Improving the service level of Akzo Nobel’s
Advanced Tinting Plant
Satisfying customers by improving due date conformance
1
Improving the service level of Akzo Nobel’s
Advanced Tinting Plant
Satisfying customers by improving due date conformance
Thesis on behalf of graduating
MSc Technology Management
January 2008
Author: Albert-Jannes Oosting
Student number: 1357549
Date: January 2008
Study: MSc Technology Management
Faculty of Economics and Business University of Groningen
Supervisors: Dr. ir. D.J. van der Zee Drs. ing. G. Nanninga Organization of study: Akzo Nobel Groot-Ammers
Supervision: Klaas Dijkstra
2
PREFACE
The following thesis is the result of research undertaken at Akzo Nobel’s Advanced Tinting Plant located in Groot-Ammers. It has been written in the light of graduating from the MSc Technology Management at the University of Groningen. The thesis is on improving the Advanced Tinting Plant’s service level which can be interpreted as improving due date conformance. In order for this to be achieved, the focus of this study has been on increasing the reliability of lead times in part of the transformation process called production.
In parallel with doing research and writing my thesis, I was also given the opportunity of being a member of a project team. Working with this team and contributing to the successful implementation of a number of improvements in production has certainly given me more insight into the workings of production systems in practice. It is my belief that this experience will not only help me in my future working career but has also contributed to my personal growth.
First of all, I would like to thank everyone working at Akzo Nobel Groot-Ammers, especially employees of the Advanced Tinting Plant, for making me feel welcome and for their willingness to answer all my questions. A lot of gratitude goes out to Klaas Dijkstra for granting me this assignment, his help with writing this thesis and for giving me the opportunity to attend courses on 5S and Total Productive Maintenance (TPM). I would also like to give a special thanks to John Cahill, for his coaching and for sharing his experience and knowledge on continuous improvement with me. Also, I would like to thank Peter Rietveld for his support during the course of this study.
Naturally, I would also like to thank my academic supervisors, D.J. van der Zee and G. Nanninga for guiding me during this study, for providing me with constructive feedback and for their flexibility.
Last but not least, I would like to thank my family and friends for their support and for just being there.
Bert Jan Oosting
3
SUMMARY
In the beginning of the study presented in this report, the ATP was experiencing serious service level problems. The base service level is a percentage expressing the share of tins that were delivered on time in full (OTIF) by a DC to its customers, measured on a weekly basis. Based on a DC’s service level, the ATP’s service level in relation to that DC is determined. First the number of tins that could not be delivered OTIF by a DC due to reasons assigned to the ATP is determined. Hereafter, this number is expressed as a percentage of the total number of tins the ATP had to supply a DC with in a certain week. By subtracting this percentage from 100, the ATP’s service level in relation to a DC is obtained.
The service level problems in relation to the ATP’s central stock point, namely the Breda DC have been considered illustrative for all DCs. In the beginning of this study, it was observed and proven that the output of the production department, i.e. the filling buffer, did not comprise the orders it should have based on the plan made by the planning department and therefore demand. Not adhering to this plan leads to orders not being delivered on their planned due date whereby the probability of manco items to arise increased. Items that are referred to as manco are out of stock in a DC and logically affect the ATP’s service level negatively. This is the reasons it was decided to focus the research on a number of processes demarked as relevant to production; this demarcation enabled the bottom-up approach taken in this study in the limited time available. The objective determined for this study was:
Provide recommendations and facilitate the implementation of actions based on these recommendations, on adapting processes relevant to production in order to structurally increase the ATP’s service level for the Breda DC to a minimum of 99,2% by constantly having a filling buffer which mix of finished production orders is enabling planned and thus timely replenishment of stocks in the Breda DC.
Subsequently this objective has been converted into the following central research question:
How can the processes relevant to production be adapted in order to constantly have a filling buffer which mix of finished production orders is according to plan?
Factual descriptions of all processes demarked as relevant to production have been provided first in the context of all of the ATP’s related processes. There after it has been analyzed what exactly is causing actual lead times in production to be longer than planned for, since this should be regarded as the primary cause for obtaining a filling buffer which mix is not according to plan. Analyzing this has been enabled by defining lead time as presented below and dividing it into an order’s start time and an order’s in-process time.
4 In total, seven generic causes have been found which might lead to an order’s actual lead time to be longer than planned for, i.e. a positive deviation from the planned lead time. The causes were found during the execution of improvement projects and by interviewing ATP’s personnel; the same goes for the more specific issues underlying these generic causes. Thereafter, these generic causes were ranked based on their occurrence and related deviations from planned lead times. This provided a starting point for determining how underlying issues could be eliminated. In answering the central research question, the following recommendations have been made:
1) Improve monitoring and control
a. Act quicker on adjustments waiting to be executed b. Anticipate material shortages
c. Anticipate tank truck arrivals
d. Act quicker on the occupation of tanks which are constraining the flow of orders through production
e. Increase turnover rate mobile tanks
2) Further research the existence of the non-packaging buffer and take note of the possible solutions already provided
3) Prevent adjustments
a. Improve raw material deliveries and storage b. Calibrate flow meters more frequently c. Set up a quality circle
d. Improve communication between the ATP’s QC department and R&D located in Sassenheim
e. Determine the FROK differently 4) Improve stock accuracy raw materials
a. Make the GL responsible for all raw material related transactions in SAP 5) Improve the scheduling and dispatching processes
a. The GL should dispatch the orders instead of the operators themselves
b. Apply a scheduling heuristic based on maximum adherence to the plan made by the planning department
6) Recommendations for future research
a. Investigate how capacity planning should came to bear; its necessity has been proven in this report
b. Determine more suitable measures than volume for determining the ATP’s daily performance in relation to weekly and monthly targets set
c. Investigate how the consequences of VOC 2010 can best be anticipate and coped with
5 the probability of manco items to arise has decreased. In relation, it has been shown that a couple of weeks after the implementation of the PPC board which has improved monitoring and control in production, the service level began to improve as is indicated below. The weeks in which the PPC board was fully implemented are delineated.
