| The introduction of open electricity markets and the fast pace of changes brought by modern information technology bring both opportunities and challenges to the power industry. Vast quantities of data are generated by the underlying physical system and the business operations. Fast and low cost communications allow the data to be more widely accessed. For electric utilities, it is becoming clear that data and information are vital assets. Proper management and modeling of these assets is as essential to the engineering of the power system as is the underlying physical system. This dissertation introduces several new methods to address information modeling and data processing concerns in the new utility environment.; Presently, legacy information systems in the industry do not make adequate use of the data produced. Hence, a new information infrastructure using data warehousing---a data integration technology used for decision support---is proposed for novel management and utilization of data. Detailed examples and discussion are given on the schema building, extract transform and load (ETL) strategies for power system specific data. The benefits of this approach are shown through a new viewpoint of state estimation. Inaccurate grid information, especially topology information, can be a major detriment to energy market traders' ability to make appropriate bids. A two-stage DC state estimation algorithm is presented to provide them with a simpler data viewpoint to make knowledgeable trading decisions. Numerical results show how the results of a DC state estimator can be accurately made available to all concerned.; Additionally, the proposed communication and information infrastructure allow for new formulations and solutions to traditional power problems. In this vein, a new distributed communication model of the power system using publisher/subscriber paradigm is presented and simulated. The simulation results prove its feasibility and show it has adequate performance under today's communication technology. Based on this model, a new state estimation algorithm, which can decentralizes computations and minimizes communication overhead, is derived using a set of overlapping areas to cover the entire network. Numerical experiments show that it is efficient, robust, and has comparable accuracy as the conventional full network state estimation. |
