Uploaded by Nicholas De Palo


The course is about the design of different production solutions.
Process/diagram/layout/type of technology description:
The factory: quality of workplace  quality of car, the work eviroment is more pleasant. They
build engine in the same place. The factory is a clean building, it looks more scientific than
industrial, there they have the department where they produce the sand (to melt aluminum, the
quality of the san dis important to avoid defects) …. . Then thers is the control process (con il coso
con la punta a rubino di TMQ). 25 days to create one ?. Then they ripassano by hand. There are a
lot of robots and robotic applications. Reparto assemblaggio (to assemble components into the
engine). (assembling by hand)  than final assembling area (assembling with the car body, that is
created 80km far from Maranello, made out of alluminum). The central body is put together by
individual craftsmen. Each part of the car is designed to save the weight  23 different materials
to make the body. 5 axes ratio device  another control. Hen they add the doors and the other
parts. Artisans control if the finish is perfetc. Then special light control to look for imperfections.
And then the body goes to maranello. Than owners can customize the car.
Painting process:
- Wash and clean
- elettrcode
- seiling application (?)  manual activity to see the imperfection od the painting just looking at
the reflections of the light.
Than final assembly line: where they add logo, than laser cutting machine to select the best part of
the leather. (just 6 FF every day produced, now 20 per day, but Ferrari is limiting the number of
car, as part of their strategy (you have to wait 1,5 years to have a Ferrari FF). Than adding of the
engine (less the 2 min).
Is the typical car maker industry? Yes in terms of technologie used, but here the context is
different, they are keeping inside all the production (also for electronic devices are designed
inside but produced outside, unlike many car makers), they make customer order the car (and
wait)  pure ETO company.
Which prodiction technologies? Funding, drilling&milling, cutting, painting, assembling. How
are this tech organized? They are presented step by step, department by department. (the
engine is made by 2 departmants, than department of interior, and so on). This company has
been designed department by department, or job shop by job shop. This production
organization is really common (putting together the tech that are similar OR putting together
the machines that are used to make the same part of the product). How many area we can
1. Foundry & body plants (body not in Maranello, where they produce just components
for the body)
2. Engine (Ferrari is one of the last car makers that are still producing engines inside) 
two main shops, fabrications of component and assembling of component + many
quality control, for a marketing reason.
3. Painting shopping side
4. Final assembly line (they produce also the inside, tailoring and cutting leather). Today
that line is double and much bigger, it’s the typical syncrono line + other check
Classification of production system. In the Ferrari case we have all this systems in place inside the
company. The main organization is in job shops. About the engine production, They have parts
production organized in jib shops (putting together machines that are doing the same stuff). There
was also example of cells and transfer lines (like authomatic assembly line for Romeo&Juliet
About the assembly, there was an assembly line, in the last part of the plant, but it’s not the onli
assembly case. Thinking about doors and wheels, they are mounted in fixed position (not example
of assembly shops).
Ferrari is a manufacturing industry, but foundry is an example of a process industry. This is not the
normal situation, this is part of the peculiarity of this particular company. The main part of foundry
is working as a batch production.
Lesson 10/10/16
It’s basically machines + tools and technologies used to produce finished goods or components.
These are resources of the PS. In some contents, like mechanical industries, the role of HR is
controlling, making set up or stuff like that, not being in front of the line, while in other industries
like tailoring, the impact of HR is much more heavy. It depends also on the geographic area you
are. The machines could be in different status: working, set up (preparing to work, like changing
some tools, the machine could be waiting for pieces ecc, it could be blocked for installation
problems, for materials of not the proper quality, for breakdowns, or because it is under
maintenance or repair. Out task eill be DESIGN the production system through some design
decision: how many persons we need, how the interactions will be, because our operators are
resources. Operators could be very skilled and could work with machines controlled by the
computer, or we could use some less sill operators when it comes to load something or stuff like
that. Production is made of single elements called Work Orders that are planned and scheduled in
the production plan, and then are “launched” and executed. Production system is dynamic: what
we plan may not be what will happen, everythingk could happen to machines, operaotrs or work
order itself (like customer ask to be delivered its product with more urgence). We won’t deal with
managing production system, but designing solutions that allow us to be more flexible. Flexibility
is a performance of the production system. Movements are important, how things are moved is a
big issue in designing. Produstion system could be organized in different kind of layout (=how the
machines are physically installed). Layouts are different according to the different performances,
there are basic definitions:
- Resources (machines and operators) are usually gathered in departments (area, production
floors, or shop floors, or genba that is a Japanese word).
