Classifying Flexibility

Nippondenso's classifications are as

follows:

• Flexibility for Product Variation Configuration, size, model, and type are levels of variability within the product itself that are increasingly difficult for the automation system to accommodate.

• Flexibility to Design Change - Minor changes are often easy to accommodate, model changes are harder, and the next generation of the product usually requires a new factory.

• Flexibility to Production Volume Change - Total volume fluctuation requires reassigning manufacturing systems to different products; gradually increasing volume normally requires buying more capacity; fluctuations among product types require reassigning production capability between the types.

In response to these needs, Nippondenso has utilized several strategies, beginning in the 1960s (see Figure 1):

FMS-0- Use specialized automation with no flexibility and make rigidly standardized products; an example is little control relays.

Fukaya notes that strong efforts at standardization occur even when automation at levels 1, 2, and 3 below is adopted. Furthermore, FMS-0 is the preferred approach and is used wherever possible.

• FMS-1 - Design the product with

several versions of each part, capable of being intermixed: 3 fronts, 4 middles, 3 backs, total 36 types; an example is a panel meter gage, in which many varieties of one basic model are made minute by minute

based on a stream of orders from Toyota (Ref 4).

FMS-2 - Design the product with a common outer shell and interchangeable interiors and provide robots and sensors as needed to make quick changes from one to another; several models of an air conditioner are made this way, all being essentially the same size.

• FMS-3 - Design product and process so intimately that one can even change the outside shell's length and diameter without affecting the automation system. The Type III alternator (see below) is an example.

Flexibility means not only the ability to switch some important factors of the product but to switch rapidly. Nippondenso has worked over the decades to cut the changeover time from hours to minutes to seconds, while at the same time increasing the range of flexibility.

The size of the product is one of the most important factors in the design of an automation system. Supporting a later change in product size without rebuilding the system is almost impossible. Yet as cars become smaller and lighter, so must their components. Only the largest cars can take the largest components; even here, however, the manufacturers are pressing for smaller components which perform more functions. One can no longer simply reduce the capability of the product for use in a smaller car while keeping the outer shell the same. The shell must shrink, too. As more varieties of cars are made, more sizes of the same product are needed, each made in smaller production volumes than before. Lower production volumes mean less efficient automation unless some way can be found to make all sizes on one automation system. Thus FMS-3 is a very difficult but important level to achieve.

Summarizing, the ultimate factory can make any quantity of any item without any penalty for switching. The disadvantage of current automation systems is that they are too focussed on a small range of models of one product. If demand for one version of alternator, say, rises while that for another falls, the underloaded line cannot help the overloaded one. Instead, one must build more lines, resulting in overcapacity and wasted investment. Several generic approaches exist to this longstanding problem: predict future demand perfectly, make superintelligent manufacturing systems that can switch, or design the products and their manufacturing processes to contain a measure of alterability. I call the latter "smart products" below. It is probably the best of the three approaches, the first being obviously unavailable and the second beyond the current state of the art except in restricted but very useful situations. Nippondenso has adopted the smart products approach and showed some interesting examples.

Classification of New Product Development Efforts

This classification is as follows:

1. Innovative, totally new product (10% of design efforts); examples include active suspension or cathode ray tube (CRT) dashboard displays.

2. Strategic new product (called Jikigata); these are major, marketshare-grabbing improvements of existing products such as radiators, alternators, and fuel pumps.

3. Semi-new products; these are, in

fact, minor improvements in performance of existing items; several such improvements come along between Jikigatas.

The Jikigata Process

Jikigata efforts are directed at products which are mainstays of the company, feed a mass production requirement for a popular car, and face important competition, thus requiring strong innovation. On top of this, such products require timely and reliable delivery. These requirements have forced the creation of new design staff organizations and close involvement of top management. While CAD and CAE have played important supporting roles, the most important element of such developments is creation of new manufacturing methods to support the "smart" flexible design. This has meant making production engineering an equal partner in the design process. Nippondenso, like many Japanese companies, maintains production engineering as a corporate level activity with a director (equivalent to executive VP) as its head. Thus the company was long prepared for the required organizational changes.

It is important to realize that this is a more sophisticated activity than mere “design for manufacture" (DFM) or "design for assembly” (DFA). A new level of automation/flexibility is being sought, and it cannot be achieved unless new manufacturing methods are created, methods which are enabled, not just eased, by the product design itself.

A Jikigata effort combines corporate production engineering and a product division's capabilities as shown in Table 1.

Product development begins after a launch decision by the New Product Development Council, which appoints a product development team (four to five engineers) and a process development team (two to three engineers). These teams work together to create the concept design specifications. Each then splits into separate activities,

The plan (Figure 2) must be challenging but reasonable. It must contain the total view and plenty of detail. It involves top management, who attend monthly follow-up meetings. Each goal has a responsible person and a list of risk-management actions. Each goal is classified as to its importance to the project and its level of risk. The importance levels range from "M" (for must have) to "W1" (want very much) to "W2" (want, but not so much). Risk varies from "A" (feasible today) to "B" (currently being studied for application to mass production) to “C” (under basic examination, not out of the laboratory yet). At each point in the schedule there is a "T" (target) date after which, if a risky process or design element has not been achieved, one of the prearranged alternates will automatically be substituted.

