Product blending is the process of allocating components to a product and the application of mathematical programming to produce the required solution.
This topic discusses the technology and methodology of fuel blend optimization.
Modes of Blend Optimization
- Single-Blend Offline Optimization – This is a process of finding the least costly component combination to keep the blend specs constant.
- Single-Blending Online control and optimization – This is a real-time adjustment of specific component rates. It enables the blend specs to be constant at all times.
- Multiple-Blend Optimization –This is the process where multiple blends are allocated multiple resources.
Processes of Achieving Blend Optimization
- Control and Manipulation of Variables-This are achieved by adjusting and manipulating various variables in a blend equation.
- Generating Sets of Blend Equations-This is done by placing the components into linear and nonlinear equations simultaneously.
- Realistic Sets of Equations-This refers to setting out both minimum and maximum specs by using equations.
- Transforming Inequality to Equality-A slack variable is added to an equation with inequalities to solve it. In this instance, the slack variable represents violation and giveaway. Again, it helps to provide a feasible solution.
- Final Sets of Equations-Minimum and maximum specs are solved as linear equations while introducing the slack variable to find a feasible solution.
- Finding an Optimum Solution-This is achieved by adding an objection term to the set of equations. The objective term has a driving force for the minimum deviation from the required specs of the blend quality, maximum profit, and minimum giveaway.
Methods of Solving Blend Equations for Optimization
- Linear equations
- nonlinear equations using linear optimization.
- Nonlinear equations using nonlinear optimization.
To ensure optimum profitability of a refinery, it should optimize its blend recipe. Depending upon the requirement of prediction accuracy and blend models, an optimization algorithm may be chosen.
Options for eLearning This Topic
#ProductBlending #Single-BlendOfflineOptimization #BlendSpecsConstant #Multiple-BlendOptimization #Single-BlendingOnlineOptimization #ControlManipulationVariables #LinearBlendOptimization #Non-LinearBlendOptimization #BlendModelsOptimizationAlgorithm #RealisticSetsEquations #GeneratingSetsBlendEquations