Printer Friendly


Have you hired or fired an industrial consultant during the last six months? Has your company reorganized twice during the last three years? Do you refer to CAD/CAM as CAD/can't? Do you employ equal opportunity robots? Have you built a special waiting room for all those factory-of-the-future salesmen sure to descend on your company? If you can answer yes to any two of these questions you probably are aware (maybe painfully) that a rapidly unfolding revolution in manufacturing technology is, or soon will be, affecting your firm's way of doing business.

This revolution's mainspring is computer-integrated manufacturing (CIM)--a cornerstone of the factory of the future. Edward J Adlard, supervisor of manufacturing software systems at Metcut Research puts it this way, "The future factory is evolving by linking and sharing data between currently available systems such as computer-aided design (CAD), computer-aided manufacturing (CAM), computer-aided process planning (CAPP), distributed numerical-control (DNC) systems, flexible-manufacturing systems (FMS), etc, which is exactly what CIM is about--communications and methods that tie together existing technologies."

During the next 50 years, CIM will become the rule, not the exception in America's factories. Distributed computerized networking supported by feedback will close the design and manufacturing loop, and provide self-correcting processes. Then all of the acronym-tagged technologies will surrender their separate identities and merge into one awesome product development Hydra.

"The payoff will come on the shop floor," insists Adlard. "It's estimated that CIM has the potential to increase machine utilization, and reduce work-in-process inventory equal to the cost of building and equipping another complete plant." Don't expect future factory in a drum

An off-the-shelf fully integrated factory isn't available today and probably never will be. It will be up to you to acquire, implement, and integrate the necessary pieces of hardware, software, and equipment.

"After visiting hundreds of plants, and examining the product lines of all major factory-automation suppliers, one fact becomes clear: Unlike the office, which is a relatively uniform environment across all industries, there is no generic factory," says Glen Allmendinger, director of factory systems research at the Yankee Group, a technology research firm. "Each systems proposal, therefore, must be customized, making the potential for developing an integrated solution much more complex.

"This complexity makes an 'off-the-shelf' solution unrealistic," he continues. "Few integration specialists have a track record for installing factory systems, and when this is broken down by specific industry and application, the numbers are even smaller (e.g., a group with experience in an automobile factory may not be of much help to an aerospace or electronics plant)."

Allmendinger suggests there are several areas that must be addressed if you are considering fully integrating your factory's automation and information systems.

* Data communications are needed for distributing design, engineering, and production data over a large network of computers and plant-floor workstations. And there must be capability to communicate outside the manufacturing computer environment, providing users with a method of linking different functional business areas together.

* Manufacturing systems and applications software are needed as a foundation for a fully integrated system and an interactive manufacturing environment.

* A manufacturing data base is required for eacy storage, retrieval, manipulation, and management of large quantities of engineering and production data.

"CIM's ultimate benefit for the future factory is establishing an integrated manufacturing data base within the company," comments Metcut's Adlard. "This data base will consist of parts and tooling geometry, machine-tool data, workpiece-material data, part/operation data, and machinability data. In the long run, the data base will facilitate collection, storage, flow, and revision of all the information needed to plan, organize, and control manufacturing. This will include not only the raw manufacturing data, but the economic data necessary for job tracking and cost analysis."

Today's leading-edge CAD/CAM systems are already demonstrating their value in creating, maintaining, and transferring such information. Built around 32-bit computers and specialized graphics software, these systems are creating large multipurpose data bases.

During the next 50 years, information regarding every phase of product development and engineering will be resident in data bases, which will routinely issue commands to robots and various automated processes on the factory floor. Local area networking

By 2034, most factories will contain an array of computer-based subsystems such as CAD, CNC, robots, material-handling systems, etc. To fully integrate these subsystems and experience phenomenal advances in efficiency, flexibility, and productivity, these islands of automation must be linked via a communications medium. On the shop floor, such a communications system, called a local area network (LAN), will allow automation to react quickly to sudden changes in process conditions (e.g., overload or device breakdown).

Until recently, most LANs were designed specifically for office environments--Xerox's Ethernet, Datapoint's ARCNET and WANGNET, are examples. Now, though, a number of factory-oriented networks have been announced: GEnet, Gould Modicon's MODWAY, Concord Data System's DataNet, and the Interactive Systems/3M * Western Digital * Allen-Bradley communications network, to name just a few.

These networks are not like those targeted for the office. First, the application environments are completely different. Unlike offices, the manufacturing work place is often harsh. This, as well as electromagnetic interference, tends to degrade data communications.

"These needs make factory communication users much more performance conscious than office users (the office LAN market is extremely cost conscious)," comments the Yankee Group's Allmendinger. "In factories, managers planning to install a network will have to forego many cost considerations, and pay a premium for reliability, data integrity, and guaranteed access time, i.e., maximum time to deliver a message on the network. This need is the best argument for token-based (deterministic) systems over CSMA/CD systems, by the way, which have been criticized for problems generated by propagation delay and random access algorithms." Quest for standards

As with most emerging technologies, the question of LAN standards arises. All of the competing vendors have developed a proprietary communications system. But, actual acceptance of any standard won't happen for at least five years.

Traditionally, a standard develops in one of three ways: one supplier dominates the market and sets a de facto standard (e.g., IBM's SNA), the government imposes a standard (e.g., via the National Bureau of Standards), or a standard is formed by committee (e.g., IEEE 802). Allmendinger believes that a factory communications standard will shape itself differently. For example, as a large user of automation (and now obliquely as a robot vendor through GMF) General Motors has created its manufacturing automation protocol (MAP) task force, which sets corporate standard protocols for local communication among plant floor systems.

A de facto standard, therefore, may be set by a dominant automation user. "This is why IBM is interested in GM's MAP--they realize that their presence in management information systems holds no weight among manufacturing veterans," emphasis Allmendinger. "Thus, Big Blue has been willing to bend much more than they ever would have considered in their traditional markets.

"Large users of automation will probably follow GM's lead and resolve the communications issue themselves. We are currently performing a study, polling the largest automation users, to answer this question.

"When standardization finally is accomplished," Allmendinger continues, "the entire production process will be coordinated and integrated from product design, through shop floor applications for fabrication, and on into applications for assembly."

He points out, "ANSI is attempting to coordinate the various standards activity in industrial automation, and is presently identifying all existing standards and projects that relate to forming standards in this area. This effort is a first glimmer of hope; however, if standards are to become effective in time to prevent further sliding of manufacturing productivity, automation users will have to take the bull by the horns (like GM) and force suppliers to create an atmosphere of flexibility.

"An interest in software standards is also growing. Ability to provide a natural interface for a variety of users is vital. IBM's introduction of AML, with its hooks to APT (combined with the firm's introduction of the XT-370 PC workstation), illustrates its awareness of the forces moving in standardization. This gives AML a head start in becoming the next industry standard for machine language. The election of Dr Samuel Korin from IBM to the chairmanship of the RIA Standards Committee ensures that IBM will play a major role in this area."

Computers, communications systems, and data-base management are the backbone of the factory of the future. Many industrial firms feel they have only five years to fully integrate their entire production process. If they miss the boat they quite probably will lose a strategic competitive advantage in their industry.

Flexible manufacturing is a loosely defined concept organized around the goal of CIM. Yet, Dr Suren B Rao, director of product development, National Broach & Machine Div, Lear Siegler Inc, feels "The only way the US will regain its lead in manufacturing is through implementation of the integrated flexible factory of the future. The degree of implementation, of course, depends on if we have a national commitment. And tremendous gains are to be made just by full implementation of available computer technology."

The notion of an FMS (flexible manufacturing system) is the final stage in the minimization of human intervention in machine-tool operations and is certain to be the dominant discrete-parts manufacturing strategy during the next 50 years. The ultimate FMS includes tiers of DNCs (distributed numerical control) and subordinate CNCs (computer numerical control), plus associated robot, material-handling, and packaging systems--all controlled as subsystems of a master factory computer. The key, though, is NC (numerical control). A machine-tool evolution

In a just-released study from Predicasts, the Cleveland-based business-intelligence firm, Senior Research Specialist Dr john Breckling forecasts that by 1995 NC will be specified on nearly 40 percent of the newly purchased machine tools. Development of CNC has paved the way for widespread realization of the benefits of NC.

"The NC principle is applied at a still higher level in DNC by linking a network of CNC machine tools with a master computer," says Dr Breckling. "Just as each CNC controls the functions of its machine, the master computer integrates performance of the entire network. The status of each subsystem is communicated to the master, or host computer. It then makes adaptive responses to problems throughout the network in terms of desired outcomes, not only for individual metalworking functions, but for the entire process.

"This approach reduces machine interdependence in integrated systems," he continues, "because the master computer can compensate for individual machine failure by redistributing the work flow. DNC in conjunction with the multiple-station transfer concept lays the groundwork for factory-wide integration--enter the FMS." Future features

What can we expect? Advances in adaptive control and in-process gaging will continue to expand use of untended FMS on second and third shifts. This ability saves not only the cost of labor but the cost of light and heat for those shifts as well. Self-diagnostic systems that stop production when failures occur are essential for such operations, however. Self-repairing functions are even more preferable, although this capability isn't anticipated before the end of the decade.

"Some of the features of an FMS will find their way into other manufacturing processes to improve efficiency and flexibility," remarks Cliffoed D Young, an FMS specialist at the consulting firm of Arthur D Little. "Transfer lines, for example, will soon start evolving towards an FMS by incorporating tool-changing devices, parts-handling devices, and advanced computer controls.

"Software is also a vitally important element. There are two areas where improvements are starting to appear. First, simulation techniques are becoming accepted as part of the systems design and specification procedure. Second, some FMS suppliers, notably Comau and White-Sundstrand, are actively pursuing development of either modular or standard software packages." Who has them? Who needs them?

FMSs are for batch manufacturing where workpiece variety is too high and production volume too low for transfer lines and other dedicated processes; at the same time, however, volume is too high for stand-alone machine tools. Even though this represents about 35 percent of the US manufacturing base today, FMSs have only just begun to be implemented.

Young reports, "Caterpillar Tractor has been the largest purchaser of FMSs, spending more than $135 million to date. Also in the first tier are AVCO, John Deere, and GE. A leading supplier, Kearney & Trecker, has stated there are 1000 potential customers in this country alone. If each one bought only four to five systems that would be almost 50 times the most optimistic view of the installed base.

"So what's the holdup? Long lead times and a soft world economy have more to do with the scarcity of installations than lack of interest. Nevertheless, results have been dramatic where the technology is being used and as these experiences become more widely known usage will grow."

Richard P Cottrell, VP, Machine Tool Group, Newcor Inc, has a different view on why industry has been slow in implementing the technology. "I think part of the reason is that standard machine-tool builders were the ones who first developed the concept and built all the early systems. Success in the standard-machine business depends on volume production of a few basic designs, so early FMSs were built by companies seeking mainly to increase markets for existing machines.

