Computer simulation of Dinitrotoluene Nitration Process

(1)

Computer simulation of

Dinitrotoluene Nitration Process

Master of Science in Engineering, Degree Programme in Chemical Engineering

Datasimulering av Dinitrotoluen Nitreringsprocess.

Moses Ruhweza

Faculty of Health, Science and Technology Subject: Process Engineering

Points: 30 ECTS

Supervisors: Lars Nilsson and Lars Stenmark Examiner: Lars Järnström

Date: 2018-01-19

(2)

Computer Simulation of Dinitrotoluene Nitration Process Moses Ruhweza

Department of Engineering and Chemical Sciences Karlstad University

Abstract

This paper presents an approach for modelling a commercial dinitrotoluene (DNT) production process using the CHEMCAD simulation software. A validation of the model was performed based on results of an experimental study carried out at Chematur Engineering AB, Sweden.

Important parameters such as fluid properties, temperature profile and other operating conditions for CHEMCAD steady state model were selected so as to obtain the crude DNT yield as well as the acid –and organic phase compositions within the same range as the reference values from the experimental study. The results showed that the assumption of the steady state model was correct, and that acid – and organic phase compositions were in good agreement, although with a slightly lower sulphuric acid concentration than that observed in the experimental study.

Also, a detailed study was carried out to analyse the effects of physicochemical conditions on the desired product yield. Both the results from the experimental study and the simulated model agree that the effects of mixed acids or heats of mixing acids contribute significantly to the energy balance.

For the appropriateness of the thermodynamics, a NRTL model was chosen and the reactor system was optimized by an equilibrium based approach, producing MNT in 99.8% yield and crude DNT in 99.9% yield. An 80.1/19.9 DNT isomer ratio of the main isomers was achieved and a reduction of by- products in the crude DNT shows a good agreement between the model and the experimental study.

Keywords: Dinitrotoluene, CHEMCAD, thermodynamics, steady state model, MNT

(3)

Datasimulering av Dinitrotoluen Nitreringsprocess.

Moses Ruhweza

Department of Engineering and Chemical Sciences Karlstad University

Sammanfattning

I denna rapport presenteras en metod för att modellera en kommersiell nitreringsprocess för tillverkning av dinitrotoluen (DNT) med simuleringsprogrammet CHEMCAD. En validering av modellen gjordes baserat på resultat från en experimentell studie utförd hos Chimärer Engineering AB, Sverige.

CHEMCAD-modellen utgår från ”steady-state” drift av anläggningen. Viktiga parametrar såsom fluidegenskaper, temperaturprofil och andra driftsbetingelser i CHEMCAD-modellen valdes för att erhålla ett utbyte av DNT samt sammansättningar av såväl syrafas som organisk fas i god överensstämmelse med referensvärdena från den experimentella studien.

Resultaten visade att antagandena i modellen var korrekta och sammansättningarna för syrafasen och den organiska fasen överensstämde med data från den experimentella studien.

Det genomfördes också en detaljerad studie för att analysera effekterna av fysikalisk-kemiska betingelser på det önskade produktutbytet. Både resultaten från den experimentella studien och data från anläggning i drift överensstämde med den simulerade modellen avseende utspädningsvärmens bidrag till energibalansen.

För att erhålla en lämplig beskrivning av reaktionssystemets termodynamik valdes en NRTL-modell och reaktorsystemet optimerades, vilket gav 99,8 % utbyte av MNT och 99,9 % DNT utbyte. Ett förhållande på 80,1 / 19,9 mellan de två huvudisomererna av DNT uppnåddes och en minskning av biprodukter i DNT produktblandningen. Detta är två exempel på en bra överensstämmelse mellan modellen och experimentstudien.

(4)

Acknowledgements

I would like to express my sincere gratitude to Chematur Engineering AB for providing me an opportunity to do my Master’s Thesis at their company.

I sincerely thank my supervisor Lars Stenmark, Development Director at Chematur Engineering AB, Sweden, for his constructive advice and supervision throughout this project.

I also wish to extend my thanks to Stefan Johansson, Technology Manager Nitroaromatics and Margareta Dahl, Manager Process Design for their insightful comments and constructive suggestions to improve the quality of this project work.

With full pleasure, I converge my heartiest thanks to the Process Department as well as all the employees at Chematur Engineering AB for their invaluable advice and wholehearted cooperation without which this project would not have seen the light of day.

Lastly, I would like to thank Professor Lars Nilsson for his exemplary guidance and monitoring throughout this project.

(5)

List of Figures

FIGURE 1:THREE ISOMERS FORMED BY NITRATION OF TOLUENE.PERCENTAGES ARE EXAMPLES OF VALUES TYPICAL

FOR ISOTHERMAL CONDITIONS. ... 4

FIGURE 2:FORMATION OF NITRONIUM ION (THE POWERFUL ELEECTROPHILE). ... 5

FIGURE 3:MECHANISM OF THE ELECTROPHILIC ATTACK ON THE AROMATIC SYSTEM ... 6

FIGURE 4:THE SIX ISOMERS OF DNT(8). ... 7

FIGURE 5:THIS FIGURE SHOWS THE OXIDATION BY-PRODUCTS FORMED DURING THE DNT NITRATION PROCESS (6). ... 9

FIGURE 6:A PROCESS SYNTHESIS PROBLEM (14,15). ... 13

FIGURE 7:THE FIGURE SHOWS THE PROCESS FLOWSHEET FOR THE DNT NITRATION PROCESS. ... 14

FIGURE 8:THIS FIGURE SHOWS A SET-UP OF MIXING TWO LIQUIDS. ... 17

FIGURE 9:WATER MOLECULES SURROUNDING A SULFATE ION. ... 18

FIGURE 10:THE GRAPH OF THE ELECTRIC POTENTIAL FOR CHARGE Q RELATIVE TO THE WATER MOLECULES AT R. .... 19

FIGURE 11:THE POTENTIAL ENERGY FOR A CHARGE Q AT DISTANCE R ... 20

FIGURE 12:A SET-UP SHOWING MIXING OF TWO LIQUIDS. ... 21

FIGURE 13:THIS FIGURE SHOWS A DIALOG BOX FOR THE PROPERTY ESTIMATION OF NEW COMPONENTS. ... 24

FIGURE 14:MOLECULAR STRUCTURE OF TOLUENE (42). ... 26

FIGURE 15:CHEMCADUNITOPS THAT CALCULATE FLOW AS A FUNCTION OF PRESSURE. ... 34

FIGURE 16:A SET-UP SHOWING THE EFFICIENCY OF A HEAT ENGINE. ... 37

FIGURE 17:THE EQUILIBRIUM REACTOR IN CHEMCAD ... 41

(8)

