Master’s Thesis in Industrial Engineering and Management, 30 credits Supervisor Ume˚a University: Victor Falgas Ravry External Supervisor: Sture Öberg, Smurfit Kappa Examiner: Gerold Jäger

(1)

Masters Thesis in Industrial Engineering and Management, 30 credits

Supervisor Umeå University: Victor Falgas Ravry External Supervisor: Sture Öberg, Smurfit Kappa Piteå

Stock and Cuts in Piteå

Standardization vs. Customization

in the pulp and paper industry

Frida Bergvall May 29, 2019

Master’s Thesis in Industrial Engineering and Management, 30 credits Supervisor Ume˚a University: Victor Falgas Ravry

External Supervisor: Sture ¨Oberg, Smurfit Kappa Examiner: Gerold J¨ager

(2)

(3)

Abstract

A major challenge for companies is to find the appropriate balance between standardization and customization. Today, Smurfit Kappa Pite˚a manufactures Kraftliner paper with a great range of grades, grammages, widths and diameters which increases complexity and total costs. The purpose of this work is to identify what savings potential may lie in reducing the number of articles. In this project we focused on the costs incurred in trimming and stocking. A standardization tool has been designed to be able to look

at different standardization scenarios. The results show that if Smurfit Kappa Pite˚a were to standardize and only produce reels with a diameter of 1400mm, they would be able to reduce their required safety stock and the number of trim reels needed to fulfill demand. Thereby annually reducing their inventory costs by 316 440 EUR and the possibility of gaining 4 020 462 EUR in net value by selling the minimized

waste and trim reels as ordered reels to European customers.

(4)

Sammanfattning

En stor utmaning för företag är att hitta den rätta balansen mellan standardisering och kundanpassning. Idag producerar Smurfit Kappa Pite˚a papper av en mängd olika kvalitéer, ytvikter, bredder och diametrar vilket

¨

okar b˚ade komplexiteten och de totala kostnaderna. Syftet med den här rapporten är att identifiera de potentiella besparingarna som g˚ar att göra om antalet artiklar minskar. I det här projektet l˚ag fokus p˚a de kostnader som uppst˚ar i samband med trimning och lagerh˚allning. Ett standardiseringsverktyg har utformats

f¨or att kunna unders¨oka utfallet vid olika standardiseringsscenarion. Resultaten visar att om Smurfit Kappa Pite˚a skulle standardisera och bara producera diameter 1400mm, skulle de kunna minska storleken p˚a deras

säkerhetslager samt minska antalet trim rullar som krävs för att tillgodose kundernas efterfr˚agan. Och därigenom ˚arligen minska lagerkostnaderna med 316 440 EUR och möjligheten att öka 4 020 462 EUR i nettovärde genom att sälja det minskade trimbortfallet och trimrullarna som orderrullar till de europeiska

kunderna.

Titel: Lager och Trim i Pite˚a

-Standardisering vs. Kundanpassning inom pappersindustrin

(5)

Glossary

In this section we will define commonly used terms and concepts that may be unfamiliar to the reader.

Kraftliner - A paper product with fresh fibers from softwood or hardwood. In this project we will regularly refer to kraftliner paper by just using the word paper.

Grade - Type of paper.

Grammage - Mass per unit of area, expressed in gsm(g/m²).

Width - Distance expressed in mm.

Diameter - Distance expressed in mm.

Production Run - Each production cycle is a new run.

Tambour - When the paper is finished, dried and pressed, it is rolled up into a so-called tambour. A tambour consists of paper rolled up on a tambour iron.

Trim - The process of planning how the reels should be cut to minimize waste and manufacture of the reels ordered by customers.

Winding machine - A machine where reels are cut down into subreels with widths and diameters matching customer specifications.

Slitters - Sharp cutting wheels. The distance between the slitters are adjusted to the specified widths for the orders.

Ordered Reel - A reel that is made against a customer order.

Trim Reel - A reel that is made to fill a gap in the trim.

Stock Keeping Unit (SKU) - A product identity that is given to each individual production item or product category. The identity is a unique number used to track inventory and sales through reporting.

Distribution Centre - A facility used to stock products that will later be re-distributed to the customer.

Cycle Stock - Cycle stock is the amount of stock necessary to meet forecasted customer demand, from the time you reorder stock until the time when it arrives.

Safety Stock - An additional quantity of an item held on top of cycle stock in inventory in order to reduce the risk of running out of stock.

Stock-Out - A situation in which the demand for an item cannot be fulfilled from the current inventory.

Lead Time - The time between closing date, the last day for the customer to enter an order in the closest production run, and arrival at the customer.

Enterprise Resource Planning System (ERP) - A software package with integrated IT systems to handle a company’s information and meet a company’s needs for governance and administration.

(8)

1 Introduction

In this section we describe Smurfit Kappa and the project background. We then present the purpose, aims and main questions investigated in this project, and provide an overview of the report structure

1.1 Smurfit Kappa Pite˚ a

Smurfit Kappa Pite˚a is a part of Smurfit Kappa Group and is the largest producer of kraftliner in Europe. Last year the factory produced 719,500 tons of kraftliner. Kraftliner is a paper product that is mainly used as the outer layer in the production of corrugated cardboard boxes. Examples of finished products are store displays and wineboxes, Figure 1. The Kraftliner paper is made from fresh fibers giving the products good printability, strength, moisture resistance and chemical and biological purity [1].

(a) Store display [2]. (b) Bag-in-box products [3].

Figure 1: Two examples of finished products.

In Pite˚a, brown and white kraftliner are produced in two paper machines, called PM1 and PM2. The process begins with the supply of raw wood material to the factory. Softwood and hardwood are converted into pulp.

At the end of the process, the pulp is mixed with water, distributed evenly on a 6.5 m wide wire mesh and then pressed and dried using heated cylinders. The resulting paper is rolled up on large tambours and when the tambours are sufficiently large, they are moved from the paper machine to a winding machine. In the winding machine, the 6.5 m wide paper is rolled to a specific diameter and cut down to smaller reels based on the customers’ orders, Figure 2 [1].

Figure 2: The paper production process [4].

After production, some of the reels are converted to coated reels at the coating unit in Karlsborg. Here, they receive a coating layer that smoothes the surface and allows for better printing results. This type of paper is used for cardboard boxes that need a high-quality finish such as consumer packaging and retail packaging. After this process the coated reels are returned to Pite˚a again [5].

(9)

All produced reels are then delivered to different distribution centres around Europe. There are a total of six distribution centres located in Pite˚a, Bremen, Terneuzen, Sheerness, Alicante and Modena, Figure 3. Deliveries are made by truck, train or boat.

Figure 3: A map with the location of the Smurfit Kappa distribution centres in Europe [4].

1.2 Background

Smurfit Kappa Pite˚a produces paper with a great variety of qualities, grammages and widths in different diameters. This is done by adjusting the machines, a higher amount of fibres gives more grammage and depending on how the paper on the tambour is trimmed it can be cut into different widths. A larger diameter is obtained by allowing more laps of paper on the core. Varying these parameters allows the customers to choose from around 2600 different items. A greater variety of production items increase the need for conversion in production, the need for trimmed reels and the number of stock keeping units (SKUs). Conversion leads to loss of production and the trimming costs are believed to be of the order of several million SEK each year [6]. The number of production items also affects inventory and customer service negatively. A way to deal with this problem is standardization.

