Computer integrated manufacturing (CIM) for welded products

TWI Bulletin, November/December 1988

Dave Waller, BSc, MSc, CEng, MIProdE, FIM, FWeldI, is head of Manufacturing Systems Department at The Welding Institute.

Manufacturing is edging into a new industrial revolution as mechanisation and dedicated automation are being superseded by computer controlled machines and systems. This offers the possibility that computers can mutually communicate and collaborate so that the sub-systems for which they are responsible can operate as a harmonised total system. The technology and implementations of such computer collaboration for manufacturing are known as Computer Integrated Manufacturing (CIM). The purpose of this first of a series of Bulletin articles is to give an initial appreciation, of CIM and its implications, to the non-CIM specialist with particular emphasis on CIM for welded products. Subsequent Bulletin articles will elaborate on some of the major CIM ingredients.

The potential benefits to the producer and consumer to be derived when CIM systems are installed will be enormous. Complete accomplishment of a total CIM system is that of the fully automated factory.

Of course such a goal as a fully automated factory, covering all operations between the reception of a purchase order to the packed product is some years from realisation. This is mainly because the proliferation of choice offered by computer systems vendors has led to a situation where machines and sub-systems, procured from different vendors mostly stand alone without the compatibility necessary to communicate with their peers or their hierarchy. The International Standards Organisation has now established a seven layer model (for functions and interfaces) which is being further developed for Open Systems Inter-connection (OSI). A well known implementation standard, coupling with the OSI model, is the Manufacturing Automation Protocol (MAP) initiated by General Motors.

The adoption of CIM techniques is not just the province of the very large organisation. Small and medium size enterprises can equally apply CIM technology and profit from the benefits.

Scope of CIM

The scope of CIM is a matter of choice for the manufacturer and can extend to take into account the whole business. The manufacturing activities normally considered are indicated in the Fig.

Steps for CIM implementation

The topics which have to be addressed to create an environment for CIM implementations are:

• An awareness of CIM, its implications and benefits;
• Responsibility for equipment definition;
• The development of a CIM strategy for the manufacturing organisation which includes:
  • A CIM system architectural model for the organisation;
  • An investment and procurement strategy;
  • An implementation plan.
• How to specify equipment and automation for CIM;
• Testing ideas before commitment using simulation techniques and test benches;
• Implementation and commissioning of CIM systems and automation;
• System performance monitoring.

CIM terminology

As the manufacturing systems specialists have grouped around CIM and established it as a technological discipline, jargon and acronyms have become part of the ethos. There is no deliberate intention to expose the reader to such 'Information Technology (IT) speak' so a glossary of terms is provided.

Retrospective view

Before going forward it is worth looking back at manufacturing system developments so far this century to look for bench marks and trends which might lead us into the future.

The consumer product which provokes more opinion and passion from the user than any other is the motor car. The competitiveness of the industry has led to major advancements of manufacturing techniques. Some car manufacturers are truly innovative and jump in with new manufacturing ideas at each model change. Some are less adventurous and are content to follow the trend of their competitors. The nett effect is a ratcheting advance of manufacturing technology.

Henry Ford I with his historic invention of the moving assembly track was an important trend setter. His simple act of ganging together consecutive assembly jig trucks essentially transferred the initiative for operation completion from the production direct operator to the track supervisor. This transfer of pace setting responsibility from worker to first line manager is a trend which will continue as an essential ingredient of CIM. However, for CIM the transfer of responsibility for pace setting and for decision making is from man to machine; that is the production schedule will be established, managed and driven by computers. The interim situation where many humans remain integrated in the computer controlled manufacturing system is a key aspect of any CIM development strategy. Human activities which need to be eventually transferred to machines, including computers and expert systems, are:

• Physical 'doing' activities;
• Intellectual activities;
• Process knowledge;
• Decision making;
• Information handling.

Until these human attributes are transferred, the human remains the driver of the manufacturing sub-system (i.e. the total system). An aspect, therefore, of CIM implementation is the need for the strict definition of the data interfaces between man/machine and machine/machine. A secondary benefit of this formality is that ongoing production is not disrupted as new machines are introduced and so releasing man for more creative roles.

Since World War II, assembly line manufacturing has been typified by demanding a production operator to repeat the same physical and mental activities at a frequency of less than one a minute for an entire shift. These syndrome inducing production operations which are still a feature of traditionally founded mass production manufacturing industry are gradually being phased out of the system by precision mechanisms and automation. This development phase has endured 40 years and continues into the CIM era. The chronological stages of these developments in the car industry are:

• Mechanised fixtures replacing manually operated fixtures;
• Automatic transfer lines for specific major sub-assemblies such as car doors using dedicated tooling and hard wired controllers;
• Transfer lines with interchangeable tooling;
• Automatic material handling for loading and unloading of components and assemblies into and out of tooling;
• Reprogrammable controllers.