"Who were the customers? Companies producing big parts with numerous operations in low to medium volume--mostly construction and farm-equipment producers. Why did they buy? Because the parts they produced were well suited to standard machining centers and because they were familiar with NC, having been among the first to adopt it.

"But," he continues, "standard machining centers are an inefficient way to produce high-volume components because only one spindle is cutting at a time. Further, high-volume manufacturers haven't had much experience with NC. In fact, many are just taking delivery on their first CNC machines. Thus, existing FMSs are not well suited to the needs of high-volume manufacturers and they have, correctly I think, avoided them."

Cottrell cites an example of a major automobile manufacturer that recently requested proposals for an fMS to build a variety of power-train components at production rates ranging from 15 to 20 parts/hr. Several machine-tool builders submitted quotes on the system. Then the automaker threw them a curve by adding a new part, increasing overall production requirements to 100 parts/hr. The builders reacted by simply adding more machining centers, which blew the cost of the project into orbit.

He suggests there's a better approach. "The concept of an FMS needs a healthy dose of special-machine thinking--I mean the ability to process a part from scratch, and develop innovative machine and tool solutions that minimize the equipment required. Transfer-machine people are used to combining many operations on a single machine, thus reducing the number needed. When this philosophy is applied to FMSs, it can also reduce time spent moving parts from one machine to another and changing tools. That's what's necessary to make an FMS profitable for high-volume producers.

"We recently developed a concept for an FMS to produce five different parts in a family of one- and two-barrel carburetor bodies," he notes as an example of his point. "The system uses 28 machining centers designed on the modular building-block principle, thus delivering only necessary capabilities. Moreover, many machines were simply eliminated by innovative thinking. An example is indexing a fixture horizontally as well as vertically. This permitted performing a reaming operation vertically, which eliminated tool sag, and also horizontally rough drilling from both sides, which eliminated an entire workstation.

"Special-machine thinking will greatly increase the application range for FMSs and also increase productivity and/or reduce costs in future applications," says Cottrell. "But there are two other ractices that need to change if such special FMSs are really going to take off. First, most potential purchasers follow traditional machine-tool buying practice in expecting concept work to be done on a speculative basis. This is fine for traditional FMSs where little creative thinking is involved, but stifling when applied to newer FMSs where system conception is a considerable cost (although still only about 1 percent of the total cost). Second, we must break the current capital-dudgeting mind-set. Sure, an FMS costs more, but years down the road you'll save a bundle when engineering changes are made. Five to ten years from now you'll save even more when a new product is introduced."

Given the advantages of systems that can maximize machine-operating time, minimize failure, and reduce labor costs--and the further advantages of systems integrating planning, production, and accounting functions--the only barrier in principle to the growth of FMSs is this mind-set about high initial costs, which actually may become no problem in the near future.

"We are used to seeing figures like $15 to $20 million (and up) per system, but that is changing," states A D Little's Young. "Consider, for example, the FMSs installed by both Citroen and Volvo in Europe for between $4 to $6 million each. If significant manufacturing benefits can be demonstrated from these cheaper systems then FMSs will be opened up for many more companies in the future."

Predicasts' Dr Breckling adds, "Revolutionary growth is predicted by most industry analysts, and some expect that within the next decade half the CNC machine tools produced will be incorporated in FMS installations. There is no doubt that flexible-manufacturing technology will significantly impact the metalworking industries for the balance of this century." Avoiding future flops

Flexible manufacturing systems are a hot topic. Their potential for reducing inventories, simplifying changeover, and increasing operating efficiencies are keys to the US regaining its position as the world's premiere industrial nation.

Unfortunately, there is widespread lack of knowledge about what to expect from an FMS and, more importantly, what commitments must be made to ensure success. We asked George Rathke, VP at Ex-Cell-O Manufacturing Systems Co, for some generic guidelines that the inexperienced and misinformed should follow when assessing the potential for an FMS. He told us the following:

* Do your homework. Analyze your product mix, volumes, and product stability. If you're producing one-of-a-kind products or vast quantities of the same product, FMSs may not be for you. A builder can help with the decision, but the basic evaluation must be performed by your own personnel. Also, the act of getting your product "house" in order is a chance to sort out marginal products, while arranging what's left into groups or families. This is a critical step in the study, by the way, and is essential in making an intelligent decision.

* Make an up-front commitment. Hire or train technically competent managers early during the feasibility study. They should be conversant with the elements and capabilities of FMSs, especially the computer and software areas. Such people can form an evaluation/justification team and eventually become your system managers, thereby speeding up implementation and training other employees during start-up.

* Evaluate the builder. Seek out and work with companies that have the financial and technical strengths to see you through the project launch and beyond. Look for turnkey capability both in machinery and the all-important software. Nothing is more disillusioning than playing referee between suppliers when none will accept responsibility for the overall system performance.

* Be prepared. Once the commitment to buy an FMS is made, a user must be prepared to hold up his end of the ensuing partnership. You must provide adequate technical and managerial staff to ensure long-term success. The systems are relatively complex, which means well-trained software, electronic, and mechanical maintenance people must be readily available. A high degree of discipline in tool setup and distribution must be established, and the people responsible for those activities must be backed by a dedicated and progressive management team.

* Start small. Flexible manufacturing will probably be new to your company, so start with a cell comprised of one or two machines, an automatic material-handling device (such as a robot), and perhaps a gaging machine. Make sure that the controls and software in the cell can be integrated into a more sophisticated system at a later date. As you acquire confidence and technical capability, your ability to evaluate and specify future systems will become more professional and the likelihood of a flop will rapidly diminish.

* Be patient. No matter how well thought out and designed an FMS is, its complexity demands start-up time and a learning curve that is much longer than stand-alone machines. Nevertheless, the increased flexibility, higher productivity, and reduced inventory that it offers will more than offset the initial inconvenience.

* Keep it simple. If this is a first experience with a computerized machining system, robots, or electronic inspection, don't load up with unnecessary sophistication. Adaptive controls, machine vision, laser inspection, automated storage/retrieval, etc, are all powerful tools, but they require expertise in programming, maintenance, and operator knowledge to run effectively. A well-planned system will allow these devices to be interfaced at a later date--when you're ready.

* Get involved. The success of an FMS, now or in the future, is a company-wide concern that greatly depends upon visible promotion and support from top management. Nothing guarantees failure faster than an apathetic boss. Make sure that technical managers are involved with the builder, and play an active role in system specifications and design. A progressive builder invites this sort of interchange. It's a way to avoid delivery day surprises, and avoid and unwanted early retirement. Industrial robots: Automated workaholics

Other than the computer, the industrial robot promises the greatest impact on manufacturing during the next 50 years. In fact, increasing material costs, greater foreign and domestic competition, and a scarcity of skilled workers is putting a premium on productivity and cost efficiency, making automation via robotics a must for firms expecting even to survive to the end of this century.

Tech Tran Corp, an indudstrial-market research firm, estimates annual unit US sales of the devices will increase from about 1700 in 1982 to nearly 38,000 per year in 1992, resulting in a total installed base of 134,000 robots by the end of that year. If this number of installations is achieved, by 1992 robots should be used in about 5 to 10 percent of all manufacturing applications for which they are suited. Technology trends

Tech Tran further anticipates robot prices falling to about half of their current levels by 1992 (in 1983 dollars). Such reductions are likely to be accompanied by the following technical developments, thereby offering potential users a greater bang for their capital-equipment buck.

* Smaller and lower weight robots--Lighter materials and more streamlined designs will allow development of much smaller robots. Platics and composites increasingly will be used.

* Stronger robots--As robots become smaller, their relative payload capacity to weight ratio will increase. New materials and designs will permit greater carrying capacities.

* Flexible grippers--It's likely that current efforts to develop a general-purpose flexible end effector will result in commercially available end-of-arm tooling that can be used for a variety of applications.

* Vision sensing--Machine vision is expected to be available in 25 to 50 percent of all robots by 1992.

* Tactile sensing--An effective high-resolution compliant array-type tactile sensor will be available by then as well. Tactile sensing will probably play a more important role than machine vision during the next five years.

* Improved control--Robot-control technology will advance in several areas during the next 10 years. For example, hierarchical control will be standard; electronic control also will be widely used; microcomputers will be placed in about 50 percent of all robots, with programmable controllers used in the simple ones; adaptive control will be available in about one third of all robots by 1992; and multiprocessing and control of multiple robots will become common.

* Programming improvements--A standard robot programming language, probably derived from a graphics language, will be in use by the early '90s. Other expected developments include: Off-line programming in up to 25 ercent of all robots (initial applications should appear between 1986 and 1987); programming with self-diagnostic capabilities; and more user-oriented programming.

* Voice actuation--It's likely that by 1992 robots will recognize simple voice commands and act on them. This would be extremely useful in inspection operations.

* Artificial intelligence (AI)--Elementary forms of AI are already available in vision systems. By 1992, advanced decision-making and feedback capabilities will be available that, while not approaching human judgment, will allow more flexibility in applications such as assembly.

Eric Mittelstadt, president of GMF Robotics Corp, agrees with these predictions and adds, "Technological trends for vendors during the next 50 years will be to diversify into areas like software and AI. Obviously, as programming, machine vision, and other sensory software develop and become less expensive, robot application growth will result. Expect robots to move faster, more accurately, and with improved repeatability. They will respond with real-time motions to complex inputs from their sensor systems, which will enable working with randomly oriented parts. And robots will gain greater mobility, making the application possibilities almost limitless.

"Instead of being restricted to a track, as they are today," he continues, "robots will randomly move around in the manufacturing setting, responding to inputs from their sensing systems. Mobility is a mere extension of technologies that already exist in rudimentary form. Its widespread use, though, depends on when it becomes cost effective. This capability has tremendous potential, especially for assembly applications."

Timothy J Bublick, manager of automation engineering at the DeVilbiss Co, believes during the next 20 years, "Robots used for finishing applications will be improved in at least two major areas. First, there will be continued development of a low-cost robot of sufficient capability, reliability and programming ease for less sophisticated finishing operations. Plus, these robots will have the flexibility to be upgraded if desired.

"The second area deals with the high-tech paint shop. With availability of CAD/CAM programming, a finishing robot's path, as well as all finishing parameters, will be constantly monitored and adjusted. The parameters might include temperature, humidity, viscosity, and flow. For example, if there is a variation in temperature, the system will automatically change paint viscosity and flow for a consistent finish." As many as 30 percent of all robots are expected to be interfaced with CAD/CAM systems by 1992.

Bublick finishes by saying, "Units, with one or two arms capable of painting, will be equipped with vision sensors to instantaneously modify the programmed path to compensate for improper part orientation, for incorrect hanging of parts, and to detect applied film thickness. Controllers will be equipped with diagnostics to detect malfunctions--even head them off before they occur. It's also possible to envision robots painting in body shops or performing outdoor paint-contracting work." Application shift

In general, expect a trend toward a broader base of robot applications. In the near future, leading applications will be materials handling, machine loading, and assembly. (Note: Spot welding, which is now an established application, will increase in number, but decrease as a percentage of total robots in use--refer to the graph.)

