Multiobjective Optimization of Hydrophobic Process for Performance Analysis of Cover Date Bunch

ABSTRACT

Technologies like bunch covers have been shown to improve quality of date fruits. It is one of the approaches which are believed effective to protect the most sensitive stage of date fruit from autumnal rain and carob date. An experimental procedure is set up to allow farmers to predict, analyze, and optimize cover bunch performance after its use in protection applications and to select the ideal bag. The cover quality is designed based on a mathematical model and further optimized using MINITAB. From this analysis, the impact of each individual parameter on the desired response parameters has been studied. The results proved that the best cover could be found in optimal conditions. Using the influential parameters as tear strength, tensile strength, weight, air permeability and resistance to water penetration, the results indicate that the cover bunch remains influenced after harvesting period by the resistance to water penetration properties and tear strength as well.

摘要

像束盖这样的技术已经被证明可以提高椰枣的质量。这是保护枣果实最敏感阶段免受秋雨和角豆枣侵害的有效方法之一. 建立了一个实验程序，使农民能够预测、分析和优化盖束在保护应用中使用后的性能，并选择理想的袋子. 封面质量是基于数学模型设计的，并使用MINITAB进行进一步优化. 通过该分析，研究了每个单独参数对所需响应参数的影响. 结果证明，在最佳条件下，可以找到最佳覆盖层. 使用撕裂强度、拉伸强度、重量、透气性和耐水渗透性等影响参数，结果表明，覆盖层束在收获期后仍然受到耐水渗透性和撕裂强度的影响.

KEYWORDS:

• Date palm
• bagging product
• RSM
• ANOVA
• multicriteria

关键词:

• 棕榈
• 套袋产品
• 多准则

Introduction

Date Bunch covering is an old practice (Alyafei et al. 2022; Awad 2007; Nadeem et al. 2022). Date bunch with appropriate cover is an essential cultural practice that decreases the rate of fruit loss mainly during harvesting period and increase the humidity around the fruits, meanwhile not protecting the date bunch may expose date fruit to a higher rate of rot, moisture and deficiency, in the ripening stage (Al-Falahy, Omar, and Thamer 2020). These limitations will negatively affect the economic assessment of the harvested dates (Awad 2007). For improving date fruit quality and yield, bunch covering alone or in combination with other treatments had been explored (Matteo et al. 2018). Different types of materials have been used including cloth, plastic film, jute bag, mosquito net, paper Kraft or nonwoven product have been used (Guedri, Jaoudi, and Msahli 2022). The requirements for the ideal cover date bunch were known as early as 1935, and elaborated in 1949 by Bliss et al in his classic work. The cover bunch must allow air circulation throughout the fruit cluster because the water vapors transpired by the date fruit surfaces can be trapped by the bag and leads to fungus infection and water injury; must be waterproof during heavy rains to exclude insects and birds, and costs only a nominal amount (Denis 2001). According to Guedri et al (Guedri, Jaouadi, and Msahli 2016), few studies have dealt with the cover properties related to its quality. The technical quality of the bagging product is the set of properties that must be achieved to satisfy the farmer and the consumer (Nadeem et al. 2022). Guedri et al (Guedri, Jaouadi, and Msahli 2020) defined the ideal cover date bunch as a multi-criteria phenomenon and its identification requires the simultaneous satisfaction of all criteria (Abidi et al. 2021). Various works have evaluated how date bunch cover could enhance date fruit quality (Ahmed et al. 2022). However, in the general literature survey, the quality of the covers has not been studied. Based on previous bibliographic studies (Guedri, Jaouadi, and Msahli 2016), it was important to carry out scientific research to improve parameters affecting the quality of cover date bunch (Alyafei et al. 2022). For that reason, the present study focused on the optimization of the parameters after harvesting period. The objective was to predict the evolution of some responses such as the tensile strength (F), tear strength (MRR), weight (Pc), air permeability and resistance to water penetration (Ra). In fact, the ANOVA and RSM approach were exploited. A plot analysis, showing the effects of hydrophobic treatment on the cover characteristics of the finished surface, was highlighted.

The objective of this approach was to optimize the responses simultaneously: tensile strengh (T), tear strength (TS), air permeability (A), resistance to water penetration (P) and weight (W). The optimal conditions were established using an experimental design, mathematical method and using superimposed contour diagrams.

Material and methods

Materials

The nonwoven cover works better than the plastic and the mosquito net in terms of cluster protection, which depreciates the quality of dates and additional treatment for these fruits. The selected nonwoven bagging products are white polyester, nonwoven, with porosity from 69.6% to 78.9% having a weight 90 g/m2. The dimension date bunch cover is about 100 cm length and 85 cm diameter at the bottom and 0.3 cm thickness.

