Document Type : Research Article
Authors
^{1} Easwari Engineering College (Autonomous), Chennai, Tamilnadu, 600089, India
^{2} B.S.Abdur Rahman Crescent Institute of Science and Technology, Vandalur, Chennai, Tamilnadu, 600048, India
^{3} Indian Institute of Technology, Chennai, Tamilnadu, 600036, India
Abstract
Keywords
Main Subjects
Research Article
Vinoth Viswanathan ^{a * } , Sathiyamurthy Subbarayan ^{a}, Ananthi Narayanaswamy ^{a}, Devi Panneerselvam ^{b}, Prabhakaran Jayasankar ^{c}
^{a} Easwari Engineering College (Autonomous), Chennai, Tamilnadu, 600089, India
^{b }B.S.Abdur Rahman Crescent Institute of Science and Technology, Vandalur, Chennai, Tamilnadu, 600048, India
^{c }Indian Institute of Technology, Chennai, Tamilnadu, 600036, India
ARTICLE INFO 

ABSTRACT 
Article history: Received: 20240417 Revised: 20240724 Accepted: 20240812 

The present study investigates the mechanical behaviors of hybrid fiberreinforced polyester composites in developing a new strengthened material. The experiments were planned as per the design of experiments, the selected input parameters were fiber length (mm), NaOH treatment (%), and fiber weight (%) and the output parameters were tensile, flexural, and impact strength conditions. A NonLinear Regression Modelling (NLRM) and Fuzzy logic model have been designed to predict and analyze the mechanical properties in unknown test conditions. Every input factor was categorized into three linguistic descriptors, while each output factor was classified into three linguistic categories. A triangular membership function was employed to define all these variables. The effectiveness of the nonlinear regression analysis and fuzzy logic model was evaluated through confirmatory experiments. The model predicted the mechanical results with an error of 7.19%, 5.38%, and 2.33% respectively. The proposed approach can significantly simplify reallife multiresponse optimization problems, thereby reducing fabrication costs and enhancing composite fabrication efficiency. 



Keywords: Hybrid fiber; Polyester; Mechanical properties; Regression analysis; Fuzzy rule. 


© 2025 The Author(s). Mechanics of Advanced Composite Structures published by Semnan University Press. This is an open access article under the CCBY 4.0 license. (https://creativecommons.org/licenses/by/4.0/) 
1. Introduction
In recent times, a burgeoning industrial consciousness has emerged concerning environmental impact, the diminishing reservoirs of petroleum resources, and a heightened emphasis on human health. This has precipitated a pivot in attention towards ecologically sustainable products, displacing the erstwhile emphasis on synthetic materials [1].
The substantial prevalence of natural fibers and alternative bioderived products has played a pivotal role in augmenting product sustainability in contrast to synthetic materials [2,3]. Furthermore, natural fibers provide notable advantages over synthetic materials, including low cost, low weight, biodegradability, higher strength, and diminished wear [46]. These mentioned attributes position natural fibers as suitable aspirants for broader utilization in biomedical, automobiles, furniture, and construction sectors, etc [78]. In recent times, plant fibers have been highly effectively utilized as reinforcement materials for preparing polymer matrix composites [9,39]. Examples of plant fiber reinforcement materials such as pineapple leaf, banana, coir, jute, roselle, luffa, and hemp, etc., can very effectively enhance the mechanical behaviors of fabricated composites [1014]. The development in rigidity is due to greater adhesion between the base matrix and fiber reinforcement contact, which is challenging when using natural fibers. The hydrophilic nature of natural fibers and the hydrophobic nature of the polymer matrix are the main causes of this behavior [15,16]. To address this concern, researchers frequently engage in surface modification of natural fibers by subjecting them to chemicalrich solutions as pretreatment. Chemically treating natural fibers enhances bond strength, thereby diminishing their hydrophilic behavior and bolstering interfacial strength [17,18]. Different types of chemical pretreatments involve significantly modify the fiber surface and dimensional stability characteristics [19]. The paddy straw fiber was immersed in 2 wt% sodium hydroxide for the duration of 48 hours [20,21]. Motaleb et al (2022) an examination of the impact of incorporating chemically treated pineapple fiber (ranging from 25 to 45 wt%) to reinforce polypropylene composite through compression molding reveals that formulations prepared with 45 wt% pineapple fiber exhibit improvements in tensile, bending, and impact strength compared to other composite preparations [22]. Soundar et al (2019) Response surface methodology (RSM) was used to organize trials in the fabrication of a hybrid composite that contained 0.78% of silica and 2 wt.% of roselle fiber to get the best flexural and tensile properties [23]. Balasubramanian et al (2016) Drilling operations analyze speed, feed, and depth of cut as input factors, and thrust force and torque as output responses. The average errors of the models, which showed a significant correlation with the experimental data, were 0.77% and 1.22%, respectively. As such, it is possible to regard the fuzzy modeling of thrust and torque as extremely precise. Moreover, within a specified range of values, fuzzy modeling makes it easier to predict reactions with confidence for every combination of cutting speed, feed rate, and depth of cut [24]. Thirumal et al (2016) The study employed fuzzy logic with the goal of forecasting the Indoor Air Quality (IAQ) parameters' transitional response under various input conditions. Comparing the suggested multiresponse fuzzy model to nonlinear regression models, the