AUTOMATIC GENERATION CONTROL IN RESTRUCTURED POWER SYSTEM

Abstract

Around the world, the electric power industry has been undergoing reforms from the traditional regulated, Vertically Integrated Utility (VIU) into a competitive, deregulated market. Market deregulation has caused significant changes not only in the generation sector but also in the power transmission and distribution sectors and has introduced new challenges for market participants. The new electricity market structure results in large number of independent players such as Generating companies (Gencos), Transmission companies (Transcos), Distribution companies (Disocs) and customers. The system operation and market management is carried by an Independent System Operator (ISO). The primary objective of the System Operator is allowing the contracted power to flow from Genco to Disco. To transport the contracted power at acceptable level of quality and reliability certain ancillary services are required by the System Operator. The US Federal Energy Regulatory Commission (FERC) has identified the six services as ancillary services, these services are, Scheduling, System Control and Dispatch, Reactive Power and Voltage Control, Frequency Regulation, Energy Imbalance, Operating Reserve - Spinning, Operating Reserve - Supplemental. One of the most important ancillary service is frequency regulation service which is procured through a separate market by many ISO's such as California Independent System Operator Corporation (http://www.caiso.com), ISO New England, Inc. (http://www.iso-ne.com), New York Independent System Operator (http://www.nyiso.com), Ontario's Independent System Operator(http://www.ieso.ca), etc. Frequency regulation is the minute-to-minute adaptation of the generator output to meet the imbalance between total supply and demand in the system. This instantaneous response of a generating unit is usually achievable through the use of the governor droop characteristics (primary control) and automatic control signals from the control area based on determining the required change (up or down) to the real power output to bring the Area Control Error (ACE) within bounds (secondary control). Frequency regulation helps to maintain interconnection frequency at nominal value, minimize differences between actual and scheduled power flows between control areas, and match the generation to the load within the control area. In the present work, issues related to frequency regulation and AGC scheme have been addressed. In order to improve stability, power quality and reliability of electric supply energy storage devices such as Flywheel, Battery Storage, Compressed air, Superconducting magnetic energy, Pumped storage hydro-electric system have been incorporated in the practical power system applications. In the present work, Superconducting Magnetic Energy Storage (SMES) system has been used to improve the dynamic response of an AGC scheme. The conventional AGC scheme cannot be implemented directly in a deregulated power system environment. Few reasons for this are: different types of transactions between Gencos and Discos, a Genco may or may not participate in the market, participation factors of Gencos may dynamically change, revised definition of Area Control Error (ACE). A competitive electricity market may have following transactions; Poolco based transactions, bilateral transactions and combination of these two transactions. A general purpose AGC model considering all type of the transactions has been proposed in this thesis. The proposed model also includes the SMES units. In an AGC scheme frequency can be controlled by speed-governor action (primary control) and by utilizing an integral controller (secondary control). The essential requirement of secondary controller in the system is to control the real power output generating units and minimizes Area Control Error (ACE). One of the important element of an AGC scheme is a controller which is used in the secondary control to eliminate the frequency error. In this thesis, PID controller has been used for this purpose. There are many methods available in the literature to obtain the parameters of PID controller. The practical model of AGC scheme consists of many non-linear elements such as Generator Rate Constraints (GRC), generator output saturation; etc. A conventional method available for the tuning of PID controller does not give satisfactory results due to the nonlinearity in the AGC model. Therefore, the parameter of PID controller has been tuned using Genetic Algorithm which in a system having non-linearity gives quite satisfactory results. In this thesis, PID controller has been used to control the real power output of the Gencos of different areas. In multi-area system, as the system parameter changes the performance of the PID controller deteriorates. Parameters of PID controller have to be 11 retuned again and again as the system parameters change. This problem may be avoided by using artificial intelligent technique based controller. In this thesis, an ANN based controller has been proposed for multi-area AGC scheme with SMES unit and suitable in restructured electricity market environment. The conventional PID controller has been replaced by artificial neural network (ANN) trained controller. The best value for controller parameters has been obtained by training the ANN off line at different load parameters using Back Propagation Algorithm (BPA). The best value of weights has been obtained by minimizing the error through Gradient Descent optimization technique. The inputs to the neural controller were the ACE (Area Control Error) and a reference load variation. The output of controller has been used to change the set point of the governor. A properly trained ANN based controller gives satisfactory results but it requires extensive training in which a large amount of data is required. Further, in this thesis, Dynamic Neural Network (DNN) based controller has been designed. The advantage of using the DNN controller is that it does not require extensive and rigorous model for optimal tuning. The DNN based controller has constructed by taking only one dynamic neuron and only one unit time delay to feedback output layer. The weight vectors of DNN based controller is found by hit and trial method. Automatic Generation Control is an important function in modern Energy Management System. Since, January 1998, new control performance standards (CPS) have been enforced in North America by North American Electric Reliability Council (NERC). NERC requires that control areas must be no less than 100% compliant with CPS1 and no less than 90% compliant with CPS2. These two indicators provide fully guidance to ensure the quality of interconnected frequency. CPS1 measures the correlation of a control area's Area Control Error (ACE) and interconnection frequency error over 12- month duration. CPS2 measures how many times a control area manages to hold the magnitude of its 10-minute ACE average within a predetermined ACE limit over month duration. In this thesis, NERC standards compliance has been established for all the three proposed controllers. Around the world, governments have been paying more attention on policies that promoting the clean or Renewable Energy Sources (RES) to serve customer demand. in Renewable energy sources, and in particular solar technology and wind power technology, are considered to be the most sustainable. In this thesis, a multi-area AGC scheme has been proposed which includes Photo voltaic (PV) and wind energy resources. The previously proposed AGC model has been modified to include PV and wind sources. The PV and wind model suitable for AGC scheme has also been developed. As the power output of these sources is not constant they produce very small fluctuation in the output frequency. These fluctuations can be minimize^ by using the energy storage systems. In this work, Flywheel Energy Storage System (FESS) has been included into the proposed AGC model. In this work, it is assumed that RES is owned by ISO. In place of per unit cost of these resources, the social cost has been considered. It is assumed that whenever regulating power is required system operator provide the maximum output of PV and wind sources. The proposed AGC model and along with the proposed controller has been successfully tested on two systems. First system has taken as IEEE 39-bus New England system containing 10 generators. The 39-bus system has been divided into two control areas based on geographical nature of the system, SMES unit has been considered in area-1. The 75-bus Indian power system has been considered as the second test system. The 75- bus system consists of 95 lines and 15 generators. It has been divided into four control areas of different ratings and having different no. of Gencos in each area. In this case, SMES units has been considers in area-1 and area-3. A deregulated electricity market scenario has been assumed in the both the test systems. To simulate the effect of RES for frequency regulation in 39-bus system, one PV unit has been considered in area-1 and one wind turbine unit in area-2. Similarly, for 75-bus system, one PV unit has been considered in area-2 and one wind turbine unit in area-4. All the simulation studies have been carried out using MATLAB SIMULINK. Though Automatic Generation Control under competitive electricity market is a vast area of study, the present work tries to fill-in some research gaps. Scope in liberalization and restructuring of power system has been addressed as a subject of matter. The issues workout in this thesis would continue to attract attention for a long time.