Assessing the impact of maintenance on the production costs

(1)

Assessing the impact of maintenance on

the production costs

Nils Plantinga S3442586

MSc Thesis Technology and Operations Managemen t Faculty of Economics and Business

First supervisor: Dr. B. de Jonge Second supervisor: Dr. Ir. S. Fazi

(2)

Page 2 of 75

Abstract

In current competitive markets, companies must constantly search for ways to differentiate themselves and to increase profits. One possibility to increase the profits is by decreasing the production costs. Because maintenance is considered as a substantial part of the total operating costs and due to its interconnections with other cost aspects of a production facility, improvements in maintenance can have a large impact on the total production costs. However, it is often difficult to justify investments in maintenance improvements because it is difficult to see the financial impact. Both theory and practice struggle with the impact of maintenance on the production costs.

In this research, we extensively investigated how the characteristics of the maintenance policies influences the impact on the production costs. Furthermore, we constructed an assessment model that can be used to assess the impact of maintenance on the production costs. This assessment model includes a framework that distinguishes the three most popular maintenance policies, purely corrective, time-based, and condition-based maintenance, by their impact on the production costs. This gives production facilities a tool that can be used to benchmark their current situation and justify investing in a specific maintenance policy. The model and framework are constructed by means of a single-case study consisting of interviews and company data of the Philips shaver production facility in Drachten. The working of the assessment model is illustrated by a case example, based on a situation at Philips Drachten.

(3)

Page 3 of 75

Acknowledgements

(4)

Page 4 of 75

List of figures

Figure 1: Types of maintenance ... 10

Figure 2: Maintenance cost, savings and profits model (Al-Najjar & Alsyouf, 2004) ... 12

Figure 3: Model that illustrates the impact of maintenance on a firms’ profits ... 14

Figure 4: Conceptual model ... 21

Figure 5: Example of typical shavers produced at Philips Drachten ... 24

Figure 6: Inventory avoiding and hiding problems (Nicholas, 2011)... 30

Figure 7: Production costs ... 37

Figure 8: Cost-scoring system ... 40

Figure 9: Importance scoring system ... 40

Figure 10: Assessment model ... 47

Figure 11: Key maintenance performance indicators (Muchiri, et al., 2010). ... 52

List of tables

Table 1: Comparison of time-based and condition-based maintenance (Ahmad & Kamaruddin, 2012) ... 18

Table 2: Structure of the framework (numbers serve as an example) ... 23

Table 3: The framework ... 41

Table 4: Example: fixed parameters and costs. ... 56

Table 5: Example: variable parameters and costs, current situation. ... 56

Table 6: Example: production costs, current situation. ... 57

Table 7: Example: estimation with framework. ... 57

Table 8: Example: variable parameters and costs, old vs new situation. ... 58

(6)

Page 6 of 75

1. Introduction

The customer of today has the luxury to choose between different companies for almost all the products he or she needs. To survive in such a competitive market, companies must differentiate themselves and constantly search for ways to increase profits in order to stay competitive. Having an efficient and effective production process is essential in this because it lowers the production costs per unit, which makes it possible to either decrease selling prices or increase profits and, therefore, stay competitive.

The existence of such a competitive market forces companies to improve internal processes in order to reduce costs. Improving maintenance can have a large impact on the total production costs, due to the substantial part of the total costs and because it is linked with many other cost factors. Especially a preventive maintenance policy enables a more efficient, effective, and reliable production process, which contributes to decreasing wastes such as waiting and inventory (Nicholas, 2011; Lee & Cha, 2016). Within a preventive maintenance policy, a maintenance action is planned in advance to prevent a system from breaking down and to improve the systems’ reliability (Nicholas, 2011; Lee & Cha, 2016). Preventive maintenance can be distinguished in time-based maintenance and condition-based maintenance. In a time-based maintenance policy, maintenance is performed at a certain time interval, while in a condition-based maintenance policy a specific parameter, or set of parameters, is monitored which triggers maintenance when a certain condition is reached. The other type of maintenance is corrective maintenance in which a system is only repaired if it is broken.

If the improvement of internal processes require investments, it is important that these investments can be justified. This also holds for maintenance investments. To justify maintenance investments, the impact on the production costs should be assessed to explore whether the investment is profitable. However, companies struggle with assessing the impact due to the complex interactions with other working areas such as quality, inventory, and production (Al-Najjar, 2007).

(7)

Page 7 of 75 There are a few studies that address the assessment of maintenance. First, Al-Najjar & Alsyouf (2004) created a model for maintenance costs, savings and profits to identify, monitor and improve the economic impact of maintenance. The second study contributing to this topic is the study of Alsyouf (2007), who gives a general formula to determine the difference in profits resulting from maintenance activities. At last, the study of Al-Najjar (2007) resulted in a model that exposes the technical and financial impact of vibration-based maintenance. Besides the studies about assessing maintenance, several studies address the impact of maintenance on certain production costs. We will further elaborate on these in Section 3. As described above, only a few studies address the impact of maintenance on the production costs and how this can be assessed. However, a complete overview of which production costs should be considered when assessing the impact of maintenance, is lacking. This overview is important because this gives a complete indication of the impact of maintenance that, in turn, helps in justifying investments in maintenance. In addition, the studies do not distinguish between different maintenance policies. This distinction is important because it exposes the differences between maintenance policies and, therefore, makes it possible to justify investing in a certain policy. Distinguishing maintenance policies makes it also possible to benchmark the current performance. This study aims to build a model that makes it possible to assess the impact of maintenance on the production costs. Part of this model is a framework that distinguishes purely corrective maintenance, time-based maintenance, and condition-based maintenance. This distinction is made by their impact on the production costs, which can be used to justify investments in a maintenance policy and to benchmark the performance of a current maintenance policy. In order to do so, we first distinguish and describe various cost factors that are affected by the maintenance policy that is used. Thereafter, we describe how these costs vary for different maintenance policies. This is done with a combination of a literature study and a single-case study. The input of the case company is both financial and qualitative data.

(8)

Page 8 of 75

How to assess the impact of maintenance on the production costs?

(9)

Page 9 of 75

2. Literature review

This literature review starts by giving a short explanation of what is considered a machine failure. In Section 2.2, we further describe maintenance and its different policies. In Section 2.3, the literature about how to assess the impact of maintenance is reviewed. Section 2.4 reviews studies related to the relationship between maintenance and the production costs. In Section 2.5, the studies that distinguishes corrective maintenance, time-based maintenance, and condition-based maintenance are discussed. The last section is an overview and puts more emphasis on the gap in the literature.

