Perishable goods production: Selecting maintenance policies in order to maximize

(1)

University of Groningen

Faculty of Economics and Business

Msc Technology and Operations Managment

Perishable goods production: Selecting maintenance policies in order to maximize

profit. A case study.

Author:

ing. J.V. Kluitenberg s2386224

Supervisor:

dr. N.D. van Foreest Co-Supervisor:

dr. ir. D. Catanzaro Company Supervisor:

dhr. E. Kreder

Leeuwarden, June 23, 2014 version: Msc-Thesis 1.0 (Final)

(2)

(3)

Abstract

The purpose of this report is to select a maintenance strategy which can be used in a multi-unit continuous production environment with perishable goods. Therefore, a comparison between the selection of a maintenance strategy for perishable and non perishable production is made. In order to select a maintenance strategy the Analytical Hierarchy Process is used. Data was obtained from a literature review and a case study at Friesland- Campina Bedum. The findings of this study show that there is no significant difference in the selection of a maintenance strategy for perishable or non perishable operations. These findings can be explained by the fact that in both cases a proactive maintenance strategy is preferred.

3

(4)

(5)

Preface

A final stage of the Msc Technology and Operations Management (and a seven year long career as a student) I hereby present my Msc-Thesis. The past half year I have conducted research in the field of maintenance management in cooperation with Royal Friesland- Campina and the Faculty of Business and Economics of the University of Groningen. The results of this research are presented in this document.

While writing this Thesis I have learned a lot about Multi-Criteria Decision Making Systems and Maintenance Management Strategies. This knowledge will hopefully contribute to a successful and promising career.

I would like to thank all the participants who contributed to this thesis. In particular I would like to thank Edwin Kreder and the maintenance department of FrieslandCampina for their valuable input and their help with collecting data. Furthermore, I would like to thank dr. Nicky van Foreest for his guidance during this project. Finally, I would like to thank my fellow students, family, and friends for their support over the last few years.

Thank you for reading.

Leeuwarden, June 2014 ing. J.V. Kluitenberg

7

(8)

(9)

Chapter 1 Introduction

Over the last decades the way we look at maintenance has significantly changed; from a time where maintenance was considered as a "necessary evil", to a time where all different kinds of maintenance strategies are being developed and it is being seen as a partnership within the company and beyond (Waeyenbergh and Pintelon2002). Nowadays, companies have to deal with the great pressure of reducing production costs (Wang, Chu, and Wu 2007). With maintenance costs that can be as high as 70 percent of the production costs (Bevilacqua and Braglia2000), there seems to be an opportunity to cut down costs in order to maximize profit. However, choosing the right strategy can be difficult because of all the different aspects involved. Moreover, consequences of inefficient strategies can exceed far beyond the direct costs of maintenance (Al-Najjar and Alsyouf 2003).

Many types of maintenance strategies are described in literature. Kothamasu, Huang, and Verduin (2006) divide these strategies into two categories. Unplanned maintenance (i.e. reactive) where maintenance is performed after a defect or breakdown (e.g corrective maintenance), and planned maintenance (i.e. proactive) where maintenance is being performed before a defect or breakdown. Swanson (2001) describes a third strategy, aggressive maintenance, which seeks to improve overall equipment operation. Furthermore, there are different types of methods to determine what kind of strategy to use. Using fuzzy multiple criteria decision-making (MCDM) technique, Al-Najjar et al. (2003) try to determine the most efficient maintenance strategy for an oil company. Moreover, Wang et al. (2007) show that a mix of different maintenance strategies can improve the reliability and availability of production equipment using the Analytical Hierarchy Process (AHP). Cho and Parlar (1991) describe different maintenance strategies which can be used for multi-unit systems (i.e. a production line with multiple subsequent stages) using a survey. Finally, Rao and Bhadury (2000) show in their case study that, within a multi-unit system, opportunistic maintenance strategies are most appropriate.

However, none of these articles take into account the perishability of products within the production line when selecting the optimal maintenance strategy, even though perishability of products can be an important factor when performing maintenance on a production line.

The goal of this research is to develop a framework to select a maintenance strategy in a multi-unit series production environment with perishable goods. To do so, a case study has been performed at the FrieslandCampina Bedum cheese factory, where (due to perishability) products need to be processed within a specified amount of time. The result of this study will be a maintenance program which can be used in a multi-unit continuous production environment with perishable goods in order to increase up-time.

Therefore, this research will look at different maintenance strategies that are available in literature and the criteria that can be used in order to make a selection. The Analytical Hierachy Process (AHP) will be used in order to select a strategy. Furthermore, this research will look at the influence of perishability on the decision process.