Layout provide company with different performances, in terms of production volume or
production variety (or mix), elements that are almost always in trade-off. We will manage this
through layout design. For example, let’s imagine that a product A, whose demand is really high (A
in Pareto ABC), while product B is not asked. The product A is to be produced in a large quantity 
we are interested in designing part pf the production system to produce just product A in a small
amount of time ( transfer line, with machines that work very fast to produce one product in
small time). For product B, on the other hand, it makes sense investing in job shops, that are
cheaper. We will always start thinking about volume and mix to produce for every firm.
Main operational performances (time) (performance = measuring how a system is working):
- Lead time, or flowing time, or throughput time, that is the amount of time spent form starting
to end of all the production activities;
Lead time is measurable from the outside, from the customer perspective, from when the
order is sent to the firm to its arrival back to the customer. From the firm point of view,
lead time is made by different action. In our course we are interested in the production
lead time, that is understanding how the machine manage the production (in toto,
inclusive of unproductive time) (vd slide “the concept of lead time”. It is also possible to rearrange the planned delivery date, in case of ritardo, making some interaction with th
customer. The point is that what happens inside is crucial in determing the customer
service quality.
- Setup time, time for the set up  it’s a matter of understanding how are production will be
able to create products (in terms of flexibility or rigidness)
- WIP, in a specific second, this measure how much work that plant has to produce (measured in
terms of material, number of pieces inside the plant in a specific moment, or it can be
measured in terms of time: time that is needed to finish what have in the plant
- Production rate, rhythm  speediness of the system, how many pieces we are able to produce
in a certain amount of time. It’s important to know the unità di misura (if it is pieces, we are
measuring a rate)
- Cycle time: time needed to produce a piece (inverse of production rate). We will design
assembly line on the cycle time given by the market.
We need to understand how the system works in its dynamic  we have a system, it’s a general
line. We know that we have 4 connected machines: materials enter in the first machines, stays
there for the processing time of the 1st machines, than it moves to the following machines. We
also the production rate of the single machines. Which is the slowest? B! B is called the bottleneck of the production flow. Being a line, the LT is the sum of all the processing time. The highest
production rate achievable is the production rate of the slowest machines of the line (we can’t go
faster since we need to pass through the bottle-neck). This means that inside the system we can
have some machines that are working less than what it was expected, because they are waiting for
the slowest machines: we could change the slowest machine, or buying a new B machines to put
in parallel with the existing one. We won’t reduce the LT, but we will increase the production rate.
Other operational performances:
- Saturation: how much the system is used.
- Flexibility: (of production, of mix, of product, of utilization). A system could be much more
flexible depending on how many products could be made, or at least how much does it take to
make a set up. We don’t need the detailed definition of Flexibility.
- Availability: the impact of breakdowns or maintenance of the machines. The machine is
available if it’s not affected by breakdown and so on.
Project is one-of-a-kind production. We have different solutions that are giving us different
performances in terms of LT (normally, with kìjob shops the production LT is higher), WIP (job
shops is a situation in which we register a higher level of WIP, for every kind of production), setup
time, utilization (normally the best is the line).
An important issue that among Led time and WIP there is a connection. Lower WIP level  lower
LT  being able to have less uncertainty and easier control of production, with planning in
advance and being able to deliver on time (not to go out of the market, kind of market
According to the situation in which we are, some elements are not up to us (as production
designer) that are operating conditions. Foer example, if you plan to buy new machines but the
owner doesn’t want to buy them  you are forced to deal with old technologies, vedi slide 30.
There are some fixed, blocked situation (like the amount of money we have to design the
production system). We will work with constraints.
Main cost categories of the production system:
We are dealing with physical stuff  physical costs, real costs.
- Cost of installation (at capital assets expenditure level, CAPEX)
When we are creating a new plants, we have to pay for the feasibility study (paying the
company that will make some preliminary analysis on the available ground, consultant
company); we have to pay for people, material needed to lunch the production, that are
operating costs but for us are part of installation
- Operating cost (impact on operating costs, OPEX)  costs that occur during the life of the
system we designed (rent, morgate, logistics service, maintenance). They are variable or semivariable costs or fixed.
