Master thesis at Voith Leeuwarden Improvement of the availability of trains and the delivery reliability of corrective maintenance

the delivery reliability of corrective maintenance

Master thesis at Voith Leeuwarden

Improvement of the availability of trains and

the delivery reliability of corrective maintenance

Report date: August, 16th 2010

Version: 1.1 (final)

Student: Jacob Strikwerda Student number: S1831526

Address: Amelandsdwinger 337

Zip code: 8911 LR

Place: Leeuwarden

Telephone number: 06 43 04 19 76

E-mail address: jacobstrikwerda@gmail.com

Organization: Voith Railservices B.V. location Leeuwarden Organization mentor: Rob de Boer

Function: Location leader Leeuwarden

Address: Snekertrekweg 19

Zip code: 8912 AA

Place: Leeuwarden

Telephone number: 058-2804870

E-mail address: rob.de.boer@voith.com University mentor (1): dr. G.A. Welker

Address: Nettelbosje 2

Zip code: 9747 AE

Place: Groningen

Telephone number: 050 363 7020 E-mail address: g.a.welker@rug.nl University mentor (2): drs. G.D. Soepenberg

Address: Nettelbosje 2

Zip code: 9747 AE

Place: Groningen

Telephone number: 050 363 7020

maintenance, which result in a large backlog. At the same time the availability of trains was lower than the demanded 93%. Both customer performances needed to be improved.

The purpose of this research is to give advice on how to improve the planning process and spare part management in order to improve the availability of trains and the delivery reliability of corrective maintenance.

Important causes of the bad performances on the delivery reliability and the availability of trains are the availability of repair capacities and the availability of spare parts. The availability of spare parts is in the current situation negatively influenced by long delivery lead times and incorrect inventory levels. The availability of repair capacities is negatively influenced by the way order information is processed and consequently the way orders are released and controlled.

Improving order information processing leads to better planning decisions which affect the delivery reliability and the availability of trains. The use of clear planning rules regarding the throughput time of work orders leads to a better delivery reliability.

Delivery lead times need to be decreased by managing the relationship with suppliers. At the same time improvements in the inventory levels must be made when lead times are still to long. Both improvements result in a higher availability of spare parts. In order to decrease lead times and manage inventory levels a supply chain management function must be installed central in the organization.

Index

2.1 The basic form ... 10

2.2 The information flow... 11

3. Problem statement... 15

4. Research design... 16

4.1 Spare parts management ... 16

4.2 Planning of repair capacities ... 16

4.3 Sub-questions ... 18

4.4 Research strategy ... 18

5. Further description and analysis of the current situation... 20

5.1 Availability of spare parts ... 20

5.1.1 Inventory levels ... 20

5.1.2 Concluding remarks inventory levels ... 22

5.1.3 (Delivery) lead time ... 22

5.1.4 Remarks about delivery lead time... 23

5.2 Availability of repair capacities... 24

5.2.1 Planning conditions and rules ... 24

5.2.2 Planning decisions (figure 11) ... 24

5.2.3 Planning problems... 26

5.2.4 Order processing ... 27

6.1.1 Delivery reliability... 30

6.1.2 Availability of trains ... 31

6.2 New found relations ... 31

7 Redesign and recommendations... 33

7.1 Spare part management... 33

7.1.1 Inventory level management ... 33

7.1.2 Delivery lead time... 34

7.2 Planning of repair capacities ... 37

7.3 Overall blueprint for the organization... 40

7.4 Implemented solutions ... 41

7.5 Reflection on the research ... 41

6

1. Introduction

End 2009 Voith Leeuwarden experienced an increase of work orders and at the same time the availability of trains was lower than the demanded 93%. The increase led to a growing backlog which was not a surprise because the warranty period of the trains expired. The manager of Voith

Leeuwarden concluded that the planning of repair capacities - personnel, docks and trains - and the management of spare parts must be changed in order to perform better.

1.1 Company background

Voith Railservices B.V. started in 2006 with the maintenance of trains in Leeuwarden. Trains owned by Arriva. This was a pilot for Voith. In 2007 another site was opened in Venlo. These relatively fast growth was due to the privatization in the public transport sector demanded by the government. In Leeuwarden there was a contract made with the manufacturer of the trains, Stadler, and Arriva which stipulated that Stadler provided a warranty of 3 years. In practice it meant that Stadler did most part of the problem fixing. Only small problems and preventive maintenance were the responsibility of Voith.

1.2 management problem

Voith focused in the beginning on preventive maintenance and some minor corrective maintenance. Modifications and time consuming problems, these were most of the time warranty issues, were the responsibility of Stadler, the manufacturer of the trains. In January 2010 the warranty period was over for almost all the trains. This led to an increase of tasks. Figure 1 shows the increase of the amount of orders for Voith.

Figure 1 the amount of orders for Voith and Stadler from August 2009 until February 2010 (per month)

7 regularly returning problems.

Planning uses the preventive schedule as starting point. Corrective maintenance is only integrated for time consuming repairs. The maintenance is planned without knowing the amount of time needed to perform the tasks. Non-extensive corrective maintenance is not considered by the planning. When there is time it will be fixed. Because more and more tasks are assigned to Voith more and more tasks are deferred because of lack of time. Also the trains become more and more unavailable for operations, the transportation of passengers by Arriva.

Spare parts are necessary for maintenance. Voith started with a basic stock recommended by the manufacturer and updated this with own insights. This basic stock is not sufficient anymore because the corrective maintenance is now more and more the responsibility of Voith. This led to a changing demand on existing spare parts and growing demand on “new” spare parts. As a result the problems with stock-outs (no spare parts available) are growing.

The current planning method resulted in less work done on-time and a growing amount of deferred maintenance. The management of Voith believes that improvements in spare part management and a better planning process can lead to a higher availability of trains and a decreasing amount of deferred maintenance.