The above development steps essentially represent the transfer of physical, decision making and some elementary information handling tasks from man to machine.

Welding robots and CIM

Welding robots are a major enabling tool and together with other programmable machines represent the final line of command of the total system.

Robots were first used for car assembly in the early 1960s in a welding role. Strictly speaking the robot was used as a programmable tool handler to position a resistance spot welding gun at its point destination. Robots are now proliferating on car assembly lines and are used for a variety of operations including pick-and-place material handling, mechanical assembly, paint spraying, sealing, spot welding and arc welding.

Most welding robots in such applications are statically positioned alongside the track and rely on the assumption that the component and weld joint are always moved into the same relative spatial position by the indexing assembly conveyor. In the early stages of robot development, the robot precision was inferior to the product dimensional tolerance. This led to improved pressed panel and assembly fixturing quality. It should be noted that once the components' geometric precision falls within the robot's performance tolerance then no intelligence is required of the robot.

The long production run of a particular car model dictates that robots only need reprogramming for every model change (which could be two or more years). Used in this way robots are essentially units of dedicated automation. As product variants are normally scheduled down the same assembly tracks each robot is either supplied with a set of programs (one for each variant) or instructed not to perform for certain product variants. Gangs of robots are co-ordinated and synchronised by line controllers.

For arc welding and for small batch quantity production the difficulties for CIM implementation are currently at their most severe. The reasons for this are:

• The process tolerances and their mutual relationships are not fully developed for different joint geometries.
• Programming a robot, even using the latest off-line preprogramming techniques, demands high cost labour and is inhibiting for small batch production runs, particularly of new products.
• Arc welding of non-fixturable products often deforms the seam path out of the tolerance demanded of the arc process. This places a demand for local robotic intelligence.

Each of these automation points is currently being tackled by research programmes for developing expert systems and sensor driven artificial intelligence of welding robots. But such difficulties should not inhibit the commitment for a manufacturing organisation to develop its CIM strategy and start to work towards CIM implementations.

Notwithstanding the important productivity gains of these 'bottom up' automation advances, the way forward is to analyse and define the manufacturing system starting with the company objectives (top down analyses).

The first stage is to assess by a system analysis technique, the current system. The second stage is to identify the future company business needs and objectives and again using a formal analysis technique, define the 'to be' system. Both of these analyses are known as top down analyses as they stem from the company global objectives.

The third stage is to see how existing sub-systems align with 'to be' system needs. Systems which are procured to satisfy local productivity gains often have very weak lateral connections to other sub-systems.

The fourth stage is that of defining and designing new sub-system requirements by using the top down analyses to derive a specification for new software and hardware. These analyses also enable comparative productivity gains to be assessed between competing candidates for investment.

The fifth stage is that of designing and specifying a communications network which will be the final integration stage, i.e. the I in CIM. The communications network is simultaneously the enabling and limiting technology of a CIM system. Once installed, all parts of the system must communicate through it and must also be capable of extension/enhancement to meet the lifetime needs of the system. Therefore selection of a communications network requires the utmost care.

Benefits

The implications and benefits of a business capable of not only automatically managing, controlling and monitoring a complex ongoing manufacturing system, but of instantaneously responding to change, cannot be overstated. The ultimate goal is for a system to be dynamically responsive to:

• Product design changes;
• New product launch and phase out;
• Production schedule changes;
• Production mix changes;
• Eventualities such as production bottle necks, breakdowns.

The ultimate benefits will be:

• Reduced manufacturing costs, lower product cost, faster throughput, detachment from direct and indirect labour inflation costs;
• Better system flexibility;
• More consistent product quality.

The problems which currently represent a barrier against CIM introduction are:

• Human skill and expertise are necessary at all levels of the manufacturing system, for management, planning, process knowledge and application. These tasks have to be defined and transferred to expert systems in order to transferthe systems' pacemaking from man to machine.
• The incompatibility of computer systems from different vendors.

Risks

So far the benefits that CIM will eventually yield have been stressed but it is important to realise the risks attached to ad hoc CIM implementations.

Firstly, the question of compatibility of new equipment, both pre-production and production, needs to be specified at each new procurement opportunity. Secondly, actual computer integration implementations can be very costly and if a CIM strategy and design have not been established for an organisation, this can lead to the possibility of investing in the wrong areas of the business. On the other hand an investment which is directed at the parts of the system where high value is added by human labour and where considerable activity time is expended, can yield the desired return. The risk of ignoring Computer Integrated Manufacturing is that non-participants will not be able to remain competitive.

Glossary of acronyms