Chemical

 Process  Design  /  Diseño  de  Procesos  Químicos  

Topic 9. Design and scheduling of batch processes

Javier  R.  Viguri  Fuente  

Eva  Cifrian  Bemposta  
Department  of  Chemistry  and  Process  &  Resource  Engineering  
GER  Green  Engineering  and  Resources  Research  Group  
This work is published under a License:
Crea>ve  Commons  BY-­‐NC-­‐SA  4.0  
INDEX

1.- Batch Processes: Introduction

2.- Single Product Batch Plants: Example, Scheduling by Gantt chart
3.- Multiple Product Batch Plants: Campaigns
4.- Transfer Policies
5.- Parallel units and Intermediate storage
6.- Synthesis of flowshop plants
7.- Sizing of vessels in batch plants
8.- Inventories
9.- Further Reading and References
PRACTICAL CHAPTER
RELEVANT TO LEARNING
1.- Batch Processes: Introduction

• Continuous processes: manufacture of commodities.

• Batch processes: specialty chemicals,

pharmaceuticals.

• Semicontinuous processes: hybrids

of batch and continuous.

BATCH or SEMICONTINUOUS * Fed-batch

processes * Batch-product
• SMALL production rates.
removal
• LARGE residence times.
• INTERMITENT product demand.
• SAFETY aspects are of great concern.
• MULTIPRODUCT facilities.
• Mostly RECIPE Based.
1.- Batch Processes: Introduction

Batch reactors used in the dairy industry. Fermentation plant producing enzymes at industrial scale.
© DCI Inc. St. Cloud, MN. © Fraunhofer.

Recipe management system Batch in a Pesticides production in batch process.

pharmaceutical company to produce generics. http://argossouthafrica.weebly.com/about.html.
http://www.kolektorautomation.com/index.php?t=referenceNews&l=en&id=60.
1.- Batch Processes: Introduction
When a batch process is used to manufacture two or more products (Multiple
Product Batch Plants), two major limiting types of plants can arise:

• FLOWSHOP (or multiproduct) plants in which all products require all stages
following the same sequence of operations.
A P1
B P2
C P3

Flowshop plant

• JOBSHOP (or multipurpose) plants where not all products require all stages and/
or follow the same sequence.
A P1
B P2
P3 C

Jobshop plant

The greater the similarity in the products being produced, the closer a real
plant will approach a flowshop, and vice versa-the more dissimilar, the more it
will approach a jobshop.
2.- Single Product Batch Plants: Example, Scheduling by Gantt chart
Example of Batch Process

Processing Steps (RECIPE):

1. Mix A + B, heat to 80°C, react to form C. Total 4 hours.
2. Mix with solvent D for 1 hr. at 20°C.
3. Centrifuge to separate the product C for 2 hrs.
4. Dry in a tray for 1 hour at 60°C.

• Amounts are also specified.

• Assume processing times independent of batch sizes.
• Neglect transfer times.
2.- Single Product Batch Plants: Example, Scheduling by Gantt chart
Scheduling  Gantt chart (time activity chart) RECIPE
Non-Overlapping operation 1. Mix 4 hrs.
4 4 2. Mix 1 hr.
Stage 1 3. Centrifuge 2 hrs.
1 1 4. Dry 1 hr.
Stage 2
2 2
Stage 3
1 1 Cycle time (CT) = time
Stage 4
between the completions of
batches.
CT Time MSpan
Non-overlapping:
• Cycle time (CT) = 4 + 1 + 2 + 1 = 8 hrs. M
• Makespan (2 batches) = 16 hrs. Poor equipment use. CT = ∑ τ j
j=1
Overlapping operation
4 4 Overlapping:
Stage 1
1 1
Stage 2 CT = max{τ j }
2 2 J=1,M
Stage 3
€
Stage 4
1 1 Makespan = total time
required to produce a given
CT MSpan Time number of batches.
€
• Cycle time (CT) = max {4, 1, 2, 1} = 4 hrs. Bottleneck unit = unit having
the longest batch unit.
• Makespan (2 batches) = 12 hrs.
• Examples with Zero-Wait (ZW) policy: transfer to units as soon as processing finished.
3.- Multiple Product Batch Plants
Product Processing Time (hrs.)
If Multiple products are produced, more alternatives Stage 1 Stage 2
f scheduling (Campaigns). A 5 2
Example: production of 2 batches of A and 2 batches of B. B 2 4

a) Single Product Campaigns (SPC)

