Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

QA For Polymer Webinar - FINAL

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Improve Process Performance with Aspen Polymers &

Reduce Energy Costs by Modeling the Entire Polymer Process


with Aspen Plus® V8.8
Combined Webinar Q&A

This document summarizes the responses to questions posed before and during the webinars.
The recording of the Improve Process Performance with Aspen Polymers webinar can be found
here and the recording of the Reduce Energy Costs by Modeling the Entire Polymer Process with
Aspen Plus V8.8 webinar can be found here. Addition questions should be directed to AspenTech
Support.

General Questions about Aspen Polymers

Q: Is Aspen Polymers compatible with Aspen HYSYS®?


A: No, Aspen Polymers is only compatible with Aspen Plus, Aspen Plus Dynamics, and
Aspen Custom Modeler.
______________________________________________________________________________
Q: Does AspenTech offer a consultancy service to use Aspen Polymers on our behalf?
A: Yes, Aspen Technology has a professional services team that can assist with Aspen Polymers.
______________________________________________________________________________
Q: Is Aspen Polymers implemented in Equation-Oriented (EO) mode?
A: No, not directly in Aspen Plus. However, you can export polymer models to Aspen Plus
Dynamics which is equation-oriented and can run in dynamic and steady-state mode. We have
open equation-based versions of the free-radical and Ziegler-Natta kinetics models which can be
used inside Aspen Plus Dynamics or Aspen Custom Modeler (ACM) (the other reaction models
run as closed-form models via procedure calls).
______________________________________________________________________________
Q: Can I model pressure relief scenarios for ZN bulk polymerization reactors?
A: For depressuring of polymer reactors we recommend using Aspen Plus pressure relief or
Aspen Dynamics pressure relief (not the simplified relief sizing in the safety environment). The
rigorous pressure relief models use batch or dynamic CSTR models so they account for
simultaneous reactions and relief. This gives better accuracy for runaway reactions or highly
exothermic reactions.
______________________________________________________________________________
Q: Can I incorporate heat transfer in the model?
A: The plug-flow reactor model includes the option of using heat-transfer specifications or
alternately a user-defined heat transfer subroutine. The heat-exchanger models work fine with
polymers. Other models, such as CSTR, can handle heat transfer indirectly, for example using a
CALCULATOR block to manipulate specified heat duty.
Q: Does Aspen Polymers include unit operations for devolatization of the polymer product?
A: You can use FLASH2 to assume phase equilibrium. There are some tricks for modeling
rate-limited devices, like wiped file evaporators using RPLUG or RCSTR or treating
devolatilization like a reaction that moves material from one phase to another. In Aspen Plus
V8.8, the dryer model was extended to simulate devolatization of solid polymers.
______________________________________________________________________________
Q: Would it be easy to check precipitation of a product in a stream where our solvent is going
to be replaced? We currently boil our solvent leading to precipitation of our compound in
water.
A: For conventional compounds, you would handle this using solution chemistry (define the
solid component as a salt) and an appropriate unit operation model, such as RBATCH if your
solvent swap is a batch operation. Polymers are handled a bit differently; we treat polymers as a
second liquid phase. The system would predict the precipitation of the polymer in the solvent by
showing the appearance of a pseudo-liquid polymer phase. This approach is used because
monomers and solvents dissolve in the polymer phase – it is not a pure like a crystalline solid.
______________________________________________________________________________
Q: When performing polymerization in a plugflow reactor with internal cooling tubes, is it
possible to accurately calculate the HT coefficient (U), such as in HEATX S&T?
A: You can apply correlations (like Nu = a x (Re^b) x (Pr^c)) to calculate heat transfer coefficients
using a simple heat transfer subroutine. The PFR model does allow you to specify multi-tube
geometry.
______________________________________________________________________________