9

(10)

Chapter 2 discusses the theoretical background by providing details about maintenance strategies and selection methods. Chapter 3 gives a description of the used methods.

Chapter 4 provides e the analysis and the selection framework. The results of the analysis are discussed in Chapter 5, after which the conclusion is presented in Chapter 6.

10 CHAPTER 1. INTRODUCTION

(11)

Chapter 2 Theoretical Background

In this chapter the different maintenance strategies are outlined in section 2.1. Different methods in order to select a strategy and/or strategies will be discussed in section 2.2.

Section 2.3 addresses several maintenance selection criteria. Section 2.4 deals with the type of production environment. Finally, a conceptual model can be found in section 2.5.

2.1 Maintenance strategies

Kothamasu et al. (2006) divides maintenance into two categories, reactive maintenance and proactive maintenance. Swanson (2001) describes a third strategy, aggressive maintenance.

Figure2.1gives an overview of the different maintenance strategies that are being discussed in this chapter.

Reactive maintenance

Reactive maintenance is performed when a machine breaks down (Kothamasu et al.2006).

There are two types of reactive maintenance, corrective maintenance (where the intention is to restore the machine) and emergency maintenance that needs to be performed in order to avoid serious consequences to this and other machines. Reactive maintenance can also be seen as unplanned maintenance and is done in an ad-hoc manner. The cost of reactive maintenance can be relatively high, because the failure can cause unforeseen damage and it is difficult to plan personnel and the ordering of spare parts, since it is unknown when they are needed (Garg and Deshmukh 2006). However, the data obtained from corrective maintenance can provide valuable data for the prediction of failure behaviour (Garg et al.

2006), which can be used in order to optimize and choose maintenance strategies.

Proactive maintenance

Proactive maintenance is performed in order to avoid breakdowns by undertaking repair activities to restore equipment into a proper condition (Swanson 2001). There are four types of proactive maintenance strategies. Preventive maintenance is being performed on a machine in order to prevent it from breaking down (Bevilacqua et al.2000). According to Kothamasu et al. (2006), preventive maintenance can be done at constant time intervals or based on the age of the machine (e.g. no maintenance is being performed in the first year).

Opportunistic maintenance can be performed when a machine has to stop because another machine is in maintenance, or when maintenance needs to be performed on a specific part of the machine. When this happens, all other maintenance activities that have to be done in the near future can also be done so that the machine does not have to be stopped again.

When using condition based maintenance (CBM) the system is measured and monitored, and maintenance is being performed when a machine or part reaches a certain condition

11

(12)

Maintenance

Reactive

Corrective

Emergency Proactive

Preventive

Time based Age based

Opportunistic Condition

based Predictive

Aggressive

TPM Other

Figure 2.1: An overview of different maintenance strategies

(Bevilacqua et al.2000; Kothamasu et al.2006). Predictive maintenance is also based on the condition of the system. But now it is forecast when the machine will reach a preselect condition (Bevilacqua et al. 2000) which allows for better planning of the maintenance activities.

Using proactive maintenance strategies can reduce maintenance cost by avoiding un- controlled damage and downtime (Garg et al.2006). However, El-Haram (1995) indicates that there are also some disadvantages to proactive maintenance. To keep on the safe side a number of unnecessary tasks will be performed. Due to human failure, equipment may suffer unnecessary damage. Moreover, planned maintenance tasks can be demanding in the amount of labour and the number of spare parts that are needed. However, these disadvantages can be brought to a minimum by determining the optimal maintenance strategy.

Aggressive maintenance

Aggressive maintenance strategies focus on improving the design and function of production equipment. This type of strategy demands a high amount of resources, commitment, integration and training, but can also lead to significant improvements (Swanson 2001).

An example of an aggressive maintenance strategy is total productive maintenance (TPM).

In using TPM several project groups integrate the relation between maintenance and production. In doing so production workers become involved in maintenance activities and become able to monitor the condition of equipment. Also, maintenance prevention teams are constantly working on improvements on installations in order to prevent maintenance (Swanson 2001).

Other maintenance strategies

In addition to the ones discussed above, there are several other maintenance strategies.

The outsourcing of maintenance is a strategy which can be used to improve the quality of maintenance due to technological advantages. Also, it can result in lower costs due to a lower number of employees in the maintenance department, and the flexibility can be increased (Martin 1997). Furthermore, Garg et al. (2006) show that maintenance is becoming a multiple disciplinary task which involves the integration of techniques like TPM

12 CHAPTER 2. THEORETICAL BACKGROUND

(13)

and RCM (reliability centred maintenance). Also, new maintenance strategies are emerging which are based on this integration of strategies (e.g. risk based maintenance (RBM) and effectiveness centred maintenance (ECM)). Using RBM, the probability of failure is set off against the consequences of failure and is used in order to minimize hazards to humans and the environment (Khan and Haddara2003). ECM concerns itself with performing the right tasks and mainly focuses on customer service and quality improvement (Garg et al.