Mancos Critical Service level Mancos Critical Service level Mancos Critical Service level
Wk 1 27 69 99,0 Wk 16 46 198 96,0 Wk 31 22 90 99,5 Wk 2 24 80 98,4 Wk 17 47 199 95,8 Wk 32 17 92 99,4 Wk 3 17 75 99,1 Wk 18 75 199 95,8 Wk 33 19 89 99,3 Wk 4 14 83 99,9 Wk 19 83 195 90,5 Wk 34 21 92 99,1 Wk 5 16 116 99,2 Wk 20 73 225 94,5 Wk 35 19 98 99,3 Wk 6 27 118 99,9 Wk 21 89 220 96,6 Wk 36 26 100 99,4 Wk 7 30 92 99,7 Wk 22 73 232 96,0 Wk 37 27 99 98,4 Wk 8 25 92 96,1 Wk 23 69 225 95,8 Wk 38 35 75 98,3 Wk 9 22 103 99,2 Wk 24 76 185 96,1 Wk 39 14 81 99,1 Wk 10 28 100 99,3 Wk 25 76 184 95,8 Wk 40 27 75 99,8 Wk 11 31 126 99,6 Wk 26 64 140 96,8 Wk 41 19 64 99,9 Wk 12 40 141 98,4 Wk 27 59 155 98,9 Wk 42 21 76 99,8 Wk 13 32 168 98,7 Wk 28 45 136 98,6 Wk 14 45 170 97,7 Wk 29 32 105 99,7 Wk 15 50 182 95,5 Wk 30 30 86 99,9
Week nr. Week nr. Week nr.
6
TABLE OF CONTENT
PREFACE... 2 SUMMARY ... 3 TABLE OF CONTENT ... 6 ABBREVIATIONS... 9 1. INTRODUCTION ... 10 1.1 Akzo Nobel... 10 1.2 BU decorative coatings... 101.3 Akzo Nobel Decorative Coatings Groot Ammers... 11
1.4 Research background... 13
1.4.1 Key performance indicators ... 14
1.4.2 Priorities and research topic ... 16
1.5 Summary... 17
2. RESEARCH DESIGN ... 18
2.1 Problem background... 18
2.1.1 Service level ... 18
2.1.2 Focussed approach ... 19
2.2 Central research question ... 22
2.2.1 Assumptions ... 22
2.2.2 Definitions ... 23
2.3 Approach ... 23
3. PLANNING AND TRANSFORMATION PROCESSES ... 24
3.1 Introduction ... 24
3.2 Master Production Schedule (MPS) ... 24
3.2.1 MPS parameters ... 25
3.2.2 Period Order Quantity (POQ)... 26
3.2.3 Material Requirements Planning (MRP) and orders ... 28
3.3 Planning production... 29
3.3.1 Creating orders ... 29
3.3.2 Releasing production orders ... 30
3.4 Scheduling production... 31
3.4.1 Scheduling rules ... 31
7
3.4.3 Adjusting the schedule ... 33
3.5 Physical transformation processes... 34
3.5.1 Products, processes and areas... 35
3.5.2 Delivering, storing and getting raw materials ... 38
3.6 Processes QC department ... 40
3.6.1 Volatile Organic Content (VOC) ... 41
3.6.2 Checks on quality ... 41
3.6.3 Adjustments and dissaprovals ... 42
3.6.4 First Run OK (FROK) ... 43
3.7 Filling and wrapping department... 43
3.8 Monitoring and control... 45
3.9 Summary... 46
4. PLANNED AND ACTUAL LEAD TIMES IN PRODUCTION... 47
4.1 Introduction ... 47
4.2 Overall situation ... 48
4.3 Lead time ... 50
4.4 Deviation from planned start time... 51
4.4.1 Calculation of deviations... 51
4.4.2 Causes for positive deviations ... 52
4.4.3 Inability to schedule production orders ... 52
4.4.4 Not scheduling orders on their planned start date ... 53
4.4.5 Timely scheduled orders requiring rescheduling ... 54
4.5 Deviation from planned in-process times... 57
4.5.1 Calculation of deviations... 57
4.5.2 Causes for positive deviations ... 58
4.5.3 Significance of the identified causes ... 59
4.5.4 Batches in need of an adjustment ... 61
4.5.5 Non-packaging buffer... 66
4.5.6 Disturbances in physical production ... 67
4.5.7 Waiting on the execution of final testing ... 69
4.6 Ranking of generic causes ... 69
4.7 Summary... 71
5. RECOMMANDATIONS ... 72
8
5.2 PPC board... 72
5.2.1 PPC board’s process element ... 76
5.2.2 Operational aspects of the PPC board ... 76
5.3 Not scheduling orders on their planned start date ... 77
5.3.1 Avenues for future research... 78
5.3.2 Adapting the processes of scheduling and dispatching ... 78
5.4 Batches in need of an adjustment ... 81
5.4.1 RM project... 82
5.4.2 Focus on quality and additional recommendations ... 86
5.4.3 FROK ... 88
5.4.4 Quality and future research ... 89
5.5 Timely scheduled orders requiring rescheduling ... 89
5.5.1 Eliminating shortages in mobile tanks ... 90
5.5.2 Shortages in raw materials and lower-sefis ... 92
6. EVALUATION ... 94
7. CONCLUSIONS ... 96
8. CRITICAL REVIEW ... 98
9
ABBREVIATIONS
ATD Advanced Tinting Plant
BU Business Unit
DC Distribution Centre
EET European Excellence Team ERP Enterprise Resource Planning FROK First Run OK
FTE Full Time Equivalent FMA Failing Materials
GL Group Leader (NL voorman) KPI Key Performance Indicator
LPC Logistics Planning and Control department
LT Lead time
LTI Lost Time Injury / accidents Lower-sefi NL Halffabrikaat
MPS Master Production Schedule MRP Material Requirement Planning OEE Overall Equipment Effectiveness OSS Out of stock
OTIF On Time In Full
PPC Production Planning and Control QC Quality Control (department)
RM Raw Material
SKU Stock Keeping Unit
Sefi Semi-finished product, i.e. batch of paint SOP Standard Operating Procedure
VOC Volatile Organic Content
10
1. INTRODUCTION
This report is written on account of a study on how to improve the service level of the Advanced Tinting Plant (ATP) located at the Akzo Nobel Decorative Coatings site in Groot- Ammers. In the following a brief introduction with respect to the company, business unit (BU) and plant will be given; also the relation with the purpose of the research presented in this report will be elucidated.
1.1 Akzo Nobel
Akzo Nobel is multinational organisation, employing around 62.000 people, based in the Netherlands. Nowadays, after the divestment of the pharmaceutical branch, it is serving customers and conducting activities in the coatings - and chemicals segment (ws Akzo Nobel).