"Machine loading/unloading and material handling will continue to dominate industrial robot applications over the next five years," says GMF's Mittelstadt, "however, the proliferation of assembly installations will mark the next full decade."

To accommodate the increasing use of robots in assembly applications, Tech Tran foresees product designers being required to redesign products for robot handling. For example, parts will need to be made more symmetrical for easy grasping. And protruding pieces of material may be added to some parts to provide a point of reference.

In addition, the increasing use of vision systems will require critical dimensions be designed around easily recognizable features. Parts also will need to be organized on assembly lines in an orderly manner. Finally, work materials will need to be able to resist damage by end effectors. Machines that see

Machine vision is a new term for the science of image analysis--a technology dating back at least 50 years, yet its most significant impact will be on manufacturing processes of the future. According to Dr Stanley R Sternberg, president of Machine Vision International, and adjunct associate professor of electrical and computer engineering at the University of Michigan, "Much of the history of image analysis has been in the materials sciences, i.e., looking at metallic microstructure or composite fibers. It wasn't used for looking at discrete parts or painted surfaces. The reason? Image analysis was time consuming and expensive."

"The older systems required what I call a clean image," comments George Gagliardi at Arthur D Little, "because the machines would get confused by stry light and shadows. The newest generation of vision systems, however, are much more tolerant of 'noisy' scenes typical of a manufacturing environment.

"The reason for the improvement concerns how picture elemtns are processed," he continues. "Up until recently most image-processing systems were designed to work on one pixel at a time. Even when using relatively fast microprocessors this was extremely slow. The newest vision systems chain many microprocessors together so they work on successive picture elements. This greatly speeds up processing. Some of the systems are even running many of these chains in parallel to further accelerate image analysis."

Dr Sternberg adds that, as with most other electronic devices, advances in the technology are accompanied by decreasing costs. "Now the cost of image analysis is low enough so individual items can be examine. And today, of course, requirements for examining items has exploded because of government regulations, more complex assemblies, and tighter tolerances. Don't expect this to change in the future." Foreseeing applications

Machine-vision systems are being considered as sensing devices for all forms of flexible manufacturing hardware, such as robots, and for 100 percent inspection of parts produced in large volumes. Inspection of smaller batches is another application area if the workpieces are extremely complex, such as those produced in the auto or aerospace industries.

In a recent industry analysis, Tech Tran reported, "As the number of vision systems increase, there will be a diffusion in the types of systems and applications in which they are used. There will be more use of the technology for robot control, with the percentage of systems used for this purpose increasing from about 4 percent in 1982 to about 22 to 30 percent in 1992. Vision systems used for inspection applications, while increasing substantially in number, should decline from about two thirds of the systems used today to about 40 percent by 1992."

A D Little's Gagliardi says, "Probably within the next five years, vision systems will be used for in-process gaging (both 2-D and 3-D), tool guidance, and process control--sort of a closed-loop approach with vision acting as the feedback. Also, thermal analysis (using infrared) of tool bits while they are buried in the cut is not unreasonable to expect. And machine vision will probably be used to reduce the amount of fixturing required for machining, especially for parts scheduled through an FMS."

Dr Sternberg adds, "Image analysis technology definitely will be put into control systems. During the next 15 to 20 years the technology will be applied to gathering data, then using it to control end effectors, e.g., devices such as a welding gun attached to the arm of the robot.

"Until now," he points out, "image analysis has been passive. In the future, vision systems will be watching the welding gun, the workpiece, the weld bead, and the robot--and controlling the entire process. Today's programmable automation is already moving to a closed-loop machine-vision system."

Looking forward from 20 to 50 years, Dr Sternberg expects closed-loop systems to incorporate sensors and become a new generation of adaptive controls. These will be the basis of all intelligent manufacturing systems by 2034. "Over the next 20 years," he notes, "software will become a larger and larger part of the development effort. Eventually a point will be reached where software complexity is so great that you simply couldn't add anymore to achieve increased function. In fact, modifications could degrade performance by adding errors. Humans couldn't test these systems because they will go through so many paths, branches, and loops that it would be impossible to analyze what they are doing.

"This generation of adaptive controls won't require us to understand the inner workings of the software or even the system. The only thing to monitor will be performance. Using image analysis, this generation of adaptive controls will learn by watching, analyzing, and being part of an application. Then the system will generate its own software to optimally perform the task."

Dr Sternberg concludes, "Vision will be the basis for achieving powerful closed-loop adaptive control in the factory of the future. It's the only thing that can collect information on the entire work environment. Application of this technology will be most prevalent in the auto industry, specially in FMSs and fixtureless assembly of products like car bodies." Tools and tooling issues

"In the cutting-tool industry, we're using computer technology to put things together that we've done manually for years," says Gene Sanders at General Electric's Carboloy Systems Dept. "One example is placing application information into a program that will eventually allow users to call up a proper cutting speed or material recommendation. This will help compensate for the expected storage of tooling engineers in the factories of the future, and help minimize service costs." Users will potentially be able to call a distributor, enter certain information and then obtain recommendations that otherwise would require a service call by an experienced technician.

Future activities in the cutting-tool industry, naturally, will be most impacted by developments in work materials. Thirty years ago there was a saying in the specialty-tool-materials business, which holds true now and probably will be true in the future. The tool materials of today will become the structural materials of tomorrow. One example is the superalloys, which are outgrowths of high-speed steel, developed to improve high-temperature creep resistance.

Sanders notes, "Recently, in the oil patch there has ben a shift to high-temperature alloys from carbon steels by manufacturers of tool joints and other down-hole drilling components. This is because of the growing activity in 'sour gas' drilling, which tends to corrode conventional materials. It's safe to assume that as more of these types of work materials are used, there will be increasing use of cutting-tool materials such as ceramics, diamond, and cubic boron nitride (CBN)." Tenacious tools

GE was the first company to announce success in reproducibly creating diamond and CBN in the laboratory and the first to manufacture these superabrasives on a commercial scale. We asked E Louis Kapernaros, general manager of GE's Specialty Materials Dept, if he thought any materials superior to diamond and CBN will be developed during the next 50 years? "I doubt it," he says. "When we first manufactured diamond at the GE Research Laboratory back in 1954, a newly developed high-pressure, high-temperature technology was used to create other materials that could, in theory, have properties superior to those of diamond. No useful material other than CBN was found. Because CBN is more resistant than diamond to chemical attack by some metals at high temperatures, it's even more super than diamond for some applications."

Both of these tool materials, though, promise tremendous potential for improving productivity, particularly where work materials are difficult to grind or difficult to cut. And that will be true of many of the materials used in 2034.

"Some of today's traditional engineering materials will be long gone by the end of the century," emphasizes Kapernaros. "The auto industry will be manufacturing high-performance gasturbine and internal-combustion engines made mostly of advanced ceramics--silicon carbide, silicon nitride, and others. These materials are extremely hard and abrasive. They can't be effectively ground or cut by conventional tool materials."

"I believe we'll see significantly increased use of ceramics as structural materials, which dictates new material removal techniques, possibly some offshoot of grinding," adds Carboloy's Engineering Manager Keith McKee. "In certain cases, such parts will be produced by injection molding to near-net shape, thereby avoiding the problems connected with machining. But, generally any time there are mating parts it's necessary to qualify the surfaces by cutting metal. The real challenge will be how to do it productively.

"Just as important," he continues, "will be the use of conventional materials for production of parts with net or near-net shapes. "This trend should impact metalworking in several ways. First, there will be opportunities for tremendous savings in raw materials. Second, and most obvious, far less machining will be required than for parts hogged from solid stock or those requiring several machining stages. Finally, machining may only be required to satisfy surface-finish requirements. Later, we will discuss the high-pressure forming technologies expected to play a dominant role in generating net and near-net shapes during the next 50 years.

Kapernaros agrees with McKee about the future of conventional workpiece materials. "Steel and cast iron will remain among the lowest cost and most versatile engineering materials available. Of course, improved steel alloys and better grades of cast irons will be coming along between now and 2034. These will have improved properties and will be more difficult to grind and cut. They will have to be ground by CBN wheels, or cut by tools made with polycrystalline CBN blanks."

McKee feels there will be widespread use of composite structural materials as well. The question is, are there diamond and CBN products that will be able to grind and cut these advanced materials? Kapernaros says, yes. "Today's superabrasive products will do the job quite well. But that isn't to say they can't be improved. An advantage of manufactured superabrasive products is that they can be engineered to give optimum results under specific production conditions on specific work materials.

"By varying the parameters used to make diamond and CBN, you can create crystals with closely controlled sizes, shapes and toughness. Surface treatments and coatings can improve performance, too.

"In the manufacture of both products," he points out, "you can control grain sizes--create die blanks with finer grained microstructures for finer drawn-wire finishes, for example. When industry has the need for better performance, it will be possible to quickly develop better superabrasive products." Machine tool impact

Several of the automotive companies are running tests to optimize geometries of tools made from CBN and polycrystalline diamond and also to establish optimum machining conditions. The data will be used to design the next generation of cutting tools and machine tools--forerunners of what will be used in 2034.

"In many cases," Kapernaros says, "grinding operations will replace machining operations on parts made of advanced ferrous metals. The ability of CBN to take heavy roughing cuts (even slice through thick heat-treating scale) makes it attractive for this purpose."

Does this mean that metalcutting operations--turning, milling and so on--will become technologically outmoded? "Not necessarily," suggests Kapernaros. "Users will have a choice between grinding and cutting operations--CBN wheels or CBN cutting tools. They'll pick the best process for the specific application they have in mind. Users will also have the choice of machining and grinding parts in either the hardened or soft state."

Practically all of today's metalcutting machines are designed to use carbide tools. Some of these machines perform adequately with CBN. But, generally they don't have the rigidity, horsepower, high wheel speeds and high grinding-fluid capacity needed to extract the full productive potential from the tooling.

Fortunately, forerunners of a new generation of machines, designed to tap this productive potential, are beginning to appear. These are usually three to five times more productive than conventional machines. The next generation of machine tools, designed to take full advantage of the productive capabilities of these tools, will perform even better. The production grinding machines will probably use CBN wheels to perform all operations, from roughing to finishing, in one setup--very possibly with the same wheel.

Future metalworking productivity improvements won't rely solely on tool or machine-tool enhancements, however. "One thing that will help the future of metalworking tremendously," suggests Carboloy's Sanders, "is more rigid quality standards for work materials. Today you often find a wide range of hardness from lot to lot. If the metal-producing industries improve work-material consistency, it will certainly improve predictability of machining." At the cutting edge

Carboloy's McKee believes, "Machining in the future will be done at very light depths of cut because of the economic necessity of using near-net-shape parts. Nevertheless, the operations may still be performed at reasonably high feed rates because of productivity requirements.

