Due to energy scarcity coupled with environment issues, it is likely to see the biggest shift in generation portfolio in the UK and world wide, stimulated by various governmental incentives policies for promoting renewable generation and reducing emission. The generation expansion in the future will be driven by not only peak demand growth but also emission reduction target. Thus, the traditional generation expansion planning (GEP) model has to be improved to reflect this change against the new environment. The policy makers need a better assessment tool to facilitate the new environment, so they can make appropriate policies for promoting renewable generation and emission reduction, and guide the generation mix to evolve appropriately over time. Since the expansion of new generation capacities is highly capital intensive, it makes the improvement of GEP quite urgent and important. The thesis proposes the GEP modelling improvement works from the following aspects: • Integrating short-term emission cost, unit commitment constraints in an emission target constrained GEP model. • Including the network transmission constraints and generation location optimization in an emission constrained GEP. • Investigating the impacts of multi-stage emission targets setting on an emission constrained GEP problem and its overall expansion cost. • Incorporating the uncertain renewable generation expansion and short-term DSR into the GEP problem and find out its potential contributions to the GEP problem. A real case study is made to determine the optimal generation mix of the Great Britain in 2020 in order to meet the 2020 emission reduction target. Different optimal generation mixes of the UK in 2020 are identified under a series of scenarios. The scenarios are constructed according to different GB network transmission capacity hypotheses and demand side response (DSR) level scenarios.

Two methodologies are presented for analyzing the choice between centralized and decentralized energy infrastructures from a least-cost perspective. The first of these develops a novel minimum spanning tree network algorithm to approximate the shortest-length network that connects a given fraction of total system population. This algorithm is used to identify high priority locations for decentralized electrification in 150 countries. The second methodology utilizes a mixed-integer programming framework to determine the least-cost combination of centralized and decentralized electricity infrastructure that is capable of serving demand throughout a given system. This methodology is demonstrated through a case study of Rwanda. The centralized-decentralized electrification paradigm is also approached from an energy security perspective, incorporating stochastic events and probabilistic parameters into a simulation model that is used to compare different development paths. The impact of explicitly modeling stochastic events as opposed to utilizing a conventional formulation is also considered Finally, a subsidy-free lighting cost curve is developed and a model is presented to compare the costs and benefits of three different financial mechanisms that can be employed to make capital intensive energy systems more accessible to rural populations. The optimal contract is determined on the basis of utility-maximization for a range of costs to the providing agency and a comprehensive single and multi-factor sensitivity analysis is performed.

Thermal power plants dominate electric power generation in Taiwan, which causes high Green House gases (GHG) emissions. CO2 is the most important greenhouse gas that cause global warming and sea-level to rise. This paper faces the relationship between CO2 limitation and power generation expansion planning (GEP) for the utility. Modify Particle Swarm Optimization (MPSO) is presented to determine the generation expansion planning strategy of the utility with independent power providers (IPPs). The utility has to take both the IPPs¡¦ participation and environmental impact into account when a new generation unit is expanded. This problem also takes into account the CO2 reduction and reliability issues, while satisfying all electrical constraints simultaneously from the supply point of view. MPSO scheme was improved to avoid getting a premature answer. Testing results shows that MPSO can offer an efficient way in determining the generation expansion planning. Generation expansion planning is an important decision-making activity in a competitive market, all utilities including IPPs need to maximize the profit while meeting the load demand with a pre-specified reliability criterion. In order to achieve the objective, utilities will perform the generation expansion planning to determine the minimal-cost capacity power addition. For better economy and efficiency, they will consider options of either constructing new generating units or purchasing electricity from other utilities or IPPs.

To date, over 84% of countries worldwide have renewable energy targets (RET), requiring that a certain amount of electricity be produced from renewable sources by a target date. Despite the worldwide prevalence of these policies, little research has been conducted on ex-ante RET policy analysis. In an effort to move toward evidence-based policymaking, this thesis develops computational models to assess the tradeoffs associated with alternatives for both RET achievement and RET policy formulation, including the option of creating renewable energy credit (REC) markets to facilitate meeting an RET goal. A mixed integer linear program (MILP), a probabilistic cost prediction model and a mixed complementarity problem (MCP) serve as the theoretical bases for the RET alternative and policy formulation analyses. From these models it was found, inter alia, that RET goals set too low run the risk of creating technological lock-in and could inhibit achievement of higher goals; probabilistic cost predictions give decision-makers essential risk information, when cost estimation is an integral part of alternatives assessment; and though REC markets may facilitate RET achievement, including REC markets in an RET policy formulation may not result in the lowest possible greenhouse gas emissions (GHG).

