abstract: In the deregulated power system, locational marginal prices are used in transmission engineering predominantly as near real-time pricing signals. This work extends this concept to distribution engineering so that a distribution class locational marginal price might be used for real-time pricing and control of advanced control systems in distribution circuits. A formulation for the distribution locational marginal price signal is presented that is based on power flow sensitivities in a distribution system. A Jacobian-based sensitivity analysis has been developed for application in the distribution pricing method. Increasing deployment of distributed energy sources is being seen at the distribution level and this trend is expected to continue. To facilitate an optimal use of the distributed infrastructure, the control of the energy demand on a feeder node in the distribution system has been formulated as a multiobjective optimization problem and a solution algorithm has been developed. In multiobjective problems the Pareto optimality criterion is generally applied, and commonly used solution algorithms are decision-based and heuristic. In contrast, a mathematically-robust technique called normal boundary intersection has been modeled for use in this work, and the control variable is solved via separable programming. The Roy Billinton Test System (RBTS) has predominantly been used to demonstrate the application of the formulation in distribution system control. A parallel processing environment has been used to replicate the distributed nature of controls at many points in the distribution system. Interactions between the real-time prices in a distribution feeder and the nodal prices at the aggregated load bus have been investigated. The application of the formulations in an islanded operating condition has also been demonstrated. The DLMP formulation has been validated using the test bed systems and a practical framework for its application in distribution engineering has been presented. The multiobjective optimization yields excellent results and is found to be robust for finer time resolutions. The work shown in this report is applicable to, and has been researched under the aegis of the Future Renewable Electric Energy Delivery and Management (FREEDM) center, which is a generation III National Science Foundation engineering research center headquartered at North Carolina State University. / Dissertation/Thesis / Ph.D. Electrical Engineering 2012

Management and supply of water in an urban water distribution system is a complex process, which include various complexities like pressure variations across the network depending on topography, demand variations depending on customers’ requirement and unaccounted water etc. Applying automatic control methods to water distribution systems is a way to improve the management of water distribution. There have been some attempts in recent years to develop optimal control algorithms to assist in the operation of complex water distribution systems. The difficulties involved by these hydraulic systems such as non-linearity, and diurnal demand patterns make the choice of a suitable automatic control method a challenge. For this purpose, this study intends to investigate the applicability of different controllers which would be able to meet the targets as quickly as possible and without creating undue transients. As a first step towards application of different controllers, PD and PID linear controllers have been designed for pump control and valve control in water distribution systems. Then a Dynamic Inversion based nonlinear controller has been designed by considering the non-linearities in the system. Here, different cases considering the effects of initial conditions used, linearization methods used, time step used for integration and selection of gains etc., have been studied before arriving at best controller. These controllers have been designed for both the flow control problems and level control problems. It is found that Dynamic Inversion-based nonlinear controller outperforms other controllers. It is well known that the performance of controllers is much dependent on the tuning of the gains (parameters). Thus in this study various alternative techniques such as Ziegler--Nichols rules (ZNPID), Genetic algorithms (GAPID) and fuzzy algorithms (FZPID) have been studied and a comparative study has been made Although with all the three gain tuning methods, required states have reached their target values, but the responses vary much in reaching to final targets. The self-tuned FZPID controller outperforms other two controllers, especially with regard to overshoots and the time taken to tune the gains for each problem. Further, an optimal DI controller is developed for the over determined case with more controls and less targets. Energy loss is considered as an objective function and normal DI controller equations are considered as constraints. Hence, an attempt is made to reduce the energy minimization in water distribution system by formulating an optimal control problem using optimal Dynamic Inversion concept. Finally, leakage reduction model is developed based on excessive pressure minimization problem by locating valves optimally as well as by setting valves optimally. For this purpose, optimization problem is solved using Pattern search algorithms and hydraulic analysis is carried out using EPANET program.