Mathematical Modeling and Simulation

Lecture 1: Introduction
Introduction: The Role of Modeling in Petroleum Refining
Welcome to the advanced course Modeling and Simulation of Chemical
Engineering Systems, focusing on petroleum and gas refining. Mathematical
modeling and simulation techniques are vital for solving complex
engineering problems, optimizing refinery operations, and improving safety
and environmental sustainability in the petroleum industry.
1. What is a Model?
Definition: A model is a mathematical or computational representation of a
physical system. It simplifies complex real-world phenomena by applying
theoretical principles to predict behavior under various conditions.
Purpose of Modeling in Refining:
1. Optimization: Models are used to optimize processes like distillation,
cracking, and reforming to maximize yield, improve product quality,
and reduce energy consumption.
2. Design and Scale-up: Before designing large-scale equipment,
engineers use models to simulate processes at a smaller scale. For
example, modeling a pilot distillation column before scaling up to
industrial size.
3. Safety and Environmental Compliance: Modeling enables
engineers to predict hazardous scenarios, such as the release of
flammable gases, and design safety systems like flare networks or
emergency shutdown systems.
Key Components of a Model:
1. System Representation: A model represents physical phenomena like
mass transfer, heat transfer, and fluid dynamics through equations
based on conservation laws.
2. Approximation: Remember, all models are approximations. Even the
best models cannot capture every detail of a complex refinery system
but provide reliable predictions under specific assumptions.
2. Types of Models
Empirical Models: Based on experimental data or operational experience.
These are useful when fundamental knowledge of the process is limited or
when complex phenomena need to be described in a simple way.
Example: Correlations for heat exchanger design, where empirical formulas
predict heat transfer coefficients based on measured data.
Semi-Empirical Models: A combination of theoretical and empirical
methods. These models apply fundamental laws but rely on experimental
data to fine-tune parameters.
Example: Reactor modeling in catalytic cracking units, where kinetic
parameters are derived from experimental data.
Theoretical (Mechanistic) Models: Based purely on first-principles and
fundamental physical laws (conservation of mass, energy, and momentum).
These models require a detailed understanding of the process and are often
more computationally intensive.
Example: Computational Fluid Dynamics (CFD) models used to simulate
fluid flow in reactors or separation units, accounting for detailed spatial and
temporal variations.

3. What is Simulation?
Definition: Simulation involves solving the mathematical model of a
process, either analytically or numerically, to predict system behavior.
Importance in Refining:
1. Cost-Effective: Simulation allows engineers to explore multiple
scenarios without the expense and risk of physical experimentation.
For example, optimizing the feed composition in a refinery can be
done in a simulation before actual implementation.
2. Safety: Dynamic simulations can predict how systems behave under
abnormal conditions, such as equipment failure or feed fluctuations,
allowing engineers to design better control systems and safety
protocols.
Simulation Methods:
1. Analytical: Used when exact solutions can be derived from the
equations. This is often possible for simple systems but becomes
difficult for complex processes.
2. Numerical: Most industrial processes require numerical techniques,
such as finite difference methods, to solve differential equations that
cannot be solved analytically. Tools like HYSYS and Aspen Plus use
these techniques to simulate complex refinery systems.

4. System Classification in Refining

Thermodynamic Systems:
1. Isolated System: No mass or energy exchange with surroundings.
Rare in refining.
2. Closed System: Energy exchange occurs, but no mass enters or leaves
the system. Example: a batch reactor during the reaction phase.
3. Open System: Both mass and energy can cross the system boundaries.
Most refinery processes, such as continuous distillation columns and
heat exchangers, fall into this category.
Phase Classification:
1. Homogeneous Systems: A single phase is present, such as a liquid
phase in a distillation column.
2. Heterogeneous Systems: Multiple phases exist, such as a liquid-vapor
mixture in a refinery’s distillation column or a reactor involving solid
catalysts.

5. Steady-State vs. Dynamic Models in Refining

Steady-State Models:
Definition: In a steady-state process, variables like temperature, pressure,
and flow rate remain constant over time. Steady-state models are useful for
designing and optimizing continuous processes that operate under stable
conditions.
Example: Designing a heat exchanger or distillation column where the inlet
and outlet conditions remain constant during normal operation.
Dynamic Models:
Definition: In dynamic systems, the process variables change with time.
Dynamic models are essential for understanding transient behavior during
startup, shutdown, or when responding to disturbances.
Importance in Refining: Startup and shutdown procedures, or responding
to changes in crude oil feedstock, require dynamic modeling to predict how
process variables like pressure and temperature change over time.
Example: Simulating the startup of a catalytic reforming unit, where
temperatures and reaction rates must be carefully controlled to avoid
damaging the catalyst.