2.1. Machine failure

Before the concept of maintenance is discussed, first a failure needs to be defined. It is important to define a failure because a failure, or indicators of a failure, will initiate some form of maintenance. When a production facility does not know what a failure is, it is often difficult to establish an effective maintenance policy. Ahmad & Kamaruddin (2012) gives a good description about what a failure in a production environment is. The study describes that a failure can be classified in three categories, namely a breakdown failure, a functional failure, and a combination of a breakdown and functional failure. During a breakdown failure, a machine is considered failed when it cannot perform its predetermined and expected function due to physical damage. When the performance of a machine, system, or component is below a certain threshold, it is called a functional failure. The last type of failure is a combination of a breakdown and a functional failure, which is especially the case for multi-unit machines. Within this research, if we refer to a failure, we mean the definition of a failure from Ahmad & Kamaruddin (2012).

2.2. Maintenance

(10)

Page 10 of 75

Purely corrective maintenance

Purely corrective maintenance, also referred to as failure-based maintenance (Velmurugan & Dhingra, 2015), repair maintenance (Nicholas, 2011), or reactive maintenance (Swanson, 2001), can be described as a form of maintenance in which a machine failure triggers maintenance activities. In other words, maintenance is only performed when a breakdown has occurred. Corrective maintenance is also referred to as run-to-failure. The maintenance policy is called “purely” corrective maintenance because in the two maintenance policies, corrective maintenance is not excluded. However, in the policy described in this section, purely corrective maintenance, only corrective actions are performed.

Preventive maintenance

Preventive maintenance, also called life-based maintenance (Velmurugan & Dhingra, 2015) and proactive maintenance (Swanson, 2001), can be described as a maintenance event that is scheduled in advance at which a planned maintenance task is performed, in order to prevent a system from failing (Laks & Verhagen, 2018; Basri, et al., 2017). The ultimate goal of preventive maintenance is to improve the reliability of a certain system or machine (Laks & Verhagen, 201; Lee & Cha, 2016). Preventive maintenance is usually distinguished into two sub-categories, namely time-based maintenance and condition-based maintenance.

2.2.2.1. Time-based maintenance

The traditional form of preventive maintenance is time-based maintenance. Within this policy, maintenance is performed at a certain time interval. There are several options to determine this interval,

Maintenance Corrective maintenance Preventive maintenance Time-based maintenance Age-based Usage-based Block-based Condition-based maintenance Inspections Continous monitoring

(11)

Page 11 of 75 namely based on age, time (block) and usage (De Jonge & Scarf, 2019). Age-based maintenance is a maintenance policy under which a machine is preventively maintained at a predetermined age or when a failure occurs, while in a block-based maintenance policy, maintenance is performed at certain intervals independent of a failure has occurred or not (Shafiee & Chukova, 2013). In the last sub-category, use-based maintenance, the maintenance interval is use-based on the running time of a machine or on the production numbers.

2.2.2.2. Condition-based maintenance

Condition-based maintenance, an upcoming form of preventive maintenance, is a form of maintenance in which maintenance actions are performed based on information that is collected through a monitoring process (Ahmad & Kamaruddin, 2012). In a condition-based maintenance policy, a specific parameter, or set of parameters, is monitored which triggers maintenance when a certain condition is reached. An example of such a parameter is temperature or vibration. Within condition-based maintenance there are generally two types of monitoring, namely through inspections or continuously monitoring.

Preventive maintenance is usually considered more beneficial than corrective maintenance because it extends the lifetime of a machine and reduces the probability of a breakdown or failure and, therefore, increases the reliability of a machine (Swanson, 2001). The increase in reliability has several other advantages which will reduce the production costs, for example the reduction of buffers (WIP) between processes and the reduction of direct labor costs due to less inefficient resources resulting from a reduction in unplanned maintenance and stoppages.

2.3. Assessment of maintenance

(12)

Page 12 of 75 frequency of planned stoppages, failures, short stoppages, and amount of rejected items due to quality, also referred to as rejects. Economic data includes parameters like maintenance costs, operating costs, profit margin, investments in maintenance, and spare parts inventory. After the required input data is determined, the next step is to determine the equations that calculate the models’ cost factors determined

in the first step. When the equations are established and the data is gathered, the output of the model can be determined. These outputs are (A) potential savings, (B) direct maintenance cost, (C) investments in maintenance, (D) maintenance savings, (E) maintenance profits, and (F) maintenance performance measures. The first four outputs (A, B, C, and D) contribute to the maintenance performance measures and help in identifying problem areas.

The model is useful for this research because it provides a comprehensive roadmap for determining the potential savings resulting from maintenance. Furthermore, the article describes the cost factors that are influenced by preventive maintenance. According to Al-Najjar & Alsyouf (2004), the costs that are influenced by maintenance are direct labor, direct materials (e.g. spare parts), and overheads (tools, instruments, training, and administration). Furthermore, the article describes a positive relationship between preventive maintenance and the overall equipment efficiency (OEE) due to the increase in availability, increase in efficiency, and decrease in idle fixed cost resources (e.g. idle machines and idle workers). However, the article does not further explain the maintenance performance measurements while this is an important aspect of determining the effect of the maintenance policy. The article also does

(13)

Page 13 of 75 not explain how the formulas in the fourth step of model should be established and what should be included or excluded.

2.4. Relationship between maintenance and the production costs

This section reviews the studies about the relationship between maintenance and the production costs. First, there is another article of Al-Najjar (2007). The article provides a framework that indicates the relationships between five different maintenance investment areas and the technical and financial impact on the operative level and the strategic level. The operative level includes the production, quality, and logistic system, together with other operational activities. At the strategic level are a company’s technical and financial objectives translated into strategies for all the working areas, e.g. quality, production, and maintenance. The framework highlights the route through which an investment is converted from a technical impact to a financial impact and eventually to a strategic impact. For example, investing in better using the available technologies will have a technical impact by reducing failurs, stoppages, and the time to repair, and gives a higher OEE. The investment has a financial impact because the technical impact result in, for example, less personnel expenses, lower production costs, and less direct maintenance costs. Eventually, the investments has an impact on the strategic level by, for example, higher savings, more profit, less delivery delay, and a bigger markert share.

