Linear Programming to Improve Raw Coal Blending: A New Twist to an Old Tool

A traditional linear programming technique is offered to optimize raw coal blending for a given plant with known performance characteristics. The model matches raw coal feed rates to a plant's assumed optimum "target" feed, subject to quality and quantity constraints. The technique's usefulness is enhanced by a novel computer program for optimizing clean coal sampling frequency. This program analyzes a plant's current sampling scheme, models alternatives, and compares results. The user is presented a menu of feasible sampling schemes which can improve plant economics significantly. Adoption of these improved sampling programs provides the plant operator with a clearer view of the cleaning potential of the plant, permitting tighter quality control of clean coal output.