Introduction
With well over 70% of the total operating cost due to feedstock and
energy costs, optimal production planning and scheduling are
crucial business activities for Olefins and Aromatics (O&A) facilities.
These activities cover many
decisions such as:
- Which feedstocks should be purchased?
- When to schedule deliveries and material transfer operations?
- How do we manage the feedstock allocation for different cracking furnaces:
- On a monthly scale?
- On a daily scale?
- How should the olefins product distribution be targeted across
different time scales?
Optience provides a rigorous optimization system implemented
with the
SCMart Suite Platform that has enabled Olefins and Aromatics
facilities to answer these questions. Our methodology ensures that
planners develop the optimum feedstock purchase and production plans
resulting in significant financial benefits. An improvement in the order
of 1-3% of feedstock cost reduction for ethylene plants can generate benefits in the order
of tens of millions of US dollars annually.
Top
Methodology
A rigorous mathematical programming approach is used to develop the
optimal solution for the O&A planning and scheduling problem.
First, a model of the supply chain is built with the SCMart Suite
Modeler. The modeling system uses a flowsheet based approach where
the model can be graphically built according to the material flow
and business process.
The mathematical representation of this material flow is then generated
automatically. This representation may be a linear, nonlinear and/or a
combinatorial (mixed integer) optimization problem, depending on the
complexity of the processing units in the supply chain. Depending on the
nature of the mathematical model, pre-tailored solution strategies are
automatically employed in order to find the optimal solution.
The key differentiators in our methodology are:
- An intuitive flowsheet based supply chain management model resulting
in ease of maintenance and technology transfer
- Seamless handling of linear, nonlinear and mixed integer systems with
efficient algorithms for large scale optimization
- Single framework for planning and scheduling models, that ensures
that the optimized plans can be effectively implemented on a daily basis
- Ease of integration with your workflow
Top
The Planning and Scheduling Solution
The Planning solution maximizes the total profit over the planning
horizon by optimizing key decision variables such as:
- The total feedstock to be purchased with break down of quantity
and quality of each feedstock along with their shadow prices in each
time period.
- The allocation of feedstock type and quantity to each furnace in
each time period.
- The inventory, flow of materials, and conditions in the processing
units at each time period across the multi-site supply chain.
- All estimated final products for sale and transfer products in each
time period.
The data needed to generate the optimal plan can be input manually, or
directly transferred from your corporate database or Microsoft Excel
file into the SCMart Planning application. A rundown of typical input
variables is provided below.
- Market demands and the available feedstock basket can be input
as fixed values or as minimum and maximum values with priority definitions.
The SCMart planning solution typically satisfies the demands that will
provide the greatest economical profit, while considering the defined
priorities.
- Plant Constraints such as limits on process variables such as flows,
conversions, coil outlet temperatures may be entered into the planning
model. Discrete constraints such as furnace availabilities, vessel
loading/unloading durations, flow switches, tank transfer constraints
can be also be entered.
- Inventories at the start of the plan period should be entered.
These include feeds, intermediates, and final products. The storage
policy such as minimum and maximum inventories may also be entered for
each time period.
- Prices and Costs need to be entered and will be used to select the
feedstock and decide on final products that maximize the profit, while
satisfying all constraints. We support a wide variety of pricing and cost
patterns.
The Scheduling application considers a finer time scale, and optimizes daily
operations such as tank and vessel transfer operations, feedstock lineup to
the furnaces to ensure that daily production constraints are met while
ensuring optimum use of materials. The results obtained in the planning
module are automatically transferred to the scheduler module, which attempts
to maintain the optimal targets and objective set from the planning step.
In case this objective cannot be achieved, the scheduling module indicates
the reason and may provide guidance to revise the plan.
The information needed for the scheduling application is similar to that
for the planning application, except that inventory and plant constraints
need to be provided on a finer time scale, typically on a daily basis.
The scheduler can also be configured to run policies such as maximum
segregation between light and heavy Naphtha. In these cases, it considers
the quantity and composition of each tank in the feedstock tank farm and
the possible flow arrangement between them to optimize the naphtha and
condensate blends to preserve segregation over the schedule horizon.
The planning and scheduling solutions are both built with the same
SCMart Suite platform providing a consistent user experience across
the workflow.
Both applications present a gamut of customizable reports, which include
the total financial picture as well as specific information such as costs
important to each user ranging from sales, logistics, feedstock purchases,
and utility costs. In addition, the solution will be able to point out the
variables that are at constraints and the opportunity cost of relaxing these
constraints.
Top
Yield Model Development
A critical step in developing accurate plans is the development of reliable
yield models. The optimal feedstock blend to a furnace can be determined
only if the cracker yields are well correlated to feed properties.
The SCMart Estimator product is utilized to develop nonlinear correlation
models for the planning and scheduling applications. For olefins complexes,
the most important correlation is the cracker yield model with respect to
feed properties, rates and operating conditions. Some implementations
include correlations for reformers and gasoline blending operations when
they are part of the olefins complex. The module can use various data
sources, from the results of rigorous simulators to laboratory measurements
to estimate the parameters. Parameters in the yield model can be updated
easily from new data, and the resulting parameters can be shared seamlessly
with the planning and scheduling model.
Top
Workflow Integration
Our applications are built with a keen consideration of the workflow and
we build customized user interfaces to meet the specific demands of your
workflow. The end goal is to make the process smoother and simultaneously
ensure optimal decisions. A typical workflow in O&A facilities involves
the traders, planners and schedulers.
Visualization Interfaces are specifically designed for each decision maker in the workflow, for example:
- The Trader defines the possible universe of feedstock purchases
based on contracts or spot market purchases. Each feedstock has an
associated price and properties. This information is combined with the
information provided by the planning department on forecasted demands and
plant capacities to develop an optimal feedstock purchase basket and
associated shadow prices. The trader then attempts to secure this feedstock
basket.
- The Planner takes the purchase basket information from the trader,
and runs detailed cases to determine the optimal line up of feeds to each
facility and cracking furnace, the blending patterns, and plant operation
conditions for specified time horizon, typically ranging from three months
to a year.
- The Scheduler takes the objectives defined by the Planner to
generate a feasible and detailed scheduling solution for each day,
with the minimum deviation from the planning objectives of maximized
profit.
The application provides feedback when the planned targets cannot be
achieved at the scheduling step.
For additional information on our O&A solutions, please
email us.
Top