The article of Al-Najjar (2007) also provides some formulas for determining the impact of maintenance, which is lacking in the article of Al-Najjar & Alsyouf (2004). The author provides formulas for the following savings Sn:

1. Fewer failures: the savings in the production cost achieved by less failure:

𝑆1= 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑓𝑎𝑖𝑙𝑢𝑟𝑒𝑠 𝑎𝑣𝑜𝑖𝑑𝑒𝑑 ∗ 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑠𝑡𝑜𝑝𝑝𝑎𝑔𝑒 𝑡𝑖𝑚𝑒 ∗ 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑟𝑎𝑡𝑒

∗ 𝑝𝑟𝑜𝑓𝑖𝑡 𝑚𝑎𝑟𝑔𝑖𝑛. (1)

2. Shorter average stoppage time: longer production time:

𝑆2= 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑖𝑛 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑓𝑎𝑖𝑙𝑢𝑟𝑒 𝑠𝑡𝑜𝑝𝑝𝑎𝑔𝑒 𝑡𝑖𝑚𝑒 ∗ 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑓𝑎𝑖𝑙𝑢𝑟𝑒𝑠 ∗ 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑟𝑎𝑡𝑒

∗ 𝑝𝑟𝑜𝑓𝑖𝑡 𝑚𝑎𝑟𝑔𝑖𝑛. (2)

3. Less short stoppages: savings in production costs achieved by fewer short stoppages:

𝑆3= [𝑃𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝑦𝑒𝑎𝑟 𝑠ℎ𝑜𝑟𝑡 𝑠𝑡𝑜𝑝𝑝𝑎𝑔𝑒𝑠 − 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑦𝑒𝑎𝑟 𝑠𝑡𝑜𝑝𝑝𝑎𝑔𝑒𝑠] ∗ 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑠𝑡𝑜𝑝𝑝𝑎𝑔𝑒 𝑡𝑖𝑚𝑒

∗ 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑟𝑎𝑡𝑒 ∗ 𝑝𝑟𝑜𝑓𝑖𝑡 𝑚𝑎𝑟𝑔𝑖𝑛. (3)

4. Higher quality production: higher production rate:

𝑆4= [

𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑦𝑒𝑎𝑟 ℎ𝑖𝑔ℎ 𝑞𝑢𝑎𝑙𝑖𝑡𝑦 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑝𝑒𝑟 ℎ𝑜𝑢𝑟 − 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠 𝑦𝑒𝑎𝑟 ℎ𝑖𝑔ℎ 𝑞𝑢𝑎𝑙𝑖𝑡𝑦

𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑝𝑒𝑟 ℎ𝑜𝑢𝑟 ] ∗ 24 ℎ

(14)

Page 14 of 75 5. Less expenses paid by the company:

𝑆5= ∑(𝐸𝑥𝑝𝑒𝑛𝑠𝑒𝑠 𝑏𝑒𝑓𝑜𝑟𝑒 − 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑠 𝑎𝑓𝑡𝑒𝑟)𝑗 𝑛

𝑗=1

. (5)

The total savings can be computed by the following formula:

𝑇𝑜𝑡𝑎𝑙 𝑠𝑎𝑣𝑖𝑛𝑔 = ∑ 𝑆𝑖 5

𝑖=1

. (6)

The study of Al-Najjar (2007) is an important contribution for this research because it gives a clear overview of the impact of several maintenance investments on both the technical and financial aspects. The study goes even further by clarifying the impact on the strategic level. In addition, the study also provides basic formulas for calculating the savings resulting from investments in maintenance. However, the study remains general on assessing the impact of maintenance. The formulas provide a general method to determine the savings, but does not exclude other factors that influence the saving. For example, the formula to determine the savings due to higher quality production (4), does not exclude other factors, which are not related to maintenance, that influence quality production rate. An example of such a factor might be better skilled personnel.

Another study that looked into the impact of maintenance is the study of Alsyouf (2007), in which the role of maintenance on a company’s productivity and profitability is studied. The study explores the effect of maintenance on the profit margin and tested the model shown in Figure 3. The author explains the relationship between having a productive maintenance policy and the production process’ efficiency and effectiveness. These two factors influence the profit margin, due to changes in productivity and price recovery resulting from maintenance. The price recovery indicates the ratio between the price of the output products and the allocated cost of the consumed inputs. The article has the focus on changes in productivity because the changes in price recovery are hard to trace. The article concludes that having a

(15)

Page 15 of 75 productive maintenance policy will increase the efficiency and effectiveness of a system. In order words, having a productive maintenance policy will result in a higher overall equipment efficiency (OEE).

Swanson (2001) performed a study in which she reflected on the performance of certain maintenance policies. Swanson (2001) performed 708 surveys at 354 plants to investigate the relationship between maintenance strategies (corrective maintenance, preventive maintenance and aggressive strategies (e.g. TPM)) and performance. The survey included questions related to maintenance practices and performance measures. The results of the analysis indicated a strong positive relationship between proactive (preventive) and aggressive maintenance strategies and performance. This implies that a proactive and/or an aggressive maintenance strategy increases the performance of a production facility. However, this study does not indicate what the influence is of preventive maintenance on the different production costs components.

The studies described above explain the relationship between maintenance and production costs on a relatively high level. There are, however, studies that go into more detail by describing the relationship between individual costs aspects and maintenance. The study of Fakher, Nourelfath, & Gendreau (2018), who developed an optimization model that takes maintenance, quality, and production into account, aims to maximize the expected profit. An important contribution of this study is that it found that increasing the level of preventive maintenance would lead to a reduction in rejects. This is an important contribution because it shows the relationship between preventive maintenance and quality control.

The relationship between spare parts and preventive maintenance is a focal topic within the maintenance literature because the availability of a spare part can increase the performance and effectiveness of a system (Barabadi, et al., 2014). Literature concludes that when a system has a preventive maintenance policy, the amount of spare parts in inventory can be reduced because the moment when these spare parts are needed, is better predictable (Smidt-Destombes, et al., 2008; Oner, et al., 2010; Bjarnason, et al., 2014). This is an important contribution because it implies that the costs of spare parts inventory will decrease as the amount of preventive maintenance will increase. However, Poppe et. al (2017) concluded in their simulation study that the number of spare parts needed will increase because the entire useful life of the components is not fully exploited.

(16)

Page 16 of 75 reduction in WIP and buffer lies in the increase in reliability due to preventive maintenance. When the reliability is increased, the size of the safety buffer could be decreased.

2.5. Comparison of the maintenance policies

(17)

Page 17 of 75 The relative benefit of condition-based maintenance over time-based maintenance will decline if the uncertainty of the failure level increases.

The contribution of the study of De Jonge, et al. (2017) for this study is twofold. First, by reviewing literature about the comparison between time-based maintenance and condition-based maintenance, the study exposes the gap in literature related to time-based maintenance and condition-based maintenance. Second, the study shows that the performance of condition-based maintenance is in most situations preferred over time-based maintenance. However, the study does not explain the impact of these policies on the production costs and it does not include the investments needed for these policies. These investments are important to include because these can differ significantly among the policies, which can influence the decision-making between policies.

