In the realm of process improvement and quality management, understanding advanced Design of Experiments (DOE) techniques is essential for achieving optimal results. Among these methodologies, the outer array stands as a powerful tool that enables organizations to create robust processes that perform consistently despite variations in uncontrollable factors. This comprehensive guide will walk you through the outer array technique, its applications, and how to implement it effectively in your quality improvement initiatives.
Understanding the Outer Array Concept
The outer array, also known as the noise array, is a critical component of Taguchi’s robust parameter design methodology. This technique recognizes that in real-world manufacturing and service environments, not all variables can be controlled. While some factors can be set and maintained at specific levels (control factors), others fluctuate due to environmental conditions, material variations, or customer usage patterns (noise factors). You might also enjoy reading about How to Understand and Calculate Between Subgroup Variation: A Complete Guide.
The outer array specifically focuses on these uncontrollable noise factors. By systematically varying these factors in a structured experimental design, organizations can identify control factor settings that minimize the impact of noise, resulting in processes that deliver consistent quality regardless of varying conditions. You might also enjoy reading about How to Understand and Apply the Different Levels of Lean Six Sigma in Your Organization.
The Difference Between Inner and Outer Arrays
To fully grasp the outer array concept, you must understand its relationship with the inner array. The inner array contains the controllable factors that you can adjust and set at desired levels. These might include machine settings, process parameters, or material specifications that you can directly influence.
The outer array, conversely, contains the noise factors that you cannot easily control but significantly affect process output. These might include ambient temperature, humidity, material lot variations, operator differences, or equipment aging. By combining both arrays in a crossed array design, you can evaluate how different combinations of control factors perform across various noise conditions.
Step-by-Step Guide to Creating an Outer Array
Step 1: Identify Your Noise Factors
Begin by conducting a thorough analysis of your process to identify all potential noise factors. Engage your team in brainstorming sessions, review historical data, and consult with operators who have intimate knowledge of daily variations. Common categories of noise factors include environmental conditions, material variations, deterioration over time, and differences in customer usage patterns.
Step 2: Select the Most Critical Noise Factors
Not all noise factors warrant inclusion in your outer array. Prioritize those that have the most significant impact on your process output. Typically, you should select two to four noise factors for your outer array to maintain experimental efficiency while capturing the most important sources of variation.
Step 3: Determine the Levels for Each Noise Factor
For each selected noise factor, establish the range of variation you expect to encounter in actual operating conditions. Generally, you will use two levels representing the extremes of expected variation. For example, if temperature is a noise factor, you might select the highest and lowest temperatures experienced during normal operations.
Step 4: Construct the Outer Array Design
Create a factorial design that includes all combinations of your selected noise factors at their specified levels. The number of runs in your outer array equals 2^n, where n represents the number of noise factors. With three noise factors, your outer array would contain eight runs (2^3 = 8).
Practical Example: Manufacturing Process Optimization
Consider a manufacturing company producing metal brackets used in automotive applications. The company wants to optimize the welding process to ensure consistent weld strength regardless of varying conditions.
Identifying Control and Noise Factors
Control Factors (Inner Array):
- Welding current: 150A or 180A
- Welding time: 2 seconds or 3 seconds
- Electrode pressure: 300 psi or 400 psi
Noise Factors (Outer Array):
- Material lot variation: Lot A or Lot B
- Ambient temperature: 60°F or 85°F
- Electrode wear: New or Worn
Sample Data Set and Analysis
The inner array consists of 8 runs (2^3 combinations of control factors), and the outer array consists of 8 runs (2^3 combinations of noise factors). Each combination from the inner array is tested under all 8 noise conditions from the outer array, resulting in 64 total experimental runs.
