A decision support model for long-term investment planning of electricity generation from the perspective of generation companies

Electrical power is vital for all societies, all economies and for daily lives of human beings. Moreover, unlike other products, electricity cannot be stored in large quantities, and should be consumed as soon as it is produced. Electrical power generation systems are obliged to meet constantly the growing electricity demand. On the other hand, the electricity needed for economic growth and social development should be continuously supplied to the consumers in sufficient amounts, in high quality, in economically affordable and environmentally sustainable conditions. In order to meet the growing demand for electricity, a good strategic planning for the growth of the electrical power system is critical. However, in the face of economic uncertainty, electrical power capacity investment decisions are high risky. The possible regulatory decisions may affect the expected income and amortization of the investments. Therefore, electrical power capacity investment decisions are partially or completely irreversible under changing market conditions. For managing these risks, researchers employ the Generation Expansion Planning (GEP) problem. Usually in a planning horizon of 10-30 years, and within the given reliability criteria, the capacity schedule (timing, sizing and technology of new power plant additions) that meets the projected load demand is acquired by the solution of GEP problem. In a vertically integrated electricity market structure, GEP has been solved centrally. In the centralized approach, GEP has been carried out by considering economic criteria, and by establishing models to determine the most appropriate technology, capacity size and timing of the additional installed capacity for meeting the growing demand at the lowest cost. However, in last two decades in many countries, the deregulation of electrical power industry has emerged. In order to ensure more efficient electrical power industry, many countries liberalized their electricity markets. Transformation from monopolies to the competitive electricity markets has created major changes in the organization and operation of the electrical power systems. Therefore, assumptions and solution approaches of GEP problems has changed. In decentralized approach, the exercise of GEP can be carried in two stages. Now, the individual investors undertake the first stage of GEP. In this stage, the individual investors make their own investment plans so as to maximize their long-term profits. In the second stage, the central authority approves or rejects the individual investment plans of investors. This central authority is generally an independent system operator (ISO) that operates and manages the transmission lines, and specifies security and reliability standards and guidelines. In some cases, there may be repetitive mechanism between the two phases in order to provide an opportunity to revise the investment proposals of individual investors. The focus of this dissertation is the first stage of GEP in a deregulated electrical power industry. Therefore, this dissertation presents some modeling frameworks for generation capacity investment planning applicable to any independent investor firm. The objective of the generation capacity investment planning is the maximization of the total present value of the investor's profit at the end of the planning horizon. The proposed modeling frameworks consider the long-term decision making process of investor firms and the discrete nature of generation capacity addition. It is assumed that the investor is willing to invest in four different technologies, namely, gas-fired, coal (lignite)-fired and hydroelectric power plants and wind turbines. The developed models are generic enough to accommodate other technologies within certain modifications in the operational constraints. Also, at the end of existing generation units' useful life, refurbishment decisions are employed. In this way, conversion of old units to the units with lower operation costs and/or green house gases emissions is modeled. The electricity markets that are emerged as a result of deregulation trends in the world could be categorized in three different types: market pools (PoolCos), bilateral contract (BC) markets and hybrid markets (HMs). In a PoolCo, such as a day-ahead market (DAM), an Independent System Operator (ISO) operates and manages the entire system, and maintains its reliability by clearing the market according to the bids from both the sellers (the generation companies) and the buyers (the retailers). In other words, in a PoolCo different generation companies (GenCos) compete not for specific customers but for the right to supply energy to the grid. BCs are flexible market models where the participants can specify and negotiate the terms and conditions of trading agreements. In a HM, the participants can directly sign bilateral contracts with others and/or bid in the PoolCos. The conducted extensive literature review reveals the fact that GEP problems that are exercised by decentralized approach, only consider the prices at the spot markets (PoolCos). Although bilateral contracts and spot market prices have similar increase trends in the long-term, usually, they are at different levels. Due to the different levels of electricity prices in different markets, the feasibility of generation capacity investments vary according to the selected market price for the calculations. Therefore, at a HM where electrical power trade is carried out with bilateral contracts and organized markets (spot markets), it is important to make GEP according to which market price. In parallel with the electricity market liberalization trends in the world, major steps have been taken for the liberalization of Turkish electricity market after the transition period between 2006 and 2009. The electrical power industry deregulation in Turkey has aimed a hybrid electricity market structure such as in Pennsylvania New Jersey Maryland (PJM), New York, New England and California electricity markets. In this dissertation, a long-term generation capacity investment problem of an independent GenCo faces in an environment where there exists a hybrid electricity