(755h) Advanced Optimization Approach for Managing Ethylene Business Planning and Scheduling Problem

Feedstock cost is one single major production cost in the ethylene business.  Managing feedstock selection is one of key factors in improving business competitiveness for an olefins business where two of the key products (ethylene and propylene) demands and prices can vary significantly with the market conditions.  For liquid feedstock based ethylene producers, managing the by-products through recycles, internal utility consumption and product sales is part of the decisions in evaluating optimal feedstock selection.  By optimally selecting and blending proper feedstock for given operation periods, we can ensure maximum benefits by matching the market conditions with our plant constraints.

In the past, we have used linear programming (LP) approach for managing planning problem.  This approach is based on open loop optimization and has served us well.  Through this process, we learned its limitation and key area where further improvements can be made byremoving some important assumptions that lead to improvement in model accuracy and more realistic representation of the real problem. 

We recently have implemented more advanced optimization approach that allows us to build nonlinear as well as combinatorial optimization models to address our previous limitations.  The new production model is based on closed loop optimization where we no longer need to make assumption on tear pricing that may affect the true optimization result as well as more accurately build the process constraints with improved yield model prediction of the overall processed feedstock (including closed-loop recycle streams) and utility balance.  Furthermore, we also integrate the olefins demands with our downstream processes; this approach allows us to globally optimize our value chain in order to obtain global profit optimization results.

With our new system, we can address our needs for monthly planning and daily production scheduling.  The model has been built in such a way that we can manage constraints at the monthly and daily level with unified approach for ensuring consistency in the direction of the optimization results.  In this presentation, we will present case studies to demonstrate the benefits of our new approach.