1.3 research approach

This research could help in giving advice about how to adapt to these changing demand regarding spare parts management and the planning of repair capacities. For this research a case-study strategy is chosen as this research wants to study a phenomenon in its real-life context. Interviews, data analyses and observations are instruments that are used during the research.

The problem exploration is described in detail in chapter 2. The research design is described in chapter 3 and represents the methods used during this research. The research strategy can be found in chapter 4. The report proceeds with a description of the current situation followed by the analyses of the problems and causes. Recommendations and a redesign of the organization is given in chapter 7.

8

2. Problem exploration

This problem exploration is made in order to sharpen the overview of the problems. The problems and causes are related to each other and the demanded performances are made clear.

Voith must assure that 93% of the trains are available for Arriva during rush hour. In the morning and evening there is a rush hour-period according to the contract. An availability of 93% means that 43.7 of the 47 trains must be available for operations of Arriva. This performance is not met for a longer period. The availability was too low for almost every rush-hour-period in January (Figure 2 and 3). Trains in red cannot operate because they are broken. Trains in blue cannot be used because they are maintained at that moment.

Figure 2, availability of number of trains per day at morning rush hour for January 2010

9 marked with a priority, based on an agreement with Arriva, and a time-period is coupled to the priority. This leads to a certain due date. The time-period per priority is the delivery speed contractual agreed with Arriva. Consequently a delivery time is set; reporting date + time-period. Meeting due dates can have an influence on the average availability. A train cannot be used if a problem need to be fixed and the train is at Voith. Thus, the average availability is affected. The performance on the delivery reliability for corrective maintenance is that 29% of the orders is done on-time. A histogram (figure 13) of those orders shows that most orders are between 10 and 5 days too early. Most orders that are too late are 5 to 10 days too late.

Figure 4: histogram of lateness (priority 1,2 and 3) of corrective maintenance from January 1, 2010 until May 17, 2010. Based on finished work orders.

10 0 5 10 15 20 25 30 35 40 45 Ja n-08 Apr -08 Jul-0 8 Oct -08 Ja n-09 Apr -09 Jul-0 9 Oct -09 Ja n-10 Apr -10 Jul-1 0 Oct -10

Aantal WO's met stock-out

Linear (Aantal WO's met stock-out)

Figure 5, stock-outs (Wo’s = workorders) per month

The blue line in figure 5 shows the amount of stock-outs per month. The black line shows the

expected trend. Because of the shift of work from Stadler to Voith the expectation is that the number of stock-outs grows not linear but exponential the upcoming months. This figure only contains some of the stock-out information because not all stock-out information has been stored or is deleted after the spare parts became available.

A closer insight into the current process of planning resources for maintenance and managing spare parts is necessary to achieve a good problem statement and research design. Subject of interest are the goods flow (operations management scan, 2008) , information flow and organizational setting. A goods flow diagram (or basic form) outlines the goods flow, the activities executed in the

transformation process, the sequence of these activities and the stock points (queues and decoupling points). Which are helpful to clarify how the spare parts and trains flow through the organization in order to achieve availability of trains and/ or delivery reliability. The information flow is important as it triggers the next step in the process of finding a problem until fixing it. Examples of triggers are: the customer that reports a problem or a mechanic that requests materials. The organization setting is also important as there are actors, e.g. mechanics or planners, involved that influence the

performances.

2.1 The basic form

11

Figure 6, goods flow at Voith

2.2 The information flow

The information flow is described in order to give insight in how the goods flow is triggered and managed.

Trains need maintenance according to a predetermined maintenance schedule. The type of

maintenance is based on the amount of driven km’s and the operations hours of the engines. Arriva sends every day the mileage of the trains and the operation hours of the engines. The planner enters this information in the Excel file Actual State of Trains (figure 7). The file then advices which trains need preventive maintenance.

The planner uses the advice about the trains that need preventive maintenance and makes a Week Plan in Excel. This Week Plan shows which trains are coming when, the mechanics that are needed and which docks are used. This Week Plan is discussed with the Shift Leaders and with the planners of Arriva and then finalized.

A train driver can fill in a trouble form when he notices a problem with a train. The trouble form is send to Voith and is picked up by the Shift Leaders that use the information to fill in the Status Report in Excel. The Shift Leaders use the Status Report to manage the service mechanic in

Groningen. This mechanic does the first inspection when problems arise and tries to solve problems immediately if possible. Arriva uses the Status Report to monitor the status of the trains that are for maintenance at Voith or the trains that need service.

12 Corrective maintenance arises when problems are discovered by the train drivers of Arriva or

mechanics of Voith . The trouble forms of the train drivers are translated into work orders in Datastream. This is done by the Administration. Mechanics and Shift Leaders can make work orders in Datastream when they find problems at the trains.

Corrective maintenance can range from replacement of a sticker until replacement of a complete axle. If corrective work takes a long time or when third parties are necessary the jobs are planned by the planner and entered into the Week Plan. The Planner and the Order Engineer use the

information in Datastream to plan the larger corrective jobs.

The smaller corrective work orders are derived from Datastream by the Shift Leaders of the maintenance department just before the train is at the docks. The smaller jobs are assigned to mechanics by the Work Engineer. This is done in Datastream.

Information about the spend hours, used materials and chosen solution is stored in the work orders in Datastream. A mechanic enters this information. The service mechanic at Groningen also reports back by phone to the Shift Leaders. The Shift Leader then updates the Status Report. Information about the spare parts that must be repaired or cleaned is not stored.

When work orders are filled with correct data by the mechanics the status is changed into Ready for Check. This is the sign for the Shift Leaders to check the work order. When the work order is ok, the status is set to Ready for Admin. Then the Administration checks the work orders again! Afterwards the Administration sets the work order op Finished.

Finished work orders are send to Navision automatically. Navision is a ERP system. Voith uses it to handle the financial transactions of the work orders. Navision also adjusts the stock level of the spare parts. The stock level is also adjusted when spare parts are received by the Warehouse.

