Production Planning Redesign

(1)

TOMINGROEP

Production Planning

Redesign

A case study at Tomin Metaal

Name: Jaap van der Ende

Student No: S1668978

University: Rijksuniversiteit Groningen

Programme: Master Technology Management

First Supervisor: Dr. M. J. Land

Second Supervisor: Dr. J.A.C. Bokhorst Date: 22-8-2011

(2)

2

Preface

As a student of the Master Technology Management, on the University of Groningen, I was given the opportunity to perform my graduation assignment at Tomin Metaal. After the workshop of the Lean Operation Research Center, I contacted Jelle Postma (the business unit manager of Tomin Metaal) and he offered me an internship. For me, it was a challenge to redesign the production planning approach for a unique and interesting company which provides jobs for many people with a distance to the labour market. Besides the subject of this thesis, I learned a lot about the different business processes of this company.

I would like to thank Tomin Metaal and especially Jelle Postma for providing me with the opportunity of this internship and I am grateful for the ideas, thoughts, and cooperation of the employees. Furthermore I would like to thank my supervisor, Martin Land for his constructive feedback, useful tips and motivational words. But the most grateful I am for his flexibility and endless patience during this research. At last I want to thank my family and friends for their help and support during the research. Special thanks are for my parents, they support me greatly during my entire study.

(3)

3

Summary

This thesis presents a study of the production planning at Tomin Metaal. Current production planning decisions at Tomin Metaal are taken in an ad hoc and unformalized way. Several earlier research projects concluded that there is a need for a more structured production planning approach. This thesis discusses the redesign of the production planning decisions. The redesign has three main objectives. These objectives are:

 Reduction of employee dependency  Reduction of lead times

 Improvement delivery performance

A planning concept is selected on the basis of the production characteristics and the performance goals of the company. The selected concept is the workload control (WLC) concept. This concept is suitable for the production characteristics of Tomin Metaal and according to the PPC literature it has the potential to reduce lead times and to improve delivery performance. The WLC concept uses a decision framework which contains three hierarchical levels. These levels are: order entry, order release and order dispatching. The decision framework is used to analyze the shortcomings of current production planning decisions. For these shortcomings new solutions are made up based on production planning literature and applied to the specific situation of Tomin Metaal. The redesign pays special attention to the specific planning issues of Tomin Metaal. These issues are: the production of fast moving product family, the production of assembly orders and the production of rush orders.

The proposed redesign is supposed to improve the delivery performance and reduce lead times. The calculation of due dates will be more accurate and therefore the delivery performance will be improved, because planners can communicate more realistic due dates to customers. The release of orders is based on the workload of bottleneck operations. This procedure maximizes the capacity utilization of bottleneck operations and therefore reduces the lead time of orders. Orders are dispatched based on their operation due dates instead of their final due dates as in the current situation. This will improve the delivery performance of orders with relatively long lead times.

(4)

4

1 Introduction

1.1 The motive of the research

The motive of this research was given by the thesis of graduate student Anne-marie van Dijk. During her internship at Tomin Metaal she investigated how the production of Tomin Metaal could be made more transparent. Part of her thesis was about the production planning. She concluded that the planning process was not transparent in terms of communication during the planning process and in terms of production performance. These insights did decide the managing director of Tomin Metaal to investigate how to redesign the production planning.

1.2 The company

Tomin Metaal, the metal division, is one of the ten business units of the Tomingroep. Tomingroep offers industrial as well as personal services. These activities occur in the domain of retail and building construction, both in conformity with the local market.

The main goal of the company is to give people with a distance to the labour market a job. Because Tomingroep works with people that have problems to find a job in a regular company, it is called a social work company, also known as a sheltered workshop or social-enterprise. The major part of the employees has a contract based on the law of sheltered employment. This law is in Dutch called: Wet sociale werkvoorziening (Wsw). Besides this group of employees, there is also a group that participates in a reintegration project. The remaining employees are regular employees. This group of employees have a so called regular job. They mainly fulfil all supporting functions within Tomingroep at departments such as Personnel & Organization, Salary administration and the production offices.

1.3 The metal division

Tomin Metaal, the metal division of Tomingroep, has two locations. The biggest location is in Hilversum and a smaller workplace is located in Almere. The total amount of employees is about 70. This research concerns the production planning of the location in Hilversum. The core business of this location is the machining and assembly of metal parts. The customer base of Tomin Metaal exists of a large variety of customers which order products in batches. Machining plate metal parts generates the most income, but Tomin Metaal produces a wide variety of orders. For several customers they realize various machining processes for stainless steel, metal and aluminium. The currently delivered activities at Tomin Metal are:

 Pre-treatment (sawing, cutting)

 Plate machining process (punching, folding)  Drilling (turn-, milling cutter- and drill work)  Welding (MIG-, TIG-, spot welding & solder)

(7)

7

1.4 Research focus

This research is focused at the production planning of Tomin Metaal. The initiator of this research is the managing director of the division retail and industry. According to him Tomin Metaal has rapidly grown over the past few years but lacks the formalisation of production planning. This lack of formalisation results in a lot of problems which are described in chapter 2.

At the moment there are two people responsible for the production planning at Tomin Metaal, the production leader and the head of the production office. Figure 1 shows the part of the management hierarchy at Tomin Metaal concerning this research. At the beginning of this research project, it was well known for the Management of Tomin Metaal that the planning process had to improve, but which problems had to be solved and which problem had the highest priority to be improved was unknown. Therefore before the research design, research question and research goals are explicated, the problems concerning the production planning will be explored in the next chapter.

(8)

8

2 Problem exploration

This chapter describes which problems the production planning of Tomin Metaal suffers. Several problems will be mentioned. A part of these problems are exposed by earlier research at the production of Tomin Metaal. Other problems perceived by management and the employees and are exposed during interviews and conversations with them. The problem exploration is focused on two kinds of problems, namely: planning process problems and lacking performance. The problem exploration is based on two earlier executed research projects at Tomin Metaal, namely the internal audits of Maas (2010) and the Master Thesis of van Dijk (2008). The results from these researches are validated and supplemented with own research.

