Reduction of lead times and optimization of inventory levels for

meeting on time delivery (OTD) using value stream mapping

-A case study for the fast runner product in IMI Hydronic Engineering, Ljung.

Master’s thesis in the Master programes of Supply Chain Management and

Production Engineering

SHASHIDHAR PAI
KRISHNA TAMILSELVAM

Department of Technology Management and Economics

CHALMERS UNIVERSITY OF TECHNOLOGY
Gothenburg, Sweden 2018, E2018:035
 Master’s Thesis E 2018:035

Reduction of lead times and optimization of

inventory levels for meeting on time delivery
(OTD) using value stream mapping

-A case study for the fast runner product in IMI Hydronic

Engineering, Ljung.

SHASHIDHAR PAI
KRISHNA TAMILSELVAM

Department of Technology Management and Economics

Division of Supply and Operations Management
Chalmers University of Technology
Gothenburg, Sweden 2018
Reduction of lead times and optimization of inventory levels for meeting on time
delivery (OTD) using value stream mapping

-A case study for the fast runner product in IMI Hydronic Engineering, Ljung.

SHASHIDHAR PAI
KRISHNA TAMILSELVAM

© SHASHIDHAR PAI & KRISHNA TAMILSELVAM, 2018.

Supervisor & Examiner: Peter Almström, Technology Management and Economics

Master’s Thesis E 2018:035

Department of Technology Management and Economics
Division of Supply and Operations Management
Chalmers University of Technology
SE-412 96 Gothenburg
Telephone +46 31 772 1000

iv
Reduction of lead times and optimization of inventory levels for meeting on time
delivery (OTD) using value stream mapping
-A case study for the fast runner product in IMI Hydronic Engineering, Ljung.
SHASHIDHAR PAI
KRISHNA TAMILSELVAM
Department of Supply Chain Management and Production Engineering
Chalmers University of Technology

Abstract
IMI Hydronic Engineering is the provider of hydronic components for Heating, Ven-
tilation and Air Conditioning systems (HVAC) systems. The main vision for the
company is to provide challenging innovative and sustainable energy efficient so-
lutions for hydronic problems with high level of delivery accuracy, customer sat-
isfaction and cost efficiency. Their product range includes various climate control
products with integrated intelligent solutions. One of their facilities is located in
Ljung. This facility produces many product families and product varieties out of
which the product flow of their fast runner product, product SK is investigated in
this thesis.

Due to the customer requirements, the company produces wide variety of hydronic
components. This leads to substantial increase in the product variants which gets
reflected in the product mix in the production. The complexity of the production
flow increases as the number of variants and the production mix increase. This in
turn affects the on time delivery (OTD). Currently the company is striving to keep
up their delivery performance with lots of inventories in the production flow.

This thesis work mainly focuses on the enhancement of On Time Delivery (OTD)
by providing suggestions for the reduction of production lead time with optimal
inventory levels in the production flow of product SK. Since the demand is seasonal,
keeping safety stock in the inventories and making the production system to reflect
to the present demand is an important parameter.

The main purpose of this thesis is to reduce the production lead time with opti-
mal safety stock and reorder points in the production flow of product SK in order
to make the production system to produce in tune with the customer demand for
that period of time.

Currently the production lead time is 9.5 days. After setting the optimal safety
stock and reorder points, the planned production lead time is 2.17 days. However
for running the production system with optimal stocks, this thesis work also sug-
gests some other improvements on the production line. The improvement functions
include placement of pacemaker and supermarkets in the production flow. The
suggestions for the improvement are shown in the future state map.

Keywords: VSM, Safety stock, Reorder points, CONWIP, OTD.

v
Acknowledgements
This report is written as a master thesis in Chalmers University of Technology in
collaboration with IMI Hydronic Engineering by the students from Supply Chain
Management and Production Engineering master programmes.

First, we would like to thank our supervisor Peter Almström from Chalmers, for
his splendid support and valuable time.

We would like to thank our industrial supervisors at IMI Hydronic Engineering,

Ljung, Anna Ehn and Fredrik Johansson for entrusting us and for their great sup-
port. We would like to thank the people involved in this thesis project in IMI
Hydronic Engineering, Ljung, for their timely help and support.

We like to like extend our thanks to our friends and family who have helped us
a lot during this project.

Shashidhar Pai and Krishna Tamilselvam, Gothenburg, 2018

vii
 Contents

List of Figures x

List of Tables xi

1 Introduction 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Goals and Deliverables . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.5 Scope and Delimitations . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.6 About the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Theoretical framework 5
2.1 Customer order decoupling point and its types . . . . . . . . . . . . . 6
2.2 OTD - On Time Delivery . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Why lead time decrement and buffer optimisation for OTD improve-
ment? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 Lead time reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.5 Reorder point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6 Safety stock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.7 Kanban systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.8 CONWIP system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.9 Value stream mapping . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3 Methodology 21
3.1 Overall research strategy . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2 Background study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.3 Literature study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.4 Data collection and validation . . . . . . . . . . . . . . . . . . . . . . 22
3.4.1 Qualitative data . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.4.2 Quantitative data . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.4.3 Data validation . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.5 Current state map . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.6 Approach for Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.7 Future state map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4 Current production 27

ix
Contents

4.1 Current production flow . . . . . . . . . . . . . . . . . . . . . . . . . 27

4.2 Current information flow . . . . . . . . . . . . . . . . . . . . . . . . . 32

5 Analysis 35
5.1 Current state map . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.2 Future state map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

6 Recommendations for the company 43

6.1 Improvements for implementation of the proposed future state . . . . 43
6.2 System level improvements . . . . . . . . . . . . . . . . . . . . . . . 45
6.3 General recommendations . . . . . . . . . . . . . . . . . . . . . . . . 46

7 Discussion and conclusions 47

Bibliography 49

A Appendix 1 I
A.1 Interview questionnaire for planner . . . . . . . . . . . . . . . . . . . I
A.2 Interview questionnaire for Supply chain
planner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III
A.3 Interview questionnaire for Industrial
Engineers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV
A.4 Interview questionnaire for quality manager . . . . . . . . . . . . . . V

B APPENDIX B VII
B.1 Symbols used in Value Stream Mapping . . . . . . . . . . . . . . . . VIII

x
 List of Figures

2.1 Overview of framework used for literature . . . . . . . . . . . . . . . 5

2.2 Kanban system vs CONWIP system (Jonsson and Mattsson, 2009) . 17

3.1 Overall research strategy for the thesis project . . . . . . . . . . . . . 21

3.2 Approach used for analysis . . . . . . . . . . . . . . . . . . . . . . . . 24

4.1 Process flow map for product SK . . . . . . . . . . . . . . . . . . . . 28

4.2 Layout of the production system . . . . . . . . . . . . . . . . . . . . . 29
4.3 Current state value stream map of product SK . . . . . . . . . . . . . 31

5.1 Future state map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

xi
List of Figures

xii
 List of Tables

2.1 Relation between desired service level and service factor (Constantin,
2016) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

5.1 Values for the current state production . . . . . . . . . . . . . . . . . 35

5.2 Values used for calculating safety stock . . . . . . . . . . . . . . . . . 39
5.3 Safety stock and reorder point values . . . . . . . . . . . . . . . . . . 40
5.4 Values for proposed future state map . . . . . . . . . . . . . . . . . . 40

xiii
List of Tables

xiv
 1
Introduction

This chapter contains a brief description about the case study company, about the
product SK, problem identification, aim and purpose of the thesis along with the
scope and delimitation.

1.1 Background
IMI Hydronic Engineering is one of the pioneers in providing all types of hydronic
components for the HVAC (Heating, Ventilation and Air- Conditioning systems). It
was started in the year 1928 and it was acquired by a British manufacturing group,
IMI Plc., in 1997. This group comprises of three brands that includes IMI Hydronic
Engineering, IMI Precision Engineering and IMI Critical Engineering. Currently
IMI Plc operates in more than 61 countries across the globe. IMI Hydronic Engi-
neering has its headquarters in Switzerland and currently 400 employees are working
in their facility in Ljung.

IMI Hydronic Engineering started adapting lean techniques and developed a strat-
egy based on lean philosophy. They started practicing it under the name IMI way
in 2010. Since then, they are training people and implementing various lean tools
and techniques in their processes and in the production. They have altered their
performance measurement indicators based on the IMI way. They have five KPI
(key performance indicators) which includes OTD (on time delivery), IT (inventory
turn), productivity, LTA (loss time accident) and COQ (cost of quality).

The facility in Ljung produces more than 100 product varieties and their main
customers are in Nordic countries, in Poland and in Germany. Because of the rise
in the number of product variants, there has been a significant increase in the com-
plexity of the production processes and inventories in the production flow. This in
turn gets reflected in the on time delivery (OTD) of the finished products.

AF product family is one of their fast running product family and it has several
product variants based on the dimensions and minor functionality changes. All the
variants of the product family AF follows the same production flow. This thesis
focuses on the production flow of product variant SK, which is one of their fast
running product for IMI Hydronic Engineering. The names of the product and the
product family has been changed because of confidentiality reasons.

1
1. Introduction

Product SK is a type of valve used for mixing the hot and cold fluids in the HVAC
systems and it has more than 30 components in its bill of materials(BOM) and
the major components are made of Brass. The dimension of the product SK is
74*110*54mm (length*height*width). There are three main elements of product
SK, that is, the main body, the other metal parts and a few rubber components.
The main body part is manufactured in the foundry and the other metal parts are
manufactured in the bar machining area. Both the foundry and the bar machining
are located within the facility and the rubber components are procured from exter-
nal suppliers. Components for product SK are assembled to the casted body part
in the final assembly in the production flow.

The company has categorized the product items based on the sales volumes and
the delivery strategy is to ship their fast runner products within a day. The com-
pany follows ‘make to stock’ decoupling strategy for faster delivery of goods to their
customers. The company strives to achieve on time delivery (OTD), in order to
provide high service level to their customers. As a result, the company has large
number of parts in the final inventory and also in many parts of the production
flow. Production lead time and inventory management are important parameters,
for achieving on time delivery (Ndubi, et. al., 2016) (Rother and Shook, 2009).

One another important cause for the huge inventories is the seasonal demand of
the products. As the demand for the HVAC products are high in the spring and
summer seasons, in order to satisfy the market requirements, the company ends up
with the large inventories in most parts of the year. One such product is SK. It
contributes nearly one third of the final sales volumes of the similar variants of the
valves that are produced in their facility.

1.2 Problem Definition

In order to satisfy the customer requirements, IMI Hydronic Engineering developed
its product portfolio to great extent in recent years. This in turn results in complex
product mixes in the production flow. Since the company has a delivery strategy
of shipping their fast runner products in one day, the current problem faced by the
company is to meet the ‘on time delivery’ (OTD).

Currently the on time delivery for the product SK varies from 65% to 93% with
the average of 84% for the past year. In their facility at Ljung, they are trying to
ensure higher on time delivery rate of 96%.

1.3 Purpose
The purpose of this master thesis project is to find the improvement potentials for
meeting the on time delivery (OTD) rate of product SK.

