Quality engineering glossary

84 terms from quality engineering, Six Sigma, reliability and lean, in plain language. Each one links to a worked guide or a free calculator.

A

Acceptance number (c)

In a single sampling plan, the largest number of defectives in the sample for which the lot is still accepted. Acceptance sampling →

Acceptance quality limit (AQL)

The worst process average (percent defective) that a sampling plan treats as acceptable over a series of lots. It describes the plan, not the quality of any one lot. AQL and OC curves →

Alpha risk (α)

The probability of a Type I error: rejecting the null hypothesis when it is true. Also the producer's risk in acceptance sampling. Type I vs Type II error →

ANOVA

Analysis of variance: tests whether three or more group means differ by comparing the variation between groups with the variation within them (F = MS between / MS within). One-way ANOVA →

Attribute data

Data counted in categories, such as pass/fail or number of defects. It needs larger samples than variables data to give the same information. Variables vs attributes →

Availability

The share of time equipment is able to run. Inherent availability is MTBF / (MTBF + MTTR). MTBF and availability calculator →

B

B10 life

The age by which 10% of units are expected to fail. Common in bearing and component specifications. Weibull calculator →

Beta risk (β)

The probability of a Type II error: failing to detect a real effect. Also the consumer's risk in acceptance sampling. Power = 1 − β. Type I vs Type II error →

Bias (measurement)

The difference between the average of repeated measurements and the reference value of the same part. Bias, linearity and stability →

Binomial distribution

The number of successes in n independent trials with the same probability p each. Used for defectives in a sample from a large lot. Probability calculator →

Black Belt

A Six Sigma practitioner who leads improvement projects full time and coaches Green Belts. ASQ's credential is the CSSBB. Black Belt study →

Box plot

A chart of the median, quartiles and spread of a data set, useful for comparing groups and spotting outliers. Graphical analysis →

C

Calibration

Comparing a measuring instrument with a more accurate standard, traceable to national standards, and adjusting or recording the difference. Calibration and traceability →

CAPA

Corrective and preventive action: a system for removing the cause of a detected problem (corrective) or a potential one (preventive). NCR, CA and PA →

Cause-and-effect diagram

A fishbone (Ishikawa) diagram that sorts possible causes of a problem into categories such as man, machine, method, material, measurement and environment. Fishbone diagram →

Central limit theorem

The means of samples tend toward a normal distribution as sample size grows, whatever the shape of the individual data. It is why X̄ charts work with non-normal data. Normal distribution →

Chi-square test

A test that compares observed counts with expected counts (goodness of fit, independence), or a sample variance with a target. Chi-square test →

Common cause variation

The natural, ever-present variation of a stable process. Reducing it requires changing the process, not reacting to single points. Control charts →

Confidence interval

A range, built from sample data, that would contain the true value in a stated share of repeated samples (for example 95%). Confidence interval calculator →

Control chart

A time-ordered plot with a center line and control limits that separates common-cause from special-cause variation. Control chart picker →

Control limits

Lines usually set 3 standard deviations from the center line, calculated from the process data. They are not specification limits. Control limits calculator →

Control plan

A document listing each characteristic to control, the method, sample size, frequency and the reaction plan when it goes out of control. Control plan explained →

Cost of quality

Prevention plus appraisal plus internal and external failure costs. Spending on prevention usually lowers the total. Cost of quality →

Cp

Potential capability: (USL − LSL) / 6σ_within. It ignores where the process is centered. Process capability calculator →

Cpk

Actual short-term capability: the distance from the mean to the nearest spec limit divided by 3σ_within. 1.33 is a common minimum. Cpk and Ppk calculator →

Cpm

Taguchi capability index that penalizes distance from the target: (USL − LSL) / (6√(σ² + (μ − T)²)). Cp, Cpk and Cpm →

Critical to quality (CTQ)

A measurable characteristic, with a target and limits, that translates a customer need into something the process can control. VOC and CTQ tree →

CUSUM chart

A control chart of cumulative sums of deviations from target. It detects small sustained shifts faster than a Shewhart chart. CUSUM and EWMA →

D

Defect vs defective

A defect is one nonconformity; a defective is a unit with one or more defects. Defects go on c or u charts, defectives on p or np charts. Attribute control charts →

Design of experiments (DOE)

Changing several factors at once in a planned pattern to find which ones affect the response, and how they interact, with few runs. DOE effects calculator →

DFMEA and PFMEA

Design FMEA analyzes how a design can fail to meet its function; process FMEA analyzes how a process can fail to produce a good part. DFMEA vs PFMEA →

DMAIC

Define, Measure, Analyze, Improve, Control: the five-phase Six Sigma method for improving an existing process. Six Sigma and DMAIC →

DPMO

Defects per million opportunities: defects × 1,000,000 / (units × opportunities per unit). DPMO and sigma calculator →

DPU

Defects per unit: total defects / units. First pass yield ≈ e^(−DPU). DPMO and sigma calculator →

E

EWMA chart

Exponentially weighted moving average chart: each point blends the new value with the previous average, which makes small shifts visible. CUSUM and EWMA →

F

First pass yield (FPY)

The share of units that pass a step the first time, with no rework or scrap. RTY calculator →

FMEA

Failure mode and effects analysis: a structured review of how something can fail, the effects and causes, and the controls, ranked for action. FMEA: RPN vs AP →

G

Gage R&R

A study that splits measurement variation into repeatability (the gage) and reproducibility (the appraisers), compared with part variation or tolerance. Gage R&R calculator →

Green Belt

A Six Sigma practitioner who works on or leads smaller improvement projects part time. ASQ's credential is the CSSGB. Green Belt study →

H

Hypothesis test

A procedure that decides between a null hypothesis (no effect) and an alternative, using a test statistic and a significance level α. Which test to use →