We have also figurative or opportunistic costs, that are not moving real money but represent the
economic value of a decision and the loss of an opportunity, a loss in revenue or in potential
- Depreciation of machines
- Inefficiency costs like overtime work costs, stockout costs (measured in terms of lost margin,
or penalties I have to pay if we had a contract problem with the customer), stock holding cost
(the storage is blocking my money), setup cost, cost related to make an activity in subcontract.
We have to solve the trade-off between level of stock and customer service level (capacità di
risposta veloce al cliente).
A relevant cost is:
- Future
- Avoidable
- Differential
We are dealing with the fact that we are designing production system that will have a long line
(10-20 years). It’s a long lasting decision. In this kind of situation is important to consider the
normal depreciation costs. For example, my company can decide to install a new plant in Italy or in
Poland. To make this decision, I have to think about taxes I should pay on the depreciation of my
How to evaluate an investment in production:
Cost analysis
Revenues or saving analysis
Cash flows analysis
Discounted cash flow analysis (net present value)
Payback time analysis/ internal rate of return
VIDEO – illy coffee
Ship transportation from south America, than truck transportation to the plant, than sacks are
empty and coffee depurato from impurities.
Can produced internally.
Illy is an example of a process industry, working with big plants to produce a cheap product, but
have other manufacturing departments. Everything is integrated to provide us wuth a specific
product that is coffee. Our design activity will start from an analysis of the demand: what do
people ask? This analysis is made in a really good way by the company (capsule ecc). The layout
was possible to see in the video.
Diagram: the relevant full process is the treatment that is done to the coffee, from raw materials
coming from all over the world. They test and analyse a lot from the very beginning to get the best
final result. There are several steps like drying, roasting and let’s say “cooking”. We can describe
this first part of the plant as automatic big machines, where the operators are skilled people that
just control, check, verify, and supervise. This is typical for all the process industry. Why investing
so much money in this kind of activity? Because people buy coffee if it is affordable  companies
needs to be efficient (high demand, low margin  economy scale needs to be applied, this is why
Illy has very big automatic plants, to cut costs (the customer is actually not interested in how the
coffee is made). A process plant can’t produce a large amount of different product  the final
product itself it’s quite the same, always coffee  process industry is limited to the technology, if
you have an assembly company you can assemble whatever you want potentially, while process
industries ate considered rigid. Anyway, we can still design a flexible production system (producing
coffee with different flavours, different cups/package). The can building process is not very often
in a process industry.
Batch design production  solution to design big plant for process industry that wants to be
flexible, having a decoupling point (coffee beans grinding)
Process industries big machines usually can have big impact in terms of availability  we can add
some machines that can be used just in case something goes wrong.
Organization  focus on the operation part. Organizations exist to make money. The thing is how
do they make money? Through transformation of input in output. They need a lot of different
resources (that are reflected in the different functions)
With the same title of “operation manager” can have different roles, like “product industrialization”
and “production” or “production” and “purchasing”.
What does an operation manager do?
- An operation manager holds responsibility for all the resources that need to be transformed in
output, to allow the process to get a physical output. Resources are people (assigning tasks,
discipline), physical assets (machines, energy, building), intangible assets (know-how, software,
other tools, guide lines, procedures)  management role (management of plant resources, to
achieve certain targets in terms of people and assets)
TARGET: some example are flexibility, wip, lead time… we have to cluster them in macro
categories (Quality, Delivery, Service, Cost  the quality of production must be good, that is for
granted, delivery means that you have to be short and accurate, service related to delivery, and
cost, we have to make money, if you don’t have leverage on the price, you have to think about
the costs, cost is OM’s field of play). OM HAS TO MAKE LIFE QUANTITATIVE  giving figures,
transforming qualitative info in quantitative data, with data you can make decision.
- He is not responsible about sales (price definition), but for costs (how much are the cost,
depending on the use of inputs)  financial role
- He has to comply with legislation, laws, standard. For ex if we have a recognition for people that
behave in a certain way, you have to apply the policy in the right way  governance role (legal
- He stays always up-to-date, having a forward looking vision, thinking about innovation and
continuous improvement linked to manufacturing and production
- He has to make sure that “life within the family” is ok: manager operation is the supplier of
someone and the customer of someone else: relationships need to be positive.