2
 1. Introduction

1.4 Goals and Deliverables

The goals and deliverables of this thesis project includes

• Using value stream mapping to understand and have an overview of the pro-
duction flow of product SK.
• Analyzing the current state map to identify factors affecting OTD for product
SK, in that production flow.
• Analyzing the identified factors and determine solutions for improving OTD
• Proposing future state map with the improvements for the product SK.

1.5 Scope and Delimitations

The scope of the thesis project is limited to the production flow of one particular
product SK in the product family AF.

Value stream is only done for the body part which comes from the foundry as it
takes the longer lead time when compared to the other parts in the bill of materials
(BOM) of the SK. The thesis project is limited to the in-house flow of materials
that is from foundry to final warehouse.

This master thesis focuses on improving the On Time Delivery (OTD). The inter-
relations between the OTD and other KPI’s are not taken into account.

1.6 About the thesis

The structure of our thesis report is explained in this section. It also contains a
brief insight of the contents covered under each chapters in this thesis.

chapter 2 - Theoretical framework

This chapter remains as a backbone of the thesis and it constitutes of the theory
relating to the concepts discussed in this Master’s thesis.

chapter 3 - Methodology
This part of the thesis report explains about the methodology in which the Master’s
thesis has been conducted.

chapter 4 - Current production

This chapter of the thesis describes about the data collected and the current pro-
duction state of product SK along with the current state map.

3
1. Introduction

chapter 5 - Analysis
This chapter of the thesis explains about the analysis of the current state with the
theoretical framework as its base. This chapter also explains about the proposed
future state map.

chapter 5 - Recommendations for the company

This part states the results of the analysis along with the improvements that are
needed for the implementation of the proposed future state map for improving the
production flow.

Chapter 7 - Discussions and conclusion

This chapter consists of discussions of the obtained results and the final conclusions
of this master thesis project. This chapter also explains about the points to consider
for the effective implementation of the proposed future state.

4
 2
Theoretical framework

This theoretical framework consists of the relevant literature pertaining to the con-
cepts used in this thesis along with the ways of improving On time delivery (OTD).
The figure 2.1 explains about the framework used for this literature study, followed
by literature about OTD and its importance along with the general concepts used.
Since the project scope is limited to in-house production of product SK, we have
limited our search to the factors affecting OTD in relation to internal processes.
The literature relating to the above aspect, especially for in-house production, was
limited. Therefore, two factors were selected and the process of selecting it is ex-
plained in later part of theoretical framework and as well as in methodology section.
The two factors are, lead time decrement and ways of having effective control of
buffers. Finally, the literature framework explains about value stream mapping, the
tool used for performing the master thesis.

Figure 2.1: Overview of framework used for literature

The framework for this literature study consists of two phases that includes identifi-
cation of improvement potentials and identification and study of factors influencing
the preferred improvement potentials. In the first phase, the concept of OTD is
explained along with its importance followed by the identification and selection of
the improvement potentials. In the second phase of the literature study, various
factors influencing the identified potentials is explained.

5
2. Theoretical framework

2.1 Customer order decoupling point and its types

Manufacturing companies can be classified by, the way the operations are carried
out based on customer orders. The level of integration between the production op-
erations and customer orders plays an essential role in managing an efficient flow.
This is done by defining customer order point and customer order decoupling point
(Jonsson and Mattsson, 2009). Customer order point refers to the point in products’
bill of material from where the material supply is based on the specific customer
order. All the production operations before that are usually generic and have the
same processes. After the customer order point the products’ appearance and char-
acteristics are customer order specific. Furthermore, it also defines the delivery lead
times which should be at least calculated from customer order point to the time
it takes to complete the entire manufacturing operations (Jonsson and Mattsson,
2009) (Karim et al., 2009). Customer order decoupling point refers to the point in
bill of material where the material supply and value added activities are based on
customer order received. This is very essential as the material planning and delivery
dates before this point are based on forecasts (Jonsson and Mattsson, 2009).

Based on the customer order decoupling point there are certain manufacturing
strategies that can be implemented for better customer order integration. They
are engineer to order (ETO), make to order (MTO), assemble to order (ATO), and
make to stock (MTS) (Jonsson and Mattsson, 2009). ETO means that the products
are made according to the specifications of the customers and the companies must
try to be flexible and adapt according to the demand of the customers, possibly,
at every stage of the production. The customer order decoupling point lies in the
earlier stages of the production flow, in this case. MTO is also quite similar to
ETO, the only difference is that certain parts, sub-assemblies, and semi-finished
products are produced or procured from suppliers before customer orders are re-
ceived. ATO means that the parts, sub-assemblies, and semi-finished products are
produced without any relation to the customer orders. These are assembled together
once the customer order is received. The semi-finished products usually have stan-
dardized items which are then assembled as per customer order. MTS means that
the company is fully aware of the customer order specifications and then manufac-
tures them to keep it in stock which is shipped as soon as the customer orders are
received. This is also possible as the products are completely standardized with little
or different operation along the flow. Furthermore, the forecasts, delivery schedules,
resource availability, inventory levels, etc are the ones that dictate the initiation of
manufacturing operations (Jonsson and Mattsson, 2009).

6
 2. Theoretical framework

2.2 OTD - On Time Delivery

Manufacturing includes all the processes and activities from order receipt to fi-
nal shipment of the product to the customer (Karim et al., 2009). For achieving
competitive advantage, the firm must focus on improving one or mix of factors or
parameters rather than concentrating on all (Tersine and Hummingbird, 1995). As
mentioned earlier the five KPI’s for the case company are OTD (on time delivery),
IT (inventory turn), productivity, LTA (loss time accident) and COQ (cost of qual-
ity). Now, if we focus on improving OTD, the other KPIs such as inventory turns
and productivity is also influenced positively when OTD is improved.

The firm should measure in terms of delivery performance, if their focus is on lead
time and quality parameters (Tersine and Hummingbird, 1995). OTD (On Time
Delivery) is one of the important measuring parameter for a manufacturing com-
pany’s delivery performance and it should be measured based on the OPD (original
promised date) to the customer (Ramachandran and Neelakrishnan, 2017). In other
words, OTD can be improved by reducing the lead times and thereby delivering
products to customers on time.

Karim et al. (2009) and Ramachandran and Neelakrishnan (2017), argue that for a
sustained improvement of on time delivery (OTD), all the processes and functional
units included in the manufacturing should be taken into account. Thorsen and Yao
(2015), states that for maintaining higher service levels, it is important to calculate
the optimal inventory ordering policies. On the other hand, Karim et al. (2009)
confronts that inventory control and optimization of inventories alone have weaker
impact over the On Time Delivery (OTD) performance when it is improved sepa-
rately and it needs to be done with the improvement of a parameter or the mix of
other parameters.

2.3 Why lead time decrement and buffer optimi-

sation for OTD improvement?
Literature relating to the improvement of On time delivery (OTD) performance, for
in-house production, is scarce. However, from the available literature, the stakehold-
ers involved and from the authors’ academic knowledge, the authors have narrowed
down to lead time decrement and optimization of buffers. The motivation for se-
lecting these two factors are also that they are much more related to the case study
company and the product selected for the thesis.

Production lead time is one of the most important parameters for delivering prod-
ucts on time and a varying lead time leads to many consequences throughout the
supply chain (Ndubi, et. al., 2016) (Wouters, 1991). For example, one of the issues
of long or varying lead times is the cost related to expediting and fixing the prob-
lems caused in scheduling. Also, the planning horizons are made for longer periods
and this leads to even more inventories, which again is an added expense to the

7
2. Theoretical framework

company (Tersine and Hummingbird, 1995). Furthermore, Wouters (1991), states

that, reducing the lead time, reduces the delivery lead time. If the customer order
decoupling point lies downstream, the delivery lead time can be improved by both
lead time reduction and optimization of buffers.As mentioned earlier, the case com-
pany works in a make to stock environment, which means the decoupling point lies
downstream near the customer. Therefore, the delivery lead time in case company
can be improved by both lead time reduction and optimization of buffers.

Excess inventories are a type of waste and it is one of the reason for the long lead
times (Rother and Shook, 2009). It is considered as a waste as it does not add value
and is, therefore, an unnecessary cost to the company. But at the same time, inven-
tory might be required for products having high service level and then, in this case,
it is not considered as a waste. Therefore, it is first necessary to understand what
is considered as waste in perspective of the company or rather, in relation to the
need of the company (Tersine and Hummingbird, 1995). Furthermore, optimizing
the buffers helps in having effective control over the production flow. To summarize,
reduction in the production lead time and excess inventories will help in enhance-
ment of delivery performance by delivering products on time to the customers with
higher service levels, thereby leading to higher On Time Delivery (OTD) rates.

The following chapters of the literature study explains about the lead time reduction
and the ways of having effective control and execution in production systems.

2.4 Lead time reduction

As mentioned earlier, production lead time is one of the most important parame-
ters for delivering products on time. Production lead time refers to the time from
when a customer order is taken until it is ready for shipping (Ndubi, et. al., 2016)
(Wouters, 1991). Lead time is divided into value added and non value added ac-
tivities and in most of the companies, the difference between value added time (the
actual time spent on the product) and non value added time is very huge. For
example, the value added time can be 4 hours and the non value added time goes
upto 4 weeks (Hopp, et. al., 1990). Therefore, lead time is an important param-
eter and reducing it is essential for having a competitive advantage. Furthermore,
(Hopp, et. al., 1990) and Wouters (1991), states that lead time reduction helps in
improving the quality, reducing inventories, reducing scrap, reducing the changes in
production, having frozen horizons i.e. having a fixed schedule for immediate few
days, so as to not depend on forecasts and all these lead to several economic benefits.

The need for lead time reduction needs to be defined as well. Mainly because,
it could vary from one company to other. It could be that the lead time needs to be
reduced to reduce inventory, thereby saving on inventory carrying costs. Another
reason could be to reduce the delivery lead time in order to fulfill customer order or
gain competitive advantage (Hopp, et. al., 1990).

8
 2. Theoretical framework

According to (Hopp, et. al., 1990), lead time is affected by three major aspects.
First is the waiting time and queuing time, second is the WIP and finally the vari-
ability in the lead time. The waiting time and queuing time are directly proportional
to the lead time and therefore, it must be dealt accordingly. Similarly, WIP and lead
time are directly proportional for a given throughput. This helps in understanding
the causes of excess lead time by identifying places with large inventories. Finally,
the variability in lead times need to be considered as it is easy to reduce the average
lead time and not keep in mind the former aspect. But in situations when there is
high or low lead times, it causes shortage or excess of products or materials, thereby
creating unnecessary issues.

(Hopp, et. al., 1990), further states certain general strategies to reduce lead times
taking into consideration the above mentioned causes, they are: analyzing the WIP,
keeping the products moving, and synchronizing the different production activities.

As mentioned previously, increased WIP leads to increased lead time and that is
the reason the WIP needs to be analyzed. But at the same time, keeping stock in
buffers might be necessary for avoiding bottlenecks or depending on the need of the
company, for example if high service level is needed for a high runner product. This
will be explained later in the section 2.5. Once the places with large buffers are
identified, it needs to evaluated if those are really needed (Hopp, et. al., 1990). If it
is in excess, then steps must be taken to ensure reducing the buffers. Now, the ques-
tion arises, as to what level it needs to be reduced to or rather what is the optimal
level of inventories that needs to be maintained. This will be explained in section 2.6.

