MANUFACTURING PROCESS PLANNING AND DESIGN:COMPUTER-AIDED PROCESS PLANNING

COMPUTER-AIDED PROCESS PLANNING

Process planning has traditionally been experience based and performed manually. A problem facing modern industry is the lack of a skilled labor force to produce machined parts as in the past. Manual process planning also has other problems. Variability among the planners’ judgment and experience can lead to differences in the perception of what constitutes the optimal or best method of production. This manifests itself in the fact that most industries have several different process plans for the same part, which leads to inconsistent plans and an additional amount of paperwork. To alleviate this problem, a computer-aided approach can be taken. Development in computer-aided process planning attempts to free the process planner from the planning process. Computer-aided process planning can eliminate many of the decisions required during planning. It has the following advantages:

• It reduces the demand on the skilled planner.

• It reduces the process-planning time.

• It reduces both process-planning and manufacturing cost.

• It creates consistent plans.

• It produces accurate plans.

• It increases productivity.

The benefits of computer-aided process planning systems have been documented in several in- dustries. Such systems can reduce planning time from days to hours and result in large cost savings.

The idea of using computers in the process planning activity was discussed by Niebel (1965). Other early investigations on the feasibility of automated process planning can be found in Scheck (1966) and Berra and Barash (1968). Many industries also started research efforts in this direction in the late 1960s and early 1970s. Early attempts to automate process planning consisted primarily of building computer-assisted systems for report generation, storage, and retrieval of plans. A database system with a standard form editor is what many early systems encompassed. Formatting of plans was performed automatically by the system. Process planners simply filled in the details. The storage and retrieval of plans are based on part number, part name, or project ID. When used effectively, these systems can save up to 40% of a process planner’s time. A typical example can be found in Lockheed’s CAP system (1981). An example of a modern version is Pro / Process for Manufacturing (launched in 1996 and since discontinued). Such a system can by no means perform the process- planning tasks; rather, it helps reduce the clerical work required of the process planner.

The typical organization of using a process-planning system is shown in Figure 17. A human planner interprets an engineering drawing and translates it into the input data format for a process- planning system. Either interactively or automatically, a process plan is produced. The plan is then used by production planners for scheduling of production and used by industrial engineers to lay out the manufacturing cell and calculate production cost and time. A part programmer follows the in- structions on the process plan and the engineering drawing to prepare NC (numerical control) part programs. The same organization applies to all kinds of process planning systems.

Perhaps the best-known automated process planning system is the CAM-I automated process planning system (CAPP) (Link 1976). (CAM-I stands for ComputerAided Manufacturing Interna- tional, a nonprofit industrial research organization.) In CAPP, previously prepared process plans are stored in a database. When a new component is planned, a process plan for a similar component is retrieved and subsequently modified by a process planner to satisfy special requirements. The tech-

nique involved is called group technology (GT)-based variant planning (Burbidge 1975). Variant planning will be discussed in more detail in the next section.

Figure 18 represents the structure of a complete computer-aided process-planning system. Al- though no existing turnkey system integrates all of the functions shown in the figure (or even a goodly portion of them), it illustrates the functional dependencies of a complete process-planning system. It also helps to illustrate some of the constraints imposed on a process-planning system (e.g., available machines, tooling, and jigs).

In Figure 18, the modules are not necessarily arranged based on importance or decision sequence. The system monitor controls the execution sequence of the individual modules. Each module may require execution several times in order to obtain an optimum process plan. Iterations are required to reach feasibility as well as good economic balance.

The input to the system will most probably be a 3D model from a CAD database. The model contains not only the shape and dimensioning information, but also the tolerances and special features. The process plan can be routed directly to the production-planning system and production-control system. Time estimates and resource requirements can be sent to the production-planning system for scheduling. The part program, cutter location (CL) file, and material-handling control program can also be sent to the control system.

Process planning is the critical bridge between design and manufacturing. Design information can be translated into manufacturing language only through process planning. Today, both automated design (CAD) and manufacturing (CAM) have been implemented. Integrating, or bridging, these functions requires automated process planning as the key component.

There are two basic approaches to computer-aided process planning: variant and generative. The variant approach is used by the computer to retrieve plans for similar components using table look- up procedures. The process planner then edits the plan to create a variant to suit the specific require- ments of the component being planned. Creation and modification of standard plans are the process planner’s responsibility. The generative approach is based on generating a plan for each component without referring to existing plans. Generative systems are systems that perform many of the functions in a generative manner. The remaining functions are performed with the use of humans in the planning loop.

Variant Approach

The variant approach to process planning was the first approach used to computerize planning tech- niques. It is based on the idea that similar parts will have similar process plans. The computer can

be used as a tool to assist in the identification of similar plans, retrieving them and editing the plans to suit the requirements for specific parts.

A variant process planning system includes the following functions:

• Family formation

• Standard plan preparation

• Plan editing

• Databases

In order to implement such a concept, GT-based part coding and classification are used as a foundation. Individual parts are coded based upon several characteristics and attributes. Part families are created of ‘‘like’’ parts having sufficiently common attributes to group them into a family. This family formation is determined by analyzing the codes of the part spectrum. A ‘‘standard’’ plan consisting of a process plan to manufacture the entire family is created and stored for each part family. The development of a variant process-planning system has two stages: the preparatory stage and the production stage.