Standardization can be defined as “the process of setting generally uniform characteristics for a particular good or service”. Standardization can be used within an organization for a number of reasons. It can help minimizing variance and thus increase predictability in production and for consumers [7]. The opposite of standardization is customization which, at its extreme, means developing and producing tailor-made products without any standardized elements. By contrast extreme standardization occurs when every product is produced, priced and promoted in exactly the same way [8].

Most of the company’s customers are internal to Smurfit Kappa which gives Smurfit Kappa Pite˚a more freedom to restrict their customer’s number of choices through standardization. On the other hand, since the customers also are part of the Smurfit Kappa Group, they have the responsibility of supplying their customers with the Kraftliner paper adequate to their needs. The end customers demand various types of cardboard boxes and therefore some level of customization is required.

(10)

1.3 The Assignment

The assignment from the company Smurfit Kappa Pite˚a is to determine how they should standardize production by reducing the number of articles and thereby mitigating the negative effects of excessive customization described earlier. To do this an appropriate balance between standardization and customization needs to be found.

1.3.1 Purpose and Aims

The purpose of this work is to identify what savings potential may lie in reducing the number of articles.

The aims are to map out the current situation at Smurfit Kappa Pite˚a, to assess the potential savings by reducing the number of articles produced by the factory, and to create a model in Microsoft Excel and Matlab where articles and costs can be modified to simulate different standardization scenarios. From the outcome of these simulations, a proposal for a new standard should be produced.

1.3.2 Thesis Questions

For the study to fulfill its purpose and aims, three research questions have been formulated. They will be used as a basis for the analysis and recommendations.

Q1. What costs are affected by standardization and how can these be estimated?

Q2. How can we model the outcomes of different standardization scenarios?

Q3. To what extent can customers switch to other articles in case of standardization, and what costs might this entail?

1.4 Scope

In this project, we will restrict our attention for the standardization proposals to the Smurfit Kappa factory in Pite˚a and to its internal European customers. The factory also sells to customers in other parts of the world, but we choose to focus on internal customers because of the possibility of changing their consumption behavior by introducing standardization measures. European customers represent the bulk of the orders for the factory. In the standardization model and simulations on the other hand, we will use data for all customers to be able to create a such as realistic simulation as possible. If the external customers were removed, we would risk having an incomplete or misleading picture of the overall effect of a standardization measure.

The simulations will be based on collected production and economic data from 2018. The decision to rely on the previous year’s production data alone is based on several factors. We wanted to use as updated data as possible and since there is not enough data from 2019, we chose 2018 to be able to study a whole year. One year’s data should be enough to be able to generalize patterns in consumption and production, a longer period of time means more data and it can become difficult to handle.

1.5 Report Overview

The report is divided into seven parts. The next section describes how Smurfit Kappa Pite˚a operates today. In Chapter 3, we give some theoretical background on standardization, cost estimation and various optimization approaches. Chapter 4 describes the methodology and Chapter 5 describes the results from the scenarios simulated by the implemented model.

In Chapter 6 we perform an analysis of the results. We conclude in Chapter 7 with our recommendation to the company.

(11)

2 Current State

In the following section we describe the current state which we have mapped out through interviews and observations.

We explain how the company has worked with standardization until today and how the operational activities are affected by the selection of articles to produce.

2.1 Standardization

Smurfit Kappa Group in Europe is constantly working to maintain and improve their high levels of standardization. The company has a Paper Product Development division part of which is focused on standardization questions, see Figure 4.

Figure 4: The hierarchical structure for the Paper Product Development [9].

A number of standardization initiatives have been implemented during the last decade. In 2015 Premium Paper Grades (PPGs) was introduced, a set of standards, covering all deliveries of standard papers for corrugated board.

These were implemented due to an observed increase in the number of stock keeping units, which until 2010 had been on a downward trend. To address this problem a standardization was performed to reduce complexity. It focused on three parameters: grade, grammage and width [10].

The PPG initiative affected the entire Smurfit Kappa Group in Europe including Smurfit Kappa Pite˚a and its internal customers. The basic idea is that if customers/intragroup corrugated plants¹choose to buy paper of a certain standard, they are rewarded with a standardization discount for some widths.

The customers can choose their standard paper from a predetermined number of PPGs and European Widths of which some are also European Standard Widths, see below.

From the European Widths:

• Each customer can choose a maximum of seven widths as their standard widths.

• They can choose a maximum of two widths below 1800 mm.

• They can choose a maximum of five widths from 1800 mm and upwards.

From the European Standard Widths:

• Each customer customer can choose their ”top three widths” and get a standardization discount per ton on PPG paper.

Some customers need exceptions due to the nature of their business and get additional width choices to be able to deliver what their customers demand. There are also regional exceptions and these must be approved prior to supply [10].

1Intragroup corrugated plants = Internal customers producing corrugated cardboard boxes.

(12)

2.2 Sales

All customers, internal and external, place their orders through their regional Sales Office (SO). The SO ascertains whether the order can be produced and delivered by a plant by the desired delivery date whereupon the order is planned in a production run².

However, most internal customers use VMI, which stands for Vendor Managed Inventory. The customer places a framework order for each article and vendor and the SO is then responsible for keeping track of inventory levels and avoiding stock-outs. This is a way of integrating suppliers and customers with one another by sharing information².

2.3 Production

Today there are five different grades produced in Pite˚a:

• Brown Kraftliner (K)

• Kraftliner White Top (KWT)

• Royal 2000 (KW2)

• Kraftliner White Top Heavy Coated (KCH)

• Kraftliner White Top Light Coated (KCL)

Properties that distinguish grades from each other are color, layer of coating and the proportion of recycled fibre. Reels in various grades are shown in Figure 5.

Figure 5: Reels in various grades [4].

2.3.1 Paper Machines

There are two paper machines, PM1 and PM2, Brown Kraftliner (K) is produced at PM1. Kraftliner White Top (KWT) and Royal 2000 (KW2) are produced at PM2. The remaining two grades are produced by putting a coating layer on already produced KWT reels, this will be explained in more detail later in the report. The difference between the papers is the relative composition of pulp mixture. The pulp mixtures are made from softwood, hardwood and recycled fibre.

A production cycle for the machines lasts one week and production runs continuously 7 days a week, 24 hours a day.

The goal is to be able to use 93% of the cycle time for efficient production. The remaining 7% of the time is set aside for scheduled and unscheduled stops in production which are expected to occur³.

2Helena Eriksson: Supply Chain Manager (Scandinavia). Smurfit Kappa. 2019. Interview 24 May.

(13)

In the paper machines the grammage is determined by the amount of fibers in each layer (a total of 2 layers on PM1 and PM2). More fibers gives a higher grammage. The transitions between grammages can be free from discard if the difference between the two grammages is less than some threshold (whose value varies between 6 and 20 gsm, depending on the grade produced). If the difference between the two grammages is larger, the paper produced during the transition will not meet the production standards, and must therefore be discarded. The lost profit from this waste can be reduced by dissolving the wasted paper and reusing it in production³.