(Hopp, et. al., 1990) states, if it is ensured that the products are moving towards the
completion of manufacturing, then the lead time and buffers will decrease. The main
reason is that, 90%-95% of the time the product is in waiting period. Therefore,
improving the production by having new machines or investing time and resources
in setup reduction will not be of much help in reducing lead times. Instead, the
focus should be on reducing the waiting times. Waiting times can be, waiting for
parts or sub assemblies, the queue time of a product. However, (Hopp, et. al., 1990),
state that setup time reduction is important for increasing capacity, but if reducing
lead time is the main focus, then reducing waiting times is essential. This is also
supported by (Wouters, 1991), who states that increasing machine utilization, in-
creases waiting time and queue time, resulting in longer lead time. One of the ways
to reduce queue time is to directly control the WIP by using tools such as Kanban
or CONWIP systems (Hopp, et. al., 1990). Furthermore, these tools also help in
synchronization of production (Hopp, et. al., 1990) and these two concepts will be
explained in the further sections.

In a pull system, the material planning, execution and control are not separate
functions but rather are initiated based on the consumption of units in the subse-
quent operations (Jonsson and Mattsson, 2009). This means that the individual
control of various functions and departments is not necessary, since all of the above
mentioned parameters are integrated at a single point. The main idea here is to

9
2. Theoretical framework

create a smooth production flow which is based on the customer demand but at
the same time it is kept as simple as possible with fewer inventories and without
much communication of information at every process. It differs from the traditional
production in the sense that there are no production orders for every department.
It is the material planning that directly initiates production (Jonsson and Matts-
son, 2009). For example, the material consumed downstream initiates production
upstream. Similarly, the order sequence in the upstream process will be match-
ing the replenishment signals from the downstream process. The degree to which
this process is possible is dependent on the size of the inventory. The reason is,
there are situations where lead times are too long for immediate delivery, setup
times are too high, higher batch sizes, ordering costs are high, etc which leads to
higher inventory and restricts direct demand delivery (Jonsson and Mattsson, 2009).

It can be seen that even in pull systems, it is not possible to function stockless.
The intermediate stock levels are also very essential mainly because, the time for
a product to go through from the first operation to the last would be very long if
there were no stocks in the middle (Jonsson and Mattsson, 2009). For example,
when material consumed at a downstream operation, it initiates production in the
preceding operation. Now, there might be a situation where there is no input mate-
rial from that preceding operation, and the reason might be because of high setup
times, higher batch sizes, etc. As this situation can occur anywhere in the flow, it
is necessary to have stocks, but at optimal levels. Another reason to have interme-
diate stock levels is that it balances out the minor variations in the production line
between any successive operations (Jonsson and Mattsson, 2009). For this purpose
intermediate buffers can be used, as sometimes there cannot be continuous flows
between all the processes (Rother and Shook, 2009).

Furthermore, these buffers should be linked to the downstream processes instead

of individually scheduling them (Rother and Shook, 2009). Linking them with the
downstream processes creates a pull based system, as it will depend on the con-
sumption of the downstream process. Whereas, in a push system, the order are
scheduled to the upstream process directly, without considering the consumption of
the products in the downstream process.

By using buffers in a pull based system, the upstream production can be controlled
from this point and this point is called the pacemaker (Rother and Shook, 2009). In
a traditional push based system, the production scheduling of each department is
controlled individually and thereby having a greater task of coordinating it. Having
only one production control point, in a pull based system, will reduce this need of
coordination and it is even more beneficial when there is mix model based produc-
tion. The pacemaker, therefore, sends signal to the upstream process based on the
demand and only then will the production start, thereby creating a pull flow (Rother
and Shook, 2009). The downstream process from the pacemaker process needs to
be in a continuous flow, up until the finished product (Rother and Shook, 2009).

10
 2. Theoretical framework

Although, there is justification for keeping stock, the main idea here is to keep
it at a very minimal level by leveling the production, minimizing the setup times
and batch sizes as much as possible. Therefore, for a pull system to work, it is first
necessary to have a proper planning environment where visibility and control can be
established. Furthermore, traditional planning methods can also be implemented in
parallel to the pull systems as in some cases it is not possible to fully implement
pull systems (Jonsson and Mattsson, 2009).

As explained earlier, for a smooth production flow the preceding operations are
initiated based on respective subsequent operations. To keep this running in an
efficient way it is, therefore, very essential to have the downstream operations lev-
eled which means that there needs to be necessary production schedules, along with
smaller batch sizes and every variant of the items should be scheduled at regular in-
tervals. This creates a smoother flow with leveled production and higher the demand
volume, the better it is for the material flow. Determining the demand volume is an-
other essential factor as these pull systems can manage +/- 10 % demand variation
(Jonsson and Mattsson, 2009).

2.5 Reorder point

To balance the supply and demand, there can be many planning methods which
can be implemented based on the manufacturing environment. It needs to answer
two questions i.e. when the orders start and when the delivery will take will place
(Jonsson and Mattsson, 2009). One of the methods is reordering point where a com-
parison method is used to initiate a manufacturing or a procurement order (Jonsson
and Mattsson, 2009). The comparison is made between the available stock and a
reference level of stock (i.e. the reorder point). If the available stock goes below the
reference stock level a manufacturing or procurement order is initiated to replenish
the respective quantity of stock. Furthermore, if any planned orders results in quan-
tity of stock going below this reorder point, a replenishment order can be initiated.
The reorder point is calculated taking into consideration the lead time required for
replenishment of the respective stock. Also, certain levels of safety stocks are calcu-
lated, taking into consideration the variations in demand, lead times and any other
unpredictable events (Jonsson and Mattsson, 2009).

It might be seem that, since material flows are managed by planning and scheduling,
this is a push flow. But at the same time, if the downstream process consumes a
unit from the buffer and then the signal is sent to the preceding process, it is a pull
flow. Another aspect to look at is whether the planning and control takes place
from a central unit or not (Jonsson and Mattsson, 2009). If manufacturing orders
are given out to each and every functions/departments it is clearly a push flow. If
it is customer order initiated or centrally controlled it is a pull flow. Therefore, the
material planning methods cannot be classified as push or a pull flow; it is their
way of use and application of it, which makes it a push or a pull flow (Jonsson and
Mattsson, 2009).

11
2. Theoretical framework

For example a reorder point may be used to order material, initiated by the subse-
quent operation after consumption of the same; which makes it a pull environment.
Similarly, the planning department can order to fill up stocks, which is needed for
future sales; and this makes it a push environment.

The ideal way to calculate reorder point is to add the safety stocks and demand
during lead time. The formula used is R.O.P = Safety stock + (D*L) (Jonsson and
Mattsson, 2009). The demand during lead time is the units that will be consumed
during the replenishment of the stock. This helps to distinguish between the vari-
ation of demand that is dependent on lead time or on other uncertain aspects. By
collecting the information on demand, consumption, variations, quantities in stock
an average value for reorder point can be obtained. Furthermore, this can be used
to maintain optimal tied up capital and delivery schedules (Jonsson and Mattsson,
2009). A fixed reorder point can be used if there is some random variation expected
with respect to demand and lead times in the future. Similarly, if there is seasonal
demand, then the reorder point must be calculated accordingly by adjusting the
deviations and safety stock. In this way, the finished goods stock can be maintained
using the reorder points.

This is used to efficiently and effectively manage inventory even when the demands
and lead times vary, as it provides a structure for how much to order and when to
order (Coyle, Bardi and Langley, 2003). Furthermore, the four ways to use reorder
point are: fixed quantity with fixed interval, fixed quantity with irregular interval,
irregular quantity with fixed interval and irregular quantity with irregular interval.
If the demand and the lead time is known then the fixed quantity with fixed interval
can be used. If the demand varies and the lead time varies then either the fixed
quantity with irregular interval or the irregular quantity with fixed interval is used
(Coyle, Bardi and Langley, 2003).

2.6 Safety stock

Determining safety stock is an essential part of calculating reorder point, as ex-
plained previously. Companies usually keep it to prevent stock outs and as a buffer
against variations in demand. Safety stock optimization, therefore, leads to sav-
ings and better inventory turnover. Most of the companies have a perspective of
stocks as an added cost, but it should be rather seen as an element that actively
contributes towards developing or maintaining specific market through high levels
of service. The main objectives of stock management is, therefore, not only reduc-
ing inventory related costs, but also, retaining customers and acquiring new market
shares through improved service level thereby providing better service and product
availability, as per the requirements of the customers (Jonsson and Mattsson, 2009).

Having a better service level, to avoid stock outs and losing sales during a replen-
ishment cycle, means having higher safety stock levels. These higher safety stock
levels is considered as a value, as the company needs to provide high service levels
to customers (Tersine and Hummingbird, 1995). At the same time it should not

12
 2. Theoretical framework

be very high that the company loses money on inventory carrying costs, but should
have an ideal balance. Therefore, companies have a trade-off between opportunity
costs and operation cost, with respect to maintaining the service level (Jonsson and
Mattsson, 2009). Reducing inventory frees up money but at the same time the
company might lose on customer satisfaction and it is complex to measure on such
intangible factors. This also depends on a lot of other factors such as customer
sensitivity to stock outs, the risks involved in maintaining high level of stocks such
as storage costs, products getting expired, the prices of the products getting lower,
etc. Therefore, optimizing stock levels is complex and specific to the type of market
environment the company is functioning in (Jonsson and Mattsson, 2009).

One of the approaches to this dilemma is the ABC analysis where the products
are classified based on the revenue they generate (Jonsson and Mattsson, 2009).
The more essential products, which generate more sales, are assigned higher service
level. Therefore depending on the revenues generated the service level is assigned.
The top 20% of the products which are of strategic importance to the company are
classified as items ‘A’ and they are assigned with a high service level of 96%-98%.
The next 20%-30% of the products which are of intermediate importance are clas-
sified as items ‘B’ and they are assigned a medium service level of 91%-95%. The
last 50%-60% of the products which are of relatively lower importance are classified
as items ‘C’ and they are assigned a lower service level of 85%-90% (Jonsson and
Mattsson, 2009). Furthermore, if required, the service level for each individual prod-
uct can be determined, but at the same time the company must have the required
technological infrastructure to achieve this aspect. Therefore, for a company to de-
cide on the target service level there is a trade-off between cost of the inventory and
cost of the stock outs. The company needs to find a balance between the two and
decide in which way to optimize their inventory, keeping in mind the costs involved
(Jonsson and Mattsson, 2009).

Based on the consumption pattern of the products, there are two ways of deter-
mining safety stocks. The first one is based on manual calculations and estimations
and the second one is based on scientific calculations and collecting maximum pos-
sible information about the uncertainty (Jonsson and Mattsson, 2009). In manual
calculations the results are not aimed at a specific parameter to optimize safety
stocks, for example, like service level, inventory carrying costs, stock out costs. It
is quite resource consuming in creating such a system and even more complicated
when revising the same. For calculating safety stocks using scientific calculations
there are various options such as calculating safety stocks based on lead times and
safety stocks based on demand.