"As depth of cut is reduced for a given feed rate, more pressure is exerted on the nose of the cutting tool. This means that tool materials will have to be stronger to withstand the machining stresses." There will be added demands for more consistency in tool materials, as well, because of the trend toward untended machining.

"I expect that insert size will become considerably smaller, too," he says. "Today, it's common for a user to consume only about 1 percent of the tool material and then throw away the other 99 percent. As the metalworking industries move to more advanced (and more expensive) tool materials, they will also find ways to use smaller inserts."

Furthermore, McKee thinks, "The tools in the factory of the future won't have to be sized. Today, we do a lot of diamond grinding to size a tool to meet finish-size specification. In the future, a tool's size characteristics, in relation to its location on the machine tool, will be done using sensors. These devices will position the tool accurately, within limits, regardless of its shape. As a result, there will be little need for precision grinding of inserts other than for tool surface-finish requirements."

"Sensors will also be used to inspect parts sufficiently to know when a tool is too worn and should be replaced," remarks Sanders. "So, automatic size-control and tool-placement devices will certainly be more commonplace in 50 years. In fact, it will be a lot sooner than that."

Both McKee and Sanders predict that automatic tooling systems used for flexible machining will eventually shift some of the buying emphasis toward toolholders and, in some ways, away from tool materials. This is because toolholders will play an increasingly important role in the system and represent a larger portion of the tooling expense. Moreover, there will be requirements for duplicates of each tool--one in the machine, one in the storage magazine, and one in the tool crib.

Also, unless very sophisticated measuring and positioning techniques are used, future users may tend to be overcautious when running a tool, i.e., it may be replaced more frequently than in today's operations. This could result in less than optimal usage, but will provide more machine uptime.

"With the advent of more automation, on-line inspection and tool-size compensation," McKee concludes, "you will probably see development of a toolholding system that permits indexing an insert without removing the tool from the machine. This will circumvent some of the cost of toolchangers and storage systems. It's an idea that has been around a long time, but never really was put together because the other links in the automation chain were missing. Still, it's an idea whose time will come." Future forming productivity under pressure

We expect most metalworking firms to seriously explore high-pressure forming processes as production economics force specifying net-shape, or at least near-net-shape, parts in the coming years. In fact, it appears that this could make high-pressure technology the preferred forming method by 2034.

Jerry Pfeffer of ASEA Pressure Systems Inc, has helped many domestic companies successfully apply high-pressure forming processes today. We asked him about the benefits of the technology. "In high-pressure forming of sheet metals," he says, "you can usually form complex parts in just one or two operations, instead of the five or six required when the same parts are formed by conventional methods. Moreover, you can create parts with sharper radii, deeper draws, and better definition."

With fewer forming operations, productivity is improved, of course. And tooling costs are reduced. "Probably the biggest benefit," Pfeffer emphasizes, "is that materials that are extremely difficult to form using conventional processes can be successfully and economically formed using high-pressure processes. This means a broader range of material options for designers who will invariably be under pressure to improve end-product performance.

"Using another high-pressure process, you can form steel plate up to 12" thick into a variety of shapes without the need for preheating the material, thus saving energy. Computer in-process measurement and control systems enable manipulating shapes exactly, thus improving quality and eliminating the need for reworking."

Pfeffer notes that yet another high-pressure process, called hydrostatic extrusion, is at least 100 times faster than conventional extruding, while capable of producing complex cross sections.

Finally, hot isostatic compaction of powders--just one of many successful applications of hot isostatic pressing (HIP)--enables production of near-net-shape and net-shape parts with dense, homogeneous microstructures from practically and metal or ceramic material or combinations of these materials. HIP also can be used to heal defects in castings, to clad parts with wear-resistant and heat-resistant materials, and for solid-state welding. Sheet and plate forming

At one time, all of the just mentioned high-pressure processes were regarded as exotic, even futuristic. They have all proven, however, practical and economic in real-world production applications. More importantly, all of these processes have growth potential--the inherent ability to cope with the engineering materials that will be used in 2034.

For example, let's talk about sheet metals. Based on past rates of materials development, it's reasonable to expect three or four new generations of sheet metals to be developed during the next 50 years. Each will have properties superior to those of preceding generations--strength, fatigue resistance, corrosion resistance, heat resistance, and so on. Today's high-pressure sheet-metal forming technologies will make it possible to form these new materials efficiently.

Pfeffer points out, "Larger-sized sheet-metal parts will probably be required by the end of this decade, and today's high-pressure forming processes can be scaled up as needed. Very large high-pressure sheet-metal forming presses should perform as efficiently as today's smaller presses.

"I anticipate that product designers will take advantage of the new materials to develop high-strength, lightweight sheet-metal parts for demanding structural applications. Complex configurations will probably be required, and parts will have to have the final formed accuracy required for easy fit-up. Today's high-pressure sheet-metal forming processes already have the required capabilities."

We asked Pfeffer aobut his expectations for high-pressure forming of heavy plate. He says, "As in the case of sheet metals, I expect stronger, more difficult-to-form plate materials to be developed. There should be no problem forming these materials using high-pressure plate-forming equipment."

Most of today's applications of high-pressure plate forming involve cylindrical shapes--large-diameter pipe for power plants, legs for offshore oil-drilling platforms, and walls for pressure vessels. The process can also form cones, hemispheres, channel sections, and parts requiring compound curvature. As design engineers become aware of these capabilities, we expect the number of applications of high-pressure plate forming to grow. New applications include forming hull plates for submarines and icebreakers, and very large turbine blades for hydroelectric plants.

Note that components produced as weldments today could be formed in one piece by high-pressure forming, improving their structural integrity and reducing fabrication costs. Costs also have been reduced by development of robot-like manipulators that easily handle heavy plate during press loading, forming, and unloading operations.

"I think high-pressure plate forming will be a workhorse of industry in 2034. Its capabilities have just begun to be exploited," comments Pfeffer. Extruding into the future

What about the future of hydrostatic extrusion? This is another process that's still in its infancy so far as the number of production applications is concerned. At the same time, it has been thoroughly proven out.

By using extremely high pressures, billets can bet extruded to simple or complex shapes in one high-speed operation, without preheating. "You can clad base materials with other materials during the extrusion process," says Pfeffer, "making it possible to give low-cost base materials wear-resistant or heat-resistant skins of more expensive materials, thereby upgrading product performance at low cost. Experimentally, we've even extruded heat-exchanger tubing with integral fins, eliminating the need for welding them."

Another advantage of the process is that it's possible to maintain extremely close tolerances and excellent surface finishes. Any defects occurring as a shape is extruded are automatically healed as the process proceeds. Shapes are thus crack-free and flawless, with consistently high properties.

Because of these benefits, Pfeffer predicts hydrostatic extrusion will replace conventional extrusion almost entirely by 2034 for extruding hard-to-form materials such as 7075 aluminum used to make thin-wall tubing. Also, since the process makes it possible to produce complex shapes economically, we suspect that extruded parts will take the place of many parts now produced by other processes. In HIP's pocket

On the basis of its capability to do many things better than conventional processes, we think that HIP will be used extensively in 2034 to produce materials and parts of superior quality. The largest, single application will be hot isostatic compaction of powders--metals, ceramics and cermets--into billets for further processing or into net-shape or near-net-shape parts.

By mixing appropriate powders, it will be possible to tailor materials for optimum performance in specific applications. These materials will have dense, homogeneous, fine-grained microstructures that can't be attained in conventional uniaxial hot-pressing processes, or in processes such as casting and forging.

We asked Pfeffer what he thought. "Hot isostatically compacted billets will be formed close to the desired shape. For example, billets could be formed of optimum shapes for hydrostatic extrusion. Because of the uniform properties of the material, these billets would be easier to extrude than forged billets. Similarly, billets used for forgings could be produced close to final shape by HIP. Again, the consistent material properties would facilitate the forging operation. In both cases, a minimum of raw material would be required to produce final product.

"Difficult-to-work materials," he continues, "will certainly be formed to final shape by HIP. As high-temperature ceramics replace metals, the process will be used increasingly--it's the only economical way for producing high-density parts, having complex shapes, from ceramics such as silicon nitride and silicon carbide."

It seems likely that net-shape parts produced by hot isostatic compaction of metal powders will take the place of parts produced by conventional casting and forging processes during the next 50 years. HIP'ed parts will be more economical to produce and they'll have consistently better properties, leading to improved service life. If a part requires highly directional properties it could be subsequently hot forged. If heat treating is required, this could be accomplished during controlled cool-down after a hot isostatic pressing cycle.

In the future, HIP will not be confined to exotic materials. Parts hot isostatically compacted from ordinary materials--say aluminum-alloy powders--have extraordinary properties due to their defect-free microstructures. The added strength and fatigue resistance will definitely upgrade part performance, making it possible to substitute inexpensive materials for expensive materials. aerospace is the future. If there's one industry that this country can count on being here decades from now, it is our aerospace industry. It will be needed to supply our aerial national defense as well as our aggressive space offense.

It will be here in the year 2000 because there will be as broad a market for its products then as there is now. And it will be using leading-edge technology then to manufacture those products, just as it does today.

In fact, if you really want to see the future of manufacturing, go to Wilmington, NC, to the General Electric Aircraft Engine Group's rotating-part plant. It's a showcase, all leading-edge technology. It is GE's best example of the "Factory-of-the-Future" concept that they are marketing to us in the manufacturing community.

The future is now at GE Wilmington. Changes are going on continually--robot carts are planned for '85, and they're going paperless in '86. Yet, if you come back in the year 2000, it will look much the same. The biggest changes are in place already. Thus, it will continue to represent the best of manufacturing's future.

Ralph Patsfall, chief manufacturing engineer, General Electric Aircraft Engine Business Group, told us a lot about their thinking and planning for their aircraft-engine factories of the future. Here's his story:

Of course there will be some refinements in the propulsion systems currently in use, but the basic jet engine of today will still be in use 20 years from now. It's a product that's here to stay. When you look at competing propulsion systems, there's nothing else right now on the horizon, nor any other company not now in the propulsion picture that we could imagine becoming significantly competitive.

Right now, we're on an upswing, our manufacturing load factors are increasing, and because we are in the defense industry, (in addition to marine, industrial, and commercial), we will see increasing orders. Sure, historically, we've had major military-engine programs canceled and had other fluctuations, but the future of our military business looks very good right now.

Our government wants to have at least two major engine suppliers competing--and Pratt & Whitney certainly makes a good competitor--but we will not only keep them on their toes, we're going to beat them! Satellite technology for the group

As far as manufacturing the product is concerned, as we go into the future, there will be more and more use of microelectronics and computers, and flexible manufacturing that will group parts somewhat differently than we have in the past. We have talked families of parts for a long time, and we have a charter-source system where we manufacture rotating parts in one section and fabricated parts in another satellite plant.