Another source of discard are transitions between the grades KWT and KW2 at the machine PM2. If everything goes as planned, this only occurs once per production cycle³.

Grammage also affects the speeds at which the machines can run and how many tons can be produced per unit of time.

When the grammage is high the machines run slowly because it is harder to drain the paper and more steam is needed to reach the dryness specification, but the produced quantity per time unit is higher. When the grammage is low, the machines can run faster, but there is a greater risk that there will be interruptions because a thinner paper breaks more easily.

2.3.2 Winding Machines

There are two winding machines, RM1 and RM2. RM1 takes care of the tambours from PM1 and RM2 takes care of the tambours from PM2. In the winding machines, the diameter and width of the finished products are determined. The paper from the tambour is unwound and run through the slitters until the correct diameter is reached. As the paper passes through the slitter, it is cut by a set of knives. The positions of the knives determines the final widths. This procedure is known as trimming and the trim planner decides in advance which widths are to be produced.

The trim planning is done once a week after the order entry has been closed, also called closing date. Every combination of grade, grammage and diameter must be planned individually. All entered orders are inserted into a program which outputs a trim suggestion. The trim planner can then make alterations manually if she thinks there is a better combination, with less waste. The goal is to achieve less than 2% of total waste during a year⁴.

There are some constraints that must be met. The sum of the planned widths must be somewhere in the range between a minimum width and a maximum width. The width of the tambour varies between 6450-6600 mm depending on grammage and the maximum trim width must always be a little lower than the width of the tambour since uneven edges must be removed. The minimum trim width must always exceed 6000 mm because too wide edge strips in the winder gives runnability problems. There are only seven knives that can be used at the same time and this means that there can be a maximum of eight different widths on each set⁴.

To be able to meet these constraints and avoid large gaps in a set, it is possible to either borrow a reel from orders in later production runs or to create a trim order. When a trim order is created, the reel is often sold at a lower price to other markets that are further away, called the Overseas markets. These customers must adapt to what is outputted in production, typically, narrow reels used to fill gaps. Most of the Overseas customers buy reels with a variable diameter which means that it is acceptable to deliver both small and large diameter. Usually they prefer paper with a lower grammage since it is possible to produce more cardboard boxes per square meter. Many produce lighter boxes that do not require a high grammage to hold for the loaded content, for example fruit boxes⁵.

3Lars Bergstr¨om: Production Engineer. Smurfit Kappa. 2019. Interview 18 February.

4Minna T¨orma: Trim Planner. Smurfit Kappa. 2019. Interview 12 April.

5Markus Rensfeldt: Sales Manager. Smurfit Kappa. 2019. Interview 24 May.

(14)

2.4 Inventory

After production, the reels are transported to different distribution centers (DCs) around Europe and from there on to the customer. The DCs are placed in Pite˚a, Bremen, Sheerness, Terneuzen, Alicante and Modena. Average lead times, times between replenishment and default modes of transportation are presented in Table 1.

DC Lead Time (Weeks) Replenishment(Weeks) Default Transport

Alicante 6 4 Boat

Pite˚a (DS0) 2,5 1 Train

Modena 2,5 1 Train

Sheerness 4 1 Boat

Bremen 3,5 1 Boat

Terneuzen 4 1 Boat

Table 1: Lead times, times between replenishment and default modes of transportation for each DC.

In Pite˚a, a model called Logistics 90 is used to calculate the size of the required stock at each DC, see Figure 6.

The quantity of each article to be held in stock is determined using average lead time, variations in demand, service level and forecasted consumption. Recall that the lead time is the time between closing date and arrival date at customer.

Figure 6: A graphic representation of Logistics 90.

The quantity of articles to hold in stock is decided using limits based on volume and variability. If the total demand for an article is too low or if it varies too much, it may not be worth stocking and is classified as a nonstandard article.

All articles with sufficiently strong and stable demand are called standard articles and considered important to have in stock at all times.

An article a is classified as standard if (1) demand for a is sufficiently large and (2) demand for a is sufficiently stable.

da≥ V olumeLimit (1)

where da is the expected demand for a.

σd/da≤ V ariabilityLimit (2)

where σd_a is the standard deviation of the demand for a.

The values for V olumeLimit and V ariabilityLimit are decided individually for each of the DCs.

(15)

Articles ordered with VMI are stored at a DC until replenishment is made by the SO. The SO uses an integrated module in the Enterprise Resource Planning (ERP) system SAP, called Supply Chain Management (SCM). SCM is a tool to help with replenishment and uses data from the system to calculate a suggestion based on customer stock level, DC stock level, lead time, forecast and safety stock. The data is updated and provides new replenishment proposals on a daily basis. If there is not enough paper at the DC to meet customers needs, the system adjusts the proposal and calculates an appropriate allocation called fair share⁶.

A good forecast facilitates replenishment since the theoretical calculations can be trusted. If the real demand deviates too much defined per region, the customer has no right to compensation if a stock-out should occur. This is an agreement between the SO and their internal customers. It is therefore in everyone’s interest to generate a forecast as realistic as possible⁶.

2.5 The Coating Unit

The coating unit is located in Karlsborg, about 130 km from Pite˚a. Here they produce Kraftliner White Top Heavy Coated (KCH) and Kraftliner White Top Light Coated (KCL). KCH and KCL are produced by adding a coating layer to already produced Kraftliner White Top (KWT) reels. When KWT reels are used for this purpose they are called base paper.

The base paper is produced and stored in Pite˚a where there is a safety stock based on the customers’ forecasts. The trim planner treats base paper differently from the other KWT reels. They always have a diameter of 1250 mm and can only use two-thirds of the total trim width of 6450 mm. The last third must be either a regular KWT customer order or a trim reel. The reason for this is that the base paper is extra sensitive and the two thirds restriction leaves the opportunity to adjust the position of the reels on the tambour after the paper is produced.

The coat orders are received in Pite˚a by the coated liner planner. She is responsible for the production of base paper, the transport by truck to Karlsborg and scheduling the production of coated reels in Karlsborg. Based on the orders, a production plan is created. It is mainly planned according to grade. The grades must lie in separate production blocks because of the different coat mixtures. After grade, the production is planned based on width and grammage. The reels are produced from wide to narrow widths and the trim planner tries to prevent large jumps between grammages to avoid waste due to conversions. For the start-up of each new production block, a start-up base paper reel is used because the first reel is always discarded⁷.

When the reels are finished and have got a layer of coat, the paper gets a higher density. The higher density affects the weights of the reels which can be a problem for some customers. They might have weight restrictions for their corrugated board machines or on other machines transporting the reels. This is why reels with a diameter of 1400 mm are currently not in production. Karlsborg has previously handled heavier reels for customers in the US and probably still has the capacity to do so but it has been a long time since this was investigated⁸.

6Helena Eriksson: Supply Chain Manager (Scandinavia). Smurfit Kappa. 2019. Interview 24 May.

7Angelica Zingmark: Coated Liner Planner. Smurfit Kappa. 2019. Interview 14 February.

8Erik Tollander: Technical Service Director. Smurfit Kappa. 2019. Interview 6 February.

(16)

3 Theoretical Framework

In this section we describe theoretical background relevant to this project. We discuss theories about Standardization, Cost Estimation, Optimization and Modeling and explain why they are relevant.

