Teoria Industrial Technologies

Job shop

Suggested when a huge amount of items different from each other is manufactured or volumes of single products are much lower than the total amount, or the production mix is very variable along the time. In a job shop, machines are grouped based on the technological processes involved. In the same department, we have machines similar in technological capability. Machines can have different characteristics (high/low performances, size, etc.) but they share the (technological) process they support. The idea of functional departments is related to this concept. Each department has its own capability, and the labor is divided into departments according to task specialization. Workers are skilled according to the technological process involved; however, this organizational choice leads us to short visibility on the entire process of transformation (the workers can’t see the final product).

Each product has its own technological routing in the system through different departments, so it is important to follow the specific sequence of the operations (process plan) in which we have to select the departments according to the product specifications. In a job shop, materials are moved according to the required product routings (e.g., from one department to another), while items are moved grouped in lots. The logistics (material handling through the factory) is characterized by high flexibility because handling must enable different types of movements. There’s no rigid interconnection through departments. This allows us to have flexibility also in stockholding points since the machine responsible for the next operation could be busy working on a different lot. There are some storing points called buffers. Inter-operational buffers can be of two types:

• There’s a buffer (e.g., by a machine) in every functional department.
• The buffer is common for all departments, known as a “system buffer.” This second choice is better because space occupation results from the aggregation of different needs (the aggregated demand from the various departments). Summarizing, we occupy less space while satisfying all the requirements.

Strengths

High flexibility: this characteristic has different dimensions (also combined). Flexibility is the ability to change or react to new situations in the environment with little penalties (in terms of performance reductions of money, time, etc.) to achieve some reachable states in a short-medium term. It can also be seen as the ability to meet some market requirements. Short to medium-term flexibility:

• Mix: the ability to meet some market requirements in terms of variety of products (characteristics) supplied in a given time. A wide range of product types can have different production volumes.
• Volume: the ability to deal with variations of aggregated volumes (staying in a fixed range of the physical possibility offered by the machines).
• Product (customization): the ability to customize some product already existing or introduce a new product in the portfolio.

Medium to long-term flexibility:

• Product (innovation)
• Expansion: the ability to have new technological capability or improve the production capacity without a layout reorganization, assuming that additional spaces were planned at the beginning of the layout design.

This flexibility comes from:

• The characteristic (flexibility) of the machines installed (so it does not depend on the job shop!)
• Material handling, e.g., trial productions. I can’t stop the flexibility that permits different productions in the same time as normal production, but I have to use some of the machines for these pre-series. It is possible thanks to the material handling that doesn’t affect normal production.
• Routing flexibility: I can define alternative product plans and sequences (of course, I have one preferable option and alternative routes that usually are not as performing as the preferable). This flexibility is also useful to adapt to changes in workloads and for the production of pre-series.

Subsequently:

• Low impact of breakdowns helped by routing flexibility and material handling flexibility: buffer avoids the propagation of the breakdown (or it reduces its impact) of the system, which does not stop due to spare machines in each area and the presence of alternative technological routing.
• Low obsolescence of the system: aging leads to decreasing performances of the machine (it is a normal degradation) and increasing costs as a consequence. In this case, product, mix, and volume flexibility avoid obsolescence because it permits changes to meet market needs.
• Low initial investments and high chance to invest in the future following demand.
• Easy know-how transfer between workers with the same competencies.

Weaknesses

Production management is difficult because many options (offered by the configuration of the job shop) imply many decisions. This fact leads to several problems:

• High WIP, as a consequence of the desire for protection, because it is difficult to schedule production and manage the working plan.
• Lead times are long (because of queues) and characterized by high variability. This variability comes from the different mixes and routes (workload changes time to time and causes queues).
• Difficulties in estimating delivery lead times make it difficult to use MRP and other similar systems that need LT as an input.
• Low utilization rate of machines: Machines are often off because they wait for materials that are in queue by another machine (bottlenecks bring slowdowns).
• Limitations in efficiency (machine efficiency).
• Quality characteristics of the products can vary for different pieces.
• The calculation of production capacity is difficult because it depends on:

• Mix of jobs that have to be manufactured (that can vary a lot during the time – long time decision).
• Technological characteristics of jobs (long time).
• Complexity of the pieces that have to be manufactured (long time).
• The possibility of having different routes (it’s a short-term decision).
• The number of machines and their state: of course, breakdowns are not predictable.
• Lot sizes: this choice depends on internal production policies (I don’t want to do many setups on the machines to avoid stopping it lots of times – if setups cost me a lot in terms of time consuming, or I can prefer production following the effective demand to avoid high stock holding costs).
• Ability to schedule jobs: I have to schedule different lots (of different products) on different machines, and it can also happen that setups depend on the sequence, so I have to choose the best time-performing sequence.