former demonstrated better prediction accuracy. Notably, the fuzzy model's absolute error percentages for temperature, relative humidity, and carbon dioxide levels were 2.05%, 3.81%, and 2.24%, respectively [25]. Sinha et al (2020) the study examined the wear characteristics of epoxy composites reinforced with abaca fiber and augmented by the inclusion of red mud filler. The investigation considered the impact of the weight percentage of abaca, red mud, and the particle size of red mud using a mathematical model (Response Surface Methodology) to optimize the sliding wear of hybrid composites. Additionally, experimental data were utilized to develop a fuzzy logic model for predicting the sliding wear of hybrid composites. The suggested fuzzy model demonstrated a predictive accuracy of 87% in estimating the sliding wear of hybrid composites [26]. Gangwar et al (2022) studied a Taguchi L27 orthogonal array to record the mechanical characteristics of kenaf fiber, NaOH concentration, and fiber immersion time. ANOVA analysis revealed NaOH concentration as the most influencing parameter, followed by kenaf wt.%. The greyfuzzy approach was used to convert multivariate features into a single reasoning grade, minimizing ambiguity and providing results close to a position value [27]. Guo et al (2022) study, a fuzzy reasoning technique are employed to identify the combination of input variables that can yield the best overall performance. The input variables include a 0° layup angle, a 260°C nozzle temperature, an 80% fiber filling rate, and a 0.6 mm layer thickness. [28].
Based on the abovementioned literature review, researchers encounter difficulties in optimizing various mechanical characteristics for composite formulations reinforced with natural fibers. Hybrid fiber (Paddy straw and pineapple leaf fiber) has not yet been reported in any literature study. Utilization of AgroWastages is used as reinforcement, turning waste into valuable resources. Hybrid Composites present a sustainable alternative to conventional materials, decreasing dependence on synthetic fibers and mitigating environmental impact. In the same context, hybrid fiberreinforced polyester composite formulations were prepared using a semiautomatic compression molding machine based on the L27 experimental design. Literature also supports the requirement for a chemical treatment to increase bond strength in fibers so the the hybrid fibers were chemically treated in varying percentages of NaOH for the duration of 24 hrs. Further influence of different types of fiber length and fiber weight percentage was also taken in this study. The mechanical properties of tensile, flexural, and impact strengths were maximized using nonlinear regression modeling and a fuzzy dual approach. The optimal combinations of fiber length, fiber weight, and NaOH concentration % for treatment.
Paddy straw fibers are collected from agricultural lands to mitigate wastage globally. The gathered agricultural waste undergoes a thorough cleaning process to eliminate unwanted materials and impurities. Similarly, pineapple leaves are also collected from agricultural fields and the fibers are extracted using a double roller fiber extracting machine. Both paddy straw stems and pineapple leaves undergo a chemical treatment process to enhance fiber surface and strength. NaOH was selected for chemical treatment due to its effectiveness compared to other treatment processes [29,30]. Various NaOH concentrations (1%, 3%, and 5%) are utilized for a duration of 24 hours at room temperature. The main objective of submerging the fibers in a chemical solution is to reduce their hydrophilic behavior, resulting in better interfacial strength with the matrix. This treatment expressively improves the mechanical characteristics of the fibers. Subsequently, the fibers are meticulously cleaned with distilled water to remove any residual chemicals. The fibers are then sunlightdried for 2 days to complete the process of enhancing their properties [38]. The matrix material used in the present study was polyester resin (thermosetting group), the accelerator (Methyl Ethyl Ketone Peroxide), and catalyst (Cobalt octoate) used as a curing agent [31]. The resin and the curing agent were purchased from Go Green Products, Chennai, India. The polyester resin properties are density 1.132 (g/cm3), viscosity @ 25°C – 470 (cp), volatile content – 36.2 (%), acid value – 25.18 (mg KOH/gm), gel time @ 25°C – 14 min [11]. Figure 1 shows chemically treated paddy straw and pineapple leaf fiber.
Fig. 1. Chemically treated fibers: (a) Paddy straw, (b) Pineapple leaf fiber
The fabrication of hybrid fiberreinforced composite preparations of chemically treated fiber (1%, 3%, 5%), fiber length (25mm, 50mm, 75mm), and fiber weight (30%, 40%, 50%) was carried out using semiautomatic compression molding machine with the mold dimensions of 300 x 300 x 3 mm [32]. The polythene sheet was located between the base plate and mold for easy removal of the composite plate. The Wax content was applied to the mold surface and the polyethylene sheet to prevent the liquid polyester from adhering to the mold and base plate surfaces. Polyester resin, combined with a catalyst and accelerator at a ratio of 1:0.015:0.015, was manually stirred for 23 minutes to prevent agglomeration [33]. Now, weighted fiber is placed in the bottom mold and mixed resin is poured into the entire mold uniformly in a circular way in the mold. Further, the upper die was compressed into the bottom mold, in the same context the temperature 50°C and 30 bar pressure was applied for 45 mins. After 45 mins the eject the fabricated composite laminate was placed at room temperature. 27 Laminates were fabricated in accordance with the L27 experimental design. The stepwise experimentation procedure is shown in Figure 2.
