In a significant stride toward sustainable materials, a study published in Scientific Reports explores the use of coconut shell biochar (CCB) as a reinforcement for polylactic acid (PLA) biocomposites. Authors Harishbabu Sundarasetty et al. developed a new framework to optimize the mechanical performance of these materials, proving that a combination of traditional experimental design and advanced machine learning (ML) offers a highly reliable method for process optimization. The core challenge in creating a high-performance composite is precisely controlling manufacturing parameters—like material composition and injection molding conditions—that govern the final product’s strength and stiffness. This research provides a robust, data-driven solution.

The study fabricated PLA/CCB composites by varying four key injection molding parameters: composition (pure PLA, 5 wt%, 10 wt% CCB), injection temperature (135∘C, 145∘C, 155∘C), injection speed (50 mm/s, 60 mm/s, 70 mm/s), and injection pressure (30 bar, 40 bar, 50 bar). By employing a Taguchi L27 orthogonal array, the researchers efficiently tested a wide range of parameter combinations, systematically gathering data on tensile strength, Young’s modulus (stiffness), and hardness. This experimental approach quickly identified the most promising material-processing combinations.

The experimental results confirmed that adding coconut shell biochar dramatically enhanced the material’s properties. Composites with 10 wt% CCB demonstrated a maximum tensile strength of 54.6–60.99 MPa and a maximum Young’s modulus of 2652–3561 MPa, a significant increase over pure PLA. For hardness, samples with 10 wt% CCB achieved a maximum hardness of 71.7 to 77 HV, clearly demonstrating the biochar’s effectiveness as a rigid reinforcement filler.

To statistically evaluate which parameters mattered most, the researchers performed an Analysis of Variance (ANOVA). The ANOVA results unequivocally pointed to composition and injection temperature as the most influential factors, showing they govern the material’s mechanical behavior.Composition was the leading factor, contributing 50.42% of the total variation, followed closely by injection temperature at 42.67%. The Taguchi analysis further showed that tensile strength improved significantly up to 5% CCB and declined as temperature increased from 135∘C to 155∘C. Injection temperature had the highest contribution at 38.58%, with composition being the second most important at 20.14%. The maximum modulus was achieved at the lowest temperature investigated, 135∘C, reinforcing temperature’s critical role in material stiffness. Composition was the single most dominant factor, responsible for a commanding 78.3% of the total variation. Hardness consistently increased with higher CCB content.

Leveraging the structured experimental data, the research employed five machine learning (ML) models—Linear Regression, Support Vector Regression (SVR), Random Forest Regression (RFR), Gradient Boosting (GBR), and Extreme Gradient Boosting (XGBoost)—to predict the mechanical properties. These models were tasked with capturing the complex, non-linear relationships that traditional models often miss.

The results were outstanding. The ensemble models, particularly Gradient Boosting and XGBoost, demonstrated superior predictive performance. They achieved R2 values (a statistical measure indicating goodness of fit, with 1 being a perfect fit) of 0.9878 for tensile strength, 0.9628 for Young’s modulus, and 0.9645 for hardness. Their R2 values for tensile strength were essentially identical, confirming their exceptional capability to accurately predict the material’s response across various manufacturing conditions. Cross-validation confirmed the models’ reliability, with prediction errors contained within approximately 5% for tensile strength, under 10% for hardness, and around 13% for Young’s modulus.

This successful integration of Taguchi DOE, ANOVA analysis, and advanced ML models establishes a powerful, data-driven framework. This methodology is not only effective for optimizing injection molding parameters but also provides a sustainable path for manufacturing high-performance, eco-friendly composites using waste materials like coconut shells.

Source: Sundarasetty, H., Louhichi, B., Alrasheedi, N. H., Sahu, S. K., Lee, I. E., & Wali, Q. (2025). Machine learning guided process optimization and sustainable valorization of coconut biochar filled PLA biocomposites. Scientific Reports, 15(34647).