3.1 Standardization

The appropriate level of standardization for a multinational company depends on what industrial sector they are oper- ating in, the type of customer they are serving, overall turnover and number of employees. A standardized product can give economies of scale in production, research and development. Instead of producing a small amount of a great variety of products, a large quantity of a single article can be produced which reduces production losses and increases machine utilization. A large quantity of a single article is easier to plan and control and entails smaller stock [11]. Theory regarding standardization can be used as a starting point to find potential savings.

On the other hand, standardization can slow down market development. Instead of competing with unique and cus- tomized products, focus may be on developing products that can be produced to lower price. A company that operates in many different markets may need to handle different cultures, economic conditions, legal and political factors [11].

These aspects must also be taken into account when making a decision to standardize.

Ideally companies would like to achieve the benefits of standardization in terms of cost reduction whilst meeting the local requirements mentioned above. Postponement, or delayed configuration, refers to the process of delaying final assembly or customization of a product or service until the last possible moment, when the final market destination or customer requirement is known. There are several benefits to this strategy. Inventory can be held at a generic level with fewer stock keeping units and greater flexibility. The same components or modules can be assembled in a variety of end products [12, p. 177-178].

3.2 Cost Estimation

Accounting is a way of communicating economic information to a company’s stakeholders and can be separated into two categories: internal and external accounting [13, p. 4]. External accounting is called Financial Accounting and provides historical information to external stakeholders such as shareholders and creditors [14, p. 7]. Internal accounting is called Management Accounting and provides information to people within the company. This kind of information is often more detailed and focuses on future projections. Managers can use this economic information to make decisions and control activities [13, p. 4]. Management Accounting can help us provide the internal economic information needed to make a decision about standardization.

For decision-making, it is important to identify which costs are relevant and will be affected [13, p. 23].

Sunk costs are not calculated because they are costs that have already incurred and are not relevant [13, p. 31].

Differential costs are the difference between the costs of two alternatives and are therefore affected by which choice is made [13, p. 32].

3.2.1 Production Costs

For a manufacturing company, costs can be divided into manufacturing costs and non-manufacturing costs. Man- ufacturing costs are constituted of direct material costs, direct labor costs, manufacturing overhead costs [13, p. 24-25].

Direct material costs are the costs of the raw materials used in production.

Direct labour costs are costs incurred from the employment of the workers who convert the raw materials into finished products.

Overhead costs are all other costs that arise during manufacturing such as rent, utilities. insurance and salaries that are not job- or product-specific.

Direct material and labour costs are the easiest manufacturing costs to estimate since they can be associated directly to a specific article. Overhead costs are often shared between product lines and it can therefore be hard to allocate these. Activity Based Costing (ABC) is a costing methodology used to allocate overhead costs to articles, or in other words, identify the relationship between articles and the resources used in their production.

(17)

3.2.2 Inventory Costs

Inventory Carrying Costs are variable costs incurred to store produced items that increase as the stock grows in size [15, p. 40]:

The Cost of Capital can be seen as an opportunity cost for the tied-up capital, the cost of not being able to invest the money somewhere else with a higher return.

The Risk Cost is the cost entailed by having goods in stock. There is always a risk that the products cannot be sold, that there will be shrinkage or scraps, and there are insurance premiums that must be paid.

Warehousing Costs are semi-solid costs which, in the short term, are independent of the stored volume and are only affected by major changes. If inventory volume increases significantly, more space and more equipment might be needed [15, p. 105].

Facility Costs are costs of leasing a storage area or building.

Equipment Costs are costs of equipment such as trucks or warehouse racking used for storage.

Labour Costs is the cost of the personnel handling the storage.

Inventory costs can be calculated using Equation (3).

Inventory Costs (EUR/t,year) = Inventory Carrying Costs + Warehousing Costs (3)

3.3 Stock Calculations

A Cycle Stock is a stock based on the average planned consumption. It arises from the fact that there is a higher quantity delivered into the stock than what is needed for immediate consumption. In purchasing and production, economies of scale are achieved by purchasing or producing large volumes at a time [15, p. 109-113].

A Safety Stock is held to be able to provide a chosen level of service and mitigate the risk of stock-outs. There may be an uncertainty in the customers’ demand and unexpected things may occur in the production or at suppliers. The safety stock helps to manage this uncertainty and the size is determined according to the desired service level one is aiming for [15, p. 109-113].

(a) The basic model of a cycle stock where replenishment is done

every week [15, p. 110, the historical image has been reworked]. (b) The basic model of a safety stock combined with the cycle stock [15, p. 110, the historical has been reworked].

Figure 7: Examples of different types of stock.

(18)

Figure 8: Service levels and the normal distribution [12, p. 45]

The cost curve in Figure 8 assumes that the demand is normally distributed, with a mean x and standard deviation σ.

When these parameters are known they can be used to calculate the probability of a given value occurring. This can be useful when deciding how much safety stock is required. Figure 8 shows how much safety stock is required if demand is approximately normally distributed. If σ is large relative to ¯x, a small incremental change in service level requires a big investment in inventory [12, p. 43-45].

Equation (4) uses the standard deviation for forecast errors and deviations in lead time to calculate required safety stock at given service level, assuming that these are normally distributed [15, p. 244].

Saf ety Stock = k ∗ q

σ_d²∗ LT + x²∗ σ_LT² (4)

where

σd= standard deviation for forecast errors.

σLT = standard deviation in lead time.

x = mean demand.

LT = mean lead time, the time between closing date and arrival at customer.

k = multiplier that can be collected from a table.

Assuming that the lead time is reliable and lacks standard deviation (i.e. σLT = 0), Equation (4) reduces to [15, p. 243].

Saf ety Stock = (z − x)√

LT (5)

z = x + kσd, where z is the quantity of stock you take into your inventory to ensure a good service level; it constitutes of x, which is the cycle stock, together with the safety stock kσd

√LT .

The Average Cycle Stock can be calculated using Equation (6).

Average Cycle Stock = repl ∗ (x/2) (6)

where repl is the time which passes between two replenishments.

The Average Stock can be calculated by adding Equations (5) and (6).

Average Stock = Average Cycle Stock + Saf ety Stock (7)

(19)

3.4 Optimization

Optimization models are used to describe complex systems mathematically and to find optimal choices of parameters with respect to some objective function. Optimization models can be a practical tool in decision making in all sorts of areas and is widely used within business, science and engineering. With the increase in computational power in recent years it is now possible to solve large problems with thousands of variables [16, p. 3].

An optimization problem can be formulated generally as follows [16, p. 4]: minimize the value of the objective function f (x) over a specified set of feasible points x ∈ S, subject to the constraints

gj(x) ≤ bj, j = 1, ..., m

where g1(x), ..., gm(x) are real-valued functions of x and b1, . . . , bmare fixed real constants.

A solution x ∈ S that minimizes f (x) is called an optimal solution and can be written as x^∗. The optimal value of f over S subject to the above constraints is denoted by z^∗.

3.4.1 The Cutting Stock Problem

An optimization problem that can be found particularly relevant to the paper industry, is the Trim-Loss Problem also known as the Cutting Stock Problem. This kind of problem belongs to a broader class of cutting and packaging problems [17]. When a large reel of paper is produced, it has to be cut into different sizes to meet the demand from customers. This can usually not be done without throwing away some of the material, see Figure 9. The goal is to find the optimal cutting scheme that minimizes this waste [18].