2.1 Lack of transparency

The Master Thesis of Van Dijk (2008) was focused on creating transparency in the production process of Tomin Metaal. Her findings were that the production process of Tomin Metaal has lost its transparency because the department did increase in sales but is still working with original production arrangements. The combination of non formal and ad hoc production arrangements with a large production process make the production process is not transparent anymore. A large part of the research how to create transparency at Tomin Metaal was about production planning.

According to van Dijk (2008) the planning process is not transparent, while transparency is needed to control the planning process. Causes for the lack of transparency are unstructured communication and decisions which are made on base of intuition. More transparency would make the production process better manageable because there will be understanding about what happens in the production process.

2.2 Inefficient planning process

(9)

9

2.3 Employee dependency

The production planning is executed by two employees and decisions are mainly made based on tacit knowledge1. Because there are no other employees at Tomin Metaal who could take over the planning tasks, the dependency on these employees is large. So the risk one of the planners is leaving the company is very large.

2.4 No feedback information

Another problem rising from the tacit knowledge based decisions is (according to the Managing Director) that he does not very little feedback information of the planning process. He announced that for him it is hard to manage the planning process with a shortage of feedback information.

2.5 Unevenly spread workload

Besides the management and the planners, the operators and foremen also announced some problems. According to them, they have to face unevenly spread workload with the current planning. Some weeks they have got very little work to do and some weeks they are very busy and overloaded with orders. Another problem in their eyes is that they often have to interrupt the production of orders because of rush orders. These interventions cause extra start up and shut down operations which reduce the productivity.

2.6 Summary

A lot of planning related problems have been exposed. The exposed problems come from one root cause namely, an unstructured and ad hoc planning process. An overview of the exposed problems, their mutual relations and their root cause is given through a casual diagram (figure 2). Unstructured Planning Process Little Feedback Information Lack of Transparency Little Responsibilities Assigned Employee Dependency Inefficient Planning Process Tacit Knowledge Redundant Tasks Little Information Administrated Unevenly Spread workload Reduced Productivity

Figure 2: Casual diagram of the planning related problems at Tomin Metaal

1_{Tacit information is held inside people’s heads. Tacit information is highly personal and hard to formalize. This}

(10)

10

3 Research approach

3.1 Redesign objectives

In the previous chapter several planning related problems are exposed. This research will not focus all problems that are exposed. According to the business unit manager, the employee dependency is a high risk and is the most urgent problem. One objective of the research is thus to reduce the employee dependency. The redesign of the planning process should reduce this problem as illustrated in casual diagram (figure 3).

Structured Planning Process Registered Information Explicit Knowledge Less Employee Dependency

Figure 3: Causal diagram of the research focus

In a more structured planning process information will be administrated, this leads to information and knowledge that is “hard” and not just in the heads of employees. In a structured planning more knowledge will be explicit; this explicit knowledge2 will make the process less dependent of the tacit knowledge of some planning employees.

Besides the employee dependency, the objective of the redesign of the production planning is to improve the production performance. According to Hopp and Spearman (2001) production planning strives for the strategic goals and the vision of a company. The vision of Tomin Metaal is to provide jobs for disabled people with the profitable production of metal products of high quality, short lead times and on time delivered. Production planning has the most influence on the time related goals. The production planning of Tomin Metaal should focus on these goals. The minimization of lead times and the maximization of the delivery performance are therefore two objectives of the production planning redesign. Summarized the redesign of the production planning at Tomin Metaal has three objectives. These objectives are:

 The reduction of the employee dependency  The reduction of the lead times of orders  The improvement of the delivery performance

2

(11)

11

3.2 Research design

The selection of a production planning concept depends on the production characteristics and the objectives it tries to satisfy (MacCarthy and Fernandes, 2000; Stevenson et al, 2005). The objectives of the production planning at Tomin Metaal are already known. The goals are the reduction of lead times and the improvement of the delivery performance. But the production characteristics of Tomin Metaal should be investigated. When the production characteristics are investigated and mapped, a production planning concept which suits the production characteristics and strives for the right performance goals should be selected. The current production planning of Tomin Metaal is ad hoc and unformalized. To structurally analyze all planning decisions, the current production planning will be described on hand of the selected planning concept. The shortcomings of the current production planning approach will be exposed and improvements for the shortcomings will be proposed. An overview of the research design is illustrated in the conceptual model of figure 4.

Production Characteristics Production Performance Goals PPC concept selection Improvement of

planning decsions Short lead times

Employee independent High delivery performance

Figure 4: Conceptual model of the production planning redesign

3.3 Thesis outline

To come to a new and improved production planning design the following phases will be gone through:

1. Investigation of the Production Characteristics

Besides the existing planning process also the production process characteristics will be mapped. The production process characteristics are investigated because the production planning design depends on the production characteristics. The investigation of the production characteristics will take place in chapter 4.

2. Production Planning Concept Selection

(12)

12

3. Analysis and Improvement of the current production planning decisions

The first objective of this phase is to identify shortcomings in the current production planning. The second objective is to give alternatives for these shortcomings. The selected planning concept will be used as framework to analyse and improve current approach to production planning. For each planning decision that has to be taken the following questions are answered:

1. How is this decision made in the current situation? 2. Which shortcomings does the current process have?

3. How should this decision be made according to the selected planning concept?

4. How should the planning concept be applied in the situation of Tomin Metaal?

The analysis and improvement of the current production planning decisions will take place in chapter 7.

4. Production Planning Redesign

In this phase a structured overview of the complete redesigned production planning will be presented. An overview of the redesigned production planning approach will be given in chapter 8.