Single product campaign • Cycle time A = 5 hrs.
5 5 2 2
Stage 1 • Cycle time B = 4 hrs.
2 2 4 4  CTA+B = 9 hrs.
Stage 2 (To repeat AABB).
• Makespan (2 batches)
A A B B Time
= 20 hrs.
CT MSpan

b) Mixed Product Campaigns (MPC)

Mixed product campaign
5 2 5 2
Stage 1 • CTA+B = 7 hrs.
(To repeat ABAB).
2 4 2 4
Stage 2
• Makespan (2 batches)
= 18 hrs.
A B A B Time

CT MSpan
3.- Multiple Product Batch Plants
With clean-up (changeover) times (e.g. 1 hr.) the results will be:

Mixed product campaign

5 1 2 1 5 1 2
Stage 1

2 1 4 1 2 1 4
Stage 2

A B A B Time

CT MSpan

• CTA+B = 9 hrs. (To repeat ABAB).

• Makespan (2 batches) = 21 hrs.

4.- Transfer policies
ZW: zero-wait, transfer immediately.
NIS: no intermediate storage. Need not transfer immediately, can store in unit.
UIS: unlimited intermediate storage, between stages in storage vessels.

In practice, plants will normally have a mixture of the three transfer policies.

Product Processing Time (hrs.)

Stage 1 Stage 2 Stage 3
A 6 4 3
B 3 2 2

Example: production of the same number of batches of

each product using a sequence ABAB.

Product Processing Time (hrs.)

j1 j2 j3
i1 t11 t12 t13
i2 t21 t22 t23

Is the absolute minimum (UIS) with:

 N  tij: processing time of product i for stage j.
CTUIS = max  ∑ niτ ij  ni: number of batches for product i.
j=1...M
 i=1  M: number of stages.
N: number of products.
4.- Transfer policies
ZW transfer
6 2 3 6 2 3 • CTA+B = 6 + 3 + 2 (slack) =
Stage 1
11 hrs. (To repeat ABAB).
4 1 2 4 1 2
Stage 2
• Makespan (2 batches) =
3 2 3 2
Stage 3 26 hrs.

A B CT A B MSpan Time

NIS transfer
6 3 1 6 3 1 • CTA+B = 6 + 3 + 1 (slack) =
Stage 1
10 hrs. (To repeat ABAB).
4 2 1 4 2 1
Stage 2
• Makespan (2 batches) =
3 2 3 2
Stage 3 25 hrs.

A B CT A B MSpan Time

UIS transfer
6 3 6 3 • CTA+B = 6 + 3 = 9 hrs.
Stage 1
(To repeat ABAB).
4 2 4 2
Stage 2
• Makespan (2 batches) =
3 2 3 2
Stage 3 24 hrs.

A B A B Time
CT MSpan
5.- Parallel units and Intermediate storage
Intermediate storage tanks between stages  can increase the efficiency or equipment utilization.
Parallel units operating out of cycle.

Example: fermentation plant. Stage 1 (fermenter) takes 12 hrs. Stage 2 (separation) only 3 hrs.
Assume zero-wait transfer and the size of the batch in each stage is the same (1000 kg).

12 12
Step 1
3 3
Step 2

A CT A Time

Alternatives to decrease CT:

a) Intermediate storage between stages.
b) Two parallel units in Stage 1 (Bottleneck unit).
5.- Parallel units and Intermediate storage
a) Intermediate storage between stages: change batch size. Decoupling the two stages so that each
stage can operate with different cycle times and batch sizes.

• CTstage 1 = 12 hrs.;
1000 kg 12 12
Stage 1 handles batches of
1000 kg.
Store 750 kg
3 3 3 3 3 3 3 3
Stage 2
• CTstage 2 = 3 hrs.;
handles batches of
A A
CT
A Time 250 kg.

b) Two parallel units in stage 1 (Bottleneck unit).

