Jump to content



Featured Articles

Check out the latest featured articles.

File Library

Check out the latest downloads available in the File Library.

New Article

Product Viscosity vs. Shear

Featured File

Vertical Tank Selection

New Blog Entry

Low Flow in Pipes- posted in Ankur's blog

3

Design Of A Cumene Production Facility


41 replies to this topic
Share this topic:
| More

#1 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 21 March 2019 - 10:24 PM

Attached File  PFD.png   326.99KB   3 downloadsGiven the output of cumene production, how do I go on about performing material balances on every single unit in a process. My initial thoughts of approaching this problem was that I look in the literature for the kinetics of the reaction from which I can assume the conversion and the ratio of the reactants in the feed to the reactor. I went to ask my professor about this question, he said that I should use the summary table of the molar flow rates for the process flow diagram I selected. I am not sure this is right because that would yield the exact numbers as in the table secondly I dont have the summary table of the flow rates.

 

Attached Files


Edited by AAAIK, 21 March 2019 - 10:28 PM.


#2 PingPong

PingPong

    Gold Member

  • Members
  • 1,270 posts

Posted 22 March 2019 - 03:59 AM

Attached File  --- Cumene Production – Change in Reactor Conditions.pdf   127.37KB   46 downloads



#3 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 23 March 2019 - 08:48 AM

I have a question that bothers me until now,  is the flowrate, temperature, pressure, vapor quality of steams entering or leaving something that should be given in a design problem or is it inferred from putting the design specifications of the reactor along with the reaction kinetics, and separation process as the input to a software?



#4 Art Montemayor

Art Montemayor

    Gold Member

  • Admin
  • 5,698 posts

Posted 23 March 2019 - 11:39 AM

AAAIK:

 

It astounds and concerns me that a chemical engineering student, once given the excellent and almost exact model of a problem solution, should be still unsure of what to do as a first step in arriving at a specified and logical answer.  PingPong, in his usual and professional manner has handed you, in my opinion, the perfect answer to your query.  He has left out only the most valuable and important engineering asset needed to simply input into a simulation program: common sense and reasonable, logical assumptions.

 

In my opinion, those are the engineering qualities being indirectly sought by the authors of the problem statement.  The nature of the problem is to generate “a preliminary design” to get an idea of the possible economics in generating cumene.

 

The information PingPong has given you is a copy of the exact basic problem statement - edited in another quiz version.  Even the unit “800” is the same - as well as the entire flow diagram and equipment tags!  Only the production rate has changed from 79.4 t/yr to 100M t/yr.  Common sense dictates that the conditions of temperature and pressure are those prevailing at the local site - which are traditionally taken to be locally ambient, given no other information.  As to the flow rate, you are given the flow rate in the capacity figure.  What is your problem?  In my opinion, you have the answers to your own questions.  Put yourself in the place of those that have to make a managerial decision in whether to invest millions of $ or not.  That is the basic, principal work of an engineer out in the process industry: deciding how to best make a profit instead of an investment loss.

 

You are being asked to make a recommendation as to whether it makes sense to invest or not, given the information you have been given.  What you are lacking are basic, common sense assumptions in order to arrive at a reasonable estimate.  That, in my opinion, is the heart and nature of this “problem” assignment - not the numerical results of any simulation program.  This type of problem is typical of what occurs out in industry and is important as a teaching tool.

 

I believe I have complemented PingPong’s excellent response with specific comments, but I’ll leave it up to him to follow up with additional remarks and possibly critiques if he feels any are needed.

 

In my opinion, you have all the information to arrive at an engineering decision.  As a lesson learned, you can never expect the computer - or much less a simulation program - to use common, rational sense to an engineering solution.  That is an input originating only in the world’s perfect computer - your god-given brain.