For safety stock based on lead times, it is calculated as a percentage of demand

during lead time (Jonsson and Mattsson, 2009). In this the quantity in the safety
stock is determined by the size of the demand and the duration of the lead time.
In this way the safety stock can be determined when the demand or the lead time
changes. But the main disadvantage is that this method does not take into consid-
eration the variation in demand or the forecast errors. For example, the calculation

13
2. Theoretical framework

requires average demand throughout the entire lead time in consideration giving the
same safety stock for that entire period. But it might be the case where the de-
mand variation is very high for a particular product and that safety stock calculated
for the average demand might not be sufficient or would be too high (Jonsson and
Mattsson, 2009).

For safety stock based on demand distributions, the above mentioned demand varia-
tions are taken into consideration by using normal distribution (Jonsson and Matts-
son, 2009). The normal distribution is defined by a mean value and standard devi-
ation which gives a symmetric distribution. Since these two factors are taken into
account, safety stock can be calculated based on demand data, average demand dur-
ing the lead time and standard deviation of demand during this lead time (Jonsson
and Mattsson, 2009). There are possibilities that the lead times can also vary, and
similar to demand, the lead time has a normal distribution which can be taken into
consideration. Hence, the required safety stock and based on it the required service
level can be much better controlled using this method.

To further enhance the reliability on the above method, ‘service factor’ (Z) is used
as a multiplier along with the standard deviation, as shown in the formula below
(Constantin, 2016).

s
LT 2
Saf ety stock = Z ∗ ( ∗ σD ) + (σLT ∗ Davg )2
T

Where,
Z is the service factor
LT is the maximum lead time
D avg is average demand
σD is the standard deviation of demand
σLT is the standard deviation of lead time
T is the total time taken for calculation (1 day)

This is done to calculate the specific required quantity which relates to the de-
sired service level. Constantin (2016) shows in table 1, the relation between the
service level in percentage and the corresponding service factor.

It can be seen that it is non linear in nature, which means that, the higher the
service level, the safety stock level will get exponentially higher. Therefore, instead
of having a fixed service factor for all products, it can be assigned on basis of the
ABC analysis of the products as mentioned earlier. For example, if the ‘A class
items’ are assigned to have service level of 96%, then based on table 1, a service
factor of 1.75 will be used in the calculations for the safety stock. Once the safety
stock is determined, it should be reviewed in regular time intervals to check if it is,
what is needed or if any adjustments is required or performing root cause analysis
if any problems occur, etc.

14
 2. Theoretical framework

Table 2.1: Relation between desired service level and service factor (Constantin,
2016)

Desired service level Service factor

50% 0.0
60% 0.3
70% 0.5
80% 0.8
85% 1.0
90% 1.3
93% 1.5
95% 1.6
97% 1.9
98% 2.1
99% 2.3
99.90% 3.1

2.7 Kanban systems

As explained in the previous section, the authorization of movement of materials is
initiated when the operation, consuming a unit of the material, more or less gives a
production signal to the preceding operation. The degree to which this is possible
depends on the size of the inventory which is used for various reasons such as lead
times are too long for direct delivery, high set up times, ordering costs are too high,
etc. This is one of the main reasons that an inventory buffer needs to be kept to
cover for these direct needs. These type of methods are called direct call offs meth-
ods as, in this material planning method, ordering takes place directly from the
consuming operation to the supplying operation which is same as the reorder point
system (Jonsson and Mattsson, 2009). Another variant of this reorder point system
is called Kanban system, and the physical form of this is called the two bin system.
The word ‘Kanban’ means card or sign in Japanese, which is a method for authoriz-
ing movement of materials based on this visual signals (Jonsson and Mattsson, 2009).

In the two bin variant of the reorder point system, as the name suggests, there
are two bins. One of which is a bigger bin and the other one is a smaller bin. From
the bigger bin, withdrawals of material takes place and from the second bin a re-
plenishment amount is withdrawn which is equal to the corresponding consumption
unit from the larger bin plus the safety stock. The consumption from the second bin
continues until the quantity is replenished in the first bin. This is again an example
of pull system where the downstream consumption of material initiates replenish-
ment order to the upstream process (Jonsson and Mattsson, 2009).

As explained in the previous section, the reorder point is calculated by adding safety
stocks and demand during lead time. This lead time is the total lead time and can
be split into two, mainly production lead time and moving lead time (Jonsson and
Mattsson, 2009). The production lead time is the time taken for the actual man-

15
2. Theoretical framework

ufacturing process on the material such as set up time, changeover time, run time,
waiting time, maintenance time, etc. The moving lead time is the time it takes to
send and order upstream from the consuming operation and also the time required
to physically moving the material from the upstream operation to the downstream
operation that initiated the replenishment order. Similarly, production Kanban is a
visual signal for a manufacturing unit to start the operation and a move Kanban is
a visual signal to move the materials from upstream to downstream replenishment
manufacturing unit.

In a pull system, there is high repetitiveness and a specific item is produced and
moved around frequently. But it does not mean that there is high volume pro-
duction (Jonsson and Mattsson, 2009). It might be the case where there are large
variants of products produced and some of them are infrequently produced, like in
a make to order or assemble to order manufacturing environment. In this case it
is very challenging to implement the traditional Kanban system mentioned above;
instead a generic Kanban or CONWIP card (explained further in the next section)
is issued (Jonsson and Mattsson, 2009). Generic Kanban is the same as the tradi-
tional Kanban system, which is used as a visual authorization to start production
to the upstream process; but with the only difference being that it does not specify
which particular product to produce. Instead it specifies by the product family and
not about the exact model or variant in it. The information regarding the specific
product, with its corresponding configuration and processes, and the sequence of
products, has to be generated and sent out separately, which is based on historical
consumption. This can, therefore, work in a mix model production environment.
Also, there is a mix of pull and push systems, in the sense that the pull mechanism of
Kanban and push mechanism of sending out the product plan is both implemented
in this generic Kanban method (Jonsson and Mattsson, 2009).

2.8 CONWIP system

CONWIP stands for constant work in process. A CONWIP system is the same
as generic Kanban system wherein the signal for production is sent upstream by
the consuming downstream process, without any specifications of the product, but
only the product family (Jonsson and Mattsson, 2009). The only difference being
that, in a generic Kanban system the upstream process receives order signals from
the successive downstream process, whereas in a CONWIP system the last process
sends order signals to the very first process in the production flow. As shown in the
fig. 1, this process, thus, secures a constant work in process and can handle mixed
model variations as well (Jonsson and Mattsson, 2009).

A CONWIP card can be used instead of a Kanban card and the last operation in
the process will send it to the first process in the operation for authorization of
production (Jonsson and Mattsson, 2009). The CONWIP card needs to be sup-
ported with a list of product specifications, configurations and sequence, which is
based on the future expected demand compared to the historical consumption used
in the case of Kanban systems (Jonsson and Mattsson, 2009). Now, the sequence of

16
 2. Theoretical framework

Figure 2.2: Kanban system vs CONWIP system (Jonsson and Mattsson, 2009)

operations is decided on the basis of first come first served and when the CONWIP
card is received at the first operation, it is produced and pushed through the next
process and so on. Also, in this system it allows to have mix model variations to
be included in the processes in between as it secures a constant work in process.
The main reason is that the signal is received from the last operations to the first,
there can be various individual operations included in between; which is the main
difference between a CONWIP and Kanban system (Jonsson and Mattsson, 2009).
The only aspect to consider in this case is the sequence of the received order should
be followed, even if there are various individual operations, as the items are pushed
through from first operation till the end.

2.9 Value stream mapping

Value stream is mapping of all the processes, irrespective of whether it’s value adding
or non value adding (Rother and Shook, 2009). It relates to the flow of products from
raw material stage to the final customer, based on customer demand (Abdulmalek
and Rajgopal, 2007). Rother and Shook (2009), further mentions the importance
of creating a value stream map as it is an essential tool which helps to visualize
the various processes in the production system, for example, casting, sand blasting,

17
2. Theoretical framework

machining, assembly, etc. The path of a product is followed from the start till the
end i.e. from the raw material stage to the finished product reaching the customer,
recording all the value adding activities and the non value adding activities (Liker
and Hoseus, 2007). As mentioned by Rother and Shook (2009), if the components
for the chosen product comes from many sources, the component which is having the
maximum lead time should be taken for value stream mapping first. Furthermore,
it helps in decision making, in the sense that, it helps in visualizing the effects of a
decision over other processes connected to it. Having a broad view like this, helps
to avoid cherry pick situations that seem right at a particular moment.

Within the production flow, apart from the material flows explained above, the
information flow is also of equal importance. The linkage between the two can also
be visualized through a value stream map. As the name suggests, the information
flow tells each process as to what to do next. The main idea is to have a flow of
information such that one process makes exactly what is needed by the next one,
nothing more and nothing less (Rother and Shook, 2009).

The initial step is to define a product family for the value stream mapping (Rother
and Shook, 2009). This is essential as otherwise it would be highly complicated
and difficult to map the flows of all product families of a manufacturing plant. It
is also necessary from the customer point of view, as they care only about certain
products and not the entire range of products that is manufactured in a plant. A
product family is a group of products going through similar processes and machine
equipments in the downstream processes. Establishing a product family from the
upstream processes might lead to grouping of many products, for example, they
might be produced in batch mode in fabrication, casting, or similar processes. A
product family must be clearly selected, the different finished part numbers in that
selected family must be defined, the quantity required by the customer, the fre-
quency of the requirement, etc should all be decided in the early stages of value
stream mapping. (Rother and Shook, 2009)

The next step is to continuously improve the identified issues in the value stream
map. This process of continuous improvements is known as Kaizen. There are es-
sentially two types that is flow level Kaizen and process level Kaizen (Rother and
Shook, 2009). Flow level Kaizen refers to the continuous improvements done for
the entire value stream, trying to create value from the flow. Process level Kaizens
refer to the continuous improvements done at the shop floor level by eliminating or
reducing waste. Flow level Kaizens are usually done at a senior management level
as it focuses more on material and information flow. Process level Kaizens are done
at shop floor level which means more focus on processes and people.

18
 2. Theoretical framework

As companies are usually organized in terms of functions and departments, it is

difficult for a person from any department to focus on the entire value stream. This
is mainly because; they tend to look more into their own functions for identifying
improvements and thereby leading to process level Kaizens. This is good in a way,
but looking at a bigger picture, for the development and improvement of the entire
flow, one should focus more on flow level Kaizen (Rother and Shook, 2009).

19
2. Theoretical framework

20
 3
Methodology

This section contains information about the methodology by which this master thesis
was performed. This section includes a short description about the strategy, data
collection for the value stream map and finally about the study and analysis.

3.1 Overall research strategy

As shown in fig. 3.1 the thesis study was followed in 7 stages. As mentioned earlier,
the current problem of the company was to meet the on time delivery (OTD) require-
ments. Once the focus for the study was decided as exploring solutions for meeting
OTD requirements, the related background study was done. Field observations and
collecting information, through various literatures, was done to understand the basic
factors related to OTD. For the literature framework, more specific articles related
to OTD and factors affecting OTD were collected. In relation to the literature, the
required qualitative and quantitative data was collected. Using the literature and
the data available the current state map was drawn. This map was analyzed and
redrawn a few times for the future state map. Based on the future state, recom-
mendations related to system level and other improvements were proposed to the
company, to improve OTD.