List of tables

TABLE 1:ISOMER CONTENT OF DINITROTOLUENE ... 6

TABLE 2:RELATIONSHIP BETWEEN THE COMPOSITION OF NITROTOLUENE INTERMEDIATES AND TEMPERATURE. ... 10

TABLE 3:SOLUBILITY OF DNT IN SULPHURIC ACID.MODIFIED FROM (11). ... 11

TABLE 4:COMPONENTS DEFINED IN CHEMCAD TO REPRESENT DNT NITRATION PROCESS. ... 23

TABLE 5:PROPERTIES OF THE THREE VERSIONS OF TOLUENE CREATED WITH DIFFERENT METHODS. ... 27

TABLE 6:THIS TABLE SHOWS THE NUMBER OF EACH MOLECULAR GROUP WHICH OCCUR WITHIN THE STRUCTURE BEING ESTIMATED. ... 28

TABLE 7:THERMODYNAMIC MODELS SUITABLE FOR REGULAR SOLUTIONS. ... 31

TABLE 8:THIS TABLE SHOWS SUITABLE ENTHALPY MODELS FOR EQUILIBRIA-K VALUES. ... 32

TABLE 9:EQUIPMENT SUMMARY TABLE FOR THE DNT NITRATION PROCESS ... 38

TABLE 10:REACTIONS USED TO REPRESENT THE DNT NITRATION PROCESS. ... 42

TABLE 11:EQUIPMENT SUMMARY TABLE OF THE EQUILIBRIUM REACTORS ... 44

TABLE 12:EQUIPMENT SUMMARY TABLE FOR HEAT EXCHANGER ... 47

TABLE 13:EXPERIMENTAL RESULTS VERSUS SIMULATION PREDICTIONS FOR SPENT ACID. ... 50

TABLE 14:THIS TABLE SHOWS THE COMPOSITION OF THE RECOVERED SULFURIC ACID. ... 50

TABLE 15:THIS TABLE SHOWS THE RECOVERED DNT. ... 50

TABLE 16:THE COMPOSITION OF THE ACID PHASE IN THE ORGANIC PHASE AND MNT YIELD ... 51

TABLE 17:THE COMPOSITION OF THE ORGANIC PHASE IN THE ACID PHASE. ... 52

TABLE 18:THIS TABLE SHOWS THE COMPOSITION OF THE RECOVERED NITRIC/SULPHURIC MIXTURE ... 53

TABLE 19:A COMPARISON OF THE COMPOSITION OF CRUDE DNT WITH THE RESULTS OBTAINED FROM THE SIMULATED MODEL AND AVAILABLE INDUSTRIAL DATA (IN PARENTHESIS) OR EXPERIMENTAL DATA. ... 54

TABLE 20:THE OVERALL ENERGY BALANCE OF THE DNT NITRATION PROCESS. ... 55

TABLE 21:THIS TABLE SHOWS THE OVERALL MASS BALANCE OF THE DNT NITRATION PROCESS. ... 56

TABLE 22:SULPHURIC ACID MIXED WITH WATER. ... 57

TABLE 23:SULPHURIC ACID MIXED WITH 50 WT.-% SULPHURIC ACID... 58

TABLE 24:NITRIC ACID MIXED WITH WATER. ... 59

TABLE 25:NITRIC ACID MIXED WITH 20 WT.-%NITRIC ACID ... 60

TABLE 26:THIS TABLE SHOWS A COMPARISON OF THE CRITICAL PROPERTIES OF THE NEW COMPONENTS WITH VALUES (SEE APPENDIX C). ... 61

(9)

1. Introduction

This master’s thesis concerns the development of a simulation model and optimization technology for industrial DNT nitration process. All of the simulation/analysis was done using CHEMCAD, version 7.1. I chose CHEMCAD because it enables one to quickly build fundamental steady-state models of chemical process. The work was done at Chematur Engineering AB.

1.1. Background

Computer simulation of dinitrotoluene (DNT) production facility is the main subject of this master’s thesis. A study regarding the use of CHEMCAD for designing and analysing process equipment and benefits of process simulation, is also a focus in this thesis.

Nitration of aromatic hydrocarbons such us benzene and toluene has been extensively studied, mainly due to its industrial importance in the manufacturing of organic synthetic compounds, and its role in the development of our present understanding of organic reactions, particularly electrophilic substitution (1–3).

Commercially, the nitration of toluene is mostly performed to produce toluene diisocyanate (TDI) via DNT and toluenediamine (TDA).

In a DNT production facility, nitration of toluene is performed in two stages with the production of nitrotoluene intermediates in the first stage which is known as mononitration and the production of DNT in the second stage, also known as dinitration.

With a sharp rise in market demand for chemicals such as TDI, many companies have implemented process simulation technology in order to maximize production capacity, improve safety and environmental management among others (4).

An example of a company that can provide such an excellent process technology service to their customers is Chematur Engineering AB (CEAB).

Chematur engineering AB has its headquarters in Karlskoga, Sweden, where it was founded by Alfred Nobel in the late 19th century. The company has an extensive experience (tracing back to the days of Alfred Nobel and his achievements) in modelling and simulation of chemical plants. This in turn has made CEAB the global provider of excellent technology and therefore a forerunner in providing engineering expertise.

In a recent work, CEAB did a research and developed their existing technology of producing toluene diisocyanate (TDI) by optimizing their proprietary pump nitration process for continuous production of Dinitrotoluene (DNT).

(10)

The DNT production was done in pilot scale experiments using a bench scale pump nitration unit installed at their Technology Centre.

This master thesis presents a computer simulation model of CEABs pump nitration process for continuous production of DNT.

1.2. Purpose

The aim of this master’s thesis:

The purpose of this master’s thesis was to build a simulation model for the production of DNT using CHEMCAD. A steady-state model is used as a basis for the simulation, and it is desired to see how well this model predicts the process characteristics (e.g. flowrates, compositions, nitration temperature, properties, equipment sizes, etc.) of the DNT nitration process compared to an experimental study.