Fig. 2. Stepwise procedure for experimentation methodology
The fabricated composite laminate cutting with the ASTM standard. Tensile strength tests for composite samples were conducted the ASTM D638 standard [34], utilizing a universal testing machine (UTM) with a gauge length of 100mm and a constant crosshead speed of 1.5 mm/min.
The flexural test was conducted using the 3point bending test in accordance with the ASTM D790 [35] standard on a universal testing machine. The flexural strength was performed by the following equation:
Flexural strength (𝜎)= 3𝑝𝑙 /2𝑏𝑡^{2} 

where p – Maximum load, l – Span length, b width of the specimen, t – thickness of the specimen.
The Charpy impact test is a standard method used to measure the impact strength. The test assesses the energy absorbed by a material during fracture. Which is determined according to ASTM D25610 [36] standards using the Charpy test. In this test, a notch with a depth of 2 mm and a subtended angle of 45° is created, and the energy released upon fracturing the sample is recorded.
Design of Experiments (DOE) is a systematic approach used to determine the relationship between factors that affect a process and the resulting output of that process. Incorporating fuzzy rules into DOE involves considering uncertainty and imprecision in the factors and their levels. The assigned input variables and their level are in Table 1.
Table 1. Input variables and their level
Level 
Fiber length (mm) 
NaOH Treatment (%) 
Fiber weight (%) 
Low 
25 
1 
30 
Medium 
50 
3 
40 
High 
75 
5 
50 
According to the full factorial design of experiments, 27 sets of interpretations were recorded and then categorized into three levels for the development of a multipurpose fuzzy model, as shown in Table 2.
Table 2. Output variables and their level
Level 
Tensile Strength (MPa) 
Flexural Strength (MPa) 
Impact Strength (KJ/m2) 
Low 
1530 
3044 
3661 
Medium 
3145 
4559 
6287 
High 
4660 
6074 
87112 
The tensile, flexural, and impact strength for 27 polyester composite laminates on different settings of input parameters using DOE were recorded and reported in Table 3.
Table 3. Experimental results using Design of experiments
S.No 
Fiber length (mm) 
NaOH Treatment (%) 
Fiber Weight (%) 
Tensile Strength (MPa) 
Flexural Strength (MPa) 
Impact Strength (KJ/m2) 
1 
25 
1 
30 
15.6 
29.6 
50.6 
2 
25 
1 
40 
21.8 
32.9 
62.8 
3 
25 
1 
50 
27.1 
40.7 
88.3 
4 
25 
3 
30 
18.7 
33.4 
59.3 
5 
25 
3 
40 
29.4 
41.4 
71.8 
6 
25 
3 
50 
36.9 
51.2 
96.4 
7 
25 
5 
30 
22.1 
40.6 
68.9 
8 
25 
5 
40 
39.2 
53.1 
83.3 
9 
25 
5 
50 
47.5 
60.2 
112.3 
10 
50 
1 
30 
21.5 
33.7 
48.4 
11 
50 
1 
40 
28.7 
43.8 
53.9 
12 
50 
1 
50 
44.6 
55.7 
78.4 
13 
50 
3 
30 
23.8 
42.4 
54.3 
14 
50 
3 
40 
33.6 
48.5 
58.1 
15 
50 
3 
50 
51.9 
61.3 
82.6 
16 
50 
5 
30 
28.4 
55.2 
61.7 
17 
50 
5 
40 
39.3 
60.6 
73.4 
18 
50 
5 
50 
60.4 
73.1 
87.37 
19 
75 
1 
30 
18.3 
32.4 
36.8 
20 
75 
1 
40 
23.6 
38.9 
41.4 
21 
75 
1 
50 
37.9 
49.5 
58.4 
22 
75 
3 
30 
21.8 
37.8 
40.7 
23 
75 
3 
40 
30.6 
41.1 
49.9 
24 
75 
3 
50 
38.1 
52.4 
62.6 
25 
75 
5 
30 
29.3 
36.8 
40.8 
26 
75 
5 
40 
32.7 
44.6 
50.3 
27 
75 
5 
50 
39.4 
55.4 
72.2 
Nonlinear regression in statistics involves modeling observational data using a function that is a nonlinear combination of model parameters and depends on one or more independent variables [25].
yu = φ (x1u, x2u, ..., xku) + εu 
(1) 
where u=1, 2, 3, k, and k represent the number in the factorial experiment. The term xiu represents the level of the ith factor in the uth experiment. The function φ is called the response surface. The residual εu measures the experimental error in the uth observation.