13

Figure 7, information flow. The arrows indicate in which direction information flows between software, excel sheets and functions.

Remarks about the information flow

There are 3 different information flows that assign work to the maintenance department.

The feedback information is entered into Datastream and the Status Rapport. It means double work. Information of broken but repairable spare parts is not stored. The lack of a central planning and control leads to inefficient information processing and fragmented information for all actors involved. As a result a lot of coordination is necessary between all the actors.

The work order information is only send to Navision when orders are completed by the

14 Organizational setting

The spare parts and information flows through the organization. This is influenced by certain actors. These are outlined in table 1 and their responsibility is also given in order to clarify which influence they have on the goods and information flow.

Function Responsibility

Warehouse employee Replenishment of spare parts.

Receiving parts and stocking them Sending defect parts

Mechanic Maintain the train. Report feedback.

Shift Leader Manage mechanics, check work and feedback of

mechanic, judge new reported problems.

Planner Plan preventive and extensive corrective

maintenance.

Order engineer Plan non-extensive maintenance and assure

availability of spare parts and tools.

Prepare special projects (Engineer to order)

Table 1 organizational setting

Remarks about the organization setting

15

3. Problem statement

In this chapter a problem statement is given. The problem statement includes the research objective and the research questions which are used to shape the research.

Voith must provide an availability of 93% of the trains for Arriva during rush-hours. Also important is the delivery reliability regarding problems found on the train. But most important is that at least 93% of the trains are available for Arriva for operations.

As the current management of spare parts and the planning of resources do not lead to a good performance on delivery reliability, it is not even set and monitored, and an availability of trains it is important to adjust these processes to assure a good performance. Delivery reliability is only important for corrective maintenance. Preventive maintenance does not have a set due date, it only affects the availability of trains. Simchi-Levi et.al. (2007) found that best-of-class performances are at least 94% regarding delivery performance to request. Voith wants to be the best-of-class and

therefore this is an appropriate goal. Research objective

Objective of this research is to give advice on how to improve the planning of repair capacities and the management of spare parts in order to guarantee a 93% availability of trains and a delivery reliability of 94% on corrective maintenance.

The advice will be presented in a report for the management of Voith Railservices. Research questions

How should Voith manage the spare parts in order to assure a higher delivery reliability of corrective maintenance and availability of trains?

How should Voith plan the repair capacities in order to assure a higher availability of trains and delivery reliability of corrective maintenance?

To find answers on the research questions the following questions must be answered: - What is the current situation (described in chapter 2 and 3)?

- What are influencing factors? The analysis of the current situation.

16

4. Research design

In this chapter the research design is discussed. The influencing factors of the performances on delivery reliability and availability of trains are discussed and represented in a conceptual research model.

In order to maintain a train docks, personnel, spare parts and the trains itself are needed. Docks and personnel have their restrictions in available time. These repair capacities are managed by the Planning. Planning also considers the availability of trains. Spare part management is done by the Warehouse.

4.1 Spare parts management

A train consists of many components. Each component consists of many parts. If one of the critical parts, parts that could cause a safety or environmental problem, is defect, a train may not be used for transporting passengers. Rustenberg et.al. (2000) say that if a system's component fails and it cannot be replaced immediately, due to a lack of spares, the system becomes unavailable until a new component is installed. Sleptchenko et.al. (2003) state that the availability of repairable technical systems, a train or it components, depends on the availability of (repairable) spare parts, to be influenced by (1) inventory levels and (2) repair capacity. Thus, the availability of trains depends on the availability of spare parts and the ability to replace or repair them. Sleptchenko et. al. (2003) also found that system availability, a train, can be improved by increasing spare part inventories to buffer against repair shop throughput times. Repair shop throughput time is the time from the moment a train cannot transport passengers until the moment it can transport passengers.

At Voith there are two important reasons why inventory is necessary:

1. Unexpected changes in customer demand; the demand patterns for corrective maintenance are irregular and sometimes lumpy.

2. Lead times; the delivery lead time is longer than when the parts are needed.

Demand cannot be influenced by spare parts management. But the delivery lead time is a reason that could be influenced. Important for Voith is how they manage delivery lead times as the delivery lead time could have an effect on the availability of spare parts.

Having the right spare parts available can lead to a higher availability of trains because waiting time for the procurement of a spare part is not necessary and the repair can be done earlier, which result in an earlier repaired train and Arriva can use the train earlier for operations. As well as the

availability of trains the delivery reliability can be higher because a problem can be fixed earlier because there is no waiting time for spare parts (figure 8). Holding more inventory or lowering the delivery lead times could lead to a higher availability of spare parts.

4.2 Planning of repair capacities

17 set delivery time: the repair must be done or “delivered”. Slack and Lewis (2008) call this delivery reliability: being able to honor the promised agreements about delivery time and speed.

Capacities can be unavailable when needed. E.g. the trains are not available for repair because they are used by Arriva, the docks are occupied or there is not enough personnel to maintain the trains. The match of these capacities with the maintenance that is demanded influences the availability of the trains and the lateness of orders. E.g. if there are no docks available, the trains cannot be repaired and therefore cannot operate. No personnel available means that trains cannot be repaired and thus cannot operate. Thus, the better the availability of capacities the higher the availability of the trains could be and the better the delivery reliability could be.

Planning rules

Planning must check if more than 93% of the trains are available when they plan trains for

maintenance. Furthermore they have to plan corrective maintenance so that the problems are fixed before the due dates expired. These planning rules influence the repair capacities. When a train cannot be maintained because Arriva needs it for transporting travelers the availability of repair capacities is lower. Thus, the planning rules influence the planning decisions which has an influence on the availability of the repair capacities.

In figure 8 the conceptual model is represented.

Figure 8, conceptual model

18 of orders that are fulfilled on or before the requested date. Planning rules have an effect on the planning decisions and this influences the availability of repair capacities. The availability of spare parts is influenced by the level of the spare parts inventory and the delivery lead time. Inventory management and lead time management are part of spare parts management. Planning rules and decisions are part of repair capacity planning.

