الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 127 |  |  |
| | | from | 127 | | |
Abstract
The problem of optimal load flow in power systems is considered. Based on Carpentier’s formulation, the pt’oblem is re¬formulated to achieve a reduction in the number of system variables, by treating all generator nodes as swing busses. Reduced system variables are also classified into control and dependent variables. The set of equality constraints is associated with the objective function using Lagrangian multipliers, and functional inequality constraints are included as penaities.
The set of control variables is iteratively adjusted, using Newton’s method, to minimize the objective function, while the set of dependent variables is eva1uated by solving the set of equality constraints after each such adjustment.
The reduction in the number of system variables and their classification as such provide a great saving in computer storage requirement as compared to other established methods. Furthermore,
the use of Newton’s method provides an excellent convergence behaviour.
The developed minimization algorithm is applied to the well known minimum generation cost and minimum system loss problem. In addition, the minimum fuel consumption, and combined fuel-cost minimization problems are defined and solved.
An optimal ordering scheme of system nodes, for use with large systems, is also developed and compared to two other effective schemes. The new scheme proved to be generally comparable to both schemes for the cases studied.
The concept of fixed penalty factor, developed in assocation with the minimization algorithm is investigated and compared to
the usual concept of monotonically increasing sequence of penalty factors. The former proved to be superior from a practical point of view.