Questions about Properties and Polymer Characteristics


Q: Is data for reactor rates included in the physical properties for the various polymers?
A: The only rate constants available from the property databases are the decomposition rates of
common initiators for free radical or emulsion polymerization.
______________________________________________________________________________
Q: Is it possible to get the chemical composition distribution (CCD) for a copolymer as HDPE
using Aspen Polymers?
A: Aspen Polymers predicts the number-average segment mole fractions (copolymer
composition). It also predicts the dyad concentrations and number-average sequence length for
block copolymers. For example, if you have styrene and butadiene, the system would predict
mole fraction of styrene repeats, mole fraction of butadiene repeats, concentration of
styrene-styrene, styrene-butadiene, butadiene-butadiene pairs and the average length of
styrene blocks and butadiene blocks.
______________________________________________________________________________
Q: Can I model bimodal molecular weight distributions?
A: Yes - or multimodal molecular weight distributions. We can track up to 10 sites.
______________________________________________________________________________
Q: Is it possible to include branching or other characteristics of the polymer (as the fraction of
the copolymer as a function of the length of the chain) in ACM using Aspen Polymers?
A: Yes, we can simulate short or long chain branching, tacticity, pendant/terminal double bonds,
cross linking, sequence lengths, dyad concentrations, and many other polymer characteristics.
Q: Can Aspen Polymers be used for polymer properties control?
A: This question is not entirely clear to me. You can relate polymer properties, such as intrinsic
viscosity (IV), melt index (MI), and density back to fundamental properties tracked in the
simulation (such as number average molecular weight and segment composition). Many
customers use process simulation to evaluate and resolve quality issues, for example by
optimizing reactor temperatures and holdups to minimize formation of undesirable by-products.
______________________________________________________________________________
Q: Can the EO solver handle polymer component attribute properties? If not, how does that
work in Aspen Plus Dynamics or ACM?
A: The reaction models return rates of change of each component and component attribute
based on current compositions, temperatures, pressures, etc. These models are called
procedures from inside ACM or Aspen Plus Dynamics (APD).

We also have open-form (equation-based) models for the Free-Radical and Ziegler-Natta models
which can be used in ACM instead of the closed form models from Aspen Plus. The stream
structure in APD is designed to track component attributes included in Aspen Polymers. We
have not extended the stream structure in equation-oriented Aspen Plus, so within the context
of Aspen Plus you must use sequential-modular simulation. (Of course you can export a steady-
state model to Aspen Plus Dynamics and run it in steady-state mode, which gives you the power
of EO solving.)
______________________________________________________________________________

Questions about Reactions and Kinetics


Q: Can I use bench scale data (batch) to estimate kinetic parameters from the Ziegler Natta
catalyst? Are there any examples on the support page?
A: Batch data can be useful for fitting the key rate constants. Propagation rate controls
monomer conversion. Other reactions such as chain transfer and termination, influence number
and weight average of molecular weight. Visit AspenTech Support for examples and more
information.
______________________________________________________________________________
Q: Can I estimate the ZN catalyst deactivation using temperature profile data from the bench
scale batch reactor?
A: Yes, if you have enough measurements you should be able to tweak the various catalyst
deactivation reactions to match the data. You will need several isothermal experiments to get a
handle on this.
______________________________________________________________________________
Q: Is it possible to model an emulsion polymerization reaction?
A: Yes, we have an emulsion polymerization module. We can also model suspension reactions
using a two-phase reactor model and any of the polymer kinetics modules.
Q: We have distillation actively separating one of the products during the reaction. Can a
similar model be built for batch reactions, where one of the products generated is continually
removed to ensure the reaction equilibrium moves forward?
A: There are two options for modeling polymer batch reactors with vent condensers or
overhead columns. The first option is to build the model in Aspen Dynamics. A second option is
to use the “clock method”, which uses a sensitivity block to take time steps around an Aspen
Plus model.
______________________________________________________________________________
Q: I work in a gas phase PP plant, and know that by using a different catalyst, the parameters
could change. How can I address that in the software?
A: Correct. Each catalyst has unique reaction properties and may generate polymers with a
different molecular weight distribution. We handle this in the model by adjusting the rate
constants of various reactions for one or more (up to ten) unique active sites. We have an
Excel-based tool that can be used to take a measured MWD and apply statistical analysis
“deconvolution” to determine the number of types of active sites expected based on the MWD
curve (each site generates the most probable distribution, so a catalyst with two sites may
generate a bimodal MWD).

