CIM and the production engineer

TWI Bulletin, March/April 1989

All enterprises make investments - in both manufacturing facilities and people. All enterprises expect a good return on their investments.

In pursuing the goal of maximising return on investment, enterprises of all sizes, from all sectors of industry, accept that careful planning is vital to success. Process planning - a fundamental part of production engineering, is no exception to this. There are now many computer aided process planning (CAPP) packages available to ease and improve this task.

This article, the third in a series, briefly explores the concepts and aims of process planning, the various CAPP systems available and how these relate to the overall concept of CIM (computer integrated manufacturing).

Definitions

Glossary

What is a process plan

A process plan ( Fig. 1) is an authoritative statement of 'how' a product shall be made. In its simplest form it comprises two elements, namely:

1. The routing sheet--which specifies the sequence in which various operations (e.g. cut, weld, paint) are to be performed.
2. The operation sheet(s) - which gives precise details of how any particular operation is to be performed (e.g. working to procedure No.123 and using fixture ref. JT001, weld joint A).

Included within these documents will be details of all jigs, tools, fixtures and machines required; process parameters (e.g. welding procedures, cutting speeds, etc) and NC/CNC programs required. All documentation carries 'identifiers' which tie it unambiguously to the product in question, and in some cases job costing data (e.g. labour grades, allowed time/piece, etc) may be included.

The degree of detail included in a process plan is generally a function of the stability and skill of the labour force (i.e. where labour turnover is high and/or skill level low, it is necessary to specify requirements in great depth). However, exceptional circumstances (e.g. expansion necessitating a rapid build-up in labour force, abnormal loss of experienced staff, etc) may also lead to a demand for more detail (assuming such plans exist in the first instance!).

Why plan manufacturing processes?

Manufacturing processes, such as may be found in any fabrication shop, are often unplanned and undocumented (except for weld procedures); they are left to the discretion of the shop supervisor or even the individual operator and are, therefore, variable. With diminishing levels of skilled craftsmen, ever increasing product complexity and greater demands being made for product consistency and quality, this informal approach is no longer acceptable. It has become necessary to plan the manufacturing process in order to:

• Optimise the use of the work force (time and skills);
• Provide a basis for the management of production;
• Ensure product uniformity through process consistency;
• Provide the basis for formal product quality assurance.

Design for manufacture

A fundamental requirement of any process plan is that the product be 'manufacturable' - plans cannot compensate for 'impossible to make' designs. Conceivably, the action of generating a process plan may identify 'impossible to make' designs before they reach the production floor, but even this is too late to avoid the expense of re-design.

The competent designer has long recognised the need to rationalise product design as the first move towards ensuring that it is manufacturable. The two most important steps in this process are generally accepted ^[2] as:

• Critical study of the product design to remove any of its attributes which may create manufacturing difficulties;
• Assessment of alternative manufacturing methods and technologies.

These steps are usually undertaken in conjunction with manufacturing, the representative of whom is the production/welding engineer. Of primary concern, is the need to ensure optimal production results (i.e. maximum output and quality at minimum cost); to this end it is vital that the product is designed to take full advantage of existing and future manufacturing processes.

Process planning systems

Process planning systems ( Fig.2) have, until recently, relied upon manual techniques to produce, issue and maintain their documentation. Modern computing systems offer considerable opportunities to ease the production engineer's clerical burden, improve the consistency of the plans and eliminate human errors. The reduction in clerical effort, which some sources ^[3] suggest is as high as 45%, frees the production engineer for more demanding and/or fruitful tasks (e.g. process development, tool design, facilities planning, etc). Moreover, given that the CAPP system may be able to communicate with the product database, there is opportunity for automated generation of process plans which offers another step towards CIM.

There are three recognised types of CAPP:

Variant (e.g. Supercapes from Scicon);
Generative (e.g. Locam from Pafec);
Constructive (e.g. C-plan from CAD Centre).

Variant process planning systems

These make use of the power of computers as very efficient filing systems. They enforce the use of standardised company documentation and allow the storage of large amounts of data concerning existing products and processes. This data can be easily retrieved when similar products are being considered; having studied the 'similar' product the production engineer is able to create a new plan by amending the existing data to account for the characteristics of the new product. The newly created plan can be stored in the same way as that from which it was created. Such systems perform best in 'similar to' situations, and are limited in so far as the quality of plan produced is dependent upon the engineer's knowledge and experience. Systems of this type have been used successfully for a variety of applications from machine shops to sheet metal and presswork shops. Their modest installation costs (£10-50K) and ease of installation make them an attractive proposition for a company seeking a first time implementation.

Generative process planning systems

These attempt to capture the fundamental knowledge of the production engineer and use this captured knowledge to generate fresh process plans for each application. Much of this knowledge exists as standard data concerning the various operations in the process (e.g. metal removal rates, tool wear algorithms, etc.), which is manipulated by various computer programs in the same way as a production engineer would. They make no attempt to use any part of the plans produced for similar products, but rather use the power of the computer to emulate human thinking to create individual plans.