System design of a job shop

When designing a job shop system, we usually have to answer some questions, such as:

• How many machines do we need to meet the demand?
• How many operators do we need to meet demand? (It depends on the automation level and also on the organizational decisions made, i.e., specialization level.)
• Where are the bottlenecks (some of the resources are limiting the production capacity)?
• What happens if the production mix changes?
• What happens if I introduce new machines?
• What are the effects of reducing lot sizes?

Rough design of a job shop

This is done to identify the typologies and number of machines to get a given product mix in the desired quantity. Steps include:

Production mix definition

• A) Identify all the product types (I can use statistical methods combined with past experience to split all the products into different categories).
• B) Estimate yearly demand for each product type by looking at the past demand for similar products (of course, there is a variable margin of uncertainty).
• C) Define the lot sizes for each product type (I base this decision on the number of lots I want to manufacture during the year).

Routing definition

• A) Define the main routing for each product type.
• B) If possible, define alternative routings.

Machine identification

Based on routings, it is possible to identify all the machine types necessary to manufacture the production mix. For each product type, I have to calculate the total time of the operations that have to be done on the same type of machine (Tij is the time required for manufacturing the product j on the machine i based on past data and NOT including setup time).

Calculate the yearly workload NHi for each type of machine i:

Tij*qj total working time (per year) required for the pieces I want to sell on the market. But I can’t consider that, unfortunately, I can have losses: with the term “Yield” we refer to every loss of production we have. First of all, not all the outputs can be good and pass the quality test: in the first part of the formula, I have to subtract the scrap rate from the total production. I can calculate the percentage of scraps the machine realizes: scraps/inputs, and in the same way, I can calculate the percentage of good production with this formula: good outputs/inputs. My losses can also comprehend the time we are waiting for materials or operators. For example, HC=0.8, we have to wait for the operator for the 20% of the time the machine is active. Ai is the total time the machine can work. Of course, it can be down for maintenance, so it won’t be available. TR represents moments in which the machine is used to produce new products for technical tests (it’s connected to the innovation aim). The result of this formula is a “gross workload.”

Calculate the number of hours available for each machine type i (of course, I have to make a comparison between the number of hours required for the expected production and the hours available):

• s is the number of shifts per day, SE explains the results of the management complexity.

Calculate the number of machines of type i necessary to manufacture the production mix, given the yearly demand. The number obtained must be rounded up or down depending on:

• Machine-type cost: if the machine type is not expensive, I can decide to buy more machines of the same type.
• Possibility to outsource the production of some product types.
• Possibility to use alternative routings for some product types.

So, I can decide to reduce the workload or improve my resources. However, I have to remember that the coefficient is just an estimate. Rounding it up will pledge the production desired. Then, I have to evaluate the number of shifts per day, computing the yearly costs of adopting 1, 2, or 3 shifts per day. The energy cost doesn’t depend on the total amount of hours (I can make a machine work for two shifts or two machines working on only one shift), but it is of course possible that there is a discount for using energy in particular moments of the day (the unit energy cost can change during the day). If we have more machines working (the case in which we produce using one shift), of course, we have higher costs related to utilities, depending on their type. So, we have a trade-off because maintaining only one shift leads to higher costs for facilities and machines, but opening more shifts involves less cost related to energy. But if I want to maintain the same shifts open for all the departments.

Of course, there are machines that can do more than one piece during the same operating cycle, but this hypothesis is not considered in our previous formulas.

Manufacturing cells

Here, we have groups of machines that support different technologies: we group together parts in part families and machines into cells. Therefore, we have to find criteria to split parts (sharing some similarities – shape or material) into part families. The machines are grouped based on the processing requirements of the part families. The main difference is that manufacturing cells are based on the parts, not on technology (as we saw for the job shop), so we have different technology capabilities in one cell. It is a product-oriented layout.

The main objective is to have separated technological routing: each product has its own routing within the cell (this is the case of complete cell independence when no inter-cell move is required) according to the idea that each product should visit only one cell, so the layout is important.