1.3. Limitations

Although the master’s thesis has reached its aims and was completed within the limits of the assignment’s due date, there were some unavoidable limitations. The vast majority of the literature was over 50 years old and some related secondary sources cited in those literature were difficult to locate and retrieve. As a result, those secondary sources couldn’t be entered in the reference list, but they are all cited in the body of the paper. Also, during the assembly of the model it became apparent that some process units could not be directly modelled. For instance, the use of an output stream that goes directly to the recycled stream in the process could not be done since CHEMCAD does not allow multiple streams to be sent directly into other process units. To solve this situation a stream divider was added to the model so the separation unit sent its output to the stream splitter, which was then the input for the decanter.

1.4. Methodology

Several methods to build a simulation model for the production of DNT (for determining the solvability of the process system) were considered: Hand calculations, spreadsheet and CHEMCAD.

Hand calculations were only used to solve easy problems and the method was not used for complex problems due to the number of calculations required and the need to re-work the entire process if design conditions were changed.

The advantage of using a spreadsheet is that it is fairly easy to update changes in it. However, it is time consuming to set-up and it can be difficult to add or change some steps.

(11)

The designing of the simulation model was divided into smaller activities: obtaining information, backbone assembly, workarounds, error checking and updates. The information gathered for use included, among others, physical properties of chemicals, experimental records and process diagrams (PFD and P&ID). Once the information was gathered, an agreement was reached on how to build the model.

The backbone of the simulation model was reviewed by my supervisor and experienced process engineers from CEAB’s process department

1.5. Organization of the report

This report consists of six chapters which will cover the designing of a simulation model for the DNT nitration process. Here is an overview of each presented chapter:

 Chapter One: presents the introduction of the thesis. This chapter also discusses the purpose of the study, the methodology of the study and its limitations.

 Chapter Two: covers the scientific literature review and relevant information used to accomplish the project work.

 Chapter Three: this chapter explains the details of the simulation development and selected methodology used for the process.

 Chapter Four: presents the results predicted by CHEMCAD

 Chapter Five: presents a discussion of the process simulation

 Chapter Six: discusses the conclusion and future work to improve this study.

(12)

2. Nitration Theory

2.1. Nitration

Nitration has a long history of industrial application and an extensive research on its mechanisms (5) . Today, nitration is the main reaction used to synthesize one of the most important and largest groups of industrial chemicals, namely aromatic nitro compounds.

A typical industrial nitration is mainly carried out by a mixed-acid reaction of concentrated sulfuric acid and nitric acid (6). The reaction generates nitronium ions (NO2+) which are added onto aromatic substrates via electrophilic substitution. In this way, benzene, toluene and phenol are converted into the simplest of all aromatic nitro compounds, namely, nitrobenzene, nitrotoluenes, and nitrophenols (3,5,7).

2.2. Nitration of Toluene 2.2.1. Mononitration

Toluene undergoes nitration on reaction with a mixture of concentrated sulfuric acid and concentrated nitric acid. In this way, three different isomers of mononitrotoluene (MNT) are formed (8). See Figure 1 for details.

Figure 1: Three isomers formed by nitration of toluene. Percentages are examples of values typical for isothermal conditions.

(13)

2.2.1.1. Generation of the electrophile

To initiate this reaction, we need to form a powerful electrophile, namely the nitronium ion (NO2+).

𝐻𝑁𝑂₃+ 2𝐻₂𝑆𝑂₄ ⇋ 𝑁𝑂₂⁺+ 2𝐻𝑆𝑂₄⁻+ 𝐻₃𝑂⁺

The first step of this reaction is protonation of the OH-group and the reason is to make it a good leaving group so that we can generate a powerful electrophile see Figure 2. After protonation, the negative oxygen forms a double bond and expels water (second step), giving a powerful electrophile, which is, as said before, the nitronium ion (3,5,8) . The sulphuric acid acts as a catalyst in this reaction.

Figure 2:Formation of nitronium ion (the powerful eleectrophile).

2.2.1.2. Electrophilic Attack on Aromatic System

The third step (Figure 3) is that the nitronium ion attacks the aromatic ring of toluene, giving an unstable intermediate (arenium ion). The nitro (-NO2) group can now be added on the ortho position.

It should be noted that when toluene undergoes electrophilic substitution, most of the substitution takes place at its ortho and para positions, because the methyl group on toluene is an ortho-para director (3).

The last step (re-aromatization) is to remove the hydrogen atom and this is done by using hydrogen sulphate (HSO4- ) or excess water in the mixture. The hydrogen sulphate grabs the hydrogen and reforms a double bond, giving ortho-nitrotoluene. All the three isomers are formed but the nitration proceeds with predominant formation of the ortho isomer, but the para and meta product is formed as well (3,8).

(14)

Figure 3: Mechanism of the electrophilic attack on the aromatic system 2.2.2. Dinitration

The second nitration step, dinitration, takes place in the same way as mononitration but it’s more difficult to achieve because of steric hindrance and deactivation of the aromatic ring by the nitro group.

As a result, a higher temperature and a higher sulfuric acid concentration is required (8).Several isomers are formed of which 2,4- and 2,6-DNT are the most important, see Figure 4 and Table 1.

Table 1: Isomer content of dinitrotoluene

Isomers Organic product

(Wt.-%)

2,4-DNT 76.1

2,6-DNT 19.8

3,4-DNT 2.25

2,3-DNT 1.23

2,5-DNT 0.54

3,5-DNT 0.08

(15)

Figure 4: The six isomers of DNT (8).

(16)

2.3. Impurities

In the industrial production of DNT, impurities are formed and the formation rates of these impurities are significantly affected by the initial reaction conditions, including nitration temperature and-, initial sulfuric and nitric acid concentration in the mixed acid (5,8).

The presence of the methyl group in toluene makes it easier to be oxidized to nitrocresols. According to (9), nitration of toluene to MNT generates on average about 0.7 wt% nitrocresols, which are mainly dinitro-para and ortho-cresol (80% 2,6-dinitro-p-cresol). Benzoic acid products, nitrogen dioxide (NO2) and, carbondioxide (CO2) are also formed due to the oxidative power of the acid (9,10).

Oxidative degradation of nitrocresols leads to the formation of nitrous acid, mainly during dinitration.

(17)

Figure 5: This figure shows the oxidation by-products formed during the DNT nitration process (6).

It is important to minimize the formation of by-products during nitration of toluene as this causes a reduction in yield (6).

2.4. Effects of physicochemical factors on nitration of toluene 2.4.1. Effect of temperature

The nitration temperature is a crucial parameter as it influences the yield of the mononitro isomers.

The nitration temperature also influences the reaction rate but at a considerably lower degree (8).