The first order polynomials formula,
yu = β0 + β1x1u + β2x2u +· · ·+βkxku + εu 
(2) 
The secondorder polynomial, specifically known as the Quadratic response surface has 3−x variables, and takes the form:
yu =β0+β1x1u +β2x2u+β3x3u +β12x1ux2u +β23x2ux3u+ β13x1ux3u+ εu + β11x1u2 + β22x2u2 + β33x3u2 
(3) 
"εu" is a term representing additional sources of variability not accounted for. This encompasses measurement errors in the response, other inherent sources of variation in the process or system, and so forth. Quadratic design models were chosen based on superior values of the coefficient of correlation and Ftest for all responses in this study using statistical software.
Fuzzy logic simulates human decisionmaking by employing linguistic reasoning, integrating a mathematical theory that combines multivalued logic, probability theory, and artificial intelligence methods. It has been proven to be effective in addressing complex problems. Mamdani and Sugeno represent two distinct types of fuzzy inference systems (FISs) within the fuzzy logic toolbox. These two types of inference systems vary in the manner in which they specify outputs. The Mamdanitype fuzzy inference produces a fuzzy set output, whereas the Sugenotype inference provides either a constant output or a linear mathematical expression [37]. In this present investigation, Mamdani FIS was used and the structure of the fuzzy logic system is shown in Figure 3.
Fig. 3. Fuzzy model for Mechanical strength parameters
Meaningful linguistic statements are selected for each variable and represented using appropriate fuzzy sets, such as "low," "medium," and "high." The concept of fuzzy reasoning for a threeinput–threeoutput multiresponse fuzzy logic model is described as follows. The fuzzy rule base consists of a group of IFTHEN statements with three inputs, x1, x2, and x3, and three outputs y1, y2, and y3,
Rule 1: If x1 is A1 and x2 is B1 and x3 is C1 then y1 is P1 and y2 is Q1 and y3 is R1
Rule 2: If x1 is A2 and x2 is B2 and x3 is C2 then y1 is P2 and y2 is Q2 and y3 is R2
Rule n: If x1 is An and x2 is Bn and x3 is Cn then y1 is Pn and y2 is Qn and y3 is Rn (4)
An, Bn, and Cn function as input membership functions for the fuzzy models, while Pn, Qn, and Rn serve as output membership functions for the multiresponse fuzzy models.
Triangular and trapezoidal membership functions were applied to both input and output conditions in the fuzzy model. The generated membership function plots for input and output variables are depicted in Figures 4 and 5, respectively. The triangular membership function (trimf) is the most basic, created using three points to form a triangle, while the trapezoidal membership function (trapmf) is a truncated triangle defined by four points.
The Fuzzy Rule Editor is employed to generate 27 sets of rules based on the specified input and output variables using the FIS editor. Within the Fuzzy model, these rules are organized in such a way that the first three columns correspond to the input variables, while the remaining columns pertain to the output variables. These rules provide a comprehensive definition of the Fuzzy Inference System, encompassing variables, membership functions, and the essential rules for computing intermediate mechanical strength parameters.
The Fuzzy Rule Viewer illustrates the entire fuzzy inference process, presenting each row as an ifthen rule following a specific syntax. In this representation, each column corresponds to a variable, and the rule numbers are positioned next to each row. The rule viewer provides a detailed microview of the Fuzzy Inference System, showcasing every calculation in great detail. Furthermore, the Surface Viewer displays the comprehensive output of the fuzzy system, encompassing the entire span of the input set.
Fig. 4. (ac) Input parameters and their membership plots
Fig. 5. (ac) Output parameters and their membership plots
3. Results and Discussion
In statistics, the coefficient of determination, Rsquared (R2), signifies the portion of variability in a dataset that the statistical model can account for. R2 values of 0.90, 0.98, and 0.94 were achieved for the fiber length, NaOH treatment, and fiber weight models, respectively. The scientific relationships for correlating the input process variables are derived from the coefficients obtained through the use of Design Expert software. The terms l, n, and w are Fiber length, NaOH treatment, and Fiber weight whereas Ts, Fs, and Is represents Tensile strength, Flexural strength, and Impact strength respectively.
Ts = 26.35 +1.45928 l +1.78056 n  0.124028 w
 0.037833 ln 0.003033 lw + 0.055417 nw
 0.011947 l2 + 0.108333 n2 +0.014167 w2
(5)
The relevance of adding new model terms beyond those in the original model was assessed using the Fvalue. It was evident from a decreased pvalue (Probability > F) that the addition of secondorder variables enhanced the model. The coefficients derived from the Design Expert software output correspond mathematically to the flexural strength (Fs) and the process variables under consideration.
Fs =  8.45185+1.73217 l + 3.825 n  0.632639 w
 0.057833 ln + 0.0006 lw + 0.024583 nw
 0.015698 l2 + 0.247222 n2 + 0.017556 w2
(6)
The coefficients produced by the Design Expert software output are used to establish the mathematical relationship between the Impact strength (Is) and the process variables under consideration.