3.4 Research restrictions and requirements

(13)

13

4 Production characteristics

The characteristics of the production influence the design of the production planning. For this reason the production characteristics of Tomin Metaal are analyzed and described. The characteristics that are investigated are the presented by literature. Several authors (e.g. Slack et al., 2007; Amaro et al., 1999; Oosterman et al., 2000) characterized a production process with different production characteristics. This chapter describes the following characteristics:

 Make to order/ make to stock  The customization degree  Product volume

 Product variety

 Process task complexity  Process flow

 The shop floor configuration

4.1 Make to order/ Make to stock.

Production systems could be characterized as make to order (MTO) or make to stock (MTS). The MTO systems offer a high variety of customer-specific products (Soman et al., 2004). The production of a MTO product starts when a customer places an order. MTS systems offer a low variety of products (Soman et al., 2004) and the production of a MTS product starts when inventory stocks fall below the reorder point

Over the year 2010 Tomin Metaal did serve 78 customers. The customers are mostly companies in the building and installation industries. Many of the customers are wholesalers, which often order large quantities. Most of these customers were served on a MTO basis. Four customers of Tomin Metaal are also served on a MTO basis. Tomin Metaal produces products on stock for several reasons.

 The customer orders repeatedly in low quantities. Tomin Metaal produces one large order and puts the rest of the products in stock for future demand.

 Some products require very little process times to be produced, but require relative long setup times. Tomin Metaal produces these orders in large batches to increase efficiency.

 Some customers desire short lead times. The customers share their forecasts and Tomin Metaal produces these products on stock.

(14)

14

4.2 Customization degree

Tomin Metaal has regular customers which order the same products repeatedly. But because the products are highly customized, Tomin Metaal can be characterized as a Repeat Business Customizer (RBC). Within a RBC there is competition and customization only for the initial customer purchase decision. After the first order and agreement to the supply contract, the customer will place repeated orders. The supplier company may choose to start producing after it receives the order or begin to produce to stock because it can be confident of selling the products in the near future (Amaro et al., 1999).

4.3 Process and product dimensions

The character of the customers and the high customization degree both influence the characteristics of the production process. Slack et al. (2007) sorted the production characteristics by four dimensions. These four dimensions are: product variety, product volume, complexity of process tasks and the process flow. On base of these four dimensions, production processes can be divided into five different manufacturing processes. Figure 5 illustrates the five different manufacturing processes on the basis of the four dimensions.

Figure 5 The five different manufacturing processes (Slack et al. 2007)

4.3.1 Product volume

Over the year 2010 Tomin Metaal produced 1.017.570 products. The average batch was 815. The product volume is not very small as in project and jobbing processes and also not very large as in mass and continues production processes. The product volume at Tomin Metaal is medium as is common in batch production processes.

Fast moving product family

(15)

15 efficient can be described as the focused factory3 concept. The fast moving product family has a volume of 170.641 products and an average batch size of 4260.

4.3.2 Product variety

In the year 2010 Tomin Metaal did serve 78 different customers which each have their own product assortment. All together Tomin Metaal did produce 562 different products. The customers operate in different industries and their products do have very different applications. Therefore very little modularity exists between the products. The only similarity is that all products are made of metal materials. The product variety can thus be characterized as high.

The fast moving product family is an exception to the general product variety. The family contains five different products. The five products within this family have many similarities. They have the same appearance but differ only in dimensions, undergo exactly the same operations and have the same process times.

4.3.3 Process tasks

Most of the process tasks at Tomin Metaal are simple. The tasks are designed in such way that they could be operated with the disabilities of the workforce. Each batch contains a different products, this makes the process tasks more diverse. The average process task complexity can be described as moderately diverse and complex.

The products within the fast moving product family differ only in dimensions. Therefore the process tasks of the fast moving production “line” are more repetitive than the regular production part.

4.3.4 Process flow

Because Tomin Metaal produces a high variety of products many different routings exist. The products have to be moved from an operation to the next operation and have to wait for other orders. This makes the process flow can be described as intermittent.

All products within the fast moving product family undergo the same operations. The throughputs of all operations in the focused factory are balanced. The balanced throughput of operations leads to a production with very little interruptions. The process flow of the fast moving product family is therefore more repetitive.

4.3.5 Summary

The production of Tomin Metaal can be divided in a regular production part and part for the production of fast moving products. The regular production produces a high variety of

3

(16)

16 products in medium sized volumes. The process tasks are simple and divided over the different operations. The process flow is intermittent because products are moved and have to wait between operations. The fast moving products are produced in larger volumes and with little variation. The process tasks are more standardized than in the regular production part. The tasks are simple and divided as in the regular part. The tasks are more repeated because the tasks do not change with the production of another product. The process flow is repetitive because the throughput of the operations is balanced. The production of the fast moving product family can be characterized as a batch process. The regular part of the production should be put somewhere in between a job and a batch process. The intermittent process flow and the high product variety belong to a job process but the large batch sizes and the moderately diverse and complex tasks belong to a batch process.

