The Production 2000+ system consists of flexible manufacturing machines and a flexible transportation system, while the latter also puts constraints on the configuration of the machines.
The flexible machines are computerized numeric control (CNC) machines with three, four or five axes. The number of axes characterizes the processing capability of a (CNC) machine. A machine with three axes is capable to position a tool, such as a drill, anywhere on a plane (with the first two axes) and to move the tool forward, e.g. in order to drill a vertical hole (with the third axis). A fourth and fifth axes would then enable a machine to bend the tool either horizontally, vertically or in both directions in order e.g. to drill a slanted hole.
Obviously, CNC machines are capable of performing a wide range of processing steps and are thus quite flexible. Nevertheless, their flexibility is not unlimited. First of all, different processing steps may require different tools. For instance, the holes to be drilled into a cylinder head must be of different size. To drill holes of different size, a CNC machine must also use drills of different size. To change the drills during processing (of a single workpiece), CNC machines usually have a tool magazine and a tool changer. But even with a large magazine full of different tools, a CNC machine cannot perform a processing step if the tool required for this processing step is missing in the magazine. The bottom line is that the operator of a manufacturing system like Production 2000+ must configure the machines such that the right tools are in the right machine at the right time. But this is acceptable for a large-series manufacturing system which should be able to change the types of products processed from day to day (and not from second to second).
Secondly, a machine with three axes cannot perform certain operations for which four or five axis are required. The machines installed for the manufacturing system thus put constraints on the processing steps that can be performed. But as long as there is at least one five axes machine, a Production 2000+ is able to process any product since the flexible transportation system can move a workpiece from any machine to any other machine of the manufacturing system (see flexible transportation system). Of course, a single five axes machine may be a bottleneck in a large manufacturing system, so that even a flexible manufacturing system like Production 2000+ must be configured with care (and foresight). But again, this is acceptable for a large-series manufacturing system.
One of the main innovations of Production 2000+ is a flexible transportation system which is suitable for large-series manufacturing. In Production 2000+, the machines are organized along a transportation system which enables each workpiece to move from any machine to any other machine, while supporting a main (high-volume) flow of material. In large-series manufacturing, the vast majority of workpieces (approximately 80%) moves from the first machine straightforward to the last machine. This main flow usually has a high-volume and must be prioritized in order to ensure the high output of the manufacturing system. The rest of the material flow, however, must also reach its goal in order to achieve the flexibility and also the robustness of the manufacturing system. To achieve this flexibility and robustness, while maintaining the ability to support a main high-volume material flow, is the main innovation of the P2000+ transportation system.
A Production 2000+ transportation system is usually organized as follows. The transportation system consists of a forward and a backward conveyor which both bypass all machines (see figure 1). Between each two machines, there is a shifting table that allows workpieces to change the direction, i.e., to move from the forward to the backward or from the backward to the forward conveyor. A workpiece may thus move from anywhere in the transportation to anywhere else.
Additionally, a shifting table may put a workpiece onto a supply conveyor which leads directly into the machine. Any workpiece that is supposed to be processed by a machine must thus move to the shifting table in front of the machine and must be moved by this shifting table onto the supply conveyor of this machine (see figure 2). On this supply conveyor, the workpiece then waits to be processed by the machine. Once it has been processed, the workpiece is put onto the output conveyor of the machine which directly leads to the next shifting table. At this shifting table, the system then has to decide where to move the workpiece next.
The overall behavior of the transportation system is such that the system always tries to move a workpiece to its next goal machine on the shortest path currently possible. That is, a workpiece coming from a machine is put onto the forward or the backward conveyor depending on the direction of the next goal machine. The system then moves the workpiece to the shifting table in front of the next goal machine and tries to put it onto the supply conveyor of the goal machine. In doing so, however, the transportation system always avoids any jams (or delays) on the forward and backward conveyors.
If a workpiece reaches the shifting table in front of the goal machine and it is not possible to put it onto the supply conveyor, for instance, because the supply conveyor is fully occupied, then the shifting table does not wait for the supply conveyor to become free, but puts the workpiece either onto the forward or backward conveyor, depending on which one is free. The motivation for this is the following: A workpiece waiting in front of a shifting table on either of the forward or backward conveyor would block any other workpieces moving to other machines. This could interrupt the supply of the other machines and must therefore be avoided at all costs. There should never be any jams on forward or backward conveyors! If a workpiece cannot move to its goal machine, then it should "circle" around the machine (on forward and backward conveyors). This behavior of the transportation system neither interrupts the supply of the machine the workpiece wants to move to. If the supply conveyor is not free, then there are already several (usually two) workpieces in front of the machine waiting for being processed. Assuming processing times of at least 30 seconds, the machine has an immediate supply of 60 seconds. This is enough time for other workpieces to circle around the machine and to arrive again at the shifting table in front of the machine. If, in-between, the machine has processed a workpiece on its supply conveyor, then there will now be sufficient space for the workpiece to be moved to the supply conveyor.
In principal, the machines of a Production 2000+ system can be organized in any way, since the transportation system is able to move a workpiece from any machine to any other machine. However, there are two constraints on the configuration of the P2000+ system which must be fulfilled in order to achieve the performance as described in the simulation and the performance test.
First of all, the machines must be organized such that most workpieces move straightforward from the beginning to the end of the transportation system. This is necessary to ensure that the transportation system is able to handle a high volume of workpieces (cf. the beginning of the section on the transportation system). Ideally, workpieces have to "move back" only in case of disturbances. Note that the Production 2000+ transportation system is able to handle any kind of material flow. It is just that the throughput of the transportation system is suboptimal if workpieces move "randomly" between machines.
Secondly, to achieve robustness, there should always be for any possible operation an alternative machine which can perform the same operation. If there is an operation which can be performed only by a single machine, then this machine will be a bottleneck in case of a disturbance. If this machine breaks down, then any workpiece requiring this operation will have to queue in front of this machine. Furthermore, any machine processing workpieces moving next (or in the future) to the disturbed machine will also be slowed down because it can no longer get rid of the processed workpieces (for a discussion of the buffering mechanism see the section on the buffering mechanism). Eventually, the whole system stops because all workpieces are waiting for the disturbed machine. As a consequence, if robustness with respect to any machine is desired, there should always be an alternative machine for any operation. A simple approach to achieve this is to organized the machines such that always two adjacent machines provide the same operations. This was the approach of Production 2000+, in particular for the prototype.
Note that alternative machines do not imply redundancy or over-capacity. It is possible to organize alternative machines such that there is no redundancy. For instance, if each two adjacent machines have the same set of operations, all machines are fully utilized if no disturbances occur (which is the case for Production 2000+). Basically, such a production configuration consists of two lines of machines. In case of a single machine failure, however, the performance of the production system does not drop to 0%, but to 50%. That is the main reason why the performance of the Production 2000+ is much better than the performance of transfer lines (see the performance test; even though we do not have the space here to discuss why this is the case).
Note also that with CNC machines it is possible to configure machines such that at least two machines are alternative machines. If a manufacturing line consists of four machines, each performing 25% of the necessary operations, the operations can be redistributed over the four machines such that the first two machines perform the first 50%, and the other two machines perform the second 50% of the operations. In total, then, the reconfigured production system has the same amount of resources and produces the same throughput. There is no over-capacity, but a lot more robustness since there is an alternative machine for any operation. Of course, it is not that simple to reconfigure a manufacturing system with CNC machines because an operation may depend on more than three axes (and machines with more axes are more expensive). But with a sufficiently large set of operations, as for cylinder heads, this is possible and was demonstrated by Mercedes-Benz in the project Production 2000+.