Power system transmission is undergoing rapid changes by the advent of renewable resources of energy, distributed generation, market integration, etc. Transmission planning is, nowadays, about building more inter-connections between adjacent regions or connecting off-shore wind farms to the grid or augmenting the network to support new path flows of energy and is still almost entirely the responsibility of regulated transmission system operator (TSO). Moreover, a well-developed transmission planning includes anticipating the generation investment decisions made by profit-maximizing generation companies (Gencos). Ensuring sustainable development of the power system necessitates coordination between TSO transmission investment with Gencos generation investments. Moreover, coordination between inter-connected TSOs in planning the network is also required in order to hunter the economic benefits of a robust and efficiently planned multi-area power system. Driven by the need for more coordination of the long-term planning of the inter-connected power systems, this thesis aims to develop models to be used in analysis of the multi-TSO multi-Genco transmission and generation planning and suggest mechanisms and coordination approaches for the better functioning of the power system. This includes mathematical models for transmission planning in the context of multiple TSOs and generation-transmission investment planning based on the game theory concepts in applied mathematics and evaluating mechanisms and approaches. / <p>QC 20161110</p>

With a significant number of states in the U.S. and countries around the world trading electricity in restructured markets, a sizeable proportion of capacity expansion in the future will have to take place in market-based environments. However, since a majority of the literature on capacity expansion is focused on regulated market structures, there is a critical need for comprehensive capacity expansion models targeting restructured markets. In this research, we develop a two-level game-theoretic model, and a novel solution algorithm that incorporates risk due to volatilities in profit (via CVaR), to obtain multi-period, multi-player capacity expansion plans. To solve the matrix games that arise in the generation expansion planning (GEP) model, we first develop a novel value function approximation based reinforcement learning (RL) algorithm. Currently there exist only mathematical programming based solution approaches for two player games and the N-player extensions in literature still have several unresolved computational issues. Therefore, there is a critical void in literature for finding solutions of N-player matrix games. The RL-based approach we develop in this research presents itself as a viable computational alternative. The solution approach for matrix games will also serve a much broader purpose of being able to solve a larger class of problems known as stochastic games. This RL-based algorithm is used in our two-tier game-theoretic approach for obtaining generation expansion strategies. Our unique contributions to the GEP literature include the explicit consideration of risk due to volatilities in profit and individual risk preference of generators. We also consider transmission constraints, multi-year planning horizon, and multiple generation technologies. The applicability of the twotier model is demonstrated using a sample power network from PowerWorld software. A detailed analysis of the model is performed, which examines the results with respect to the nature of Nash equilibrium solutions obtained, nodal prices, factors affecting nodal prices, potential for market power, and variations in risk preferences of investors. Future research directions include the incorporation of comprehensive cap-and-trade and renewable portfolio standards components in the GEP model.

No description available.