market such as in Turkey is modeled and solved. To the best of our knowledge, there is no study that considers operations of GenCos in bilateral contract and day-ahead markets together for long-term individual generation capacity investment planning of a GenCo. Considering the already published works in the literature, the proposed models in this PhD thesis is novel in the sense that the operations of the GenCo in bilateral contract and day-ahead markets are incorporated in the GenCo's long-term investment planning by using shortly time steps. In this way, long-term sales strategy of the GenCo is considered in long-term investment planning. Furthermore, usage of yearly and monthly time steps minimize the overall risks associated with the long-term planning by helping the decision makers to revise their investment plans against changes in the context of a continuing deregulation process. In the dissertation, three different long-term generation capacity investment model for 10-year planning horizon are developed. All tests were performed on a 2.67 GHz Intel I5 processor with 4 GB RAM, running 64-bit Windows. Three different mathematical models in the dissertation were implemented in GAMS 23.5 using the CPLEX Academic Studio 12.0.1 solver. The solutions of these models provide essential strategic planning information to the decision makers of the GenCos such as in which planning period which generation technology should be invested for how much capacity, in which planning period which generation unit should generate how much power, in which wholesale electricity market how much of the generated power should be sold and, finally which existing generation unit should be rehabilitated at the end of its economical life. In the dissertation, the first model named as "a long-term investment planning model for the independent GenCo" is a mixed integer linear programming model that provides investment decisions for candidate generation units and rehabilitation decisions for existing units by using yearly periods during the planning horizon. By using binary decision variables in the mathematical model formulation, discrete unit sizes of generation technologies have been considered in long-term investment planning of the GenCo. The profitability of already existing generation units and new capacity additions, and electricity wholesale transactions of the GenCo have been included in the mathematical model. The second model in the PhD dissertation is named as "a long-term investment planning model that considers mid-term planning for the independent GenCo". In the context of deregulation such as in Turkey, not only the long-term risks are high but also medium-term risks are high. In order to minimize the overall risks associated with long-term planning, a monthly time-step is employed in this model. Moreover, splitting planning horizon into monthly sub-periods allows mid-term planning decisions such as power sales and generation unit maintenance planning to be taken along with investment decisions. In addition, the proposed model is novel in the sense that mid-term power sales and generation unit maintenance planning decisions could be taken along with investment decisions within the framework of the GenCo's long-term investment planning problem. The last model developed for the purpose of the PhD dissertation is named as "a fuzzy long-term investment planning model for the independent GenCo". The long-termed investment decisions of the GenCo cannot be examined without considering the environment, where the GenCo do business. The changes in the economic environment and in the energy industry would affect supply and demand in the electricity market. Therefore, those changes would affect the optimal value of the GenCo's objective function. In such a fuzzy environment, decision maker of the independent GenCo might not want to actually maximize present value of its total profit. Therefore aiming to reach some aspiration levels which may not even be definable crisply would be more realistic. In addition to the fuzziness in the economic environment, changes in the nature affect especially energy output of power plants using renewable energy sources. Turkey is located in the Mediterranean Basin that is predicted to be severely affected by the adverse effects of global climate change, particularly in terms of water resources. It is expected that water resources of Turkey are threatened by climate change impacts such as an increase in summer temperatures, a decrease in winter precipitation particularly in western provinces, a loss of surface waters, an increase in the frequency of droughts, land degradation, coastal erosion and floods. Also, it is expected that the frequency, intensity and duration of extreme weather events such as drought will increase in the South, Southeast and West regions where the important part of Turkey's hydroelectric potential exist. Hence, the expected maximum energy output of hydroelectric units becomes fuzzy. In addition to the fuzziness in the energy output of hydroelectric units, GenCos generally can tolerate some increase in their investment budgets during long-term investment planning. Due to the fuzziness in the objective function, budget and maximum hydroelectric output constraints of the GenCo's long-term investment planning problem, a fuzzy mathematical model with a fuzzy objective and fuzzy constraints is developed. Then, this fuzzy decision making problem is solved by using Zimmermann approach. Also, for investigating which generation units are preferred to hydroelectric units in investment decisions, sensitivity analysis are performed by increasing the allowance for deviations (tolerance) in the maximum energy output of hydroelectric units. In the literature, the effects of climate change on hydroelectric power plants' energy output and financial performance are examined. However, to the best of our knowledge, the substitution between hydroelectric power plant investments and other generation technology investments is not investigated in the literature. Thus, it can be said that another contribution of this PhD thesis is investigating how investment decisions of the GenCo are affected, if the fuzziness in the energy output of hydroelectric units increases by the adverse effects of climate change.