Customers who use many different catalysts (to produce different product grades) often use a
CALCULATOR block with a single user parameter to switch grades. For example, IF (GRADE=1)
THEN…ELSE… and within each section of the logic set the rate parameters for grade 1, grade 2,
and so on. You can do something similar outside the model using Excel and Aspen Simulation
Workbook. Alternately, define a REACTION model for each catalyst and give the catalyst unique
names. Reference all the REACTION models in each reactor. The user would specify which
catalyst is used in the feed stream and the model would then assign the rates based on the
catalyst that is actually present.
______________________________________________________________________________
Q: Can we model Polymer thermal decomposition reactions in Aspen Plus?
A: Yes, Aspen Polymers includes a general kinetic model (segment-based power law) that can be
used to simulate pyrolysis (thermal cracking) of polymers back to simple molecules. Several
customers have applied this model to simulate recycling and gasification schemes. The
step-growth model can be used to simulate hydrolysis (steam cracking) or alcoholysis of
polyesters, nylons, and other step-growth polymers.
______________________________________________________________________________
Q: What approach is used for polymerization kinetic modeling? Is moment method,
monte-carlo, or other numerical methods used? I was told that Aspen Polymers can only
perform the method of moment approach, is that right? How do you get the MWD, because
method of moment can only get the average MW not MWD, unless the MWD function
assumed beforehand?
A: Most of our reaction models use the method of moments to track the number and average
molecular weight, branching frequency, and other polymer attributes. We also have an option
of tracking the full MWD for additional polymers (Free-Radical, Ziegler-Natta, Metallocene, and
Emulsion). These models use “method of instantaneous properties” developed by Professor
Archie Hamielec of McMaster. This method assumes that each site produces the most probable
distribution at each instance of time, allowing one to perform numerical integration to track all
the elements of the molecular weight distribution. The method does not work when you have
extensive branching (network formation) or when you have reversible reactions (ionic & step-
growth).
Q: My company is researching a new set of polymer reactions. Can a custom set of properties
be created in Aspen Polymers?
A: The reaction models are very flexible. There is a general purpose “segment-based power law”
model that you can use to define reactions in terms of conventional components and polymer
segments. We often use this model together with built-in models (like Free-Radical). The
segment-based reactions can be used to capture thermal scission or other special reactions.
There are a wide range of built-in component attributes to track properties (branching, dyads,
pendent double bonds, terminal double bonds, etc.). There are also five “user attributes”, each
with ten elements that could be used to track additional conserved quantities (you would need
to use a user kinetics routine to return rates for these).Of course another option is to use Aspen
Custom Modeler which is completely flexible.
______________________________________________________________________________

Questions about Polymers as Solids Below the Melting Point


Q: Is the vapor feed to the distillation column a "vent" stream or a "product" stream?
A: Aspen Plus includes a batch reactor and a batch distillation model. The batch models allow
for charge (batch) feeds, as well as continuous feeds and products. Accumulator models are
used to simulate receivers for the vent/distillate streams.
______________________________________________________________________________
Q: What are the dynamic functionalities (if any) of solids handling from Aspen Plus?
A: We have not yet extended Aspen Plus Dynamics to address solids. We do have an example of
a batch dryer that uses a clever trick in Aspen Plus to emulate dynamics. The example is posted
on aspenONE® Exchange.
______________________________________________________________________________
Q: Have you modeled a dense phase transfer? Also, have you modeled extrusion processes?
A: We had a pilot program about 10 years ago to develop rigorous models for extruders which
could calculate shaft work required, viscous heating, etc. We had some success but ended up
shelving the project because we could not get enough detailed geometry data from extruder
vendors and because the models were complex. Generally, most customers just treat extruders
as plug-flow reactors (using RPLUG).