4.4 Shop floor configuration

To produce this high variety of products, the production layout of Tomin Metaal is divided into functional work areas (See appendix C). The different processes are clustered into several departments; such a layout is called a process layout. The advantage of a process layout is that it can readily process any product that requires work in any of the departments, regardless of the product’s production volume or routing (Nicholas, 1998). In a process layout a job travels through the work areas according to its routing sequence of operations. The production floor of Tomin Metaal has about eight main operations, these operations are:

1. Sawing (of metal bars and profiles) 2. Cutting (of metal plates in parts) 3. Punching (part out of metal plates) 4. Folding (of metal parts)

5. Drilling (holes in metal bars and profiles)

6. Metalworking (pressing and die cutting of metal plates) 7. Welding (MIG-, TIG-, spot welding and solder)

8. Assembly (Mechanical assembly, Packaging/sending and remaining activities)

Oosterman et al (2000) distinguish four kinds of shop configurations on the basis of the direction of material flow. The four configurations are a Pure Flow Shop, the General Flow Shop, the Restricted Job Shop and the Pure Job Shop. In a Pure Flow Shop, work travels in one direction through a sequence of operations in a strict order, unlikely in a MTO company. In a General Flow Shop work still travels in one direction but jobs are allowed to visit a subset of operations, permitting limited customization, relevant to a RBC. Job Shop routing sequences are random; jobs can start and finish at any operations allowing complete freedom and customization. The Restricted Shop is defined as providing variable routing sequences and constant routing lengths

(17)

17 Sawing Welding Folding Cutting Punching Drilling Assembly Metalworking Warehouse

Figure 6: Routing diagram of Tomin Metaal

4.4.1 Converging routings

The operations in the production process are performed in two kinds of machinery namely: the bars and section machinery and the metal plate machinery. The bars and section part of the production contains the sawing and drilling operations. And the metal plate part of the production contains the cutting, punching, pressing and folding operations. These two sections can produce complete products or finished products. Two or more semi-finished products are welded and/or assembled to one semi-finished product. The different routings of two or more sub-parts that come together at one point (joining process) are called converging routings.

4.4.2 Bottlenecks

The busiest part of the production process is the plate steel machinery. Most orders are waiting in front of the punching and folding operations. The folding and punching operations are also the only operations of the production process which are subcontracted at other metal processing companies because of lack of capacity. So the bottlenecks of the metal plate routing are the punching and folding operation. The two production steps have a strong relation because 834 percent of all orders that undergo the punching operations do have the folding operation as next production step. The bottleneck at the bars and section routing is at the sawing operation, the first production step in this routing.

4.5 Summary

The production facility of Tomin Metaal has a production layout and with a General Flow Shop configuration. The production facility contains a focused factory. These characteristic of the facility makes it possible to produces a high customized products for more 70 customers. The high customization leads to a high product variety. The process tasks are simple. The bottlenecks of the production process are at the beginning of the routings. Tomin Metaal has one product family which is order more repeatedly and in larger amounts

4_{Measured by logging the order overview (appendix A) over three weeks. A total of 152 orders were taken into}

(18)

18 than the other products. To produce the products of this family more efficient, Tomin Metaal set up production line with exist of unchained operations.

(19)

19

5 Production planning concept selection

The selection of a production planning concept depends on the goals and the production characteristics of the company. The production planning goals for Tomin Metaal are high delivery reliability and short lead times. The characteristics of Tomin Metaal are investigated and mapped in the previous chapter. Tomin Metaal can be described as a company which produces a high variety of products in a job shop with general flow shop configuration and operates in a repeated business customizer market. In this chapter a PPC concept that suits the characteristics and the goals of Tomin Metaal will be selected.

Stevenson et al. (2005) investigated the applicability of existing planning approaches for companies with different customization degrees and shop floor configurations. They also investigated the applicability of PPC approaches for companies comparable with Tomin Metaal, namely companies with a general flow shop configuration, operating in a repeated business customizer market. According to Stevenson et al. (2005), this kind of companies has specific operational issues and PPC systems should cope with these issues.

RBC companies offer a high variety of customer specific products. An appropriate PPC concept has to cope with this high level of customization, which entails in many products, variable routings and numerous set ups. At RBC companies, customers place orders and request due dates for these orders. A requested due date should be checked whether it is realisable, otherwise a due date should be proposed. Thus the ability of a PPC system to assist in the calculation of competitively short, realistic delivery dates is important. To meet the callculated due dates, the release and dispatch of orders should be focused on due date adherence. In summary, Stevenson et al. (2005) propose the following criteria as the main requirements of a PPC system in the MTO sector:

1. Inclusion of the Entry stage for delivery date determinations.

2. Inclusion of the Release and Priority dispatching stages, focusing on due date adherence.

3. Ability to cope with non-repeat production, i.e. highly customized products. 4. Ability to provide planning and control when shop floor routings are variable.

According Stevenson et al. (2005) there are five applicable production planning concepts for RBC’s with a general shop floor configuration namely; Continuous Work in Progress (CONWIP), Material Requirements Planning (MRP), Paired-cell Overlapping Loops of Cards with Authorization (POLCA), Theory of Constraints (TOC) and Workload Control (WLC). The selection of an applicable planning concept will thus be a choice between these concepts. The approach and the applicability of these planning concepts will be briefly described. The applicability of the concepts mostly relies on the research of Stevenson et al. (2005).

5.1 Theory of Constraints

(20)

20 The bottleneck-oriented is approach designed to control the one and only bottleneck in a production routing. As described in chapter 4 the production process of Tomin Metaal contains routings whit multiply bottlenecks resources. This makes the TOC concept less appropriate for Tomin Metaal.

5.2 Constant Work-In-Process

CONWIP is a system that maintains a constant maximum amount of work in process (WIP). This is usually done by using cards that are attached to an order at the beginning of the production process. When an order is processed at the last workstation, the card is released and sent back to the beginning of the system, where it is attached to the next order to be processed. No order can enter the system without its corresponding card (Dogger et al., 2010)

According to Spearman et al. (2001) the basic CONWIP approach will only work well in production processes with constant routings, similar process times, no significant setups and no assembly orders. This basic approach can be extended with several additions which make CONWIP also work well in processes with setup times, assembly orders and more variety in routing and process times. However, under CONWIP some standardization of products is needed because if the number of cards is to regulate the level of WIP, the workload represented by each card will have to be similar. This makes CONWIP less applicable for companies producing a high variety of products. Besides that CONWIP cannot provide the necessary control at the entry and release levels.

5.3 Material Requirements Planning

MRP assumes the production of standard products with well-known Bill of Materials and product routings. MRP also assumes that it is possible to forecast the future demand of these standard products as a basis for the Master Production Schedule. As a result MRP does not always generate feasible plans, and as Kanet (1988) explains, can lead to high WIP and long cycle times.