Is = 72.87644 + 0.663522 l + 3.30389 n
 2.77957 w  0.060167 ln  0.014433 lw
+ 0.046542 nw  0.004404 l2 +0.248194 n2
+ 0.061261 w2
(7)
The input and output variables undergo fuzzification and are expressed using membership functions. The output variables are categorized into three levels each, resulting in a formulation of 27 fuzzy rules (Table 4).
Table 4. Fuzzy set of rules
Rule no 
Fiber length 
Treatment 
Weight (%) 
Tensile 
Flexural 
Impact 
1 
Low 
Low 
low 
low 
low 
low 
2 
Low 
Low 
medium 
low 
low 
medium 
3 
Low 
Low 
high 
low 
low 
high 
4 
Low 
medium 
low 
low 
low 
low 
5 
Low 
medium 
medium 
low 
low 
medium 
6 
Low 
medium 
high 
medium 
medium 
high 
7 
Low 
high 
low 
low 
low 
medium 
8 
Low 
high 
medium 
medium 
medium 
medium 
9 
Low 
high 
high 
high 
high 
high 
10 
Medium 
Low 
low 
low 
low 
low 
11 
Medium 
Low 
medium 
low 
low 
low 
12 
Medium 
Low 
high 
medium 
medium 
medium 
13 
Medium 
medium 
low 
low 
low 
low 
14 
Medium 
medium 
medium 
medium 
medium 
low 
15 
Medium 
medium 
high 
high 
high 
medium 
16 
Medium 
high 
low 
low 
medium 
medium 
17 
Medium 
high 
medium 
medium 
high 
medium 
18 
Medium 
high 
high 
high 
high 
high 
19 
High 
Low 
low 
low 
low 
low 
20 
High 
Low 
medium 
low 
low 
low 
21 
High 
Low 
high 
medium 
medium 
low 
22 
High 
medium 
low 
low 
low 
low 
23 
High 
medium 
medium 
medium 
low 
low 
24 
High 
medium 
high 
medium 
medium 
medium 
25 
High 
high 
low 
low 
low 
low 
26 
High 
high 
medium 
medium 
medium 
low 
27 
High 
high 
high 
medium 
medium 
medium 
The fuzzy rule viewer for the confirmation test is depicted in Figures 6 and 7. To validate the fuzzy system, an experimental value (Fiber length 75, NaOH treatment 5%, and fiber weight 50%) was provided, yielding output values (Tensile strength 37.5 MPa, Flexural strength 51.7 MPa, and Impact strength 74.2 KJ/m^{2}) that closely match the experimental results.
Fig. 6. Fuzzy rule reading
Fig. 7. Fuzzy rule viewer of mechanical strength parameters
The interaction effects of input parameters on the responses were examined using surface plots. The interaction effects of Fiber length, NaOH treatment, and Fiber weight on Tensile strength, Flexural strength, and Impact strength are shown in Figures 8,9,10.
Fig. 8. Surface view plots for tensile strength: (a) NaOH treatment vs Fiber length, (b) Fiber weight vs Fiber length,
and (c) Fiber weight vs NaOH treatment
Fig. 9. Surface view plots for Flexural strength: (a) NaOH treatment vs Fiber length, (b) Fiber weight vs Fiber length,
and (c) Fiber weight vs NaOH treatment
Fig. 10. Surface view plots for Impact strength: (a) NaOH treatment vs Fiber length, (b) Fiber weight vs Fiber length,
and (c) Fiber weight vs NaOH treatment
Confirmation experiments were conducted under six distinct sets of input conditions. The experimental data and the predicted values obtained from both the mathematical model and fuzzy logic were compared (see Table 5).
% of error = (Experimental value  Predicted value) × 100 / Experimental value
It was obtained that the average absolute error percentages for input parameters in the Nonlinear regression model are shown in Table 6.
Tensile strength = 7.19 %
Flexural strength = 5.38 %
Impact strength = 2.33 %
It was obtained that the average absolute error percentages for input parameters in the fuzzy model are shown in Table 6.