Wilmington is a good example. It is sort of a clone of what we have in Evendale, OH, where the base technology was developed. With the charter-source system, you zero in on a family of parts and aim all planning, strategies, and new technology on that product line. In rotating parts, we started using NC technology way back in 1956, probably one of the earliest installations in the US. That has put us in good stead because we used that base technology to grow into centralized computer control systems. Over the next 20 years, that trend will continue. Rotors held together by friction

In rotor welding, we got started in the early '70s with monolithic inertia-welded rotors because it yields a much more stable rotor than bolted connections. Today, we have a few parts, very few, that are welded by electron beam, but nearly all of our rotors are inertia welded. And I don't feel that this will change much in the future. We are the only ones with inertial welding in high production--at least the volumes we have in the jet-engine business.

There is a size limitation when you go to much larger parts, so it might be more practical to use electron-beam methods, but wherever inertia welding can be used in future applications, it will be used. We have developed that skill to a very high degree; we can hold accuracies to [plus-or-minus] 0.005" on concentricity and [plus-or-minus] 0.020 on lateral position of the web. That's the real trick to the success of inertial welding. The ultimate job shop

We are actually the largest small-lot manufacturer in the country. Our D700 engine will probably reach 100 engines/month at some point, as we did with the J79, however this year we will be up in the 50 to 60/month range. But when you look at the total range of engines we manufacture, there's a large number of engine models. In some cases, such as the building of spares for the J79, the rate is very low, like 2 or 3/month. So even at $3-billion/year sales, we are a small-lot manufacturer.

And this is why the computer has really helped and will be the key to our future evolution in manufacturing. When you can group axisymmetric parts and prismatic parts, ignoring families to some extent, but more by process and strict geometry factors, you can benefit greatly.

The trend in the future will be the use of group technology to an even greater extent, coding of components along those lines, computer process planning, and of course NC-automated machinery to complete the task. We will be adding flexible manufacturing systems that will handle specific groups of parts with varying geometries and varying lot sizes. We will have the ability to introduce new product lines a lot faster, easier, and be able to make design changes with a lot more flexibility.

And remember, the machine doing the work doesn't know whether it is a military, marine, industrial, or commercial-airline part. The benefits will be the same in all areas, and a particular machine may see in a month's time all three categories. That's the way our plants are laid out. This allows us to cut cycle times and introduce product improvements quicker that improve the performance of our engines at less cost, and all these are direct benefits to our customers, whether it's the us government, an airline, or some gas pumper. The evolution of computer controls

Looking back historically, when NC first came in, we were dealing with thermionic devices that were highly unreliable. In the '60s when solid-state electronics came into being and was followed shortly thereafter by microelectronics, the vast improvement in reliability that they provided expanded the use of electronic NC tremendously.

To take advantage of the computer and microelectronic technology, we moved from individually NC-controlled machines to computer control to distributed NC control that controlled from a central location by downloading through a workstation.

Today that workstation has added a CRT and alphanumeric keyboard to bring the operator into the picture. And we can now add modules to the system like centralized computer-process planning in which the computer plans the sequences of operations, and the types of tooling that's needed at each workstation.

We have maintenance diagnostics, in some cases with sensors on the machine telling you that you have a coolant problem, lubrication is low, a switch is bad, etc; or a menu system that leads the operators through an analysis of machine condition and automatically generates a work order for the maintenance department to dispatch the right skill to fix the machine.

You can also add a central quality planning and control system that is very valuable, and production controls, scheduling, inventory control--all these things can be added to this network to get the best of today's technology. Modularly moving into the future

When you look to the future modernizing of this industrial base, you attack it on a modular basis. It starts with doing a proper job of planning, such as we did in transferring Evendale technology to the new Wilmington plant and tieing it all together electronically.

The GE CALMA interactive graphics system on which parts are designed is also a key factor. Computer-aided design will be coming to other components and product production plants in the future. We don't relay on drawings in our rotating parts plant. The alphanumeric information and the electronic geometric data is transmitted electronically to our manufacturing organization. They use this electronic data base to do the machine-tool programming, the computer-process planning, the quality planning, the designing of fixture and tools to hold and machine the part, all parts of an integrated system. This is done completely in our rotating parts facility, and is not yet installed at our airfoil, fabrication, and casing installations.

Thus, in the future we will be expanding on this capability in our satellite plants. Our strategy is a modular, pay-as-you-go plan. When you develop a module, it then must be transportable to our other operations. With this strategy, we have done well in the areas of networking and computer-information systems. Closing the loop

Our goal in the future will be to close the loop on those operations--make them real-time controlled. We will use probes for direct measurement to determine what kind of automatic offsets are necessary to generate the dimensions you're cutting with a machine tool, or on a continuous basis sense the dimension you're creating to compensate for tool wear. We're testing closed-loop machining now on pilot equipment under shop conditions on a real-time basis, where dimensions are sensed either intermittently or continuously. We can determine what corrections to make after a rought cut, for example, so that the finish cut will be absolutely precise.

That's one key thing coming in the future. We have about 106 NC machines in Evendale and 110 NC machines in Wilmington all under direct numerical control. Eventually, all of these machines will be put on closed-loop machining, and this will take time. We're looking at the '85 or '86 time period for completion.

The SME/University of Michigan Delphi study said that by 1985-1990 only 15 percent of machine tools would be NC. We are buying 85 percent NC tools today, so some companies are way ahead and others are 20 years behind. Adding the physicals

In computer communication and the transfer of data and information, we've done a good job, but where industry in general has fallen down is in adding the physical system. We have the electronic information, but we need the tools to make such things happen as automation of parts storage, movement of parts, transfer to workstations, and automatic loading; all under central control.

This is the next step at Wilmington. We currently have a computer-controlled warehousing system, a large stacker crane under computers control, and we are adding the physical transporting robot cars to move material to and from the workstation, and also the fixtures and cutting tools.

You must plan ahead to use universal pallets and fixtures so they will handle the complete family of parts. Such a system can run a 1-pc lot size just as efficiently as a 50-pc lot size. Or you can mix 50 different geometries as long as they are all the same group technology. Parts will be prefixtured.

If you go beyond the next 20 years, I think you will see microelectronics and the computer becoming the nerve center for the factory, and the kinds of decisions that are made today by the management of our factories will be made by the computer. It will be a matter of keeping those systems running, generating the proper software, developing the decision trees and data bases that are necessary. It will be all real-time, closed-loop rechnology and this could be 50 or 60 years away. People power

This will mean new skill requirements for our labor force, but the microcomputer will help here also because it will be used to train people in these new skills of software, software engineering, system engineering, maintenance, etc--all at the higher level of sophistication it will take tokkep thse systems running.

I see a displacement of skills, but I don't see any shrinkage in the number of people. Automation's net effect for us will be to bring in more business. It increases your throughput, decreases cost, creates new business opportunities, and help you gain in employment and create the need for more new skills. The effect in the near term--the next 20 years--will be an increase in employment.

Now, there are a lot of people who will disagree with me on that. But certainly in aerospace where we must deal with a lot more precision and highly critical parts, quality will be as good as we have today, and it will get better because of the closed-loop benefits and the kind of trend analysis the computer can provide. But this will require more monitoring. There will be more people interacting with the system, and making decisions.

There will be a continual changing and upgrading of programs. The need id for software programmers and engineers, systems engineers, and skilled maintenance people.

In the area of management, new structures drive the ownership for quality down to the lowest levels of the organization. This is really what those who use participative teams are trying to do, to get everyone to assist in efficiency programs. Unions, some of them at least, seem to be reluctant to participate, but they seem to be coming around. This simply has to come if you are going to be competitive. Materials of the future

In the materials area, we will see the use of more composites. Our Albuquerque plant right now manufactures composites, and we're putting in an automated installation there for making structural composites. This takes some different processes than the machining of rotating parts, but conceptually it is the same idea.

Ceramics are coming also, but I see problems in how to handle their brittle nature. The development of processes that could give them more tolerance to small defects is critical. Additional research will be necessary before they will be useful in the sizes of turbine engines we make, although the smaller automotive applications do have some potential, particularly supercharger impellers. We see some good possibilities in lightly loaded static components, like turbine shrouds. But applications in high reliability rotating parts are well into the future.

I don't see any big changes in the makeup of rotating parts in the engine, other than the use of somewhat different metal-processing techniques, such as the use of precision castings in rotating parts to a greater exrent. As we move into composites, ceramics, carbon/carbon materials, and the coatings that are necessary in some cases, these will require new or different processing techniques. Process of the future

We're using electron-beam welding on a few components now, and I anticipate an expanded use of this joining technique in the aerospace industry in the future. It gives you a high penetration in difficult superalloys, a very dep weld without the preparation and grooves required by arc-welding methods.

The use of lasers will expand dramatically in all areas. We are installing our second laser blanking machine, cutting blanks that would normally be done on a press with a die set. The laser saves the high cost of the dies; all you have to do is program the movement of the head. This flexibility and the nesting efficiencies that save materials make lasers much more economical for our use. Generally, the laser will be used much more extensively for measurement.

Electrochemical machining is another potential growth area, one that cycled up years ago and then came bank down. ECM will again be useful in forming monilithically bladed discs that are made as a unit rather than the dovetailing of blade elements between rim and hub in present turbines. EDM will also be used more. Machine tools of the future

The machine-tool industry will move toward multioperation machines. We've had many machining centers for milling, yet very few that do turning, milling, drilling, tapping, etc. So the trend will be to do more and more operations on a single machine to eleminate queuing problems and transfer problems. The more operations you can perform on a single machine--within reason, of prohibitively expensive--the higher your productivity.

Machines will have to be built to accommodate the loading of parts, fixtures, and cutting tools. This will be designed into the machine, not a burden for the user to develop on his own. Some thought will need to be given to how to marry universial tooling to a machine so that you can use effectively a variety of parts within the same group technology.

As you look further into the future, as artificial intelligence becomes more available, expert systems will be knowledge-based and "think" like experts. A good example today is maintenance diagnostics where the computer has a decision tree and interrogates the operator from a menu, and when he answers, it moves on to the next decision. That will become much more sophisticated. You see that already in the medical profession today.

As this expands, you can visualize the control of the whole factory that way. When a machine goes down, the system decides which other machine to put on-line and use in its stead, reschedules the operation for maximum efficiency, and all these decisions could be left to the computer. With adequate sensing devices out on the floor to close that loop, the computer can run the whole factory. Robots become more custom-tailored

I think that in the next 20 years you will see more use of the robot as a robot, and less as a simple human replacement. Eventually, and this goes beyond that 20-yr period, the robot won't have the physical limitations of either the human being or robot of today. It will have artificial intelligence, much more flexibility of motion, and it won't look like a simple articulated copy of the human arm. It will be tailored for a specific industry and a particular product line, and designed with much greater physical capabilities. This would involve completely intelligent systems with sensors for real-time control, and surpass today's human robot or articulated-arm capabilities. Sole survivor in airframes? In commercial aircraft, there may be only one manufacture of large planes by the year 2000. When you ask the Boeing people about this, they continue to maintain that their two competitors--McDonnell-Douglas and Airbus--are healthy, but the former is expected to phase out of commercial aircraft soon, and the future of the latter is hanging by political threads.