2006).

2.2 Maintenance strategy selection methods

In literature, different methods are used to select maintenance strategies. Some of these strategies will be discussed below. Many optimization models have been developed (e.g.

simulation, multiple criteria decision making (MCDM) (Al-Najjar et al. 2003) and mixed integer linear programming (MILP) (Ashayeri, Teelen, and Selenj 1996)) and are used in order to select maintenance strategies. According to Garg et al. (2006), all of these methods generally include the following phases:

1. a description of a technical system, its function and importance;

2. a modeling of the deterioration of the system through time and possible consequences for this system;

3. a description of the available information about the system and actions open to management; and

4. an objective function and an optimization technique which helps finding the best balance.

The outcome of these phases is generally a maintenance program. A good maintenance program defines different strategies for different machines (Bevilacqua et al.2000) in order to get an optimal economic result. The effectiveness and accuracy of maintenance policies can be used to determine its efficiency (Al-Najjar 1997).

The Analytical Hierarchy Process (AHP)

Garg et al. (2006) provides an overview of a series of studies in which different maintenance methods are used and compared. Looking at these studies the best method for this research appears to be the analytical hierarchy process as used by Bevilacqua et al. (2000). In their study they are using this method to select the optimum maintenance program for more than 200 units in an oil refinery. In doing so, they have created a framework that can be used for similar facilities within the company. By using a criticality index (CI), the optimum strategy can be chosen based on preselected maintenance criteria. Other methods described by Garg et al. (2006) are more complex, or are aimed at the design stage of a production process. Studies which used a simulation type method where used in order to reduce maintenance and spare parts costs. Since this study will only look at maintenance strategies, the AHP method is used. Moreover, this method is also chosen because it is used in studies similar to this and has been proven to provide satisfying results.

2.3 Maintenance selection criteria

Depending on the production environment, different criteria are used in order to determine a maintenance strategy. Bevilacqua et al. (2000) point out several criteria in their case study which can be used in the selection of a strategy:

2.2. MAINTENANCE STRATEGY SELECTION METHODS 13

(14)

• Safety

• Maintenance importance for the process

• Maintenance cost

• Failure frequency

• Downtime length

• Operating conditions

• Machine access difficulty

• Spare part availability

Depending on the situation, different criteria will be selected. During this case study, interviews were conducted in order to determine the criteria that have to be considered in the case of perishable goods. These interviews were semi-structured with open questions.

The interviewees were employees of the maintenance department.

2.4 Type of production environment

Little literature can be found on the type of products of the production process which needs to be maintained. However, in the case of FrieslandCampina Bedum perishability is an important factor, because the milk which is delivered at the plant needs to be processed within a certain amount of time. Disruptions within this process can cause planning problems, and milk may need to be allocated to other locations which can result in capacity problems. Furthermore, several stages within the production process are not allowed to be idle for more than 15 minutes due to quality constraints. If these 15 minutes are exceeded, the products require an additional quality inspection. Depending on the amount of time the machine has been idle, this can lead to the rejection of products and additional disposal costs.

2.5 Conceptual model

Figure2.2shows the conceptual model of this research. Using a strategy selection method, the most suitable maintenance program for different types of operations can be determined.

The overall goal of the research is profit maximization. A higher throughput yields more profit, as all products that are being produced can be sold. In this study looked at the maintenance program in order to maximize throughput. When selecting maintenance strategies for the maintenance program, input criteria such as discussed in selection 2.3 were used. As mentioned, different methods and strategies are discussed in literature but none take into account the perishability of products. Therefore, this figures separately in the model.

14 CHAPTER 2. THEORETICAL BACKGROUND

(15)

Maintenance strategy selection

Maintenance program Throughput Profit maximization

Perish-ability

Input criteria and parameters

(-) More complex (+) More accurate

(+) High throughout yields more profit

It is important that the right parameters are chosen

(+) a good maintenance program will increase througput

Figure 2.2: Conceptual model

2.5. CONCEPTUAL MODEL 15

(16)

(17)

Chapter 3 Methodology

For this research, a case study was performed at FrieslandCampina Bedum to gather the data which was needed to determine a maintenance strategy. The first part of this chapter discusses the methods that were used for the selection of a maintenance strategy. The second part of this chapter discusses the data collection method.