Tensile strength = 6.61 %
Flexural strength = 6.22 %
Impact strength = 9.09 %
Table 5. Comparison of experimental values with predicted values
Run 
Fiber length (mm) 
NaOH treatment (%) 
Fiber weight (%) 
Experimental values 
Predicted values 
Predicted values 

Ts 
Fs 
Is 
Ts 
Fs 
Is 
Ts 
Fs 
Is 

1 
25 
3 
30 
18.7 
33.4 
59.3 
17.9 
33.9 
59.4 
18.6 
35.3 
44.1 
2 
25 
3 
50 
36.9 
51.2 
96.4 
39.9 
51.1 
97.5 
37.5 
51.7 
104 
3 
50 
1 
30 
21.5 
33.7 
48.4 
22.8 
38.5 
47.1 
18.6 
35.3 
44.1 
4 
50 
5 
40 
39.3 
60.6 
73.4 
41.6 
58.6 
70 
37.5 
68.3 
74.2 
5 
75 
3 
30 
21.8 
37.8 
40.7 
20.9 
34.2 
39.9 
18.6 
35.3 
44.1 
6 
75 
5 
50 
39.4 
55.4 
72.2 
45 
57.4 
69.7 
37.5 
51.7 
74.2 
Table 6. Comparison of error percentage
Run 
Fiber length (mm) 
NaOH treatment (%) 
Fiber weight (%) 
Predicted values 
Predicted values 

Ts 
Fs 
Is 
Ts 
Fs 
Is 

1 
25 
3 
30 
4.34 
1.45 
0.26 
0.53 
5.68 
25.63 
2 
25 
3 
50 
8.08 
0.19 
1.09 
1.62 
0.97 
7.88 
3 
50 
1 
30 
6.44 
14.39 
2.72 
13.48 
4.74 
8.88 
4 
50 
5 
40 
5.91 
3.20 
4.62 
4.58 
12.70 
1.08 
5 
75 
3 
30 
4.16 
9.44 
1.87 
14.68 
6.61 
8.35 
6 
75 
5 
50 
14.24 
3.66 
3.44 
4.82 
6.67 
2.77 
Average error percentage 
7.19 
5.38 
2.33 
6.61 
6.22 
9.09 
The comparisons of experimental and predicted values of Tensile, Flexural, and Impact strength are shown in Figures 11, 12, and 13 respectively. The accuracy of NLRM (Nonlinear regression model) compared with fuzzy model prediction was studied using these validation plots. Include surface plots comparing experimental and predicted values to visually assess the model’s performance across different conditions.
Fig. 11. Comparison of experimental and predicted values of Tensile strength
Fig. 12. Comparison of experimental and predicted values of Flexural strength
Fig. 13. Comparison of experimental and predicted values of Impact strength
4.Conclusions
This paper showcases the optimization of diverse mechanical characteristics in hybrid fiberreinforced polyester composite formulations. The methodology integrates nonlinear regression analysis and fuzzy logic to facilitate the identification of optimal input fabrication parameters. A total of 27 experiments were performed to evaluate different input parameter configurations, during which multiple mechanical properties were documented. The input fabrication parameters encompassed hybrid fiber length (mm), NaOH concentration in a chemical solution (%), and fiber weight percentage (%). The corresponding output characteristics included tensile strength, flexural strength, and impact strength. The study analyzed the interaction effects and significance of variables on the outputs through surface plots. Additionally, a systematic fuzzy model was formulated to predict the responses within the specified range of conditions. The optimal input parameter levels that maximize tensile strength, flexural strength, and impact strength are as follows: a fiber length of 75, NaOH concentration at 5%, and a fiber weight of 50%. The Nonlinear Regression Model (NLRM) accurately predicted the mechanical characteristics and validation results indicated an error percentage of 4.96% for all the models, whereas the error percentage ranged up to 7.30% in the fuzzy model. This suggested approach has the potential to significantly simplify the complexity inherent in reallife multiresponse optimization problems, ultimately reducing fabrication costs and enhancing efficiency in composite fabrication. In this composite can be used in a wide range of applications including automotive parts, building materials, packaging, furniture, and consumer goods, demonstrating their versatility and broad market potential. The biodegradability of many agricultural fibers improves the composites' environmental friendliness and aids in waste disposal at the end of their life cycle.
The authors would like to thank Dr Krishna Kanth Pulicherla, Scientist, TDT Division, DST, Delhi, India, for their valuable suggestion to do this work, and also the authors wish to express their sincere thanks to E. Devarajan, farmer, Chinnakavanam, Thiruvallur (District), Tamilnadu, India, for the collection of paddy straw fibers.
The author(s) disclosed receipt of the following financial support for the research, authorship, and publication of this article. This article is supported by the scheme of Innovation, Technology Development, and Deployment (1819), Department of Science and Technology  Delhi. Ref no: DST/TDT/WM/2019/78 Consortia (G).
The author declares that there is no conflict of interest regarding the publication of this article.
[1] Nimanpure S, Hashmi SA, Kumar R, Nigrawal A, Bhargaw HN, Naik A, 2018. Sisal fibril epoxy composite a high strength electrical insulating material. Polymer Composites, 39(S4), pp. E2175E2184.
[2] Rafiqah SA, Khalina A, Harmaen AS, Tawakkal IA, Zaman K, Asim M, Nurrazi MN, Lee CH, 2021. A review on properties and application of biobased poly (butylene succinate). Polymers, 13(9) p.1436.
[3] Platnieks O, Gaidukovs S, Thakur VK, Barkane A, Beluns S. Biobased poly (butylene succinate), 2021. Recent progress, challenges and future opportunities. European Polymer Journal, 161, p.110855.