CT = max {τ ij / NPj }
j=1...M
12 12
Stage 1- I
tij: processing time of
12 12
Stage 1- II product i for stage j.
3 3 3 3 NPj : number of parallel units.
Stage 2
M: number of stages.
A A A
€ Time
CT
• CT = max {12 / 2, 3}
= 6 hrs.

• The cycle time has been halved  can reduce the batch size to 500 kg.
• 4 fermenters eliminate all idle times.
6.- Synthesis of Multiproduct Batch Plants (p. 196 Biegler (et al.) 1997)
• STRUCTURAL DECISONS:
a) Assignment of tasks to equipment.
b) Number of parallel units or intermediate storage.
• SIZING DECISIONS:
a) Equipment sizing.
• SCHEDULING DECISIONS:
a) Campaigns and transfer policies.
b) Length of production cycles.
c) Sequencing at products.
Assignment of tasks to equipment:
Recipe  Successive tasks.
Ex. Single product batch process with four processing tasks.
Task N° Task 1  Task 2  Task 3  Task 4
Operation (t) Mix (2 hrs.) React (4 hrs.) Mix (1 hrs.) React (2 hrs.)
Material C/S C/S C/S S/S
1. The simplest alternative is to assign each ZW policy
task to one processing device: Stage 1
2

4
Stage 2
1
Stage 3
2
Stage 4

• Cycle time = 4 hrs. • 4 pieces of equipment. Time

6.- Synthesis of Multiproduct Batch Plants
2. The Second alternative is to assign Tasks 3 and 4 to one single piece of equipment, namely
to the stainless steel (S/S) reactor. In this alternative, the cycle time remains unchanged in 4
hours despite the fact that we have eliminated one piece of equipment. This alternative is
clearly superior to the previous one. Thus, a simple design guideline that we can postulate
is: "Merge adjacent tasks whose sum of processing times does not exceed the cycle time".

ZW policy
2
Stage 1
4
Stage 2
3
• Cycle time = 4 hrs. Stage 3

• 3 pieces of equipment. Time

3. A Third alternative is all tasks merged in one piece of equipment  the jacketed stainless
steel vessel that can perform the four tasks.

ZW policy

9
Stage 1

• Cycle time = 9 hrs. Time

• 1 piece of equipment.

Economic evaluation  The best Alternative

7.- Sizing Batch Processes: Single Product Plants
Example: demand of 500,000 lb/yr. of product C. The plant is assumed to operate 6000 hours per
year. The recipe for producing product C is as follows:

Recipe in weight:
1. Mix 1 lb. A, 1 Ib B, and react for 4 hours to form C. The yield is 40% and the density of the
mixture, rm is 60 Ib/ft3.
2. Add 1 Ib solvent and separate by centrifuge during 1 hour to recover 95% of product C. The
density of the mixture is rm 65 Ib/ft3.

Mass Balance 1 lb Solv

1,2 lb
1 lb A A, B
2,24 lb
A, B, solvent
0,8 lb C
1 lb B

0,76 lb C
product

Define Size Factors, Sj, for each stage j:

Sj = volume vessel j required to produce 1 lb of final product.
7.- Sizing Batch Processes: Single Product Plants
• STAGE 1: specific volume v = 1 / rm = 1/60 = 0.0166 ft3/lb mix.
Size Factor: S1 = 0.0166 ft3/lb mix · [2 lb mix / 0.76 lb prod.] = 0.0438 ft3/lb prod.

• STAGE 2: specific volume v = 1 / rm = 1 / 62.5 = 0.0153 ft3/lb mix.

Size factor: S2 = 0.0153 ft3/Ib mix · [3 lb mix / 0.76 lb prod.] = 0.0607 ft3/lb prod.

