Tuesday, July 24, 2012

Various IT Applications in Plant Operation and How Linear Programming (LP) is used in Refinery

Linear Programming (LP) application software is used for process plant. It is designed to provide plants with an economic advantage in today’s highly competitive environment. This system uses feedstock properties, plant models, and economic considerations to help planners maximize profitability over a board operating range-both in conceptual and design stage as well as to optimize the operation. It takes into consideration of all the constrains and variables expressed in the form of linear equations. The ideal applications of LP model is where:
  • There are many potential solution
  • Certain objectives to be optimized
  • Interconnectedness between the variable elements of the system.
Oil refineries face an enormous number of options in their operations:
·         Which crude oils to refine
·         What processing conditions to use
·         Which products to sell
·         How to blend them from the intermediate components
There is a straightforward objective to arrive at optimized solutions: the profit. The operations of the refinery are intrinsically interconnected: it is a sequential process with one decision affecting the other. For example choosing to process one crude means that you have less processing capacity available for others. Thus the problems which a refinery faces have the characteristics of a LP solution.
A typical structure of LP software for optimization of a refinery (conceptual stage) as well as optimization of operation of existing refinery is shown in figure below.

Linear Programming (LP) Software Structure

Product Demand Pattern: The refinery need to be optimized not to exceed a specific product demand pattern of the market. This is normally a fixed parameter and called ‘constraint’ in LP modeling.
Product specification: It is fixed for a particular country or region, depending on the standard specification of salable products in the market. These are also called ‘constraints’ in modeling for optimization.
Selection of Process Units and Their Capacity: This gives the largest sets of variables. There is a wide range of choice of the processing units. Each gives a particular yield of products and particulars properties of the products to meet the specifications. The final product quantities are arrived at by the blending the intermediate products from various process units to meet the product specification.
Investment Costs: It will depend on the selection of process units, as each process unit will have different investment costs proportionate to its capacity.
Operating Costs: This again will depend on the selection of process units, each of which will have different operating cost heads like utility consumption, manning requirement etc.
LP modeling is carried out in the following manner:

·         Mass balance equations between process units, overall product balance and heat balance are expressed in linear equations.
·         Constraint equations such as product demand and specification by blending of components (intermediate products from the process units) are also used as linear equations. It defines the constraints of which products should be produced in the refinery and in how much quantity.
·         Process unit models predicting yield and quality of products based on crude oil characteristics, are built into a modern LP optimization software.
·         Equations for capital cost variation with capacity of the process units, cost of operation of each unit.
·         Overall cost optimization equations form the complete matrix of equations.
Non-linear models of processes (to give yield of products and product properties) and blending correlation for the properties form separate modules.
The parameter to be optimized normally is investment or profit margin.
Versatile LP software with built-n database and process models are available today.
Such models give option to change:-
·         Crude oil and product prices
·         Product specifications
·         The quantity of products
·         Plant sizes and operation modes
Thus a lot of business sensitivity factors can be studied using such models.
LP models are today used for:
  • Optimization of configuration of new refineries
  • Planning daily, weekly, monthly and long term operation of existing refineries
  • Optimization of operation of individual units
  • Evaluation of different types of crude oils



There are lots of functions required for operating data and other parameters which need to be collected and processed in operation of a plant. For these functions, Advanced Process Control (APC) System are used which combines with Distributed Control System (DCS) with process dynamic model and management into one.

Distributed Control System (DCS) In Process Plant

DCS has three essential features:
  1. DCS distributes its functions into smaller sets of semi-autonomous sub-systems covering specific process or geographic areas of the plant complex.
The functions generally are:
·         Data collection
·         Process control
·         Process analysis and Supervision
·         Storage and Retrieval of Information
·         Presentation of Information and Reports
2.  The second is to automate the manufacturing process by integrating advanced regulatory control logic and procedural languages with advanced applications packages, expert systems, including information to support such manufacturing enterprise application as:
·         Production scheduling and dispatching
·         Preventive and predictive maintenance scheduling
·         Information exchanges with business and logistics application
3.     The third characteristic is the system aspect of the DCS, which organizes information    flow between the constituent parts so that a single automation system unifying the semi-autonomous sub-systems is created.
4.     DCS has been extensively for all round application in operation, process control, maintenance, equipment availability etc.
Advanced Process Control Hierarchy (APC) (DCS/DDCS)

Dynamic Simulation Model and Advanced Process Control (APC)

Building the system model involves entering the details about each item in the process system. Much of the information needed to build the model is obtained during the design stage. It is always to create the model during the design stage and keep the model current through startup and operation.
But 100% safety is not possible and failures do occur. There are multiple independent safety layers and SIS as shown in figure.
Safety Layers Emergency Response Plan
As shown in figure, the final layer is that of emergency response. Every major hydrocarbon facility must have an Emergency Response Plan.

Dynamic model predicts responses of various equipment and process parameters due to any change in:
·         Feedstock quality or quantity
·         Operating conditions
·         Utility parameters (e.g. fuel gas quality for the furnace)
·         Price of Products
The software can have in built process optimization system. It calculates the new sets of operating conditions required for each part of the flow system to get the requisite yield and quality of products in the most economic way.
The program allows the operator to calculate new control set points to achieve optimum performance, carry out studies and determine where problems are occurring and what are the reasons.

In Advanced Process Control System, the model transmits the corrective actions required to the plant control system, which automatically resets the plant operating parameters.

Get Notified for new Tutorials:
*Check your email to confirm your subscription*


Post a Comment


© 2011 PIPING GUIDE - Designed by Ankit | ToS | Privacy Policy | Sitemap

About Us | Contact Us | Write For Us