During the preparatory stage, existing components are coded, classified, and later grouped into families (Figure 19). The part family formation can be performed in several ways. Families can be formed based on geometric shapes or process similarities. Several methods can be used to form these groupings. A simple approach would be to compare the similarity of the part’s code with other part codes. Since similar parts will have similar code characteristics, a logic that compares part of the code or the entire code can be used to determine similarity between parts.

Families can often be described by a set of family matrices. Each family has a binary matrix with a column for each digit in the code and a row for each value a code digit can have. A nonzero entry in the matrix indicates that the particular digit can have the value of that row. For example, entry (3,2) equals one implies that a code x3xxx can be a member of the family. Since the processes of all family members are similar, a standard plan can be assigned to the family.

The standard plan is structured and stored in a coded manner using operation codes (OP codes). An operation code represents a series of operations on one machine / workstation. For example, an OP code DRL10 may represent the sequence center drill, change drill, drill hole, change to reamer, and ream hole. A series of OP codes constitutes the representation of the standard process plan.

Before the system can be of any use, coding, classification, family formation, and standard plan preparation must be completed. The effectiveness and performance of the variant process-planning system depends to a very large extent on the effort put forth at this stage. The preparatory stage is a very time-consuming process.

The production stage occurs when the system is ready for production. New components can be planned in this stage. An incoming component is first coded. The code is then sent to a part family search routine to find the family to which it belongs. Because the standard plan is indexed by the family number, the standard plan can be easily retrieved from the database. Because the standard

plan is designed for the entire family rather than for a specific component, editing the plan is una- voidable.

Variant process-planning systems are relatively easy to build. However, several problems are associated with them:

• The components to be planned are limited to similar components previously planned.

• Experienced process planners are still required to modify the standard plan for the specific component.

• Details of the plan cannot be generated.

• Variant planning cannot be used in an entirely automated manufacturing system, without ad- ditional process planning.

Despite these problems, the variant approach is still an effective method, especially when the primary objective is to improve the current practice of process planning. In most batch manufacturing industries, where similar components are produced repetitively, a variant system can improve the planning efficiency dramatically. Some other advantages of variant process planning are:

• Once a standard plan has been written, a variety of components can be planned.

• Comparatively simple programming and installation (compared with generative systems) is re- quired to implement a planning system.

• The system is understandable and the planner has control of the final plan.

• It is easy to learn and easy to use.

The variant approach is the most popular approach in industry today. Most working systems are of this type, such as CAPP of CAM-I (Link 1976) and Multiplan of OIR (OIR 1983).

Generative Approach

Generative process planning is the second type of computer-aided process planning. It can be con- cisely defined as a system that automatically synthesizes a process plan for a new component. The generative approach envisions the creation of a process plan from information available in a manu- facturing database without human intervention. Upon receiving the design model, the system is able to generate the required operations and operation sequence for the component.

A generative process-planning system consists of the following important functions:

• Design representation

• Feature recognition

• Knowledge representation

• System structures

Knowledge of manufacturing has to be captured and encoded into computer programs. A process planner’s decision-making process can be imitated by applying decision logic. Other planning func- tions, such as machine selection, tool selection, and process optimization, can also be automated using generative planning techniques.

A generative process-planning system contains three main components:

• Part description

• Manufacturing databases

• Decision-making logic and algorithms

The definition of generative process planning used in industry today is somewhat relaxed. Thus, systems that contain some decision-making capability in process selection are called generative sys- tems. Some of the so-called generative systems use a decision tree to retrieve a standard plan. Generative process planning is regarded as more advanced than variant process planning. Ideally, a generative process-planning system is a turnkey system with all the decision logic built in. Since this is still far from being realized, generative systems currently developed provide a wide range of capabilities and can at best be described as only semigenerative.

The generative process-planning approach has the following advantages:

• It generates consistent process plans rapidly.

• New components can be planned as easily as existing components.

• It has potential for integrating with an automated manufacturing facility to provide detailed control information.

Successful implementation of this approach requires the following key developments:

• The logic of process planning must be identified and captured.

• The part to be produced must be clearly and precisely defined in a computer-compatible format.

• The captured logic of process planning and the part-description data must be incorporated into a unified manufacturing database.

Part-Description Methods for Generative Process-Planning Systems

Part description forms a major part of the information needed for process planning. The way in which the part description is input into the process-planning system has a direct effect on the degree of automation that can be achieved. Since the aim is to automate the system, the part description should be in a computer-readable format. Traditionally, engineering drawings have been used to convey part descriptions and communicate between design and manufacturing. Understanding the engineering drawing was a task suited for well-trained human beings and initially not suitable for direct input for process planning. The requirements of the part-description methods include:

• Geometrical information

• Part shape

• Design features

• Technological information

• Tolerances

• Surface quality (surface finish, surface integrity)

• Special manufacturing notes

• Feature information

• Manufacturing features (e.g., slots, holes, pockets)

Before the representation method is decided, the following factors have to be determined:

• Data format required

• Ease of use for planning

• Interface with other functions, such as part programming and design

• Easy recognition of manufacturing features

• Easy extraction of planning information from the representation

Some of the representations used in a generative process-planning system include: GT code, line drawing, special language, symbolic representation, solid model, CSG, B-Rep, feature-based model. Extract and decompose features from a geometric model.

• Syntactic pattern recognition

• State transition diagram and automata

• Decomposition

• Logic

• Graph matching

• Face growing