Here is a simplified representation of one inner array combination tested across the outer array:
Inner Array Run 1: Current = 150A, Time = 2s, Pressure = 300 psi
Outer Array Results (Weld Strength in kg):
- Run 1 (Lot A, 60°F, New electrode): 245 kg
- Run 2 (Lot A, 60°F, Worn electrode): 230 kg
- Run 3 (Lot A, 85°F, New electrode): 235 kg
- Run 4 (Lot A, 85°F, Worn electrode): 218 kg
- Run 5 (Lot B, 60°F, New electrode): 240 kg
- Run 6 (Lot B, 60°F, Worn electrode): 225 kg
- Run 7 (Lot B, 85°F, New electrode): 228 kg
- Run 8 (Lot B, 85°F, Worn electrode): 215 kg
For this particular combination of control factors, the mean weld strength is 229.5 kg with a standard deviation of 10.2 kg. This analysis would be repeated for all eight inner array combinations.
Analyzing Outer Array Results
The primary objective of outer array analysis is to identify control factor settings that maximize the signal-to-noise ratio (S/N ratio). The S/N ratio quantifies how much the desired signal (target performance) stands out from the background noise (variation due to noise factors).
Calculating Signal-to-Noise Ratios
Depending on your quality characteristic, you will select an appropriate S/N ratio formula. For the “larger is better” characteristic (such as weld strength), the formula is:
S/N = -10 × log₁₀(mean of 1/y²)
For the welding example above, after calculating S/N ratios for all eight inner array combinations, you would identify which combination of welding current, time, and pressure produces the highest S/N ratio. This combination represents the most robust process settings that will deliver consistent weld strength despite variations in material lots, temperature, and electrode condition.
Implementing Outer Array Findings
Once you have identified the optimal control factor settings through your outer array experiment, implementation should follow a structured approach:
Validation Testing
Before full-scale implementation, conduct confirmation runs using the optimal settings under various noise conditions to verify that performance matches your predictions. This validation step ensures that your experimental findings translate to actual process improvements.
Process Standardization
Document the optimal settings in standard operating procedures and train all relevant personnel on the new process parameters. Establish control charts to monitor ongoing performance and ensure that the process remains within acceptable limits.
Continuous Monitoring
Implement a system for tracking process performance over time. Even robust processes may drift due to factors not captured in the original experiment. Regular monitoring allows you to detect changes early and take corrective action before quality issues arise.
Benefits of Using Outer Arrays
Organizations that effectively implement outer array techniques in their process improvement efforts realize numerous benefits. First, they achieve more consistent product quality with reduced variation, leading to fewer customer complaints and warranty claims. Second, robust processes require less adjustment and troubleshooting, resulting in improved productivity and reduced scrap rates. Third, understanding how noise factors affect your process enables better risk management and proactive problem-solving.
Additionally, the knowledge gained from outer array experiments often reveals unexpected interactions between control and noise factors, providing insights that guide future innovation and process development efforts.
Common Challenges and Solutions
While outer arrays provide powerful insights, practitioners often encounter challenges during implementation. The most common issue is the resource requirement, as crossed array designs can result in numerous experimental runs. To address this, consider using fractional factorial designs that reduce the number of runs while still capturing the most important effects.
Another challenge involves accurately simulating noise conditions in an experimental setting. Some noise factors, such as long-term deterioration, cannot be easily reproduced. In such cases, use accelerated testing methods or historical data to estimate the effects of these factors.
Take Your Quality Improvement Skills to the Next Level
Mastering advanced techniques like outer arrays requires comprehensive training and practical experience. The principles covered in this guide represent just one aspect of the robust toolkit available through Lean Six Sigma methodologies. Whether you are beginning your quality improvement journey or seeking to enhance your existing skills, formal training provides the structured knowledge and hands-on practice necessary for success.
Lean Six Sigma certification programs offer in-depth coverage of Design of Experiments, including outer arrays, Taguchi methods, response surface methodology, and much more. These programs combine theoretical understanding with real-world applications, preparing you to lead transformative improvement initiatives in your organization.
Enrol in Lean Six Sigma Training Today and gain the expertise needed to implement sophisticated quality improvement techniques that deliver measurable results. Professional certification demonstrates your commitment to excellence and positions you as a valuable asset capable of driving operational excellence. Do not let your competitors gain the advantage. Start your Lean Six Sigma journey today and transform the way your organization approaches quality, efficiency, and customer satisfaction.