Thus, Boeing seems to be in an enviable position, if they were not at the mercy of the market--the presently unregulated and unprofitable airline industry. Boeing's commitment to high technology is well known, but to explore in detail their views on how they will be making planes in the future, we interviewed Gary Michaelson, director of manufacturing operations, Boeing Commercial Aircraft Co. Here are his insights on Boeing's manufacturing plans:

We predict that we will be building airplanes at similar rates for the next 20 years. We hope to see improvement from present levels as the airline profitabilities get healthier. We expect the market life of our newer models--the 757 and 767--to be 20 years, with various improvements. Of course, we hope to introduce some new models during that time as wel. But we expect that the aircraft in our current inventory will continue to be produced during that time span.

We expect to maintain our market share or improve on it, but looking at it from the manufacturing standpoint, it is not that stable because of the influences of the marketplace, i.e., the profitability of the airlines and how that affects our production rate. We have an especially competitive environment with Airbus and McDonnell-Douglas, and we might see some emerging competition years from now from Japan, possibly merging with other countries to produce smaller aircraft. Both aluminum and composites will grow

We expect to be building essentially aluminum aircraft for some time to come, but I expect major growth in advanced composites and newer aluminum alloys. We certainly have that capability. Yet we must maintain our capabilities to build the aluminium aircraft we have in production today. That includes the 757, 767, 737, as well as the 747. The latest derivative is the 737-3000 we just introduced. We still make a few 707s and 727s.

Of these, the 757, 767 and the 737-300 now have about 3 percent of their structure made of advanced composites, and some amounts of titanium and high-strength steel are used in landing-gear beams and forgings. Aluminum predominates for 65 to 70 percent of the plane.

The advanced composites include graphite epoxy, Kevlar epoxy, or some combinations of both, and are in secondary structure applications like control surfaces, tail fetherers, etc.

You must recognize that the commitment to those materials for those aricraft was made in the mid '70s. We were at an area of risk where we could only commit to secondary situations because of the newness of the materials then.

Today, we would be looking from a position of more confidence because of our developmental and service history with these materials on the 757 and 767. There was an extensive amount of durability and fatigue data generated before the first plane was delivered for FAA qualification and certification. We have composite programs dating back to the late '60s, working with NASA and the Air Force. Tape-laying automation cuts cost

These advanced-composite material costs are much more expensive than aluminum. We pay something like $50/lb for graphite/epoxy material compared to only $5/lb for advanced aluminum alloys. Currently, most of the advanced-composite material is fabricated in our factories by hand-layup operations.

Our new emphasis on tape-laying automation is a thrust to get into a more productive mode to get the total part costs down and make these materials more competitive with aluminum. We must have this automation to be able to afford the weight savings we can gain from these materials.

We are currently using a tape-laying machine for 737-300 components. This is an Ingersoll machien like the one used at General Dynamics. Cincinnati Milacron is now introducing a tape-laying machine that can work with contoured surfaces. We stimulated them to do that because we feel that that's where the biggest cost benefits are going to be: to be able to lay the tape material directly on the mold that we cure that material with, without the need to transfer it. The Ingersoll machine has essentially a flat capability for laying tape at fairly high speeds. It's a lot like adding multi-axis capabilities to a machine tool. The same technologies apply as far as the movement of the equipment is concerned.

As this technology has matured, the key factor is improvement in the ability to handle the tape material as well as better control in the manufacture of the material itself. As the process becomes automated, we must have better control of the material quality since the machine is not as effective as human beings are in adapting to variation in the material.

Sensors that we've developed allow us to inspect the tape for foreign objects entrapped in the material, and track the width and alignment of the material. We want to assure that as we lay that expensive material down that it is in the right place, that it has no foreign debris in it, and that it has no laps, gaps, or broken fibers. Then, when it's all done, we don't have to slow down our productivity by needing total inspection of the unit.

We look to automating our tape-laying process into a family of processes similar to what we have now on the metal side. Compared to carving parts from aluminum, we feel that we can save from 20 to 30 percent in final part weight; and in waste, the 5:1 ratio of "buy to fly" in aluminum drops to numbers as good as 1.1:1. Of course, the high cost of the material forces you in that direction, and the inherent design flexibility helps. Pultrusion gains more pull

Derivative composite processes are filament winding, molding, and a process we pioneered called pultrusion. This whole family is nearing the production stage now.

Pultrusion is similar to extrusion in metals. We collect a series of tape materials in a tooling system and pultrude various structural-stiffener shapes like half sections, I sections, angles, or bars. This is economical way to generate constant-section shapes for stiffening ribs and spars, without needing autoclave finishing operations. The material can be partially cured for later curing in concert with the balance of the component, like a rib or spar, or it can be pultruded in the fully cured state for complete angles or Ts.

Complex tapers are a little beyond the capability of today's equipment. Within limits, the elements can be shaped into arcs, but compression creates wrinkling or distortion. There are other ways to make contoured frame sections. Composites have big future

When we look ahead at the future composition of composites in aircraft, I would say it's reasonable to expect anywhere from 50 to 65 percent of the aircraft structure would be from advanced composites. I see the potential of replacing a lot of our aluminum primary wing and fuselage structures in the next 20 years.

The skin of the aricraft will also be advanced composites. Wings will be manufactured as an integral piece, rather than manufacturing skins and mechanically fastening them to the wing structure as we do in aluminum now. This would mean fewer parts to control and manufacture, and the ability to compete very nicely with the aluminum aircraft we build today.

We recently received an Air Force contract from Wright-Patterson to investigate methods to make large composite fuselage structures. So the interest is both military and commercial. We have already certified with the FAA a primary structure of advanced composites--a 737 horizontal stabilizer in August of '82. We will be flying five of these in service-evaluation programs

on Alaska and Delta Airline aircraft. This is the next big step. Manufacturing futures look bright

As we look ahead, metalworking basics will not change too much but processes will. We will continue to use machining, mechanical fastening, and forming, but the manner in which we do these processes will evolve from stand-alone NC machines and riveters of today into more sophisticated manufacturing systems.

To get productivity up and costs down, you must bring a systems-engineering approach to integrate these stand-alone systems, reducing overall costs, floor space, and the flow time necessary to manufacture parts. We will see manufacturing systems over the next few years that are very flexible, productive, and cost effective. We have one FMS system currently in our aerospace company, and we expect to implement in the next year ro so an FMS system in commercial-aircraft fabrication.

This is how manufacturing will evolve for us in not only machining, but also forming, fabricating interior components, electrical wiring, etc. These same machining concepts can be applied.

Our computer-controlled shot-peen forming, fabricating interior components, electrical wiring, etc. These same machining concepts can be applied.

Our computer-controlled shot-peen forming system is being used to contour all of our wing skins on the 757 and 767 aircraft. We've added the missing link--before we had to take these wing parts off the vertical carriers and lay them down in horizontal fixtures to manually check the contours. We are currently implementing a machine to automatically check those contours in the vertical position. Downstream someday would be real-time checking as you peen the skin shape. Networking that really works

Our computer networking will continue to evolve. I'll never be fully satisfied. We will continue to integrate our CAD and CAM system, add further sophistication in DNC systems, and we're still only seeing the front end of the full implementation of these sophisticated systems. Sure we had the most complex changeover of any industry when we went ot CAD/CAM, but we still have a way to go to stay up with the state of the art.

We're tied together in some areas with CAD/CAM and are incrementally picking off those targets of opportunity as time goes on. The complete aircraft is defined electronically in the data base, it's just a matter of using this information more fully. We've already integrated it into some of our floor-panel and tubing areas, but there are still opportunities to move toward more direct numerical control and away from NC programming steps and the generation of machine-control tapes to get DNC implemented more extensively throughout the manufacturing process.

Twenty years from now, our network will be radically different. You will see the coming together of the various networks we have currently into an integrated data base labeled "computer-integrated manufacturing" or CIM. I don't know of anyone with CIM in a complex application like aircraft at this time, but they are certainly evolving in that direction. Vendor relationships will be stable

The kinds of things we subcontract out now, we will be subcontracting in the future. We have a very good vendor base, and I expect we will add to that as newer supplier capabilities emerge. Our 747 is 60 percent contracted out now, and the current four major programs are about 50 percent or more.

As a company, we are already more oriented toward assembly than manufacturing, with the exception of those things we call our "lifeline components." This includes the critical wing components, flap tracks, and other key components. People will always count

Over the past 20 years, we've made big inroads in productivity improvements. We will continue to evolve and take advantage of the technology that becomes available through robotics and manufacturing systems. We're going to play a key role in not only developing those systems, but also defining the requirements that others will have to meet to provide us with equipment to make those productivity strides.

In our recent negotiations with the union, we've put in some fairly unique provisions for retraining people, and for the identification of new technology as it evolves. We believe that keeping people aware of what we see coming is helpful both for us and the labor force. Big automation cost benefits ahead

As we get more productive and automated, our costs should be significantly reduced--40 to 50 percent, I feel, is obtainable in the next 15 to 20 years. With that kind of a cost reduction, we would make ourselves much more competitive than anyone else will be able to do, and as a result we should be able to sustain higher production rates and a stable work force. Sure, that's optimistic, but if I weren't an optimist, I wouldn't be in research and development.

It won't be easy, or a sure thing to happen. But it is within reach. A lot of things will be involved--time is one, and money is another. The right attitudes of management, especially, is required to make these things happen, plus cooperation and teamwork within the organization. Filling the ME shortage

One key area we've identified is the talent and skill of the people we're going to need to make these manufacturing systems happen. We're a little concerned about the small numbers of universities--and engineers coming from them--that have manufacturing-engineering training. The quality and numbers of these people is a potentially limiting factor in our future. So we are working very hard with a number of universities right now. The complexity and sophistication of the manufacturing environment is going to grow at a rapid rate, and we must step up as an industry and as a country to meet that challenge with the proper emphasis on the manufacturing side.

A lot of students today are not interested in the manufacturing side, they want to take things like aerodynamics, chemistry, or what have you. That's not all bad, but they should have more exposure to systems engineering and the complexity of the equipment and processes in the advanced manufacturing area.

Exposure to computer systems and computer sciences is essential, of course, to anyone we will be hiring in the future for involvement in our manufacturing technology. And we're seeing a lot better background and training in those areas for the colleges today. But they need more exposure to what the manufacturing engineer does for industry. It isn't very well exposed or defined. People think of manufacturing engineers more as blacksmiths than engineers. A certain snobbery still exists in many of the college faculties that must be overcome. Training for the future

We have an in-house training program for manufacturing engineers, and provide a lot of on-the-job training where we're in the build-up mode. We have learning centers in most of our facilities where off-hour training is available. We encourage people to go back to school for updating and retraining, and provide undergraduate training programs and paid educational programs, as well as graduate opportunities.

