Many statistical models are available to define an in-control manufacturing process. In the simplest terms, an in-control manufacturing process is the successful integration of all production factors, combining in an integrated manner, to produce a sufficient quantity of product at an acceptable quality level. When any of these factors is compromised, the process requires corrective action.
If the initial process was documented with quantifiable metrics, methods, and processes, re-establishing control is a relatively simple process. A comparison of current operational status to the initial manufacturing benchmarks often reveals the effects of changes to the system. Alterations in manufacturing methods, tooling, or even personnel can produce unanticipated results. A return to the previous status quo is as easy as turning back the page. Documenting the changes based on accepted statistical methods should re-establish control.
Unfortunately, many manufacturing processes have evolved or matured in the dark, and are based on anecdotal information, unrecorded assumptions, and tribal knowledge. In such instances, previous success is difficult to replicate, as there is little if any objective evidence to serve as guidelines for recreation of control. Rather than attempting to impose order on random circumstance, it's time to create a new and documented process.
Establishing metrics for the manufacturing process is the first step in establishing the basics of control. It's impossible to win if there's no way to keep score. Acceptable benchmarks for output, efficiency, and quality must be clearly defined. Production levels are often defined by factors outside the production process. Customer delivery expectations, the need to maintain manufacturing flow, and basic economic factors very often define output. The variable factors that will define and establish control are quality and efficiency of operations. To define these benchmarks, it may be necessary to reappraise current inspection systems and other quality assurance practices used to evaluate product conformance.
Traditional attribute inspection is a valuable tool for determining conformance, but is of little value in establishing process variation. By definition, attributes inspection is at the limits, and there is only Go-Not Go data. Degradation of process, root causes of variation, and significant sources of change are not measured. Entire process shifts that undermine both efficiency and quality are not detected in real-time, but only as product reaches the limits of acceptability. Variables inspection, or actual process measurement, will not only verify conformance, but has the ability to document change and detect subtle but significant process shifts. It forms the basis for statistical evaluation of capability, and analysis of process error as it actually occurs. Further, process measurement with variables-inspection systems can optimally target processes within product tolerance, for maximum production capability and output.
Several years ago, we helped evaluate and document manufacturing operations and methods for a new automotive engine design using variables inspection tools. Given the high output level expectations, new methods of manufacture, corporate acceptance standards, and the basic design necessity to minimize thread form error and deviation, basic variables thread-inspection practices were instituted during the manufacturing verification process. Initial sampling showed excessive and unacceptable form and size deviation. Processes were evaluated and modified to ensure acceptable product. The interrelated effect of tolerances, tooling, and methods was documented and incorporated into the process plan. Data-driven changes to the process directly related to actual measurement of real-time variation formed the basis for an efficient manufacturing process that continued during the production process.
As these methods continued on the floor, unexpected benefits became evident. As tapping operations were documented, it became apparent that tooling degradation did not occur at a uniform rate. Degradation was exponential. When the process reached a certain point, tap wear accelerated, and process control was lost. Based on this documented evidence, taps and other tooling were replaced at predetermined intervals before control was lost. Efficiency, output, and quality were all maintained at the expected and acceptable levels.
Change is inevitable in any process. Inspection systems that cannot detect process shift and document subtle but significant change, that do little to detect basic process error, have no place in today's manufacturing environment. Establishing, correcting, and maintaining processes is a basic responsibility in any production environment.
This article was first published in the March 2010 edition of Manufacturing Engineering magazine.