Within the coatings segment, Akzo Nobel is global market leader in the production and sales of paint, as it has been developing and supplying innovative coatings products for more than 200 years. Also, this segment is employing 74% of total Akzo Nobel employees and is responsible for 62% of Akzo Nobel’s total revenues (Beyond paint, 2007). The coatings segment comprises five business units, distinguished according to market segments, e.g. marine coatings. One of these five is the globally operating BU for decorative coatings; this is also the largest contributor in terms of revenues for the total coatings segment. This BU has been defined by Akzo Nobel as:
A global leading supplier of architectural paints, lacquers, special wood care products, tools and equipment for both professionals and consumers, as well as a leading European supplier of Building Adhesives. (Beyond paint, 2007)
The BU decorative coatings comprises five regional and one market segment-orientated sub-business units (SBU). One regional SBU is Europe Central, which is the heading under which all the belonging country-based management teams fall, such as the one for the Netherlands. Also, falling directly under the BU is a functional management team. Since representatives of the Groot-Ammer’s site have to report to both management teams and this organisational structure is thus utilizing functional and divisional chains of command simultaneously, it can be depicted as a matrix structure (Daft, 2001). For a complete overview of the organisational structure of the coatings branch of Akzo Nobel, reference is made to appendix 1.
1.2 BU decorative coatings
11 The view expressed in the recently developed strategy for the BU decorative coatings (G.O.L.D. – Rising to the global challenge, 2007), is that this is foremost achieved by offering customer value. Since the goal for every country, site and plant within the BU Decorative Coatings, is to maximize profit, this might not completely align with the optimisation of the whole. Because of a certain level of competitiveness between the sites, obstacles might have to be overcome in transferring volume between them, leading to sub-optimisation. This notion is interesting to look into for the whole BU decorative coatings; however this lies outside the scope of the study presented in this report. Namely, the scope of this study lies within Europe and particularly within the Netherlands
Even though there is an element of competitiveness within the BU decorative coatings, there is an awareness that all plants are contributing to the same company’s profits. This statement is explicated by the existence of the European Excellence team. The members of this team are several sector leaders or plant managers and a number of improvement managers; they come from plants all over Europe. They meet twice a year, every time on a different site location, to share best practices and facilitate learning.
1.3 Akzo Nobel Decorative Coatings Groot Ammers
The site in Groot-Ammers comprises two plants, namely the Wall Paint/Multicolour (WP) plant and the ATP. In addition, a warehouse is present fulfilling the following two functions:
1. Raw materials storage and distribution to the production plants 2. Finished product storage and distribution to customers
To enable the ATP, the WP and the warehouse to function as intended, a number of support functions/departments are present. Examples are a financial -, human resource - and Logistics, Planning and Control (LPC) department.
ATP
The ATP is supplying its distribution centres with brand specific and multiple sized tins, ranging from 0,125 litres till 25 litres, filled with paint produced in the ATP. Also, vessels of 200 litres and containers of 1000 litres are filled occasionally. Just a few examples of brands the ATP is making paint products for are Sikkens ®, Flexa ®, Cetabever ®, Herbol ® and Trimetal ®.
12 ATP’s ability to produce paint in small batch sizes, thereby allowing for small orders to be filled and thus fulfilled.
The number of paint recipes (including all colours) converted into items is about 750; the majority of it is solvent born, while the rest is water born. The difference between 1600 items and 750 recipes can be explained by the fact that multiple items are based on the same recipe, due to branding and the filling of multiple sized tins. The forecasts over 2007 expressed in tons of required output add up to 8700 ton. As can also be derived from the decorative coatings BU’s definition, the end-customers for the ATP’s products are present within the professional (trade-items) and the Do-It-Yourself (retail-items) market.
The main departments in the ATP are the production- and filling department. In the production department paint can be made in several batch quantities enabled by the availability of multiple sized production tanks. Also, materials which cannot be ordered from external suppliers are made in this department, referred to in the remainder of this report as lower-sefis. The term sefi is derived from semi-finished product; a regular sefi is pertaining to a material (paint) that is filled into tins.
After a batch has been approved by the process technology /quality control department, in the remainder of this report referred to as QC department, it is used to fill tins in the filling department. It can be stated that the filling department is the core of the transformation process at the ATP, this is where everything, in addition to the paint itself, comes together to create finished products, i.e. tins filled with paint. ‘Everything’ in the former, can be divided into lids, stickers, packaging material and either supplied lithographic tins or tins labelled in the ATP. The packaging machines are connected to the filling machines by a conveyor belt and are therefore an integral part of the filling department. Stacking of packaged products onto a pallet is a manual process; every running packaging machine has an operator standing behind it. After a pallet has been stacked according to specification, it is transported to the wrapping department, where a stacked pallet with finish products is made ready for transportation to a DC.
13
Figure 1 ATP’s transformation process
As can partly be derived from this figure, a number of support functions/departments within the ATP are enabling the two main departments to run accordingly, i.e. planning, labelling tins, printing stickers, technical services, transportation and the afore mentioned QC department. During the season of high market demand, the plant is running in three shifts: a morning-, an evening- and an overlapping dayshift. In this situation, the ATP is employing around 57 people who equal 34 Full Time Equivalents (FTE1) of personnel with a fixed contract and 23 FTEs of third-party personnel. These figures are however subject to change because the plant is running in dayshift during the season of low market demand. In this situation, the workforce has been reduced by (temporarily) letting third party worker go.
1.4 Research background
The result of this study is expected to contribute to an improvement in the ATP’s performance; therefore it is important to explain what exactly constitutes performance. In the next sections, the key performance indicators (KPI) and the relation of these KPIs with the research topic are presented.
1
FTE: a measurement equal to one staff person working a defined full-time work schedule. In the ATP for
14 1.4.1 Key performance indicators
Senior management of Akzo Nobel has determined five KPIs that apply to all production sites and plants within the BU Decorative Coatings. These KPI’s are enabling performance measurement on important aspects of the business and allow for future goals to be set. Below, the five KPI’s, which fall under the heading of operational excellence, are explained and related to the ATP.
1) Service level
The base service level is a percentage expressing the share of tins that were delivered on time in full (OTIF) by a DC to its customers, measured on a weekly basis.
