Question 8 of 8: Computer Aided Process Planning, Part Coding and the Stages of a CAD System

Question 8: Computer Aided Process Planning, Part Coding and the Stages of a CAD System (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Part (a) — advantages of computer aided process planning. Process planning converts a design into the sequence of operations, machines, tools, fixtures, cutting conditions and times that will make it. Done manually it is slow, it is inconsistent between planners, and it embodies knowledge that leaves the firm when the planner retires. A CAPP system addresses each of those directly. Its first advantage is consistency: two similar parts receive similar routings, because the plan is generated from stored logic or from a retrieved standard plan rather than from an individual's habits, and the resulting standardization of tooling and methods reduces set-up variety across the shop. The second is productivity of the planning function, typically a several-fold reduction in planning effort and lead time, which matters most in job-shop and short-run work where planning cost is a real fraction of part cost. The third is rationalization: because the system either retrieves a plan for a part family (the variant approach, built on a group-technology code) or synthesizes one from machining knowledge and part features (the generative approach), it drives the firm toward fewer, better methods and exposes duplicate parts that could be eliminated. The fourth is integration, which is the point of the word in the syllabus title: the plan produced is a data structure, so it can feed the routing file for material requirements planning and capacity planning, the standard cost and estimating system, the tool-management system, the numerical-control programming system and the shop-floor documentation, all without re-keying. The fifth is currency and control: an engineering change is applied once, to the logic or the standard plan, and propagates to every affected part, with revision history retained. Sixth is better estimating, since consistent operation times taken from a standard-data file make quotations and scheduling more reliable. The honest limitations are the effort of building and maintaining the knowledge base, the dependence of the variant approach on a well-designed coding scheme, and the fact that generative planning remains practical only within restricted part families.

Part (b) — levels of part coding. Group technology assigns each part a code that captures its design and manufacturing attributes so that families can be recognized. Coding schemes are distinguished by their structure and by their depth. The three structures are the hierarchical or monocode, in which the meaning of each digit depends on the value of the digits before it, so the code branches like a classification tree; it packs a great deal of information into few digits and suits design retrieval, but it is harder to learn and to interrogate. The chain-type, attribute or polycode has digits in fixed positions each with a fixed meaning — digit 3 is always the material, digit 5 always the principal bore — which makes it easy to search on a single attribute and easy to teach, at the cost of a longer code. The hybrid or mixed code, used by most commercial systems such as Opitz and MICLASS, combines a short hierarchical form code with chain-type detail digits, and is the usual practical compromise. Depth of coding then forms a second, orthogonal dimension. At the shallowest level a code records only design attributes: basic shape class, principal dimensions and ratios, material, tolerance and surface finish, which is what a designer needs in order to retrieve an existing part instead of drawing a new one. A middle level adds manufacturing attributes: the major and minor operations required, machine type, work-holding, batch size and annual quantity, which is what a planner needs to form machine cells and retrieve routings. The deepest level extends to operational and logistical attributes — tooling, set-up and run times, inspection requirements, and in a fully integrated plant the identifiers carried on a bar code or radio-frequency tag that link the physical part to its record as it moves. The engineering judgement is to code deeply enough to serve the intended use and no deeper, since every additional digit must be assigned correctly for every part, forever.

Part (c) — the main stages of a computer aided design system. A CAD system supports four stages, conventionally described as the design process itself made computable. The first is geometric modelling: the designer builds a mathematical description of the part — wireframe, surface, or in modern practice a parametric solid model with constraints and a feature history — from which views, sections and mass properties follow automatically. The second is engineering analysis: the model is passed to analysis modules for mass property calculation, interference and tolerance checking, kinematic simulation, and finite element analysis of stress, deflection, vibration or heat flow, so that performance is predicted before anything is cut. The third is design review and evaluation: layering, zooming, sectioning, animated assembly checks and automatic interference detection let the design be examined against its requirements, and it is here that design for manufacture and assembly rules are applied and that the engineering change is formally reviewed. The fourth is automated drafting and documentation: dimensioned drawings, bills of material, tolerance schedules and the neutral files (IGES, STEP) or native data that downstream systems consume. In a computer integrated manufacturing environment a fifth stage is implicit and is the reason CAD appears in this syllabus at all — the model is the master record, and it is passed on to computer aided process planning, to numerical-control part programming, to fixture and tooling design, and to the coordinate measuring machine that will inspect the finished part, so that geometry entered once is used by every function that needs it.