We're trying to put all the tools in place and allow our people who are so motivated--and hopefully, we're motivating them in that direction--to take advantage of these programs.

But for the future, we believe that that's the wrong end of the horse to feed. We must get more of our schools and colleges more strongly emphasizing the manufacturing arena. Most of their laboratory equipment is very old and outdated. Because of budget cuts or whatever, they are far less modern than our factories are. It's hard to train people on advanced NC systems without access to real equipment. We are trying to stimulate people on all levels--the university level, the government level, and within industry itself. Automotive factories of the future

The future of the US auto industry is one of the most important issues facing us today, both as members of the metalworking community and as citizens. For ech of the five US carmakers, the biggest question is which of them will survive the next 10 to 20 years. Will there be any North-American car company left by the year 2000?

Answering that question today is impossible. But there is good reason to try to anticipate what we can about the US's single-most important manufacturing industry. Its success or failure will affect us all.

Clearly the role of less developed countries will grow. Their low labor rates represent a considerable threat, particularly in automotive components. However, the high cost of capital investment suggests joint ventures are the most practical solution to survival. The real question is how quickly can auto factories of the future be built, and overcome the obstacles of inertia, training, and technology?

Few doubt that the US will remain the world's largest consumer of automobiles by the year 2000. The real question is where will those cars be built? Designed for 2001

But first, what will the cars of the future look like? What will they cost? Will the car body be injection molded in one huge piece of plastic like a toy car? Will the car evolve into a throwaway appliance, or be built to last for decades? Will it get 200 miles/gallon? Or 20 miles/kilowatt?

Ask the auto industry these questions, and they don't answer. Either they don't know or don't want you to know what they know. They have good reasons. Their answers would affect an awful lot of people: consumers, auto-industry suppliers, auto workers, unions, taxpayers, etc--not to mention their competitors in the auto industry.

Marvin Cetron, president, Forecasting International, is willing to make predictions. He sees cars ultimately being assembled in Mexico, South America, Korea, and Italy or Africa. That's it! Four car companies! Labor costs will eliminate the rest of the contenders.

The car of 1990, he predicts, will be 92 percent nonmetal--fiberglass and plastic--and it will have a ceramic engine. Then, depending on what happens in improvements in energy/density ratio, by the end of the '90s he says we will see battery or fuel-cell-powered cars because by then they will have the car's weight down to 1000 lb. A small engine will be used for peak power, but the basic car will be battery powered. And this car will last for an average of 20 years.

As he explains, "The auto industry that I see is headed for a state of flux. I gave this projection to the SAE conference last fall, and they were all shook up. Most people knewthis, but the auto industry people said, 'My Goc! You're scaring all our suppliers!'

"I don't mean to scare anybody. I'm just saying, 'Hey, the numbers are there. Take a look! Don't you people plot any points?' And they replied, 'No, that's not our job. Our job is to respond to the types of things the industry wants.' So I said, 'Man, you've got to do more than that! You've got to look into your future!'

"That future is that the automobile today lasts for seven years and ten months, which is longer than at any time since World War II. It would last for 10 years and three months in 1990 with simple extrapolation. The Volvo today lasts 16 years! So, you can see why I feel that our cars will certainly last 20 years by the year 2000.

"Part of the reason is not just that cars are being made better, but it's because people are just not getting rid of their old cars as quickly as they once did. They keep them for their kids now. So, looking at the auto industry in 2000 when we're going to be manufacturing fewer cars that last much longer, the biggest sellers of used cars will be Avis and Hertz!" Breaking even

We have already seen a major transformation in the auto industry, during this past recession. As GM's Chairman Roger Smith told the Wall Street Journal in December, "We took our industry apart and put it back together again."

The industry that emerged is one that can break even on one third fewer sales. In three years, Detroit laid off thousands of workers (many of these were white-collar jobs), squeezed suppliers for millions of dollars in concessions, canceled or delayed over a dozen new products, and closed enough plant space to house a small city. Altogether, the Big Three chopped more than $10 billion out of their annual costs.

As the journal concluded, "While slashing costs, automakers have avoided passing on their savings to customers; in fact, shielded by quotas on Japanese imports, they have raised prices sharply. At the same time, they have benefited from customers' renewed preference for big-ticket, higher-profit cars. So, with sales far surpassing the lower break-even points, automakers this year have been able to collect average profits of more than $2500 for every vehicle that they sell beyond the break-even level." How much can you automate?

More and more, Us auto companies are acknowledging that keeping their huge US manufacturing base competitive in world markets as they enter the 21st century involves more than throwing microchips and robots at the problem.

The futility of countering the $1500 to $2000 cost-per-car advantage enjoyed by Japanese automation is best illustrated by noting that $2000 is 85 percent of GM's average labor cost per vehicle in 1982, when production was at 60 percent of capacity. Total elimination of labor costs would hardly be more than a break-even proposition, even if we ignore the enormous capital investment it would take to build a workerless plant.

Furthermore, a closer look at the Japanese auto industry reveals that despite their use of advanced robotic assembly, they are essentially less automated than most US auto plants. Toyota Motor Co Ltd, for example, the most profitable of the Japanese giants, relies far less on automatic equipment than its arch-rival Nissan Motor Co Ltd. But this doesn't deter US automakers from testing the far reaches of technology. Japanese ideas come to California

Given equal weight with technology right now is some serious reappraisal of management techniques. The most visible of such efforts today is GM's joint venture with Toyota Motor in Fremont, CA, a Japanese-managed assembly plant that will turn out Toyota-designed small crs for sale through Chevrolet dealers.

According to GM's Roger Smith, the main purpose of the venture is educational. It will give GM a chance to see if Japanese management techniques work in the US environment. It also gives Toyota a chance to try out its system with US workers organized by the UAW.

Yet, it appears unlikely that Toyota's kanban system of a central assembly plant surrounded by supplier plants (many partially owned) can be duplicated here. The Fremont plant will depend heavily on components shipped from Japan. Also the ups and downs of the California marketplace will be a far cry from the market-insulating effect of the strong export program and supportive Japanese trade policy that Toyota enjoys at home. Saturn and its satellites

Plans for GM's Saturn Project--a late 1980's small-car program--promise revolutionary plant organization systems and state-of-the-art technology, demonstrating that Detroit remains enamored with manufacturing gadgets as the key to future profits and survival.

Saturn will rely on highly automated manufacturing modules linked by interactive computer networks to coordinate production, inventory, and volume. Theoretically, this would combine the advantages of high-level automation with the flexibility of the cottage industry, some owned by GM and some by suppliers. The State of Michigan, which stands to benefit from the satellite concept, has already set up a manufacturing technology research institute near Ann Arbor to help suppliers deal with the technology of interfacing with such a manufacturing complex.

Ford and Chrysler have also told major suppliers they are going after long-term relationships and are willing to sign single-sourcing agreements. New vendor relationships

It is rapidly becoming mandatory for suppliers to maintain computer linkups with primary manufacturers. The rapid adoption of computer-design terminals means the auto industry can cut dies directly via NC or transmit drawings electronically. To use these drawings, a supplier must be able to establish a computer link.

Paul Guy is director of the manufacturing Engineering and Systems Office at Ford. He points out, "We have an aggressive program to bring our suppliers into more of a partnership relationship. We want them to feel that they are more of an extension of our manufacturing base than perhaps they have felt in the past.

"This means linking them electronically with many of our engineering systems here at Ford so that they can communicate with us quicker and more accurately than the old systems of shipping bundles of blueprints back and forth, which introduced all kinds of opportunities for error, misunderstanding, and miscommunication. So we have established, particularly in our body tooling area, direct electronic links with a number of locally based suppliers.

"With machine-tool suppliers, we are stressing a sort of partnership agreement. WHoever we place business with must stay with us through the launching of that new machine tool or transfer line until a certain level of production efficiency is reached. As part of the original procurement order, we must mutually agree on what that objective is. We recognize that we, as the user of the equipment, have as much equity in reaching a productive efficiency level as the man who builds the equipment."

Ralph Behler is program director of the Advanced Product and Manufacturing Engineering Staff, General Motors Technical Center, Warren, MI. He agrees that vendor relationships in the future will become much more closely tied. "Not only will our product-engineering and manufacturing-engineering functions within GM be linked closer, but our relationships and communication systems with vendors will be much more responsive and faster. We will be able to communicate as well with a supplier as we do among our own divisions.

"This new technology is becoming available very rapidly, and the people who can understand it the fastest and apply it in ways that make good business sense are going to be in the strong competitive positions. The ones that hang back will have serious problems." Lower volumes demand higher flexibility

In the past, Us automakers have pursued low-cost production through intensive investment in laborsaving equipment, recouping these costs through high-volume model runs. Volume sales of low-priced cars were essential, but the low-priced segment has become fragmented by import competition and buyer preference for the higher-priced domestic product. Complicating this has been the hot and cold market shifts for diesel engines, convertibles, subcompacts, and hatchbacks.

Thus, manufacturing planning today demands flexilibity and the reduction of break-even volumes. GM's Buick City complex in Flint, MI, is an attempt to convince suppliers to build light assembly plants and warehouses close to main assembly plants. The bait is long-term exclusive contracts and outsourcing some assembly processes. This is seen as a prelude to the Saturn philosophy of building whole subsections of a car outside and leaving only final bolting together for GM's assembly plant.

The design of the Pontiac Fiero also points to modular assembly--a cage-like framing system that supports prefabricated panels hung on the exterior-like wallpaper. This permits major assemblyline reorganization and the prepainting of panels. Modular models and methods

Ford's Guy explains the new industry philosophy, "Yes, we see a definite fragmentation of the market. Ford has realized that we won't have the same economies of sale that we had as recently as five or six years ago. We will be dealing with smaller, more fragmented markets, and therefore our manufacturing operations will have to be designed around more flexibility and less fixed, or dedicated processes of five or ten years ago.

"Smaller volumes mean more tooling changeovers and a greater need for flexibility to change from one vehicle product line to another. Thus, the focus in manufacturing technology is to design new systems in our new factories that are more flexible, and to design systems that can be retrofited into our current facilities that will provide more flexibility for us to change in less time and at less investment cost."

Paul Guy continues, "There has been a lot of dialogue in the past two or three years between our automation engineers people in terms of modular concepts of vehicle assembly, and how to simplify the structure of the car so that it lends itself to more automated assembly.

"The structure in many cases inhibits us from assembling many of the trim and chassis parts. So we're looking at the structure of the car as a whole so that we can make the next generation more amenable to trim-and final-chassis assembly automation. Such things as how do we get the seats in automatically, how do we get the headliners in, the quarter-trim panels, the door panels, etc--all parts that have historically been installed manually.

"How can we design these in the future so that these components can be installed by machine? We haven't got all these answers yet, obviously. There are many technological problems to be resolved.