Experimental treatments

The cover bunch was treated with a water repellent agent to offer protection against rain and the carob moth. A solution of Tubicoat hp27 (water repellent supplied by CHT Bezema) was prepared to ensure the hydrophobic property with concentration from 20 g/L to 50 g/L. A padding process was chosen to develop the optimal recipe with a 75% pick up after bath treatment with a calendar having finishing rolls, a pressure from 1 bar to 2 bars, and a speed value from 10 rpm to 30 rpm. Then, heat treatment was used to thermally treat the bags for 2 minutes. The main requirements for the hydrophobic compounds are their ability to form on each of the fibers a continuous film and its ability to ensure the stability of the waterproofing effect due to a strong chemical bond with the textile material, or because of their insolubility in water. In the proposed treatment, the water repellency properties of the fabric are obtained through the formation of a thin film. The film protects the fibers against tearing and pulling.

The present study evaluated by the following methods:

• Resistance to water penetration (ISO 20,811)

• Tensile strength (ISO 13,937)

• Tear strength (ISO 9073:3)

• Air permeability (ISO 9237)

• Weight (ISO 9073:1)

Statistical analysis

In this study, in order to ensure the reliability of the obtained experimental information, all tests were replicated three times. The average values of the measurements were used for statistical analysis to determine the effects of Tubicoat hp27 concentration, speed and cylinder pressure (three influence factors) on cover characteristics. Standard deviation for statistical analysis was estimated by means of Microsoft Excel 2007.

In order to obtain the optimized parameters of treated nonwoven cover, the analysis was conducted by statistical analysis with the help of Minitab 16 software.

The Minitab 16 made it possible to define the individual desirability functions of each response to be measured. It takes into consideration:

• The values of all the answers

• The weight associated with each answer

• The values taken by the individual desirability functions

• The value taken by the global desirability function.

Results and discussions

The present study comprises the use of bags to protect date fruit from insects, rain, wind, birds, and sunburn. These covers are prepared of flexible nonwoven materials that allow free transmission of air throughout the fruit cluster.

Developing an empirical model

The empirical model for tensile strength (T), tear strength (TS), weight (W), resistance to water penetration (R), and air permeability (A) of the cover bunch are a function of Tubicoat hp concentration (C), pressure (P), cylinder speed (S) which is mentioned in Equation 1–(1) $\begin{array}{rl}& \mathit{Tensile}\mathit{\hspace{0.17em}}\mathit{strength}\mathit{\hspace{0.17em}}\left(\mathit{T}\right)\mathit{\hspace{0.17em}}=\mathit{\hspace{0.17em}}23,867+2,113\mathit{C}+2,459\mathit{P}+2,871\mathit{S}+6,738\mathit{C}\ast \mathit{C}+6,545\mathit{P}\ast \mathit{P}\\ & -7,080\mathit{V}\ast \mathit{V}-\mathit{\hspace{0.17em}}6,050\mathit{C}\ast \mathit{P}-1,475\mathit{C}\ast \mathit{S}-2,168\mathit{P}\ast \mathit{S}\end{array}$(1) Equation 5.(5) $\begin{array}{rl}& \mathit{Air}\mathit{\hspace{0.17em}}\mathit{permeability}\mathit{\hspace{0.17em}}\left(\mathit{A}\right)\mathit{\hspace{0.17em}}=\mathit{\hspace{0.17em}}2614,9167+190,4000\mathit{C}-5133,8333\mathit{P}-8,3917\mathit{S}-2,3167\mathit{\hspace{0.17em}}\mathit{C}\ast \mathit{C}\mathit{\hspace{0.17em}}+\text{break}\mathit{\hspace{0.17em}}2545\mathit{\hspace{0.17em}}\mathit{P}\ast \mathit{P}+7,6025\mathit{V}\ast \mathit{V}\\ & -14,1333\mathit{C}\ast \mathit{P}-0,6667\mathit{C}\ast \mathit{S}-142,5000\mathit{P}\ast \mathit{S}\end{array}$(5)

The developed model relates on regression coefficients.