On top of all this, Nippondenso aimed at reducing the development time from the customary 6 years to 4, by overlapping product and process development activities. For the Type III

alternator the development time apparently was 5 years.

Along with this elaborate planning process, Nippondenso has some "useful tools." These comprise the usual CAD/CAM/CAE software, plus value engineering, group technology, and variation reduction, plus Nippondenso's own design for assembly evaluation method, a variety of system engineering aids like discrete event simulation and process failure modes and effects analysis (FMEA), and quality management methods (design reviews, quality control (QC) techniques, and the Taguchi method). Calling these “useful tools" reveals Nippondenso's priorities: get the methodology in place first, then support it with tools.

All of the debates and tradeoffs involved in these efforts are carried out by experienced people. When there is a major problem a top executive decides. Design is vulnerable to change, often forced by the actions of a competitor. In alternators and air conditioners, where Nippondenso dominates, competitors' actions are less disruptive of the design schedule, but in brake systems where Nippondenso does not dominate, the schedule is more vulnerable. The availability of top management and their willingness to take the responsibility and make decisions quickly are crucial. In this sense, Nippondenso is like Nissan and other companies who organize to absorb change during the design process rather than try to resist it.

Note, too, that Nippondenso is willing to use the overlapping tasks method even on projects with lots of technical risk. Overlapping brings the risk of more change, but Nippondenso and Nissan both feel that changes forced by outside pressures such as competitors' actions are more severe. This fact slightly counters Prof. Kimura's feeling that only "understood" processes and products could be approached this way. Fukaya was quite clear on this point, and said that Concurrent Engineering (CE) (joint operation of product and process design teams with monthly follow-up by top management) was the way to accomplish it. They all agree that it is based on human communication and experience and wish for computerized versions of CE. I did not hear them suggest any ways to create them. Nippondenso's production engineering people are also sympathetic to the idea that computer aids will help this process and fervently wish for such help, but they do not see it becoming available soon and do not think it will be a dominant feature of their success. Yet they are developing several effective computer tools and see where others might be introduced. See below for a summary of these.

Development of the Type III Alternator

The main components of an alternator are the stator, the rotor, the twopiece cast outer case, and the rectifier assembly. The goals of the redesign were to produce an alternator that could be made in several lengths and diameters on the same fabrication and assembly equipment. Important changes in the design of all four components were required. Some were relatively easy, such as cutting different diameter grooves on different size cases. Redesign to permit assembly from one direction was also not too difficult to achieve. Others required considerable innovation, such as making different diameter

stators. This was done by coiling stator laminations stamped from long strips of steel (Figure 3) rather than stamping rings from steel sheets and stacking them up. (The amount of scrap material is also drastically cut this way.) The wire windings for the stator are formed separately from the stator itself and pushed radially outward into the grooves in the stator rather than being wound in the stator rather than being wound in place in the stators. Changing the diameter of the windings is easier this diameter of the windings is easier this way. Most of the size changes can be made almost without stopping the manufacturing equipment.

Altogether 74 new manufacturing technologies were developed. This project occurred in the early 1980s.

The resulting design comes in three main sizes with capacities ranging from 35 to 80 amps. Within each size there are about 250 variations.

These alternators are assembled on automated assembly lines that use mostly specialized automation for the assembly moves themselves plus robots to feed the parts from trays to the assembly stations. A few simple fixture changes, accomplished manually, support changes in product model. Nippondenso built these lines in 1987 after seeing a film in 1980 made in 1977 by our group at Draper demonstrating complete robot assembly of Ford alternators.

The spirit of these innovations can also be seen in the way Nippondenso redesigned radiators a few years earlier redesigned radiators a few years earlier (Ref 13). A major feature was machines that could switch sizes of components in a few seconds, plus a snap-together assembly method that eliminated the need for fixtures in the different sizes. The cost of the fixtures was saved but more importantly the time required to switch from one size set of fixtures to another was eliminated. Some of these techniques were pioneered by General Motors (GM) Harrison Radiator Division but not put in place as completely as at Nippondenso.

Radiators and alternators are clear examples of "smart products," being designed so that the challenging manufacturing strategy of conquering variety could be achieved without basic advances in manufacturing knowledge. Innovative manufacturing methods were indeed made, however. Deciding how to partition the problem into product innovation and process innovation clearly required a single team working together from the start of the project. Success would have been unlikely if process engineers had merely critiqued the product engineers' design and would have been impossible if the process equipment had been merely purchased from vendors after product design was complete.

Use of Computers in the Design Process

Nippondenso has a large CAD/ CAM/CAE activity, combining their own software development and use of commercial software. The system they have developed is similar to several commercially available "frameworks" in the sense that it supports many application programs as long as they respect certain data conversion protocols, but there is no true common database. In addition to this core system, there is the typical array of CAE plus a range of software that supports production preparation and production control.

The goals of CAD/CAM/CAE are stated as

⚫ improving the efficiency of product development

shortening the lead time for new products

⚫ making it easier to design product variants

helping create smaller and lighter. products