3.1 Used methods

This research focuses on two machines within the cheese production line. The first machine which was selected operates at the beginning of the production process. At this stage the products within the process are highly perishable (e.g. 15 minutes). The second machine that was selected is operative at the end of the process where the perishability of the products are negligible (e.g. weeks or months). This way a comparison can be made between maintenance strategies for machines for perishable and non perishable goods. In order to select a maintenance strategy for these machines this research goes through the stages described in table 3.1. The results of the first analysis will be a list of maintenance strategies which can be used in the two cases which were selected from the process.

Phase Action Method

First Provide an overview of the available maintenance strategies and the conditions which are involved;

Literature review

Second Provide an overview of the criteria that a strategy should meet for the two specific cases;

Interviews Third Compare the available strategies based on selec-

tion criteria and select the most suitable for each case.

Setting up framework using the analytical hierarchy process

Table 3.1: Phases in order to create framework

The second stage was choosing the selection criteria which were needed for the analytical hierarchy process (AHP). These selection criteria were determined by conducting interviews. The result of this stage was a list of criteria which was used in order to select a maintenance strategy.

The third stage provided a more detailed decision making process by using the analytical hierarchy process (AHP). In doing so, the criticality of different components within the selected machines were taken into account. The AHP consisted of the following steps:

17

(18)

1. Determining relevant criteria;

2. Selection of maintenance strategies to be compared;

3. Making a hierarchy structure of the decision making problem. (i.e. mapping how the input parameters are affected by each other);

4. Weighting the input parameters against each other in order to determine the criticality;

5. Making the AHP framework;

6. Performing the analysis.

The outcome of the analysis was a score for the analyzed maintenance strategies. This score was used to select a maintenance strategy.

3.2 Data collection

The data was collected via three means. First, a literature review was performed (as discussed in chapter 2) Then, the data which was needed for the AHP framework was collected through a case study at FrieslandCampina Bedum. Furthermore, semi-structured interviews were held with employees of the maintenance department. Finally, there has been made use of data provided by the mainenance department of FrieslandCampina Bedum.

This data concerned maintenance tasks, failure and repair behaviour, and the costs involved in downtime and product waste.

18 CHAPTER 3. METHODOLOGY

(19)

Chapter 4 Analysis

This chapter describes the analysis peformed for this study. First, the maintenance selection criteria which are needed for the Analytical Hierarchy Process are discussed. Second, the Analytical Hierarchy framework and the manner of calculating the criticality of the different criteria is described. Finally, different maintenance strategies are compared, using the AHP Matrix.

4.1 Setting up the framework

To determine the criteria for selecting a maintenance strategy, we conducted interviews at FrieslandCampina with personnel of the maintenance department. Furthermore, we conducted a literature interview and used the criteria discussed in section 2.3 as input.

Based on this, we indicated 8 criteria that were used in the Analytical Hierarchy Process.

We limited the number of criteria to 8 to keep the number of comparisons limited to 28.

These eight criteria are discussed in section 4.1.1. The maintenance strategies that are used in the comparison are discussed in section 4.1.2.

4.1.1 Maintenance selection criteria Safety

Safety concerns the safety of personnel, the environment, the production equipment and consumers when a failure occurs. In case of a failure, personnel may be injured through the failure itself or during maintenance activities. Also, there can be an impact on the environment or even on public health (i.e. consumers) when there is a failure which concerns quality. Safety is considered because it is highly valued by FrieslandCampina.

Maintenance importance to process

The maintenance importance to the process indicates the influence a machine has on the performance of the production line. When downtime of a machine also leads to downtime of other machines the maintenance importance becomes bigger. The maintenance importance to the process is considered because of the muli-unit series production character of the production line.

Maintenance costs

Next, the costs associated with preforming certain maintenance tasks (e.g. personnel, downtime cost and spare parts costs) are considered. While taking into account the costs of maintenance, we similarly have to keep in mind that it is often hard to quantify the

19

(20)

financial benefits of maintenance. The maintenance cost are considered because of the high costs of downtime and spare parts.

Failure frequency

The failure frequency is linked to the mean time between failures (MTBF) (Bevilacqua et al.2000). This criteria is considered because a higher failure frequency will result in a more disruptive production process.

Downtime length

The downtime length is linked to the mean time to repair (MTTR) (Bevilacqua et al.

2000) and is considered because of the perishability of the products and the influence of the downtime on the rest on the production process.

Production loss costs

The production loss costs are the costs caused by the inability to continue production or production related activities (e.g. through cleaning the equipment). Furthermore, the production loss costs are the costs caused by the waste of products due to quality issues or perishability. The production loss costs are considered because of the perishability of the goods involved.