(1) $\begin{array}{rl}& \mathit{Tensile}\mathit{\hspace{0.17em}}\mathit{strength}\mathit{\hspace{0.17em}}\left(\mathit{T}\right)\mathit{\hspace{0.17em}}=\mathit{\hspace{0.17em}}23,867+2,113\mathit{C}+2,459\mathit{P}+2,871\mathit{S}+6,738\mathit{C}\ast \mathit{C}+6,545\mathit{P}\ast \mathit{P}\\ & -7,080\mathit{V}\ast \mathit{V}-\mathit{\hspace{0.17em}}6,050\mathit{C}\ast \mathit{P}-1,475\mathit{C}\ast \mathit{S}-2,168\mathit{P}\ast \mathit{S}\end{array}$(1)

(2) $\begin{array}{rl}& \mathit{Tear}\mathit{\hspace{0.17em}}\mathit{strength}\mathit{\hspace{0.17em}}\left(\mathit{TS}\right)\mathit{\hspace{0.17em}}=\mathit{\hspace{0.17em}}20,4099-0,2968\mathit{C}-34,4698\mathit{P}+1,8550\mathit{S}+0,0093\mathit{C}\ast \mathit{C}+13,6063\mathit{P}\ast \text{breakP}-\mathit{\hspace{0.17em}}0,0488\mathit{V}\ast \mathit{V}\\ & -0,2793\mathit{C}\ast \mathit{P}+0,0018\mathit{C}\ast \mathit{S}+0,1037\mathit{P}\ast \mathit{S}\end{array}$(2)

(3) $\begin{array}{rl}& \mathit{Weight}\mathit{\hspace{0.17em}}\left(\mathit{W}\right)\mathit{\hspace{0.17em}}=\mathit{\hspace{0.17em}}10,1745+1,7566\mathit{C}-27,1333\mathit{P}+4,2588\mathit{S}-0,0078\mathit{C}\ast \mathit{C}+20,0833\mathit{P}\ast \text{breakP}-0,0540\mathit{V}\ast \mathit{V}-\mathit{\hspace{0.17em}}0,6533\mathit{C}\ast \\ & \mathit{P}-0,0232\mathit{C}\ast \mathit{S}-0,8500\mathit{P}\ast \mathit{S}\end{array}$(3)

(4) $\begin{array}{rl}& \mathit{Resistance}\mathit{\hspace{0.17em}}\mathit{to}\mathit{\hspace{0.17em}}\mathit{water}\mathit{\hspace{0.17em}}\mathit{penetration}\left(\mathit{R}\right)\mathit{\hspace{0.17em}}=\mathit{\hspace{0.17em}}36,9240-3,8860\mathit{C}-1,9633\mathit{P}\mathit{\hspace{0.17em}}+\mathit{\hspace{0.17em}}8,8704\mathit{\hspace{0.17em}}\mathit{S}\mathit{\hspace{0.17em}}+\mathit{\hspace{0.17em}}0,0629\mathit{\hspace{0.17em}}\mathit{C}\ast \mathit{C}+\\ & 22,3150\mathit{P}\ast \mathit{P}-0,1776\mathit{V}\ast \mathit{V}-0,8853\mathit{C}\ast \mathit{P}+0,0349\mathit{C}\ast \mathit{S}-1,3830\mathit{P}\ast \mathit{S}\end{array}$(4)

(5) $\begin{array}{rl}& \mathit{Air}\mathit{\hspace{0.17em}}\mathit{permeability}\mathit{\hspace{0.17em}}\left(\mathit{A}\right)\mathit{\hspace{0.17em}}=\mathit{\hspace{0.17em}}2614,9167+190,4000\mathit{C}-5133,8333\mathit{P}-8,3917\mathit{S}-2,3167\mathit{\hspace{0.17em}}\mathit{C}\ast \mathit{C}\mathit{\hspace{0.17em}}+\text{break}\mathit{\hspace{0.17em}}2545\mathit{\hspace{0.17em}}\mathit{P}\ast \mathit{P}+7,6025\mathit{V}\ast \mathit{V}\\ & -14,1333\mathit{C}\ast \mathit{P}-0,6667\mathit{C}\ast \mathit{S}-142,5000\mathit{P}\ast \mathit{S}\end{array}$(5)

The ANOVA results for all responses (T, TS, W, R, A) are shown in Table 1 and Table 2. It consists of model with correlation with various parameters.

Table 1. The experimental DOE factorial design.

Std. order	Run order	C	P	S	T (N)	TS (N)	W (g/m2)	A (mm/s)	R (mmH2O)
10	1	35	2	10	20,8	5,356	59	2010	66,3
5	2	20	1,5	10	24,1	2,41	43	2070	63,24
13	3	35	1,5	20	25,5	2,24	44	1910	53,04
2	4	50	1	20	45,6	15,65	62,8	2024	100
6	5	50	1,5	10	27,2	6,18	43	2110	41,82
15	6	35	1,5	20	21,9	17,18	60,2	1998	87,72
4	7	50	2	20	37	8,36	45	1700	79,56
8	8	50	1,5	30	20	9,32	44	1920	75,99
3	9	20	2	20	40,8	16,3	61,8	2130	88,74
7	10	20	1,5	30	22,8	4,47	57,9	2280	76,5
14	11	35	1,5	20	24,2	5,37	58,2	1660	63,24
9	12	35	1	10	10,13	6	56	2160	32,64
1	13	20	1	20	25,2	15,21	60	2030	82,62
11	14	35	1	30	30,2	7,17	57	5920	59,16
12	15	35	2	30	32,2	8,6	43	2920	65,16

Table 2. The ANOVA results for all responses of nonwoven cover bunch.