Figure 3.1: Overall research strategy for the thesis project

3.2 Background study

As mentioned earlier, the background study was done to understand the issues
related to OTD. In this, the shop floor processes of various departments in the man-
ufacturing plant were observed. The observations were done directly by the authors
while going through the processes and taking down notes. Furthermore, information
was also collected by interacting with the shop floor team leaders, supervisors and

21
3. Methodology

workers. The main aim here was to understand the process flow. The next step was
to get to know about the information flow and the ways in which the production
was controlled. This was done by having some initial unstructured interviews with
the managers and department heads. Throughout this process, various literature
relating to OTD were referred to and accordingly the background study was carried
out. Once the process flow and the information flow was known, the next step was
to have in depth literature study.

3.3 Literature study

The main purpose of the literature study was to provide an academic framework
and base for the thesis project. Continuing from the background study, literature
relating to factors affecting OTD were taken into consideration. It was found that
literature relating to the improvement of On time delivery (OTD) performance,
for in-house production, was scarce. However, from the available literature, the
stakeholders involved and from the academic knowledge of the authors, the scope
for literature study was narrowed down to lead time decrements and optimization
of buffers. The motivations for selecting these two factors were explained previously
in the literature framework. The next step was to find literature relating to lead
time and buffers and the way in which it can improve OTD. Furthermore, literature
regarding control and execution of the production system was included as it further
supports improvement of OTD. These articles were collected from various databases
and sources that include Science direct, Web of science,IEEE, Scopus and chalmers
library.

3.4 Data collection and validation

All the relevant data for the thesis work was collected from the company’s ERP
(enterprise resource planning) system, through interviews and also through direct
observations. The data collection is divided into two parts, that is collection of
qualitative data and the collection of quantitative data.

3.4.1 Qualitative data

The main aim of the qualitative data collection was to collect all relevant data and
valid information about the problems in the production flow of product SK, in re-
lation to OTD. Qualitative data was collected via 6 interviews from the employees
at the company. Interview time was one hour per interview and the questionnaires
were formulated with the help of the literature framework. Interviews were con-
ducted with the supply planners, industrial engineers, team leaders and with the
procurement head of the company. The motivation for selecting supply planners
was that they have direct influence over the control and execution of information
flow of product SK. They are responsible for placing the orders, taking into con-
sideration the capacity and demand. The reason to interview industrial engineers
was that they are responsible for the production flow of SK. Furthermore, the team

22
 3. Methodology

leaders were also interviewed as they work closely with the production flow. The
rubber and plastic components are procured from the external suppliers, the authors
interviewed the procurement head in order to understand the impact of the supplier
lead time on the production flow.

All the interviews were semi structured and the questionnaires were sent to the
respective people well before the interview. The focus of these interviews was to
collect as much information as possible regarding the production and information
flow of product SK. The aim was to narrow down and to find the factors affecting
OTD through these interviews. The questionnaires are shown in the appendix A.

Since all the interviews were semi structured, the interviewees explained about the
problems and its consequences in their perspective. But at the same time, the inter-
views gave the authors insights about the root cause of the problem which in turn
helped us a lot in the analysis part.

3.4.2 Quantitative data

The main idea behind the collection of quantitative data is to collect all the empirical
data pertaining to the production flow of product SK. The current state map should
be a snapshot of the existing production system (Rother and Shook, 2009). The data
is collected by walking along the production flow of product SK. The data relating
to few inventory levels is collected as an aggregate of all the snapshots taken from
ERP system for one particular year (2017). Snap shots for those inventories are
based on the average time spent by the products in that particular inventory. Few
empirical data relating to the forecast and the past sales data are collected from the
planning department.

3.4.3 Data validation

Sometimes the data available in the ERP systems might not be up to date, so manual
collection and verification was necessary (Rother and Shook, 2009). After collecting
all the required quantitative data of the current state value stream map, we validated
the data by checking with the people who are involved in the production of SK. This
validation ensures the reliability of the data collected through the field observation
and from the ERP systems.

3.5 Current state map

Plant level Value stream mapping was done for product SK from the raw material
to the final warehouse. The authors used value stream for visualizing the current
production and also to visualize the future state with the proposed improvements.
Data for the current state map includes the data pertaining to both the information
and material flow (Rother and Shook, 2009).

23
3. Methodology

Current state map was drawn with the help of the manually collected qualitative,
quantitative data and from the ERP system. Usually when the components for the
selected product comes from multiple ways, the flow which comprises of maximum
lead time is taken into account for drawing the value stream mapping and analysis
(Rother and Shook, 2009). For product SK, since the body part is having maximum
lead time, it is considered for drawing the current state map. The data for current
state map is collected by walking through the production line and by asking ques-
tions to the people who are involved in the production (Rother and Shook, 2009).
But since it is a batch production, the production is not continuous and the pro-
duction for product SK is performed on different days, according to the production
schedule, in different departments. Therefore, the data was collected on different
days.

The data collection for the current state map was collected from the warehouse
to the starting point of the production flow. During data collection phase, current
state maps were drawn in an A3 paper by hand first. Then after collecting all the
empirical data for the current state map, it was validated with the people in the pro-
duction flow. Then again the map was further analyzed for improvements and the
above mentioned procedure was repeated, Finally after number of iterative maps,
the final current state map was digitized.

3.6 Approach for Analysis

Figure 3.2: Approach used for analysis

24
 3. Methodology

Rother and Shook (2009) state that the value stream is done to identify and eliminate
sources of waste by implementing a future state value stream. Rother and Shook
(2009) further state that one of the ways to draw the future state is through a list
of questions mentioned in section 3.7. Therefore, after drawing the current state
map, the approach for analyzing the current state map was to follow the steps as
shown in fig. 3.2. The problems were identified using the future state questions,
which will be explained in the next section. Once the problems were identified, the
possible related solutions were found using literature. These solutions were then
discussed with the stakeholders for their suggestions and approval. The focus was
also to consider the feasibility of the solution in relation to OTD and as per the
availability of the resources. Once a solution was decided, the entire process was
repeated and then the final future state map was drawn and digitized.

3.7 Future state map

Future state map is drawn with the aim of eliminating the wastes and also to pro-
duce according to the customer requirements in future (Rother and Shook, 2009).
Future state map was drawn after the analysis of the current state with the scien-
tific data collected from the literature study. Analysis was done with the aim of
producing product SK according to the customer demand and also to deliver it on
time by eliminating the wastes such as excess lead time and excess inventories in
the production flow of SK. Future state map was drawn by consulting with a team
of Industrial engineers, supply chain planners and the lean manager and also by
answering the following questions from Rother and Shook (2009). The procedure
followed for drawing the future state map is described in the previous section 3.6.
The analysis of the following questions is explained in section 5.2.

1. What is the real customer demand?

This question is really important for the company to manufacture according to
customer demand. Producing to real demand within the available time helps
in eliminating most of the wastes in the production. Knowing the present
demand and altering the production rate accordingly ensures on time delivery
and reliability of the production flow.

2. Will the produced goods be delivered directly to the customer or

produced to final warehouse? From where the customer pulls the
finished goods?
Many factors needs to be considered before answering this question. How-
ever company’s on time delivery strategy is one among the main factor to be
considered. The other factors such as demand variation, ability to react to
the demand and the production lead time are need to be considered while
answering this question.

25
3. Methodology

3. Where can the continuous production flow be implemented?

This question is about having the continuous production flow without buffers
in between the processes. But as mentioned earlier, if buffers are needed, then
it should be optimized so as to have continuous production flow. The pro-
duction rate of different processes and the layout needs to considered as well.
Order replenishment systems and the order releasing and the receiving points,
should also be considered.

4. Where can the supermarkets and buffers can be placed in the flow?
This question is about the need for having the supermarkets and buffers in
the production flow. Placing the supermarket in the production flow helps in
regulating the complete flow downstream. Factors such as lead time, ability to
deliver it on time, delivery policies, safety stock and reorder strategies needs
to be decided while answering this question.

5. Where should the pacemaker be placed to control the production

flow?
This question describes about having the control point in the production with
which the production control will place the order based on the current demand.
The order for the upstream processes will be placed from the pacemaker pro-
cess. Many factors such as type of production (batch, line etc.), ability of the
production system to reflect to the demand and traceability needs to be con-
sidered. The answer for this question also explains about the level of inventory
control and the safety stocks needed as the demand is seasonal.

6. What are the process improvements needed?

This question is about the process improvements that are needed for imple-
menting the future state map. Answer for this question justifies the feasibility
of achieving the future state and also allows the value stream team of the
company to make implementation plans.

Number of iterations for improvements, were done in the current state map
to create the future state map, and for which the above process was followed.
For example, when a future state map was drawn, the improvements were
discussed with the team in terms of its feasibility and practical implications.
The feedback from the team was noted down and was used to create a new
future state map. Similarly, the future state map in this report has been
finalized by the continuous improvements and feedback from the team.

26
 4
Current production

This section contains information about the current production state of product SK.
This section includes the brief explanation about the product itself and the current
production and information flow along with the current state map.

The case company started to implement pull systems for material replenishment
in order to produce to the customer demand. Kanban is suitable with only smaller
fluctuations in demand (Jonsson and Mattsson, 2009). Since the demand and the
production lead time are fluctuating, the company needs its production system to
react to the customer demands quickly in order to ensure the higher on time delivery
rates. For implementing the pull system, they started using kanban cards between
the production processes. However, because of the unpredictable lead times and the
lack of frozen horizon for production orders, using kanban cards for material replen-
ishment did not give proper results in the initial testing stage itself. One another
reason for that is the lack of modern electronic tools for material replenishment and
traceability. Currently, the material replenishment is done by 3 planners, who have
to, sometimes, physically go to the shop floor, to check the material in the inventory.
They have to check in the ERP system as well and verify if the data matches with
what is actually available at the shop floor. Since the departments are not connected
to each other, the departments behave as isolated islands resulting in piling up of
inventory in many places.

4.1 Current production flow

As mentioned earlier, product SK comprises of more than 30 components which
comes from three different sources. All the metal parts are manufactured in-house,
whereas the rubber and the plastic components are procured directly from external
suppliers and supplied to the various parts of the production flow by the internal
logistics department. The metal parts for product SK comes from two different
sources.

The body part of SK which has the maximum production lead time is die casted
in-house in the facility at Ljung. All other metal parts are manufactured in the bar
machining area. The parts for the bar machining area are manufactured by machin-
ing bars which are procured from the suppliers. The parts from the bar machining
area are sent to the 4 sub assembly stations before being used in the final assembly.
The parts from the bar machining area are manufactured in large numbers, in a

27
4. Current production

batch size of approximately 12000. This is mainly because of the large setup and
changeover times of the machines at the bar machining facility. Another reason for
such huge batch size is that the components from bar machining area are relatively
small and the cost for storing such small parts are negligible when comparing to the
body part.

If the components for the chosen product comes from many sources, the compo-
nent which is having the maximum lead time should be taken for current stream
value stream mapping (Rother and Shook, 2009). The body part for the product
SK has the maximum production lead time, so the production flow of body part is
followed and the value stream for the body part is drawn.

