Optimum Blending of Coal by Linear Programming for the Power Plant at Seyitömer Coal Mine K.Erarslan, H.Aykul, H.Akçakoca & N.Çetin
Dumlupınar University, Department of Mining Engineering, 43100, Kütahya, Turkey
ABSTRACT: In this study, a linear programming model is developed to determine the optimum coal blend in terms of quality and quantity. Coal with various features is mined from different panels of Seyitömer Lignite Coal District and fed to a nearby power plant. The quality of the coal is extremely variable through the horizontal and vertical directions, which entails the precise planning of coal blending during the mining and stockpiling stages. Otherwise, a large penalty has to be paid to the power plant. In this study, the objective is to match the calorific values required by the power plant. The quality features and production capacities of coal from different panels are determined and are used in quality constraints. The power plant requires coal in two groups, which are of different qualities and quantities. Therefore, two linear programming models complementing each other are developed in order to determine the blending conditions that satisfy the needs of the plant. The models are introduced and solved m the LINDO package program. Reasonable solutions are obtained and optimal amounts of blending are handled. The model also allows the evaluation of coal panels of low quality.
1 INTRODUCTION' Linear Programming (LP) is one of the most widely used methods of operation research for decision problems. This method is a reasonable and reliable procedure for determining the optimum distribution of resources, optimal production, minimum cost, maximum profit, etc., which comprise the objectives (Öztürk, 1997). In this method, decision parameters to make the objective optimal are linear or assumed to be linear (Taha, 1992, Hillier and Liebermann, 1995). The general form of the problem is formed by objective function and subjected constraints; (1)
where; Z= objective of the model Cj = coefficient of/* decision variable (1=1,2 Xj = / * decision variable
coefficient of decision variable (i=l,2,..,m) b, = limited resource for iJ* constraint LP approximation is widely used in mining as well as in other industrial fields. Open pit limits, production scheduling, material flow in processing plants, blending, equipment selection, method selection, transportation, etc., are its main applications (Chanda and Wilke, 1992, Dijilani and Dowd, 1994, Huang, 1993, Mann and Wilke, 1992, Meyer, 1969, Smith and You, 1995). However, investment, planning, or selection, in other words any actions requiring decision, can be optimized. Especially in open pit mines and underground mines feeding coal to power plants, the quality and quantity of coal is crucial because the burner blocks of power plants are designed according to specific features of coal. Inability to match coal quality and quantity to these specific features results in either penalty costs for the coal enterprise or a decrease in the power plant's efficiency. In addition, inconsistent coal features lead to wear in the power plant's burning units and all integrated components. In this respect, coal-producing enterprises try to match their coal features to power plants' specifications by blending and homogenizing coal extracted from different panels and levels. Satisfying the requirements of the plant is achieved by selective 719
mining and/or blending. In this study, a relevant case is considered. Seyitömer coal enterprise in Kütahya, Turkey has problems of quality and quantity in supplying the nearby power plant. A well-planned and organized blending procedure and, accordingly, production plan is necessary. In this paper, the problem, is modeled in terms of linear programming and reasonable solutions are obtained. 2 SEYİTÖMER COAL ENTERPRISE AND...