Figure 9: An illustration of the trim-loss problem [17].

(20)

Mathematical Formulation

Aim: given a set of orders and a set of cut patterns, find an optimal cutting scheme with respect to trim loss [16, p. 224].

Set 1

I=set of order widths.

A cut pattern j is legal if it satisfies Equations (8) and (9).

X

i∈I

iaij≤ Wmax, with xj∈ Z , x^j> 0 (8)

(The width of each pattern must be less than or equal to the maximum width of the raw paper reel.)

X

i∈I

iaij≥ Wmin, with xj∈ Z , xj> 0 (9)

(The width of each pattern must be less than or equal to the minimum width of the raw paper reel.) Set 2

J =set of all legal cut patterns.

Parameters

aij=number of reels of width i cut in pattern j.

bi=demand for order reels of width i.

Wmax=maximum width for each pattern.

Wmin=minimum width for each pattern.

Decision Variables

xj=number of reels cut using pattern j.

Objective Function

min z =X

j∈J

xj (10)

(Minimize the total number of reels z.) subject to the constraints:

X

j∈J

aijxj≥ bi, ∀i ∈ I (11)

(The number of reels of width i must be greater or equal to the demand for order width i.) xj≥ 0

xjinteger

Solution Approach

The problem as described above is a integer linear programming problem. However, one should add as a remark that it is computationally too onerous to generate the whole set J and optimize over it. The algorithmic strategy one adopts instead is to start with a set of trivial legal cut patterns J0. One solves the linear optimization problem over J0, and then checks whether there is a simple extension J1of J0which would give a better solution, in which case one solves the linear programming problem with J1replacing J0 as the set of cut patterns. This procedure is iterated, generating sets of cut patterns J1, J2, ..., Jf until one has a set of cut patterns Jf for which there exists a solution that is (within a small error tolerance factor) optimal with respect to the overall set J.

(21)

4 Methodology

This part describes the methods used to answer our thesis questions.

4.1 The General Approach

The approach of this project was to start by collecting qualitative data through interviews and quantitative data from internal databases, documents and files. In parallel, theories about standardization, cost estimates and linear optimization were investigated as part of a literature study. The collected data was used together with the theoretical framework from the literature study to gain a better understanding of possible standardization scenarios. This information was then used to build and implement a model. Once this was done, we analyzed the outcome from different scenarios and made a final recommendation to the company.

4.2 Data Collection

The data collection came in two parts: Interviews and Quantitative Data.

4.2.1 Interviews

Interviews are a way of collecting qualitative data that is hard to obtain through performance records or observations.

Interviews can be divided into two categories: structured and unstructured interviews. A structured interview is when the same questions are asked in the same order to be able to compare the answers. An unstructured interview starts from a few general questions and branches off in different directions depending on the answers from the participant and the follow-up questions from the interviewer. It is possible to get useful and more detailed information than from for example a questionnaire but this method requires more preparation [19, p. 23-24].

In this project, unstructured interviews were conducted to avoid closing down perspectives. They were conducted mostly orally but also via email exchanges and through video calls. The latter were used to interview Smurfit Kappa employees working in other offices or abroad. In preparation for each interview, questions were tailored to each interviewee and to their field of expertise within the company. Most of the interviews were recorded with the Voice Memo app on iPhone and were later transcribed.

The interviews were used so as to understand the current standardization situation, both for Smurfit Kappa Pite˚a as a plant but also for Smurfit Kappa as a large global group. The qualitative data from the interviews gave an overall picture of the production process and stock control system. The information was also used to investigate what opportunities there are to make changes.

4.2.2 Quantitative Data

Historical data about production, customers and billing was gathered from DivePort, an internal database that retrieves its information from Northern Lights, Pite˚a’s Enterprise Resource Planning (ERP) system. But some information was also gathered directly from Northern Lights. The data was then ’cleansed’ by identifying incomplete or incorrect parts of the data and then modifying or deleting them. Examples of cleansed data are records of reels with an impossible weight or without article information. The cleansed historical data was then used as an input in our model.

The other type of gathered data consisted of internal documents and files about customers restrictions, inventory calculations and trimming results. This information was used to build the model but also as a part of the results.

(22)

4.3 Modeling

A model was created incorporating the information from the data collection and literature study. From this information we decided to simulate how standardizing diameter could improve the trim process and reduce stock levels. Once the model was produced we simulated costs for different scenarios in order to make recommendations to the company. The results from the simulation runs can be compared to see how the costs for trim and inventory would differ in the various standardization scenarios.

The model was implemented in Microsoft Excel and Matlab. The advantage of using Excel is that all employees at the company have access to the program, which is positive as it provides high availability. Unfortunately Matlab is not avail- able at the company at the moment but it is possible to purchase access if the company considers the model useful. In Excel one can insert notes in the spreadsheets explaining how it works which is advantageous from a knowledge-transfer perspective. All code was commented, so as to facilitate the handover to Smurfit Kappa at the project’s end. When a copy of the model is handed over to the company its engineers will be able to develop it further or alter it as needed.

4.3.1 Model Construction Input:

Historical production data and economic data for all reels produced at PM1 and PM2 for a chosen period of time.

Output:

Trim: Potential net value.

Inventory : Potential inventory costs savings.

The interface is built in Microsoft Excel, where you can create tables that can be retrieved from Matlab where the simulations are run, see Figure 10. The model can be divided into four parts:

1. Create Tables (Historical Data), the user selects which type of paper to study and the model will create tables from the historical data used as input in the model.

2. Run Simulation (Without Standardization), the user can run the simulation with the tables created from the historical data and receive an output with theoretical results.

3. Create Tables (Standardized Data), the user selects which standardization scenario to study and the model will create tables by modifying the historical data so that it mimics the chosen scenario.

4. Run Simulation (With Standardization), the user can run the simulation with the tables created from the standardized data and receive an output with theoretical results.

Figure 10: The model interface in Microsoft Excel. Step 1 and step 3 can be performed directly from here, while step 2 and step 4 must be performed in Matlab.

Tables

Tables are created using Visual Basic for Applications (VBA). VBA, is a limited version of the Visual Basic software language for writing macros in Microsoft Excel. In VBA, a custom user interface can be created by a so-called UserForm.

The UserForm allows the user to choose which type of paper and standardization scenario to study, see Figure 11 for an example. The choices are used when new tables are created. Currently it is only possible to standardize from 1250 mm to 1400 mm or vice versa. What happens at this type of standardization is that reel weights and quantities are recalculated to match historical demand.

(23)

(a) A paper type choice

(b) A standardization scenario choice

Figure 11: Two examples of choices.

Simulation Trim Input:

A table with production week, grade, grammage, width, diameter, weight, quantity, net value Europe, net value Overseas.

Output:

Waste and number of reels.

Trimming costs are simulated using Matlab. As mentioned in the theory chapter, the stock cutting problem can be used to find an optimal cutting scheme and minimize waste. In this project, we used an algorithmic approach to the stock cutting problem to determine a theoretical cutting scheme for the ordered reels for both the historical and in the standardized data. From these trimming schemes we then calculated the trim outcome and compared the results.