• If One Unit per Stage, ZW policy transfer:

- Cycle time = Max {4, 1} = 4 hrs.  N° Batches = 6000 h / [4 hrs / batch] = 1500 batches.
- Batch Size Product i (Bi): B = 500,000 lb prod. / 1500 = 333 lb prod.
- Size Vessel 1: V1 = S1 · B = 0.0438 ft3/lb prod. · 333 lb prod. = 14.6 ft3.
- Size Vessel 2: V2 = S2 · B = 0.0607 ft3/lb prod. · 333 lb prod. = 20.2 ft3.

• Bottleneck Stage 1  2 Parallel Units in Stage 1:

- Cycle time = max {4 / 2, 1} = 2 hrs.  N° Batches = 6000 / 2 = 3000 batches.
- Batch Size = 166 lb.
- Stage 1: 2 vessels, V1 = 7.3 ft3.
- Stage 2: 1 vessel, V2 = 10 ft3.

VT = 24.6 ft3, 3 vessels < VT = 34.8 ft3, 2 vessels.

Reduction in the investment cost depending on the cost correlation.
7.- Sizing Batch Processes: Single Product Plants
• Sizing depends on how the plant is scheduled.
• Simplest Alternative: Single Product Campaigns with fixed production cycle.

Example: demand 500,000 lb/yr of A, and 300,000 lb/yr of B. The plant is assumed to operate 6000
hours per year. We will select arbitrarily a production cycle of 1000 hours (42 days), which implies
[6000 / 1000 = 6 campaigns] over one year.

Processing Time Side Factors

Product
(hrs.) (ft3/lb prod.) One Unit per stage, ZW policy

Stage 1 Stage 2 Stage 1 Stage 2 tA 4 tB 4

A 8 3 0.08 0.05

B 6 3 0.09 0.04 1000 hrs. Time

Cleanup times: 4 hrs. A to B, B to A.

Cycle time A = CTA = 8 hrs., Cycle time B = CTB = 6 hrs.;

Effective time for production in each cycle = tA, tB ; tA + tB = 992 hrs.

How to allocate the production of A and B (i.e. selecting tA, tB) during this time horizon?
7.- Sizing Batch Processes: Single Product Plants
A simple solution is to use as a heuristic the same batch size for all products. The batch size Bi or
product i is given by:

Bi = Production I (Pi) / n° batches i = Production I (Pi) / [production time ti / CTi]

Production per campaign (Pi)  PA: 500,000 / 6 = 83,333 lb ; PB: 300,0000 / 6 = 50,000 lb.
Applying the heuristic of equating the batch sizes and constraining the production times to 992 hours
yields the two equations:
- Linear Equations: 83,333 / [tA / 8] = 50,000 / [tB / 6] ; tA + tB = 992.
- Solutions: tA = 684 hrs. ; tB = 308 hrs. ; BA = BB = 974 lb.

The required volumes for each product in the two stages (Vij = Sij Bi):

Product Volumes Vij (ft3)

Stage 1 Stage 2

A 77.9 48.7

B 87.7 39.0

The largest volumes to be selected in each stage are given by: Vj = max{Vij }
i=1,N
 V1 = 87.7 ft3 ; V2 = 48.7 ft3.
8.- Inventories
Selection of the Production Cycle (PC): Trade-off.
Fraction of transition or cleanup times vs. Inventories.

PC ↓  ↓ Inventory (products available frequently) + ↑ fraction of the transitions.

PC ↑  ↑ Inventory (production less frequently) + ↓ fraction of the transitions.

Example: demand 500,000 kg/yr of A, and 1,000,000 kg/yr of B. 8000 hrs. horizon time.

Product Processing Time (hrs.)

Stage 1 Stage 2

A 5 3

B 3 4

a) Single Product Campaigns (SPC):

- Assume production cycle = 1000 hrs.  8 campaigns.
- Same Batch Size. PA = 62,500 kg., CT = 5 hrs. ; PB = 125,000 kg, CT = 4 hrs.
B = 62,500 / [tA / 5] = 125,000 / [tB / 4].
tA = 385 ; tB = 615 hrs. ; B = 812 kg.
tA + tB = 1000 hrs.

N° batches A = 62,500 / 812 = 77 ; N° Batches B = 125,000 / 812 = 154.

8.- Inventories
14 days 25 days
A B
tA = 385 tB = 615

Time
1000 hrs.