Many companies make use of Enterprise Resource Planning (ERP) systems. ERP systems are comprised of a suite of software modules with each module typically responsible for a separate business function, or a group of separate business functions (Muscatello et al., 2003). The planning function of ERP systems is based on the MRP approach. When companies already use an ERP system for other business functions the choice to use it also for the production planning is very attractive. But according to Stevenson et al. (2005) the widely availability of MRP does not mean it has wide applicability. The MRP approach has several issues. ERP vendors have tended to adopt a universal approach, rather than gearing the associated software to any particular industrial sector. This makes the implementation of tailored MRP systems complicated and costly. Another disadvantage of the MRP approach is the lack of capacity planning whilst determining due date quotations at the entry level.

5.4 Paired cell overlapping loops of cards with authorization

(21)

21 According to Stevenson et al. (2005) POLCA is a very appropriate concept for scheduling and balancing the production on the shop floor. But its lack of entry and release planning levels makes the POLCA approach less applicable.

5.5 Workload Control

WLC is a sophisticated PPC solution specially designed for the needs of the MTO Industry. WLC uses a pre-shop pool of orders to reduce shop ﬂoor congestion, making the shop ﬂoor more manageable, consisting of a series of short queues (Stevenson et al., 2005).

WLC is the most appropriate PPC approach for a company that produces a high variety of products in a job shop with restricted flow shop configuration and operates in a RBC market. The WLC addresses all PPC levels required by a RBC and can be an effective method of controlling lead times and adherence due dates, accommodating non-repeat production and variable routings.

5.6 Concept selection

With the selection of a production planning concept, the findings of Stevenson et al. (2005) are followed and the WLC concept is chosen. The WLC is suits the best with the production characteristics of Tomin Metaal and it also suits the performance goals of Tomin Metaal because in PPC literature it has been proven that the WLC is be able to reduce lead times and improve the due date adherence of orders (Land, 2004; Oosterman et al. 2000).

The other goal of the redesign is to reduce employee dependency. Employee dependency occurs because the planning decisions are made on the basis of tacit knowledge. The PPC literature describes a structured decision framework of the WLC concept. A successful implementation of this framework could convert the tacit knowledge in explicit knowledge. This conversion will reduce dependency of the planning employees.

(22)

22

6 Production planning concept description

The WLC concept is selected to form the framework for the production planning redesign. In this chapter the general WLC approach will be described.

WLC is defined by Henrich (2005) as a comprehensive shop floor control concept especially developed for job shop production. The load oriented production control policy intents to establish short and predictable lead times in order to improve delivery reliability. According to Slomp, Bokhorst and Germs (2009), WLC can be seen as the most sophisticated pull system, since workload norms for each workstation on the shop floor can keep the workload at each workstation constant. This capability is defined by Land and Gaalman (1996) as the balancing capability of a pull system. Figure 5 shows which decisions have to be made at which stage of the process.

6.1.1.1 Entry level

During the entry level two input decisions and one output decision have to be taken. The input decisions are: order acceptance and the due date assignment. The output decisions relate to medium term capacity adjustments.

Order acceptance

The acceptance of an order is generally not simply a matter of saying yes or no, but it will be preceded by an order quotation process. Within the concept of WLC, order acceptance provides the first opportunity to control the input of work in a job shop (Land, 2004)

Due date assignment

When the customer does request a specific due date it should be checked whether it is reliable, otherwise a due date should be proposed. The due date promise should preferable

(23)

23 be calculated on the basis of the material delivery times, pool times and operation throughput times.

6.1.1.2 Release level

The decision to release an order is relies on two aspects, namely the urgency of the order itself and its influence on the shop floor situation. The influence on the shop floor situation is determined by comparing workloads with norms. Workload norms can be defined for each capacity group and are usually expressed in time units. They should guarantee a small but stable buffer of work in front of the resources within the capacity groups. A stable buffer allows for constant operation lead times. In turn these constant lead times are used for determining accurate planned release dates. (Henrich, 2005). The decisions made in this level include ‘which orders’ should be released to the shop floor and ‘when’ (Ebadian, 2009).

6.1.1.3 Priority dispatching level

The dispatching decision is concerned with the choice which order should be processed next at a workstation after the operation of another order has been completed. The stations use a priority rule to sequence the orders in their queues (Land, 2004). The only input decision made at this level relates to sequencing the orders in the queue of each workstation to reach the corresponding promised delivery dates (Ebadian, 2009)

6.1.1.4 Capacity management

(24)

24

7 Analysis and improvement of the current production

planning

In this chapter the current production planning5 at Tomin Metaal is analysed. The shortcomings of the production planning will be exposed. This research is focusing on the reduction of the employee dependency, the reduction of the lead times and the improvement of delivery performance. Therefore the analysis of shortcomings is focused on the same aspects. To solve the shortcomings, practical solutions will be proposed. These solutions will be based on the WLC literature. This chapter follows the structure of the WLC concept. The first three paragraphs discuss the input control levels of the WLC concept then the output control decisions are treated. Afterward some complicating issues regarding the use of the WLC concept are treated. These issues are specific for the characteristics of Tomin Metaal. An explanation is given how could be dealt with these issues. For each planning decision that has to be taken, the following questions are answered.

1. How is this decision made in the current situation?

2. Which shortcomings does the current decision making process have? 3. How should this decision be made according to the WLC concept?

4. How should the WLC concept be applied in the situation of Tomin Metaal?

7.1 Entry level

As described in 6.1.1 two main input decisions have to be made at the entry level: the order acceptance and the assignment of due dates. During this level orders can be accepted or rejected. Orders can be rejected for several reasons. Customers and Tomin Metaal could be in disagreement about the price or delivery date.