[4] Hashmi SA, Naik A, Chand N, Sharma J, Sharma P, 2011. Development of environment friendly hybrid layered sisal–glass–epoxy composites. Composite Interfaces, 18(8), pp. 67183,
[5] Vinoth, V., Sathiyamurthy, S., Saravanakumar, S. and Senthilkumar, R., 2024. Integrating response surface methodology and machine learning for analyzing the unconventional machining properties of hybrid fiber‐reinforced composites. Polymer Composites, 45(7), pp. 60776092. https://doi.org/10.1002/pc.28180
[6] Mohammed L., Ansari M.N., Pua G., Jawaid M., Islam M.S., 2015. A review on natural fiber reinforced polymer composite and its applications. International Journal of polymer science, 2015(1), p.243947.
[7] Karimah A., Ridho M.R., Munawar S.S., Adi D.S., Damayanti R., Subiyanto B., Fatriasari W., Fudholi A., 2021. A review on natural fibers for development of ecofriendly biocomposite: characteristics, and utilizations. Journal of materials research and technology, 13, pp. 24422458.
[8] Jayan, J.S., Appukuttan, S., Wilson, R., Joseph, K., George, G. and Oksman, K., 2021. An introduction to fiber reinforced composite materials. In Fiber reinforced composites (pp. 124). Woodhead Publishing.
[9] Mahesh, V., Joladarashi, S. and Kulkarni, S.M., 2021. A comprehensive review on material selection for polymer matrix composites subjected to impact load. Defence Technology, 17(1), pp. 257277.
[10] Saravanakumar, S., Sathiyamurthy, S., Ananthi, N. and Devi, P., 2023. Optimization of drilling characteristics of Al2O3 and boiled eggshell filleradded hybrid bio composite from agriculture residue. Biomass Conversion and Biorefinery, pp. 115.
[11] Sathiyamurthy, S., Vinoth, V., Ananthi, N. and Devi, P., 2024. The effect of fiber stacking sequence on mechanical and morphological behavior of paddy straw/pineapple leaf fiberreinforced ortholaminated polyester hybrid composites. Proceedings of the Institution of Mechanical Engineers, Part E: Journal of Process Mechanical Engineering, 238(1), pp. 463473.
[12] Saravanakumar, S., Sathiyamurthy, S., Vinoth, V. and Devi, P., 2024. Effect of alumina on epoxy composites with banana fiber: mechanical, water resistance and degradation property analysis. Fibers and Polymers, 25(1), pp. 275287.
[13] Sathiyamurthy, S., Vinoth, V., Saravanakumar, S. and Devi, P., 2023. Investigation on mechanical behaviour of calcium carbonate and groundnut Shell filler added Cocos Nucifera Fiber reinforced polyester composites. ARPN Journal of Engineering and Applied Science, 18(12), pp. 14251432.
[14] Thiruchitrambalam, M., Athijayamani, A., Sathiyamurthy, S. and Thaheer, A.S.A., 2010. A review on the natural fiberreinforced polymer composites for the development of roselle fiberreinforced polyester composite. Journal of Natural Fibers, 7(4), pp. 307323.
[15] Rajeshkumar, G., Seshadri, S.A., Ramakrishnan, S., Sanjay, M.R., Siengchin, S. and Nagaraja, K.C., 2021. A comprehensive review on natural fiber/nano‐clay reinforced hybrid polymeric composites: Materials and technologies. Polymer Composites, 42(8), pp. 36873701.
[16] Nurazzi, N.M., Harussani, M.M., Aisyah, H.A., Ilyas, R.A., Norrrahim, M.N.F., Khalina, A. and Abdullah, N., 2021. Treatments of natural fiber as reinforcement in polymer compositesa short review. Functional Composites and Structures, 3(2), p.024002.
[17] Vinoth, V., Sathiyamurthy, S., Ananthi, N. and Elaiyarasan, U., 2022. Chemical treatments and mechanical characterisation of natural fibre reinforced composite materialsa review. International Journal of Materials Engineering Innovation, 13(3), pp. 208221.
[18] Gholampour, A. and Ozbakkaloglu, T., 2020. A review of natural fiber composites: Properties, modification and processing techniques, characterization, applications. Journal of Materials Science, 55(3), pp.829892.
[19] Kabir, M.M., Wang, H., Lau, K.T. and Cardona, F., 2012. Chemical treatments on plantbased natural fibre reinforced polymer composites: An overview. Composites Part B: Engineering, 43(7), pp.28832892.
[20] AliEldin, S.S., Abd ElMoezz, S.M., Megahed, M. and Abdalla, W.S., 2021. Study of hybridization effect of new developed rice straw mat/glass fiber reinforced polyester composite. Journal of Natural Fibers, 18(8), pp. 11941206.
[21] Megahed, M., AliEldin, S.S., Abd El Moezz, S.M. and Abdalla, W.S., 2020. Synthesis of developed rice straw sheets and glass fiberreinforced polyester composites. Journal of Composite Materials, 54(23), pp.33813394.
