Nguyễn et al. (2024) — Taguchi Robust Design of Slag-Based Engineered Cementitious Composites

Citation

Phương Hoàng Nguyễn, Se-Eon Park, Huy Hoàng Nguyễn, Youngsang Kim, Bang Yeon Lee (2024). Influential factor analysis of slag-based engineered cementitious composites using Taguchi robust method. Journal of Structural Integrity and Maintenance, 9(1), 2317529.

DOI: 10.1080/24705314.2024.2317529
Atlas layer: supporting
Related Victor Li book chapter: Chapter 9: Green ECC & Chapter 4: Micromechanics-Based Tailoring
Source PDF: primary_data/nguyen-2024-influential-factor-analysis-of-slag-based.pdfIJP07824N_Influential factors of ECC_JSIM.pdf`
Extracted text: secondary_data/full_texts/nguyen-2024-influential-factor-analysis-of-slag-based_full_text.mdsecondary_data/full_texts/IJP07824N_Influential factors of ECC_JSIM_full_text.md`
Source note: secondary_data/source_notes/nguyen-2024-influential-factor-analysis-of-slag-based_source_note.mdsecondary_data/source_notes/IJP07824N_Influential factors of ECC_JSIM_source_note.md`

Why this paper matters

Applies the Taguchi robust design ($L_8$ orthogonal array) and ANOVA to systematically analyze the effects of slag content (10 vs 50 wt%), $w/b$ ratio (0.40 vs 0.50), crumb rubber (0 vs 10 wt%), and PE fiber volume fraction (1.25 vs 1.50 vol. %) on slag-based ECC (S-ECC). Formulates multiple regression models that accurately predict the optimal high-strength and high-ductility mixture (S0.1-W0.4-R0-P1.5: $f_c = 72.37\text{ MPa}, \sigma_{tu} = 9.24\text{ MPa}, \epsilon_{ts} = 7.35\text{ \%}, \omega_c = 110.6\ \mu\text{m}$) with under 5 % error.

Main contribution

Taguchi $L_8$ Slag-ECC Experimental Matrix: Formulated and tested 8 mix designs, achieving compressive strengths of 35.12–77.10 MPa and direct tensile strain capacities of 5.36–7.35 %.
ANOVA Sensitivity Hierarchy:
Compressive Strength ($f_c$): Governed by water-to-binder ratio $W$ (46.53 %) and crumb rubber $R$ (40.84 %).
Tensile Strength ($f_{ts}$): Governed by slag $S$ (37.68 %) and crumb rubber $R$ (30.21 %).
Tensile Strain Capacity ($\epsilon_{ts}$): Governed by water-to-binder ratio $W$ (47.64 %).
Crack Width ($\omega_c$): Governed by PE fiber content $P$ (49.83 % reduction contribution).
Predictive Multi-Linear Equations & Optimization: Derived empirical models that precisely matched experimental properties of the optimal S0.1-W0.4-R0-P1.5 composite ($f_c = 72.37\text{ MPa}, \epsilon_{ts} = 7.35\text{ \%}$, error $<5\text{ \%}$).