Matlab has a pre-built code for solving a geneneric instance of the Stock Cutting Problem. The code was modified to suit the specifities of our problem. For example, it needed a minimum width constraint, see Equation (9). Other additions to the code were algorithms that created lists with quantity, weight and width for each trimming session, algorithms that were able to plot and visualize the results and the possibility to get more information from the output.

The algorithm includes the following steps:

1. Import data from Excel.

2. Go through the table and create arrays for each combination of week, grade, grammage and diameter.

3. Trim by combining the ordered reels in the most optimal way by finding optimal patterns and using a minimum number of reels. Trim reels are inserted when needed to fulfill all constraints. Recall that trim reels are used to fill a gap in a pattern.

4. Sum waste and trim reels for all weeks.

5. Calculate the net value for the resulting waste and trim reels.

The simulation is run for the historical scenario and standardized scenario using the tables created in Excel. The results from the runs can be compared to see how the costs for trim would differ.

Assumptions and Limitations:

There are some sub-grades that must be trimmed and produced in a certain order so as to meet certain specifications (recycled fiber-free for example, or no stains). Another exception is that base paper is trimmed separately from other KWT reels and can only use two thirds of a set. These exceptions are not taken into account by the model, in which all reels with the same grade, grammage and diameter can be combined.

The algorithm that generates patterns for the trim initializes the first set of patterns by using each of the widths a maximum number of times to ensure that each of the widths is included in at least one pattern. When the initial patterns are created, the minimum trim width is not taken into account. One problem with this is that in some instances the algorithm may not find a better solution than the initial one and therefore outputs a pattern that does not satisfy the inequality, Equation (9).

The model can only choose trim reels from the list of widths created for each trimming session. Therefore, if there are few widths on the list there are also few alternative trim options.

(24)

Simulation Inventory Input:

Tables for each DC with demand per week and article, service level, volume limit, variability limit, warehousing cost, net value Europe.

Output:

Stock levels.

Inventory costs are simulated using Matlab. The algorithm includes the following steps:

1. Import data from Excel For each DC and article a:

2. Calculate Safety Stock, using Equation (5).

3. Calculate Safety Stock, using Equation (5).

4. Calculate Average Cycle Stock, using Equation (6).

5. Calculate Average Stock, using Equation (7).

6. Check if article a is a Standard Article that should be stocked, using Equation (1) and (2).

7. Sum stock levels for all articles.

8. Calculate the inventory cost for the required safety stock, using Equation (3)

The simulation is run for the historical scenario and standardized scenario using the tables for each DC created in Excel.

The results from the runs can be compared to see how the costs for inventory would differ.

Assumptions and Limitations:

The service level, the volume limit and the variability limits are decided individually for each of the DCs. These values are not exact and in this project they are estimated with the help of the supervisors at Smurfit Kappa Pite˚a. Adjusting these affects the end result.

(25)

5 Results

This section presents the results of the simulations carried out and a compilation of data regarding the customers’

limitations and restrictions. It also includes a comparison between the built model and the current trim program.

5.1 Simulations

Below are the results of standardizing the diameter from small (1250 mm) to large (1400 mm), and thus only producing reels with the same diameter. In the simulations we assumed that customers’ demand for the same paper could be transferred. The number of ordered reels input into the model is lower in the standardized case to maintain the same level of demand/tonnage, since reels with a diameter of 1400 mm are heavier. The simulations are based on production data from 2018. To estimate how the costs for trim and inventory differ we used net value. We wanted to take into account all parameters that affect the final price to customer. A description of how the net invoiced and the net value is calculated from the order price is presented in Appendix A.

5.1.1 Trim

Table 2-4 present the trim outcome during a year for K, KWT and KW2 if all grades are simulated separately. Inspection of the tables reveals that standardization of diameter decreases waste in all three cases. It also decreases the number of trim reels needed to satisfy demand.

Waste in percent is the proportion of ton not used to produce reels. Reels are the total number of used reels, including both ordered reels and trim reels.

K

Historical Standardized Waste in percent (t) 1.95% 1.27%

(6 969 t) (4 477 t)

Reels 152 087 140 391

-Ordered reels: 143 615 135 668 (344 194 t) (344 604 t)

-Trim reels: 8 474 4 723

(12 960 t) (7 745 t) Table 2: Trim result for grade K.

KWT

(3 142 t) (4 704 t)

Reels 56 596 46 468

-Ordered reels: 50 640 44 267 (127 790 t) (127 893 t)

-Trim reels: 5 956 2 201

(10 543 t) (4 704 t) Table 3: Trim result for grade KWT.

KW2

(4139 t) (2816 t)

Reels 64 959 59 168

-Ordered reels: 54 554 51 201 (159 163 t) (159 288 t)

-Trim reels: 10 402 7 967

(19 647 t) (16 212 t) Table 4: Trim result for grade KW2.

(26)

Potential Gain (Net Value Trim Reels)

Grade Reduced Trim Reels Net Value Europe - Net Value Overseas Potential Gain

(t) (EUR/t) (EUR)

K 5 215 41 213 822

KWT 5 839 62 362 023

KW2 3 435 58 199 236

Total: 775 081 EUR Table 5: The potential net value gain from trim reels, when standardizing diameter from small (1250 mm) to large (1400 mm).

Potential Gain (Net Value Trim Waste)

Grade Reduced Trim Waste Net Value Europe Potential Gain

(t) (EUR/t) (EUR)

K 2 492 588 1 465 296

KWT 1 423 668 950 564

KW2 1 323 627 829 521

Total: 3 245 381 EUR

Table 6: The potential net value gain from trim waste, when standardizing diameter from small (1250 mm) to large (1400 mm).

(27)

5.1.2 Inventory

Table 7-12 present the inventory level outcomes for Alicante, Bremen, Terneuzen, Sheerness, DS0 and Modena. The DCs are located in different places and they stock different articles and therefore should be treated independently. As metioned in Section 2.4, the volume limits, variability limits and service levels are local to each DC and are estimated with the help of the supervisors at Smurfit Kappa Pite˚a.

Table 13 present the potential inventory costs savings of standardizing from small (1250 mm) to large (1400 mm) diameter.

Alicante

Replenishment: 4 Volume Limit: 200 t Lead Time: 6 Weeks Variability Limit: 250%

Service Level: 96%

Historical Standardized

Average Stock (t) 5 857 5 710

-Safety Stock (t) 5 058 4 911

-Average Cycle Stock (t) 799 799

Standard Average Stock (t) 2 120 2 281

Share of Standard Articles 13% (18/137) 18% (21/118) Table 7: Stock Levels for Alicante.

Bremen

Replenishment: 1 (Weekly) Volume Limit: 200 t Lead Time: 3 Weeks Variability Limit: 150%

Service Level: 96%

-Safety Stock (t) 12 589 11 642

-Average Cycle Stock (t) 1 274 1 274

Share of Standard Articles 14% (92/640) 18% (93/518) Table 8: Stock Levels for Bremen.

Terneuzen

Service Level: 96%

historical Standardized

Share of Standard Articles 9% (27/301) 11% (28/263) Table 9: Stock Levels for Terneuzen.