Regarding your question on a dense phase transfer, it’s not clear what your question is. For
mass transfers between polymer and liquid phases, we deliver an example in V8.8 showing how
to model solid/liquid extraction columns for the nylon-6 process.
______________________________________________________________________________
Q: What about extrusion and pneumatic conveying of pellets, for example LDPE? Can you
model a dense phase solid transfer?
A: We can model pneumatic transfer of polymer particles, gas/solid and liquid/solid separations,
and other operations involving polymer particles. The solids models in Aspen Plus V8.8 have just
been extended to address solid polymers. Use the PIPELINE model in Aspen Plus to simulate the
transfer lines.
______________________________________________________________________________
Q: Can you explain the depressurization of reactors containing solids?
A: You can model reactive pressure relief involving polymers using Aspen Plus or Aspen Plus
Dynamics.
Q: Can solids modeling handle chemical reactions, such as crystallization?
A: For conventional solids, you use CHEMISTRY to define crystallization. For example:
NA+ + CL - NaCl(s)
H2O(l)  ICE(s)
The “Ksalt” is used to specify the solubility equilibrium.
Currently, the crystallizer is limited to a single solid. We are evaluating an extension of the
model to address multiple competing crystalline products. You can model semi-crystalline
polymers by either specifying the parameter POLCRY (mass fraction crystallinity) or by moving
some polymer from MIXED to CISOLID using a reaction model or RSTOIC.
______________________________________________________________________________

Questions Specific to Webinar Demonstrations


Q: During the section that covered reactions, the presenter showed a database with typical
reactions and some kinetics associated with the reactions. Is that what it was, or were those
kinetics specific to the problem?
A: There is a database of suggested reactions to get you started, but you should adjust that set
of reactions for your model. The kinetics for those need to be specified and are not suggested by
Aspen Polymers.
______________________________________________________________________________
Q: Is the flash tank calculation equilibrium-based or is there a full transport model?
A: By default, Aspen Plus assumes phase equilibrium in the reactor models. However, we have
examples showing how to account for mass-transfer limited reactors. This may require some
user models, however.
______________________________________________________________________________

Questions about Examples and Further Resources


Q: Do you have any specific model of Spheripol Process of PP and Sclairtech Process of Swing
PE Plant?
A: We have a basic PP example delivered with the product. We have developed many high
fidelity Spheripol and Unipol models for our customers through our services organization.
______________________________________________________________________________
Q: How can I access the best practices for polymer modeling generated by AspenTech?
A: We publish examples of reactor models and some industrial process models. A set of
examples are delivered with Aspen Plus. For versions V8 or higher, go to the Examples button
on the Get Started or Resources ribbon tab and look for a sub-folder called ”polymers”.
Additional examples are posted in the support knowledgebase and via aspenONE Exchange. We
also offer classroom or on-site training for polymers, as well as services projects to build turn-
key models.
______________________________________________________________________________
Q: Do you have polymers in solution examples?
A: Yes, we have many examples available in Aspen Plus. Go to the Resources ribbon, click the
Examples button, and navigate to the Polymers folder.
Q: Are there any examples of Aspen DMCplus and Aspen Polymers for a control application?
A: No, unfortunately there isn’t.
______________________________________________________________________________
Q: Can Aspen Plus be used for PET production?
A: Yes, we have a set of very detailed models for PET and PBT. The PET models include
examples of the Zimmer, Hitachi, Toray, and DuPont processes. The models use user prop-set
properties to calculate and return predictions for common measurements, such as intrinsic
viscosity, acid value, DEG content, and many other factors. The models include dissolution of
solid pTA (equilibrium or mass-transfer rate-based). The polymerization reactors use a
mass-transfer rate-based approach (there is a kla parameter that you adjust to match the
reactor performance). The reaction kinetics considers an antimony catalyst, phosphate
inhibitors, and other additives such as Co, Mg, Mn. We offer the models “as is” on request. You
may send David an email at David.Tremblay@aspentech.com for more information.

You might also like