• Assume Constant demand rates, dP:

dA = 62,500 / 1000 = 62.5 kg/h. ; dB = 125,000 / 1000 = 125 kg/h.

• Production rates, pP:

pA = 62,500 / 385 = 162.3 kg/h ; pB = 125,000 / 615 = 203.3 kg/h.

• Inventory profile for A:

0 – 385: accumulation rate = pA – dA = 162.3 – 62.5 = 99.8 kg/h.
385 – 1000: depletion rate = – dA = – 62.5 kg/h.
From 0 – 385 produced 62,500 kg,
sold 24,257 kg.  Average Inventory
38.423 kg = area under curve.
A
1
I=
τ
• ∫ f (t)dt

0 hrs. 385 hrs. 1000 hrs.

€
8.- Inventories
• Inventory profile for B:
0 – 385: depletion rate = – dB = –125 kg/h.
385 – 1000: accumulation rate = pB – dB = 203.3 – 125 = 78.3 kg/h.

0 hrs. 385 hrs. 1000 hrs.

Average Inventory = area under curve.

I = 24,077 kg

If Inventory cost 1.25 €/kg yr  Inventory Cost = 1.25 (19,211 + 24.077) = 54,110 €/yr.
€
8.- Inventories

5 3 3 5 • CT (ABB) = 12 hrs.
Step 1
3 4 4 3 • N° Cycles = 8000 / 12 = 667 cycles.
Step 2
 667 batches A; 1333 batches B.
A B B A Time • Batch Size = 500,000 / 667 = 750 kg (vs. 812 kg).
CT

• Stage 2 = 12 hrs.
3 4 4 • If daily deliveries: Accumulation =
Step 2
A: 2 · 750 = 1,500 kg.
A B B Time B: 4 · 750 = 3,000 kg.
• Inventory costs = 1.25 (1,500 + 3,000) = 5,625 €/yr.

If clean-up times ≠ 0 Mixed Product Campaigns (MPC) will require

longer batch sizes but will still require lower Inventories.
9.- Further Reading and References
• Biegler, L.; Grossmann, I. & Westerberg, A. (1997): «Systematic methods of chemical
process design». Prentice Hall.

• Seider, W.; Seader, J.; Lewin, D. & Widagdo, S. (2010): «Product and process design
principles. Synthesis, analysis and evaluation». 3rd Ed. John Wiley & Sons.

• Majozi, T. (2010): «Batch chemical process integration. Analysis, synthesis and

optimization». Springer.
PRACTICAL CHAPTER

• Examples of Batch process to obtain CT and makespan.

RELEVANT TO LEARNING (I)

• What are the reasons for designing a batch process instead of a

continuous process?

• Flowshop and Jobshop processes.

• What is a recipe? What are CT and Makespan?

• Campaign (SPC, MPC) and Transfer policies (ZW, NIS, UIS).

• Alternatives for solving Bottleneck units.

• Characteristics of the synthesis of Multiproduct Batch Plants.

• In a Semi-Batch or Batch process, in what cases will it make sense to

use process units in parallel?
RELEVANT TO LEARNING (II)
• Given the processing times of 3 products A, B and C in the table, deter-
mine the Gantt charts, the makespan and the cycle time to manufacture 2
batches of A, 1 batch of B and 1 batch of C in the following cases:
- No Intermediate Storage (NIS) with the AABC sequence and the BAAC sequence.
- Unlimited Intermediate Storage (UIS) with the AABC sequence and the BAAC
sequence.

• Dados los tiempos de procesado de 3 productos A, B y C de la tabla,

determinar las gráficas de Gantt, el makespan y el tiempo de ciclo para
fabricar 2 lotes de A, 1 lote de B y 1 lote de C en los casos:
- No Intermediate Storage (NIS) con la secuencia AABC y la secuencia BAAC.
- Unlimited Intermediate Storage (UIS) con la secuencia AABC y la secuencia
BAAC.

Processing Time (hrs.) (Cleaning Time = 0)

Product/ Step Step 1 Step 2 Step 3

A 5 4 3

B 3 1 3

C 4 3 2