7.1.1 Order acceptance

The purpose of this step is to classify the arriving orders based on the customers’ characteristics in order to make the allocation of available capacity to the arriving orders easier. Only orders that can be produced profitably should be accepted. In other words, the categorization of arriving orders to reach short and reliable delivery dates is crucial so that the production system will be able to plan the available capacity more efficiently (Ebadian, 2009). Tomin Metaal is a RBC company. Within a RBC company a new product is accepted only at the initial agreement. After the agreement, the customer will place repeated orders of the product.

7.1.1.1 How is this decision made in the current situation?

In the current situation a classification of orders is not formalized. Identifying desirable and undesirable products is done by the experience of the employees at the production office. The employees do not use an overview or method to prioritize customers. Employees estimate the importance of customers now on the basis of their knowledge about the order frequency, the amount of products ordered and the amount of sales of the customer.

5

(25)

25

7.1.1.2 Which shortcomings does the current decision making process have?

In the current situation the sales man en the planning employees check whether or not new products can be produced with the resources of the production facility. Less attention is paid on how these products influence the workload on the production floor.

The salesman is searching for customers with products which could be produced at the low utilized operations. Generating more workload for low utilized operation is a good intention to increase output with the same resources. Unfortunately, it is not taken into account how much workload the same product will generate for the scarce resources. When the same product also generates a lot of workload for the bottleneck resources the order is less attractive.

7.1.1.3 How should this decision be made according to the WLC concept?

In PPC literature is described how MTO companies can prioritize their customers in a more structured way. Huiskonen et al. (2003) classifies MTO customers on the basis of their margin contribution. A high margin contribution will make the customer more attractive.

7.1.1.4 How should the WLC concept be applied in the situation of Tomin Metaal?

Tomin Metaal can use the margin contribution to prioritize their customers but two other factors can be added. These factors are specific for Tomin Metaal. These factors are the amount of Wsw labour hours6 spend for the production of the order and the amount of processing time the order has to spend on the bottlenecks of the production process.

It is obvious that, in case two orders have the same production requirements; the one with the largest margin contribution is the most desirable. To get a straight attractiveness index, the margin contribution should be divided by the number of work shop hours. But because the bottlenecks are the most scarce and therefore most expensive resources, the hours an order spends on a bottleneck, should also be taken into account. The fewer hours an order has to spend on a bottleneck, the more attractive the order is. Therefore one factor which determines the attractiveness of an order is the margin contribution per bottleneck hour. Another factor which should be used in the attractiveness analysis is the amount of Wsw hours. This factor is specifically applicable for Tomin Metaal because one of the main goals of the company is to maximize the number of Wsw jobs. The more Wsw hours are required for the production of orders the more Wsw jobs can be created. The Wsw hours should also be divided by the total work shop hours of the order to get a straight attractiveness index. To determine the attractiveness of orders the following priorities are made up:

1. Very attractive orders: This group consists of the most important orders. This group exists of orders with large margin contribution per bottleneck hour and a large amount of Wsw labour hours.

6

(26)

26 2. Moderately attractive orders: This group exists of orders which also require a large amount of Wsw labour hours as a fraction of the total shop floor hours. The fact that these orders contribute relatively little margins per bottleneck hour, makes them moderately attractive.

3. Less attractive orders are orders which require little Wsw labour hours. These orders do not fit in the portfolio of a sheltered workshop.

7.1.2 Estimating delivery dates

At the moment the employees at the production office are having contact with the customers and communicate the assigned due date. The production leader is the one who takes the final due date assignment decision. The production Leader assigns the due dates on the basis of his tacit knowledge of the production process (capacity) and his tacit knowledge about the current workload on the production floor. The order overview (see appendix A) is his tool to estimate the delivery dates. The information the order overview displays is the number of orders, the order sizes, the routing of the order and the progress of the order. The order overview is updated once a week.

The production leader is the only person who can estimate the lead times of new orders. No other person can assign the due date because the information needed for this decision is not administrated. Here the dependency risk becomes visible.

Information about the capacity and the current workload has to be explicitly available and therefore has to be administrated. It is also seems hard to estimate the lead times of orders. A more structured procedure to estimate due dates would probably improve the accuracy of the due date estimation. A more accurate due date would reduce the tardiness and lead times of orders.

Cheng and Gupta (1989) describe different endogenous due date assignment methodologies. In the case of endogenous assignment methods, the due dates are set internally by the scheduler as each order arrives on the basis of job characteristics, job shop status information and an estimate of the throughput time (Cheng and Gupta, 1989). To calculate a due date, the throughput time has to be estimated. The WLC concept creates steady throughput times. Controlled release must keep the queues of jobs on the floor, as indicated by direct loads small and steady. Control of these direct loads creates predictable operation throughput times which allows for a good timing of order release and facilitates delivery date promising at the entry level. The entry level guards against excessive fluctuations of the total of accepted work (Land 2004). Stevenson et al. (2006) describe a formula to calculate the due date of an order. The formula sums the material lead time, the total work content, a pool delay and a queuing time constant.

(27)

27 needed for this estimation is based on his tacit knowledge. The information to estimate should be made explicit. The same components as in the formula of Stevenson et al. (2006) could be used to calculate due dates at Tomin Metaal. The due dates of orders could be estimated by sum all the lead time components of an order. These components are: Material delivery time, the pool time, the queuing times and the processing times. Because in the WLC concepts workloads are kept constant the operation queuing times will not exceed a certain norm. This norm will bound the maximum operation throughput time of an order. The maximum possible throughput could be used as a constant to make the due date calculation more simple. Because it is posible that pool times and material delivery times would overlap each other, the maximal combination of the pool time and the delivery time should be calculated. The due date calculation formula for Tomin Metaal will be:

Due date = order arrival date + Max(material delivery time, pool time) + operation throughput times of all operations in the routing + slack.

Order arrival date is the date a customer places an order at Tomin Metaal.