Figure 4.1 shows the process flow map followed by the body part. The process
flow of Body part of product SK starts in the foundry where it is been die casted
and sent to the sand blasting process. The simplified layout of the path travelled
by the body part is shown in fig. 4.2. The sand blasting machine is also located
in the foundry area. After sand blasting the casted body parts are sent to either
the Brother machines or the Riello based on the planners decision and machine
availability. The Riello machine has a longer lead time, compared to the Brother
machines, and reasons for that is explained later. After machining, either from the
Riello or the brother machine, the part is washed as shown by process ‘washing 2’.
After washing, the machined body part is stored near the final assembly until the
next order for final assembly of product SK is placed. In the final assembly the body
part is assembled with the other parts and the final finished product SK is stored
in the warehouse.

Figure 4.1: Process flow map for product SK

As shown in the current state map, in figure 4.3, the production flow of the body
part starts from die casting, in machine number 15. The molten metal is fed into
the casting machine with the help of the forklifts from the melting furnace when it

28
 4. Current production

Figure 4.2: Layout of the production system

is needed. This machine is dedicated for the product family AF. However, during
seasons of high demand, few varieties of AF will be casted in machine number 24 for
satisfying the demand. However as far as the product SK is considered it has been
casted in machine number 15. The quality of the product SK is very good compared
to the other variants of AF. The casting machines are running all the three shifts
for satisfying the demand.
After die casting, the products are collected in the pallet and it is sent to the sand
blasting process in a batch of 600 products. The sand blasting machine is also lo-
cated inside the foundry department. Even Though the setup time and the cycle
times are relatively small, since the sand blasting machine is not dedicated for AF
family, the batches have to wait in the queue for more than half a day to get sand
blasted.

After sandblasting process the parts are sent out of the foundry and stored in the
inventory before the machining process. From the inventory it is routed to two ma-
chines that is to the Riello or to the brother machines based on the availability.

29
4. Current production

The reason for this is the availability of the Riello. Even though the Riello is
really fast, it is used by more than 70 product varieties. One more reason is the
complexity of the machining process steps. In the Riello, the machining is done in
two steps, that is operation 30 and operation 50. Once the casted body is sent into
the Riello, operation 30 is carried out and after that a batch of 200 is collected and
washed in the washing machine near the Riello machine and it is stored temporar-
ily near the Riello machine till the complete batch is machined and washed in the
washing machine. Then the entire batch is sent again to Riello and the operation
50 is carried out and the products are washed again and stored in the supermarket
before final assembly. Even though the machining operation in Riello is really fast,
as two operations are carried out in the same machine, the parts after finishing the
operation 30 have to wait for approximately 1.3 days for carrying out its finishing
operation that is operation 50.

The Brother machines only need one operation and the finished products are just
washed once. One of the major drawback was the high cycle time. The capacity of
the brother machines is approximately 32 products per hour whereas Riello is able to
machine nearly 200 products per hour but it has a waiting time of 1.3 days between
its two operations. Recently, Industrial engineers at the company have worked on
the improvement of cycle time of the Brother machines and it is now in the testing
phase. Once the change has been implemented completely, the new capacity of the
machines will be 60 products per hour with a set up time of 2 hours. The engineers
claim that, after the implementation of the improvements, the Brother machines can
accommodate all the product variants of AF to be machined in brother machines
without being machined in the Riello. Machined parts from both the Riello and
the Brother machines, are washed and randomly checked for quality and stored in
the inventory before the final assembly. Since the quality for product SK is really
good, it is stored directly in the inventory before final assembly. For the other prod-
uct variants of AF, its been sent to the visual inspection area and collected in the
supermarket before final assembly.

30
 4. Current production

Figure 4.3: Current state value stream map of product SK

31
4. Current production

After machining and washing, the products are consumed in the final assembly
whenever it is needed. Since the pallet in the warehouse can accommodate only 480
parts per pallet, the batch size in the final assembly is 480. The assembly line is
automatic expect for the first station, which needs human resource for feeding the
valves into the assembly line and also for inserting few parts into the body initially
before passing to the assembly station 2. After finishing the assembly, the valves are
tested and laser marked automatically. After laser marking, the valves are sent to
the packing station. The faulty valves in the testing are collected separately and are
repaired near the assembly station manually. The valves which have good quality
are packed by the packing operator and sent to the final warehouse through the
automatic conveyors. The packed valves are collected at the receiving station in the
warehouse, where they are scanned for feeding the information in the ERP system
and then stored in the required place in the warehouse. This helps in easy retrieval
of valves when the customer order is placed.

4.2 Current information flow

Currently all the processes in the company are operating as isolated islands. There
are three planners who decide the amount of parts to be produced under each de-
partment. The production orders are sent to the corresponding departments every
week. Limited amount of time is considered as a frozen horizon and the produc-
tion orders are changed in the last minute before final production because of many
reasons. The information is sent directly to the department heads and also to the
team leads by emails and by informing them via phone calls. The planners check
the inventory information via ERP systems. Even though the ERP systems are
updated correctly there is still the prevalence of mismatch between the inventory
values in the ERP system and the physical presence. The consequences include both
the waiting times and also over production resulting in inventories in many places.
The information flow is shown in the current state map (figure 4.3).

As mentioned before, there are three planners involved in scheduling the processes.
One planner is responsible for the production scheduling of die casting, sand blast-
ing and machining processes. The second planner is responsible for the production
scheduling of sub assembly and the final assembly and the third planner is responsi-
ble for bar machining area. All the planners place the production orders to respective
departments with the help of the forecast from the sales and operation plans from
the top management. They also keep the track of the inventories in their individual
departments separately. Currently, the Kanban cards for the production are con-
trolled by the respective planners and there is a prevalence of mismatch between
the production schedule and the quantity ordered from kanban. The production is
carried out on the basis of the sales and operation plans and the planners release
the kanban cards accordingly.

32
 4. Current production

The traceability is one of the major issues in the flow. There is one single production
order from the foundry till the final machining process. The production orders are
updated in the ERP system only when it is finished. Since the process operations
are based on the different production control planners, it gives the possibility for
traceability issues, which is explained in the following sections.

33
4. Current production

34
 5
Analysis

This section of the report explains the analysis of the current production of product
SK along with the future state map.

5.1 Current state map

Efficient production flow can be achieved by improving the throughput time (Gabriel
et. al. 1991). Excess inventory in the production is one major contributor for long
lead times (Rother and Shook, 2009). The current state map shown in figure 4.3
clearly visualizes the amount of time spent on each process and the inventory in
between the processes. The average production lead time or the throughput time
is 9.5 days. The production flow starts from the die casting, ends in the final ware-
house and after that the products are shipped to various customers based on their
need.

Even though the production lead time is more than a week, the value added time for
the product is only 140 seconds. The major part of the production lead time is from
the inventories in various part of the production flow of SK. From the current state
map, as shown in figure. 4.3, there are 5 major inventories in the production flow
which includes a buffer between die casting and sand blasting (0.8 day), between
sand blasting and both the machining processes (3.2 days), between the two opera-
tion step when the machining is done in Riello (1.3 days) , between the washing and
final assembly (0.7 days) and finally finished product inventory (1 day). Table 5.1
Shows the details about the value added time and the time spent in the inventory.

Table 5.1: Values for the current state production

Total time Value added time number of buffer locations Time spent in buffers
9.5 days 140 seconds 5 9.5 days

The maximum quantity is in between the sandblasting and the machining process.
From this buffer, casted and sandblasted body parts are sent to both the Brother
and the Riello machines based on the availability. Currently, machining the body of
SK in the Riello takes longer time in the flow since it has two separate operations
that is done in the same machine. The parts which have finished the first operation
are kept in the inventory for approximately 1.3 days and later on it is sent for the

35
5. Analysis

second machining operation. Even though the cycle time for both the machining
operations and the washing operations are very low, because of the waiting time
of 1.3 days, it makes it lengthy. Whereas, since more than 70 product variants
are machined in the Riello machine, availability for machining the product SK is
also a main reason for the inventory of around 3.2 days before the machining process.

On the other hand, the Brother machines are in the testing stage for the updated
double chuck mechanism, to hold and machine two products at a time. This is
another reason for the inventory of 3.2 days after the sand blasting process. The
industrial engineers at the company explained that after the implementation of the
new improvement, the brother machine will be able handle all the products of AF
and it will remain as a dedicated machine for the AF product family. One main
advantage of the Brother machine is that all the machining operations are done in
a single step. This helps in eliminating the waiting time, as compared to the Riello.

After the machining and washing, the machined body parts lie in the the inventory
for approximately 0.7 days and this is because of the production planning made for
the final assembly.

5.2 Future state map

As mentioned earlier in the problem definition, currently the on time delivery for
the product SK varies from 65% to 93% with the average of 84% for the past year.
The main aim while formulating the future state map was achieving on time delivery
rate of 96% by reducing production lead time and by optimizing the inventories in
the flow. This part of the report answers the questions mentioned in the methodol-
ogy (Section 3.7). This part of the report also examines the other factors relating
to feasibility issues before the future state is being proposed. The demand for the
product family AF which passes through the same production flow is 260260 for this
year. The available production time is 100 hours per week as most of the processes
are running 3 shifts per day, the takt time is 70 seconds per product. However, the
yearly demand for the product SK is around 87500 and the average daily demand
for the product SK is 350 per day as the total number of working days for a year in
the company is 250.

As mentioned in the current production (section 4), since the company’s deliv-
ery strategy is make to stock, the products are stored in the final warehouse and
then supplied to the customers in various parts of the Nordic countries, Germany
and Poland. Even though the average daily demand for the product SK is 350, the
standard deviation of demand per day is 84. So, during high seasons, the company
has a demand of approximately 436 products per day and on the other hand in the
low seasons the demand for the product falls to 266 products per day. So in order to
supply during high seasons, the company is following the make to stock, customer
order decoupling strategy. Wouters (1991), states that the reduction in the lead
time will have positive impact on the delivery lead time. So reducing the lead time
is very much essential for achieving higher on time delivery rates.

36
 5. Analysis

According to (Rother and Shook, 2009), inventories should be placed in the pro-
duction flow for the following reasons : a place where the need for the upstream
process to operate in the batch mode, difference in cycle time between the processes
in the value stream and the uncertainties in the lead time. Based on the steps
explained in section 3.7, in the future state value stream map of product SK, the
inventory is placed in two places. This is mainly because of the batch production
processes and also uncertainties in the lead time. New buffer placements are shown
in the future state map in figure 5.1.

37
5. Analysis

Figure 5.1: Future state map

38
 5. Analysis

Coyle, et. al., 2003, argue that for the company with both demand and lead time
variations, reorder point system with fixed quantity and variable time is the best
suited material replenishment system over other types like Kanban. Constantin
(2016), states that having optimized safety stock and inventory levels is essential for
achieving higher delivery rates with low tied up capital. Therefore, having unnec-
essary inventory will lead to poor delivery performance. On the other hand, many
companies have higher safety stocks for achieving higher service levels to the cus-
tomers. However, optimization of safety stocks helps in achieving high service levels
and having low tied up capital as well (Constantin, 2016). So in the future state
map, safety stock is placed in both the buffers in the flow. This safety stock is to
provide higher service levels for the delivery.