Evidence summary

Mechanical Properties Across $L_8$ Array:
S0.1-W0.4-R0-P1.25: $f_c = 77.10\text{ MPa}$, $\sigma_{tu} = 9.23\text{ MPa}$, $\epsilon_{ts} = 6.93\text{ \%}$, $\omega_c = 141.0\ \mu\text{m}$ (Tables 5–7, Pages 4, 6).
S0.1-W0.4-R0-P1.5 (Optimal): $f_c = \mathbf{72.37\text{ MPa}}$, $\sigma_{tu} = \mathbf{9.24\text{ MPa}}$, $\epsilon_{ts} = \mathbf{7.35\text{ \%}}$, $\omega_c = \mathbf{110.6\ \mu\text{m}}$, 53.2 cracks.
S0.1-W0.5-R0.1-P1.25: $f_c = 35.12\text{ MPa}$, $\sigma_{tu} = 5.84\text{ MPa}$, $\epsilon_{ts} = 5.36\text{ \%}$.
S0.5-W0.4-R0.1-P1.5: $f_c = 48.53\text{ MPa}$, $\sigma_{tu} = 6.27\text{ MPa}$, $\epsilon_{ts} = 6.21\text{ \%}$.
ANOVA Contributions:
$f_c$: $W$ (46.53 %), $R$ (40.84 %), $S$ (11.20 %) (Table 8, Page 7).
$f_{cr}$: $W$ (52.81 %), $P$ (34.46 %) (Table 9, Page 7).
$f_{ts}$: $S$ (37.68 %), $R$ (30.21 %), $W$ (17.20 %) (Table 10, Page 7).
$\epsilon_{ts}$: $W$ (47.64 %), $S$ (18.24 %), $P$ (11.92 %) (Table 11, Page 7).
$\omega_c$: $P$ (49.83 % negative contribution) (Table 12, Page 7).
Regression Equations:
$f_c = 156.43 - 24.68 S - 201.10 W - 188.40 R + 0.88 P$ (Eq. 7, Page 7).
$f_{ts} = 9.86 - 4.06 S - 11.00 W - 14.50 R + 3.04 P$ (Eq. 9, Page 7).
$\epsilon_{ts} = 8.32 - 1.30 S - 8.40 W - 4.00 R + 1.68 P$ (Eq. 10, Page 7).
Optimal validation: Predicted $74.84\text{ MPa} / 7.36\text{ \%}$ vs Experimental $72.37\text{ MPa} / 7.35\text{ \%}$ (Table 13, Page 9).

Linked Atlas nodes

04_material_systems/green_ecc.md
02_concepts/flaw_design.md
05_experiments/crack_width_distribution.md
05_experiments/direct_tensile_test.md

Relationship to Victor Li book

Primary book anchor remains Victor Li (2019), Engineered Cementitious Composites (ECC).
Directly supports Chapter 9 (Green ECC) and Chapter 4 (Micromechanics) by providing a statistical Taguchi ANOVA and multi-regression optimization protocol for slag-cement PE-ECC, establishing that $w/b$ governs matrix strength/ductility (46.5–47.6 % contribution) while PE fibers govern crack tightness (49.8 % contribution).

Claim-evidence rows to add

Atlas node	Claim	Evidence summary	Page/Figure/Table	Status
`04_material_systems/green_ecc.md`	Taguchi optimization of slag-ECC identifies S0.1-W0.4-R0-P1.5 as optimal, achieving 72.37 MPa compressive strength and 7.35 % direct tensile strain capacity	Experimental tension and compression tests verified $f_c = 72.37\text{ MPa}$ and $\epsilon_{ts} = 7.35\text{ \%}$	Page 2317529:1 & 9 / Table 5, 6 / Fig. 2a	verified_from_pdf
`02_concepts/flaw_design.md`	ANOVA reveals water-to-binder ratio (46.53 %) and crumb rubber (40.84 %) are the two primary negative factors governing compressive strength in slag-ECC	ANOVA statistical table confirmed 46.53 % ($W$) and 40.84 % ($R$) contributions	Page 2317529:4 / Table 8 / Fig. 3a	verified_from_pdf
`05_experiments/crack_width_distribution.md`	PE fiber volume fraction exerts a 49.83 % statistical contribution toward tightening crack widths ($\omega_c = 110.6\ \mu\text{m}$) in slag-ECC	Cracking pattern measurements and ANOVA verified 49.83 % contribution to crack width reduction	Page 2317529:7–8 / Table 7, 12 / Fig. 3e	verified_from_pdf

Verification status

PDF preserved: yes (in primary_data/IJP07824N_Influential factors of ECC_JSIM.pdf)
Text extracted: yes (PyMuPDF, 11 pages)
DOI verified: yes (10.1080/24705314.2024.2317529)
Page/figure/table verified: yes (all checked in PDF text)
Claim-evidence matrix ready: yes

Cautions

Crumb rubber addition above 10 wt% causes severe compressive strength loss.
Slag replacement at 50 wt% slightly lowers tensile strength compared to 10 wt% slag.