Electricity supply-demand matching : An integrated approach

Abstract

Electricity is the most preferred source of energy because of its quality and convenience. This has made the electric power industry one of the fastest-growing sectors in most developing countries. Even though the electric sector typically receives a major share in the budgetary allocations of these countries, the addition of new generation capacity has not kept pace with the increase in demand for electricity. Capital scarcity in the public sector, lack of enthusiasm among private and foreign investors, and strong opposition from environmentalists have together contributed to the slow pace of new generating capacity addition. As a result, we have severely constrained electricity systems in most developing countries. Electricity is generated as and when needed and cannot be inventoried like manufactured goods. The level of electricity supply has to always be maintained at the level of demand. Therefore, either the supply has to be matched to the demand or vice versa. An electricity system with sufficiently large installed capacity will have this kind of flexibility of supply-demand matching. A system with inadequate installed capacity cannot have this kind of luxury and has to manage with shortages. The problem of shortages gets compounded when there are additional inadequacies in the form of a lack of fuel supplies and inadequate water in the reservoirs. Under these situations, matching the ever-increasing demand with existing supply alone becomes unattainable. Therefore, it becomes imperative for the utilities to look for other alternatives to achieve an effective match between supply and demand. This thesis addresses the issue of supply-demand matching in the case of constrained electricity systems. The main objective of any electrical power system is to supply electricity as demanded by the consumers. Due to various reasons, as explained above, the supply system would sometimes not be able to supply the requisite electricity. This may be due to capacity or energy shortages. In these situations, utilities need to adopt non-supply options (methods of supply interruptions and power rationing) to overcome the shortages. However, the non-supply options are short-term measures to achieve a match between supply and demand. In the long run, and when the shortages become unmanageable with non-supply options, new supply options have to be considered. Traditionally, new supply is added to the system to meet the forecast future demand. However, in the case of constrained systems, new supply is required to bridge the existing gap between demand and supply along with meeting the forecast future demand. To include existing supply, non-supply, and new supply options in the planning exercise, it is necessary to consider the total electricity system as consisting of two sub-systems, namely, the electricity supply system (utility) and the electricity demand system (consumers). Once implemented, the implications of including these options in the planning will be felt by both the utility and the consumers. Thus, to achieve an effective match between supply and demand, we have the following alternatives: 1. Meet the varying demand as it comes, through various supply options. That is, adjust the generation level to the varying demand level. This is possible only if the supply potential is adequate to meet any level of demand. In this case, the supply is matched with the demand. 2. Reduce the demand through various non-supply options so that the demand reaches the level of supply potential available. This is done if the supply potential available is inadequate to meet the demand. In this case, the demand is matched with the supply. 3. Install new generation capacity to meet the additional demand which is not met by the existing supply options. However, this is a medium- to long-term measure and will not give immediate benefits. In this case, the supply is increased to match it with the demand. Costs are involved with the adoption of any of the above options. To make a cost-effective match between supply and demand, one needs to minimize the total cost of using the above options. However, the cost of supply and cost of new supply are incurred by the supplier of the electricity (utility), and the cost of non-supply is incurred by the consumer. Thus, the objective of minimizing the total cost consists of costs incurred by two different parties. Under this situation, the research objective is to look at the total system consisting of both the supplier and the consumer, and minimize the total cost of the system rather than minimizing the cost of any one party. In short, the approach is to analyze the existing electricity supply options, non-supply options, and new supply options as an integrated system with an objective to match effectively and efficiently the supply and demand for electricity. Since the terms supply, non-supply, and new supply of electricity are either new or used in differing contexts, our interpretation of these terms is as follows: Supply of electricity is defined as the electricity that is supplied to the consumers as demanded, using the existing supply options. Non-supply of electricity is defined as the electricity that is not supplied to the consumers when demanded, using various non-supply options (rationing measures). New supply of electricity is defined as the electricity that is supplied to the consumers as demanded, using new supply options. The supply options could include electricity generation from hydel, coal thermal, nuclear, and natural gas power plants, as well as imports from other utilities. The non-supply options include the imposition of power cuts and energy cuts (power rationing) on some identified customers, and resorting to load shedding (cutting the supply) at specific hours for a given geographical area. The new supply options could include new hydel, coal, nuclear, natural gas, and other energy source-based power plants. A review of the existing literature showed a lack of an integrated approach to managing electricity shortages. Non-supply options have mostly been included in planning exercises from the point of view of system reliability or for allocating shortages among various consumer categories. New supply options have been considered to meet forecast future demand for electricity, not for bridging the existing gap between supply and demand. An approach to integrate supply, non-supply, and new supply options in order to achieve a match between supply and demand by managing shortages has not been attempted in the literature. More specifically, the objectives of this thesis are as follows: 1. Develop Representative Load Curves (RLCs) to gain an understanding of the varying demand pattern for electricity. 2. Develop a mathematical model for planning and scheduling various electricity supply options, non-supply options, and new supply options to achieve an effective matching of electricity supply and demand-both in terms of power and energy-at minimum cost. 3. Use the model to study the effect of changes in supply availability and hourly demand on the performance of the electricity system in the state of Karnataka. 4. Develop some likely future scenarios and apply the model to study the applicability of various options to supply-demand matching under different situations. 