4.3 Sub-questions

The following sub-questions help with proving the conceptual model and improving the current situation.

- how is the spare part management and the planning of repair capacities currently organized? - what factors influence the spare part management and the planning of repair capacities? - how can spare part management and planning of capacities be improved, so that the

availability of trains and delivery reliability is better?

4.4 Research strategy

This sub-chapter describes the used research methods and why they are used.

This research is done at Voith Leeuwarden and its purpose is to find improvements for Voith Leeuwarden. Welker and Broekhuis (2009) state that “a case study is a logical research strategy if you want to study a phenomenon in its real-life context; studying the actual practice and

understanding the phenomenon”. Because we want to understand the current situation it is important to reflect it with known literature on spare part management and planning of capacities. Therefore literature is used to clarify situations.

Observations are made to describe the current situation, data is analyzed to give a closer insight into problems and interviews are held to clarify situations. Table 2 shows an oversight of how data is gathered.

Variable Instrument Where/ who

Inventory levels Interviews

Data analysis.

Warehouse employee, Order Engineer, Maintenance engineer, Financial controller. Configuration and usage of ERP-software.

Delivery lead times Interviews Data analysis

Warehouse employee Purchasing data in the ERP-software.

Planning decisions and rules Interviews

Observation

19 Interviews

Data analysis.

Warehouse employee. Reported stock-outs.

Availability of repair capacities Interviews Planner, Shift Leaders, Location manager.

Availability of trains Data analysis. Interviews

Status Reports (Excel). Maintenance engineer, Location manager. Delivery reliability Data analysis Work order reporting and

finishing dates compared with the due dates.

20

5. Further description and analysis of the current situation

In chapter 2 a start with the description of the current situation is made. This chapter continues from that point and certain findings are further explained in this chapter. First spare part management will be discussed and secondly the planning of repair capacities.

Spare part management includes the inventory levels and lead time of spare parts and the availability of spare parts. The planning of repair capacities includes the planning rules, decisions and the availability of repair capacities.

5.1 Availability of spare parts

Already known from the problem exploration is that inventory management is based on experience. Voith started with an initial stock that was based on an advice of Stadler and own expectations. Now the Warehouse employee determines the inventory level based on experience. The amount of parts purchased and the moment of purchasing are based on experience or on request of the mechanics, the order engineer or maintenance engineers. A prominent cause of the experienced based purchasing process is that there is no inventory management system used.

5.1.1 Inventory levels

The inventory levels influence the availability of spare parts. The situation regarding the management of inventory levels is discussed here.

The Warehouse employee does not rely on Navision because the necessary parameters regarding inventory and purchasing are not set. Which finds its origin in the way Voith assigned inventory management responsibilities. The Warehouse employee is responsible for replenishment activities, but the responsibility for determining the needed inventory is not addressed.

There are more problems with inventory levels. The usage of a spare part cannot be booked on a daily base because Datastream reports that the inventory level is lower than zero. About 25 corrections per month have to be made because the physical inventory level is higher than the inventory level in Navision.

The following causes are found: - conflicting information systems - special contractual agreements - cannibalizing

- defect part registration - received parts registration - no forecasting used Conflicting information systems

21 (figure 9).

Figure 9, lag in days between work-done (status: Ready for Check) and finishing of orders in Datastream (status: Finished)

Table 3 shows a simplified example of the inventory levels and how received and used items influence the inventory levels.

Day Physical Datastream Navision

Physical used Received 1 5 5 5 2 3 3 5 2 3 3 3 5 4 3 3 5 5 8 3 8 5 6 8 8 8

Table 3 the stock levels that do not match (an order is finished in Datastream on average the 4th day)

Special contractual agreements

Voith allows Stadler to use spare parts from its Warehouse. Stadler must replace these parts within 4 months. The usage of spare parts by Stadler from the Voith Warehouse is not booked. As a result the amount of physical parts is lower than the stock level in the ERP system. Again corrections need to be made when periodic counting is done for the accountant. Voith has two options to pay the use of the spare parts, they can deliver a replacing part or they can just pay.

Cannibalizing

When a spare part is stripped from a component, it is called cannibalization. An example of cannibalizing is using a hard drive disk from a computer. As a result the component is not usable because it is not complete. Incomplete components are not registered and are still seen in the information systems as usable. The chance of a stock-out is higher.

Defect parts

22 Receiving a part next week could mean that there is no purchasing necessary. Voith does not have a good insight in the defect parts because the parts that are send via Stadler because of warranty issues are not registered and parts that are send by the Warehouse are only registered on paper. Received parts

Not all parts that are received are booked. This is caused by the Warehouse employee who does not always registers a received part when it is received because of warranty or repair. And it happens that goods are already used by a mechanic and there is no delivery note and the Warehouse employee thinks he gets less delivered while there were more delivered.

Forecasting

The Warehouse does not receive forecasts for the preventive and corrective maintenance. Only when stock-outs occur the demand is known or when someone asks to order parts because of a special project. The missing forecast leads to excessive inventory levels for some parts while other parts have a too low inventory level. The inventory level is managed by visually checking the regular used items.

Availability of spare parts and availability of trains

Information about stock-outs cannot be linked to the non-availability of trains. In most cases the non-availability of trains is caused by engine problems. An engine is a large component that cannot be replaced at once. This is a time-consuming process and a spare engine is not always available because of the high value or because the spare engine is out for repair. An engine as a whole is not the responsibility of spare part management. These “projects” are the responsibility of the

Maintenance Engineers. In theory spare part management can have a positive influence on the availability of trains. But engines are not part of spare part management and it is not important to improve the availability of spare parts in order to improve the availability of trains.