Besides the above-described study, several other studies consider maintenance optimization models in which the different policies are compared. For example the study of Crowder & Lawless (2007) that compared corrective maintenance with the optimal time-based maintenance policy and an optimal condition-based maintenance policy. The study concluded that condition-based maintenance could provide a company with substantial savings compared to an age-replacement policy (time-based) and a corrective maintenance policy. Around the optimum of the condition-based maintenance policy, the expected cost per unit time varies little with respect to the inspection time and the replacement/repair threshold. This gives a company some flexibility with regard to the inspection time.

(18)

Page 18 of 75 preferred option. The full potential of predictive maintenance is not investigated in the work of Zio & Compare (2013). The study shortly describes the practical differences between the maintenance policies. First, a corrective maintenance policy is the simplest policy because it does not require any information, algorithm, and instrumentation. Second, in the case of time-based maintenance, a database with the failure times of a certain machine or system is required but still no instrumentation is needed. However, in the case of a condition-based maintenance policy, software and instrumentation (hardware) is needed to support the monitoring that is required in a condition-based maintenance policy. At last, for a predictive maintenance policy, besides the software, hardware, and algorithms, other technologies are needed in order to predict when maintenance is needed.

Another relevant study is the study of Ahmad & Kamaruddin (2012) who compared condition-based and time-based maintenance from a practical point of view. The study focused on the challenges of implementing the policies and especially explored the issues related to the differences in data requirements and collection, data analysis, and the eventually the decision making process. An overview of this comparison is given in Table 1.

Table 1: Comparison of time-based and condition-based maintenance (Ahmad & Kamaruddin, 2012)

Comparison criteria Time-based maintenance Condition-based maintenance Data requirements and

collection

Theory/principle

Uses failure time/user-based data

Practical issues

• The recording of failure time data is not always available

• Very sensitive due to incorrect recording and censoring effects

• Failure dataset difficult, time-consuming, and expensive to gather (Dekker & Scarf, 1998)

Uses any parameter that indicates condition of equipment

• High data collection costs (sensor, training, etc.) (Kothamasu, et al., 2006)

Data analysis/modelling Theory/principle

Uses the reliability theory, assuming the Bathtub curve

Practical issues

• Unrealistic assumptions

• Operating conditions assumed to be constant • Only effective for equipment in deteriorating

state (increasing failure rate)

Deterioration modelling

• Large data samples required

• Complex data cleaning process required Decision process Theory/principle

Use of the optimization approach

Practical issues

• Difficult to model and interpret • Decision model is not always stable • Time-consuming in model development • Most time spend on mathematics than practical

application

Use of failure estimation/prediction approach and comparison with predetermined failure limits

• Not enough planning time in the current condition evaluation-based method

(19)

Page 19 of 75 Based on the comparison, Ahmad & Kamaruddin (2012) conclude that from a principle point of view, applying condition-based maintenance is more realistic than time-based maintenance due to better data availability and accuracy, clear objectives for evaluating the condition of a system, and applying condition-based maintenance is simpler because time-condition-based maintenance is condition-based on the optimization approach. These conclusions are contradictory to the general conclusions of the studies related to condition-based maintenance. Most studies, e.g. the study of Zio & Compare (2013), conclude that time-based maintenance is easier to implement, compared to condition-based maintenance. Condition-based maintenance is assumed more complex due to expensive and complex monitoring systems and because of analyzing condition data is considered as time-consuming and complex. Furthermore, most studies also state that besides time-based maintenance, condition-based maintenance also requires optimization techniques because the moment of failure should constantly be optimized.

2.6. Overview

(20)

Page 20 of 75 policies. The main conclusion is that the performance of condition-based maintenance is usually better than time-based maintenance. However, the three policies are not distinguished by their impact on the production costs while this is important knowledge for a company because, as described in the introduction, distinguishing between these three maintenance policies makes it possible to justify investments in one of these policies.

Similar to the studies of Al-Najjar & Alsyouf (2004) and Al-Najjar (2007), this study aims to build a model that can be used to assess the impact of maintenance on the production costs. In addition, this study aims to give a complete overview of all the production costs that are impacted by maintenance. This lacks within the models of Al-Najjar & Alsyouf (2004) and Al-Najjar (2007). An important input for our model are the studies that only consider the relationship between maintenance and one cost aspect. We will further use these studies within the literature study in Section 4. Furthermore, this study will expand the assessment model by adding a framework that distinguishes purely corrective maintenance, time-based maintenance, and condition-based maintenance. This framework will indicate what the impact is of these three maintenance policies on the different production costs by means of a score. The impact of these three maintenance policies is not yet discussed within the literature. A reason for this is that condition-based maintenance is a relatively young maintenance policy compared to corrective maintenance and time-based maintenance, and that it is difficult to quantify the impact of preventive maintenance because it is complex and time-consuming to explore which failures are prevented.

(21)

Page 21 of 75 This framework is conceptually given in Figure 4. The relationships, A, B, and C, are one way. This means that the three policies affect the production costs, but the production costs do not affect the maintenance policies. The framework aims to score the three maintenance policies by their impact on the production costs. It, therefore, aims to explore the strength and the type of the relationships A, B, and C. Besides the relationships, this research explores which production costs are impacted by maintenance.

A B C Condition-based maintenance Corrective maintenance Time-based maintenance Production costs

(22)

Page 22 of 75

3. Methodology

In this section, the methodology of this research is provided. The goal of this section is to provide an overview and an explanation of the steps taken within this study. This section provides the information that makes it possible to replicate this study. Section 3.1 describes the design of this research and explains why a combination of a literature study and a single-case study is chosen. Section 3.2 gives a description about the case that is used within this research. Section 3.3 describes how the data is collected in order to build the assessment model. Section 3.4 explains how the data is analyzed.

3.1. Research design

This study aims to build and refine a theory. A theory is a set of constructs, definitions, and propositions that are linked in order to present a systematic view of a certain phenomenon (Karlsson, 2016). By specifying the relationships among variables, the behavior of the phenomena can be explained and predicted. The theory that is built is the framework that distinguishes purely corrective maintenance, time-based maintenance, and condition-time-based maintenance by their impact on the production costs. This theory is added into a model that can be used to assess the impact of maintenance on the production costs. In addition, this study also extends/refines a theory by building a model that consists all the production costs because, like indicated in the literature review, current models do not encompass all the production costs. Therefore, this study is both theory building and theory elaboration/refinement. Karlsson (2016) describes that a case study is suitable for both refining and building a theory because case studies help to mature an upcoming theory. Based on the literature review, it is assumed that the theory related to this study is upcoming.