6. The Importance of Dynamic Behavior in Refining

1. Process Control: Dynamic modeling is crucial for designing control
systems that maintain product quality and safety. For example,
controlling the reflux ratio in a distillation column to keep product
purity within desired limits.
2. Handling Disturbances: Refineries often face disturbances such as
fluctuations in feed composition or energy supply. Dynamic models
help predict how these disturbances will affect the process and how to
adjust operational parameters accordingly.
3. Start-up and Shutdown Procedures: Transient operations like
starting up or shutting down a reactor or distillation column involve
rapid changes in temperature and pressure, which must be carefully
managed to avoid equipment damage and ensure safety.
4. Example: Modeling a refinery flare system to handle excess gas
during an emergency shutdown. Dynamic simulation ensures the
system can safely handle pressure surges.
7. Systematic Model Building
Building a model involves several key steps, particularly in the complex
systems found in refineries.
1. Problem Definition: Define the scope and objectives of the model.
Are you modeling a specific unit (e.g., a distillation column) or an
entire refinery process? What are the key variables you want to predict
(e.g., temperature, pressure, composition)?
2. Identify Controlling Factors: Determine the key process variables,
such as reaction kinetics, heat transfer, mass transfer, and fluid flow.
For example, in a distillation column, the controlling factors might
include feed composition, reflux ratio, and operating pressure.
3. Evaluate Problem Data: Collect the necessary data, such as physical
properties (e.g., heat capacities, viscosities), reaction rates, and
experimental results. Data from laboratory-scale experiments can be
scaled up using models to predict industrial performance.
4. Construct the Model: Use the conservation laws for mass, energy,
and momentum to develop a mathematical model. Incorporate
equations for chemical reactions, phase equilibria, and heat/mass
transfer.
5. Solve the Model: Use analytical or numerical methods to solve the
equations. Software like HYSYS or Aspen Plus can handle the
complexity of refinery processes, allowing for the simulation of
different operational scenarios.
6. Verify the Solution: Check the model against known solutions or
experimental data to ensure it accurately reflects the physical system.
For example, compare the predicted product composition from a
simulated distillation column with actual plant data.
7. Validate the Model: Run the model under various conditions to see if
it provides accurate predictions across a range of scenarios. For
example, validate a dynamic model of a reactor by simulating startup,
shutdown, and disturbance scenarios.
8. Conservation Laws in Petroleum Refining
Conservation laws form the backbone of process modeling, ensuring that
key physical quantities are accounted for correctly in the system.
1. Mass Conservation: Mass cannot be created or destroyed. In a
refinery, this means that the total mass entering a system (e.g., crude
oil entering a distillation column) must equal the total mass leaving
(products like gasoline, kerosene, diesel, etc.) plus any mass stored
within the system.
2. Energy Conservation: The first law of thermodynamics states that
energy cannot be created or destroyed, only transferred or converted.
In refining, this law governs how heat and work are transferred within
a process, such as the heat required to vaporize a liquid feed in a
distillation column.
3. Momentum Conservation: This applies to fluid flow in pipelines,
reactors, and separation units, where pressure drops and flow rates
must be calculated. For example, pressure drop calculations are crucial
for designing the piping network in a refinery.

9. Batch vs. Continuous Processes in Refining

1. Batch Processes:
Operate in discrete steps, with raw materials added at the beginning,
processed, and then removed. Batch processes are inherently
dynamic, as process variables change over time.
Example: Batch polymerization, where reactants are charged into
a reactor, processed for a specific time, and then removed.
2. Continuous Processes:
Operate continuously, with materials constantly fed into the system
and products removed. Most large-scale refinery processes are
continuous and are typically modeled under steady-state conditions.
Example: A continuous distillation column that separates crude oil
into different fractions like gasoline, diesel, and kerosene.
10. Lumped vs. Distributed Parameter Systems
1. Lumped Parameter Systems:
Assume perfect mixing, where all process variables (temperature,
pressure, concentration) are uniform throughout the system. These
systems can be described using ordinary differential equations
(ODEs).
Example: A perfectly stirred tank reactor (CSTR) where the
temperature and composition are assumed to be the same
throughout the reactor.
2. Distributed Parameter Systems:
Recognize that process variables may vary across the system,
requiring the use of partial differential equations (PDEs) to account
for spatial and temporal changes.
Example: Heat exchangers, where temperature gradients exist
between the hot and cold streams, or distillation columns with
temperature and composition gradients along the height of the
column.

11. Software Tools for Simulation in Refining

1. HYSYS and Aspen Plus: These are industry-standard simulation
tools used to model and simulate the steady-state and dynamic
behavior of refining processes. They include libraries for various unit
operations (e.g., reactors, distillation columns, heat exchangers) and
allow for rigorous simulation of complex refinery configurations.
2. MATLAB: Often used for custom modeling and more complex
mathematical simulations that require flexibility in coding. MATLAB
is useful for solving systems of equations, performing optimization,
and running advanced simulations for process control.
3. Computational Fluid Dynamics (CFD): CFD tools simulate fluid
flow and heat transfer in more complex geometries. These are useful
for detailed equipment design, such as optimizing the flow distribution
in a reactor or heat exchanger.
12. Practical Application: Case Study
Let’s consider a case study of a refinery undergoing an upgrade to improve
the yield of diesel from a distillation column. Using simulation, the
engineers can:
1. Model the existing column to assess current performance.
2. Simulate changes in feed composition to evaluate how the column
responds.
3. Optimize the reflux ratio and tray configurations to maximize
diesel yield.
4. Perform a dynamic simulation to test how the column responds to
disturbances like changes in feed temperature or pressure.
By the end of the simulation study, the engineers can propose modifications
that are validated through the model, ensuring that the physical changes to
the refinery will result in the desired improvement without costly trial-and-
error experiments.

Conclusion: The Future of Modeling in Refining

As we look to the future, modeling and simulation will continue to be
invaluable in advancing petroleum refining. Emerging challenges such as
increasing energy efficiency, reducing carbon emissions, and integrating
renewable energy sources will require ever-more sophisticated models and
simulations. M.Sc. students in petroleum and gas refining engineering will
be at the forefront of these advancements, using tools like HYSYS, Aspen
Plus, and MATLAB to solve the industry's most pressing problems.