Based on a DC’s service level, the Groot Ammers site’s and more specifically the ATP’s service level in relation to that DC is determined. First the number of tins that could not be delivered OTIF by a DC due to reasons assigned to the ATP is determined. These reasons can be categorized into:
1) Untimely replenishment of items’ stock
2) Replenishments with an insufficient quantity of items 3) Replenishments with bad quality items which cannot be sold
Hereafter, the number derived is expressed as a percentage of the total number of tins the ATP had to supply a DC with in a certain week. By subtracting this percentage from 100, the ATP’s service level in relation to a DC is obtained. The way the ATP’s service level is derived from a DC’s base service is more extensively described in section 2.2.1.
Since the ATP is supplying two DCs in the Netherlands and 12 DCs located abroad with its tins, the ATP is subject to a total of 14 service level measurements. Herein the focus is on the service level in relation to the Breda DC. This focus is explained by the fact that this service level is the most important one because around 60% of the ATP’s output is transported to the Breda DC, making it its central stock point. Another reason is that the situation regarding this service level is indicative for the other 13. Therefore, in the remainder of this report, the term service level is pertaining to the ATP’ service level in relation to the Breda DC, unless indicated otherwise.
2) Costs
15 3) Quality
Quality is an overall heading under which a number of aspects are noticeable. The first is pertaining to the products (types of paint) itself and the narrow specifications it is subject to. This explains the existence of the QC department, which has to test and make sure that the paint filled into tins is of the quality the end-customer may expect and thus falls within the specifications set by R&D. Related is the (process) quality measurement called FROK (First Run OK); on a high level this percentage expresses the quality of the paint making process and all utilized instruments within it.
The third is the quality of the deliveries made to the DCs, measured by the number and type of incoming complaints. Examples are a shortage in actual delivered items compared to the amount stated on the transportation ticket and a wrong sticker on the boxes containing the paint tins. The fourth is the quality experienced by end-customers, measured by the number and type of incoming complaints. These complaints can for instance pertain to the quality of the paint itself.
4) HSE
Stands for health, safety and environment; it means that the products must be manufactured, handled, transported, used and disposed of safely without unacceptable risk for man and environment. This is expressed by a weekly check on housekeeping in the ATP, i.e. operators are expected to keep their working areas clean.
Also, regular measurements are taken on the level of solvents and dust in the air and action is undertaken if these prove to be below the standards set. Furthermore, main risks are fire (prevented by the obliged use of a grounding clamp), chemicals (Sire Sahi code), physical endurance (ergonomics) and internal transport. Personal protection means are widely available and operators are informed on the obligatory nature of using these safety glasses, masks, gloves, ear plugs and safety shoes. Naturally, related goals to the above are a reduction of illness rates and bringing the number of Lost Time Injuries (LTI) to zero.
5) Stocks
16 1.4.2 Priorities and research topic
Logically, within the context of offering customer value, service level is emphasized, i.e. making sure that a requested product of the right quality is always available to end-customers. This perspective has not always been leading, because the previous primary focus was on cost reduction (Breukelman, 2004; Rijckevorsel, 2006). At present, cost reduction is the second priority for the site in Groot Ammers and thus the ATP. More concrete, this implies the necessity for the ATP to always keep its output sufficient in terms of content and volume, thereby keeping the stock levels of all items at the desired levels in every DC. This is to be achieved however by keeping expenditures within the allocated budget, therefore costs cannot be ignored.
When an item’s stock in a DC is not of a sufficient level, the item is either referred to as ‘manco’ or ‘critical’. When an item has become manco, it means that the DC in question is out of stock (OOS) of that particular item and is therefore unable to supply the retailer or wholesaler, which in turn cannot sell the product to its customers. When an item is critical, it means that the level of stock in the DC in question for that particular item is below its safety-stock level; in other words, this item is on the verge of becoming manco and its safety-stock requires replenishment on short notice.
The number of manco and critical items at the end of a week in 2007 are presented together with the weekly achieved service level in table 1. A red indication on service level, means that the achieved service level is less than the minimum set target of 99,2%. The average service level measured from January up to July 2007 is 97,6% with a standard deviation of 2,15%. From both the number of red indications in table 1 and the average service level and its standard deviation, the need for improvement can be derived. The goal to be achieved and herein the contribution of this study is lying, is getting the service level up to a minimum of 99,2% structurally.
Manco's Critical Service level Manco's Critical Service level Manco's Critical Service level
Wk 1 27 69 99.0 Wk 13 32 168 98.7 Wk 25 76 184 95.8 Wk 2 24 80 98.4 Wk 14 45 170 97.7 Wk 26 64 140 96.8 Wk 3 17 75 99.1 Wk 15 50 182 95.5 Wk 27 59 155 98.9 Wk 4 14 83 99.9 Wk 16 46 198 96.0 Wk 28 45 136 98.6 Wk 5 16 116 99.2 Wk 17 47 199 95.8 Wk 29 32 105 99.7 Wk 6 27 118 99.9 Wk 18 75 199 95.8 Wk 30 30 86 99.9 Wk 7 30 92 99.7 Wk 19 83 195 90.5 Wk 31 22 90 99.5 Wk 8 25 92 96.1 Wk 20 73 225 94.5 Wk 32 17 92 99.4 Wk 9 22 103 99.2 Wk 21 89 220 96.6 Wk 33 19 89 99.3 Wk 10 28 100 99.3 Wk 22 73 232 96.0 Wk 34 Wk 11 31 126 99.6 Wk 23 69 225 95.8 Wk 35 Wk 12 40 141 98.4 Wk 24 76 185 96.1 97.613 2.1475
Week nr. Week nr. Week nr.
Standard Deviation
Average
17 The second priority for the ATP is keeping overall costs compared to overall output at a minimum; this is measured on aggregate and expressed by the actual cost per litre of paint. This plant level measure is a foremost a consequence of another plant level measure, namely the productivity of the plant. Productivity is measured as:
Output (tons) / FTE / time (week)
From the beginning of 2007, this measure has laid in the range of 4,1 – 4,8; the goal to be achieved has been set at a minimum of 5,0. The first possibility in achieving this is lowering the amount of FTEs whilst keeping output at a constant level, thereby increasing the overall efficiency of labour and lowering (direct) costs. Even though this statement does apply to the ATP, it is not true in all circumstances. Because, one can imagine an investment in a new machine which is enabling a plant to decrease the amount of FTEs whilst keeping output volume constant, thereby increasing productivity. In this case, (direct) labour costs might be reduced, but the (initial) costs associated with the new machine might be larger than the labour costs were in the previous situation, thereby increasing overall costs. The second possibility is increasing output volume with the same amount of employees or FTEs.
The goal of this study is to make recommendations on how to improve the service level. However, where variables are identified as having an influence on service level performance, whilst having an influence on productivity at the same time, such a relationship will be acknowledged and explicated in this report.
1.5 Summary
18
2. RESEARCH DESIGN
In the previous chapter, the organisation, the transformation process under consideration and an outline on the research background have been presented. In this chapter this is converted into a concrete research design; including the problem background, the central research question and the research approach combined with a demarcation.
2.1 Problem background
As mentioned in section 1.4.2, the main goal of this research is getting the ATP’s service level for its central stock point located in Breda up to a minimum of 99,2% structurally. Thereby the objective for this study has been obtained:
Provide recommendations and facilitate the implementation of actions based on these recommendations, to increase the ATP’s service level for Breda to a minimum of 99,2% structurally.
The addition of the term structurally is explicating the long term nature of this objective for the ATP. Also, it is setting a preliminary requirement for the recommendations to be made. 2.1.1 Service level
The basis for the ATP’s service level calculation is the discrepancy between the number of tins requested and actually delivered on a specified date by the Breda DC to its customers, measured on a weekly basis. An implication is that a requested quantity of an item, which is delivered in full but too late is regarded as fully missed, hence the notion of specified date is the former. The service level of the DC to its customers is calculated using this data. Multiple causes can be appointed for causing a shortage in the actual delivered number of tins; these causes have been categorised into:
- Missed due to commercial reasons - Missed due to reasons in production - Missed due to reasons in distribution
19
Demand Missed -Commercial (%) -Production (%) -Distribution (%)
Service level production plant (ATP) Full OTIF delivered by DC (Breda) 100 25 -15% -5% -5% 95% 75%
Table 2 ATP service level derivation
Thus, the ATP’s service level is determined by the number of tins associated with the items that have become manco (out of stock in the Breda DC). An item becoming manco is either due to the ATP’s inability to replenish the stock at the right time with the right quantity or indirectly due to actual sales exceeding forecasted sales. Since a certain level of flexibility is expected from the ATP, it remains responsible for manco items which have exceeded their forecast by less than 50%. Marketing & sales are responsible for the number of tins associated with the manco items subject to forecast deviations larger than 50%, since they are responsible for forecast accuracy.
By setting the required service level of the ATP at a minimum of 99,2%, a margin is installed allowing the DC some but little room for error, in obtaining its service level goal of 98%. 2.1.2 Focussed approach
Improving service level pertains to total plant performance; consequently it is very probable to find a large number of influential factors on this measure within the plant. This notion combined with the fact that this study is to be finished within five months, calls for a focussed approach.
As can be derived from paragraph 1.2.1, the core of the ATP’s transformation process is the department where filling, packaging tins and stacking a pallet takes place, i.e. the filling department. Performance of this department is measured by a daily calculation of the Overall Equipment Effectiveness2 (OEE); basically a measure of efficiency.
The fact that this core department was constraining output (being the bottleneck in the ATP’s operations) was not ignored in previous years, indicated by the numerous projects that have been undertaken in this department (Breukelman, 2004; Rijckevorsel, 2006). These projects, have led to an increase of the OEE pertaining to the whole department from 15% to 30% sustain ably over 2003 to 2006. Individual filling machines however regularly achieve an OEE of 50% and higher. An example of a project which has made the OEE increase is the implementation of SMED principles whereby set up times have been significantly reduced (Breukelman, 2004).
2 OEE theoretically is a measure incorporating the availability, the performance efficiency (the product of
20 In addition, continuous improvement through “Head-up” is still undertaken in this department on a weekly basis to eventually arrive at the overall 40% OEE target for 2007.
The trajectory in the transformation process in front of the filling department has received considerable less attention in previous years since this was not the bottleneck. In figure 2, this trajectory is presented on a more detailed level than in figure 1.
Figure 2 Order trajectories in front of filling department
The filling buffer is calculated daily and is expressed as the amount of volume ready for filling at the start of a working day, i.e. six o’clock in the morning. The filling buffer is expressed as a summation of all the quality approved batches (production orders) for which all necessary packaging materials are also available. The total volume of quality approved batches for which necessary packaging materials are not available is expressed as the non-packaging buffer. The filling buffer is a decoupling point. An implication of the filling department being the bottleneck is that this buffer is always of a level whereby the four filling lines cannot deplete it, i.e. there is always enough paint to keep the whole filling department running.
However, this reasoning is not relevant anymore in the ATP’s current operation. What is noticed by the author and confirmed by management and employees, among whom is the planner of the filling department, is that occasionally filling lines are shut down, in order to prevent complete depletion of the filling buffer. Appendix 3 shows that over May and June 2007 alone, 120 available filling hours have not been used for filling due to an insufficient level of the filling buffer. This occurrence is probably a logical consequence of the increased efficiency of the filling department; they are able to fill more volume per hour than before. This is not necessarily an unwanted situation, since fulfilment of demand would just indicate a matter of overcapacity.
21 providing the filling department with an insufficient output, i.e. an insufficient filling buffer. Insufficient pertains to its mix of finished production orders and therefore also to its volume. Its mix is determined by the release of production orders into the plant (the transformation process starts in production) and the flow of these orders through the processes within production and thus by lead times in production. Its volume is a direct consequence of its mix. In addition to the costs associated with manco items, potential productivity is not utilized because available FTEs in the filling department are not always able to produce end-products which have been requested by customers, having a negative effect on costs.
Based on the above, it has been decided that the area in which most can be gained with respect to service level improvement is production, which is thereby providing a demarked research area. This has led to a conversion of the initial objective into a more focused and demarked objective for this study:
Provide recommendations and facilitate the implementation of actions based on these recommendations, on adapting processes relevant to production in order to structurally increase the ATP’s service level for the Breda DC to a minimum of 99,2% by constantly having a filling buffer which mix of finished production orders is enabling planned and thus timely replenishment of stocks in the Breda DC.
The demarcation with respect to processes relevant to production is further explicated in figure 3, wherein these processes lye in the delineated area. This figure is based on the support processes dealing with production and filling orders and the departments wherein these orders are used to direct physical transformational processes.
22 If the filling buffer would fulfil the criteria stated in the objective, it would have been higher in times of high demand. This implies that there would have been no need to shut filling lines down due to an insufficient filling buffer and the number of manco items would have been lower, thereby a higher service level would have been achieved. A possible side-effect of the working of this scenario is that the filling buffer could end up constantly comprising a certain amount of volume whereby the filling department would become the bottleneck again.
Also, a requirement of the recommendations to be made is that they should be of a structural nature. The implementation thereof is thus implicating change, an aspect that therefore cannot be ignored and will receive attention when appropriate.
2.2 Central research question
For the objective, presented in the previous section, to be reached, a central research question aligning with this objective is developed. This has lead to the following formulation of the central research question:
How can the processes relevant to production be adapted in order to constantly have a filling buffer which mix of finished production orders is according to plan?
In order to structure the research to reduce the complexity of a research approach solely based on one central question, the following decomposition into sub questions has been formulated:
1) What does the transformation process relevant to production comprises? 2) How is the ATP’s transformation process planned, measured and controlled?
3) What is causing actual lead times to be longer than planned lead times within production?
4) Which causes are most significant? 5) How can these causes be eliminated?
6) What effect is the implementation of recommendations having on the ATP’s KPI’s? 2.2.1 Assumptions
23 Another assumption is that the capacity of the filling department is always sufficient in terms of enabling timely replenishment of stocks. Quantitative prove has been provided to support this statement. However, in reality under capacity in the filling department also occurs.
The implication of these assumptions has led to the formulation of sub research question 3. Since timely deliveries are planned for and not constraint by the filling department, it implies that actual lead times are longer than planned lead times in production.
2.2.2 Definitions
Lead time pertains to the processes relevant to production in this study; therefore first production will be defined followed by the lead time definition relevant in this study:
Production: “The physical transformation processes in front of the filling department - the processes within the production and QC department - with the filling buffer as their result.” Lead time: “The amount of time that exists between the moment an order is ready to be scheduled for production and the moment when this order becomes part of the filling buffer”
2.3 Approach
The approach taken in answering the sub research questions and eventually the central research question is presented in table 3.
What From By what Where
Question 1 Practice Interviews Chapter 3 Documents
Question 2 Practice Interviews Chapter 3 Question 3 Analysis Interviews Chapter 4
Historical data Data collection
Question 4 Analysis Literature review Chapter 4 Historical data
Question 5 Redesign Practical activities Chapter 5 Literature review
Question 6 Analysis Review of results Chapter 6
Table 3 Research approach
24
3. PLANNING AND TRANSFORMATION PROCESSES
3.1 Introduction
The research area is demarked on processes relevant to production, thereby not only including transformation processes but also planning processes. Therefore it is feasible that first a factual description of both these processes is provided, serving the purpose of providing a basis for analysis. This is expressed by the first two sub-research questions:
1) What does the transformation process relevant to production comprises? 2) How is the ATP’s transformation process planned, measured and controlled?
These questions will be answered simultaneously in this chapter, within a framework based on the sequence of processes in reality. This framework presented in figure 4 is also providing this chapter with the necessary structure by means of indicating the paragraphs in which processes will be described.
Figure 4 Framework for describing relevant processes to production
The processes that do not fall within the demarked research area are nevertheless described in order to provide a context for the processes that will be subject to analyses in chapter four.
3.2 Master Production Schedule (MPS)
25 The MPS is indicating the quantities of each item the ATP is required to produce at some point in the future. The MPS is a plan and a plan is a formalization of what is intended to happen at some time in the future. SAP is able to make the MPS because it is containing data on most aspects of the business, is able to utilize this data in algorithms and provides a linkage between several production locations, DCs and suppliers. By generating and constantly updating the MPS, SAP is materializing demand (forecasts and confirmed customer orders) into orders for the ATP.
3.2.1 MPS parameters
Generating and updating the MPS is performed using all item’s three parameters, which are presented in the following sub sections.
Forecasted demand
Item specific forecasts are made on a monthly basis by products managers; managers assigned to a portfolio of items of e.g. a particular brand. In making and adjusting these forecasts, they are using historical sales data and information provided by customers on e.g. planned promotions.
To obtain a daily forecast for an item, its monthly forecast is divided by the number of working days. By dividing an item’s current absolute stock level by its daily forecast, its stock can be expresses as the number of day’s forecasted demand can be fulfilled. In addition, confirmed customer orders for an item are also taken into account in generating the MPS.
Stock levels
From the ATP to end-customers, products move through two intermediate stages, namely a DC and a retailer/wholesaler. In both stages stock is held to anticipate upon actual end-customer demand. This reduces the risk of not being able to deliver what end-customers are requesting. However, holding stock is accompanied by costs. Therefore, stock levels for all items in the Breda DC have been determined wherein a balance is sought between costs and the ability to fulfil actual customer demand. These stock levels are:
- Safety stock levels (minimum stock levels) - Maximum stock levels
26
ATP Min. stock (days) Max. stock (days)
A (50%) 6 24
B (40%) 9 52
C (10%) 9 100
Total
Table 4 ABC classification and stock levels
Items’ expected turnover volume and value are utilized in the making of this classification because the amount of stock held is subject to ROI targets.
Delivery lead time
The third parameter is the system delivery lead time, referred to as planned delivery lead time. This is the number of days planned for the ATP to make the ordered quantity of items after an order has been released plus a number of days for transporting the item to a DC; this can be seen in figure 5.
Figure 5 Planned delivery lead time
The number of days a production order is planned to be in production depends on the type of paint to be produced and thus varies. Furthermore, the day after a production batch has been approved, is the day on which the items are planned to be filled and wrapped. The day after that is the first day of the planned transportation time which varies depending on the location of the DC to be supplied. The planned transportation time to the Breda DC is one day. 3.2.2 Period Order Quantity (POQ)
SAP is utilizing the parameters presented in the previous section in its POQ algorithm for making the MPS. The POQ algorithm is defined by Nicholas (1998, p.150) as an analytical model for lot-sizing process batches and purchase order quantities wherein the number of orders is reduced by putting restrictions on the frequency of orders, i.e. it starts with predetermined order frequencies.
This is illustrated in figure 6 and comprises the following steps:
1) Comparing the actual stock level with forecasted and confirmed demand 2) Determining the time on which the stock needs to be replenished
27 5) Determining the planned start of an order for the ATP
Figure 6 Example ideal POQ workings
In relation to ideal POQ workings, there are two possible scenarios that cause manco items to occur; the one most relevant for this study and illustrated in figure 7 is when actual delivery lead times deviate significantly from planned delivery lead times. The second scenario is when actual sales deviate significantly from forecasts; this is illustrated in appendix 4.