#5 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 23 March 2019 - 04:33 PM

Well, I am sorry if my question was dumb to you sir, it seemed reasonable for me to ask such a question but that does not matter.  I want  you to be aware that  I am a student in an educational system where the number of lectures per core subject namely (fluid mechanics, thermodynamics, heat and mass transfer, control, chemical reaction engineering, and economics, design etc) are 9-12 which is by no means an appropriate amount of material to benefit from a course. I am trying to learn and understand Chemical engineering, I desire to do so before finishing my bachelor degree. I thank you and Pingpong for your time to help, I will read  the pdf pingpong sent again carefully and ask the 'right' question



#6 Art Montemayor

Art Montemayor

    Gold Member

  • Admin
  • 5,698 posts

Posted 23 March 2019 - 10:49 PM

AAAIK:

 

Please read my comments carefully.  I certainly am not admonishing you or criticizing you.  That is not my intention.  Rather, I have tried to specifically show you where you are not putting emphasis on your learnings regarding chemical engineering.  You have not focused on my claim that the answer to your uncertainty of how to solve the problem.  I insist that the answer on how to solve it is in the document that PingPong has furnished to you.

 

I also claim that the flowrate you mention that you need is specifically given to you as the capacity expected of the proposed process size.  Do you agree with me?  Please respond.

 

I also claim that you can easily and rationally claim local ambient conditions (pressure and temperature) for your feed(s), utilities, and products.  The temperatures within the cumene process are given by the process reactions and steps.  The simulation program should yield the majority of this information - and you can check this data with the information PingPong furnished.  In other words, your computed answers should be almost a capacity scaleup of what PingPong uploaded.  Do you also agree?

 

I believe that I have responded to your query and I believe that your instructor was trying to point you in that direction as well.  Once knowing this information you have acquired the basic data to do the actual simulation or hand calculations to obtain the size of the equipment needed and from there obtain the capital and operating requirements to get to the final economic evaluation as to the rate of return on the investment.

 

As I see it, your first obstacle has been obtaining all the necessary basic data needed to initiate the routine process calculations/simulations needed to identify the process capital and operating costs - and this initial step is what I have tried to help you identify and learn as a method on how process engineers would deal with such a challenge.  Engineering is a very challenging study and we all know that.  Included in learning engineering is the fact that we are basically professional problem solvers.  In the real world, not all of the basic data is given or available to us.  However, in order to arrive at a reasonable estimate we often apply common sense and rationality to obtain estimates to problems.  Note that you were not given a detailed listing of ALL the classical needed basic data for design.  However, it is noted to you that the result you should obtain is a PRELIMINARY DESIGN for a decision based on estimated capacity and conditions.  Therefore, it is only reasonable to deduce that your results will be interpreted as an estimate and will never be used for final design or fabrication.  You should declare this in your final report.  Common sense tells me that if not all the required basic data is given to me, then I must resort to using rational, estimated data that I believe can be reasonably applied to obtain the estimated equipment sizing and costs.  And I should qualify this in my report.  To confirm this kind of rationalization, check with your instructor.

 

Please do not feel admonished or scolded.  We are here to help students and not to criticize them when we see they need help or advice.  We can be critical of how a student attacks a problem, but we don’t criticize them.  You have asked questions that are answered simply because they involve only the ability to employ common sense and rational thinking which should be a characteristic of all engineers.  You are a student and if you learn to employ these traits you will profit from them.  Please let us know if we have not been of help or if we have failed to communicate successfully.

 

And, I emphasize, your questions were not “dumb”.  There are no dumb questions in engineering.  That’s how we all learn.



#7 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 24 March 2019 - 08:38 AM


 

 

I grasped the idea of the problem statement and what’s expected as the outcome (whether its profitable to invest or not).

 

I went to my instructor and had a conversation on how to start, he said to me “you should do an overall mass balance on each unit of your process flow diagram”.  To which I replied, “can I use the flow rate from a summary table of properties of streams someone sent to me for the exact process flow diagram and production rate or should I assume an overall conversion and benzene:propylene ratio?”  He said, “yes you could assume selectivity or conversion and a benzene:propylene ratio but these assumptions should be justified”.  After doing so, he said to me: “you are required to do mass balances on each unit.  In the reactor assume a single pass conversion and, in distillation, how much of the feed is distilled and their concentrations.  After finishing material balance, do energy balance using any software of your choice.  Then comes, after the design of the reactor and separation process, an estimation of the equipment”.


Edited by Art Montemayor, 24 March 2019 - 09:40 AM.
Spelling, composition, grammar


#8 Bobby Strain

Bobby Strain

    Gold Member

  • Members
  • 2,704 posts

Posted 24 March 2019 - 09:01 AM

This exercise is well beyond the capabilities of a student. Seems like you have a lazy instructor.