In addition to the single-case study described above, this study also includes an extensive literature study. The literature study is deliberately separated from the literature review. The available literature about assessing the impact of maintenance, the literature about the relationship between maintenance and the production costs, and the literature that compares corrective maintenance, time-based maintenance, and condition-based maintenance, was reviewed within the literature review. The literature study is an actual part of this study and is not a review of the available literature. The goal of the literature review was to establish a gap that this research aims to fill, while the aim of the literature study is to assist in building the assessment model and the framework.

(23)

Page 23 of 75 positive relationship, the dependent variable, a certain production cost, behaves in the same direction as the independent variable, a maintenance policy. This implies that the cost aspect will increase when a preventive policy is used, compared to a corrective policy. In a negative relationship, the dependent variable behaves in the opposite direction of the independent variable. Which implies that the costs will decrease as there is further invested in a preventive maintenance policy. The production costs, impacted by maintenance, are explored by a mix of a literature study and a single-case study. First, literature is used to find most of the production costs that are impacted by maintenance. After the literature is consulted, these production costs are compared with the production costs of our single-case study and, if necessary, the list of production costs is complemented. The same holds for the relationships between these production costs and maintenance. Like shortly described in the literature review, there are studies that explain the relationship between several production costs, e.g. spare parts and WIP. However, not all the production costs are explored within the literature. Therefore, the single-case study is used to determine the missing relationships.

The second part of this research explores how the impact on the production costs differs per maintenance policy. This part of the study is done by means of a single-case study because it is fully theory building. As described in Section 2, there are no studies that compare corrective maintenance, time-based maintenance, and condition-based maintenance by their impact on the production costs. That is why this part of the study is assumed theory building, which makes a single-case study suitable. The result of this part is a framework that distinguishes the three maintenance policies by their impact on the production costs, determined in the first part. The characteristics of the three maintenance policies that can influence the impact on the production costs, as discussed in Section 2.5, is used to reflect on the differences between the output of the single-case study and the literature review. Table 2 shows the structure of the framework. The values entered in this structure only serve as an example. The results of this section also indicate whether there are differences between the expected relationships and the results from the single-case study.

Table 2: Structure of the framework (numbers serve as an example)

(24)

Page 24 of 75 After the framework is built, the other steps of the model are determined in the third step. The model is based on the single-case study and on the literature. There are studies that provide useful input for the model, see the literature review in Section 3. Together with the data collected from the case company, an assessment model is established. This model can be used to assess the impact of maintenance on the production costs and, therefore, will help in justifying maintenance investments. Furthermore, the framework that distinguishes the three maintenance policies by their impact on the production costs makes it also possible to choose the right maintenance policy for a certain situation. As previously described, a single-case study is suitable for building and extending a theory. Therefore, in order to build the assessment model, we did a single-case study.

After the framework and the assessment model are made, the working of the model is illustrated by a numerical example, based on a real situation from our case company. This example follows the steps of the assessment model and shows what needs to be done within these steps. Furthermore, the example highlights the importance of preventive maintenance.

3.2. Case description

This section provides a brief explanation of the case company and explains why this company is suitable for this research. The company that is used within this single-case study is the shaver production facility of Philips in Drachten. At this production facility, four different types of shavers are produced resulting in around 450 different shavers (SKU) in all kinds of colors. The factory has over a thousand different machines of which most are molding machines that produce plastic parts for the shavers. Besides producing shavers, the facility recently started producing baby bottles and parts for electric toothbrushes. This study mainly used the shaver production part as input. This facility produces around 8.4 million shavers annually. It should be noted that Philips Drachten is a production facility that delivers shavers to Philips. Therefore, Philips Drachten has little influence on the selling price, i.e. it can only decrease their production costs to increase the profit of Philips.

(25)

Page 25 of 75 What further differentiates the facility from other production facilities is its high degree of Lean. When a production facility is aiming to become Lean, management drives the organization towards continuously optimizing the production process by reducing or removing sources of waste (Nicholas, 2011). To illustrate the degree of Lean, other companies use Philips Drachten as an example for their own production processes. As an example, between 2016 and 2018 around 80 companies visited the factory to have a look at the effects of Lean.

Also worth mentioning is the combination of machines and human labor. Philips Drachten uses highly sophisticated machines and has a high degree of automation. However, it still also uses a lot of human labor to perform certain operational tasks, while these tasks could also be automated. In total, Philips Drachten has around 2000 employees.

Philips Drachten is considered suitable for this study because they are actually facing the problem that initiated this research. Namely, Philips Drachten struggles with justifying investments in maintenance. Maintenance is an important department for Philips Drachten because the facility consists of highly sophisticated machines, which makes it attractive to improve maintenance policies. Furthermore, due to the high degree of Lean, the facility aims to produce with no waste. If a maintenance policy reduces the number of failures, this will result in, for example, a reduction of safety buffers and WIP. Therefore, maintenance contributes to reducing waste. These two aspects make maintenance an important contributor to the total production costs of the facility.

3.3. Data collection

As previously described, this study consists of a combination of a literature study and a single-case study. This section describes how and which data is collected from the literature study and the case company. For this study, the input from the case company is used in several ways. First, the case company is used to determine which production costs are impacted by maintenance. Second, based on the input from the case company, the framework, which shows how the impact of the maintenance policies differs, is made. Third, the case company provides input for which steps needs to be included in the model to determine the impact of maintenance on the production costs. At last, based on data from the case company, a numerical example is given.

(26)

Page 26 of 75 Furthermore, literature is also used to reflect on the input for the framework that distinguishes the three maintenance policies by their impact on the production costs. At last, literature also provides input for the steps that need to be included in the model to determine the impact of maintenance on the production costs.

Input from the case company stems mainly from interviews with maintenance managers and with financial managers. The interviews are semi-structured, which means that the interviews contain both structured and unstructured questions. Structured questions are predetermined, while unstructured questions raise during the interview. The respondents of the interviews were given a description and a list of definitions that is shown in Appendix 1. The input from finance is used to determine the production costs of the facility and to compare these with the findings in the literature. The maintenance managers provide input for both the relationships between a production cost aspect and one of the three maintenance policies, but also for the steps that need the be included in the assessment model. Furthermore, the maintenance managers were asked to score the impact of the three maintenance policies on the production costs. The differences between these scores highlights the differences between the impacts of the three policies on the production costs. This is useful knowledge when a company specifically wants to reduce a certain cost aspect by means of investing in maintenance. At last, when the assessment model and the framework are made, these are tested by means of a numerical example. The case company provides data for this example. This numerical example is added to show how the model works and to test whether the scores of the framework are representative.

3.4. Analysis

This section describes how the collected data is further analyzed in order to build the assessment model. First, the data from the interviews with financial managers is compared with the results from the literature study. This gives a full picture of which production costs a production facility has and how these are categorized. This step results in a conceptual model that indicates the relationships between a maintenance policy and the production costs.

