<p> Electricity has become an invaluable commodity for the rest of humanity such that nations irrespective of their classification in the world economy will find it difficult to function without it&rsquo;s reliable supply. For nations such as the United States and the rest of the developed world, sustainable electricity supply is no longer optional. It has become a race for survival and maintenance of the very fabrics of those societies that made them who or what they are. So, whenever there is a disruption of electricity supply due to major natural disasters, the electric utility industry in the United States marshal thousands of first responders. These first responders always answer to the call of duty to face the challenge of restoring this valuable service to affected communities within the shortest possible time. In addition to the human element, electric grid restoration methods after disasters have depended mainly on the ability of intelligent electronic devices (IEDs) to communicate vital grid information with each other for system status. At one end are field devices and at the other end are human operators through outage management systems (OMS) with considerable command and control capabilities using Supervisory Control and Data Acquisition (SCADA) processes. Traditional use of centralized SCADA for system restoration during natural disasters takes too long and presents serious constrains on field workforce especially those on mutual assistance. In this study, we present a hybrid multi agent system (MAS) form of electric grid disaster response management that decentralizes the SCADA functions. The proposed system forms a Mobile Coordination and Restoration Center (MCRC) model that allows the different restoration agents the autonomy to execute restoration functions per outage demand after a disaster. The choice of agent location is modelled on the concept of Facility Location and Relocation Problem &ndash; under Uncertainty (FLRP-U) to identify optimum grid nodes that minimize distance travel and response time for field restoration crews. The model considers a dynamic approach that identifies agent locations based on outage demand changes and minimizes the total weighted distance for first responders. Using systems engineering (SE) concepts, an encompassing viewpoint is presented. The resulting architecture will examine the different agents and subsystems to help establish a technical framework that is logistical for future electric utility disaster response managers. This could be adopted by disaster managers in different settings to achieve improved restoration performance.</p><p>

<p> Demand response (DR) is considered one of the most reliable and cost-effective solutions for smoothing the electric demand curve of systems under stress. DR programs encourage customers to make changes in power consumption habits in response to electricity price incentives. A well designed autonomous scheduling system for households that are part of the smart grid can result in numerous benefits to all the players in the electricity market. </p><p> Distribution intelligence can be used to anticipate and moderate electricity usage, resulting in lowered production costs. When using this communication network, each entity may send and receive local and global data in a timely fashion, enabling customers to monitor their own electricity usage. Within a smart home, the energy management system is connected to smart appliances, thermostats, and other devices via a home area network (HAN). The HAN balances the electricity demand within the household and prioritizes between appliances and electric devices to modulate electricity usage and to ultimately reduce costs. </p><p> With a collection of rich and timely data, players in the power system can make better decisions to improve reliability, to optimize energy usage, and to reduce energy costs for themselves and for the system. Advanced metering infrastructure (AMI) creates ample opportunities to effectively address peak demand periods using pricing incentives, such as in DR programs and time-of-use (ToU) pricing, which ultimately reduce utilities operating costs. Electricity usage is thus reduced during peak hours with appliances and devices operating at other times, ensuring that electricity production is more evenly distributed throughout the day. </p><p> This dissertation presents a smart home energy management system (SHEMS) using a limited memory algorithm for bound constrained problems known as L-BFGS-B, along with time-of-use (ToU) pricing to optimize appliance scheduling in a 24-hour period. The allocation of energy resources for each appliance is coordinated by a smart controllable load (SCL) device embedded in the household's smart meter. SCL guarantees automation of the proposed SHEMS and prevents manual participation of customers in demand response (DR) programs. The model is simulated on a population of 247 residential prosumers with solar photovoltaic (PV) systems based on 15-min interval electric load data from a residential community in Austin, TX. After clustering households based on their electricity profiles, the proposed optimization model is performed. Simulation results showed that the proposed autonomous scheduling system reduced cumulative energy consumption for customers across the different clusters. In addition, when households were grouped based on their respective category according to the ToU pricing scheme, the simulation reported a notable decrease in total energy consumption from 65.771 kWh to 44.295 kWh; as well as a reduction in the cumulative cost of energy from $6.550 to $4.393 per day. Simulation results confirmed that the proposed algorithm effectively improved the operational efficiency of the distribution system, reduced power congestion at key times, and decreased electricity costs for prosumers.</p><p>