(18)

Pictet’s study, as cited in (11), showed that when nitrating toluene with a mixture of nitric and sulfuric acids at lower temperatures, relatively more para-nitrotoluene could be obtained than at higher temperatures. A similar study by Orlova, also as cited in (11), concluded that a lower nitration temperature causes an increase of the para-nitrotoluene content while the meta-nitrotoluene and the ortho-nitrotoluene content decreases Table 2.

Table 2: Relationship between the composition of nitrotoluene intermediates and temperature.

Temperature (°𝑪)

Composition of the product

Ortho-isomer Para-isomer Meta-isomer

30 56,9 39,9 3,2

60 57,5 38,5 4,0

For safety reasons and for the purity of the product, it is of great importance to keep the nitration temperature as low as possible and constant. Using too high temperature causes the reaction to proceed violently and by-products, especially oxidation products, are easily formed (11). Therefore, the nitration temperature should not exceed 40 ℃ during mononitration and 70 ℃ during dinitration, since above this “safety” limit both the methyl group and the aromatic nucleus are attacked oxidatively, leading to an increased formation of by-products, including nitrocresols and nitrophenols (11) .

2.4.2. Effect of mixed acids

The composition of the mixed acid depends on the compound being nitrated and the number of nitro groups to be introduced. For instance, if more nitro groups are to be added (e.g., during dinitration), then the acid concentration should be higher.

The ratio of the nitric acid, sulfuric acid and water should be chosen wisely. Otherwise nitration of toluene might be incomplete. Since water is formed during nitration and it dilutes the mixed acid, the amount of sulfuric acid must be chosen in such a way that it binds up all the water formed (11).

It is preferable to use a very slight excess of nitric acid (e.g., 1-2% in both nitration stages), to avoid oxidation processes (11).

Higher concentration of sulfuric acid increases the rate of the reaction by increasing the concentration of the electrophile, nitronium ion (3).

2.4.3. Effect of Spent Acid

The acid leaving the dinitration stage is re-used for the nitration of toluene to mononitrotluene. This in turn affects the nitration reaction and the formation of the nitrotoluene intermediates (11).

(19)

One method of adding the acid from dinitration to mononitration stage is by mixing the spent acid with concentrated nitric acid and sulfuric acid.

A disadvantage of this method is an increment of temperature, mainly due to the heat of dilution of the mixed acid (11).

2.4.4. Effect of nitro compounds solubility

The solubility of nitro compounds is an important factor in the nitration process.

The more easily the organic phase dissolves in the acid phase, the higher the reaction rate, and the degree of nitration that can be obtained in a given time.

For example, nitrobenzene and trinitrotoluene (TNT) dissolve easily in concentrated sulfuric acid.

However, TNT dissolves with difficulty in mixed acids but its solubility might be high when the content of nitric acid falls to a few percent, as in the spent acid (11).

The solubility data for DNT in sulphuric acid of various concentrations are tabulated in Table 3.

Table 3: Solubility of DNT in sulphuric acid. Modified from (11).

Concentration

% H2SO4

Solubility of DNT in sulphuric acid g DNT/ 100 g sulphuric acid

40°𝐶 50°𝐶 70°𝐶

80 - 2.5 3.8

83.6 3.6 4.7 6.3

88.7 10.0 12.8 -

90 - 16.8 20.0

3. Process simulation

In this modern age of powerful computers, the role of process simulation in the chemical industry has grown immensely, and with good reason.

Process simulation is a computer presentation of a real-world process plant or system by a mathematical model which is then solved to obtain information about the performance of the process (12).

(20)

3.1. Importance of simulation

Process simulation allows engineers to model processes in extreme detail without having to spend the time, manpower and money for physically testing the design in a real-world industrial environment.

For instance, consider being asked to design a distillation column to produce a mixture of benzene and toluene into an overhead product containing 95% benzene and a bottom product containing 90%

toluene. This process can be designed by hand calculations (e.g., calculating the condenser and reboiler duties, mass and energy balances and estimating tray efficiencies) or by physically building a pilot plant of the process. However, when the design conditions are changed (e.g., 95% toluene and a feed rate of 850kg/h instead of 750kg/h), it takes time and becomes costly to test the potential designs.

With the help of commercial process simulators (e.g., Aspen Plus, Aspen HYSYS and CHEMCAD) however, a tremendous amount of time and money can be saved.

Therefore, in the ever more competitive world of processing and manufacturing, process engineering services are no longer complete without the presence of process simulators (13).

Process simulators are extensively being used as powerful tools to increase, among others, the production capacity, profits of a company, competitiveness and to reduce the build-up time for new manufacturing process. It can also be used to reduce the capital equipment costs by optimizing the process.

Chemical engineers use process simulation tools among others Aspen Plus and CHEMCAD to design complex process plants such as large-scale process and manufacturing industries, where energy use is measured in megawatts, costs and profits are measured in hundreds of millions of dollars and materials measured in thousands of tons.

Another benefit of process simulation tools is that it allows chemical engineers to predict capital cost expenditures, evaluate optimization options and determine the overall effects of potential process changes in one area. Furthermore, chemical engineers rely on process simulation to answer what-if questions asked by operational staff or management of the process plant. All in all, chemical engineers use process simulation tools to accurately predict the outputs of the process when the process inputs and outputs are given.

(21)

Figure 6: A process synthesis problem (14,15) .

3.2. Selection of a thermodynamic method

Just like the foundation of a building, the methods used for estimating thermodynamic and transport properties determine the perfect conditions of a chemical process simulation. These days, the process industry or engineers rely on using simulators to perform their computations, and all commercial simulators today are equipped with a countless of property packages with property estimation methods such as NRTL, SRK, UNIQUAC, and many more. It is of great importance to know which property package is appropriate for one’s process computation. The objective of this section is to provide some accurate and deep understanding into the performance of those property packages and selection of a proper method to represent the various physical and chemical phenomena under a given set of operating conditions where mononitration and dinitration occurs. Another aim is to enable the reader to make the correct selection of property method.

CHEMCAD has several property packages that each consist of different computational methods that are used to estimate thermodynamic and transport properties (16).

What kind of thermodynamic and transport properties are of interest in process simulation?

If we take a look at the pump nitration process which is Chematur’s process for production of DNT, we find that the process involves separating, moving of fluids, vaporizing etc. (5,8).

To select the correct property method, the pump nitration process was analysed and the properties required to execute some computations were identified (Figure 7, appendix D).

(22)

Figure 7: The figure shows the process flowsheet for the DNT nitration process.