(27)

Page 27 of 75 Therefore, the deviations between the answers from the different maintenance managers is explored, as well as the differences between the answers and the literature.

4. Assessing maintenance

In this section, the assessment model is developed. This assessment model makes it possible for a production facility to assess the impact of maintenance. Compared to other assessment models, this model includes a framework that distinguishes purely corrective maintenance, time-based maintenance, and condition-based maintenance by their impact on the production costs. This framework facilitates the decision-making regarding which maintenance policy a production facility should choose. In order to build the assessment model, the first step is to determine which production costs a typical production facility has and in what manner these are impacted by maintenance. This is being discussed in Section 4.1. In Section 4.2, the framework is given and in Section 4.3, the assessment model is made. At last, in Section 4.4, we provide a numerical example that shows the working of the assessment model.

4.1. Production costs

This section explores which production costs a production facility has and how these production costs are impacted by maintenance. These production costs form the basis for the assessment model and the framework. It should be noted that throughout this section, the increase in availability resulting from preventive maintenance is often referred to as a cost savings. This is only the case when the demand is higher than the current capacity. Throughout this section, this assumption is made.

Overheads

The first production costs under consideration are overhead costs, also called overheads. Overheads are described as the indirect costs associated with the production of a certain product. According to Drury (2012), the overheads are the costs of manufacturing apart from direct labor and material cost. For this research, the costs of indirect labor, inventory, and quality control are not considered part of the overheads because maintenance has a direct impact on these costs. In this research, the overheads are the costs of equipment, buildings, production planning and scheduling, energy, and supplies.

(28)

Page 28 of 75 availability, decrease in products below quality, and increase the useful lifetime of a machine (Al-Najjar & Alsyouf, 2004; Nicholas, 2011). The increase in availability and the decrease in products below a quality level make it possible to produce more products in the same timeframe. These two aspects are usually measured by means of the overall equipment effectiveness (OEE). This measure translates the availability, quality rate, and efficiency into a percentage. Having an effective maintenance policy will result in a higher OEE. Furthermore, the increase in useful lifetime generates more production time for a certain machine. The increase in production, results in lower equipment costs because the fixed price of a machine can be spread out over more produced products, resulting in a lower price per produced product.

The energy costs of a production facility are the costs associated with the consumption of energy for a certain machine or system (Ozbayrak, et al., 2004). The costs of energy are considered a mix of variable and fixed costs because it can be linked to the direct production of a product, while also having a fixed part. It is directly linked to the running time of a certain machine. Obviously, if only one product is produced, less energy is required compared to the production of thousand products. On the other side, there is always a fixed part of these costs because a factory always consumes a bit of energy while not producing anything, e.g. due to lightning. Al-Najjar & Alsyouf (2004) also mention that the energy costs are related to maintenance. The study describes that failures and short stoppages can result in disproportional energy consumption. Garcia-Sanz-Calcedo & Gomez-Chaparro (2017) came to the same conclusion and concluded that preventive maintenance will result in a lower energy consumption. An effective maintenance policy will decrease the variable part of the energy costs because it reduces failures and short stoppages. Furthermore, due to the increase in availability and decrease in produced products below quality standards, the fixed part of the energy costs can be divided over more products, resulting in a lower production price per product.

Other overheads such as supplies (e.g. office items and protective equipment), planning and scheduling, and other administrative costs are also considered to go down when an effective maintenance policy is applied. The main reason for this is again the increase in production due to the increase in availability and reduction in rejects (Al-Najjar & Alsyouf, 2004; Alsyouf, 2007). These costs are mostly fixed, or have a fixed part, which can be spread over more products, which in turn will lower the cost per product produced.

Direct labor

(29)

Page 29 of 75 operators of a machine, that can be specifically and exclusively linked to a certain ‘cost object’ or product (Drury, 2012). The best-known example of a direct labor cost are wages of the operating staff. Maintenance affect the costs associated with direct labor in two ways. First, as previously mentioned, the increase in availability and decrease in rejects will lead to a higher OEE, and thus a decrease in cost per unit produced.

Second, in the case of a preventive maintenance policy, either time-based maintenance or condition-based maintenance, maintenance actions are planned. This reduces the direct labor costs because no, or less, operating staff is needed during that maintenance period. If a failure occurs, the planned operating staff usually needs to be paid regardless whether they are effective or not. During maintenance actions, operating staff is considered ineffective because the machines that they operate are not available. The costs associated with this ineffectiveness could be avoided when a preventive maintenance policy is used because the operating staff is simply not scheduled during the maintenance actions. Therefore, as already described in the article of Al-Najjar & Alsyouf (2004), having an effective maintenance policy will reduce the direct labor costs.

Indirect labor

The costs associated with indirect labor are classified as the labor costs, e.g. wages of employees, which cannot be directly linked to a certain product or the production of a product (Drury, 2012). For example, the costs associated with purchasing staff and material-handling staff. In accounting, the indirect labor costs are often assumed part of the overheads because it is difficult to link those to the production of a product. It should be noted that maintenance personnel are not included in this category, but are covered by the maintenance category.

(30)

Page 30 of 75

Inventory

Inventory is all stock of any item or resource, e.g. raw material, work-in-progress (WIP), or finished goods, within an organization. The costs associated to inventory can be divided into four different categories, namely purchase costs, ordering costs, inventory-carrying costs, and shortage costs (Samak-Kulkarni & Rajhans, 2013). It should be noted that within this research, spare parts inventory is considered separately because it has a different relationship with maintenance than the other inventories. Nicholas (2011) defines the inventory costs as holding costs that include the costs of storing space, paperwork and handling, insurance, security, and pilferage (stealing).

Inventory is considered as one of the seven wastes described by Toyota and is assumed the ‘root of all evil’ (Nicholas, 2011). This is because inventory is considered as hiding and avoiding problems. This is often illustrated by a ship (a production facility) that sails in high water (inventory) under which the rocks (problems) are hidden; see Figure 6 for an illustration. When the water is high enough, the ship does not see the rocks. When the water level is lowered, the rocks become visible and the ship needs to find ways to deal with them. This comparison illustrates the effect inventory has on a production facility.

Nicholas (2011) describes that inventory is used to maintain an uninterrupted flow in the case of machine breakdowns, delivery delays from suppliers, material defects, and long and costly setups. Lowering the amount of inventory makes suchs problems visible, as shown in Figure 6. Solving these problems eventually results in a higher process efficiency. Furthermore, having a preventive maintenance policy will decrease the number of breakdowns and the number material defects, and makes a system more reliable.