Machine access difficulty

The machine access difficulty indicates the amount of effort needed to preform maintenance tasks to the equipment. The amount of effort can be high when certain operations need to be done at places that are hard to reach, or when they need to be performed within specific care zones (e.g. special clothing and tools need to be used in certain parts of the factory). The machine access difficulty is considered as this difficulty was expected in the maintenance of the machines studied.

Spare part availability

The availability of spare parts influences the criticality of the machine, since a lack of spare parts increases the impact of breakdowns. This criteria is considered because spare parts are not always available.

As expected, many of the criteria indicated by the maintenance department of Friesland- Campina are already described in literature. When looking at the perishability of the product, the interviewees indicated that unexpected downtime length (i.e. MTTR) plays an important role on product quality. When a failure takes more than 15 minutes, the quality of the products currently on the production line need to be inspected. In the worst case scenario, the production line cannot continue and all the machines need to be emptied and cleaned by hand, wasting a lot of valuable products and increasing labour costs.

Furthermore, the mean time between failures also is important since unexpected downtime also influences the time available for repairs. When the maintenance tasks that have to be performed are not critical, the production can continue its current batch before the maintenance is performed. This way, the critical time of 15 minutes can be exceeded without affecting the product.

20 CHAPTER 4. ANALYSIS

(21)

4.1.2 Maintenance strategies

Based on the interviews and literature review, we chose three different maintenance strategies for our comparison. We limited the number of strategies to three in order to keep the number of comparisons limited to 24.

Condition Based Maintenance

Condition Based Maintenance (CBM) was chosen because this strategy describes best what the current maintenance procedures are. In the current situation maintenance staff or operators perform inspection rounds during operations. By looking and listening to equipment parts that are in a condition other than normal (and, therefore, need to be replaced) are searched for. Depending on the condition of the parts, the maintenance department determines the time when the maintenance task needs to have been performed.

Together with the production planners and/or operators it is determined when a time-slot is available to perform the maintenance.

Corrective Maintenance

The second maintenance strategy we chose is Corrective Maintenance. This maintenance strategy was chosen because FrieslandCampina Bedum plans to employ this strategy on one of the production installations in the near future. The reason for this is that when corrective maintenance is used, equipment can run without stopping until a failure occurs.

In doing so, the production time of the machine is stretched as long as possible.

Time Based Maintenance

The third maintenance strategy is Time Based Maintenance (TBM). This maintenance strategy is a basic preventive maintenance strategy, which is used in many industries (Ahmad and Kamaruddin 2012). TBM was chosen because it can be easy to use when failure behavior of parts is available. Moreover, it can provide an alternative for situations where monitoring of equipment is not possible.

4.2 The Analytical Hierarchy Framework

After selecting the maintenance selection criteria and the maintenance strategies the hierachy structure of the AHP was determined, as shown in 4.1. Using pairwise comparison all the section criteria are compared one on one using the weight scale of Saaty and Vargas (2001) which is shown in table 4.1.

To perform the pairwise comparison analysis we conducted a survey among employees of the maintenance department. The survey used can be found in appendix A.

4.3 Preforming the analysis Level 1

Table 4.2 shows the results of the pairwise comparison. By performing an inconsistency check on the matrix we were able to identify that the matrix did not meet the consistency requirement of being ≤ 0.10 (which, according to Saaty et al. (2001) is necessary). There- fore, we had to correct the data by altering the most inconsistent values in the matrix in order to meet the ≤ 0.10 constraint (Saaty et al. 2001). The results of this correction are given in table 4.3. In order to determine the criticality of the selection criteria we calculated the normalized weight of the matrix in table 4.4by dividing each element in the matrix with the sum of its column.

4.2. THE ANALYTICAL HIERARCHY FRAMEWORK 21

(22)

Maintenance selection

Maintenance MTBF cost Maintenance

importance to process

Safety MTTR Production

loss cost

Machine acces difficulty

Condition based maintenance Corrective

maintenance

Time based preventive maintenance

Figure 4.1: The analytical hierarchy process

Using the normalized matrix we can calculate the Criticality Index (CI) by calculating the avarage weigth in each row. The corresponding criticality is given in table4.5

4.4 Preforming the analysis Level 2

Level 2 of the analysis contains a pairwise comparison between the three maintenance strategies and the eight criteria. The level 2 comparisons are rated based on literature and are explained in this section. An overview of this comparison including the normalized and idealized scores can be found in table4.6.

Safety

Corrective Maintenance can cause safety and health hazards due to unexpected breakdowns (Tsang 1995), while Preventive maintenance strategies increase safety and reliability of equipment. Therefore, Corrective Maintenance yields a negative score compared to the other strategies.