"We are striving for closer relationships between the guy who designs the product and the guy who produces it. More and more, we are linking up in teams of product and manufacturing engineers when new products first surface so that the manufacturing considerations are taken care of up front during the early phases of the design." The end of mass production?

On the subject of flexibility, GM's Ralph Behler feels, "I think our approach at GM in manufacturing car elements will remain much the same for the next 20 years. What we will be doing is developing computer-integrated engineering and manufacturing systems. These will be very flexible and a lot more responsive than they are now to shifts in market demand."

Does that mean mass production as we know it will come to an end? "Yes, I think the use of dedicated single-purpose automation will diminish dramatically over the next 20 years, but there still may be places where we will use it. We will develop facilities that are much more responsive to market conditions. Design, process planning, and the whole factory complex will be tied together, fully integrated, and much more automatic and flexible than they are now.

"When you walk the shop floor 20 years from now, you'll see essentially the same machines we have now, but the differences will be inside the control boxes on these machines and inside the control room where the manufacturing decisions are being made.

"All plants will have electronic communication systems and there will be little or no paper required. Machine diagnostics will be available at each machine or wherever the repairs need to be made. The computer will be underlying almost every aspect of the manufacturing system." An FMS in your future?

Will the auto industry have use in the future for flexible manufacturing? Ford's Guy thinks so, "We are looking at flexible-manufacturing methods, even though they have not traditionally been geared to hgh-volume manufacturers like Ford. But we are beginning to see the need for such systems that have very high inherent flexibility.

"We are looking at the possible use of such systems for variants of high-volume engine or power-train components. There are pockets of customer demand for high-performance engines and high-torque transmissions, and maybe the market is only 50,000 units/year or 75,000 or 100,000. In the past, those derivatives of our standard product line would be sourced outside, so we're looking at how this technology might be applied to produce lower-volume products profitably inside the company, and produce not just production parts, but also prototypes and service parts on the same manufacturing system.

"FMS is very expensive; it takes an incremental initial investment to get into that type of business, but it is being looked at. It also takes a very long-term perspective. It calls for more foresight than maybe we've shown in the past."

GM's Behler agrees on the role of FMS. "FMS will be part of our strategy, in both meanings of the term--flexible-machining systems as part of the flexible-manufacturing systems we envision. Although FMS right now is very low volume, I think the technology will advance into higher-volume applications." Materials will evolve slowly

On the subject of what cars will be made of and how that will affect the production process, Ford's Guy said, "For the rest of this decade, the trend is obviously to lighter materials, but I would not characterize the changes this might have on our production processes as major at this point."

Adds GM's Behler, "In the mix of materials that will be going into cars in the future, we are already seeing the use of composites and plastics in new areas. We will see a great deal of competition and opportunities for alternate materials. It's still tough, though, to speculate what percentage of the car in the year 2000 will be metal, but we can forecast with certainty that the material options we have as designers will be greatly expanded. Those people with knowledge and leadership in understanding these new materials and processes are going to be competitive.

"Despite all these alternatives, however, the advanced capabilities of the computer will enable the designer to analyze his options more fully and optimize his material choices very qhickly. This is a plus-plus situation--more choices, but more ways to quickly evaluate these choices. And there's no question about it, the consumer will benefit by getting a better product." Lost foam finds a home

Points out Ford's Guy, "Certainly one of the key things we're looking at is near-net-shape processing. In the early casting or forming of the product, we want to get much closer to final shape than we have in the past. An example of that is the evaporative casting process that uses Styrofoam cores into which we pour the molten aluminum. The Styrofoam evaporates into the sand and you're left with a metal replica of that core. The process yields much more accurate castings than we were able to obtain from sand-core methods, dramatically reducing the amount of machining we need to bring the cylinder head or engine block to its final form.

"Processes like this and powdered-metal technology, cold forging, etc are being looked at as an integral part of the design-review phase. We must ask ourselves how we can design the part so that we can manufacture it in less steps, to get it to its final form without all the expensive processing we employed in the past. The manufacturing engineer is now contributing his ideas on design changes that will accommodate his more sophisticated manufacturing processes."

CM also is counting on the lost-foam process. They are using it now to produce the cylinder heads for the Olds 4.3 liter V6 diesel, and are planning to use it for the Saturn four-cylinder fuel-injected cast-aluminum engine. They exect that the Saturn blocks will hold dimensions so precisely that machining will be reduced by 40 percent. Assembly gets it together

Automating assembly operations will take some time, ford's Guy concedes. "Major changes in final assembly techniques won't occur until we construct a new plant. It's very difficult within the constraints and limitations of an existing plant to totally revise the processing. But in our planning for the next generation of assembly plants, certainly some different technologies will be applied. For example, we see ways now to increase the level of automation in our painting activities.

"Today our body shops are very heavily automated, the Ranger truck plant in Louisville applies 95 percent of the spot welds automatically, so it would be pretty hard to do much more automating there. But there are painting opportunities, and also in areas where we apply sealers manually. Opportunities in the trim and final assembly area, though, are limited at this point. Certainly in the next five to ten years, as the level of technology improves and we acqure robotic devices with greater accuracy and more flexibility than we have today, it will be possible to automate more and more of those operations." On the subject of assembly, GM's Behler adds, "Assembly philosophies will change with tne new alternatives provided by adhesives. I don't see any radically different assembly technologies 20 years out. There will simply be a lot more practical choices in material, manufacturing technique, and assembly. This will be quite exciting, I think, for our engineers in the future." Lasers measure up

Is there a future for lasers and other space-age technologies in the auto industry? "Oh, yes," says Ford's Guy. "We ahve a number of laser applications today and notice that laser technology is beginning to challenge the present domination of robotic articles in the trade press. We have 150 laser devices in our manufacturing facilities and development laboratories, ranging in application from heat treating to metalcutting to inspection. The dominant applications are low-energy lasers used for measurement and verification functions, and the high-energy CO2 and VAG lasers are used for metal-removal and heat-treatment applications. It's a growing technology we will use more and more."

Ralph Behler of GM feels laser technology will be used primarily as an inspection tool and as a heat-treat tool integrated into the machining process. "Lasers will be a factor in the factory of the future, and used in a variety of ways: bar coding, machining, machine vision, etc; the possibilities are almost limitless. We can see laser technology coming, and in a lot of ways, it's here." Software strategies needed now

One major problem area is in programmable automation--the total proliferation of software and the lack of any substantive standardization in the industry. "This creates for us a rather difficult problem in training skilled-trades and technical-support people in this myriad of control software," Ford's Guy admits. "We have urged many of our machine-tool and robotics suppliers to try to get together through some industry organization to begin some standardization, so that the various controllers we use on the plant floor can begin to communicate with one another without going through very expensive and elaborate software protocol converters.

"This is very difficult for us, and very difficult for any one machine-tool company to take on this task. But certainly the machine-tool industry, particularly those who support us, could be doing a lot more than they are."

GM's Behler agrees, "Getting international standards for communicating and protocol is going to be very important. General Motors is working on this with some computer companies. Adaptive scheduling systems will be needed to tie together plants, and there's a lot of work to be done, mostly in software. Manufacturing complexes must run on a real-time basis, and we must be able to decide where to locate all the data being generated, how much of it should be kept, and what to do with it. The needs here are much less obvious than the 2-dimensional hardware problems." New skills required

Computer inroads on the assembly line are demanding retraining of much of the existing maintenance force. Ultimately, the ratio of skilled to unskilled workers is expected to shift in favor of the former as assembly-plant work becomes more of a maintenance job.

On the one hand, skilled workers have traditionally been the most militant branch of the UAW, and the future seems to point to growing membership. Countering this is decentralized management that attempts to incorporate labor into the decision-making process, hopefully diffusing this militancy.

The GM Saturn project was the first time that a joint study group was set up with the UAW to get early input from workers on manufacturing planning. Recent contracts with the UAW strengthen job guarantees for its current work force, while joint union-management-training ventures proliferate, all aimed at softening the impact of technology on workers.

Paul Guy explains Ford's manufacturing automation philosophy. "There is a consistent efort throughout Ford to automate, whether by robot or other forms of programmable automation, particularly in the US. We are aggressively looking at ways to reduce the fatiguing, boring, and monotonous jobs that we have historically given our production operators. We want to use our human resources more for their mental capabilities than their muscle and brawn. This substitution yields more consistent results and better quality.

"Technology is certainly a big factor in our manufacturing strategies here at Ford. But the human-resource utilization factors is as important, if not even more so--how we manage people, how we utilize people, how to get them more committed to the company and its objectives, how we treat people--all of these factors are equally important in the manufacturing arena.

"This is an area receiving tremendous attention, particularly employee involvement, participative management, keeping trained and able people on the job, and the whole aspect of training. All the technology in the world (and this is part of the problem we've experienced), is of no use to us unless we have people who know how to apply it and maintain it. What's happened in the last five y ears is that there is so much new technology coming on stream so quickly that it has provided us with a tremendous challenge to bring not only our hourly and skilled-trades people, but our manufacturing-engineering people up to speed with what it is, what its capabilities are, and where it can be applied. We have a tremendous task right now of just taking what's available without creating any new technology, and putting it to use."

Ralph Behler of GM agrees. "No question about it, the people part, the upgrading of the labor force, will be extremely important. There is a continuing and large training job to be done. That includes everyone from engineers to skilled-trades people.

"The computer that is helping to create some of these problems may also have an important role in helping us with this training crises, once we figure out how to use it effectively to train people. I think this is a broad problem, and that educational approaches in universities everywhere will change.

"The things that are happening, and the pace of it all, will demand that we all become more responsive. Industry realizes they have a responsibility for training and retraining, and certainly GM intends to be very responsible in this area." But some feel that service stinks

Marvin Cetron, the prognosticator, is also an irate consumer. To him, training the manufacturing labor force may not be the key problem. "I hate like hell to say this, but our biggest problem is not in the General Motors plant, or in the Chrysler or Ford plants. Quality may be put in, as they say, before the product goes out, but it sure isn't kept up when the car gets out to the service people. They are simply not qualified to take that car back up to the same performance levels it was at when it left the plant.

"The whole service business--the people in the garages and appliance centers today--does not measure up to the machines they service. If you're fortunate, you get a good car in the beginning, but if you have to take it back and get things fixed, they won't know what to do except replace parts. If the problem is not that simple, you've got big trouble. A big help here eventually will be diagnostics built into the machine that tells the repair person. Take out this, put in that, and please don't touch anything else!"
COPYRIGHT 1984 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1984 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Insight: 2034
Publication:Tooling & Production
Date:Mar 1, 1984
Previous Article:Insight: 2034 - the next 50 years.
Next Article:Marketplace.

Related Articles
The next 50 years in metalworking.
Insight: 2034 - the next 50 years.
60 years of solutions.
The Unspoken Dialogue: Understanding Body Language and Controlling Interviews and Negotiations. (Reviews).

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters