CONTRIBUTIONS TO THE MODELING OF MANUFACTURING PROCESSES FOR THE IMPLEMENTATION OF THE KANBAN METHODOLOGY IN THE AUTOMOTIVE INDUSTRY

Abstract

The scientific paper presents an original contribution regarding the modeling of the manufacturing processes for the implementation of the KANBAN methodology in the automotive industry with immediate applicability in the area of Electronic Control Units (ECU). We present our own point of view regarding this modeling, taking into account a pertinent analysis of the early phase of production for the first samples. It starts from the analysis of the processing time at each individual process station with the final goal of so-called "balancing the production line". The research analyzes the advantages and disadvantages of the principles underlying the Kanban methodology, emphasizing those applied in Just-In-Time management, with close to zero stocks. It was also insisted on the results obtained following the production flow modeling for the application of the KANBAN methodology

Headnote

Abstract - The scientific paper presents an original contribution regarding the modeling of the manufacturing processes for the implementation of the KANBAN methodology in the automotive industry with immediate applicability in the area of Electronic Control Units (ECU). We present our own point of view regarding this modeling, taking into account a pertinent analysis of the early phase of production for the first samples. It starts from the analysis of the processing time at each individual process station with the final goal of so-called "balancing the production line". The research analyzes the advantages and disadvantages of the principles underlying the Kanban methodology, emphasizing those applied in Just-In-Time management, with close to zero stocks. It was also insisted on the results obtained following the production flow modeling for the application of the KANBAN methodology

Keywords: Kanban, Just-In-Time, Station, Electronic Control Unit, Process, cycle time, Input, Output, Automotive, Production, Quality, Supermarket.

1.Introduction

Traditional manufacturing strategy is driven by 'Push system' with aimed to keep large of inventory of product according to customer forecast. However, this has created big problem to people on floor in dealing with high of WIP inventories, unsynchronized production processes and producing unnecessary stock. Due to that, established company like Toyota Motor Corporation has moved to next level of manufacturing approach or strategy by adopting Kanban system [1].

Nowadays, many companies have faced customer pressure to produce products with high value, to deliver quality product at a competitive price. They have to focus to meet these needs as a requirement to remain and stay successful in today's market [3].

To achieve world class manufacturing level, many companies have moved toward becoming lean by adapting Just in time (JIT) practice in their manufacturing system. The Kanban pull system under JIT approach offer great benefits. The purpose of this system is to link manufacturing activities to market demand. Many research carried out has shown that maximum benefits can be gained in manufacturing areas such as reduction of inventory, reduction in lead time, improvement of value added time, increased productivity of process and also improving product quality [5].

2.Kanban System

The Kanban method was created in 1947 by Taiichi Ohno, who defined the Toyota production system. The reason was Toyota's significantly lower productivity compared to its American competitors. The basic idea was to organize the flow of materials in production according to the principle of the supermarket: if the customer takes parts from the shelf, exactly the same number of parts will be refilled. The maximum number of pieces on the shelf is calculated and limited close to the actual consumption of the customer. By using this systematic demand to manage production, the stock could be reduced to the minimum required quantity of parts to keep production running. As a result, the return speed of the material and the flow of material have been dramatically improved [6].

The principles presented in Figure 1 are:

* push principle (traditional thinking): each stage of the process is scheduled based on planning times and safety stocks. The production sequence is planned in as much detail as possible. Stock buffers are used to compensate for process instability;

* flow principle (lean thinking - high volume production): processed items are passed immediately from one process stage to another, without stagnation between them. Production orders are sent at the beginning of the flow;

* Pull principle (lean thinking - low volume production): orders are "pulled" through the process according to customer consumption. The impulse to execute orders starts at the end of the process chain. Thus, customer consumption controls the market.

The name Kanban comes from Japanese and means "Card". These cards are the basic elements of a Kanban system and are used to simplify the transfer of information and control of work systems. The general objective is to achieve a better flow in the work systems by reducing and limiting the work, which is carried out in parallel [7].

To be in line with customer expectations, work in progress must be extracted from an "accumulator" based on customer order priorities (Figure 2). In production, it focuses on the flow of materials and "pulls" material through cards from the supermarket. In administration and R&D, they focus on tasks and extract tickets from pending tasks. The major rule is the new work as input is started only if there is a customer order and free working capacity (people, material, etc.) in the working system.

To stay competitive in the automotive industry, it is not possible to accept any more Push Systems which is creating a lot of waste like inventory, movements, waiting time and quality issue. It must be a change from Push to a Flow Systems pulled by the customer to fulfil these expectations and to keep cost to competitive levels. Kanban is responsible for the design, management, and improvement of flow systems. This has led to more productivity and flexibility in day-to-day work and creates more value for the customer, the company, and the employees.

3.Identifying Nonconformities during Processing using the Kanban Methodology

In the last period, the automotive industry has developed a lot, increasing the complexity of the vehicle. The major car manufacturers have decades of experience and knowledge in the field of engineering and car construction, which are mainly mechanically operated. Now, the industry has reached a turning point where electronics and software are replacing mechanical hardware as the most critical and valuable vehicle components.

Vehicle complexity continues to grow as vehicle functions expand to higher levels of on-board automation, networking, and processing.

Manufacturers are now able to manage the massive complexity of modern vehicles while reducing time to market just to stay competitive [8].

To be competitive in the market, the manufacturers of electronic components in the automotive field, implement various methods on the production line, for modelling such as Kanban. Since it is not produced in stock, the quantity of electronic components produced is dictated by the requirements of the final customer. The Pull principle is a Lean thinking, dedicated to low volume production where orders are "pulled" according to customer consumption. The impulse to execute orders starts from the end of the process chain in Figure 2. To produce electronic components, Figure 1 is amended in the form of Figure 3.

The supply of the first "Pick and place" process station is made with raw materials consisting of Printed Circuit Board (PCB) and basic electronic components such as: diodes, resistors, capacitors, integrated circuits, transistors, and others. This process station assembles the basic electronic components on PCBs. The product at the end of the process is the Printed Circuit Board Assembled (PCBA). This is the input element in the "Supermarket" and is also a Kanban signal. The processing time of the product at the first processing station, "Pick and place" is about 72 seconds. The next process station is "Assembly". Here the input elements are the PCBA from the previous station together with the housing. The output is the final product in house. The average processing time at the assembly station is about 69 seconds. For the moment, the average processing times at the two processing stations are comparable.

The last process station in the production chain is the "End of Line" (EOL) station. This is an electronic product testing station on the production line. The input element to this process station is fed from the Supermarket. The processing time at this station is about 125 seconds.

The automotive industry manufactures millions of vehicle parts every day and end-of-line (EOL) testing - the last checkpoint before a product leaves the factory - ensures that those parts are built to the appropriate specifications. This makes EOL testing one of the most crucial steps in automotive precision manufacturing [2].

Dr. Andreas Himmler, Senior Product Manager from dSPACE, believes that "EOL testing benefits companies as a whole, since fewer flaws mean fewer warranty claims or recalls in the future. Moreover, it helps to develop the quality image of products and companies" [4].

The problem of processing time at the EOL process station is the very long processing time, about two times longer than the rest of the process stations. This non-compliant process results in fewer parts of the product and is reflected in the capacity of the production line. The clock (frequency) of the production line must be 85 seconds, to meet the minimum quantity delivered to the final customer. If the clock time is greater than 85 seconds, 125 seconds in this case, in the end, the customer is affected.

Because the cycletime (processing time) of the EOL station is much longer, it becomes the weak link in the production chain, it becomes Bottleneck. After the first product tested, it can fill the accumulator capacity of the Supermarket, and the outlet of the upstream station, the assembly station, has no place to store the products. This is also reflected in the previous station, upstream, at "Pick and Place". In this way, the production line was blocked at the EOL station. The line clock is equal to the longest processing time of all processing stations. In this case, it's 125 seconds.

4.Improvement Proposals

To make processing time more efficient at the EOL process station, two scenarios are available:

* addition with additional testing equipment;

* modelling of the processing station to dramatically reduce processing time.

The first approach, by adding additional test equipment, is not an easy task. The high costs of an equipment, somewhere around a few hundred thousand euros, must be considered because they are produced in very limited series or manually. Also, the delivery time of the equipment is between 10 and 15 weeks. To balance the production line, in this case, two more test equipment's need to be purchased, as in Figure 4.