The conomatic is a bottleneck in the production process and, therefore, of great importance to the process. Since Corrective Maintenance can cause unexpected downtime (Garg et al.

2006), the maintenance importance to the process yields in a negative score. Both Time Based Maintenance and Condition Based Maintenance can be planned and, therefore, yield an equal score for maintenance importance to the process.

Maintenance costs

Preventive maintenance costs are lower in terms of manpower and spare parts (Charles, Floru, Azzaro-Pantel, Pibouleau, and Domenech 2003). Corrective Maintenance can lead to high costs due to lost production, secondary damage and restoration to operation conditions (Tsang1995). Therefore, Corrective Maintenance yields a negative score compared to the other strategies. Because Time Based Maintenance can cause unneccesarily perfomed maintenance (El-Haram 1995), it yields a negative score compared to Condition Based Maintenance.

(23)

Intensity of importance

Definition Explanation

1 Equal importance Two activities contribute equally to the

objective

2 Weak

3 Moderate importance Experience and judgment slightly favor

one activity over another

4 Moderate plus

5 Strong importance Experience and judgment strongly favor

one activity over another

6 Strong plus

7 Very strong or demonstrated importance An activity is favored very strongly over another; its dominance demonstrated in practice

8 Very, very strong

9 Extreme importance The evidence favoring one activity over an-

other is of the highest possible order of af- firmation

Reciprocals of above non-zero

If activity i has one of the above nonzero numbers assigned to it when compared with activity j, then j has the reciprocal value when compared with i

A reasonable assumption

Rationals Ratios arising from the scale If consistency were to be forced by obtain- ing in numerical values to span the matrix

Table 4.1: AHP Score table (Saaty and Vargas 2001)

Criteria Safety Maintenance

Maintenance cost

Failure frequency

Downtime length (MTTR)

Production loss cost

Machine acces difficulty

Safety 1,00 7,00 7,00 7,00 7,00 7,00 7,00 7,00

Maintenance importance to process

0,14 1,00 1,00 0,33 0,33 0,33 1,00 0,17

Maintenance cost 0,14 1,00 1,00 1,00 0,50 0,25 3,00 0,33

Failure frequency 0,14 3,00 1,00 1,00 1,00 2,00 3,00 0,33

0,14 3,00 2,00 1,00 1,00 3,00 7,00 5,00

Production loss cost 0,14 3,00 4,00 0,50 0,33 1,00 0,33 0,33

Machine acces difficulty 0,14 1,00 0,33 0,33 0,14 3,00 1,00 0,25

Spare part availability 0,14 6,00 3,00 3,00 0,20 3,00 4,00 1,00

Total 2,00 25,00 19,33 14,17 10,51 19,58 26,33 14,42

Table 4.2: AHP Pairwise comparison matrix (Inconsistency of .17)

Failure frequency

Preventive maintenance strategies reduce the failure frequency of equipment (Ahmad et al.

2012). Therefore they yield a positive score compared to Corrective Maintenance.

Downtime length

Corrective Maintenance strategies can lead to a high amount of downtime (Ahmad et al.

2012). Therefore it yields a negative score compared to the other strategies.

Production loss costs

Corrective Maintenance strategies can lead to unexpected production loss time (Ahmad et al.2012). Although other maintenance strategies also cause production loss time, these times are often known beforehand. At FrieslandCampina production loss costs are also made when products are rejected by quality control. This can happen when the production line has to stop for more than 15 minutes (e.g. in the case of preforming corrective

4.4. PREFORMING THE ANALYSIS LEVEL 2 23

(24)

Criteria Safety Maintenance importance to process

Maintenance cost

Failure frequency

Safety 1,00 7,00 7,00 7,00 7,00 7,00 7,00 7,00

0,14 1,00 1,00 0,33 0,33 0,33 1,00 0,25

0,14 3,00 2,00 2,00 1,00 2,00 3,00 2,00

Total 2,00 23,00 19,33 14,67 10,67 18,67 21,33 11,58

Table 4.3: AHP Pairwise comparison matrix corrected for inconsistency (Inconsistency of .10)

Criteria Safety Maintenance

Maintenance cost

Failure frequency

Safety 0,50 0,35 0,36 0,48 0,63 0,42 0,36 0,54

0,07 0,05 0,05 0,03 0,05 0,03 0,05 0,02

0,07 0,10 0,10 0,07 0,09 0,12 0,21 0,23

Table 4.4: AHP normalized matrix

maintenance). Therefore, Corrective Maintenance yields a negative score compared to the other two strategies. Since both Condition and Time Based Maintenance are planned in advance the production loss costs are considered equal.