MOOG. The feeling of the people was good. Look at the video.
1. You have to understand what you are doing to decide how to do it
2. This company has ETO (the trigger point is customers’ orders) and MTO (just for some product,
or for re-buying of the same product, when the design has already been made) processes,
following Wortmann classification.
3. The product we will talk about is about flight simulation, used to train pilots. This simulator is
fully electric (electric motion, you need a motor, a screw) (while the existing one were idraulic
 new opportunity).
4. Around week 10 (when uslenghi entered) the demand was getting higher and higher. Problems:
the output was not consistent (look at the different up and downs); offer was really lower than
CASE STUDY: the process has to offer more  resizing in terms of capacity. Production process and
system were already there.
Production system was made of multi elements (tools, people, machines, areas). There were 3 main
area (input like data about demand, drawing, bill of material, physical input like raw material;
manufacturing cycle, made by steps thread grinding, heat, end shaft, od grinding, assembling;
resources like operators than run the machines, defiyne priority, indirect support (the material has
to arrive on time and so on), machines and space, logistics and movements).
The constraints were the number of m^2 of space, number of operators (we have to control the
cost), installed machines (they are expensive, they require space and time when they are new to be
installed and start working properly), product features, all the products must look like the same.
1. demand forecast has to be checked to see if it is really accurate. Do I trust the data I am
taking all my action on? Is the info reliable? Is the demand going to be stable? What is the
long term picture? A forecast can be wrong by nature, by you have to take into account your
data  demand was proved reliable
2. are other data ok? In terms of bill of material, designs, manufacturing cyle: the way of
manufacturing a major number of pieces can be different, to pursue efficiency. Maybe, if we
want to pump up the volume, maybe manufacturing cycle needs some changes
1. Raw material: do we have a problem in stock? The first graph is about demand and volume
at stock. The demand was flat (per week, at the beginning), the managing of raw material
was “point of reorder”  EOQ(?). In the second graph, the demand changed: how to
manage the inventory? The point of reorder are the same (EOQ) but sometimes happened
that they went out of stock  running out of raw material is a big problem, they had to
review the managing of te inventory. Last picture is about old purchase strategy and new
one. At the beginning they were overlapped, than purchasing were closer in time (higher
frequency) and higher in volume. From a time perspective, they had to react very quickly,
running the production much above the demand to fill the previous weeks missing in
production, than they can rebalance the two dimension (production and demand). The same
goes for raw materials  buying more RM closer in time. The question is: is the supply chain
able to meet this change? Up to know, we have the theroretical model, but OM have to deal
with the reaction of suppliers, that depends on the capability they had. In this case, initially
the most qualified supplier was not able to support them, and they had to look for a new
supplier to qualify (?, 1:05)
It is defyined in work order. Each product has its peculiarity, each product code has its own work
quarter, and for each operation is defyined some fixed performance, this info need to be
transformed in something physical by the operator. The cycle is made up of 5 operation, and KPI are
cycle time and set up time. In this case, OD grinding was the longest activities (in cycle time). The
question is how many hour do we need to produce enough product (in machine hour). They found
the hour that was missing (108), and they calculated the available capability that was higher than
what they need. But problems can be looked form different distances  the factor was not doing
just the simulator product, and they had also other customers. At the end of the day they discovered
that they need more hours than the available ones, but they had constraints. Different alternatives:
- More capacity needed: I add more capacity  buy another machine, hire new operators.  not
efficient, too much money and too time required and no space available for new machines, no
more operator because no machines enough and for financial problem.
Working on the cycle, on the way we use resources: reducing grinding hour, making the machine
go faster, reducing set up (with lean) and so on  there are trade off, the machine could require
more maintenance for example. Or maybe it is possible to skip some OD grinding steps 
reviewing the way my product is designed (and so manufactured). You can’t find the solution on
the first shot, you have to do different trials
Dividing the grinding phase in two smaller phases, zooming more in. Some steps are more critical
to function that others, and need to be controlled more. These steps could be outsourced,
renting capacity from the others (OD grinding is something that can easily find outside). It
becomes a supply chain issue (arranging suppliers, logistics,..)
