltage-corrected and load busses (3 to 6) as current-corrected. Bus 1 is considered the swing bus.

Power (Mechanics)
Electric power distribution

application/pdf

70

Publisher

South Dakota State University

COinS