5.1.2 Concluding remarks inventory levels

The inventory levels are not correct because there is no forecasting used, the information systems mismatch and purchasing is based on experience. Which results in stock-outs for some spare parts, and possible high inventory levels for other spare parts. The way tasks and responsibilities regarding inventory management are set, are the core problem of the mentioned problems.

In practice there is no link between the availability of spare parts and the availability of trains. But in theory it still can. For the critical spare parts, that affect the availability, the stock-levels are

sufficient, and maybe too high.

5.1.3 (Delivery) lead time

One of the influencing factors on the availability of spare parts could be the delivery time. The current situation regarding lead times is discussed in this chapter.

23 ordered spare parts are not yet at Voith. This is a snapshot of the stock-outs at May 25, 2010.

Figure 10 histogram of lateness caused by stock-outs snapshot of 25-05-2010

The history for the past years about stock-outs is not known. Discussions with Order Engineer and Warehouse manager showed that stock-outs always were the case. In May, June and July 2010 the number of stock-outs is between 30 and 35 orders. The increase in stock-outs as expected in figure 4 is happening.

At Voith there are a few tasks, responsibilities and authorities assigned to manage purchasing and inventory activities. These are shown in table 4.

Task Responsibility Authority

Replenishment of known parts Warehouse employee < € 500

Purchasing of new parts Warehouse employee < € 500 always based on an offer of the supplier

Advice about inventory level about “new”parts

Order Engineer Only advice

Inventory level management Warehouse employee No agreements made.

Table 4 organization of the spare part management functions

The basic tasks of replenishing are integrated in the organization. But there are no tasks assigned regarding the selection of suppliers or negotiation with suppliers. Inventory is managed by the Warehouse Employee but he does not make calculations about what the proper inventory level is. Thus there is a lack in the organization of the inventory management and purchasing activities.

5.1.4 Remarks about delivery lead time

24

5.2 Availability of repair capacities

An important influences of the customer performances availability of trains and delivery reliability is the availability of repair capacities.

The availability of repair capacities is influenced by the planning decisions that are taken by the planner, Shift Leaders and mechanics. The planning rules also influence the planning decisions. The current situation regarding the planning of repair capacities is described in this chapter.

5.2.1 Planning conditions and rules

There are many conditions that limit the availability of capacities. These conditions need to be checked during the planning process. But first there are two main planning rules that are derived from the contract with the customer Arriva.

Planning rules

First priority is to assure more than 93% availability of trains. Which means that preventive maintenance has to performed as the train may not drive when it is not maintained. Problems that reduce the availability are also top priority as they can affect the performance.

Second priority is to fix problems before their due date. In this stage the availability of trains must also be considered.

Planning conditions

Preventive maintenance is not allowed during the two rush hour periods. The rush hour periods are from 06:00am till 09:00am and from 15:00pm till 18:00pm. It is allowed to carry out corrective maintenance during rush hours. This means that for corrective maintenance the first train is

delivered at 9am, but when there is a lot of corrective maintenance to do the train can be demanded from 7am by Voith.

Voith has two rail docks in the Workshop. One of the docks (A) allows to work beneath the train without the need of lifting the train up. The other one (B) is a regular rail track. Trains leaf the docks the same way as they entered the docks. The difference in docks leads to an important rule: the preventive maintenance has to be on dock A as the mechanics need to work beneath the train. The maintenance department works in 2 shifts. From 7:00am until 03:45pm and the late-shift starts at 02:00pm until 10:45pm. So, there is an overlap of almost 2 hours. Shift Leaders use this time to transfer their work and the extra shift leader can perform other tasks that are necessary while the other one leads the shift. The second shift is a lot smaller in terms of personnel.

The already mentioned performance of 93% or higher is also a planning rule. Voith cannot demand all trains but has to assure a certain availability. Therefore, the first priority is to have more than 93% of the trains available.

5.2.2 Planning decisions (figure 11)

25 advice of the Excel Plan Actual State of the train (figure 11). This plan is based on the preventive schedule as agreed with the customer and the manufacturer of the train. Mechanics are assigned to the preventive jobs and the extensive corrective jobs. The amount of time that is reserved to finish these jobs is based on experiences and estimations of the Planner. The availability of docks is considered during the planning process. The conceptual plan is discussed with the Shift Leader on Thursdays and adjusted if necessary. On Friday the conceptual plan is approved by the traffic controllers of Arriva and is finalized. The traffic controllers already receive the conceptual plan on Thursday after the adjustments of the Shift Leaders are made. When important changes are made, Arriva also receives an update.

The non extensive corrective jobs are assigned to mechanics by the Order Engineer via Datastream after the Week Plan is finalized. Thus the planner does not have good insights in what is happening after the work is planned. This method results in two ways of assigning work to the Maintenance department. One way via Excel, and the other way via Datastream.

The service department in Groningen handles problems that restrict the operations of trains. They receive their work assignments by phone and via the Status Report in Excel. This is managed by the coordinating Shift Leader. They report back via Datastream and phone to the coordinator.

26

Figure 11 information flow

5.2.3 Planning problems

During the research a lot of problems relating to planning decision making were found. These are described in this paragraph.

The Planner and the Order Engineer do not have a good insight in the due date of a corrective maintenance order because Datastream only works with reporting dates and priorities. In practice also the reporting date is not judged. A short view of the total list of problems by the Planner leads to the decision of planning a problem or doing nothing. Thus, there is no report date sequencing used! A good example of a planning problem is the project “paint damage”. For this project there are not enough docks available because the trains need to stay for 48 hours in a dry room. Therefore this project is not executed and now the customer complaints why it takes so long. On the other hand the customer does not want to hand over the trains because they need them for operations. Finding the right balance between the availability of docks and trains is very difficult and executing this project right now results in a better delivery performance, but a lower availability of repair capacities (trains and docks) is the consequence.

27 perform the corrective maintenance.