Resources can be features that you use for daily life in production, like a system to move products
(trolleys are not enough, there would be too many trolleys). How to handle this? Loading trolleys
with more pieces and then using technology to azzerare the weight.
Time is important. (1:25) Maybe it is possible to have a shorter time to the customer, using MTO
and ATO (but we have to move to the x axe to change something in how the product id thought.
SECOND CASE – john crane – oil and gas, petrolchemical plants – pipes and pumps
We are on the vertical axe, product management decided to come to a new design, for efficiency
and to meet the desire of the market. They made a prototype, that was tested  they decided to
go for the new one.  life cycle of the product (1:37)
Prototype  ETO, that they wanted to move till the manufacturing to order. They want to develop
the manufacturing cycle toward MTO, without putting in stock.
Multi variables problems (it involves inputs, manufacturing cycle, resource). Constraints are similar
to the previous case, but we have also crtical product feature (that can’t be changed for
differentiation element that are competitive advantage on the market). Decision criteria: cost target
and volume.
- Bill of material: it is preliminary. It has to become more accurate as I move from prototype to
introduction and then to maturity (product life cycle).
- Supply chain: we don’t have detailed agreement in the first part
- The same as up.
- They used the resources that were already available.
Technology is in the centre of all the variables. It is something that I need to choose from the very
Elements to assess which is the best decision: data and numbers, production volume, target cost,
availability to invest, and product requirments.
Manca l’ultima parte, va troppo veloce.
Risk = probability that something bad happens. You don’t have to worker other big loads, but then
there are other details to be considered.
In a plant we usually have resources (=machine and operators). These resources interact: operators
do the set-up, do maintenance, load something. This two resources may not be mutually available
in the same time. For example, there are more machines than operators  operators are needed.
This problem is called “interference”, caused by the different availability of resources, and it should
not be overlooked. We have to design that system in the best possible way, trying to anticipate this
problem to avoid it.
- How many operators to put in the plant? How many machines do we have to buy?
- Where are the bottlenecks: machines or operators?  we don’t have to consider pnly how many
machines, but how many people as well.
We’ll take the case of an existing plant, with a certain # of machines and operators. So we have to
understand the cause of the level of productivity, and re-conduct it to machines and operators.
We’ll check if the bottleneck is in the machines or in the operators.
Situation: let’s imagine that we are checking what is ongoing on a production system. The graph
shows three resources (machines/departments/factories M1, M2, M3) and time (Gantt chart). We
are representing how a specific resource has used their time. M1 was operating in a small amount
of time, in a status in which worker was required (like set-up). Then, M1 was lunched and was
operating on its own. Then, worker moved to the M2 and lunched it again, and then M3 to do the
same. After some hours, M1 finished to work. The operator “was asked by the machine” for some
activities (like uploading, changing stuff…). In the picture everything is going well. Operators and
machines don’t interfere at all.
Second case: the worker is asked at the same time on two machines (M3&M1  M1 has to wait for
him, and also M2 have to wait). Interference happens when the operator is needed in more than
one resources, and some of them have to wait for him. We have to consider the dynamism of the
system to understand this kind of issues (that could cause lack of productivity).
This is the problem setting  problem solving?
- Hiring new operators (with obviously a cost)
- Increasing the productivity using outsources and contractors;
- Changing the list and the sequence of the work orders (scheduling approach).
- DESIGN PROPERLY THE SYSTEM: When we design a system, we have to take as reference point
the worst case possible that could happen, according to the best knowledge we have on the
systems in that moment, to a design a system that works safetly and is able to satisfy the original
request. A production system is leading along the time, we have to be available to understand
its dynamism (Gantt chart).
Interference is a matter of evaluating how a system is working, in term of performances, basing on
the problem that is intrinsically there. We’ll use the measuring approach to get a rough estimation
of time wasted due to the interference, that could be measured ex post (with the Gantt), or ex ante,
to anticipate it. If I have to design the system, I have to ask myself in advanced how much time could
be wasted and how I could reduce it (state space method, with simple basic assumptions like the
fact that machines work independently one from another (that actually is a strong hypothesis since
it’s not the same for every system), and there is exclusivity among states (the machine is working or
not working at all, tertium non datur, linear approach).