5. Study the implications of these options on decision-making and policy formulation in different situations. The electricity system in the state of Karnataka, which is faced with both capacity and energy shortages, has been considered as a classic example to study all aspects of supply, non-supply, and new supply of electricity. Due to inadequate availability of supply potential, this system currently adopts rationing measures like power cuts, energy cuts, and load shedding to achieve some kind of match between supply and demand. Also, the addition of new generating capacity is negligible due to severe capital shortages faced by the state. The thesis presents an overview of non-supply of electricity, including the costs associated with it. There are a number of reasons that warrant non-supply of electricity. Utilities managing constrained electricity systems adopt different types of non-supply options to overcome shortages. Costs-actual or perceived-are incurred by consumers whenever there is non-supply of electricity. There is a need for estimating the cost of non-supply in electricity planning. Fortunately, there are a number of methods available in the literature to estimate this cost. Even though estimation is difficult and influenced by many factors, the cost of non-supply has found use in a variety of applications. Multiple Discriminant Analysis has been used to group the 365 load curves of the year 1993� into nine Representative Load Curves (RLCs). This was done to characterize the varying demand for electricity. The nine RLCs approximately represent the hourly demand for electricity during the year 1993� in the state of Karnataka. Our integrated Linear Programming (LP) model uses these RLCs. Further analysis of these RLCs helped identify the factors likely responsible for continuous variations in electricity demand. These factors are: (a) Seasonal variation in demand, which causes significant fluctuations throughout the year. (b) Variations in agricultural demand, which depend on agricultural operations and factors such as cropping patterns and monsoon. (c) Variations in industrial demand, which depend on the level of industrial activity. (d) Rationing measures like power cuts, energy cuts, and load shedding adopted to reduce demand. (e) Changes in water heating and air cooling demand in residential and commercial sectors. An integrated LP model was developed for planning and scheduling various options to match electricity supply and demand on an hourly basis. The objective function was to minimize the annual total cost of supply-demand matching. The cost is the sum of the cost of supply, cost of non-supply, and cost of new supply. The major features of the LP model were: (i) Treatment of non-supply options as decision variables. (ii) Inclusion of equations for estimating power and energy shortages, and allocation of these shortages to non-supply and new supply options on an hourly basis (representative hours). (iii) Representation of hourly electricity demand in the form of representative hourly loads. A computer program (coded in FORTRAN 77) was developed to generate the LP model from data in Mathematical Programming System (MPS) format. The model in MPS format was then passed to the Optimization Subroutine Library (OSL) on an IBM-SP2 machine to obtain an optimal solution. The problem size was 5,062 rows (constraints) and 3,576 columns (variables). Initially, the LP model was applied to the base case scenario (assuming the demand met by the supply system in the year 1993� as the actual demand for electricity, i.e., no shortages in the system). The base case results indicated that with efficient management, the existing supply system could have met a higher level of demand in 1993�. The optimal hourly allocation of demand to various options showed that the available power of any option at any given hour was always higher than the demand it was meeting. Sensitivity analyses indicated that the results were somewhat sensitive to changes in the values used for the costs of supply through existing supply options. However, the results were not very sensitive to changes in other cost parameters (cost of non-supply and new supply), and hourly demand for electricity (except, of course, during peak hours). The performance of the system was further tested for the impact of changes in the level of demand and supply availability with respect to the applicability of various options in achieving a match between supply and demand. The results indicated substantial influence in this aspect. In addition, four scenarios were constructed for future years to evaluate the appropriateness of the various options considered in the study under different situations. These scenario analyses also provided ample proof of the model抯 ability to choose appropriate options to match supply and demand. Different conditions. As desired, the scenario analyses proved the ability of the model to provide answers to questions regarding the type of options, the quantum of contributions, and the time of activation needed in order to match supply and demand on an hourly basis (for representative hours). The major contributions of this thesis are summarized as follows: 1. An integrated approach to electricity supply-demand matching, as reported in this thesis, is unique in the sense that it considers options available under various alternatives梟amely, existing supply, non-supply, and new supply. This approach is most suited for severely constrained electricity systems that must manage both power and capital shortages. 2. A new term, non-supply of electricity, has been introduced in this thesis. An overview of non-supply and the cost of non-supply of electricity has been discussed. Reasons for non-supply of electricity, types of non-supply, methods to estimate the cost of non-supply, factors influencing these estimates, and some estimates found in the literature have been presented. 3. Multiple discriminant analysis has been used to classify the 365 load curves into nine Representative Load Curves (RLCs). Further analysis of these RLCs helped identify factors likely responsible for continuous variations in electricity demand. Some of the important factors identified include seasonal factors, variations in demand from industrial and agricultural sectors, water heating demand from residential and commercial sectors, and rationing measures adopted by the utility. 4. The development of RLCs has shown a new way of representing hourly electricity demand, which was traditionally done through annual load duration curves. The RLCs offer advantages over annual load duration curves by providing additional information regarding daily and seasonal variations in demand, and the approximate time of occurrence of particular demand levels. 5. The integrated Linear Programming (LP) model developed in the study has been demonstrated as a planning tool for: (a) optimal scheduling of existing supply options, (b) estimation of likely shortages in future years, (c) rational allocation of shortages to various non-supply and new supply options, and (d) scenario analyses. 6. The scenario analyses carried out in this study provided ample proof of the model抯 ability to choose appropriate options to match supply and demand under different conditions. We hope that the results of this thesis will help utilities (suppliers of electricity) make rational tactical decisions in the context of severely constrained electricity systems. The results reported here may also be used to support the rationalization of government policies, particularly with respect to tariff structures in electricity supply.