22-Mec-B4 Integrated Manufacturing Systems · December 2016
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Paper format. 07-Mec-B4 — Integrated Manufacturing Systems, National Exams December 2016. Three hours, open book, any non-communicating calculator permitted. Six questions are printed; any five constitute a complete paper and all questions are of equal value, so each is worth 20 marks on a five-question basis. Only the first five questions appearing in the answer book are marked. All six are solved here.
Reference texts. The paper draws on the operations and facilities side of manufacturing engineering rather than on process metal cutting, so the useful shelf is:
Canadian practice is assumed throughout: handling and lifting design is governed by the applicable provincial occupational health and safety regulation and by CSA standards (for example CSA B335 for lift trucks), and quality records are kept to satisfy ISO 9001 as adopted by CSA.
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) — a situation favouring centralized inspection. Centralized inspection means that work is brought to a separate inspection department, staffed by inspectors who report outside production, rather than being examined at or beside the machine that made it. The situation in which it is clearly preferable is a low-to-medium volume precision job shop making close-tolerance machined components for a regulated customer — aerospace structural fittings, hydraulic valve bodies, or bearing races — where final acceptance depends on measurements that cannot honestly be made on the shop floor.
Several features of that situation point the same way. The measuring equipment is expensive and delicate: a coordinate measuring machine, a roundness tester, an optical comparator and a surface-finish instrument together represent a capital cost that can only be justified once, and their accuracy depends on a controlled environment, since a temperature excursion of a few degrees moves a steel dimension by more than the tolerance being verified. That argues for one temperature-controlled metrology room rather than instruments scattered along a shop floor full of coolant mist and vibration. The measurements also demand specialised skill in fixturing, datum interpretation and geometric dimensioning and tolerancing, and a small group of specialists in one place can be trained, supervised and kept proficient far more easily than inspectors dispersed among departments.
Independence matters as much as capability. When inspection results determine whether a lot is shipped, the inspector must not report to the supervisor whose output is being judged, and physical separation makes that independence visible to the customer and to the auditor. In regulated work the inspection record is a deliverable in its own right — a pressure part made to the ASME Boiler and Pressure Vessel Code and registered under CSA B51 must carry a traceable record of its dimensional and material acceptance — and centralised inspection produces that record as a matter of routine. Finally the practical conditions favour it: the parts are small enough to carry, the volume is low enough that transport time is a small fraction of throughput time, the part variety is high so no single gauge would be busy at any one machine, and some tests are destructive or hazardous (hardness traverses, sectioning, radiography, chemical analysis) and belong in a laboratory whatever else is decided.
The price paid should be stated plainly, because it is what makes the choice a judgment rather than a rule. Centralized inspection adds transport and queueing, which lengthens throughput time and ties up work in process; and it delays feedback to the operator, so a process that has drifted keeps producing scrap until the lot reaches the inspection crib. The usual resolution is a hybrid: first-piece and in-process checks by the operator at the machine, with final acceptance and all instrumented measurement centralised. In a high-volume line making one part, the balance tips the other way entirely, and floor inspection with operator-maintained control charts is the right answer.
Part (b) — the value of data processing equipment and computers in the quality control programme. Computers contribute at every stage of the quality cycle, and the contributions are worth separating because they are of different kinds.
At the data capture stage, electronic gauges, coordinate measuring machines and vision systems deliver readings directly into the quality record, which eliminates transcription error, makes 100 per cent measurement economically possible where it previously was not, and timestamps every reading so that a result can afterwards be tied to a machine, an operator, a tool and a material lot.
At the analysis stage the computer does what a clerk cannot do quickly enough to be useful. Control charts are plotted in real time and tested automatically against the full set of run rules, not merely against the three-sigma limits, so a drift or a stratification pattern is signalled while the process is still running. Process capability indices are computed on demand from the same data, and histograms and normal probability plots are available without extra work. Beyond charting, the computer makes the heavier statistical tools practical: analysis of variance to separate machine, operator and material effects; designed experiments to find the settings that reduce variation; regression to relate a process variable to a quality characteristic; and gauge repeatability and reproducibility studies that quantify how much of the observed spread is the measurement system rather than the product.
At the control stage the computer closes the loop. In-process gauging feeding tool-offset corrections keeps a machining process centred without operator intervention; automatic sorting diverts non-conforming units; and acceptance sampling plans with their switching rules are administered consistently, which is something that rarely survives manual administration. At the record stage the value is traceability: inspection results, material certificates, calibration histories and non-conformance reports are stored, indexed and retrievable, which is what makes lot recall bounded, what satisfies an ISO 9001 audit, and what constitutes the documentary defence in a product liability action.
Finally, at the management stage, the computer summarises. Pareto analyses of defect type and cost, trend reports by product and department, vendor quality ratings, and a cost-of-quality account split into prevention, appraisal, internal failure and external failure turn a mass of individual results into a small number of decisions. Management by exception only works if something is sorting the exceptions, and that is the computer.
Part (c) — how statistical quality control promotes understanding and appreciation of quality control. The central contribution is that statistical quality control makes variation visible. Before a control chart exists, quality is discussed in adjectives; afterwards it is discussed in numbers that everyone in the argument can see plotted on the same sheet of paper. That change of language is the beginning of understanding.
The second contribution is the distinction between chance causes and assignable causes, which is the single most useful idea a shop can learn. Once people can see that a process in control still varies, they stop chasing individual readings and stop blaming operators for variation the system imposes; and once they can see a point outside the limits, they accept that something specific has changed and is worth finding. It also apportions responsibility honestly: assignable causes are usually local and can be fixed by the people at the machine, while common-cause variation is built into the process and only management can spend the money to reduce it. That is a lesson about the organisation, not just about statistics.
Third, the chart converts the operator from the object of inspection into a participant in control. An operator who plots the points sees the effect of a tool change or a material change immediately, and the immediacy of that feedback is what builds genuine appreciation: quality stops being something the inspection department does afterwards and becomes something the process does now. Fourth, capability analysis forces a shared and factual conversation between manufacturing, design and sales, because a capability index below one says plainly that the tolerance and the process are incompatible and that someone must change one of them; arguments about whether the tolerance is "too tight" end when the number appears.
Fifth, statistical quality control demonstrates the economic case. It shows that screening inspection cannot create quality, only sort it, and it lets the saving from a process improvement be measured rather than asserted, which is what earns quality work its budget. Sixth, it provides objective evidence of progress, and a chart whose limits visibly narrow over a year sustains motivation in a way that exhortation does not. Finally, it carries the same understanding outward: a supplier who submits control charts with a shipment is communicating capability, and a customer who reads them understands the supplier far better than any certificate of conformance would allow.