Inadequate market intelligence Ineffective logistics Line imbalance Poor layout Poor scheduling Machine breakdowns Poor quality Poor organization Long setups Production facility Inventory

(31)

Page 31 of 75 This increase in reliability will result in a decrease in WIP because less safety buffer is needed (Gupta, et al., 2001; Nicholas, 2011). This reduction will lower the inventory costs.

When a production company has storage space at their facility, which is owned by the company itself, it is often difficult to determine the savings resulting from a reduction in inventory. The reason for this is that the space that is built to function as a storage location is already paid. Therefore, reducing inventory will only result in unused storage space. However, this unused storage space generates the opportunity to improve other processes, to increase production numbers by extending processes, or to introduce new business opportunities. The increase in production will eventually decrease the cost of a producing a shaver because the fixed costs can be divided over more products. On the other side, reducing inventory directly impact the type of inventory costs that is related to administration and transportation. When a production facility rents storage space from an external party, reducing inventory will directly result in a saving because less rented space is needed.

Spare parts inventory

The relationship between spare parts and preventive maintenance is a focal topic within maintenance literature. This is an important topic because the availability of a spare part can increase the performance and effectiveness of a certain system (Barabadi, et al., 2014). Literature about the relationship between preventive maintenance and spare parts concludes that in a preventive policy, the average amount of spare parts in inventory can be reduced. This is because the moment when these spare parts are needed is better predictable and the frequency of unexpected failures is lower (Smidt-Destombes, et al., 2008; Oner, et al., 2010; Bjarnason, et al., 2014). This is an important contribution for this research because it implies that the costs of spare parts in inventory will decrease, as the amount of preventive maintenance will increase. However, Poppe et. al (2017) concluded in their simulation study that the number of spare parts needed will increase because the entire useful life of the components is not fully exploited. Therefore, based on the literature, increasing the number of preventive maintenance actions will result in a decrease in the costs of spare parts inventory, while the costs for spare parts themselves will increase. The latter will be further explained in Section 4.1.6.

(32)

Page 32 of 75 for example the study of Rausch & Liao (2010) who built and tested a heuristic to determine the optimal base-stock level of spare parts for a based maintenance policy. However, whether a condition-based maintenance policy results in less spare parts on stock compared to a time-condition-based maintenance policy is not yet investigated. One could assume that there will always be a chance that corrective maintenance (CM) is needed. Therefore, for both a time-based and condition-based maintenance policy there will always some spare parts on stock. However, under a condition-based maintenance policy there is some uncertainty in the moment that a certain deterioration threshold will be exceeded. This uncertainty could be reduced by adding a planning time, but this makes the policy less effective. Thus, it is expected that the amount of spare parts in inventory is higher under a condition-based policy. The relationship between the three policies and spare parts inventory is therefore as follows:

𝑆𝑝𝑎𝑟𝑒 𝑝𝑎𝑟𝑡𝑠 𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦: 𝑇𝐵𝑀 < 𝐶𝐵𝑀 < 𝐶𝑀. (7)

Number of spare parts

A spare part is a part of a machine that is not yet used but is needed as soon as the original part is broken, not in the right condition, or when due to another reason replacement is needed. The costs associated with the spare parts are the costs of buying the spare part (acquiring costs). These include for example, the costs of the component itself, administrative costs, and transportation costs.

The cost of a spare part should not be confused with the inventory costs to keep them in stock. The number of spare parts in inventory will reduce in a preventive maintenance policy, like described in the previous section. On the other hand, the number of spare parts needed will increase because the entire useful life of certain components is not fully exploited, i.e. these are preventively replaced (Poppe, et al., 2017). This implies that during a certain period, depending on the preventive maintenance interval, more spare parts are used than when purely corrective maintenance was performed. However, in their simulation study, Poppe, et al. (2017) assume a single-component machine. One could imagine that in a multi-component machine, a failure of a single-component can cause failures of other components. Therefore, we made a distinction between a single-component machine and a multi-component machine.

4.1.6.1. Single-component

(33)

Page 33 of 75 time-based maintenance, more spare parts are needed than a corrective or condition-based maintenance policy because the condition of a certain component is ignored and maintenance is planned on a predetermined interval. This increases the chance of preventively replacing a component that is still in the right condition. This relationship is thus as follows:

𝑆𝑝𝑎𝑟𝑒 𝑝𝑎𝑟𝑡𝑠 𝑛𝑒𝑒𝑑𝑒𝑑 (𝑠𝑖𝑛𝑔𝑙𝑒 − 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡): 𝐶𝑀 < 𝐶𝐵𝑀 < 𝑇𝐵𝑀. (8) 4.1.6.2. Multi-component

When collateral damage, i.e. the failure of one component results in the failure of other components, is considered, it is expected that the costs for spare parts are the highest under a purely corrective maintenance policy. This is because the frequency of failures are the highest under this policy compared to the other policies. Under a purely corrective policy, maintenance is only performed in the case of a failure. Therefore, the chance of collateral damage is the highest. We expect that in the multi component situation, the costs are the lowest under a condition-based maintenance policy. Under a condition-based maintenance policy, the maintenance actions are performed under a variable time interval because it is based on the condition of that component. Therefore, the maintenance policy is more efficient. Under a time-based maintenance policy, the interval between maintenance actions is fixed. The actual condition of the component is ignored under a time-based maintenance policy, which increases the chance of replacing a spare part that is still in the right condition. This eventually results in an increase in the used spare parts. This results in the following relationship:

𝑆𝑝𝑎𝑟𝑒 𝑝𝑎𝑟𝑡𝑠 𝑛𝑒𝑒𝑑𝑒𝑑 (𝑚𝑢𝑙𝑡𝑖 − 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡): 𝐶𝐵𝑀 < 𝑇𝐵𝑀 < 𝐶𝑀. (9)

Quality control

The costs associated with quality control are all costs that can be directly linked to the actions and inspections needed to guarantee that a product is fulfilling its requirements (ISO 9000, 2015). Cost of quality control can be divided into prevention costs, appraisal costs, and failure costs. Prevention costs are the costs referring to preventing quality issues, appraisal costs are the costs associated with detecting quality issues, and failure costs are the costs related to rework or when it is mistakenly delivered to the customer (Farooq, et al., 2017). Scrap related costs are also part of this category, which is defined as the cost of defective products, which are not according to a certain standard and cannot be sent to the market (Shivajee, et al., 2019). In this research, these products are also called rejects.

(34)

Page 34 of 75 impacts. This is an important contribution because it shows the relationship between preventive maintenance and quality control. However, the study does not explain the reasons behind these reductions. One explanation might be the increase in stability, which is one of the effects of preventive maintenance. Furthermore, in practice, failures or deteriorations are often not immediately visible but show themselves in the form of scrap or quality related production loss. Preventive maintenance will help in avoiding these situations because a large part of the failures is prevented.