28 The scenario depicted in figure 7 in relation to processes demarked as relevant to production, is the focus of this study and an analysis on this scenario is presented in the next chapter. 3.2.3 Material Requirements Planning (MRP) and orders
The orders derived from the MPS are thus for the ATP to deliver and therefore produce a certain quantity of an item on a certain date. Since an item comprises several parts, sub-orders for producing and purchasing these parts also have to be generated. An order for the ATP can therefore be decomposed into several sub-orders; this decomposition is presented in figure 8 by means of an example. In this figure can also be seen that the sub-order for production enables the filling of multiple items. This can also be derived from figure 6 wherein finishing one order leads to the replenishment of multiple items’ stock.
Figure 8 Order decomposition
The creation of these sub-orders is also executed by SAP wherein Materials Requirements Planning (MRP) software is utilized. The optimum production batch quantity in relation to items’ stock replenishment needs is determined by utilizing the POQ algorithm in conjunction with the MRP software. Namely, this software links together all information about the parts that go into a finished product by means of Bill of Materials (BOM). In addition to figure 8, purchase orders for raw materials and production orders for lower-sefis required to make the production batches are also generated by MRP software.
29 with a specific start and end date, and related (preliminary) filling distributions. Furthermore, these planned orders are categorized according to capacity groups in the production department where the production order is to be made. This categorization is dependent on the production batch’ quantity and the required processes for it to be made (based on routing data in SAP). The listing of planned orders starts with orders to be made in four weeks and theoretically ends at that moment in time for which there is no forecast present in the system.
3.3 Planning production
The ATP’s planning system can be depicted as a push system since, materials are processed in batches according to a schedule for each workstation (ATP’s capacity groups in production), then moved (pushed) downstream to the next workstation (ATP’s filling department) where they are processed according to another schedule (Nicholas, 1998, p.262). This is further confirmed by the fact that both production and filling orders have a start and end date.
The process of planning is defined as determining which orders should be made at each workstation on a specific date. Following this definition for planning production, this constitutes the daily release of paper production orders by the ATP’s planning department. Before a production order can be released, the whole order has been converted from a planned into a created order by the planning department; the latter process will be described first. 3.3.1 Creating orders
The list of planned orders in SAP is per capacity group in production and sorted by planned start dates. This is not a fixed list however, as it might change due to orders placed by customers of the DC, whereby need for replenishment might become more urgent. Nevertheless, several planned orders for all capacity groups are converted into created orders by the ATP’s planning department on a daily basis.
By creating an order, all necessary parts are allocated to that order based on its BOM, i.e. raw materials, tins, lids, labels and stickers. Thereby it is also the trigger for the SAP system to initiate calculations using its MRP software and derive proposals for raw- and packaging materials to be ordered. Although order creation is the trigger, planned orders are also accounted for in these calculations.
The main consideration while creating orders is the proposed start date by SAP for production. Converting planned orders into created orders for the planner entails:
- Making sure that planned orders to be made in the coming four weeks are created - Adjusting proposed start dates for certain batches so that the workload for a capacity
group is evenly distributed
30 A number of raw materials and the majority of tins (80%) are delivered using VMI (Vendor Management Inventory); the creation of orders is automatically triggering orders for suppliers to deliver a particular quantity of raw material or a particular number of tins on a specific date. The remaining raw materials, tins and lids are ordered by the ATP’s planning department based on proposals provided by SAP. These proposals are the result of the former mentioned calculations which are functions of present stock levels in SAP, expected stock movements and required stock levels for a part in the future. Labels and stickers are ordered in the same way, only this is done by the LPC department.
3.3.2 Releasing production orders
Following previously mentioned definition of planning, it can be stated that releasing production orders is equivalent to planning production. Since production orders are released two days before its planned start date, production is planned two days in advance; this can also be seen in figure 6. In this process, SAP provides support to the planner by indicating production orders’ planned start dates and also provides a priority based on the current stock levels of items that order is meant for to replenish. Nevertheless, the planner decides which production orders will be released, thereby he is making the plan.
Before the actual release of production order the planner checks if everything needed to finish the whole order is present or will be delivered in time according to SAP and/or confirmed by the supplier. It occurs that e.g. tins for a particular item cannot be delivered on the planned date by the supplier. If the production order for that particular item has to be released based on the urgent replenishment of another item, the filling distribution of the created order is adjusted (the item is removed or interchanged for another) and the production order is released. Another possibility is that the release of the production order is postponed until a confirmation is given by the supplier on the delivery date of the requested tins.
Two additions to the above have to be made. The reason production orders are released two days before its planned start date is that on the day after the release, the necessary raw materials are delivered by the onsite raw material warehouse to the ATP. The day after that is when the production order is supposed to be taken into production. The second addition is that through the described mechanisms in this paragraph, high demand is converted into the need to release a large amount of orders on a daily basis. This need is satisfied by the planner since the plan includes all production orders proposed by SAP which are complying with the indicated checks. Thereby he is assuming that production orders are started on the planned start and planning against infinite capacity.
31 day also denotes the first day of a production order’s lead time, following the definition provided in section 2.2.2. The way production is scheduled is the topic of the next paragraph.
3.4 Scheduling production
For any working day, a plan is made by the planning department indicating which orders should be made at each workstation. Workstations are capacity groups in production wherein a capacity group is a collecting of tanks and dissolvers/mixers; from hereon referred to as areas. The classification of these areas is based on machine resources needed to make different types of paint and the volumes of available production tanks. For further detail, reference is made to section 3.5.1.
In figure 9 can be seen that on day 1, the production plan for day 3 is made by the planning department, as described in the previous paragraph. The orders released on day 1 are given to the GL on the next day; his task is to schedule these and the production orders still waiting for production for day 3. The process of scheduling has been defined as allocating human resources (production operators) to production orders for a specific date. Machine resources are not included in this definition because production orders have already been allocated to an area. The notion of orders still waiting for production is implying the existence of a backlog comprising production orders. Therefore, by scheduling the GL has the ability to change the plan made by the planning department, hence the schedule made by the GL is taken into production on day 3 in figure 9.
Figure 9 Planning and scheduling cycle
3.4.1 Scheduling rules
32 This distinction is made because all three GLs perform the scheduling task since scheduling production for the next day is done by the GL who is working the evening shift. Since the shifts change every week, the GL who is scheduling also does.
