5.5.16.2.5.1 Key Concepts of Taguchi Design
Contents
Orthogonal Arrays
Origin provides standardized orthogonal arrays denoted as L(runs). Common arrays include:
- L4, L8, L12, L16 for 2-level factors
- L9, L18, L27 for 3-level factors
- Mixed-level designs like L18 (2¹ × 3⁷) for combinations of 2-level and 3-level factors
- The notation L9 (3⁴) means 9 experimental runs with 4 factors at 3 levels each.
Signal-to-Noise (S/N) Ratio
Origin calculates S/N ratios to measure quality characteristics:
- Larger-the-better (maximize response, e.g., strength)
- Smaller-the-better (minimize response, e.g., defects)
- Nominal-the-best (target specific value)
- Dynamic (functional relationship between signal and response
Control vs. Noise Factors
- Control Factors: Parameters you can adjust (e.g., temperature, pressure)
- Noise Factors: Uncontrollable variations (e.g., humidity, material batch variation)
Static vs. Dynamic Designs
| Type | Description | Application |
|---|---|---|
| Static Taguchi Design | Optimizes a single target response value | Maximizing yield, minimizing defects, achieving target dimensions |
| Dynamic Taguchi Design | Optimizes the relationship between input signal and output response | Sensor calibration, process control, image processing |