5: Combining Merit Indices
Introduction to Combining Merit Indices
Discussion on combining three merit indices that represent different objectives for safely designing pressure vessels.
Objective: Ranking specific alloys within a defined region of materials based on these indices.
Importance of Normalization and Weighting
Normalization: Converts merit indices into dimensionless units for combination.
Necessary since indices typically have different units.
Weighting: Assigns relative importance to merit indices.
Equal weighting (e.g., 50-50 split) is possible, but sometimes certain indices are more crucial (e.g., 70-30 split).
Normalization Schemes
Maximum Value Normalization
Method One: Normalize values by the maximum merit index value in the list.
Formula: ( \text{Normalized}(m_i) = \frac{m_i}{max(m)} )
Results in each value representing a percentage of the best material.
Composite Merit Index Calculation:
Combined index: ( m^* = w_1 \cdot \text{Normalized}(m_1) + w_2 \cdot \text{Normalized}(m_2) + ... )
Standard Deviation Normalization
Method Two: Normalize using mean and standard deviation.
Formula: ( \text{Normalized}(m_i) = \frac{m_i - ext{Mean}(m)}{\text{Standard Deviation}(m)} )
Indicates how many standard deviations a material's value is from the mean.
Negative values: Indicate below-average performance; positive values indicate above-average performance.
Composite Merit Index Calculation: Similar to Method One using normalized values.
Example Comparison of Normalization Methods
Merit Index One: Values from 0 to 10,000; Material B as the maximum value.
Merit Index Two: Values from 90 to 100; Material B also the maximum.
Calculating Normalizations
Maximum Value Normalization Results:
Material B normalized to 1.0; Materials A and C show lowest values.
Standard Deviation Normalization Results:
Material B maintains the highest positive value; Material C shows significant negative value.
Ranking Results from Composite Indices
Both normalization methods rank Material B as the best.
Rankings for other materials may differ:
Example: Material C's rankings flip between the two methods due to differences in perceived value spread.
Key Insights on Normalization Methods
Maximum Value Method: Doesn't distinguish enough between close values, treating nominal differences as less significant.
Standard Deviation Method: Recognizes that even small absolute differences can represent larger relative distinctions, affecting material rankings.
Conclusion
Understanding how to combine merit indices through normalization and weighting is crucial in material selection for design purposes.
The lecture covers different approaches, their calculations, and implications for ranking materials effectively.