3.3. Analysis of the DNT nitration process 3.3.1. Pump

A pump increases the pressure of a liquid stream by adding work to it. Like most pumps, a centrifugal pump converts input power to kinetic energy of the fluid by accelerating fluid in an impeller (17,18).

Therefore, the required properties include heat capacities, liquid density and pressure.

3.3.2. Separator

In the design of the separator, it was necessary to understand how the chemical components partition themselves between the two liquid phases and to set up the mass relationships of the phases. Also, it was important to understand how much liquid and vapour are produced at the operating temperatures and pressures. This means that the separation process required the properties of vapour and liquid densities, enthalpies and pressures (17,18).

3.3.3. Heat Exchangers

Heat exchangers in the pump nitration process allow the fluids to be cooled. The properties required to represent the cooling process are: liquid vapour pressure, heat of vaporisation, liquid heat capacities, densities, more (19).

3.3.4. Reactors

The reactor allows the reactants to undergo a chemical reaction. Hence, the required properties include heat of formation, heat of reaction, enthalpies, densities etc. (20).

(23)

3.4. Thermodynamic models

CHEMCAD has over 50 phase equilibrium-K (K-value) models and about twelve enthalpy models stored in its library. In other words, the phase equilibrium-K and enthalpy models are methods used for the prediction of vapour-liquid or vapour-liquid-liquid phase equilibrium (called the phase equilibrium-K) and the heat balance, which is called the enthalpy-H (16,21,22).

It should be noted that a selection of an inappropriate property method, leads to convergence problems and erroneous results. Therefore, it was fundamental to consider, among others:

The process species and compositions.

The phases involved in the system.

Temperature and pressure operating ranges.

Nature of the fluids.

To understand why a particular property method is prefered in any process simulation, you must understand some thermodynamic relationships among others:

Fugacity Activity Equilibrium Enthalpy 3.4.1. Fugacity

Fugacity is the tendency of a substance to prefer one phase (liquid, solid and gas) over another. In other words, fugacity is sort of or acts as a correction factor of pressure in real systems (23,24).

Fugacity can be estimated or determined from gases that are closer to reality than an ideal gas. Real gases behave differently from ideal gases. An ideal gas is made up of molecules whose only interactions are elastic collisions. Therefore, these molecules have no intermolecular forces between them contrary to real gases. As a result, the ideal gas law may not hold for most gases and vapours encountered in reality (25,26).

To solve this problem and accurately calculate chemical equilibria for real gases, pressure is replaced by fugacity. So, fugacity is related to how non-ideal a gas behave and it is derived from equation of states (e.g. Van der Waals, NRTL, UNIQUAC, and SRK) or other expressions that can describe non- ideal systems.

(24)

For two-phases of a species to be in equilibrium, the pressure, temperature and the chemical potential must be equal in both phases. Similarly, for pure species co-existing as liquid and vapour, if they are to be in equilibrium, then the temperature, pressure and fugacity in both phases must be the same (27).

3.4.2. Activity

Activity is a ratio of fugacity to the fugacity of the standard state of a material (pure component, mixture or solution) at the same temperature and pressure (28,29). The variation of the activity of component (activity coefficient) with temperature and composition is important in thermodynamic process because it used to determine the Gibbs energy of mixing of a component, which in turn is used to determine the equilibrium state of any chemical reaction.

3.4.3. Equilibrium

At equilibrium, all thermodynamic properties such us free energy (U), Helmholtz free energy (A), Gibbs free energy (G), amongst others., are minimized (30). To minimize the free energy functions, i.e., A, U, and G we need to have a method for determining vapour-liquid/liquid-liquid equilibrium(31).

(25)

3.5. Enthalpy changes upon mixing

When two liquids for example sulphuric acid and water are mixed, the resulting enthalpy is not necessarily the sum of the pure component enthalpies since the unlike interactions between molecules is most likely different than the like interactions. In other words, if the H2SO4 – H2O interactions are stronger than the H2SO4 – H2SO4 and H2O - H2O interactions, then the mixing process will be exothermic. The task of this section is to study the change in enthalpy that occurs when mixing occurs.

The change of enthalpy upon mixing two liquid streams is shown in Figure 8

Figure 8: This figure shows a set-up of mixing two liquids.

Two different inlet streams with moles of liquid 1 (n1) are mixed with moles of liquid 2 (n2) in a mixer and the resulting mixture leaves the process unit at a temperature T3. The energy balance for this process can be defined as follows:

𝑞 = (𝑛₁+ 𝑛₂)ℎ₃− 𝑛₁ℎ₁− 𝑛₂ℎ₂ (1)

Heat of mixing may either be positive or negative. If it is positive then the reaction is endothermic (meaning heat absorbed because the mixture has a higher enthalpy than the pure component) and if its negative then the reaction is exothermic (heat given off because the mixture has a lower enthalpy than the pure component).

From the conclusion above, you can tell from different types of molecular interactions in case heats of mixing will contribute significantly to the energy balance. Nonideal mixtures have a fairly large heat of mixing. For instance, the mixing of sulphuric acid and water produces so much heat because the energy level of the system goes down and it releases heat.

(26)

In details, let us assume a very concentrated solution like 18 moles of sulfuric acid in one litre of solution. The small amount of the water molecules will surround the sulphate ions very quickly since there are so many sulphate ions. Therefore, the initial energy that is given off is enormous and of course that energy will raise the temperature, causing the solution to boil vigorously.

Figure 9: Water molecules surrounding a sulfate ion.

The positive ends of water molecule attract themselves to the negative ions (sulfate ions) in the solution, resulting to a lower energy state. So, the enthalpy of dilution is a negative quantity i.e., it is an exothermic reaction which expels heat (energy is taken away from the ions and expelled to the solution). Figure 9, shows one way to look at why that happens.

The negative charge has an electrical field around it and so that means that the electrical potential V around the negative charge increases as the positive charge gets closer to it (34).

𝑉 =𝑘𝑄 𝑟

(2)

where k is Coulomb’s constant

(27)

Figure 10: The graph of the electric potential for charge Q relative to the water molecules at r.

The electric potential energy for a charge q at r is then given as:

𝑈 =𝑘𝑄𝑞 𝑟

(3)

(28)

Figure 11: The potential energy for a charge q at distance r

Notice that the charge q is positive which means that the electrical potential energy goes down as the charge gets closer. In other words, the potential energy gets expelled and turns into heat when the positive charge comes closer.

Again, in a mixture (e.g., water and an acid solution) the enthalpy of dilution is negative as a result of a lower energy state.