 

Bobby



#9 breizh

breizh

    Gold Member

  • ChE Plus Subscriber
  • 4,178 posts

Posted 24 March 2019 - 09:14 AM

Hi AAAIK,

Every chemical engineering  project is based on Mass and Energy balances . You have been lucky that PingPong shared with you his file . This will save you some time but you still need to perform the calculation with all the assumptions by any means .

Definitely your task is big but there is little option , your instructor is the boss!

 

Good luck

Breizh



#10 PingPong

PingPong

    Gold Member

  • Members
  • 1,270 posts

Posted 24 March 2019 - 11:50 AM

As I happened to come across this particular student problem not too long ago, I recognized the PFD posted by the topic starter.

 

It probably originates from Turton's book:

https://books.google...epage&q&f=false

and nowadays seems to exist in different forms.

 

See also:

 

Attached File  --- Problems at the Cumene Production Facility.pdf   283.86KB   27 downloads

 

Attached File  --- Cumene Production.pdf   390.12KB   25 downloads

 

which contain additional info that could be useful to AAAIK.

 

 



#11 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 27 March 2019 - 12:59 PM

For the overall  mass balance I assumed the selectivity to be equal to 33 (this was based off the numbers in the molar flow rate summary table pingpong has sent). I converted the desired production rate from mass to moles, from which I calculated the extents for reaction 1 and side reaction 2 and the molar flow rates of the feed (benzene and propylene) assuming 100% conversion of reactants. But I am not sure of this assumption (100% conversion). At this point ( I think my professor asked to perform the overall material balance as if we are still in the Input / Output Structure of Flowsheet but I am not sure) should I include the recycle stream of benzene unreacted, and how to I determine its amount if I were to take to recover and recycle all the unreacted benzene. If what I did earlier is right, Should I proceed to do material balances over each unit in the process flow diagram using the feed flow rates I obtained from the overall material balance deciding a single pass conversion for the reactor and reiterate?


Edited by AAAIK, 27 March 2019 - 03:41 PM.


#12 PingPong

PingPong

    Gold Member

  • Members
  • 1,270 posts

Posted 27 March 2019 - 03:30 PM

But I am not sure of this assumption (100% conversion).

100 % conversion of reactants is not possible.

For example: part of the propylene is lost in <9>

 

All reactions are limited by equilibrium. This is an exothermic reaction so the lower the reactor temperature the further the equilibrium moves to the right and consequently higher the conversion would be. Reactor temperature however also has an impact on kinetics.

By recycling benzene the B/P ratio increases and that helps to move equilibrium to the right. Moreover the higher the B/P ratio the less diisopropyl benzene byproduct will be formed. However a high B/P ratio increases reactor size and unit energy consumption.

 

To reach equilibrium one would need an enormous reactor, so in practice there will be a so called Approach To Equilibrium (ATE) based on a practicle reactor size. I don't know what ATE is normal for this particular reaction but you could try to back calculate it using the data from the material balance table and applicable thermodynamics for equilibrium constant calculation.


Edited by PingPong, 27 March 2019 - 04:17 PM.


#13 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 30 March 2019 - 09:35 AM

PingPong:

 

But why are you bringing up the reactor design at this stage?  I know that the amounts of unreacted benzene, propylene and propane impurity in the propane are governed by the reactor design (Mode of operation, Reactor type, packed or not, temperature and pressure) All these design variable do affect the per pass conversion of propylene, which in turn affect the decisions of should we recover and recycle unreacted propylene or vent it with the neutral impurity propane.. etc. 



#14 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 30 March 2019 - 09:47 AM

I adjusted input-ouput structure of the flowsheet, I assumed that the all the fresh propylene is converted ( but the per pass conversion of propylene <1)



#15 Bobby Strain

Bobby Strain

    Gold Member

  • Members
  • 2,704 posts

Posted 30 March 2019 - 06:37 PM

20 years ago the zeolite process converted all the propylene fed to the reactor. That probably hasn't changed.

 

Bobby



#16 PingPong

PingPong

    Gold Member

  • Members
  • 1,270 posts

Posted 31 March 2019 - 05:09 AM

The cumene process scheme that AAAIK is using is not based on the modern liquid phase reaction over zeolite catalyst, but on the old high temperature vapor phase reaction over (presumably) SPA catalyst.