There is almost no literature available that has the focus on the differences between the quality costs of a time-based and a condition-based maintenance policy. However, one could imagine that the costs of quality control are a bit higher under a time-based maintenance policy because there is a higher level of uncertainty whether a failure occurs or not. To cope with this uncertainty, more money will probably be spend on securing that the output complies with certain requirements. Furthermore, because the frequency of failures is lower under a condition-based policy, the amount of rejects will also be lower.

Raw material

Raw material is defined as the basic material from which certain products are produced. The actual definition of a raw material can differ per company. The output of a company can be the raw material for another company. An example of a raw material is crude oil of which plastic can be made. It is important to highlight the difference between this production cost aspect and the aspect of inventory, in which raw material is also mentioned. This production cost includes the acquisition of the raw material and the cost of the raw material itself. It does not include the costs for storage, which is included in the inventory cost aspect. Reason for this is that the costs for inventory will probably be influenced much more than the costs for raw material itself. In most studies, raw material is assumed as a variable cost because it directly depends on the number of products that are produced (Mondal, et al., 2009; Shivajee, et al., 2019; Thomas, 2016). It is expected that the amount of raw material needed throughout a certain period will slightly decrease if the maintenance activities are improved. The main reason for this is the previously described reduction in rejects. Therefore, the percentage of first-time-right products will increase. This will decrease the amount of raw material needed to produce a certain number of good products, resulting in a higher efficiency and a lower price of unit produced.

Maintenance

(35)

Page 35 of 75 maintenance staff and operators, software for data analysis, monitoring equipment, and maintenance tools. It is important to mention that the maintenance costs are not part of the overheads and that maintenance staff is not considered as indirect labor. Because it is expected that the maintenance costs behave different from the overhead and indirect labor costs, these costs are considered separately. Furthermore, the costs of maintenance are included within this research because these behave differently among the three maintenance policies. By including the maintenance costs in the model, the full effect of implementing a certain maintenance policy is shown. Ignoring these costs would not give a full picture of the costs and savings associated with switching to another maintenance policy. To assess these costs, we make a distinction between the costs for maintenance labor and for maintenance tools or equipment. This distinction is made because we expect that the relationship with preventive maintenance is different between these two.

4.1.9.1. Maintenance labor hours

The maintenance labor hours are the hours that a technician spend on dealing with maintenance activities. However, in this category, the labor hours that technicians spend on analyzing failure and deterioration data in order to determine the preventive maintenance interval, is also included. We expect that the maintenance labor hours will be the lowest under a condition-based policy due to the lowest frequency of failures and because maintenance can be planned at the most convenient moment. We expect that the decrease in failures resulting from the condition-based maintenance policy will outweigh the increase in labor needed for analysis. For the time-based maintenance policy, the frequency of failures is slightly higher compared to a condition-based maintenance policy. Therefore, the costs will be higher under a time-based maintenance policy. Due to the high frequency of failures and that the maintenance actions cannot be planned, the cost will be the highest under a purely corrective maintenance policy. Thus, we expect the following relationship:

𝑀𝑎𝑖𝑛𝑡𝑒𝑛𝑎𝑛𝑐𝑒 𝑙𝑎𝑏𝑜𝑟 𝑐𝑜𝑠𝑡𝑠: 𝐶𝐵𝑀 < 𝑇𝐵𝑀 < 𝐶𝑀. (10)

4.1.9.2. Maintenance tools/equipment

(36)

Page 36 of 75 basic form of maintenance, we expect that the costs for tools and equipment would be the lowest. For the maintenance equipment costs, we expect the following relationship:

(37)

Page 37 of 75

Conceptual model

Figure 7 provides an overview of the expected relationship, either positive or negative, between preventive maintenance and the various type of production costs. A negative relationship, indicated by a minus sign, implies that an increase in preventive maintenance will decrease that typical cost aspect. A positive relationship is shown by a plus sign, which means that an increase in preventive maintenance will result in an increase in that typical cost aspect. In Figure 7, preventive maintenance, either time-based maintenance (TBM) or condition-based maintenance (CBM), is compared with corrective maintenance. The framework in Table 3, which will be discussed in Section 4.2, shows the strength of the relationship.

Preventive maintenance (TBM & CBM) Overhead costs Direct labor costs Indirect labor costs Inventory costs Spare parts inventory costs

+

Spare parts costs single-component Spare parts costs

multi-component Quality control costs Raw material costs Maintenance labor costs Maintenance tools/equipment costs

(38)

Page 38 of 75

Production costs Philips Drachten

The input for this section is derived from both an interview held with the finance manager of Philips Drachten and from consulting financial charts of Philips Drachten. At Philips Drachten, the costs of the factory are given in a profit and loss (P&L) chart. In this chart, the actual costs are compared with the budgeted costs, in other words, the expectation (budgeted costs) is shown against the reality (actual costs). The budgeted costs are based on previous expenses from previous years and expected investments or budget cuts.

The different types of production costs that Philips Drachten distinguishes in their P&L are similar to the costs described in the previous sections. Philips Drachten only divides the costs in a smaller number of categories, namely five different categories. The first category is called material. In this category, the bill of materials (BOM) indicate what type of materials are needed for that product. These costs reflect the sum of the materials used per product multiplied by the quantity produced during a specific month. This gives the material costs for that certain month. It should be noted that the costs for scrap are also included in these costs. Philips Drachten holds a certain factor for scrap that is included in the amount of material needed to produce one product.

The second category is direct labor. This category is similar to the category described in Section 4.1.2. In the case of Philips Drachten, direct labor includes the costs for salary, pension, and insurances. These costs are budgeted and are again compared with the actual costs in the P&L chart. The results show whether more or less money, compared to the budget, is spent on this production cost.

The costs for indirect labor is the third category. Within Philips Drachten, the costs of indirect labor are the costs of the supportive departments. These include logistics, cost engineering, finance, human resources, IT, quality, and maintenance. Furthermore, Philips Drachten also has costs for personnel, especially maintenance personnel, who are on standby. These costs were not mentioned before but are an important contribution to this research. This is because having a preventive maintenance policy will probably reduce these costs because less technicians on standby are needed. The reason for having technicians on standby is that an unexpected failure could happen during the time of the day that the required resources are not available, e.g. at night or in the weekends. Reducing the chance of having a failure will reduce the number of personnel on standby.

(39)

Page 39 of 75 lifetime. This results in a certain cost per month. Comparing to our classification of production costs, this category is included in the overheads.

Assessing the impact of maintenance on the production costs