Optimization approach for online identification of harmonic resonance due to pending Volt/VAr operation, An

Abstract

The emphasis on creating a more efficient distribution system has led many utility companies to employ dynamic voltage and VAr compensation (Volt/VAr) applications that reduce energy demand, generation, and losses associated with the transmission and distribution of energy. To achieve these benefits, Volt/VAr applications rely upon algorithms to control voltage support equipment, such as transformer load tap changers, voltage regulators, and capacitor banks. The majority of these algorithms utilize metaheuristic programming methods to determine the Volt/VAr scheme that produces the most energy efficient operating conditions. It has been well documented that the interaction between capacitor bank reactance and the inductive reactance of a distribution system can produce parallel harmonic resonance that can damage utility and customer equipment. The Volt/VAr controlling algorithms that account for harmonics do so in an indirect manner that can mask harmonic resonance conditions. Unlike previous research endeavors, the primary focus of the method described within this dissertation is to identify Volt/VAr schemes that prevent harmonic resonance due to capacitor bank operation. Instead of a metaheuristic approach, the harmonic resonance identification algorithm relies upon constrained mixed integer nonlinear programming (MINLP), which is more suited for analyzing impedance characteristics created by the energized states of a system of capacitor banks. Utilizing a numerical approach improves the accuracy of identifying harmonic resonance conditions, while also reducing the complexity of the process by exclusively relying upon the system's admittance characteristics. The novel harmonic resonance identification method is applicable to distribution systems that are dynamically reconfigured, which can result in a number of unknown harmonic resonance producing conditions, a feature unavailable with existing controlling algorithms. The ability to identify all harmonic resonance producing configurations based upon a required compensation level also provides a utility company with a means to determine if voltage support shall be purchased from an outside source. Documentation within this dissertation describes the engineering and mathematical theories that support the MINLP dependent harmonic resonance identification algorithm.