3582277604

Elcctrical & Computer Emaneering; An International Journal <ECU) Vohime 2. Number 1. Nlarch 2013

1. Introduction

Hydrothcrmal coordination (HIC) mcans dctcrmination of thcrmal powcr and hydro powcr such that total system generation cost is minimum whilc satisfying the system constraints. Howcvcr the operating cost of hydro plant is ncgligiblc. so our objcctivc is to mini mi ze the total fucl cost of thcrmal units. subjcctcd to the water availability constraint and cncrgy balancc condition for a givcn period of timc. It is basically non-lincar programming problem involving non-lincar objcctise timetion and a mixturcof linearand non-lincar constraints.

The HTC problem is dividcd into thrcc typcs hascd on the duration of intcrval: thcsc arc long. mid and shoit-rangc problcms. Mid and long-rangę HIC’ is for one or morc ycars on a wcckly or monthly basis |1). The linal output is the amount of water to be dischaigcd at cach hydroclcctric plant throughout the coming wcck. Short-range planning on the other hand is conccrncd with distributing the generation among the availablc units ovcr a day or wcck. usually on an hourly basis. satisfying the operating constraints. In short-range problcms a fixcd water head is assumed and the net head variation can be ignored for rclativcly large rcscrvoiis, in which casc powcr generation issolcly dependent on the water dischargc [ I ].

Earlicr. a widc varicty of optimization tcchniqucs havc bccn applied to solving HTC |2) problcms such as dynamie programming (DP). gradient scarch methods etc. But. bccausc of the dnrwbacks. such as drastic growlh of computation. dimensionality roquircmcnt. and insccurc convcrgcncc properties thcsc local optimization tcchniqucs arc not suitablc for HI C problem.

In rcccnt ycars. duc to their flcxibility. \crsatility and robustness in sccking global optimal solution. heuristie optimization tcchniqucs arc gaining lot of interest.

Simplc gcnctic algorithm is capablc of localing the ncar optimal Solutions, but it takes large numberof iterations to comcrgc and taking large CPU timc. In addition to simplc GA. therc cxists a number of advanccd GAs. designed to havc advanccd fcaturcs.

A numberof heuristie tcchniqucs likc GAs. simulatcd annealing. cvolutionary strategy. particie swarm optimization havc bccn attempted for solution of HTC. Ali thcsc tcchniqucs uscd large number of variablcs which not only depend on number of generators. but also on number of intcrvals in the planning horizon. |l) presented optimal gamma bascd GA for solving HTC problem, which rcduccs the number of variablcs to be considcrcd to a minimum. GA has also bccn applied for solution of HTC. |2) proposcd an clficicnt method for optimal schcduling of fixcd head hydro thcrmal plants. A fast computational gcnctic algorithm for solving the cconomic load dispatch problem of thcrmal units has bccn presented in |3). |6) prcscnts short term h>xlro thcrmal coordination by Lagrangian rclaxation method using dual optimization tcchniquc.| 10) presented a fast cvolutionary tcchniquc for short-term hydrothcrmal schcduling. 111 ] prcscnts HTC using GA. A unit bascd cncoding schcmc is uscd. which restriets the applicability of GA to largc-scalc systems. In unit bascd cncoding the chromosomc length inercase with incrcascof numberof units in the system.

In the present paper a fast computation GA |3] is uscd considcring the optimal gamma bascd approach for short term hydrothcrmal coordination. The paper prcscnts a mcthodology to ovcrcomc the long computation times involvcd with simplc Gcnctic Algorithms.