When cellular manufacturing is applied, it may lead to two cases:

• Re-arrange existent equipment on the factory floor (e.g., machines).
• Operate with new equipment, often incorporating various forms of flexible automation (e.g., from machines, material handling equipment to FMC/FMS). In other words, a typical question related to system design is required – “which machines and their associated parts should be grouped together to form cells?” – Before rearranging existing equipment on the factory floor or incorporating flexible automation.

Strengths

• Rationalization of material flows: the flow is linear, and the machines are close in distance because they are visited by the same parts. Having close machines leads to visual control made by operators.
• Setup time reduction, because we have similar technological parameters and so require similar tools. Economically speaking, we use the same tool for different products, leading to cost reduction. Having short setup time allows us to produce lots very similar to the effective demand, avoiding stocks and ensuring a quick flow through the system. Additionally, if we have similar shapes, we can have a generic fixture that is beneficial for the parts of a family.
• Production management is easier, also thanks to the visual control made.

Compared to the job shop, we have:

• WIP reduction, because we do not need as high protection as a job shop, since operators can better control the operations; lot sizing is low, and we have a lack of stranded material flows.
• LT reduction (also considering variability), because flowing time is reduced by the rationalization of material flows, and transportation time is lower. We can also have “overlapping” when, due to the closeness of the machines, we can move units (that are fewer in number than the production lot) from a machine where they’ve already completed all the processes to the next one, anticipating the next manufacturing process and making this flow “continue”; the fewer the products manufactured in a cell, the shorter the queues and the lead time.
• More reliable estimations of delivery Lead Time: we have control on LT (that is also reduced) and on its variability.

In the cell, workers are fundamental:

• Job enlargement + job enrichment for employees: workers have tasks regarding every machine in the cell with different capabilities; they have responsibility for these tasks and can make suggestions and decisions, so they should have problem-solving competences.
• Teamwork within the cell: the same people are assigned to a particular cell, making it easy to build up a team.
• Unification of product and process responsibilities.
• More control on the quality characteristics of the products: this depends on team working, but it is also due to the quick feedback operators have.

Weaknesses

• Difficulties with workload balancing within cells: in different planning periods, production volumes may change, so workloads change. How can I adapt to this change? In the job shop, it is a simple problem because we have routing flexibility, but cells are built to keep cell independence, leading to rigidity in routing options (I want to maintain flows in different cells separated). Of course, this affects lead times and delivery performances as well.
• Problems related to production mix variability.
• Difficulties with the application to the whole stages of the production chain: not all stages can apply manufacturing cells. When the cells are not applied to every stage, other systems have characteristics that mitigate the benefits (and weaknesses) related to the cells.
• In some cases, the necessity of more machines than in a job shop (leading to higher costs in machine acquisition): if we aggregate production volumes (in the job shop), it can happen that I’m aggregating products with high workloads and products with low workloads, compensating for this aggregation. In manufacturing cells, I separate different production flows in different cells, so I have to provide machines of the same type in different cells, and I cannot compensate for production volumes anymore. The total workload on a machine of type i comes from the following formula: NHi = ∑NHi,c with i as the index of a machine type and c as the index of a cell.
• Difficulties managing technological operations outside the cells: In the case of a perfect separation between cells, I do not require operations outside the cells, but if a product requires a machine outside the cell, it creates a disturbance and affects performance. Inter-cell movements lead to lower performances in terms of LT: operators in cell2 are lending their resources (for some time) to other flows coming from a different cell!
• Problems related to breakdowns: routing flexibility is limited. If I have only one machine of type i in a cell, and this machine breaks down (ex: machine A in cell2), what can I do? I do not want to move flows to another cell, so I can only wait until the machine is totally repaired. This could be a problem because if I have put machine i in a particular cell, most of the products (or all - at worst) related to this cell should pass through machine i.

It is not the costing issue that is driving the decision to install manufacturing cells, but the main drivers are the quality and time performances!

System design

I have to identify which machine type I have to include in every cell, so I have to identify which product type is related to every cell. Group technology is an approach trying to manage diversity by exploiting similarities between parts. These similarities are the starting point to make a rationalization between the different product types. The group technology steps are the following:

• Data collection regarding the production mix and technological routings (not different from the job shop approach): we have to look through experiences and the new production expected to collect the information needed.
• Classification of products: we already know the different product types, but here we are identifying relevant characteristics to build the system.