Advanced Tools and Methods for Quality Control - OSS 5

This flashcard set covers experimental design techniques, Evolutionary Operation (EVOP), tolerance redesign principles, and response surface methodologies based on the OSS 5 lecture transcript.

Multi-level Column Technique (Τεχνική πολύ-σταθμικών στηλών)

A technique used to study a factor with multiple levels by 'sacrificing' specific columns (and their interaction columns) in an experimental design to define the required levels for the new factor.

Confounding Technique (Τεχνική σύγχυσης)

An experimental strategy where the effects of factors or interactions expected to be non-significant are confounded with others to reduce the number of required trials.

Dummy Level Technique (Τεχνική εικονικών στάθμεων)

A method allowing a factor with $k$ levels to be placed in a column designed for $m$ levels (where m > k) by repeating one or more of the factor's levels, typically those with lower experimental or economic costs.

OA27(3^{13})

An orthogonal array design from Table A11 (page 81) consisting of 13 columns and 27 experimental trials, used for economic experimental designs.

Evolutionary Operation (EVOP)

A continuous improvement method used to optimize production processes (like fabric strength) through systematic, small changes during actual production without stopping the manufacturing process.

Central Reference Point (Κεντρικό σημείο αναφοράς)

The set of initial optimal conditions (e.g., $30\%$ cotton and $40\,g$ resin) used as a baseline in Evolutionary Operation from which changes are measured.

Total Tolerance ($T_0$)

The overall allowable deviation from a target value, defined by the sum of squares of individual component tolerances: $\sum_{i=1}^{n} T_i^2 = T_0^2$.

Contribution Ratio ($P_F$)

A metric representing how much a factor's tolerance contributes to overall variability, calculated as $P_F = 100 \times \frac{AT_F - \beta\varepsilon_F \times MATY}{OAT}$, where $AT_F$ is the sum of squares for factor $F$ ($SSF$), $MATY$ is the mean square error ($MSE$), and $OAT$ is the total sum of squares ($SSTO$).

Tolerance Equation

An inequality used to determine the reduction or expansion factor $a(F)$ required for individual component tolerances to achieve a specific reduction in total process variability.

Response Surfaces (Επιφάνειες Απόκρισης)

Mathematical models and 3D visualizations used to determine the relationship between an experimental response ($Y$) and multiple factors ($X_i$) to identify the optimal combination of values.

Response Surface Regression Equation

The quadratic model expressed as \hat{Y} = \theta_0 + \sum_{i=1}^p \theta_i X_i + \sum_{i=1}^p \theta_{ii} X_i^2 + \sum\sum_{i < j} \theta_{ij} X_i X_j.

Box-Behnken

A specific type of experimental design mentioned for use in creating response surfaces and optimizing quality control parameters.

Estimation of Interaction in EVOP

Calculated by comparing the change in response for one factor while the other is at a low level against the change when it is at a high level, using the formula $\frac{(M_2 - M_5) - (M_4 - M_3)}{2}$.

Change of Mean Estimation (EVOP)

The difference between the average of all test points ($M_1$ through $M_5$) and the central reference point mean ($M$).