POWER NETWORK RESTRUCTURING AND PRICING

Abstract

Since inception, the worldwide growth of power industry had been monopolistic in nature. Over the last couple of decades, itwas being observed that the government controlled and regulated monopolies have not been able to fulfil the growing electricity demand at affordable rates. Competition was recognised as a means to achieve cost efficiency and efficacy. This led to the initiation of restructuring process in this integrated industry. Restructuring of power systems, in the present form, includes vertical unbundling of the integrated utilities into separate generation, transmission, distribution and retail components. Corporatisation and infusion of private equity has been adopted as means to reduce the market power of existing utilities. Restructuring also involves the establishment of open market structure and power exchanges, to provide for free trade of electricity. In the restructured systems, generation and retail functions operate under competitive conditions. The high fixed cost and nature of the network industry has necessitated it to retain its monopolisation. Competitive conditions can be created by providing open network access to the consumers for contracting power from the supplier of their choice. This requires network independence, and along with that a fair system of network access. The scenario calls for regulatory control over network access. Regulation aims to establish network access on a non-discriminatory basis, which necessitates the development of fair network access tariff mechanisms. This is a new issue for the industry that has been historically vertically integrated, as earlier there was no trade of network services, and consequently no need for their pricing mechanism. Restructuring of the power systems and the evolving market structure, has thus brought into focus, cost allocation and pricing mechanism for these networks. Many countries around the world have initiated power sector reforms and restructured their power industry. The thrust is to develop competition and create efficiency. As for other countries, India also has implemented a slew of legislative and policy reforms. These include the Electricity Act 2003, National Electricity Policy 2005 and Tariff Policy 2006. These aim to restructure the vertically and horizontally integrated power industry to an open market, multi-buyer, multi-seller structure associated with an independent regulatory framework. These reforms have made an impact on the market structure and subsequently on the network pricing models. During this evolving market scenario, network pricing mechanisms are experiencing a series of issues relating to fairness and implementation. The issue of developing charging mechanisms for using a monopolistic infrastructure is a well known theoretical problem in economics. Various solutions, ranging from marginal to average costing, have been proposed for the same. The problem for electricity networks however, is more complex than other infrastructural problems. Abstract Electricity network pricing pursues a number of complementary objectives. The stress on different objectives varies with the adopted market model and the regulatory scenario. This makes the task of developing universal ly acceptable pricing mechanisms difficult. Generally, the first stage of network pricing involves deciding global remuneration for the utility, while the second stage allocates the cost among the users by a specified pricing mechanism. The present pricing mechanisms for distribution utilities do not consider the enormous diversities of the network. This is a stumbling block for pricing analysis. Offering consistent cost allocation for utilities performing in different areas, in a variety of asset replacement cycles, and serving a customer variety with different nature of distribution circuits and inherited network characteristic, is thus an uphill task. In such a situation, the concept of reference network can be used for distribution network pricing. Reference network development involves assimilation of actual network data, determination and modelling of major disaggregation parameters, classification of actual feeders into coherent groups, construction of the representative feeder for each group, and evaluation of reference network for its compatibility with the actual network. In a reference network, each representative feeder is supposed to be the best fit for a specific group of actual feeders. The present work focuses on deployment of an indigenous methodology to develop and validate a reference network fora large distribution network of an Indian utility. The system under study represents a large district in India having vast potential for wind and solar power generation. The network is distributed over an area of 22,850 square kilometres, inhabited by 3 million people, and serves 338 thousand customers. This system has 220 kV, 132 kV and 33 kV sub-stations and lines, along with eleven hundred 11 kV and some 33 kV load feeders. Data obtained from the transmission and distribution utilities is correlated to develop a comprehensive network database. With an understanding of the actual network and network planning, a skeletal network is developed. All the network feeders are classified based on the disaggregation parameters of connection voltage level, point of connection, type of feeder and average distance between laterals. This results in a small set of 30 feeders that have been topographically adjusted to provide the reference network. The developed network is a weakly-meshed radial network with various levels of bifurcations. The location, voltage level, peak load, power factor and category-wise energy consumption characterise the feeder loads. Feeders associated with the reference network are characterised by their voltage, capacity and impedance ratings, along with their length, connected customers and number of circuits involved. The transformers are specified by their voltage, capacity and impedance ratings. Reference network development is validated with the help of comparative performance evaluation of overall network load and transformer