The capacity problems are long-term problems that needs to be managed by the location manager and management team because they decide how much personnel to hire. This is calculated on the basis of assumptions about the preventive maintenance. But the corrective maintenance tasks cannot be judged that well because there is no information available. Voith just started with handling all the new orders after the warranty was over so this is also based on rough estimations. This information is also needed by the planning. The way this information is available is described in the next section.

The histogram (figure 12) about the lateness shows that orders with a low priority (95 days fixing time is used to make the situation clearer) are almost always finished the same day (-95 days) as the problems were found which means that the capacity of the personnel is used for less important problems and therefore cannot be used for orders that are more important. Trains are only available for a certain time-slot. It means that when doing work orders with a lower priority earlier then higher prioritized orders the train has to leave, while orders with an earlier due date are not completed.

Figure 12: histogram of lateness of corrective maintenance from January 1, 2010 until May 17, 2010. With priority 4 set to a repair time of 100 days. Based on finished work orders.

5.2.4 Order processing

Demand is known to the Planning and later on the Maintenance Department via the ordering process. The ordering process influences the decision making process of the Planner, Shift Leaders and mechanics. This process is described in this paragraph.

28 of business processes: people are working on a product or service, people are passing a product or service they worked on to a fellow worker or a customer, or people are interacting with a computer.

Figure 13 order processing started with customer. Scheduling is not the last step of the process but the last part left out because it does not add to the discussion.

Welker (2004) describes four order processing activities. These are: quotation, acceptance, order entry and scheduling. Arriva places a request via a problem-report because they found a problem and sends this to the Shift Leaders. The Shift Leader makes a judgment of the problem and categorizes it. In the quotation and acceptance processes there is no judgement made about the needed capacities. Thus, the information in Datastream is not usable for the Planner other than as a notification that something needs to be done. Then the problem-report is send to the Administration who enters it into Datastream. The entered orders go to the Planner/ Order Engineer and are scheduled by them. In case there are materials needed the Order Engineer checks if they are available.

29 14). Here the quotation and acceptance is done by the mechanic and sometimes the Shift Leader is consulted.

Figure 14 order processing started with mechanic. Scheduling is not the last step of the process but the last part left out because it does not add to the discussion.

5.2.5 Concluding remarks

The alignment of the planning is well organized between the Planner and the Shift Leaders of Voith and the Traffic Controllers of Arriva. The planner does not have a complete oversight because the Order Engineer plans non-extensive work after the planning is formalized. Information about the status of the train and the trouble forms from Arriva are entered in two systems. Thus, a double entry of the same data. The information of the driven km’s and operated hours are also entered in an excel while it is also automatically entered into Datastream.

30

6. Empirical model

A conclusion of the relations in the current situation is given in paragraph 6.1. And new found relations are discussed in chapter 6.2.

6.1 found relations

For spare part management and the planning of repair capacities the found relations are discussed.

6.1.1 Delivery reliability

The performance on delivery reliability is low, 29%. This is caused by the non-availability of mechanics, one of the repair capacities as described in paragraph 5.2.3.

Availability of repair capacities, planning decisions and rules

Not surprising as Sleptchenko et.al. (2003) already state that the availability of repairable technical systems, a train or it components, depends on repair capacity. In the conceptual model the influence of availability of repair capacities had an influence on the availability of trains and the delivery reliability. This relation is evident. This is also the case for the relationship between the availability of repair capacities and the availability of trains. Because there were no mechanics available in order to repair.

The availability of repair capacities was negatively influenced by the way the Planner, Order Engineer and the Maintenance Department decided to perform jobs or not. These wrong planning decisions are caused by non clarified planning rules and more important a lack of order information. This new found relation is discussed in paragraph 6.2.

Availability of spare parts, inventory levels and (delivery) lead times

From snapshots of the stock-outs it can be concluded that the (non) availability of spare parts does have an influence on the delivery reliability. How big this influence was could not be determined exactly as the data about stock-outs was not stored. The relationship between the availability of spare parts and the delivery reliability is logical as Sleptchenko et.al. (2003) already state that the availability of repairable technical systems, a train or it components, depends on the availability of (repairable) spare parts.

31 In practice there was no relation between the availability of spare parts and the availability of trains. This can be explained by the fact that failures on components are not fixable with new spare parts, but more by servicing the whole engine by a specialist for example. In these cases we do not have the technical capabilities, machines, and skilled personnel to fix the problem.

The planning decisions regarding the availability of trains are clear for the whole organization. First priority is to have enough trains available. That the performance is low is mostly caused by not functioning engines, which is out of scope (6.1.1). The planning decisions taken and consequently the availability of repair capacities are fully tuned to getting trains available.

6.2 New found relations

During the research there were also new relations found that influence the performances. These are discussed in this paragraph. Figure 15 is the conceptual model which is a graphical representation of the new discussed relations combined with the found relations in paragraph 6.1.

Order processing

The information that is processed (or not) leads to planning decisions made by the Planner and Order Engineer and the Maintenance Department. Currently the information is not processed correctly. Information about demanded due date and needed skills, working hours and materials is missing. Soepenberg et.al. (2008) state that production planning and control decisions relate to input control (order acceptance/delivery date promising, order release and priority dispatching) or output control (adjusting capacities) decisions. Without the correct order information the decision making is very difficult as can be seen at Voith. Thus better order processing leads to better planning decisions which as already described influences the delivery reliability via the availability of repair capacities. Tasks and responsibilities

32

33

7 Redesign and recommendations

In order to improve the availability of trains and the delivery reliability there are several improvement areas. They are divided in spare part management issues and planning & release oriented issues.

7.1 Spare part management

Spare part management must find a balance between inventory holding costs and unavailability costs according to Cavalieri et.al. (2008). Unavailability costs are costs made when trains cannot operate. But holding inventory is not the ultimate mean to reach the spare parts management goal. This is: supplying spare parts when demanded.

