« iepriekšējāTurpināt »
based on a stream of orders from Summarizing, the ultimate factory Toyota (Ref 4).
can make any quantityofanyitem withNippondenso's classifications are as
out any penalty for switching. The disfollows:
• FMS-2 - Design the product with a advantage of current automation sys
common outer shell and interchange tems is that they are too focussed on a • Flexibility for Product Variation - able interiors and provide robots small range of models of one product.
Configuration, size, model, and type and sensors as needed to make quick If demand for one version of alternaare levels of variability within the changes from one to another; sev- tor, say, rises while that for another product itself that are increasingly eral models of an air conditioner falls, the underloaded line cannot help difficult for the automation system are made this way, all being essen- the overloaded one. Instead, one must to accommodate.
tially the same size.
build more lines, resulting in overcapac
ity and wasted investment. Several • Flexibility to Design Change - Minor • FMS-3 - Design product and pro- generic approaches exist to this long
changes are often easy to accommo- cess so intimately that one can even standing problem: predict future date, model changes are harder, and change the outside shell's length demand perfectly, make superthe next generation of the product and diameter without affecting the intelligent manufacturing systems that usually requires a new factory.
automation system. The Type III can switch, or design the products and
alternator (see below) is an example their manufacturing processes to con• Flexibility to Production Volume
tain a measure of alterability. I call the Change - Total volume fluctuation Flexibility means not only the abil- latter "smart products" below. It is requires reassigning manufacturing ity to switch some important factors of probably the best of the three systems to different products; grad- the product but to switch rapidly. approaches, the first being obviously ually increasing volume normally Nippondenso has worked over the unavailable and the second beyond the requires buying more capacity; fluc- decades to cut the changeover time current state of the art except in restricted tuations among product types require from hours to minutes to seconds, while but very useful situations. Nippondenso reassigning production capability at the same time increasing the range has adopted the smart products between the types. of flexibility
approach and showed some interesting
The size of the product is one of the examples. In response to these needs, most important factors in the design of Nippondenso has utilized several strate- an automation system. Supporting a Classification of New Product gies, beginning in the 1960s (see later change in product size without Development Efforts Figure 1):
rebuilding the system is almost impos
sible. Yet as cars become smaller and This classification is as follows: FMS-0 - Use specialized automa- lighter, so must their components. Only tion with no flexibility and make the largest cars can take the largest 1. Innovative, totally new product (10% rigidly standardized products; an components; even here, however, the of design efforts); examples include example is little control relays. manufacturers are pressing for smaller active suspension or cathode ray Fukaya notes that strong efforts at components which perform more func- tube (CRT) dashboard displays. standardization occur even when tions. One can no longer simply reduce automation at levels 1, 2, and 3 below the capability of the product for use in 2. Strategic new product (called is adopted. Furthermore, FMS-O is a smaller car while keeping the outer Jikigata); these are major, marketthe preferred approach and is used shell the same. The shell must shrink, share-grabbing improvements of wherever possible.
too. As more varieties of cars are made, existing products such as radiators,
more sizes of the same product are alternators, and fuel pumps. • FMS-1 - Design the product with needed, each made in smaller produc
several versions of each part, capa- tion volumes than before. Lower pro- 3. Semi-new products; these are, in ble of being intermixed: 3 fronts, 4 duction volumes mean less efficient fact, minor improvements in permiddles, 3 backs, total 36 types; an automation unless some way can be formance of existing items; several example is a panel meter gage, in found to make all sizes on one automa- such improvements come along which many varieties of one basic tion system. Thus FMS-3 is a very diffi- between Jikigatas. model are made minute by minute cult but important level to achieve.
Number of Automated Lines
Figure 1. History of FMS in Nippondenso. The bulk of the visit focussed on the enlarges to about 20 members each, alternator the development time apparJikigata for a new alternator. Note, too, that Nippondenso is will- stators. This was done by coiling stator Radiators and alternators are clear ing to use the overlapping tasks method laminations stamped from long strips examples of “smart products,” being even on projects with lots of technical of steel (Figure 3) rather than stamp- designed so that the challenging manurisk. Overlapping brings the risk of more ing rings from steel sheets and stacking facturing strategy of conquering varichange, but Nippondenso and Nissan them up. (The amount of scrap mate- ety could be achieved without basic both feel that changes forced by out- rial is also drastically cut this way.) The advances in manufacturing knowledge. side pressures such as competitors' wire windings for the stator are formed Innovative manufacturing methods were actions are more severe. This fact slightly separately from the stator itself and indeed made, however. Deciding how counters Prof. Kimura's feeling that pushed radially outward into the grooves to partition the problem into product only “understood” processes and prod- in the stator rather than being wound innovation and process innovation ucts could be approached this way. in place in the stators. Changing the clearly required a single team working Fukaya was quite clear on this point, diameter of the windings is easier this together from the start of the project. and said that Concurrent Engineering way. Most of the size changes can be Success would have been unlikely if (CE) (joint operation of product and made almost without stopping the process engineers had merely critiqued process design teams with monthly manufacturing equipment.
and comes up with an action plan to ently was 5 years.
meet the specification. Once the plan Along with this elaborate planning The Jikigata Process
is approved it is condensed to a single process, Nippondenso has some "use
sheet of paper and given to everyone. ful tools.” These comprise the usual Jikigata efforts are directed at prod. These 40 to 50 engineers stay with the CAD/CAM/CAE software, plus value ucts which are mainstays of the com- product until the end of the project, engineering, group technology, and pany, feed a mass production require- later being joined by about 100 manu- variation reduction, plus Nippondenso's ment for a popular car, and face impor- facturing equipment designers. The most own design for assembly evaluation tant competition, thus requiring strong specialized one-third of the machines method, a variety of system engineerinnovation. On top of this, such prod. (by cost) are made in-house, while the ing aids like discrete event simulation ucts require timely and reliable deliv- more ordinary ones are built by con- and process failure modes and effects ery. These requirements have forced tractors.