STATE SPACE METHOD: we have a list of step to be done with some formulas.
BASIC ELEMENTS: let’s take a system made up by 3 machines and 1 operator, that could be in
different state (working, waiting, asking for operator and so on). Let’s imagine that operator des juts
set-up. We use a table to define the possible situation in which the system can be (state space table,
with game theory at the base). 1 is referred to the status of “machine working without operator”; 0
is referred to the status of “machine requiring operator” or “service request”. How many possible
state could happen along the time (TIME is our design variable, because the system is dynamic)? 
combination.(𝑛𝑘), 𝑛 = 3; 𝑘 = 0, . . ,3 = 8. In the first case, the system is working; in the last case the
system is blocked. In the middle, we find situation of different level. The table is useful to have an
estimation of time wasted due to the fact that we might have interference. Which is the worst
We can transform the table in one time, the time expected to be wasted due to interferences,
applying a procedure.
1) Model the state space table (si,j)
2) Transform the table in probability (pi,j)
3) Evaluate the probability of each state (vd formula produttoria)
4) Evaluate the expected interference due to each state;
5) Calculate the expected interference time due to all states;
6) Understand the impact of the total wasted time to the overall capacity of the system (reduction
in the rate of production);
7) Calculate utilization rates of machines and operators.
To calculate the probability, we have different approaches.
- If we have knowledge about the time in which the worker is required, I calculate the total time,
then the time the worker is requested, then average;
- Occurances: how many time the worker is called? How many time the system works without the
worker?  I use the state space table.
In any way, this is just a simplification of real cases (rough estimation), because usually interferences
depend on many factors:
- Production mix;
- Technological characteristic of product types;
- Number of machines and their state;
- Schedule and ability to schedule (work orders list).
Firms are introducing some IT tools, software, to check on line and on time the current status of
each system, but in this course we won’t design an IT system.
It is a concept with different facets. We’ll take the most general perspective: availability refers to
resources. Let’s take a system or a machine: that machine in a specific time has a brake or a failure.
Maybe it is not properly designed, maybe it’s a matter of the production flow that causes some
defects, maybe the operators don’t work properly, or maybe for physical reason.
The impact of these failures on the system are measured under availability or unavailability, if the
system is working or not, without having failure. There are specific ways to measure availability
(lead time to repair or how often the failure happens). For us, availability is something given. We’ll
design the system knowing the fact that the system could be more or less available.
The availability of a component of the system has an impact on the whole system, in many cases,
but not always. A production system with a peculiar design, according to which it could be more
easily influenced by components’ failure. Design a system (or knowing the design), we can have an
expectation on its availability, we can have an estimation of how many problems could occur and
what their impact could be in terms of availability.
If I have a production systems made by two machines, theoretically speaking I can improve the
availability adding new machines. Adding a machine costs  which is the proper trade-off between
cost and availability?
Inside the production system, a lot of movements, flows, dynamic elements happen, that shouldn’t
be overlooked.
Ex: we have M1, then a buffer (decoupling point, stocking position), then the production flow goes
through the 2nd phase, where M2, M3 and M4 are in parallel (one work order could pass from m1m2-m5 or m1-m3-m5 and so on). We can decide among m2, m3, m4 which to use.
If m2 has a failure, the impact is less important than the case in which m1 has a failure, thanks to
the job flow (?). The availability of the system is fully based on the availability of m1, while it is less
based by the availability of the second department, that is made up by 3 machines.
So, up to know we have to:
- understand the concept of availability
- understand the impact of availability of each component on the whole system.
Let’s take the example of the job shop (= 3 machines that are doing a certain amount of production.
Remember: production capacity = production rate = throughput rate) made by three machines +
operator. We want to check and to have an idea of the impact of the single component of the system
in terms of availability. We have to fill another table, focusing on the machines and on the specific
working situation of a single machine, looking if in a certain moment the machine could have a
failure (possible state: working, not working just because it is broken – we don’t consider set-up and
staff). Best case: all the machines work, no impact in terms of reduction of capacity. Worst case: all
the machines are broking, no production capacity at all. In the mid, we have different situations.
Our point is to make sure that our design allows a good availability.