3.6. Heats of dilution of mixed acids

2 kg of pure water at 21.1 ℃ was mixed adiabatically with 1 kg of 80% wt.-% sulphuric acid solution at 21.1℃.

(29)

Figure 12: A set-up showing mixing of two liquids.

The mass is conserved and if we assume that there is no accumulation in the mixing unit then the:

Total mass balance: 𝑚₃ = 3𝑘𝑔

To find the composition at the outlet stream, a balance equation was used based on sulphuric acid.

Notice that the balance equation can be used on either sulphuric acid or water.

𝑥_𝑆𝐴𝑚₁ = 𝑥_3𝑆𝐴𝑚₃ → (0.8 ∗ 1) = (𝑥_3𝑆𝐴∗ 3) (4)

𝑥_3𝑆𝐴= 0.27 → 𝑥_3𝑊 = 0.73 (5)

To find the outlet temperature of the mixture, an energy balance equation is needed.

Recall that for an adiabatic process 𝑑𝑞 = 𝑑𝐻 = 0

If a change in enthalpy is zero, that implies that the inlet enthalpy is the same as the outlet enthalpy (32).

𝐻_𝐼𝑁= 𝐻_𝑂𝑈𝑇 (6)

(30)

So, as seen earlier from equation (1) the specific enthalpy of the acid mixture can be solved from

𝑚₁ℎ₁+ 𝑚₂ℎ₂ = 𝑚₃ℎ₃ (7)

Since the inlet streams are both at 21.1 ℃, we should be able to use an enthalpy concentration diagram to find the specific enthalpy of these solutions at a specific temperature and concentration.

3.6.1. Enthalpy concentration diagram

Reference 55 shows the specific enthalpy of the solution in units of kJ/kg as a function of mass fraction of sulphuric acid. The specific enthalpy of the solution is shown for several isotherms (each curve represents different temperatures). In this case, both the inlet streams are mixed at a temperature of 21.1℃. The second stream is pure water, so the mass fraction of sulphuric acid is zero and the isotherm is 21.1 ℃. The specific enthalpy can then be found by tracing the curve up to when the x- axis is equal to zero. By visual inspection, this is approximately ℎ₂ = 99 𝑘𝐽/𝑘𝑔. The first stream is at 80 wt.-% and also at 21.1 ℃. The 21.1 ℃ isotherm can also be traced up to where the x-axis is equal to 80 wt.-%. This gives approximately ℎ₁ = −240 𝑘𝐽/𝑘𝑔, also by visual inspection. The values can then be substituted in equation (24) and the specific enthalpy of the mixture ℎ₃ = −14 𝑘𝐽/𝑘𝑔 . Therefore, the temperature at which sulphuric acid solution of 𝑥_3𝑆𝐴 = 0.27 mass fraction equal to ℎ₃ = −14 𝑘𝐽/𝑘𝑔 can be found at the intersection of a horizontal line from ℎ₃ = −14 𝑘𝐽/𝑘𝑔 , and a vertical line from 𝑥_3𝑆𝐴 = 0.27. By visual inspection the outlet temperature is 48 ℃ since it is in- between 37.80 ℃ 𝑎𝑛𝑑 65.60 ℃ .

The found values by hand calculation are quite consistent with CHEMCAD values. The predicted temperature by CHEMCAD was 48.90 ℃ (less than 2 % error). Many solutions with different compositions of mixed acids were modelled with CHEMCAD and the overall values agree with literature values.

(31)

3.7. Component specification

In CHEMCAD, process engineers often use the word component when they talk about chemicals.

This section will illustrate how to create or add a new component into the databases of CHEMCAD and the different steps as part of creating the new component.

The components present in feed streams and possible products are defined in the components specification menu and are listed in Table 4.

Table 4: Components defined in CHEMCAD to represent DNT nitration process.

Name MF Name MF

Toluene C7H8 2,3-Dinitrotoluene^a C7H6N2O4

m-Nitrotoluene C7H7NO2 Sulfuric acid H2SO4

o- Nitrotoluene C7H7NO2 Nitric acid HNO3

p- Nitrotoluene C7H7NO2 Nitrous acid^a HNO2

2,4-Dinitrotoluene C7H6N2O4 Water H2O 2,5- Dinitrotoluene C7H6N2O4 Nitric oxide NO 2,6- Dinitrotoluene C7H6N2O4 Nitrogen dioxide NO2

3,4- Dinitrotoluene C7H6N2O4 Carbon oxide CO 3,5- Dinitrotoluene C7H6N2O4 Carbon dioxide CO2

Nitrogen N2 m-Dinitrobenzene C6H4N2O4

Oxygen O2 2,4,6-Trinitrotoluene C7H5N3O6

Benzene C6H6 o-Xylene C8H10

Nitrobenzene C6H5NO2 6-Nitro-m-Cresol^a C7H7NO3

Nitrobenzoic acid^a C7H5NO4 4,6-Dinitro-o-Cresol^a C7H6N2O5

Nitrocresol^a C7H6NO3 Cresol C7H8O

a Chemicals that did not exist in the databases of CHEMCAD. MF: molecular formula.

Defining or specifying chemical components is something that typically precedes the initialisation of process simulation (35). CHEMCAD is equipped with a component database known as CHEMCAD component database/library which has over 1500 chemical components from Design Institute for Physical Properties (DIPPR) databases (16,21). However, not all chemical components are available in CHEMCAD component database. Therefore, the user must define the missing chemical components before making use of them in CHEMCAD.

(32)

To define or create the chemicals that are not present (User-defined components) in the CHEMCAD library, it is important to know the basic properties of the chemical component. For instance, when you define the normal boiling point and molecular weight of toluene CHEMCAD will estimate the rest of the missing properties based on different estimation methods among others, Joback and UNIFAC group contribution method (36,37).

High accuracy is not claimed but if more chemical properties are provided, for instance the specific gravity or API gravity, the other properties predicted or generated by CHEMCAD have an acceptable percent error (less than 2% error).

Figure 13: This figure shows a dialog box for the property estimation of new components.

3.8. Creating new components

It is of great importance to make sure that the properties of chemical components are being estimated appropriately. In fact, the selection of a proper method to estimate the properties of a component is one of the most important tasks that will affect the rest of the simulation. Therefore, it is important to consider the choice of methods used to estimate different chemical properties.

(33)

The estimation of chemical components in CHEMCAD relies mostly on the group contribution method. These methods are employed because it is not always possible to find experimental values of properties for the chemical components of interest.

In these estimation methods, different specific formulas are used to estimate certain physical and thermodynamic properties based on the groups present in the chemical structure (37–39).