Machine access difficulty

When maintenance is being performed, the physical access to the machine is similar for the three maintenance strategies. When inspections need to be carried out, however, the machine access difficulty can become harder for certain parts. For these parts the Time Based Maintenance strategy is slightly favourable compared to Condition Based maintenance. Since Corrective Maintenance requires immediate access to a machine, due to the criticality of the failure, this strategy has the worst maintenance access difficulty.

Time Based Maintenance needs to be planned in cooperation with the production planning department, which sometimes causes difficulties. Therefore, we rated this inbetween the two other strategies.

When maintenance is performed using a maintenance schedule, the overall costs of spare parts can be 10% to 30% less (Charles et al.2003) compared to performing maintenance without a schedule. Furthermore, scheduled maintenance provides the opportunity to order parts when they are needed instead of keeping them in stock. Because spare parts are more critical with Corrective Maintenance this yields a negative score compared to the other strategies. When looking at Condition Based Maintenance, the spare part availability depends on the lead time of the spare parts and the time frame in which the maintenance needs to be performed. Since the time frame with Time Based Maintenance is already known this strategy yields a slightly higher score.

(25)

Criteria Critically (CI)

Safety 0,45

Maintenance importance to process 0,04

Maintenance cost 0,05

Failure frequency 0,09

Downtime length (MTTR) 0,12

Production loss cost 0,07

Machine acces difficulty 0,05

Spare part availability 0,13

Table 4.5: AHP Critically Index

4.5 AHP Results

The results of the Analytical Hierarchy Process can be seen in table4.7. In order to prevent dominance of a single factor we calculated a normalized and an idealized score (Saaty et al.

2001). Both the normalized and the idealized scores show that Time Based Maintenance is the best solution according to the AHP, followed by Condition Based Maintenance. The Corrective Maintenance strategy scores significantly lower than the preventive maintenance strategies and is, therefore, considered least suitable.

In order to test the sensitivity of the model, we changed several values to see whether this would affect the outcome. When lowering safety to 0,20 and dividing the rest of the score between the other criteria, the values of the model changed but the outcome remained the same.

4.6 Reference framework

In order to compare a machine that processes highly perishable goods with one that processes lowly perishable goods we also conducted a pairwise comparison survey for a machine at the packaging department. The resulting criticality index is more or less similar to that found in table 4.8. During the interviews we noticed that maintenance priority is higher for the Conomatic. These findings were also supported by the survey, in which the criteria are compared one on one (i.e. safety on machine A vs safety on machine B). Except for safety, all criteria are considered less important for machine B. However, since the relations among these factors are more or less the same this results in an almost identical score.

4.5. AHP RESULTS 25

(26)

* The better the strategy the higher the score Corrective maintenance

Condition based maintenance

Time based preventive maintenance

Normalized Idealized

Safety

Corrective maintenance 1,00 0,25 0,33 0,13 0,28

Condition based maintenance 4,00 1,00 1,00 0,46 1,00

Time based preventive maintenance 3,00 1,00 1,00 0,42 0,91

Maintenance cost

Failure frequency

Machine acces difficulty

Table 4.6: AHP Matrix level 2

- Safety Maintenance

Maintenance cost

Failure frequency

Score

Criticality 0,45 0,04 0,05 0,09 0,12 0,07 0,05 0,13

Normalized

Corrective maintenance 0,13 0,12 0,16 0,09 0,14 0,09 0,54 0,10 0,14

Condition based maintenance 0,46 0,32 0,59 0,45 0,43 0,45 0,16 0,33 0,42

Time based preventive maintenance 0,42 0,56 0,25 0,45 0,43 0,45 0,30 0,57 0,43

Idealized

Corrective maintenance 0,28 0,22 0,27 0,20 0,33 0,20 1,00 0,17 0,29

Condition based maintenance 1,00 0,57 1,00 1,00 1,00 1,00 0,30 0,59 0,89

Time based preventive maintenance 0,91 1,00 0,43 1,00 1,00 1,00 0,55 1,00 0,90

Table 4.7: AHP Score overview

(27)

Criteria Machine A Machine B (CI)

Safety 0,45 0,46

Maintenance importance to process 0,04 0,04

Maintenance cost 0,05 0,05

Failure frequency 0,09 0,08

Downtime length (MTTR) 0,12 0,12

Production loss cost 0,07 0,07

Machine acces difficulty 0,05 0,05

Spare part availability 0,13 0,12

Table 4.8: AHP CI Comparison

4.6. REFERENCE FRAMEWORK 27

(28)

(29)