The inputs for the scheduling process are the same every day, namely: 1) Stack of orders sorted per area in production
2) Next day’s available production personnel
Based on these inputs, the allocation of production operators to areas is determined. Herein, operators are assigned to their fixed capacity group as much as possible. Logically related to the skill sets of the operators with a fixed contract, three out of nine are always working in the same area. Noteworthy is the color maker, who is one of these three operators and the only one who is allowed to schedule his own work. Some operators are working in teams of two, wherein usually the second member is a third party operator. In addition, all GLs are taking physical discomforts of their personnel into account, i.e. an operator with a sore back won’t be dumping heavy bags of powder into tanks.
Hereafter, available production orders are assigned to an operator (or a team of operators) whilst keeping fixed tank availability into account. Tank availability plays a role because within capacity groups, tanks are assigned to the making of particular types of paint, e.g. waterborne paint, and these tanks are not always empty. The availability of tanks is either estimated by a GL or checked by looking into them.
Still, assuming exactly the same inputs, it is probable that the result of the scheduling process differs between the three GLs. From interviews held with all three GLs, it became clear that every GL is balancing aspects of the schedule to be made in a different manner. These aspects are:
1) The total volume of lower-sefis and sefis that will be produced
2) The adherence to the planned start dates set by the planning department 3) The allocation of production operators to capacity groups
33 3.4.2 Dispatching production orders
After the decisions presented in the previous section have been made, the production schedule comprises an allocation of production orders to operators which in their turn have been allocated to areas in production. Therefore, the production schedule does not indicate which orders should be made first, which should be made second, etc. The process of setting the priorities on the orders comprising the schedule has been defined as dispatching. In the ATP, dispatching is foremost being done by the production operators themselves. Most of the time, an operator is free to determine the sequence in which he will make the e.g. four orders that were assigned to him. Occasionally however, the GLs are setting a priority on a production order.
3.4.3 Adjusting the schedule
During the schedule’s execution day on, situations might arise which are requiring the schedule to be adjusted by the GL. The following situations have been identified:
1) An operator calls in sick at the beginning of the day or gets sick during the day
2) A tank truck delivering raw materials arrives unexpectedly, requiring a production operator to offload it
3) The GL making the schedule was not aware of an operator’s day-off
4) A production batch cannot be made because there is no tank available for it to be made in
5) A production batch cannot be made because a required raw material or lower-sefi is not available
6) A production batch cannot be made because a piece of equipment is experiencing a breakdown
7) An operator in the previous shift did not finish the last batch he was making due to lack of time
8) An operator in the previous shift did not start an order at all due to lack of time
34 operator for some time or gets totally lost. When it is eventually discovered that an order is totally lost, the planning department has to print out the order once more.
Since the schedule is made for a whole day, adjusting the schedule requires communication between the GLs and the operators during a shift and between both GLs during the transfer of shifts. For the latter, half an hour is available each day.
Since the processes of planning, scheduling and dispatching production orders have been presented, the next paragraph is on what this transformation process exactly entails.
3.5 Physical transformation processes
A finished production order is equivalent to a quality approved batch of paint which is ready to be lid off into tins by the filling department. In this paragraph, the physical processes are described whereby a batch of paint is produced, the processes whereby a batch is approved on quality is the topic of the next paragraph. Since the paint itself constitutes the main part of the ATP’s finished products, first a brief introduction is given on the characteristics of paint and its main components.
Paint is used to protect (preventing rot, rust and chemical erosion) and embellish numerous products. For paint to fulfil these functions in line with customer needs it is required to be produced according to the specifications / recipes set by R&D; this is where paint and its way of production are developed after extensive testing.
The total number of producible recipes by the whole of Akzo Nobel is vast compared to circa 750 produced in the ATP. However, in generic terms paint consists out of a number of basic components (Kooistra, 1991):
• Binders; binders are solid or low viscous materials which are usually mixed with solvents; their function is to ensure the makeup of a dry layer of paint.
• Pigments; pigments are dry powdery materials; their function is giving paint the required colour.
• Solvents; solvents (30-35 percent of paint) are fluids which are giving paint the required viscosity. A certain viscosity is needed for making paint applicable to surfaces. Anorganic solvents such as water are increasinly replacing organic solvents such as terpentine.
• Additives; additives are substances added in small quatities; their function is to minimalize negative features of paint (e.g. an anti-yellowing additive) and/or to enhance positive features (e.g. dryers whereby the pace of drying is increased). • Extenders; extenders are relatively cheap, powdery materials that for instance have an
35 These are all basic components; per type of paint these components have been determined more specifically in terms if their chemical compostion and in overall proportions. In the next section the required processes to produce a batch of quality approved paint in the ATP are presented.
3.5.1 Products, processes and areas
Products/batches made in the production department are eligble for clusterin on numerous criteria, e.g. quality, colour, gloss level, production time. In the context of describing distinct processes, it is sensible to cluster products according to differences in required processes that occur before a batch is part of the filling buffer. This notion is enforced by the fact that distinguished capacity groups in SAP are also partly based on this clustering. This has lead to the identification of the following product groups:
1) Lower-sefi’s 2) Primers
3) Lacquers; both waterborne and solventborne
36
Figure 10 Processes per product group
Regardless of the type of product, a production order for a particular batch comprises: 1) The sequence of obligatory and optional processes
2) The amount is of time a process is required to take, e.g. disperging for half an hour 3) The intervals on which certain raw materials need to be added
37
Figure 11 Areas within production; volume and number of tanks; flow of product
In the caption, the total quantity and volume of the tanks per area is presented. From this can be derived that there are two types of tanks, fixed tanks (connected to the floor) and mobile tanks. Furthermore, the flow of either type of product until it reaches the filling department is also made visual. Lines connecting fixed tanks can be regarded as piping through which products are pumped; another possibility is for the product to be led off into mobile tanks at the bottom of a fixed tank. This is possible because every area is located on the first floor except for the mixing department and the ATD, indicated by a bold delineation.
In addition, the difference between a finishing tanks and buffer tank connected to the filling department is that in the former a batch is still being produced after it has been pumped over, whereas the latter is merely meant for temporary storage. Lower-sefis stored in mobile tanks are virtually used in every area, therefore these connecting lines have been left out.
Within areas, tanks are allocated to a range of products within a group which is based on the following features of the product:
1) White and light colored versus clear and dark colored 2) Solvent borne versus waterborne