There are five different methods used to estimate properties of pure components with CHEMCAD (electrolyte, combustion solid, hydrocarbon-pseudo-component, Joback and UNIFAC method) but only three of them will be discussed.

3.8.1. Defining a Hydrocarbon pseudo-component

This method estimates from correlations suited for hydrocarbon pseudo-components. It requires only the normal boiling point, molecular weight and the specific gravity as input data to generate reasonable properties of a molecule. This method was not chosen for creating new components and therefore will not be considered further.

3.8.2. Estimating by the UNIFAC method

The UNIFAC method is a functional group contribution method that estimates physical properties for all kinds of compounds(mainly organic compounds) based on their molecular weight and molecular structure (40,41). CHEMCAD has data for all functional groups present in a molecule and therefore generates an accurate value of critical properties among others temperature (TC) and pressure (PC).

3.8.3. Estimating by the Joback method

The Joback Method is a group contribution method similar to the UNIFAC method.

When the molecular weight, normal boiling point and the number of each molecular group which occur within the structure being estimated is provided, the Joback method estimates other properties with a much higher accuracy (37).

Consider; toluene which has the chemical structure as shown in Figure 14.

To estimate properties for toluene you would:

 Choose a “CH3” group and enter a 1 for the occurrence

 Choose a “=C<” group and enter a 1 for the occurrence

 Choose a “=CH-” group and enter a 5 for the occurrence

(34)

Once you have entered the basic properties and your structure’s groups, CHEMCAD

estimates the properties and creates the new component.

Figure 14: Molecular structure of toluene (42).

3.9. Validating the component estimation method(s)

New components (see Table 4) were created because they did not exist in CHEMCAD’s databases.

Before creating the new components, multiple copies of other components were created based on toluene which is a well-defined component in CHEMCAD. This was done to compare and select the best estimation method.

Table 5 shows the properties of three versions of toluene that were created using three different predictive methods, Psuedo-component method, Joback method and the UNIFAC method.

(35)

The three methods considered the estimation of density, critical properties, enthalpy and other properties when the mass flow rate, pressure and temperature were fixed to be 100kg/h, one bar and 30 C respectively.

Further, the molecular weight and the normal boiling point of toluene were given as input data for all the three versions of toluene.

The properties of toluene were then generated by CHEMCAD and used as reference.

Firstly, the pseudo-component method was selected and the basic properties of toluene, i.e. molecular weight and the normal boiling point were entered manually. Also, the specific gravity of toluene was specified to give more predicting power. CHEMCAD then used the information to create a new component. The predicted component did not have the density parameters because the density was predicted from the specific gravity and the boiling point instead of being calculated by temperature dependent equation. This is not a very accurate predictive method for aromatic compounds and therefore was not used. Note that the pseudo-components use a method which is used for only simple alkenes.

Secondly, the group contribution methods were selected and the number of groups which occur within the structure of toluene were entered manually in CHEMCAD. These two methods could predict the critical properties, liquid density, the ideal gas heat of formation and Gibbs of formation among others.

It turned out that reliable results could be obtained with the Joback method. Also, the UNIFAC method gave a reasonable result. However, the UNIFAC method depended critically on the manual input of specific gravity. As a result, the Joback method was chosen over the UNIFAC method.

The critical properties were estimated by equations listed below.

Table 5: Properties of the three versions of toluene created with different methods.

Comparison of toluene library properties with CHEMCAD’s estimation methods Methods

Name Library Pseudocomponent Joback UNIFAC

Temp C 30 30 30 30

Press bar 1 1 1 1

Enth MJ/h 14.057 -41.130 12.473 -1.5201

Mass flow kg/h 100 100 100 100

(36)

Liq vol m3/h 24.3254 24.3254 24.3254 24.3254

TC C 320 314.563 320.795 317.074

PC bar 41.08 40.463 41.143 40.092

CP Kj/kg-K 1.712 1.709 1.712 1.888

Density kg/m³ 859.351 857.979 858.951 858.049

Th cond W/m-K 0.131 0.127 0.133 0.134

3.9.1. Molecular descriptors

The number of molecular descriptors on the components that were created (see Table 6) was determined and entered manually in CHEMCAD. Once that was done the Joback method was selected and the estimations of density, ideal gas of Gibbs of formation, critical properties (e.g. TC and PC) etc.

were of a precision comparable to the actual properties of the pure components. The Joback method was chosen because the it predicts properties based on the molecular structure (43).

Table 6: This table shows the number of each molecular group which occur within the structure being estimated.

Occurrences

Joback Groups NCâ DNCâ DNTâ Cresol NBA NiA

=C< (ring) 3 4 3 3 2 -

-NO2 1 2 2 1 1 -

-CH3 1 1 1 1 -

-OH (phenol) 1 1 - 1 - -

=CH- (ring) 1 2 3 3 4 -

-O- (nonring) - - - -

>C=O (nonring) - - - - 1 -

-N= - - - 1

-OH (alcohol) - - - - 1 1

=O - - - 1

aNC= Nitrocresol, NiA=Nitrous acid, DNC= Dinitro cresol, DNT=2,3-Dinitrotoluene, Cresol=6-Nitro-M-Cresol

(37)

To process the chemical components mentioned above, it is important to understand how the state of the components change with respect to pressure and temperature.

Since the DNT nitration process involves liquid phases, another consideration in selecting a suitable thermodynamic model for the simulation was to know what was happening in the liquid phases. Liquid solutions are classified into five categories:

Ideal solutions Regular solutions

Polar solutions (highly non-ideal) Electrolyte solutions, and

Special solutions.

Electrolyte and special solutions are not considered in detail here.

3.9.1.1. Ideal solution

Ideal solution is a solution in which the entropy of mixing is assumed to be

∆𝑆_𝑚𝑖𝑥 = −𝑅 (𝑥_𝐴𝑙𝑛𝑥_𝐴+ 𝑥_𝐵𝑙𝑛𝑥_𝐵) (8)

where 𝑥_𝐴𝑎𝑛𝑑 𝑥_𝐵 are the mole fractions of two components, and the enthalpy of solution is zero.

∆𝐻_𝑚𝑖𝑥 = 0 (9)

As a result, the vapour phase of ideal solution behaves as an ideal gas (low pressures) and all the molecules in the liquid phase are made up of species of similar molecular size and chemical nature (44).

It should also be added that these are typically solutions with no intermolecular forces of attraction that form ideal mixture in which the activity coefficient becomes close to one throughout the concentration range.