Safety stock at the specified service levels can be calculated efficiently with the
help of the standard deviation (Constantin, 2016). According to Constantin (2016),
since both the lead time and the demand have uncertainties, formula used for cal-
culating the safety stock is

s
LT 2
Saf ety stock = Z ∗ ( ∗ σD ) + (σLT ∗ Davg )2
T

Where,
Z is the service factor
LT is the maximum lead time
D avg is average demand
σD is the standard deviation of demand
σLT is the standard deviation of lead time
T is the total time taken for calculation (1 day)

Constantin (2016), states that the optimum service level for the fast runners should
be 96-98%. Here for calculating the safety stock, in both the places, the service
level, as required by the management is 96%. The safety stock in the buffer before
final assembly is 1273 products and in the final warehouse is 1502 products. Table
5.3 shows the numbers used for calculating the above mentioned safety stock values.

Table 5.2: Values used for calculating safety stock

Terms in formula Supermarket before final assembly Warehouse
LT - maximum lead time till that point 5 days 7.7 days
D avg - average demand 350 / day 350 / day
σD - standard deviation of demand 350 / day 84 products / day
σLT - standard deviation of lead time 2.0 2.4
T - total time for calculation 1 day 1 day

Reorder point (R.O.P) is the sum of the safety stock and the lead time demand,
whereas the lead time demand is the product of the demand per unit time and the
average lead time (Jonsson and Mattsson, 2009).

39
5. Analysis

R.O.P = Safety stock + (D*L) Where,

D is the demand per unit time
L is the lead time

The reorder points for warehouse is 4198 and the reorder point before final assembly
is 3023. Table 5.3 shows the reorder point values in the two buffers.

Table 5.3: Safety stock and reorder point values

Lead time in demand Safety stock Reorder point
Warehouse 2695 1503 4198
Before final assembly 1750 1273 3023

Rother and Shook (2009), argue that placement of pacemaker determines the level
of capacity changes in the upstream processes and, also, after the pacemaker, all the
downstream processes, should be continuous. Therefore, in the future state map,
the pacemaker is placed before the final assembly. The pacemaker controls the up-
stream processes starting from the die casting process. Once the inventory at the
pacemaker falls below the reorder point, it sends a signal to die casting process.
After die casting, the products are pushed in the subsequent downstream processes
i.e. through sand blasting, machining, washing and finally to the inventory at the
pacemaker. Constant work in process (CONWIP) ensures the constant workload in
the process in between and also it allows the production flow for a better produc-
tion mix the CONWIP loop (Jonsson and Mattsson, 2009). Since the AF product
family has to be produced in the same machines, having CONWIP with reorder
point system will hold good for IMI. In the same way if the inventory in the final
warehouse runs below the reorder point, the products are ordered from the final
assembly. Final assembly pulls the required amount of parts from the supermarket
before the final assembly. This is shown in the future state map in the figure 5.1.
The proposed total production lead time in the future state map is around 2.17 days
and with a total safety stock of 2776 products. With this safety stock and with the
average daily demand of 350 products, the company can sustain up to 7.5 days with
only the final assembly alone. Table 5.4 shows the values of the proposed future
state map.

Table 5.4: Values for proposed future state map

Total time Value added time Number of buffer locations Time spent in buffer
2.17 days 164 secs 2 2.17 days

40
 5. Analysis

However, the total production lead time is based on the safety stock and value added
time. The products which are present in the super markets except the safety stock
is not taken into consideration as it varies with time in the production. Even after
considering that, both the buffers contains parts upto the reorder point level (total
7221 products), the throughput time is around 5.8 days which is lower than the
current production lead time of 9.5 days.

The total value adding time is 164 seconds, this slight increase is because of the
usage of brother machines for the above mentioned reasons. However industrial en-
gineers at the company are continuously working for reducing the cycle time and
set-up time further.

41
5. Analysis

42
 6
Recommendations for the
company

This section contains the recommendations for the company to achieve the proposed
future state. The first subsection explains about the process level improvements
needed for the implementation of the proposed future state. The following subsection
describes about the system level improvement suggestions.

6.1 Improvements for implementation of the pro-

posed future state
For the implementation of the proposed future state and also for further tuning of
the production flow of product SK, it calls for several process level improvements.
After the analysis of the current state the authors came up with the suggestions for
the future improvements. The improvements include:

1. Introduction of pacemaker and CONWIP loops along with fixed quantity ir-
regular interval reorder point systems in the production flow of product SK,
as mentioned in section 5.1. The quantity is fixed and the time is variable
in the sense that, if the quantity in the buffer falls below reorder point, only
then the production orders are initiated according to the demand. Hence, it
is not initiated on a timely basis but at irregular intervals. This introduction
of pacemaker and CONWIP along with reorder point system loops helps in
ensuring the control over the material flow and also reduces the chance of in-
ventories getting stacked in the flow, except the planned super markets.

2. In the production flow of SK, all the machines in the production flow are dedi-
cated to handle the product family AF except the sand blasting machine which
handles the products from various product families. Currently this causes the
casted products from the die casting area to wait in the queue, before sand
blasting machine, which is operated on a first come first serve basis. This
causes the waiting time of 0.8 days in the production flow. The company
should work on prioritizing their faster runners over the other products pass-
ing through the sand blasting machine. Simple way is to have two parallel
waiting lanes before the sand blasting machine, that is one for fast runners
(high priority line) and the other line for normal products (low priority line)

43
6. Recommendations for the company

that are passing through the sand blasting machine. By doing this the waiting
time of 0.8 days can be reduced to great extents.

3. In the production flow of product SK, the Brother machine is having the maxi-
mum setup time and cycle time compared to the other processes in the flow. In
the production flow of SK, sand blasting and packing are considered as a sub-
sidiary processes for the die casting, machining and final assembly processes.
In this case die casting, machining and final assembly are the main processes
in the flow. Out of the three main processes brother machine is having the
highest cycle time and set up times. In long run there chances of accumula-
tion of products before the brother machines. However, the future state map
(figure 5.1) shows only the processing lead time, the non value adding times
and support times such as transportation time between the processes are not
considered in this thesis. The transportation times between the sandblasting
and the machining processes is comparatively high, since the distance between
the departments is far and the products have to be transported by the inter-
nal logistics personnel. This must be considered during the cycle time analysis
process, since this might become a reason for the accumulation of sand blasted
parts before machining process.

4. Leveling of load is one option to keep the low inventory level in the supermar-
kets. The leveling of the loads can be done by placing the production orders
based on the demand and also based on the present conditions in the inventory
for the products of the AF product family. This helps in preventing piling up
of inventories in the flow for particular product variants when they are not in
demand.

5. All the other products in the production flow of AF product family, except
product variant SK have to pass through the visual inspection before storing it
in the supermarket, near the final assembly. Even though this is not included
in the thesis scope, we would like to point out that, the visual inspection of
these other products, might cause delay in the production lead time of prod-
uct SK. This might affect the CONWIP loop when the production schedule
of product SK comes after the production schedule of other products in the
product family AF. This also causes transportation waste since the products
from washing is transferred to the visual inspection area first and then it is
transported back to the supermarket near final assembly. One way to avoid
separate visual inspection is to have the visual inspection near final assembly.
This might help in maintaining the lead time for all the product variants of the
product family AF and also it helps in eliminating the transportation wastes.

6. For increasing the flexibility in the production and for having the best suited
production mix, the production batch size should be optimized. Currently, for
product SK the batch size is 4500 till final assembly and for final assembly it is
480, since one pallet can accommodate only 480 product SK’s. Batch sizes for
production should be optimized to have levelled production in order to reflect

44
 6. Recommendations for the company

quickly to customer demand and also to reduce the unnecessary staking up of

inventories for the products which are not in demand.

6.2 System level improvements

During the interviews and during the collection of quantitative data, the authors
came know that the ERP system at the company is devoid of many modern function-
alities. For example, the traceability of products between processes is very difficult
as the production order numbers are same from the foundry till the machining pro-
cess. Due to this, the ERP system shows as if the entire process, between foundry
and final assembly, is just one operation. One of the main issues arising due to this
is that of the veracity of the information regarding quality. The information about
quality, for example scrap values, are fed in the system after the machining pro-
cess. It then becomes difficult to identify, where exactly in the flow has the defect
occurred. Another issue is that the defective products are initially included in the
inventory counts and entered in the ERP system. But later on, after defects are
identified, the inventory has to be revised again. This results in loss of time and
resources in managing the situation and furthermore the production schedule has to
be changed to make up for the scrap generated.

One of the solutions is to have information regarding the products entered in the
ERP system, after every process is completed in the flow. The current ERP system
is not able to handle this level of product information details and therefore, the
upgrading the ERP system, considering the future state requirements, can be im-
plemented. Another solution is to split up the production orders for every process in
the production flow. This will help in improving the traceability and keeping track
of real time information.

Currently, the information flow is such that it is operating in isolated island as

explained in section 4. As shown in the future state map, the information flow is
now limited to the first and the last process with the pacemaker at the buffer be-
fore the final assembly. The pacemaker is the point at which the information flow
regarding production is controlled (Rother and Shook, 2009). The production con-
trol now has to check the inventory levels at only this buffer. When the inventory
levels go below the reorder point, a CONWIP card is initiated and sent to the first
process i.e. the foundry. As explained earlier in section 2, CONWIP card does not
specify exactly which products to produce. Therefore, the specification list should
be sent by the production control to the foundry regarding the specific product to
produce depending on the customer demand. It is, therefore, the responsibility of
the production control to check, or have a mechanism to check, the reorder point
levels as well.

Similarly, when the inventory levels in the final buffer goes below the reorder point,
the production control has to send replenishment orders to the final assembly. The
final assembly now will be using products from the inventory at the pacemaker and

45
6. Recommendations for the company

when that inventory goes below reorder point, the procedure, as explained previ-
ously, has to be followed. Furthermore, the production control has to be responsible
for both the inventory levels and customer demands. Based on this, they have to
maintain the level of each product variety in product family AF, as explained ear-
lier. For example, in one day 1042 SK products can be produced, but the demand
might be only 350. Therefore, to avoid all the wastes, it is necessary to take into
consideration the customer demand. This also helps in freeing up the resources for
the other products in the product family AF. The optimal level of each product
variety should be maintained and thereby having a levelled inventory.

6.3 General recommendations

Apart from the above mentioned recommendations, there are some general tools
which can be implemented to further improve the efficiency of the flow.

Currently, there are standardized work procedures, but it can be further improved
in the sense that it can be made more visual. For example, the work procedures
are available at the workstations, but an improvement would be to have a visual
representation of the task to be performed at that particular station. Furthermore,
any new work procedures, should be communicated directly to the workers or to
the one concerned, as it might be the case where the instructions are updated and
visually represented but it might be overlooked by the worker.

To cultivate a culture of continuous improvements (kaizens), everyone should be

encouraged to give ideas regarding their own space of work, at least in the initial
stages. Currently, this process is going on, such as, the feedback from workers
are taken into consideration. But at the same time it is not documented and this
results in the improvements ideas not being potentially implemented elsewhere in
the processes.

46
 7
Discussion and conclusions

This chapter of the thesis report explains briefly about the research statements and
the how the future suggestions relate to the purpose of the thesis. This part also
showcases the achieved results and the final conclusions of this thesis project.