I

I-MR chart

Individuals and moving range chart, for one measurement at a time. Limits: X̄ ± 2.66 × MR̄. I-MR chart →

K

Kaizen

Continuous improvement through small, frequent changes, often in focused multi-day events. Kaizen events →

Kappa

A measure of agreement between appraisers on attribute decisions, corrected for agreement expected by chance. Above about 0.75 is usually good. Attribute agreement and kappa →

L

Lean

A system for delivering value with less waste, built on flow, pull, standard work and continuous improvement. The 8 lean wastes →

Linearity (measurement)

How bias changes across the measuring range of a gage. Bias, linearity and stability →

M

MSA

Measurement system analysis: checking that a measurement system is accurate, precise and stable enough for its use before trusting its data. MSA for Green Belts →

MTBF

Mean time between failures for a repairable item: operating time / number of failures. MTBF calculator →

MTTR

Mean time to repair: total repair time / number of repairs. MTBF calculator →

N

Normal distribution

The symmetric bell-shaped distribution described by its mean and standard deviation. About 99.73% of values fall within ±3σ. z table →

Number of distinct categories (ndc)

How many groups of parts a gage can reliably tell apart: 1.41 × PV / GRR. Five or more is the usual target. Gage R&R calculator →

O

OC curve

Operating characteristic curve: the probability that a sampling plan accepts a lot, plotted against the lot's percent defective. AQL and OC curves →

OEE

Overall equipment effectiveness: availability × performance × quality. OEE calculator →

P

p chart

A control chart for the fraction defective, with limits p̄ ± 3√(p̄(1 − p̄)/n). It handles varying sample sizes. Control limits calculator →

p-value

The probability, if the null hypothesis were true, of a result at least as extreme as the one observed. It is not the probability that the null is true. p-value explained →

Pareto chart

A bar chart of categories sorted from largest to smallest, with a cumulative line, to find the vital few. Pareto chart →

PDCA

Plan, Do, Check, Act: the basic cycle of trying a change, measuring it and standardizing what works. PDCA vs DMAIC →

Poisson distribution

The number of events in a fixed interval when they occur independently at a constant average rate λ. Used for defects per unit. Probability calculator →

Poka-yoke

Mistake-proofing: a device or method that prevents an error or makes it obvious at once. Poka-yoke →

Power

The probability that a test detects an effect of a given size when it is really there: 1 − β. Sample size calculator →

Pp and Ppk

Performance indices: like Cp and Cpk, but using the overall standard deviation, so they include shifts between subgroups. Cpk vs Ppk →

PPAP

Production part approval process: the evidence package showing a supplier's process can consistently make parts that meet requirements. APQP and PPAP →

R

Regression

Fitting an equation that predicts a response from one or more inputs. R² is the share of variation it explains. Regression calculator →

Reliability

The probability that an item performs its function for a stated time under stated conditions. Reliability calculator →

Repeatability

Variation when one appraiser measures the same part several times with the same gage (equipment variation, EV). Gage R&R explained →

Reproducibility

Variation between the averages of different appraisers measuring the same parts (appraiser variation, AV). Gage R&R explained →

Resolution (DOE)

How badly effects are aliased in a fractional factorial. Resolution III aliases main effects with two-factor interactions; IV keeps main effects clear; V keeps two-factor interactions clear too. Full vs fractional factorial →

Rolled throughput yield (RTY)

The product of the first pass yields of every step: the chance a unit goes through the whole process with no defect. RTY calculator →

RPN

Risk priority number in FMEA: severity × occurrence × detection, each rated 1 to 10. RPN vs Action Priority →

Run rules

Patterns on a control chart, such as runs, trends and points near the limits, that signal a special cause even inside the limits. Western Electric rules →

S

Sigma level

A defect rate expressed as a z-score, conventionally including a 1.5σ shift: 3.4 DPMO is 6σ. DPMO and sigma calculator →

SIPOC

Suppliers, Inputs, Process, Outputs, Customers: a one-page high-level map that sets the scope of a project. SIPOC and process mapping →

Six Sigma

A data-driven improvement approach aimed at reducing variation and defects, usually run through DMAIC projects. Six Sigma certification guide →

Special cause variation

Variation from a specific, identifiable source outside the normal process, signaled by points beyond the limits or by run rules. Run rules →

Specification limits

The limits a characteristic must meet, set by the customer or the design. They are not control limits. Process capability calculator →

Stability (measurement)

How the bias of a gage changes over time, usually checked with a control chart of a reference part. Bias, linearity and stability →

T

t-test

A hypothesis test for means when the standard deviation is estimated from the sample: one-sample, two-sample or paired. t-test calculator →

Takt time

Available working time divided by customer demand: the pace at which units must be finished. Takt time calculator →

Type I and Type II errors

Type I: rejecting a true null hypothesis (α). Type II: failing to reject a false one (β). Type I vs Type II error →

U

u chart

A control chart for defects per unit when the number of units inspected varies: ū ± 3√(ū/n). Control limits calculator →

V

Value stream map

A diagram of the material and information flow for one product family, with times, inventory and waste, used to design a better future state. Value stream mapping →

Voice of the customer (VOC)

Customer needs and expectations, gathered through interviews, surveys and complaints, before they are translated into CTQs. VOC for Green Belts →

W

Weibull distribution

A flexible life distribution. Shape β under 1 means early failures, about 1 random failures, over 1 wear-out. Weibull calculator →

X

X̄-R chart

Control charts for subgroup averages and ranges: X̿ ± A₂R̄ and D₃R̄ to D₄R̄. Used for subgroups of about 2 to 10. X̄-R vs X̄-S →

Z

z-score

How many standard deviations a value is from the mean: z = (x − μ) / σ. z table →

Test yourself on these terms