These are often referred to as knowledge based systems and have two fundamental difficulties to overcome:

• That of capturing, accurately, the production engineer's knowledge and representing this as system rules (usually referred to as knowledge engineering).
• That of presenting product data to, and extracting it from, the system. To a human, a simple sketch usually conveys clearly and unambiguously all that is required: regrettably generative CAPP systems do not accept sketches. Thus to be truly effective these systems have to be developed so that they work directly with CAD data; this is a major integration problem.

Systems of this type range, in cost, from £50-500K, and may take from 1-5 years to install fully; not surprisingly there are few true examples in operation. However, it is this area which is most likely to offer the means for truly automatic generation of process plans, leading to integration of CAPP and CAD. ^[4] It is for this reason that TWI has recently embarked on a study of the applicability of commercial knowledge based systems (expert system shells) to this task.

The best examples of this system are to be found in the metal cutting industries; however there is some indication that they are starting to be used in high quality fabrication areas.

Constructive process planning systems

Systems of this type are a combination of the variant and generative systems. They combine the powerful databanks of existing plans - which may be accessed and manipulated at will - with the ability to 'generate' items such as speeds and feeds from preprogrammed standard data. Whilst ultimately the quality of plans produced depends upon the production engineer, these systems offer more help to him than do conventional variant systems, thus endowing more consistency, better quality and greater speed.

With installation costs ranging from £30-100K, these systems represent a sound investment for serious users and have found favour in small to medium batch, high quality machining and assembly operations.

CAPP/CAPE/CAM

These three terms are regularly confused, misquoted and misunderstood. CAPP has been explained, so a brief explanation of CAPE and CAM may prove useful.

CAPE is generally used to cover the area of industrial computing which has specialised in developing systems for the production of NC and CNC machine programs directly from the CAD database of the part concerned;

CAM is usually accepted as a combination of CAPP, CAPE and any other computer system applied for the benefit of manufacturing (e.g. material requirements planning).

The majority of packages within existing CAM systems are of the stand alone variety; they do not represent CIM. Organisations which have achieved integration of the various CAPP, CAPE and other packages have done so by one of three approaches:

• Using a turnkey system from one of the larger computer system suppliers or software houses;
• Designing, producing and implementing their own packages and interfacing software;
• Using stand alone packages which communicate via custom designed interfaces (software and hardware), or a fourth generation language (4GL).

Whatever technique is used, two requirements must be satisfied:

• The current situation must be carefully analysed and future requirements carefully planned and specified;
• It should use a product database as the hub of the system.

Analysis techniques, such as SADT (structured analysis and design technique), are well understood, ^[5] and have been dealt with by other articles in this series. ^[6] They provide a means for undertaking a thorough 'top-down analysis' of needs from which to develop plans of future requirements. Database construction and management techniques are likewise well understood and documented. ^[7,8] However, the implications of having all the information (sales, design and manufacturing) relating to a product in one product database are perhaps less well understood, and therefore worthy of further consideration.

The product database

Traditionally the product, as defined by the CAD system, has been distributed to various parts of the organisation in the form of drawings and bills of materials (BOMs). Not only is such documentation wasteful in terms of time and paper, but it is also extremely dangerous. Once received by an end user the documentation is regarded as sacrosanct, even though by the time it is received it may already be out of date. Furthermore, as times goes by this (possibly out of date) information forms the base for further information relating to such things as sales orders, manufacturing processes, costs and material, all of which is filed away in the records of the various originating departments.

The concept of a product database ( Fig.3) requires that the product design forms the core of a homogeneous database, to which is added all of that information currently held within the records of such departments as production engineering, manufacturing, purchasing, quality assurance, estimating, etc. Each end user department has access to this database at all times and therefore has true knowledge of the current situation.

The problems of putting all of the company information into a single database (which may for sheer capacity reasons be physically distributed) are considerable (e.g. computing, data integrity and database management). The need for politically unbiased management of this database is no less daunting; however, the potential benefits justify the effort. Key benefits are:

• Totally up-to-date information instantaneously available;
• Vastly reduced storage requirements for documentation;
• The potential for total electronic data exchange, that is to say a major step along the road to computer integrated manufacturing.

What next?

Any enterprise embarking on CIM will need to study its current as is status and plan, with extreme care, its intended to be status in respect of the functions, data transfer systems and business strategy needed. CIM systems require very close ties between any departments which have interacting systems. This requirement has led some organisations ^[3] to re-think their whole structure. In order to accent the need for a unified approach, they have created small business teams to focus on products; these teams consist of members of all disciplines who are brought together (physically) with the product and its database as the central reference, the aim being to achieve much better cross fertilisation and a common sense of purpose.

The role of the production engineer in all this activity will vary with enterprise; however, as process planning for which he is responsible, stands at a major 'information junction' - between design (CAD) and manufacturing (CAM), he will ultimately have an important part to play.

Process planning may not be considered 'first priority' but it will acquire increasing significance as the CIM goal approaches.

References