The material delivery time is the time it takes for raw materials and parts to be ordered and received so that production can begin.

Pool time can be obtained by sum all the process times of the gateway operation with an earlier due date in the pool. At the production floor of Tomin Metaal there exist three order pools exists, one pool for each gateway station.

Operation throughput times: Constant queues lengths in front of each operation should lead to constant operation throughput times. The throughput times of all stations in the routing have to be summed, to get the total shop floor time.

Slack time: a slack time should be added to get a safety margin for the delivery of orders and to compensate for disturbances.

7.2 Release level

Which job has to be released to the production floor should be decided at the release level. In the WLC concept orders are only released when they would not cause the workload norms to be exceeded.

7.2.1 How the decision is be made in the current situation?

At Tomin Metaal there are three operations where an order can start its routing namely: Punching, sawing and cutting. In the current situation, the decision to release an order is based on the shortest time to due date and the availability of the capacity on the specific operation.

Stock orders

(28)

28

7.2.2 Which shortcomings does the current decision making process have?

In the current situation orders are released when capacity is free at one of the gateway operations. The release decision is based on the due date of an order. In cases a choice has to be made to release one order or another, the orders with the tightest due date is released first. Orders with short lead times and a tight due dates are released earlier than orders with long lead times and relative loose due dates. These situations reduce the delivery performance of orders with long lead times. Orders with longer lead times should be released earlier to be delivered on time.

Stock orders

The current release decision treats stock orders the same as customer orders. An order is released on base of a tightest due date rule. This procedure could lead to situations where stock orders get priority to customer orders. This will have negative influence on the delivery performance of customer orders.

7.2.3 How should this decision be made according to the WLC concept?

At the release decision, the orders in the pool are selected to enter the production floor. The WLC concept uses operation due dates derived from the planned station throughput times: the planned operation throughput times for all operations in the routing of an order are subtracted from its due date. The orders in the pool are sequenced in order of their operation due dates of the gateway operation. In this sequence the jobs are considered for release. An order will be released, if its release does not cause the workload of any station to exceed the workload norm, otherwise the next order in the pool is considered (Land, 2004).

Stock orders

In the PPC literature is a solution proposed to produce stock orders and customer orders on the same production floor. This solution is called Capacity Oriented Materials Management (Corma). Corma is an operations management principle that enables mixed manufacturers to balance work in process against limited capacity and short deliveries. (Schönsleben, 1995). Corma releases stock replenishment orders earlier than needed, that is before inventory levels hit the order point. An early order release is considered as soon as there is available capacity in work centers. Replenishments orders of make-to-stock products are treated as filler loadings. They fill in capacity not required by other orders. However, this may mean that make-to-stock products are produced earlier than required. Orders that are released early are scheduled without priority. They are performed when capacity is available on the required operations and there are no more urgent orders to process (Schönsleben, 1995). If inventory stocks fall faster than expected the latest due date of some of the stock replenishment orders may be advanced. Alternatively, if stocks fall more slowly than expected the latest due date will be postponed (Schönsleben, 1995).

7.2.4 How should the WLC concept be applied in the situation of Tomin Metaal?

(29)

29

Release to the Sawing operation

The sawing operation is the gateway of the bars and section routing. The orders which undergo this routing have to be released at the sawing operation. Because the sawing operation is the bottleneck of the bars and section routing, the release of orders should only take the workload of the sawing operations into account. Orders should be released in order of their operation due date and should only be released, if they do not exceed the workload of the sawing operation.

Release to the punching operation

The punching operation is an of the two gateway operations of the plate steel section. The punching operation is high utilized, which causes a naturally pool in front of the operation. The operation is strongly related to the folding operation. The next operation after the punching operation is in 83 percent the folding operation. The other 17 percent of orders are completely finished after the punching operation. The folding operation is also a highly utilized operation. The operations downstream the folding machines are less utilized. This makes the change that high amounts of WIP will arise in front of these downstream operations relatively small.

The release decision at the punching has therefore only to the balance the workload of the punching and the folding operations. The orders in the pool are sequenced in order of their operation due date. In this sequence the orders should be considered for release. An order should only be released, if it does not exceed the workload of both operations, otherwise the next order in the pool should be considered.

Release to the cutting operation

After the cutting operation, the next operation that orders undergo could be the folding or the metalworking operation. The bottlenecks in this routing are the cutting and the folding operation. The release decision at the cutting operation should balance the workloads of the folding operation. An order should only be released at the cutting operation, if its release does not cause an exceeding workload of the both operations, otherwise the next order in the pool should be considered.

Release decision in general

In the case of Tomin Metaal the bottleneck operations are the gateway operations, so the order release can be almost completely controlled by the direct load of the bottlenecks. The only bottleneck, which is not a gateway operation, is the folding operation. The folding operation has no pool of orders because there are no orders which begin their routing at the folding machines. All the orders that arrive at the folding operation come from two other operations, the cutting operation and the punching operation. The workload of the folding operation should also be balanced by the release decision. Orders which will not exceed the workload for the folding operation should be released.

(30)

30 outsourced. The outsourcing of orders after the release decision is an emergency act. Normally the decision to outsource should be taken during the entry level.

Stock orders

Stock orders should not get an assigned due date but they should be used as filler orders to maximize capacity utilization of the gateway operations. The orders should get lower priority if the stock level is not below reorder point. Filler orders are produced when there are no customer orders in the pool for a specific gateway operation. Or in cases the release of customer orders is not possible because of certain reasons, for example when required raw materials are not yet arrived.

When inventory stocks fall below the reorder point, stock orders should not be treated like a filler orders anymore. Then the production of the stock orders should get a normal priority. When the inventory stock comes close to the safety stock level, the stock orders will get high priority.

7.3 Priority dispatching

The dispatching decision is concerned with the choice which order should be processed next at a workstation after the operation of another order has been completed (Land, 2004). The influence of priority dispatching is small according the WLC literature (Kingsman and Hendry, 2002). Because at Tomin Metaal the bottlenecks are at the front of the routings, the orders will spend little time in waiting queues after they are released. The influence of priority dispatching will therefore be extra small in case of Tomin Metaal. However, some dispatching rules exists that still improve the average lateness and some that reduce the variance of due dates (Land 2004).

7.3.1 How is this decision made in the current situation?

Tomin Metaal dispatches orders on base of tightest due date rule. Orders with relative tight due dates will get priority to orders with long lead times but relative loose due dates.

The current priority dispatching rule reduces the delivery performance of orders with long lead times. Orders with relatively long lead times should get higher priority to be delivered on time.

Orders should be dispatched on base of earliest operation due date. The operation due date is defined as the order due date minus the planned throughput times for the downstream operations (Land 2004). With this priority dispatching rule, orders with relative long lead times will be released and finished earlier. This will improve the delivery performance.

7.4 Capacity management

(31)

31

7.4.1 How is this decision made in the current situation?

Tomin Metaal has limited possibilities to adjust the capacities the production process. Most employees of Tomin Metaal work on base of the sheltered employment law and are therefore not allowed to work over time or in shifts. Outsourcing is good option because the products of Tomin Metaal are simple and undergo metal processing operations which are very common in the Metal industry. Tomin Metaal uses two ways to adjust production capacity. These two capacity adjustments options are:

1. Employees can be reallocated from an under loaded to an overloaded resource 2. The production of orders can be outsourced when capacity of resources is

overloaded.

In the current situation the current situation some orders are outsourced when they are already released on the production floor. In these situations raw materials are already ordered and have to be moved to the subcontractor.

7.4.3 How should this decision be made according to the WLC concept?

According to Stevenson et al. (2006) any outsourcing should be planned as early as possible, usually at the entry level, before materials have been ordered.

Realistic and accurate due date assignment will reduce the need for outsourcing. But in cases customers ask for tight delivery dates which cannot be achieved by Tomin Metaal, the production of orders should be outsourced. Tomin Metaal should plan the outsourcing of orders at the entry level.

7.5 Complicating planning issues

Besides the input and output control decisions there are some issues which require specific attention in the production planning of Tomin Metaal. These issues are the production of assembly products, the planning of rush orders and the production of the fast moving product family.

7.5.1 Rush orders

The productions of Tomin Metaal gets frequently rush orders of important customers. A decision that has to be taken is how to deal with these rush orders.

In the current situation the rush orders at Tomin Metaal have high priority on the production floor. When a rush order arrive at the production floor, the production of other orders is aborted. The machines are quickly adjusted for the production of the rush order.

The current approach to deal with rush orders leads to extra setup times and WIP.

7.5.1.3 How should the decision be taken according to the PPC literature?

(32)

32 also required an increase in the planned throughput times for other orders. (Hendry et al. 2008).

Tomin Metaal could directly apply the solution of Hendry et al. (2008). At Tomin Metaal a part of the capacity is not should not be used in the production planning. The planning employees should determine the size of the part that has to be reserved for rush orders. The due date assignment must take the presence of rush orders into account and base the due dates of orders on the capacity reserved for non rush-orders.

7.5.2 Assembly products

Some orders at Tomin Metaal exist of two or more subparts which are assembled or welded together to one end product. The problem with sub part is that they could have different lead times but have to arrive at the same time at the converging operation. The production planning of subparts needs therefore extra attention.

Tomin Metaal has a couple of assembly orders with two different routings which are converging together. These orders are released at the same time. If the subparts have different shop floor times, the foremen let the subparts arrive synchronously by dispatching the orders in upstream operations.

At Tomin Metaal there is one product family that considers the majority part of the assembly orders. The products in this family exist of three subparts. All the three subparts have to be punched at first and have to be folded later. The three folded parts have to be assembled together. The subparts have the same production routing. The production planning has chosen to punch the subparts together out of one plate of steel and then all the three punched parts are folded synchrony on three folding machines. The subparts arrive at almost at the same time at the assembly operation.

The current approach to deal with subparts works quite well. The subpart at the production of Tomin Metaal arrive pretty synchrony at the assembly operation. This way of producing minimizes the throughput time of the orders. The problem of this method is that the decision to punch three different subparts out of one plate of steel gives a little waste of materials. Punching one kind of subpart out of a plate gives a more efficient nesting. In this case more subparts can be punched out of one plate of steel.

(33)

33

At Tomin Metaal orders are already released synchrony. However this method leads to extra loss of raw materials, the recommendation is to continue this approach because all things considered, the advantage of short lead times outweigh the loss of raw materials.

7.5.3 Production of the fast moving product family

(34)

34

8 Redesign of the Production Planning Decisions

In previous chapter all the planning decisions are described and analyzed following the structure of the WLC concept. For each planning decision as described in chapter 8, an overview will be given. The overview tells how the employees should take decisions based on explicit information. The overviews are presented in excel spreadsheets. A restriction of the redesign is that the new planning process has to work with the same tools and systems as in the current situation. In the current situation the decisions are taken with the aid of an ERP system and a excel spreadsheet (the order overview, Appendix A). The redesigned planning process does also make use of this equipment. The description of the redesigned production planning follows the same structure as the WLC framework.

8.1 Order Entry

8.1.1 Order acceptance

When new products have to be accepted, the attractiveness should be investigated. To determine the attractiveness of orders the following factors have to be taken into account:

1. The margin contribution per bottleneck hour

2. The ratio of Wsw labour hours to total shop floor hours

On hand of these factors a new product can be placed in the attractiveness matrix of figure 7.

Figure 8: The product attractiveness matrix

Less attractive products should only be accepted when they are ordered by a customer with an attractive order portfolio.

8.1.2 Due date assignment

Production Planning Redesign