(28)

Sheerness

Service Level: 96%

-Safety Stock (t) 14 700 12 904

Share of Standard Articles 20% (109/537) 26% (108/408) Table 10: Stock Levels for Sheerness.

DS0

Replenishment: 1 (Weekly) Volume Limit: 200 t Lead Time: 2.5 Weeks Variability Limit: 150%

Service Level: 96%

Share of Standard Articles 12% (70/602) 13% (68/522) Table 11: Stock Levels for DS0.

Modena

Replenishment: 1 (Weekly) Volume Limit: 200 t Lead Time: 2.5 Weeks Variability Limit: 150%

Service Level: 96%

Standard Average Stock (t) 469 688

Share of Standard Articles 3% (7/269) 4% (10/233) Table 12: Stock Levels for Modena.

Potential Savings (Inventory Costs)

DC Inventory Costs (EUR/t,year) Reduced Safety Stock (t) Potential Savings(EUR)

Alicante 83.13 147 12 208

Bremen 86.63 947 82 016

Terneuzen 84.48 196 16 548

Sheerness 84.8 1 796 152 286

DS0 84.28 350 29 524

Modena 103.08 231 23 840

Total: 316 440 EUR Table 13: The potential inventory costs savings of standardizing diameter from small (1250 mm) to large (1400 mm).

(29)

5.2 Customer Restrictions

Figure 12 and Figure 13 present information about customer restrictions.

Figure 12: A summary of the internal customers maximum diameter restrictions. Those included are internal customers who have purchased paper at some point during the period 2016–2018, a total of 123 customers. The result shows that out of the total number of customers there are only 11 of them, about 9%, whose plants are not currently able to handle diameter 1400 mm and above. These are marked in red.

Figure 13: A summary of the internal customers maximum weight restrictions. For some of the customers, there is no information on weight limitation. For the remaining 99 customers, the result shows that there are only 31 of them, about 31%, whose plants are not currently able to handle reels that are heavier than 3 ton.

These are marked in red.

(30)

5.3 Model comparison

To ensure that the model calculates what it is supposed to, a comparison was made with the trim program that is currently being used by the trim planner at Smurfit Kappa Pite˚a. Table 14-16 present the results from these tests. To be able to compare the model to the current trim program, waste was calculated in percent of meter not used instead of tonnage not used. Another exception is that the trim planner can borrow reels from later production runs if needed, these are called borrowed reels. The number of borrowed reels can in this case be compared to the number trim reels since they are needed to fill gaps.

The tests were done at two occasions:

Grade: K, KWT, KW2 Grammage 135.

Run: 915 and 918(PM1), 965 and 968(PM2)

K-135

(Run 915) (Run 918)

Trim Program Model Trim Program Model

Waste in percent (m) 1.58% 0.50% 0.99% 2.17%

Total number of sets 189 185 192 206

Total number of reels 586 569 585 628

-Ordered reels: 494 494 544 544

-Trim reels: - 75 - 84

-Borrowed reels: 92 - 41 -

Table 14: Resulting trim for all ordered reels, grade K, in Run 915 and 918.

KWT-135

(Run 965) (Run 968)

Waste in percent 1,04% 1,64% 1,11% 2.12%

-Ordered reels: 504 504 270 270

-Trim reels: 1 9 - 37

Table 15: Resulting trim for all ordered reels, grade KWT, in Run 965 and 968.

KW2-135

(Run 965) (Run 968)

Waste in percent 0.41% 1.75% 0.66% 2.54%

-Ordered reels: 574 574 382 382

-Trim reels: 148 127 41 36

Table 16: Resulting trim for all ordered reels, grade KW2, in Run 965 and 968.

(31)

6 Analysis

In this section we perform an analysis of the results by answering the research questions that were formulated at the beginning of the work. We also discuss the model’s validity and reliability.

6.1 Q1. What costs are affected by standardization and how can these be esti-

mated?

A standardization decision affects costs throughout the supply chain, from the supplier to the end customer. For Smurfit Kappa Pite˚a, the interviews indicated that the costs that are most affected by a standardization are production costs and inventory costs.

Production

Machine efficiency is affected by the grammage of the paper being produced. Grammage decides the speed at which the machines can run and how many tons can be produced per unit of time. The contributions per unit of time are worst for low grammages, the optimal level can be found somewhere in the middle/upper range.

The transitions between two grammages may result in discard if the difference between two grammages is too big. Remov- ing a grammage may therefore cause the discard to increase since it becomes more difficult to make smooth transitions.

But removing extreme ones can reduce the number of sharp transitions.

Trim Optimization is affected by the widths and diameters of the paper being cut. The maximum trim width is around 6450 mm and the ordered reels give a number of combinations to choose from. Reels that are 2100 mm are well adapted to the machines in Pite˚a because they fill the entire width, narrower or wider orders often require a narrow reel to fill the last gap. Narrow reels are often more difficult to sell due to lower demand, which also affects the price. Since two different diameters cannot be trimmed together, the number of possible combinations increases if one of them can be replaced by the other.

Producing wider reels and/or larger diameter gives less reel movements since fewer units are produced. This not only affects the production but the entire supply chain.

Inventory

Inventory Carrying Costs increase as the stock grows in size since capital that could have been invested elsewhere is tied up as stock. It also increases the risk that reels will be damaged or lost.

Warehousing Costs are semi-solid and therefore not directly affected by the stored volume. But it is important that the stored articles are what customers need. Articles with a high variability can be difficult to stock since they might have a very high demand during some periods, which can lead to a too high replenishment. When consumption goes down, the stock is larger than it should be.

A reduced number of SKU:s lowers the variation in demand and decreases the required safety stock and is thus desirable.

Standardization is a way of achieving such an SKU reduction. Two customers can buy paper of the same grade, grammage and width but it they use different diameters, they cannot use each other’s reels. If there is only one diameter, the customers can even out the variations in each others demand. One of the customers may produce more than forecast while the other produces less.

(32)

6.2 Q2. How can we model the outcomes of different standardization scenarios?

The outcome from a standardization scenario can be difficult to model since there are many different variables to handle, rapidly leading to large and complex problems. In this project we chose to reduce the number of variables by focusing on how a diameter standardization would affect trim optimization and inventory. The goal was to be able to show a general pattern for standardization and fewer SKUs.

The stock cutting problem was used to model and visualize the trim outcome. The results from Section 5.1.1 show that trim waste and trim reels decreases for all inspected grades. The biggest waste reduction was found for KWT were the waste decreased from 2.30% to 1.28%. A reason for this may be that a large proportion of the KWT reels is produced with small diameter, since they are used to produce coat reels.

The inventory outcome was modeled using established inventory models. The results from Section 5.1.2 show that the required safety stock and average stock decreases for all DCs. At the same time the standard average stock increases and this means that there is a bigger share of the articles that are worth stocking.

The reductions from the simulations show that there are savings potential in reducing the number of articles. Calcula- tions estimating the magnitude of these savings are presented in Section 5.1.1 and 5.1.2.

6.3 Q3. To what extent can customers switch to other articles in case of stan-

dardization, and what costs might this entail?

For a standardization decision to be made, it is of importance that internal customers are able to adapt their operations to the new conditions. The results from Section 5.2 show that if the diameter were to be standardized, the biggest problem would be the weight that a reel with a large diameter entails.