[22] Motaleb, K.A., Shariful Islam, M. and Hoque, M.B., 2018. Improvement of physicomechanical properties of pineapple leaf fiber reinforced composite. International Journal of Biomaterials, 2018(1), p.7384360.
[23] Soundhar, A., Rajesh, M., Jayakrishna, K., Sultan, M.T.H. and Shah, A.U.M., 2019. Investigation on mechanical properties of polyurethane hybrid nanocomposite foams reinforced with roselle fibers and silica nanoparticles. Nanocomposites, 5(1), pp.112.
[24] Balasubramanian, K., Sultan, M.T.H., Cardona, F. and Rajeswari, N., 2016, October. Machining analysis of natural fibre reinforced composites using fuzzy logic. In IOP Conference Series: Materials Science and Engineering (Vol. 152, No. 1, p. 012051). IOP Publishing.
[25] Thirumal, P., Amirthagadeswaran, K.S. and Jayabal, S., 2015. Fuzzy rule based prediction of IAQ characteristics in air conditioned car. Iranian Journal of Science and Technology Transactions of Mechanical Engineering, 39(M2), pp.437450.
[26] Sinha, A.K., Narang, H.K. and Bhattacharya, S., 2021. Experimental determination, modelling and prediction of sliding wear of hybrid polymer composites using RSM and fuzzy logic. Arabian Journal for Science and Engineering, 46, pp.20712082.
[27] Gangwar, S., Dubey, N.K. and Pathak, V.K., 2022. Experimental investigation and multiobjective optimization of multiple mechanical characteristics for chemically treated kenaf fibre reinforced epoxy composite using grey fuzzy logic. Proceedings of the Institution of Mechanical Engineers, Part L: Journal of Materials: Design and Applications, 236(1), pp.166179.
[28] Guo, A., Liu, C., Li, S., Zhou, X., Wang, J., Wang, S., Qu, P. and Hu, Y., 2022. Water absorption rates and mechanical properties of material extrusionprinted continuous carbon fiberreinforced nylon composites. Journal of Materials Research and Technology, 21, pp. 30983112.
[29] Vinoth, V., Sathiyamurthy, S., Ananthi, N., and Jayabal, S., 2024. Mechanical characterization and study on morphological properties: Natural and agro waste utilization of reinforced polyester hybrid composites. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 238(19), pp. 95779588.
[30] Heidarinejad, Z., Dehghani, M.H., Heidari, M., Javedan, G., Ali, I. and Sillanpää, M., 2020. Methods for preparation and activation of activated carbon: a review. Environmental Chemistry Letters, 18, pp. 393415.
[31] Naguib, H.M., Kandil, U.F., Hashem, A.I. and Boghdadi, Y.M., 2015. Effect of fiber loading on the mechanical and physical properties of “green” bagasse–polyester composite. Journal of Radiation Research and Applied Sciences, 8(4), pp.544548.
[32] Ramprasath, R., Jayabal, S. and Sathiyamurthy, S., 2014. Optimization of mechanical behaviors of red mud particulated coir–vinyl ester composites using intuitive analysis based on response surface methodology and simulated annealing. Transactions of the Indian Institute of Metals, 67(6), pp.971977.
[33] Jayabal, S., Sathiyamurthy, S., Loganathan, K.T. and Kalyanasundaram, S., 2012. Effect of soaking time and concentration of NaOH solution on mechanical properties of coir–polyester composites. Bulletin of Materials Science, 35(4), pp.567574.
[34] Saravanakumar, S., Sathiyamurthy, S., and Vinoth, V., 2024. Enhancing machining accuracy of banana fiberreinforced composites with ensemble machine learning. Measurement, 235, p.114912.
[35] Ashok, K.G., Kalaichelvan, K. and Damodaran, A., 2022. Effect of nano fillers on mechanical properties of luffa fiber epoxy composites. Journal of Natural Fibers, 19(4), pp.14721489.
[36] Ashok, K., Ajith, D., Bibin, C., Sheeja, R. and Nishanth, R., 2022. Influence of nanofiller lignite fly ash on tribomechanical performance of sansevieria roxburghiana fiber reinforced epoxy composites. Journal of Natural Fibers, 19(13), pp.60006014.
[37] Saravanakumar, S., Sathiyamurthy, S., Ananthi, N. and Devi, P., 2023. Optimization of drilling characteristics of Al2O3 and boiled eggshell filleradded hybrid bio composite from agriculture residue. Biomass Conversion and Biorefinery, pp.115.
[38] Vinod, A., Sanjay, M. R., Siengchin, S., & Fischer, S. 2021. Fully biobased agrowaste soy stem fiber reinforced bioepoxy composites for lightweight structural applications: influence of surface modification techniques. Construction and Building Materials, 303, p.124509.
[39] Vinod, A., Sanjay, M. R., Suchart, S., & Jyotishkumar, P. 2020. Renewable and sustainable biobased materials: An assessment on biofibers, biofilms, biopolymers and biocomposites. Journal of Cleaner Production, 258, p.120978.