Question 4 of 7: CAD, Process Planning, CAPP, Routing Sheets and Group Technology

Question 4: CAD, Process Planning, CAPP, Routing Sheets and Group Technology (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) — CAD against the drawing board. The decisive advantage of computer-aided design is that the design exists as a database rather than as a picture. Because the geometry is stored numerically, it can be measured, analysed, transformed and transmitted without being redrawn. Concretely, this yields: accuracy limited by the model rather than by pencil width; instant modification and revision control, so a change propagates to every view and to the bill of materials; reuse of standard parts and features from libraries; automatic generation of section views, assemblies and exploded illustrations; interference and clearance checking on assemblies; direct coupling to finite-element analysis, kinematic simulation, tolerance analysis and mass-property calculation; and direct coupling to manufacturing, where the same model drives NC tool-path generation, inspection programs and rapid prototyping. Documentation and archiving improve, distributed teams can work on one model, and the design-to-production interval shortens substantially.

The limitations are real and worth stating. The capital cost of software, hardware and above all training is high, and productivity typically falls during the first months of implementation. A CAD system does not design: it records decisions, and a poor concept modelled beautifully is still a poor concept. The apparent precision of a screen model can breed false confidence, particularly in tolerance and fit questions where the model shows nominal geometry only. Long-term data archiving is a genuine problem because file formats and the software that reads them have shorter lives than the products they describe, and translation between systems loses information. There are security and intellectual-property exposures that a locked drawing cabinet did not have. Finally, ease of modification encourages proliferation of near-identical parts unless a classification discipline is imposed — which is one motive for group technology in part (e).

Part (b) — the purposes of process planning, and the computer's role. Process planning determines how a designed part will be made. Its purposes are to select the sequence of processes and operations that will produce the specified geometry, tolerances and surface finish; to assign each operation to a machine or work centre; to specify the tooling, fixtures, gauges and cutting parameters; to establish standard times for each operation, which become the basis of scheduling, capacity planning and standard costing; to determine the raw material form and size; to identify the inspection points and methods; and to do all of this at the lowest cost consistent with the design intent. It is also the point at which manufacturability problems are detected and fed back to design, so process planning is a control on design as well as a consequence of it.

Computers enter at every one of those points. They store and retrieve standard plans and machining data; they hold the machine, tool and material databases that a plan must reference; they apply decision logic to select processes and sequence them; they compute speeds, feeds and standard times from formulas rather than from tables read by eye; they generate the route sheet and the operation sheets as formatted documents; and they pass the resulting data onward to MRP, scheduling, costing and NC programming without re-keying. The two architectures that do the plan generation itself are the subject of part (c).

Part (c) — two types of CAPP system. The variant or retrieval type stores a standard process plan for each part family. A new part is coded, usually with a group-technology code; the code identifies the family; the family's standard plan is retrieved and displayed; a planner edits it to suit the specific part. Its features are a family classification scheme, a library of standard plans, retrieval logic and an editing environment. It is inexpensive to implement, quick to bring into use, and effective wherever parts genuinely cluster into families; it cannot plan a part unlike anything it has seen, it perpetuates whatever was in the standard plan, and it still requires a planner.

The generative type synthesises the plan. It reads a feature description of the part — holes, slots, pockets, faces, with their tolerances, finishes and material — and applies encoded decision logic, typically decision trees, decision tables or expert-system rules, to choose the processes capable of each feature, sequence them under precedence rules, assign machines, select tools, and compute parameters and times. Its features are a feature-recognition or feature-based input, a process-capability knowledge base, an inference mechanism and an optimisation criterion. It can plan a new part with little or no human intervention and can optimise rather than imitate; it is very expensive to develop and validate, it needs a rich feature model, and its knowledge base must be maintained continuously as the shop's equipment changes. Most commercial systems are semi-generative hybrids that generate the sequence and retrieve the details.

Part (d) — the routing sheet. The routing sheet (route sheet, operation sheet, traveller) is the primary output of process planning and the document that accompanies the order through the shop. A typical sheet carries: part number, name, revision level and the drawing reference; material specification, raw stock form and size, and the quantity in the order; the order number and due date; then, one line per operation in sequence, the operation number, a description of the operation, the work centre or machine on which it is performed, the tooling, fixtures and gauges required, the set-up time and the run time per piece, and the inspection or quality requirement at that step. It usually also carries the sign-off spaces where each operator records completion, quantity good and quantity scrapped, and the department transfer points.

It is necessary for four reasons. It is the instruction that tells the shop what to do and in what order, so that the same part is made the same way every time. It is the authorisation and routing document that moves the job physically from work centre to work centre and prevents operations being skipped or done out of sequence. It is the data source for the rest of the manufacturing system: the standard times on the routing sheet drive capacity planning, scheduling, load levelling and standard costing, and MRP explodes lead times from it. And it is the record: signed off operation by operation it becomes the traceability document showing what was made, on which machine, by whom, with what yield — which is what a quality system such as ISO 9001 requires and what a failure investigation begins from. Without it, routing lives in the memory of whoever moves the tote, and nothing downstream can be planned.

Part (e) — group technology. Group technology is the manufacturing philosophy that identifies and exploits similarities among parts. Parts are classified by their design attributes (shape, size, material, tolerance) and their manufacturing attributes (processes required, machine sequence, tooling, batch size) and grouped into families whose members can be made by the same or a similar sequence of operations. Classification is usually formalised in a coding scheme — monocode (hierarchical), polycode (chain-type, each digit independent) or a hybrid such as the Opitz code — or is derived from production-flow analysis of existing routings.

It was developed because batch manufacturing is intrinsically inefficient. Mid-twentieth-century studies found that in a typical job shop a part spends about ninety-five per cent of its throughput time waiting or moving and only a few per cent on a machine, that shops carried thousands of part numbers of which large numbers were near-duplicates, and that process-type layouts — all the lathes here, all the mills there — forced every job onto a long, criss-crossing path. Group technology was the response: if parts can be grouped into families, then the machines that make a family can be grouped into a cell, and the batch shop acquires some of the flow characteristics of a production line without giving up its variety.

The advantages follow directly. In design: retrieval of an existing similar part prevents the creation of yet another near-duplicate, standardises features and tolerances, and shortens design time — a designer can often find that eighty per cent of a "new" part already exists. In process planning: family standard plans make variant CAPP possible at all, and planning time falls sharply. In manufacturing: cellular layout shortens material travel, cuts set-up times because successive parts in a family need similar fixtures and tools, reduces work in process and throughput time, allows smaller lots and therefore lower inventory, simplifies scheduling to a cell rather than a shop, and gives the cell team clear ownership of quality. In tooling and purchasing: standardisation reduces the number of distinct tools, fixtures and raw stock sizes carried. In estimating: costs for a new part can be inferred from the family. The costs to weigh against these are the effort of classifying an existing part population, the capital tied up in duplicating machines across cells, and the loss of flexibility when demand for one family collapses.