Again, we have to follow a process, but according on CAPACITY and not time ad before. We find the
expected availability of the system, result of the expected reduction of the production capacity of
the system (TPC = theoretical production capacity, it is the capacity that in theory our system should
be able to provide).
How can I have the stochastic knowledge to setting probabilities for availability? I can use historical
data, or specific knowledge about technologies and machines. When we design a new system, we
have to decide how many machines to buy, when we look for machines we have to check their
“theoretical availability” provided by suppliers. The procedure will not give us an extremely precise
answer. Behind the procedure, we’ll use the table and probability knowledge.
Availability depends on many factors in real life:
- Technological issues
- Number of machines and states
- Ability to schedule preventive maintenance
Production processes is divided in process industry and part production (or manufacturing industry).
We can divide again manufacturing in fabrications and assembly. Fabrication is a list of processes
that can’t be regenerated (when you drill a part, it’s drilled); in assembly actions are not so
We’ll start from fabrications.
A job-shop it’s a way to organize a production system in terms of machines and operators, where
machines are grouped on the basis of technological processes involved (similar machines in the
same department). This is the most intuitive organization. The good point of this approach is that
it’s easier to fix machines problems without moving machines. Shop is another way to say “area”.
These shops are able to work step by step on the different production activities. Shops are easier
way to control what’s going on with a particular kind of machines, it’s a way to manage better
resources. In fact, it’s a way to make the operators a flexible resource, that can work on different
machines of the same shops, that are similar in terms of technology.
Let’s look at the example, we have 5 departments (with technology A, B, C, D, E). Normally,
technologies in Job-shops have to be mixed to create a product (one single shop is not enough to
create a finished good). In this way, with shops I can produce a large variety of different products,
changing the sequence of shops visited. Job-shop production system has a high variety of different
outputs. This is an advantage but can turn into a drawback. In fact, we may have too many backand-forth between the departments, physically inside the factory. We need to spend time to move
products and to organize and plan it, to manage the sequence of the different work order, to balance
the utilisation of resources being able to satisfy the demand, without boosting lead time. This is
difficult. It’s difficult to see where products are, to decide the priority between different products
that are waiting for the same machine.
Job-shop layout is the most used, because it is the most flexible (especially for SME) for different
industries (textile, printing, mechanical, electronic and electro-mechanic system). Thanks to
flexibility it’s easy to use the lever of the volume (partially), and also of product mix. As far as
expansion is concerned, If I need extra capacity in a department, it’s easy to add new machines.
As a consequence, because of these benefits, this approach is characterized by:
- a high availability (low impact of breakdown): if we have a problem in one of the machine in a
department, I am always able to produce thanks to the presence of other similar machines;
- low obsolescence of the system: if department A is getting old compared to competitors, I just
have to buy a new version of technology A, I don’t have to redesign the whole system.
There are some weaknesses too:
- managing in the proper way the resources  limitation in global machine efficiency (using the
system as much as I can, in the best way) (in a flexible system, I’ll have efficiency problems);
- qualitative characteristic of products can be different for different pieces, this happens always
because of complexity in management, because it’s not enough to certify one time the
production, due to the fact that I change the sequence to vary the mix  it’s difficult to certify
the quality of production;
- production management is difficult
o high WIP: in the logistics I have to realize to move things, most of the time I need to have
buffer, area where to stock for a while the WIP that are waiting to be processed in the
next shops;
o Lead times are long: because WIP is high (directly connected elements), we have high
waiting time;
o Difficulties in estimating delivery lead time and production capacity;
o Low utilization of machines, because in job shop 80% of saturation is a good number (20%
of machines are under maintenance, or doing set up, or just waiting for products or
operators). Job shop is the typical contest where we face interference problem.
To recap, a job shop is the easiest system to imagine and design (just applying common sense), but
is one of the most difficult to be executed and managed, once it is in place.
How can we manage this system? Planning some “safe zone” that guarantee a proper work.
Practically speaking, we have to answer to this list of questions:
1) How many machines (and which type) do I need?
2) How many operators do I need to meet demand?
3) Where are the bottleneck? (we always have bottleneck, it’s better to leave them in the less
critical department)
4) What happen if the mix changes?
5) What if a machine breaks down?
6) What’s the effect of reducing set up times or lot sizes;
7) What if I add another machines in the system?