analysis (FMEA), and quality managethe creation of new design staff organi- The plan (Figure 2) must be chal- ment methods (design reviews, quality zations and close involvement of top lenging but reasonable. It must contain control (QC) techniques, and the management. While CAD and CAE the total view and plenty of detail. It Taguchi method). Calling these “usehave played important supporting roles, involves top management, who attend ful tools” reveals Nippondenso's priorthe most important element of such monthly follow-up meetings. Each goalities: get the methodology in place first, developments is creation of new man- has a responsible person and a list of then support it with tools. ufacturing methods to support the risk-management actions. Each goal is All of the debates and tradeoffs "smart" flexible design. This has meant classified as to its importance to the involved in these efforts are carried out making production engineering an equal project and its level of risk. The impor- by experienced people. When there is a partner in the design process. tance levels range from “M” (for must major problem a top executive decides. Nippondenso, like many Japanese have) to “W1" (want very much) to Design is vulnerable to change, often companies, maintains production engi- "W2” (want, but not so much). Risk forced by the actions of a competitor. neering as a corporate level activity varies from “A” (feasible today) to “B” In alternators and air conditioners, with a director (equivalent to executive (currently being studied for applica- where Nippondenso dominates, comVP) as its head. Thus the company was tion to mass production) to“C” (under petitors' actions are less disruptive of long prepared for the required organi- basic examination, not out of the labo
basic examination, not out of the labo- the design schedule, but in brake syszational changes.
ratory yet). At each point in the sched- tems where Nippondenso does not It is important to realize that this is ule there is a “T” (target) date after dominate, the schedule is more vulnera more sophisticated activity than mere which, if a risky process or design ele- able. The availability of top manage"design for manufacture” (DFM) or ment has not been achieved, one of the ment and their willingness to take the “design for assembly” (DFA). A new prearranged alternates will automatic responsibility and make decisions level of automation/flexibility is being cally be substituted.
quickly are crucial. In this sense, sought, and it cannot be achieved unless On top of all this, Nippondenso Nippondenso is like Nissan and other new manufacturing methods are created, aimed at reducing the development time companies who organize to absorb methods which are enabled, not just from the customary 6 years to 4, by change during the design process rather eased, by the product design itself. overlapping product and process devel- than try to resist it.
A Jikigata effort combines corpo- opment activities. For the Type III rate production engineering and a product division's capabilities as shown in Table 1.
Table 1. A Jikigata Effort Product development begins after a launch decision by the New Product Corporate Production
Product Group Development Council, which appoints
Engineering a product development team (four to five engineers) and a process develop- System Section
Planning Center ment team (two to three engineers). Processing Section
Product Engineer These teams work together to create Materials Section
Manufacturing Dept the concept design specifications. Each Machinery & Tools Dept
Quality Assurance then splits into separate activities,
the product engineers' design and would follow-up by top management) was the Altogether 74 new manufacturing have been impossible if the process way to accomplish it. They all agree technologies were developed. This equipment had been merely purchased that it is based on human communica- project occurred in the early 1980s. from vendors after product design was tion and experience and wish for com- The resulting design comes in three complete. puterized versions of CE. I did not hear main sizes with capacities ranging from them suggest any ways to create them. 35 to 80 amps. Within each size there Use of Computers in the Nippondenso's production engineer- are about 250 variations.
Design Process ing people are also sympathetic to the These alternators are assembled on idea that computer aids will help this automated assembly lines that use mostly Nippondenso has a large CAD/ process and fervently wish for such help, specialized automation for the assem- CAM/CAE activity, combining their but they do not see it becoming avail- bly moves themselves plus robots to own software development and use of able soon and do not think it will be a feed the parts from trays to the assem- commercial software. The system they dominant feature of their success. Yet bly stations. A few simple fixture changes, have developed is similar to several they are developing several effective accomplished manually, support commercially available “frameworks” computer tools and see where others changes in product model. Nippondenso in the sense that it supports many might be introduced. See below for a built these lines in 1987 after seeing a application programs as long as they summary of these.
film in 1980 made in 1977 by our group respect certain data conversion proto
at Draper demonstrating complete robot cols, but there is no true common dataDevelopment of the assembly of Ford alternators.
base. In addition to this core system, Type III Alternator
The spirit of these innovations can there is the typical array of CAE plus a
also be seen in the way Nippondenso range of software that supports proThe main components of an alter- redesigned radiators a few years earlier duction preparation and production nator are the stator, the rotor, the two- (Ref 13). A major feature was machines control. piece cast outer case, and the rectifier that could switch sizes of components The goals of CAD/CAM/CAE are assembly. The goals of the redesign in a few seconds, plus a snap-together stated as were to produce an alternator that could assembly method that eliminated the be made in several lengths and diam- need for fixtures in the different sizes. • improving the efficiency of product eters on the same fabrication and assem- The cost of the fixtures was saved but development bly equipment. Important changes in more importantly the time required to the design of all four components were
switch from one size set of fixtures to • shortening the lead time for new required. Some were relatively easy, another was eliminated. Some of these products such as cutting different diameter techniques were pioneered by General grooves on different size cases. Redesign Motors (GM) Harrison Radiator Divi- • making it easier to design product to permit assembly from one direction sion but not put in place as completely
variants was also not too difficult to achieve. as at Nippondenso. Others required considerable innova
• helping create smaller and lighter tion, such as making different diameter