7.1.1 Inventory level management

Determining the correct inventory level is a complex process as many inputs are necessary. Inputs about the risk of a stock-out, the inventory holding costs, the transportation costs, delivery lead times, demand variability are the most important ones. Cavalieri et.al. (2008) made a decision-making framework for managing maintenance spare parts which takes all these influences into account. The objective of the paper is to provide a stepwise decision-making path in order to orienteer an industrial manager on how to pragmatically handle the management of maintenance spare parts in a company. This framework could support Voith by addressing all the relevant issues and actors that are involved in determining the correct inventory levels.

Framework for setting up inventory management of spare parts

The framework is structured into five sequential steps: part coding, part classification, part demand forecasting, stock management policy and policy test and validation.

Part coding is already done and present in the information systems, but needs an update as not every part is present in detail. Part classification must be done in terms of demand patterns, the re-usability of spare parts and the turnover of stock-items. The complete description is available in the paper. Demand forecasting is the third step. This is currently not possible as the Datastream and Navision do not share this information and the parameters about inventory and purchasing are not set correctly in Navision. But it could be possible if the interface between the two systems would be adjusted. Another important input for forecasting is the risk-level of a stock-out. This is information that needs to be delivered by Maintenance Engineers and suppliers as they can judge how long a part can work.

Cavalieri et.al. (2008) state that “Before putting in practice all the results of the previous phases of the decision-making framework, tests using a discrete-event simulation may be carried out to make sure that the selected policy is the most suitable one.”. This is an important security measure to assure a good implementation of the framework and the final step.

The paper of Cavalieri is concentrated on keeping the right level of inventory. But keeping inventory means taking more (financial) risks because inventory can become obsolete and it costs more money as it must be stored and managed. This risk is increased by the fact that demand at Voith is

34 customers’ orders and materialize high revenue in periods when demand is high, but firms often end up with excess inventory in periods when demand is low.

An important solution to reduce inventory risks is risk pooling. It is one of the most powerful tools used to address variability in the supply chain is the concept of risk pooling (Simchi-Levi, Kaminsky & Simchi-Levi, 2007). Risk pooling is aggregating demand across locations and it reduces the demand variability because it becomes more likely that high demand from one location (Leeuwarden) will be offset by lower demand of the other location (Venlo). At Voith there is an opportunity to apply risk pooling.

Virtual pooling

Especially virtual pooling can be an effective strategy as the locations Leeuwarden and Venlo already share the same inventory management system and use the same parts. Anupindi et.al. (2001) use the term “virtual pooling” to refer to the inherent pooling that exists when demand at one location can be met from inventory at another location through transshipments. Virtual pooling in the case of Voith is sharing inventory between the locations Leeuwarden and Venlo but storing the inventory at both or one location.

Inventory level information

To enable forecasting, Datastream, Navision and the interface between these two information systems needs to be adjusted. In essence: every transaction that leads to a lower or higher stock level must be registered almost instantaneously. Also important is the registration of defect spare parts. It is possible that these spare parts can be re-used after reparation or cleaning. Therefore the need of purchasing could be affected. Also warranty agreements can result in getting a new part without the need of purchasing.

Working with two types of stock is an option. The usable parts are put in the usable inventory. The parts that need further actions before they can be used must be put into a non-usable stock, otherwise it looks like an available part but actually it is not the case.

7.1.2 Delivery lead time

Keeping inventory is one solution, which also involves certain risks. But why keeping inventory when the needed parts can be there before the Maintenance department needs them! One of the main reasons to keep inventory is that supply lead time is longer than the demanded lead time of the customer. When the demand lead time is lower, the need to keep inventory can be lower. Thus, delivery lead time reduction can also have a positive effect on the height of the inventory levels which results in less inventory costs.

Delivery lead time reduction can be achieved by negotiating with suppliers about delivery lead times. These tasks are currently not assigned to actors within the organization!

35

Figure 16 the new organizational setting of spare part management

The yellow area’s are the decision areas and the responsible functions are mentioned (e.g.

maintenance engineer). In table 5 the functions and the issues they need to address are mentioned.

The management of Voith and Voith head office in Germany already know that there is a gap in the supply chain management function. Therefore a project is started to improve supply chain

36

Function Issue

Maintenance engineer Risk judgment of operation time of parts. This will be input for demand forecasting.

Supply chain manager (not assigned)

What can I agree with my suppliers regarding delivery times, holding inventory, transportation options?

What kind of demand is there for spare parts, seasonal influences, demand patterns, etc.?

What is the aggregated demand for the two sites? Warehouse employee Do I have space available to store all the parts? Maintenance department How much do we (think we) need?

Order engineer What is the forecast of spare parts in the short run?

Management How much budget is there available?

Table 5 Decisions areas and functions

As can be read in table 5 the amount of actors in the spare part management process is higher than only the Warehouse Employee and Order Engineer. The expert knowledge of Maintenance

37

7.2 Planning of repair capacities

Voith is, as described in chapter 2.1, a make-to-order company. But in practice they are not focused on delivering those ordered work. If it is the Planner, Shift Leader or Mechanic, nobody focuses completely on delivery dates. To improve the delivery reliability performance there are two major decision areas.

Input and output control decisions can influence both the average lateness and the variance of lateness (Soepenberg et.al, 2008) which are the measures of lateness. Figure 17 shows the input and output control decisions and the relation with the flow of orders through the organization.

Figure 17 input and output control decisions.

Acceptance and delivery date promising and medium term capacity control

The delivery date promising is fixed for Voith. The contractual agreements determine the type of priority and a due date is calculated from the report date + priority-time. But in the acceptance phase improvements can be made. By setting the correct report dates and priority the delivery date

promising becomes available for the Planner and Shift Leaders. In this acceptance phase the quantification of needed trains, docks, personnel in hours and spare part must be done. This

information helps the Planner to actually plan if there are no constraints that prevent the problem to be fixed.

This information is also necessary to determine the amount of capacity needed, which are the output control related.