Rough design of a job shop
Why rough design? This will be the approach we’ll follow. We’ll take some design decision that will
be the basic for later entering in deeper level of decision. At first, we have to decide how many
machines to buy, further specific decision (like (??)..) that an industrial manager should not
overlook, we won’t explore them in this course. This is why we make a “rough” design, we create a
design common to all the different kind of products and technologies.
Step 1
We have to know what the system has to produce 
- Identifying type of product;
- Estimating yearly demand for each product;
- Defining lot sizes for each product type.
In our exercises this step will be just given data, in real life it’s much more complex, we need to
collect a lot of info to make the assumption needed to design.
Step 2 – routing definition (operating cycle)
We have to design a system where products flow, according to the activities that have to be done
(routing, from the rout that has to be followed). In our exercise this will be in a given table; in
realty it’s much more complex again (which products, how these products are produced, how
many hours and so on). The worst case in our exercise will have 10 products (10 routing); a normal
mechanical SME has more than 100 routings  assumptions and simplifications needed.
If possible, it’s useful also to define alternative routing.
Step 3 – machine identification
Example of mechanical SME: that kind of company (that produces machines) work with a
catalogue of machines, that is presented to customers (other companies that are looking for
provider of machines), that can ask for customization. After few month, the production machines
are delivered to the customers. Our exercise will start just asking how many machine we need to
buy (we won’t decide which tech we need, which type of machines, we won’t look for a provider,
we won’t sign a contract and so on). This step is the most difficult one in real life.
 Remember that under this order of steps there is logic, we start from product mix and not from
machine identification because the first think I have to think of is “I have to make profit selling
products  what machines do I need?”, I am not interested in machines in se, without a purpose.
Step 4 & 5
Collecting info about the time that is needed for productions (time for operation). Time is our
variable! We might be told that we have to make x products per years, that one product needs to
be produced in 1 minute  we have to calculate the time needed in total. The element to
consider to know how much time we need to work for is (total time for production):
- Quantity Q (of j products  different kind of products);
- Tij processing time (1 min)
- Failure: estimation of system availability;
- Roles of the operators (people don’t work always in the same way);
- Set up time, an issue I have to face (and how often I have to do set-up);
- Scrap rate (the demand is Q, if we have 10% of scrap rate, I have to produce more in order to
sell Q)
- (transportation time, it’s not part of the formula, but in other version or in specific contest
where transportation is directly connected with the machines it might be considered);
- trial rate: example, normally the production facilities are used also to make prototype and
experiments, that “steal” time for the normal production. This is not a problem but we have to
consider it anyway;
The formula is a sum (for all the different products that pass through the shops). HC, A and TR are
“safe coefficient”, used to increase the production to be safe in case something wrong happens
that will increase the time needed. These coefficients are given in our exercise; in real life we can
found it manuals or in companies’ report or from suppliers (tests and experiments).
The first factor of the addition is about “real production” or “effective production”. 3600 and 60 at
the denominator because we have to turn production time from seconds to hours and set up time
from minutes to hours. We add to production the number of set-up, in the second factor of the
addition (I usually do set-up every time I change the lot of production).
 we find the number of hours needed for production according to best knowledge we have and
according to available safe coefficient;
Step 6
Now we have an estimation of time needed. Now I have to decide how much time I want to work:
do I want to open my system 24 h/d, 7/7, or do we expect to work two shits per day, 5 days per
week, 50 weeks per year?
Available hour come from working
hour (shifts, days..) and a safe
coefficient (scheduling efficiency  I
can’t expect that everything will
happen in the same way we planned,
some activities will be inefficient by
definition (especially for job shops,
where a good SE is 80%, I am accepting that 20% of available production will be wasted as a
normal thing!!).
Step 7
Step 8
This procedure is done in alterative way. If I discover that I need X machines, I could decide to buy
them, or to increase the shifts or to go for outsourcing, or to reduce expected set up, and so on.
We have to consider availability of resources, cost of resources (or of outsourcing), cost pf set up,
cost of warehouses (mantenimento a scorta). These decisions are at operating level.
Maybe I can change the routing.
We have to use a formula that helps me in evaluating the number shifts/day, computing the
different yearly costs.
Remember that we have to design machines and not products.