Manufacturing Systems Types

Let us first look at the essential components of Manufacturing Systems:

The process of designing a manufacturing system therefore must engage upon the design of each of the above four components AND their integration.

Notice that this figure is pretty much consistent with Prof Sohlenius's architecture of Manufacturing systems, which is partially reproduced in the following figure. While the following figure implicitly assumes the important role of the human in each of the modules, I have explicitly placed it separately in the figure above, partly because it helps in highlighting the importance of planning the human aspects of the system.

(source: Lecture notes, Prof Gunnar Sohlenius)

We first look broadly at these four elements, and subsequently, we shall study each of these aspects in somewhat more detail.

Informal definitions:

Physical Systems refer to all physical aspects of a manufacturing system, including the factories, including the facilities, machines, tools etc., the raw materials, the material handling systems, the work in process, as well as the products.

The Operation refers to all aspects of decision structures that determine how the system functions. For example, does it use a Toyota style pull production, or does it depend on an MRP II system with forecasts driven production? How does the plant manager determine the size and sequence of the jobs to be done on each machine, on each day?

The Information in a manufacturing system refers to all data that will be accessed by some function/person/decision-maker/software etc., and whose value may be used deciding upon an action. Examples include design data, machine data, tool data, inventory status, process data, vendors, clients, personnel data and so on. It is likely that for any MS of reasonable complexity, one will need automated data handling facilities, e.g. a DBMS. I will also loosely include in this definition, mechanisms that are required for the flow of information, that is, Information Technology (IT). This includes communication protocols (such as MAP, TOP, ISO-OSI), etc.

Humans refers to all personnel, vendors, customers, etc. Personnel related issues include: what is the capability level of available labour, what is the working culture (1-shift, 2-shift availability), how many do we need to hire for a given MS, what is the level of training needed, what policies will lead to better working environment etc. Customers are another essential human element in the design of MS.

We begin the study with the physical systems. I will use the following classification of manufacturing systems, which uses the material flow type as its basis:

Product based: These are manufacturing systems that are designed specifically for a fixed product (or product type). The underlying principles are those of interchangeability, and division of labour. These systems may be highly automated, or human-labour intensive. However, in every case, they must operate at relatively high throughput rates.
```
Further divided into:


	Continuous production (e.g. chemicals, food processing etc.)


	Discrete part production. Further divided into:

		Assembly lines

		Transfer Lines
```
Process-based: these are facilities where machines of the same type are grouped together physically. Some examples may be seen in job shops, workshops, prototype makers, tool makers.
Flexible Manufacturing systems may loosely be categorized as highly automated versions of process-based systems.
Cellular: Cells are formed using GT, and typical cells are small clusters of a mixed bunch of machines that can handle a particular family of products. Such layouts are used when the batch sizes of orders are not large, but throughput times required are lower.
Fixed position: For single units of a large item (e.g. construction project, MTR line, ship building etc.)

The above are all the physical production systems. In addition, we shall look at material handling systems, including transportation machines and inventory handling systems.

Once we have a physical system all installed, we need to worry about the operational aspects. This includes, among other things, Production Control. Several important things here include:

Lot sizing;

Scheduling;

Process Planning;

When we design the physical systems, we are concerned with the capability to produce the designed part. When studying the operational aspects, we are more concerned with the efficiency at which we are working. Therefore, it is essential for us to know our goals, when we make operational decisions. Some typical descriptors used for Manufacturing Systems include:

WIP (Work in Process): The number of parts that are currently in the shop floor, either being worked upon by a machine/operator, or waiting at a buffer or in a queue.

Production rate: number of finished parts being produced by the system in unit time.

Throughput time: the time that a part spends in the system from the moment it is released from the inventory to the time it leaves the system.

Usually, we would like to make operational plans that are "good". That means we should know what we mean by "good", especially since often, different desirable objectives tend to be conflicting [which is what gave rise to teh famous saying: Good, Cheap, Fast. Pick any two.] Here are some definitions and some typical objectives:

Assume that jobs coming into the system are identified as J_i.

Due date, D_i: date when the job is expected to be completed.

Completion time, C_i: time at which J_i is completed.

Flow time, F_i: length of time J_i is in the shop.

Lateness, L_i = ( C_i - D_i).

Tardiness, T_i = max( 0, L_i).

Typical objectives include: minimize average flow time, minimize number of tardy jobs, minimize average tardiness, minimize the makespan (makespan = time to complete all the jobs), minimize the maximum tardiness etc.

Once we define our goals, we can test out which operational method (heuristic) gives us the best performance in order to achieve our goal(s). We can do so by actually testing our operation plans on the shop floor, or we may study their perfomance using models. A good designer will model his MS before implementing it: it is almost always cheaper to do so!

Therefore, we shall take a brief look at methods used to model manufacturing systems: in particular, simple mathematical models, and simulation.

Why are operational performance tests important to MS Design ?

- since they can give a good estimate of potential bottlenecks, and also be used to refine designs before implementation.

In the Human part, we will look at two important aspects:

What the are conditions under which the workers perform the best? How can this information be used to design the work pattern of workers, and working system of manufacturing systems (team oriented systems)?
Why is the customer an important part of the operations of the MS ? What implications does this participation of the customer have on the operation and organisation of the MS.