loadings. The costing analysis for the network is also provided. Abstract With the growing complexity of networks and increased number of transactions occurring , due to deregulation, development of a fair pricing model has become a contentious issue. Embedded cost of networks forms a large part as compared to their incremental cost, and hence draws much interest. The present thrust of embedded cost allocation approaches has been on model development based on individual network usage pattern. Objectivity of aim makes it difficult to identify the fairest criterion on which a network usage model should be based for any specific network condition. For users with different perspectives, the traditional methods may seem not to provide well-justified and fair solutions. Power network development can be assumed to be a result of cooperation between the utilities. This results in an infrastructure benefitting from "Economics of Scale", thus necessitating allocation of cost savings among the utilities. The situation can be represented as a cooperative game, with the loads and generators being players in the game. This thought has induced the application of cooperative game theory for network embedded cost allocation. In this approach, users make rational decisions in a competitive situation, wherein each participant aims at reaching an outcome that is advantageous to him, to the maximum possible extent. The cooperative game approaches of Shapley value and nucleolus provide unbiased and stable solution to a cost allocation problem. Considering these factors, cooperative game theory algorithms have been developed for embedded cost allocation of power networks relating to various scenarios involving a 3- bus, 6-bus, 11-bus practical transmission system, and a 6-bus rural distribution system. A novel cooperative game algorithm for the transmission network cost allocation between loads and generators has been developed. Shapley value and nucleolus based cost allocation approaches have been appl ied for the same. Another algorithm has been proposed for the validation of cooperative game based cost allocation of distribution networks. The approach is applied on a distribution system for cost allocation between loads. As compared to transmission systems, cost allocation for the distribution system has less difference between the traditional and game approaches, primarily due to lack of reverse power flow. In another analysis, the impact of incorporating distributed generators of different capacities, on cost allocation to loads and generators has been studied. The game approaches offer lower cost allocation for the distributed generators, when compared to the allocation from the traditional MW-Mile method. These justified solutions promote the development of distributed generators. Thereafter, a novel algorithm is proposed for understanding the implications of distributed generator placement on distribution network cost allocation. This allocation considers the impact of varying load and generation profile, on allocation reflected at network buses. The game models offer stronger locational signals as compared to the MWiii Abstract mile method, when the placement impact of solar and wind distributed generators is assessed. The traditional cooperative game approaches are based on peak loading by players. It is unrealistic to assume equal network usage by the loads and generators having equal peak capacities but varying load/generation profiles. The traditional game approaches fail to recognise this inherent probabilistic nature of players. Four novel probabilistic approaches based on Shapley value and nucleolus models have been developed for cost allocation. The approaches are broadly based on the existence probability of players and coalitions in a game. The value of playing a particular role in a game is assessed based on a probability factor ascertained by accounting for practical capacity and load factor considerations. For evaluating allocations due to probabilistic Shapley value and probabilistic nucleolus approaches, the characteristic values are based on the existence probability of the particular coalition causing that power flow. Generalised Shapley value provides a weighted sum of the marginal cost, arising from the fact that the probability of a player joining a coalition is the product of existence probability of a particular coalition already existing and the probability of the new player joining in. The method believes that all orders of players are not equally likely to occur, and thus their contribution must depend on the existence probabilities of various players. Banzhaf value approach also assigns an expected marginal allocation to the players, the allocation arising from the subjective belief that a player is equally likely to join any coalition of a specific size. The probabilistic approaches are tested on a 2-bus, 2-generator and a 6-bus, 3- generator system for cost allocation between loads and generators. The generalised Shapley value offers the strongest reflection of relative usage probability of loads and generators. But this method offers the least stable solution among the probabilistic approaches. All other probabilistic approaches offer largely stable results, though less stable than the normal nucleolus approach. Considering these facts, probabilistic Shapley value and probabilistic nucleolus are understood to offer better overall results. The probabilistic approaches are reflective of practical conditions, as incorporation of probabilities reflects actual energy usage rather than indicative peak usage. Thus, these models provide an equitable and economically justified cost allocation. The proposed methodologies help to understand, how the choice of a way to measure game values, affects the stability of allocations obtained. Though the field of network pricing is a vast area of study, the present work tries to fill-in some research gaps. The subject matter addressed is relevant for those interested in liberalisation and restructuring process of power sector in various countries. The issues handled in this work would continue to attract attention for a long time.