Transmission pricing and congestion management are two important aspects of modern power sectors working under a deregulated environment or moving towards a deregulated system (open access) from a regulated environment. The transformation of power sector for open access environment with the participation of private sector and potential power suppliers under the regime of trading electricity as a commodity is aimed at overcoming some of the limitations faced by the vertically integrated system. It is believed that this transformation will bring in new technologies, efficient and alternative sources of power which are greener, self sustainable and competitive. There is ever increasing demand for electrical power due to the changing life style of human beings fueled by modernization and growth. Augmentation of existing capacity, siting of new power plants, and a search for alternate viable sources of energy that have lesser impact on environment are being taken up. With the integration of power plants into the grid depending upon the type, loca- tion and technology used, the cost of energy production also differs. In interconnected networks, power can flow from one point to other point in infinite number of possible paths which is decided by the circuit parameters, operating conditions, topology of network and the connected loads. The transmission facility provided for power transfer has to recover the charges from the entities present in the network based on the extent of utilization. Since power transmission losses account for nearly 4 to 8% of the total generation, this has to be accounted for and shared properly among the entities depending upon the connected generation/load. In this context, this thesis aims to evaluate the shortcomings of existing tracing methods and proposes a tracing method based upon the actual operating conditions of the network taking into account the network parameters, voltage gradient among the connected buses and topology of the network as obtained by the online state estimator/load flow studies. The concept proposed is relatively simple and easy to implement in a given transactional period. The proposed method is compared against one of the existing tracing technique available in literature. Both active and reactive power tracing is handled at one go. The summation of partial contributions from all the sources in any given line of the system always matches with that of the respective base case ow. The AC power flow equations themselves are nonlinear in nature. Since the sum of respective partial flows in a given branch is always equal to the original ow, these are termed as virtual flows and the effect of nonlinearity is still unknown. The virtual flows in a given line are complex in nature and their complex sum is equal to the original complex power flows as in the base case. It is required to determine whether these are the true partial flows. To answer this, a DC equivalent of the original AC network is proposed and is called as the R - P equivalent model. This model consists of only the resistances as that of original network (the resistances of transformers and lines neglecting the series reactance and the shunt charging) only. The real power injections in a AC network i.e. sources into respective buses and loads (negative real power injections) are taken as injection measurements of this R P model and the bus voltages (purely real quantities) are estimated using the method of least squares. Complex quantities are absent in this model and only real terms which are either sums or differences are present. For this model, virtual flows are evaluated and it has been verified that the virtual real power contributions from sources are in near agreement with the original AC network. This implies that the virtual flows determined for the original network can be applied for day-to-day applications. An important feature of the virtual flows is that it is possible to identify counter ow components. Counter flow components are the transactions taking place in opposite direction to the net flow in that branch. If a particular source is produces counter flow in a given line, then it is in effect reducing congestion to that extent. This information is lacking in most of the existing techniques. Counter flows are useful in managing congestion. HVDC links are integrated with HVAC systems in order to transfer bulk power and for the additional advantages they offer. The incremental cost of a DC link is zero due to the closed loop control techniques implemented to maintain constant power transfer (excluding constant voltage or constant current control). Consequently, cost allocation to HVDC is still a problem. The proposed virtual power flow tracing method is extended to HVAC systems integrated with HVDC in order to determine the extent of utilization of a given link by the sources. Before evaluating the virtual contributions to the HVDC links, the steady state operating condition of the combined system is obtained by per-forming a sequential load flow. Congestion is one of the main aspects of a deregulated system, and is a result of several transactions taking place simultaneously through a given transmission facility. If congestion is managed by providing pricing signals for the transmission usage by the parties involved. It can also be due to the non-availability of transmission paths due to line outages as a result of contingencies. In such a case, generation active power redispatch is considered as a viable option in addition to other available controls such as phase shifters and UPFCs to streamline the transactions within the available corridors. The virtual power flow tracing technique proposed in the thesis is used as a guiding factor for managing congestions occurring due to transactions/contingencies to the possible extent. The utilization of a given line by the sources present in the network in terms of real power flow is thus obtained. These line utilization factors are called as T-coefficients and these are approximately constant for moderate increments in active power change from the sources. A simple fuzzy logic based decision system is proposed in order to obtain active power rescheduling from the sources for managing network congestions. In order to enhance the system stability after rescheduling, reactive power optimization has life systems to illustrate the proposed approaches. For secure operation of the network, the ideal proportion of active power schedule from the sources present in the network for a given load pattern is found from network [FLG] matrix. The elements of this matrix are used in the computation of static voltage stability index (L-index). This [FLG] matrix is obtained from the partitioned network YBUS matrix and gives the Relative Electrical Distance (RED) of each of the loads with respect to the sources present in the network. From this RED, the ideal proportion of real power to be drawn by a given load from different sources can be determined. This proportion of active power scheduling from sources is termed as Desired Proportion of Generation (DPG). If the generations are scheduled accordingly, the network operates with less angular separation among system buses (improved angular stability), improved voltage profiles and better voltage stability. Further, the partitioned K[GL] matrix reveals information about the relative proportion in which the loads should draw active power from the sources as per DPG which is irrespective of the present scheduling. Other partitioned [Y ′ GG] matrix is useful in finding the deviation of the present active power output from the sources with respect to the ideal schedule. Many regional power systems are interconnected to form large integrated grids for both technical and economic benefits. In such situations, Generation Expansion Planning (GEP) has to be undertaken along with augmentation of existing transmission facilities. Generation expansion at certain locations need new transmission networks which involves serious problems such as getting right-of-way and environmental clearance. An approach to find suitable generation expansion locations in different zones with least requirements of transmission network expansion has been attempted using the concept of RED. For the anticipated load growth, the capacity and siting generation facilities are identified on zonal basis. Using sample systems and real life systems, the validity of the proposed approach is demonstrated using performance criteria such as voltage stability, effect on line MVA loadings and real power losses.

[pt] Este trabalho explora a estrutura de decomposição de um problema de planejamento da expansão da geração hidrotérmica, utilizando uma integração entre uma Decomposição de Benders modificada e um Progressive Hedging. Consideramos uma representação detalhada das restrições cronológicas de curto prazo, com resolução horária, baseando-se em dias típicos para cada etapa. Além disso, representamos a natureza estocástica de uma política operacional hidrotérmica multiestágio por meio de uma Regra de Decisão Linear otimizada, garantindo decisões de investimento compatíveis com uma política operacional não antecipativa. Para resolver este problema de otimização em grande escala, propomos um método de decomposição de Benders aprimorado com várias instâncias do problema mestre, onde cada uma delas é reforçada por cortes primários além dos cortes de Benders gerados a cada candidato a solução do mestre. Nossa nova abordagem permite o uso de termos de penalização de Progressive Hedging para fins de regularização. Mostramos que o algoritmo proposto é 60 porcento mais rápido que os tradicionais e que a consideração de uma política operacional não antecipativa pode economizar, em média, 8.27porcento do custo total em testes fora da amostra. / [en] This paper exploits the decomposition structure of the hydrothermal generation expansion planning problem with an integrated modified Benders Decomposition and Progressive Hedging approach. We consider a detailed representation of hourly chronological short-term constraints based on typical days per month and year. Also, we represent the multistage stochastic nature of the hydrothermal operational policy through an optimized linear decision rule, thereby ensuring investment decisions compatible with a nonanticipative implementable operational policy. To solve the resulting large-scale optimization problem, we propose an improved Benders Decomposition method with multiple instances of the master problem, each of which strengthened by primal cuts and new Benders cuts generated by each master s trial solution. Additionally, our new approach allows using Progressive Hedging penalization terms for regularization purposes. We show that our method is 60 percent faster than the traditional ones and also that the consideration of a nonanticipative operational policy can save, on average, 8.27 percent of the total cost in out-of-sample tests.