Studies on security evaluation by pattern recognition and unit commitment by dynamic programming in power systems

Technical Report · Mon Jan 01 04:00:00 EST 1973

OSTI ID:7339515

Pang, C K

The security evaluation problem is defined in the framework of pattern recognition. The proposed criterion (F-function) for the selection of variables as features has the advantage of being invariant as far as any scaling is concerned. It also provides a measure of the classification errors that could result when using a variable as the only feature. The single ranking method which renders a straight forward and quick procedure in selecting features is given. The first use of second order and multiple-function classifiers in security classification is reported. Training of classifier is done by the proposed optimal search algorithm which is adapted from a solution of linear inequalities. Simulation results on the CIGRE 225 kV system show that such a pattern recognition approach to steady state security evaluation is feasible. Additional uses of security functions obtained from pattern recognition are outlined and discussed. The effect of many of these uses cannot be achieved as easily by other existing methods. The solution to an optimal short term (about 24 hours) unit commitment problem by dynamic programming is developed. The unit commitment determined is optimal in the sense that total fuel cost is minimized. Total fuel cost includes generation and startup costs. Also, considerations are given to time-dependent start-up cost of generating units, network constraints, spinning reserves and security requirements. The introduction of linear time-dependent startup cost simplifies the computation. Computing time required is not excessive and its feasibility for use on actual systems is illustrated by an example.

• Research Organization: Purdue Univ., Lafayette, Ind. (USA). School of Electrical Engineering
• OSTI ID: 7339515
• Report Number(s): TR-EE-73-3
• Country of Publication: United States
• Language: English

Related Subjects

200301* -- Power Transmission & Distribution-- AC Systems
EHV & UHV-- (-1989)
24 POWER TRANSMISSION AND DISTRIBUTION
AC SYSTEMS
COMPUTERS
CONTROL
ECONOMICS
EHV AC SYSTEMS
ENERGY CONSUMPTION
FUEL CONSUMPTION
OPERATION
PATTERN RECOGNITION
PERFORMANCE
POWER DEMAND
POWER GENERATION
POWER SYSTEMS
POWER TRANSMISSION
PROGRAMMING
RELIABILITY
SIMULATION