 

It originates from Turton's book and is used by others who sometimes made changes to it in order to "optimize" the process.

For example Luyben does this in a chapter of one of his books:

https://books.google...AERNoC&pg=PT179

As there are always pages missing in google books I will also post the relevant part here:

Attached File  Luyben Chapter 9 - Design and Control of the Cumene Process.pdf   1.26MB   26 downloads

 



#17 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 03 April 2019 - 07:59 PM

Decisions that determine the input/output structure flowsheet

Vent gaseous reactant, propylene and gaseous feed impurity, propane so that the inert-propane gas does not accumulate in the gas recycle loop.

Remove DIPB and use it as part of the fuel supply. It is more economical to pay for the increased raw material cost (propylene) that was converted to DIPB than to oversize all equipment in that recycle loop.

Recover and recycle excess Benzene 

 

Decision That determine the Recycle Flowsheet

Since both reactions take place at the same temperature and pressure and with a catalyst, only one reactor is required

Use of Excess benzene leads to increased selectivity which can be seen from Selectivity= rate (desired)/Rate (undesired).

 

 

Formulation of the Recycle material as a function of the dominant design variables, selectivity, molar ratio of Benzene to Propylene, mole fraction of propyelene in  feed and vent.

FG-Molar flow rate of Propylene+Inert  impurity (propane)

FFB- Molar flow rate of Fresh Benzene

FB- Molar Flow rate of Benzene 

RB- Molar Flow rate of Recycle Benzene

PC, PDIPB, PG- Molar Flow rates of Cumene, DIPB, fuel gas (Vent)

YFP, YPP- Mole fractions of propylene in feed and vent

Selectivity definition I used= Moles of Desired product formed (Cumene)/Moles of Undesired Product (DIPB)

 

The images below are my attempt to formulating the recycle material balance. To check if I am heading in the right direction, If I were to plug in values for the design variable the resulting numbers should be equal or close to the flow rates in the table as I assumed for instance, that 100% of benzene is recovered and recycled. But I am facing a difficulty, I tried to estimate the equilibrium coversion by setting the mole balance and concentration equation for all the components at the reactor outlet and considered a PFR reactor operating isothermally but I cant proceed from this point.

Another Difficulty I am facing is what composition can I assume for the propylene in the vent and on what basis 

 

 

Attached Files



#18 PingPong

PingPong

    Gold Member

  • Members
  • 1,270 posts

Posted 04 April 2019 - 04:59 AM

But I am facing a difficulty, I tried to estimate the equilibrium conversion by setting the mole balance and concentration equation for all the components at the reactor outlet and considered a PFR reactor operating isothermally but I cant proceed from this point.
I don't think anybody understands what you are doing.

Moreover: what do you mean by "equilibrium conversion" ? That would require equilibrium constants both for the cumene and the DIPB reactions. To account for that better use a process simulator.

Another Difficulty I am facing is what composition can I assume for the propylene in the vent and on what basis
The composition of the vent gas depends on the composition of the reactor effluent and the temperature and pressure in separator V-802. Doing this by hand would be very time consuming, again better use a process simulator.

#19 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 10 April 2019 - 09:01 PM

I am doing a Material balance on the reboiler and condenser on the benzene distillation column, on what basis can I choose the reflux ratios.



#20 PingPong

PingPong

    Gold Member

  • Members
  • 1,270 posts

Posted 11 April 2019 - 04:39 AM

A reflux ratio is not something you can choose.

It results from the product specifications (cumene in distillate, benzene in bottoms), the number of theoretical stages, and the pressure and consequently temperature the column is operating on.

 

In this case I suggest you simply use McCabe-Thiele for separation between benzene and cumene.



#21 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 20 April 2019 - 09:37 AM

I am now designing the Packed bed reactor using polymath, I wrote down the differential equations ( Mole balance equations, Reactor temperature, and temperature of coolant in terms of z (the length of the reactor). I assumed the length of the reactor z= 6m. But, I have a problem with working this way  shouldn't I be looking for the design parameters concerned with the sizing of the reactor such as diameter, weight of catalyst which are unknown to me,how can I find them, because in order to the temperature, molar flow rate of species, pressure profile profile versus the length the design parameters should be specified. 