If the liquid phases are ideal solutions and the vapour phase is an ideal gas, Raoult’s law describes the distribution of species between the phases as:

𝑃_𝑖 = 𝑌_𝑖𝑃 = 𝑥_𝑖𝑃_𝑖⁰ , 𝑎𝑡 𝑡𝑒𝑚𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒 𝑇 (10) Where:

𝑃_𝑖 is the partial pressure,

(38)

𝑃_𝑖⁰ is the vapour pressure at saturation 𝑃 is the system pressure.

The vapour-liquid equilibria can be determined using Raoult’s law:

𝐾_𝑖 = 𝑃_𝑖⁰ 𝑃

(11)

For these systems, it is recommended to use the ideal vapour pressure model (VAP model) for phase equilibrium-K and SRK for enthalpies. An example of this is the acetonitrile/nitromethane system.

3.9.1.2. Regular solutions

Unlike ideal solution, regular solution assumes that the enthalpy of mixing is not zero. It depends on compositions.

∆𝐻_𝑚𝑖𝑥= 𝛽𝑥_𝐴𝑥_𝐵 (12)

where B is the interaction parameter. It should be noted that the entropy of mixing is the same as that for the ideal solution.

These systems are best modelled using SRK, Peng-Robinson, API SRK, BWRS and MSRK. It is recommended to use Peng-Robinson and SRK for all hydrocarbon systems, except for certain processes with wide boiling values and heavy hydrocarbons at low pressures.

MSRK is recommended for chemicals such as branch-chained hydrocarbons, halogenated hydrocarbons, some polar compounds, etc.

CHEMCAD calculates equilibrium values using fugacity coefficients 𝐾_𝑖 = 𝑌_𝑖

𝑋_𝑖 = ∅_𝑙𝑖 ∅_𝑣𝑖

(13)

(39)

Table 7: Thermodynamic models suitable for regular solutions.

Equilibrium Enthalpy Process Application

PR or SRK PR or SRK All hydrocarbon systems for pressures > 10 bar

GS LK Hydrocarbon processes with a wide boiling range −18℃ 𝑡𝑜 430℃

ESSO LK Processes with heavy end hydrocarbons at pressures > 7 bars Temperatures (90℃ 𝑡𝑜 200℃)

MSRK SRK Branch-chained hydrocarbons, halogenated hydrocarbons and polar compounds

3.9.1.3. Polar solution

These are solution systems in which the non-ideal behaviour of the liquid phase arises from molecular interactions. The vapour phase is assumed to be a regular solution and CHEMCAD allows the user to model the system with activity coefficient methods, which require binary interaction parameters (BIPs) for accuracy (45).

The methods that suit for activity coefficients include NRTL, UNIFAC, Van Laar, UNIQUAC, Wilson, T.K. Wilson, Margles and GMAC. It is recommended to use NRTL, UNIQUAC and Wilson when the data available is sufficient (>50%) and UNIFAC when the data is incomplete (<50%).

3.9.1.4. Electrolyte solution

These solutions are treated as true species (molecules and ions) and CHEMCAD requires binary interaction parameters (BIPs) for accurate modelling.

The recommended thermodynamic model is NRTL and the enthalpy model is LATE.

For the DNT nitration process (non-ideal solutions), the main thermodynamic model (K model) selected for this study was the non-random-two-liquid (NRTL). NRTL is a K-value model that uses binary interaction parameters (BIPs) and serves to model the non-ideal behaviour in the liquid phase.

It is a suitable model because the system exhibits vapour-liquid-liquid equilibrium and operates at a pressure of less than 20 bars with an assumption of the media as an electrolyte (46). Other suitable models for two liquid phases are UNIQUAC and UNIFAC LLE.

The enthalpy model chosen was latent heat. Further, the nitration process required an equation of state approach. The Redlich-Kwong-Soave (SRK) equation of state with the phase option of

(40)

vapour/liquid/liquid was used to calculate the phase equilibria during nitration (47,48). SRK was also chosen to calculate the fugacity.

A general recommendation when selecting an enthalpy method is given in Table 8:

Table 8: This table shows suitable enthalpy models for equilibria-K values.

If the phase equilibria-K method is: Use this for enthalpy:

Peng-Robinson (PR) Peng-Robinson (PR)

BWRS BWRS

SRK.API SRK, MSRK, TSRK, VAP, ESD, SAFT SRK

ESSO, Grayson-Streed (GS) Lee-Kesler (LK)

NRTL, PSRK, WILS, T.K. Wilson, UNIQ, VANL, HRNM

Latent Heat

Amine Amine

PPAQ SRK or Latent Heat

(41)

3.10. Process Flowsheet modelling

The process flowsheet window in CHEMCAD allows the user to construct the flowsheet graphically.

CHEMCAD is equipped with an array of process units, connectors and controllers from which a process plant and its sections can be designed.

3.11. Modelling process equipment 3.11.1. Feed streams

The feeds were designed to provide a feed flow control system in which the other feeds (nitric acid, recovered sulphuric acid, nitric/sulphuric mixture and recovered DNT feed) were remotely controlled from the toluene feed, a master control feed. In other words, the feed flow control system was used to maintain the flow rate of the master control feed at a specified proportion relative to that of the other feeds in the DNT nitration process.

3.11.2. Pumps

The pump in CHEMCAD was used to increase the pressure of the fluid stream. The desired pressure was controlled by the piping system and the process involved, whereas the total flow rate in each stage was controlled by feed-backward controllers in the downstream process units, unit operations (UnitOps). A feedback controller is a mathematical controller that is used to adjust the required variables. The total mass flow rate of each loop was adjusted by the controller until the temperature increase over the reactor system was approximately 10 ℃.

Prior to using the feed-backward controllers, two process units (a stream reference and a feed-forward controller) were combined to connect the loop. The process units connect two flows to maintain a defined ratio and enable the simulation to forward all the properties of the upstream process units to downstream units.

There are many different types of pumps in CHEMCAD. However, these can be classified in two basic types: positive displacement pumps and centrifugal pumps. The type of pump used in CHEMCAD to model the DNT nitration process was a centrifugal pump with an electrical motor as a drive.

3.11.3. Centrifugal Pump-Solution Principal

The two liquid phases enter the centre of the pump and exit perpendicular to the inlet. On the inside of the centrifugal pump, there’s an impeller which rotates at thousands of revolutions per minute (rpm) and behind the centrifugal pump there’s an electrical motor which is responsible for spinning the shaft

Computer simulation of Dinitrotoluene Nitration Process