Release and short term capacity control

38 Priority dispatching and daily capacity adjustments

When the orders are released the Shift Leaders take over the responsibility. In this Phase the Shift Leader also needs to know what the planning rules are and which order has a higher priority or not. This information is given by the Shift Leader itself and mechanics in the acceptance phase.

In the new production planning & control way it is better to force the maintenance department to execute what is demanded. Clear determination of the due dates is essential. Even important is that everyone knows what the planning rules are.

Order status

The decision steps are linked with the order status. From figure 17 a few order statuses can be derived. These are linked to the current order statuses in table 6.

Input/ output decision Current state Desired state

Entry Released Entered

Released Released Released

Processed Released In execution

Finished Finished Finished

Table 6 order statuses

In the current situation the orders that are entered, the released and the orders that are processed al have the same status. A better insight in which status orders really have can be realized through using an extra order status type. When orders are entered they receive the new order status “Entered”. When they are released the Planner puts them on “Released”. When they are executed they receive the status executed and when they are finished by the administration they receive the status finished.

Control mechanisms of order progress

Achieving high delivery reliability is a combination of controlling average throughput times and controlling the progress of individual orders (Land, 2004). Voith does not track the progress of individual orders and does not control the average throughput times regarding corrective

maintenance. Voith just sees what it can do and the rest is neglected. This is the case for planning and Shift Leaders. In the current situation the absences of control mechanisms is negatively influencing the delivery reliability. But when done correctly the control mechanisms can influence the delivery reliability positively. When entering the correct data into Datastream this system is a good tool to control what the progress of the work orders is.

39 Organisatie

Bedrijfsproces

= Voith

= Order bewaking

ACTOR ACTIVITY DIAGRAM

. . . . Customer Shift Leader Order engineer Planner Workorder request Datastream Werkorder request

Requested work orders

Acceptance and quotation

Service

High priority problems

Check/ repair Update information

Feedback Conditie= indien prio

hoog

Extra preparation

necessary indien nodig

Request for order engineer

Discussion/ action planning

Action information updated

Work order ready for Shift Leader when the train is at Voith Execution for

SL Management of work

Information update and work order finishing

Statusreport downloaden

Send statusreport

When not ready go back to Order Engineer

40

7.3 Overall blueprint for the organization

The changes in the planning process and spare part management result in a different basic and information flow (figure 19) than the old one.

Figure 19 the new information flow

41

7.4 Implemented solutions

To improve the control on the released work the Planner and Order engineer fill in the planned start date of an order in Datastream. This gives the Shift Leader a tool to control the progress of work. The Week-Plan is now used as a graphical tool and the Maintenance department can find the needed information in Datastream.

Problems regarding the registration of stock-outs and internal processing speed are resolved. Clear guidelines and better registration in Datastream leads to a much faster response of the Order Engineer and Warehouse Employee.

The Service Department had troubles with connecting with Datastream and therefore they did not enter information directly in Datastream. This problem is fixed by giving them the right internet connection at location Groningen. Now the information does not need to be entered by the Shift Leader in the Status Rapport in Excel. The double entering of information is already changed regarding the “Status Rapport”. The information can now be entered in Datastream and with a simple download a report is generated for the customer. Thus, no more double entry of information for the Shift Leader.

A temporary plan for registration of the returned parts is available, and a final solution with the registration in Navision is on the way.

In order to streamline all changes for the Order engineer, Planner and Shift Leaders a weekly meeting is set-up. This meeting is also used for discussion between the departments and leads in many cases to finding solution for other problems. In the past half year there were no regular meetings between the maintenance and support office department.

A project is started up by central management in order to improve the purchasing process. This project encompasses a lot of the issues addressed in chapter 7.1.

7.5 Reflection on the research

With this report the management has insight in the causes of bad performances. With the given advice the management has the ability to improve the performances on delivery reliability and availability of trains by removing the causes. The objective of this research was to give an advice on how to improve the planning of repair capacities and the management of spare parts in order to guarantee a 93% availability of trains and a delivery reliability of 94% on corrective maintenance. Therefore, the conclusion is that with this advice the management is able to improve the

42

References

Anupindi, R., Bassok, Y. and Zemel, E., A general framework for the study of decentralized

distribution systems, Manufacturing and Service Operations Management 3 (4) (2001), pp. 349–368. Cavalieria, S., Garettib, M., Macchib, M. and Pintob, R., A decision-making framework for managing maintenance spare parts, Production Planning & Control, Vol. 19, No. 4, June 2008, 379–396.

Department of Operations, Faculty of Economics and Business, University of Groningen, Operations Management Scan. Digital copy on request available.

Eynan, A. and Fouque, T., Capturing the Risk-Pooling Effect Through Demand Reshape, Management Science, Jun2003, Vol. 49 Issue 6, p704-717.

Hoekstra, S., Romme, J., Integral Logistic Structures: Developing Customer-oriented Goods Flow, McGraw-Hill, London, 1992.

Land, M.J., Workload control in job shops, grasping the tap, Ph.D. Thesis, University of Groningen, Labyrinth Publications, Ridderkerk, the Netherlands, 2004.

Rustenburg, W. D., van Houtum, G. J., Zijm, W. H. M., Spare parts management for technical systems: resupply of spare parts under limited budgets, IIE Transactions, Oct2000, Vol. 32 Issue 10, p1013-1026.

Simchi-Levi, D., Kaminsky, P. and Simchi-Levi, E., Designing and managing the Supply Chain: Concepts, Strategies and Case studies, Mcgraw-Hill, London, 2007.

Slack, N. and Lewis, M., Operations Strategy, Prentice Hall, London, 2008.

Sleptchenko, A., van der Heijden, M.C., van Harten, A., Trade-off between inventory and repair capacity in spare part networks, Journal of the Operational Research Society, 2003, Vol. 54 Issue 3, p263.

Soepenberg, G.D., Land, M., Gaalman, G., The order progress diagram: A supportive tool for

diagnosing delivery reliability performance in make-to-order companies, Int. J. Production Economics 112 (2008) 495–503.