The purpose was to find improvement potentials for decreasing the production lead
time and optimizing the inventories for the enhancement of OTD. The initial step,
in this, was to identify the issues in the production flow which affected the lead times
and this was approached by using value stream as a tool. Value stream mapping is
a widely used tool for visualizing the production flow in relation to the customer de-
mand. Since, in this method, the mapping is done for the entire processes involved,
it is much easier to visualize the value added time, bottlenecks and constraints in
the production flow. The product SK was chosen for this purpose and the reasons
for the same have been explained in the earlier sections.

Considering the various parts involved in the product SK, the longest lead time was
observed in the ‘Body’ part of the the product SK. Following this body part, value
stream mapping was done which included all the processes related to it. It was
observed that the products from all the individual processes were pushed into the
inventory and the time spent by the products in the inventory was also too high.
This was one of the causes for increased lead time and thereby not able to deliver
on time. Also, it was noticed in the information flow that the each department had
their own planners who sent production plans. The information is not integrated
and are working as isolated functions, which is nothing but a push flow.

The next part of the purpose was to have control over the production flow and the
approach was to first stabilize the processes and then have control and visibility
across the production flow. This was done by establishing the reorder points and
safety stocks. The main reason for this was variability in lead times and demand
in the production flow and the that the manufacturing strategy was make to stock.
The reorder point was chosen before the final assembly as it served as a decoupling
point between push and pull processes. The push process being from foundry to
washing and the pull process being from the final assembly to the warehouse. For
the safety stock calculations, which is part of the reorder point, the lead time devi-
ations, demand deviations and service level factors were considered while proposing
the solution.

47
7. Discussion and conclusions

Initially the feasibility of kanban was checked, but as there is a lot of demand vari-
ation, and therefore, it cannot be implemented. The CONWIP system, which is
similar to the generic kanban system, was proposed as it secures a constant work in
process between the first and last process, with mix model variations. The CONWIP
loop is proposed to be placed from the reorder point, back to the foundry, as this
was a push process and using this method a pull mechanism can be introduced. It
is therefore a combination of push and pull. Based on the proposed future state the
expected lead time is 2.17 days. Also, with the proposed safety stock, it is expected
to sustain up to 7.25 days without production of SK products as explained in the
analysis.

The shortcomings of this thesis is that there is consideration for only the body part
of SK and that only one KPI (OTD - on time delivery), was analyzed. Furthermore,
the effects of other products in the AF product family on SK product group, in
relation to the production flow, was not analyzed as it was out of the thesis scope.
Furthermore, the sequence of production in the CONWIP loop was not calculated
as this was not part of the thesis scope as well. The main idea here is to also have a
balanced quantity, based on the customer demand, of all the varieties of the prod-
uct so that there is no pile up in the final inventory. These two factors should be
taken into consideration, along with the potential new ones, while implementing the
proposed future state map.

48
 Bibliography

Abdulmalek, F. and Rajgopal, J. (2007). Analyzing the benefits of lean

manufacturing and value stream mapping via simulation: A process sector
case study. International Journal of Production Economics, 107(1), pp.223-236

Cawdery, M. (1976). The Lead Time and Inventory Control Problem.

Operational Research Quarterly (1970-1977), 27(4), p.971

Constantin, A. (2016). Inventory management, service level and safety

stock. Journal of Public Administration, Finance and Law

Coyle, J., Bardi, E. and Langley, C. (2003). The management of busi-

ness logistics. Mason, Ohio: South-Western/Thomson Learning

Grewal, C. (2008). An initiative to implement lean manufacturing using

value stream mapping in a small company. International Journal of Manufac-
turing Technology and Management, 15(3/4), p.404

Jonsson, P. and Mattsson, S. (2009). Manufacturing, planning and con-

trol. London: McGraw-Hill Higher Education

J. Hopp, W., L. Spearman, M. and L. Woodruff, D. (1990). Practical

strategies for lead time reduction. American Society of Mechanical Engineers,
3(2)

Karim, A. and Arif-Uz-Zaman, K. (2013). A methodology for effective

implementation of lean strategies and its performance evaluation in manufac-
turing organizations. Business Process Management Journal, 19(1), pp.169-196

Karim, M., Samaranayake, P., Smith, A. and Halgamuge, S. (2009). An

on-time delivery improvement model for manufacturing organisations. Interna-
tional Journal of Production Research, 48(8), pp.2373-2394

Lewis, J. (2005). Identifying 7 types of waste. Upholstery Manufactur-

ing, pp.20-24

Liker, J. and Hoseus, M. (2007). Toyota Culture: The Heart and Soul
of the Toyota Way

49
Bibliography

Marlin, P. (1986). Manufacturing lead time accuracy. Journal of Opera-

tions Management, 6(2), pp.179-202.

Ndubi, S.O., Iravo, M. Ochiri,G. (2016). Effect of lead time variability

on inbound logistics performance in Safaricom Limited. International Aca-
demic Journal of Procurement and Supply Chain Management, 2 (2), 179-205

Ramachandran, G. and Neelakrishnan, S. (2017). An approach to im-

proving customer on-time delivery against the original promise date. South
African Journal of Industrial Engineering, 28(4)

Rother, M. and Shook, J. (2009). Learning to see. Cambridge: The Lean

Enterprise Institute

Shah, R. and Ward, P. (2003). Lean manufacturing: context, practice

bundles, and performance. Journal of Operations Management, 21(2), pp.129-
149

Tersine, R. and Hummingbird, E. (1995). Lead-time reduction: the search

for competitive advantage. International Journal of Operations Production
Management, 15(2), pp.8-18

Thorsen, A. and Yao, T. (2015). Robust inventory control under demand

and lead time uncertainty. Annals of Operations Research, 257(1-2), pp.207-236

T. Gabriel, J. Bicheno, J.E. Galletly, (1991) "JIT Manufacturing Simula-

tion", Industrial Management Data Systems, Vol. 91 Issue: 4, pp.3-7

Wouters, M. (1991). Economic evaluation of lead time reduction. Inter-

national Journal of Production Economics, 22(2), pp.111-120

Zijm, W. and Buitenhek, R. (1996). Capacity planning and lead time

management. International Journal of Production Economics, 46-47, pp.165-
179

50
 A
Appendix 1

A.1 Interview questionnaire for planner

General
1. Brief description about your job profile
2. What are your views on the current production flow of Product SK and product
family AF?
3. What are the current departmental goals?
4. What are the long term, intermediate and short term plans?
5. Brief description about the coordination required in supply and movement of
material and labour.
Scheduling
1. Brief description about the production scheduling process.
2. What are the decision making process for production scheduling?
3. What are the current challenges in it, according to you? what way is it creating
bottlenecks in production, if any?
4. Production planning is partly based on forecasts. Since it is not accurate ,
how is it managed?
5. How are the batch size decided? For example, in the foundry, machining, etc
6. How is the coordination between the different processes/ departments (ex.
between foundry, machining, etc) managed, in relation to batch size?
7. How are the reorder points fixed?
Logistics
1. Brief description about the logistic/distribution process
2. What are the current challenges in it, according to you?
3. In what way is it creating bottlenecks in production, if any?
Production control
1. Brief description on the production control methods, currently in place?
2. How can it be improved?
3. Any views on the current manufacturing operations flexibility, if any?
4. According to you, what are the initial steps required for creating a pull flow,
at IMI?
5. There is one production order for foundry, chuck and CNC. It becomes difficult
to know where in the process are we. Any solutions/ideas to this?
6. Quality checks are after machining. Defects are occurring at an early stage
but detected only after machining, which means the entire process before ma-
chining is waste. Can there me more quality checks?

I
A. Appendix 1

Inventory and utilization

1. How is the inventory level maintained/ optimized?
2. Is there any economic order quantity set, to minimize all inventory costs?
3. How economic and balanced utilization of machines are achieved?

II
 A. Appendix 1

A.2 Interview questionnaire for Supply chain

planner
General
1. Brief description about you and your job profile
2. What is your position and job responsibilities? - Job profile
3. What are your views on the current production flow of product SK and product
family AF?
4. What are the current departmental goals?
5. What are the long term, intermediate and short term plans?
6. Brief description about the coordination required in supply and movement of
material and labour.
Supply chain
1. Brief description about the supply chain process
2. What are the decision making process involved in it?
3. What are the current challenges in it, according to you?
4. In what way is it creating bottlenecks in production, if any?
5. What is the supply lead time for key components of product SK and product
family AF?
6. Is there any level of uncertainty involved in the entire supply process?
7. According to you, is there a pull flow or a push flow?
8. Where in the supply chain can there be a differentiation between a pull and
push flow?
9. How can the flow be improved?
10. What are the initial steps that are the most important to achieve this?
11. IMI is aiming for lean implementation, how will it affect the supply chain
processes?
Inventory and utilization
1. How is the inventory level maintained/ optimized?
2. Is there any economic order quantity set, to minimize all inventory costs?
3. How economic and balanced utilization of machines are achieved?
4. Production planning is partly based on forecasts. Since it is not accurate, how
is it managed?
5. How is the batch size decided? For example, in the foundry, machining, etc
6. How is the coordination between the different processes/ departments (ex.
between foundry, machining, etc) managed, in relation to batch size?
7. How are the reorder points fixed?

III
A. Appendix 1

A.3 Interview questionnaire for Industrial

Engineers
Inventory and utilization
1. Say us few words about the company
2. Say about your designation and your job profile
3. Say about the products that are manufactured in IMI?
Inventory and utilization
1. What are your views on the production flow of product SK in Ljung?
2. Say about your views on the production flow of compact modulator
3. What are the constraints that you feel in the production of product family AF
?What are the current challenges in it, according to you?
4. According to your point of view, which is the main constraint / bottleneck in
the flow?
5. How do you prioritize the bottlenecks in the flow?
6. How do you feel about the present production batch size?
7. How the batch sizes are fixed for the processes? Is it based on the forecast?
8. How the reorder points are fixed for the processes?
9. How do you feel about the production lead times for AF product family ?
10. What do you feel about the current inventory and super markets at various
places in the production flow?
Future Improvements
1. Do you have any suggestion for improvement in the production flow of product
family AF?
2. How do you feel about having the pull and continuous flow for the product
family AF?
3. What are the hindrances that you think that could come on the way while
doing adopting the pull and continuous flow?

IV
 A. Appendix 1

A.4 Interview questionnaire for quality manager

General and current process
1. Brief description about you and your job profile?
2. What are your views on the current production flow of product SK and product
family AF?
3. What are the current departmental goals?
4. What are the long term, intermediate and short term plans?
General and current process
1. Brief description about the quality process
2. How do you feel about the current quality problems of product SK and product
family AF?
3. How many times the quality is been verified in the total production flow of
AF?
4. Quality checks are after machining. Defects are occurring at an early stage
but detected only after machining, which means the entire process before ma-
chining is waste. Can there me more quality checks?
5. What are the current challenges in it, according to you?
6. In what way is it creating bottlenecks in production, if any?
7. Say about the future action plans for the quality improvement (short and long
term goals)?

V
A. Appendix 1

VI
VII
B. APPENDIX B

B
APPENDIX B

B.1 Symbols used in Value Stream Mapping

VIII