#22 PingPong

PingPong

    Gold Member

  • Members
  • 1,270 posts

Posted 21 April 2019 - 09:09 AM

I suppose you need to make a general reactor model that allows you to adjust the length and diameter of the tubes, number of tubes, and tube pitch.

 

Catalyst volume (and weight), pressure drop, heat transfer area, reactor diameter, et cetera, is then resulting from those choices. You keep adjusting (trial and error) until you reach a satisfactory result.



#23 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 22 April 2019 - 01:05 AM

The general mole balance equations based on the length of the reactor in differential form

d(FP) / d(Z) = -A*(rhoCat)*(k1*CP*CB+k2*CP*CC) #Molar flowrate of Propylene

 

 

 Is A the total cross sectional area of the tubes or is it the Cross-sectional area of a single tube?

 

I applied the energy balance equation relating reactor length and temperature of the reactor and the coolant as well. But I am missing the data for the mass flow rate of coolant and heat capacities as a function of Temperature for Cumene, Benzene, Propane, Propylene, and DIPB. So do you suggest I set the temperature of the coolant as constant to simplify matters.

I also used Ergun Equation to account for the pressure drop in the packed bed reactor (PBR). But I am also lacking  the viscosity for the stream entering the temperature. 

 

I also can't relate these equations differential equations to the number of tubes so that I can adjust it to reach the desired conversion of propylene and satisfactory temperature outlet ( below the temperature limit of 400 degrees Celsius)


Edited by AAAIK, 22 April 2019 - 01:21 AM.


#24 PingPong

PingPong

    Gold Member

  • Members
  • 1,270 posts

Posted 22 April 2019 - 07:13 AM

The reactor is a plug flow reactor with catalyst in many parallel tubes, so you really have many parallel identical reactors, each handling part of the total feed. For example: if you would have 100 parallel tubes then the feed to each tube would be 1 % of the total feed. In other words: for n tubes the feed to each tube is 1/n th of the total reactor feed.

 

The coolant is boiling water generating steam at a constant pressure and temperature (steam table). The amount of steam produced depends on the duty transferred from the catalyst tubes to the boiling water. Obtain wetted tube area, heat transfer coefficient U, and LMTD for each part dZ of the tube.

 

For physical properties of the gas mixture you can use software.

Or do it by hand, for example using the well known book The Properties of Gases and Liquids by Poling, Prausnitz & O'Connell.



#25 AAAIK

AAAIK

    Veteran Member

  • Members
  • 34 posts

Posted 23 April 2019 - 01:45 AM

1) I wrote down the mole balance equations for Propylene, Benzene, Cumene and DIPB based on the reactor length in differential form 

FA-Molar Flow rate of Benzene

Where

A=N*pi/4*D^2

N-Number of Tubes

D-Tube Diameter

From the problem Statement, the units of the activation energy are kcal/mol, the uhits of concentration are mol/l the temperature is in Kelvin and the rate is in mol/gcat*s.

2) Rate equations 

r1 =K1CACB mole/g cat sec

r2=K2CACB mole/g cat sec

3) Expressing rate equations as a function of (Flow rate, Temperature and Pressure)

CB-Cncentration of Benzene

 

CT0- Feed Gas Total Concentration

T0-Inlet Temperature

P0-Inlet Pressure

4) Energy balance-For steady State operation and no shaft work in terms of reactor length 

T-Reactor Temperature

5) Evaluating the energy balance parameters:

Heat of reaction

 

 

Heat Transfer Coefficient term

U*a=4/D*U

In the Energy balance (Temperature variation with reactor temperature), Does A express the crossectional area of a tube and the flow rate in the denominator (Flow rate*molar heat capacity) the flow rate in one tube?

No info about the temperature of the medium Ta is given in the problem statement so do I just assume saturated water at the inlet pressure from the inlet pressure and use the heat of reaction given off to generate steam?

There are 5 coupled Differential equations that must be solved simultaneously I ran polymath but the reactor temperature decreased by 1 kelvin (nearly invariant).

Attached Files

  • Attached File  1.png   3.21KB   0 downloads
  • Attached File  2.png   2.88KB   0 downloads
  • Attached File  3.png   8.6KB   0 downloads
  • Attached File  4.png   4.82KB   1 downloads

Edited by AAAIK, 23 April 2019 - 01:48 AM.





Similar Topics