Chapter 5 Discussion

5.1 Strategy selection

Using the Analytic Hierarchy Process, we assigned scores to three different maintenance strategies. There is a significant difference between the Corrective Maintenance score and that of the two preventive maintenance strategies. Even though Time Based Maintenance scored highest, this does not necessarily mean that all maintenance should be performed in this manner. Since both Time and Condition Based Maintenance scored high on the AHP, we think a combination of both strategies will result in a good maintenance program. The choice between the two strategies should be made based on the parts of the machine that will be in need of maintenance. Ahmad et al. (2012) also discusses the combination of these strategies. In order to perform Time Based Maintenance, the expected lifetime of certain parts needs to be assumed. While conducting interviews, we noticed that manufacturers are mostly disinclined to give an indication of the lifetime of the parts they produce. In such cases, the expected lifetime of parts can be estimated through user based data (Lee, Ni, Djurdjanovic, Qiu, and Liao 2006). However, this causes the strategy to be suitable only for parts of which the failure behaviour is known, or parts of which the failure behaviour can be estimated without this data.

According to Bloch and Geitner (1983), 99% of failures are indicated by signs before the actual failure occurs. In combination with the high score on the AHP, this makes Condition Based Maintenance an attractive strategy. In order to preform Condition Based Maintenance continuous monitoring of equipment is required Ahmad et al. (2012), which is often done by sensors or other types of indicators (Campos 2009). As witnessed at FrieslandCampina Bedum, much of the monitoring is done by operators and maintenance staff by looking and listening to equipment. In some cases Time Based Maintenance is more suitable, since it is not always possible to manually monitor the condition of specific parts. In order to determine which parts of a machine need to be maintained by using either a Time Based or Condition Based maintenance strategy a further analysis should be performed. Within this analysis the ability and reliability of the monitoring of the part should be determined in order to select the best strategy (e.g. when monitoring is difficult, the Time Based strategy might be more suitable than the Condition Based strategy).

5.2 The influence of perishability

Using the Analytical Hierarchy Process this study has not found any differences in the selection of a maintenance strategy for perishable or non perishable goods. Although the costs of rejected products can be much higher when the product is perishable, both machines suffer similarly from unexpected downtime. During the interviews we conducted we did see that maintenance at the beginning of the process (when the product is highly

29

(30)

perishable) has much higher priority than maintenance at the end of the process (when the product is less perishable). However, this does not change the selected maintenance strategy in order to get the highest up-time of equipment.

When looking at the results of the APH, this lack of difference between machine A and B could have been caused by the factor safety as this almost counts for 50% in the framework. Although there is a difference in perishability for both machines, this only plays a part when there is unexpected downtime. When comparing the other criteria, both processes had similar conditions which were weighted on a similar scale. While no difference was found between the two best strategies, there is a difference between the scores of Corrective Maintenance on the two separate machines. Machine B receives a better score for Corrective Maintenance but still only reaches less than half of the scores other two strategies.

30 CHAPTER 5. DISCUSSION

(31)

Chapter 6 Conclusion

Using the Analytical Hierarchy Process, this research has selected a maintenance strategy (mix) for the Conomatic machine at FrieslandCampina Bedum. Through a combination of both Condition and Time Based Maintenance we think the highest amount of up time is possible, while keeping in mind all the selected criteria. By applying the AHP to multiple machines (which processed either perishable or non perishable goods) we did not find any significant changes in the values of the models with the highest scores. Comparing the lower scores we noticed that Corrective Maintenance is less suitable for perishable production than for nonperishable production. However, in both cases this strategy is not the favored option. Therefore, based on this case study we can conclude that there is no significant difference between a perishable and non perishable production concerning the selection of maintenance strategies. In order to maximize profit we recommend FrieslandCampina Bedum not to switch from Condition Based Maintenance to Corrective Maintenance, as this change would cause an estimated cost increase of 30.000 to 50.000 due to the rejection of products and other negative effects of unplanned downtime.

In order to improve the Corrective Maintenance activities more time can be spend on monitoring the condition of equipment. Furthermore, digital monitoring equipment can be used for parts that are hard to inspect. We also encourage the maintenance department to train operators in monitoring equipment, for in this way more frequent inspections can be made. When monitoring is impossible, we recommend performing Time Based Maintenance.

31

(32)

Perishable goods production: Selecting maintenance policies in order to maximize

Perishable goods production: Selecting maintenance policies in order to maximize

profit. A case study.

Abstract

Contents

Preface

Chapter 1

Introduction

Chapter 2

Theoretical Background

Chapter 3

Methodology

Chapter 4

Analysis

Chapter 5

Discussion

Chapter 6

Conclusion