OPTIMAL STATE ESTIMATION OF ELECTRICAL POWER TRANSMISSION SYSTEMS

ELECTRIC POWER TRANSMISSION SYSTEMS ARE LARGE ELECTRICAL NETWORKS THAT DELIVER ELECTRICITY PRODUCED BY THERMAL AND HYDROELECTRIC GENERATORS. BECAUSE THESE SYSTEMS ARE SO LARGE AND DISPERSED, IT IS EXTREMELY DIFFICULT TO KNOW HOW THE SYSTEM SI BEHAVING AT ANY MOMENT IN TIME. IN SPITE OF EXTENSIVE METERING AND POWERFUL COMPUTERS, THE BEST RESULT IS TO ESTIMATE THE STATE OF THE POWER SYSTEM. SINCE CERTAIN ELECTRICAL LAWS THAT CONSTRAIN AND GOVERN ITS BEHAVIOR MUST BE OBEYED, THE STATE ESTIMATION PROBLEM IS A VERY LARGE OPTIMIZATION PROBLEM THAT SEEKS TO DETERMINE THE MOST LIKELY ELECTRICAL STATE OF THE NETWORK GIVEN THE DATA AVAILABLE. CURRENT POWER SYSTEM STATE ESTIMATORS SOLVE A SIMPLIFIED OPTIMIZATION PROBLEM THAT IGNORES CERTAIN ASPECTS OF A REAL UTILITY'S PRACTICE. FOR EXAMPLE, COMMON SYSTEM DEVICES SUCH AS TRANSFORMER TAPS, PHASE SHIFTERS, OR SERIES CAPACITORS ARE EITHER TREATED AS FIXED VARIABLES OR ARE HEURISTICALLY ACOMMODATED IN THE OPTIMIZATION PROCESS. PHYSICALLY DISCRETE VARIABLES AND INEQUALITY CONSTRAINTS IMPOSED BY PHYSICAL OR OPERATIONAL LIMITS ARE IGNORED. CURRENT SOLUTION METHODS ARE BASED ON FIRST DERIVATIVE INFORMATION AND IGNORE THE IMPORTANT EFFECTS OF SECOND DERIVATIVES WHICH CAN HAVE A PROFOUND IMPACT ON COMPUTATIONAL ROBUSTNESS. THESE SHORTCOMINGS RESULT IN AN INDUSTRY-WIDE DISSATISFACTION WITH THE CURRENT STATE-OF-THE-ART, IN SPITE OF SIGNIFICANT GAINS IN COMPUTATIONAL SPEED MADE BY ALGORITHIC RESARCH AND FASTER COMPUTERS. RESEARCHERS ARE CONSTRUCTING A STATE ESTIMATION MODEL THAT DIRECTLY INCLUDES ESSENTIAL CHARACTERISTICS THAT ARE NOW IGNORED, WITH THE PREMISE THAT A MORE ACCURATE AND ROBUST MODEL IS OF GREATER CONCERN THAN A FASTER SOLUTION METHOD. A SOFTWARE PROTOTYPE FOR NUMERICAL EXPERIMENTATION SERVES AS A BENCHMARK DEBUGGER FOR COMMERCIAL SOFTWARE DEVELOPMENT.