The individual processing time on each EOL station is also 125 seconds, but in parallel, on 2 stations we will have in 125 seconds, 2 tested products. So, in total, using 2 test equipment's, we will have 2 products in 125 seconds. In other words, the average processing time is described in (1):

(ProQuest: ... denotes formula omitted.) (1)

in which:

* T_total - Total cycle time spent at EOL test process

* T_EOL - Cycle time spent at one EOL test process

* N - Number of EOL test process

So, the total EOL cycle time will be 62.5 seconds per product, processing time. This cycletime is much better and is comparable to that of previous process stations.

Now the "slowest" processing station will be the one with the longest processing time. In this case it is the "Pick and Place" station, with 72 seconds.

The touch of the production line will now be equal to the slowest station and is equal to 72 seconds. A major improvement is observed, from 125 seconds to 72 seconds, almost twice as good.

The second approach is to modelling of the processing station to dramatically reduce processing time.

This approach does not include additional cost for the purchase of new equipment, but a change in process execution methodology is required. It must be investigated why the process takes so long to process.

After the cause is found, a solution must be found to fix the problem, and then the solution will be implemented.

In other words, the automated functional testing process in the production flow must be modelled to fit the clock of the production line. Only then can it be said that the Kanban methodology can be implemented in the production process. After remodelling the EOL process station, the cycle time was reduced to 85 seconds in Figure 5, which satisfies the production line clock.

5.Conclusions

After modelling the EOL test process in order to optimize it, as shown in Figure 5, it can be seen that the processing time at the "End Of Line" process station is much shorter than in the previous situation (without modelling), but still 13 seconds longer than the slowest "Pick and place" process station.

The EOL process station remains the slowest station, but the new test (processing) time is no longer critical because it falls within the new processing time, equal with 85 seconds as the requirement of maximum production line clock should be, in order to not affect the end customer.

By reducing the test time, budget (cost to buy a new equipment) and time savings were made, no longer being necessary to supplement with another additional test equipment and its delivery time between 10 and 15 weeks. However, it takes (cost of innovative work by an employed engineer) between 2 and 4 weeks to implement a new test procedure for reduce the cycle time spent on testing the electronic product per test run. The hired engineer (local human resource) is much cheaper than an expensive equipment bought from a third party.

KANBAN can better be applied in production only if all the process stations from the production flux have cycle time balanced or comparable.

The above EOL cycle time reduction meets this Kanban requirement. By reducing the processing time, at the last station, it means that the customer can request materials more often without blocking the capacity of the supermarket.

References

References

[1] Bonvik, A., Gershwin, S., (1996). Beyond KanbanCreating & analyzing lean shop floor control policies, Manufacturing and Service Operations Management Conferencewww.donis.ro. accessed on:03.05.2012

[2] Acerta. (2022). Why wait to catch quality issues at end-of-line testing? https:// acerta.ai/blog/anintroduction-to-automotive-end-of-line-testing/

[3] Agus, A., Hajinoor, M., (2012), Lean production supply chain management as driver towards enhancing product quality and business performance: Case study of manufacturing companies in Malaysia", International Journal of Quality & Reliability Management, 29(1), pp. 92121.

[4] dspace. (2022). End-of-Line Testing.<https://www.dspace.com/en/inc/home/applicat ionfields/> foo/ end-of-line-testing.cfm

[5] Singh, B., Garg, S., Sharma, S., and Grewal, C. (2010), Lean implementation and its benefits to production industry, International Journal of Lean Six Sigma 1(2), pp 157-168.

[6] Sirang Klankamsorn, (2020), Determination of the Number of Kanban for Automotive Axle Production, Conference: 2020 IEEE 7th International Conference on Industrial Engineering and Applications (ICIEA). doi: 10.1109/ICIEA49774.2020.9101930

[7] Singh, N., Shek, K. and Meloche, D. (1990), The Development of a Kanban System: A Case Study, International Journal of Operations & Production Management, Vol. 10 No. 7, pp. 28-36. doi:10.1108/01443579010140498

[8] Burcicki D, (2020), Automotive Industry On Course To Disruption And Evolution, https://semiengineering.com/automotiveindustry-on-course-to-disruption-and-evolution/