Optimization of Unit Commitment Problem Using Genetic Algorithm

Introduction

Electrical power systems are designed and operated to meet the continuous variation of load demand. In power system minimization of operation cost is very important. The most important problem with electric power generation scheduling is unit commitment. Unit Commitment means coordinating short-term unit generation to meet load demand forecasting. The problem of unit commitment is a complex problem of optimization. It has a variable integer as well as a constant variable (Kumar et al., 2014). The Unit Commitment (UC) function in the power system refers to the question of optimization to evaluate the on/off states of generating units that reduce operating costs for a given time horizon (Zhao et al., 2006). The committed units must meet the system's forecast demand and spinning reserve requirement at the minimum operating cost, subject to a wide set of operating restrictions.

In practice as well as in state-of-the-art research on deterministic UC models, the underlying security assumptions (i.e., contingency) network, and operational constraints for alternating current (AC) power systems have varied the most significantly described by (Stott & Alsac, (2012). Garver., (1962) improves the foundational formulation of UC. Several influential approaches have concentrated on mixed-integer linear programming (MILP) with more close convex null representations of non-network thermal unit constraints operation described in several research papers (Carrion & Arroyo, 2011); (Ostrowski et al., 2012); (Morales-Espana et al., 2013); (Jabr, 2012). Unit Commitment in linear (DC) network with losses are formulated by (Morales et al., 2009) and (Wu & Shahidehpour, 2010) and without is formulated by (Feizollahi et al., 2015).

Padhy, (2004) explain how the resulting commitment schedules neglect the constraints of reactive power dispatch and AC power flow, which must subsequently be compensated for via corrective and generally ad-hoc processes. The resultant interaction schedules ignore the constraints of reactive power dispatch and ac power flow, which must consequently be balanced by corrective and typically ad-hoc processes explained by (Bhardwaj et al., 2012). It is the energy control center's most critical feature that specifies the on / off status as well as the actual power output of the generator. Parashar & Swankar, (2013), proposes a methodology to reduce device running costs over the planning cycle subject to various physical operation and computation constraints. Columbus & Simon, (2012), explain unit Commitment (UC) as a nonlinear mixed integer optimization problem. Power system engineers also need to understand the economic decision-making tool for to accomplish the economic goal (Galli, 2019).

With rapid development of the world economy and the exhaustion of the fossil fuels, renewable energy sources (RES) has received significant attention in the current society. Several authors (Wei et al., 2013); (Danwen et al., 2016); (Shin et al., 2017) explain the characteristics of renewable energy sources as a non-pollution and inexhaustible. The wind energy has become one of the renewable sources of electricity generation. Author Hirpara & Sharma, (2020), utilizes the Fokker-Planck method, a mathematical stochastic method, to analyse the noise-influenced wind turbine-generator system. The solar energy is another most popular example of renewable sources of energy. To make a full utilization of the output power of a PV solar cell operating at the maximum power point (MPP) is described in detail by Yatimi, H. (2018).  Some researcher work on a practical case study such as (Mohamed et al., 2017) for 20 MW Egyptian PV solar power plant with battery backup system. Singh et al. (2019) evaluate the performance of a 4 kW Isolated Solar Powered Lab with IoT Energy Management System. Some applications based on solar energy is also discussed by (Singh et al. 2018) and (Salima et al. 2018).

Various emerging concepts influence logistics management (Baporikar 2020), particle swarm optimization, fuzzy logic controller (Ramadan & Altamimi 2017) and genetic algorithm etc are general purpose stochastic and parallel search tool that can be used as a technique of optimization to obtain near-global optimal solutions to the problem. The Genetic algorithm is inspired by the ideas of the natural selection and survival of the fittest in biology and evolution. Lambora et al. (2019), present review on genetic algorithm application. Many authors such as (Meshram et al.,2016) and (Xua, et. al. 2018) explain the Genetic algorithm applications in different areas.