31% of the internal customers cannot handle a reel that is heavier than 3 ton.

9% of the internal customers that would have trouble handling a diameter that is larger than 1400 mm.

The most common reasons for limitations are corrugator machines and trucks. A reel weighing over 3 ton can be too heavy for the reel stand used at the corrugator and for trucks unloading or moving reels. Costs that a standardization might entail are investments to solve these problems and remove existing constraints.

Standardization should also benefit the customers. For example, a reel with a large diameter consists of more layers of paper, lasts longer and decreases the need for conversions in production. Another positive aspect is fewer reels and fewer movements.

Depending on type of cardboard box and the area of use, different ’recipes’ are used to give the cartons a various of properties. Diameter and width can be standardized without affecting these, but changes in the supply of grades and grammages may force customers to find other combinations. Postponement is a strategy that can be used to maintain end customers’ flexibility whilst achieving the benefits of standardization. Smurfit Kappa Pite˚a belongs to a global group and most of theirs customers are corrugators and not end customers. This enables them to produce paper that can be can be converted to a variety of end products and differentiate as late in the chain as possible.

6.4 Validity and Reliability

To test the validity of the model, that it actually measures what it is supposed to, it was compared to result from the actual trimming program used at Smurfit Kappa Pite˚a. The conclusion that can be drawn from the results in Tables 14- 16 is that the performance of the built trim model is similar to the (manually optimised) current trim program, and hence that it is reasonable to use the trim model when modelling various standardization scenarios.

A reason why the model performs worse than the actual trimming program when it comes to trim waste can be that the model sometimes allows patterns that are narrower than 6000 mm. This is adjusted manually in the actual trimming program which is not possible in the model.

To test the reliability of the model, more test trim need to be performed since reliability means that the result from a test should be the same for repeated measurements.

(33)

7 Recommendations

This is the last section were we present our final recommendations to the company based on the results and analysis.

From the results and analysis in this report conclusions that can be drawn that it is worth looking into the possibilities for a future diameter standardization at Smurfit Kappa Pite˚a. According to our model it is possible to reduce annual inventory costs by 316 440 EUR and to gain 4 020 462 EUR each year in net value by selling the minimized waste and trim reels as ordered reels to European customers.

Before making a decision it must be investigated how much investment is required to remove existing customer constraints and compare that investment to the numbers in this report. The coating unit has previously been able to handle reels with larger diameter but it this must also be investigated more closely.

The standardization model can be developed further by adding standardization scenarios or improving the already existing ones.

(34)

References

[1] Smurfit Kappa Pite˚a. smurfit kappa pite˚a, 2018. https://www.smurfitkappa.com/vHome/se/Kraftliner (Retrieved 2019-01-22).

[2] Smurfit Kappa Pite˚a. Products and services, 2018. https://www.smurfitkappa.com/products-and- services/displays/counter-displays (Retrieved 2019-01-28).

[3] Smurfit Kappa Pite˚a. Products and services, 2018. https://www.smurfitkappa.com/products-and-services/bag-in- box (Retrieved 2019-01-28).

[4] Smurfit Kappa Pite˚a. Company presentation [internal material], 2018. Intran¨atet (Retrieved 2019-01-28).

[5] Smurfit Kappa Pite˚a. Kraftliner fullbestruken (cote plus) (kch) - europa, 2018.

https://www.smurfitkappa.com/vHome/se/Products/Sidor/Heavy Coated Kraftliner Cote plus KCH Europe.aspx (Retrieved 2019-01-22).

[6] Sture ¨Oberg. Logistics director, January 2019. Project Specification.

[7] Teoh Guan Chengb Kenny Sambasivanc Murali Md. Sidind Samsinar Agha Kasiria, Leila. Integration of standardization and customization: Impact on service quality, customer satisfaction. Journal of Retailing and Consumer Services and loyalty, 35:91–97, 2017.

[8] Malhotra Naresh K Baalbaki, Imad B. Standardization versus customizations in international marketing: An investigation using bridging conjoint analysis. Journal of the Academy of Marketing Science, 23(3):182–194, 1995.

[9] Smurfit Kappa Pite˚a. Organization chart [internal material], 2018. Intran¨atet.

[10] Smurfit Kappa Pite˚a. Standardization memo [internal material], 2014. Internal Communication.

[11] D. Vrontis. Integrating adaption and standardisation in international marketing: The adaptstand modelling process.

Journal of Marketing Management, 2003(19):283–305, 2013.

[12] M Christopher. Logistics and Supply Chain Management. Pitman Publishing, 4th edition, 2011.

[13] C.M Drury. Management and Cost Accounting. Springer-Science+Business Media, B.V., 3rd edition, 1992.

[14] Anderson S.W. Maher M.W. Lanen, W.N. Fundamentals of Cost Accounting. McGraw-Hill Companies Inc., New York, 3rd edition, 2011.

[15] Aronsson H. Ekdahl B. Oskarsson, B. Mordern Logistik - för ökad lönsamhet. Liber AB, Stockholm, 4th edition, 2013.

[16] Nash S. G. Sofer A. Griva, I. Linear and Nonlinear Optimization. Society for Industrial and Applied Mathematics., Philadelphia, 2nd edition, 2009.

[17] Chan Wook A. Pant M. Ali, M. Trim loss optimization by an improved differential evolution. Mathematical Problems in Engineering, 2013:1–8, 2013.

[18] Hill Wong D. S. Jang S. S. Yen, C. H. Solution of trim-loss problem by an integrated simulated annealing and ordinal optimization approach. Journal of Intelligent Manufacturing, 15:701–709, 2004.

[19] Patricia Pulliam. Phillips and Cathy A Stawarski. Data collection : planning for and collecting all types of data.

Pfeiffer, San Francisco, 2008. E-book.

(35)

Appendix A Net Value

Figure 14: A description of how net invoiced and net value is calculated from order price.

Master’s Thesis in Industrial Engineering and Management, 30 credits Supervisor Ume˚a University: Victor Falgas Ravry External Supervisor: Sture Öberg, Smurfit Kappa Examiner: Gerold Jäger

Stock and Cuts in Piteå

Standardization vs. Customization

in the pulp and paper industry

Contents

Glossary

1 Introduction

1.1 Smurfit Kappa Pite˚ a

1.2 Background

1.3 The Assignment

1.4 Scope

1.5 Report Overview

2 Current State

2.1 Standardization

2.2 Sales

2.3 Production

2.4 Inventory

2.5 The Coating Unit

3 Theoretical Framework

3.1 Standardization

3.2 Cost Estimation

3.3 Stock Calculations

3.4 Optimization

4 Methodology

4.1 The General Approach

4.2 Data Collection

4.3 Modeling

5 Results

5.1 Simulations

5.2 Customer Restrictions

5.3 Model comparison

6 Analysis

6.1 Q1. What costs are affected by standardization and how can these be esti-

mated?

6.2 Q2. How can we model the outcomes of different standardization scenarios?

6.3 Q3. To what extent can customers switch to other articles in case of stan-

dardization, and what costs might this entail?

6.4 Validity and Reliability

7 Recommendations

References

Appendix A Net Value