Response	Analysis of Variance
T(N)	Source DF Seq SS Adj SS Adj MS F P Regression 9 867,11 867,109 96,345 1,98 0,234 Linear 3 150,02 150,017 50,006 1,03 0,455 Square 3 543,19 543,187 181,062 3,72 0,096 Interaction 3 173,90 173,905 57,968 1,19 0,402 Residual Error 5 243,58 243,577 48,715 Lack-of-Fit 3 236,93 236,930 78,977 23,76 0,041 Pure Error 2 6,65 6,647 3,323
TS(N)	Source DF Seq SS Adj SS Adj MS F P Regression 9 195,45 195,45 21,717 0,67 0,719 Linear 3 15,37 15,37 5,125 0,16 0,921 Square 3 161,16 161,16 53,719 1,65 0,291 Interaction 3 18,92 18,92 6,308 0,19 0,896 Residual Error 5 162,82 162,82 32,564 Lack-of-Fit 3 38,66 38,66 12,887 0,21 0,884 Pure Error 2 124,16 124,16 62,079
W(g/m2)	Source DF Seq SS Adj SS Adj MS F P Regression 9 633,9 633,9 70,43 1,26 0,418 Linear 3 188,5 188,5 62,84 1,13 0,421 Square 3 228,7 228,7 76,25 1,37 0,353 Interaction 3 216,6 216,6 72,20 1,30 0,372 Residual Error 5 278,5 278,5 55,70 Lack-of-Fit 3 122,5 122,5 40,83 0,52 0,708 Pure Error 2 156,0 156,0 78,01
A(mm/s)	Source DF Seq SS Adj SS Adj MS F P Regression 9 11,173,207 11,173,207 1,241,467 1,62 0,310 Linear 3 4,243,939 4,243,939 1,414,646 1,84 0,256 Square 3 4,813,699 4,813,699 1,604,566 2,09 0,220 Interaction 3 2,115,569 2,115,569 705,190 0,92 0,495 Residual Error 5 3,836,175 3,836,175 767,235 Lack-of-Fit 3 3,774,679 3,774,679 1,258,226 40,92 0,024 Pure Error 2 61,496 61,496 30,748
R(mmH2O)	Source DF Seq SS Adj SS Adj MS F P Regression 9 3430,1 3430,1 381,1 1,62 0,308 Linear 3 766,5 766,5 255,5 1,09 0,434 Square 3 2186,7 2186,7 728,9 3,10 0,127 Interaction 3 476,9 476,9 159,0 0,68 0,602 Residual Error 5 1173,8 1173,8 234,8 Lack-of-Fit 3 538,5 538,5 179,5 0,57 0,689 Pure Error 2 635,3 635,3 317,7

Nonwoven bagging samples expressed with six input parameters have been evaluated and predicted by RSM on this experimental field of interest. Response surface methodology is a sequential approach which allows us to depict the response efficiently and approach the optimal region. It is a useful method analyzing of properties in which several responses require to be optimized simultaneously. We have demonstrated that the error values between experimental and analytical properties of date bunch cover show the effectiveness of the theoretical model in prediction. The model is obtained using the Design Expert Software. The analyses obtained when applying the ANOVA method with 95% of confidence level. Values of significance p value in all the models have to be less than 0.05. This value indicates that models are adequate and the terms have a significant effect on responses for the nonwoven products.

Contour plot analysis

Using surface response plots facilitate the comprehension of the relationships between the effective parameters. The simultaneous optimization of the parameters analyzed of nonwoven cover was made to have bags treated with the Tubicoat hp27 with optimum conditions. The Figures 1–5 show the plots of all responses of tensile strength, tear strength, weight, air permeability and resistance to water penetration. After hydrophobic treatment, the nonwoven cover remains breathable keeping the main hygienic properties. The cover bunch acquires the waterproof property, but the inter-fiber space remains untie, which allows the circulation of the air of the textile fabric to be maintained. The main analysis is employed to investigate the contribution and effects of hydrophobic parameters on the responses as shown in Figure 1–5. In the plots, the x/y axis demonstrates the values of each parameter at three levels and z-axis the value of each response. The 3D responses surfaces are the graphical representations of the regression equation. The main goal of response surface is hunt efficiently for the optimum values of the variables such that the response is maximized. Each contour curve indicates an infinitive number of combinations of two test parameters with the other maintained at its respective zero level. The highest predicted value is mentioned by the surface hold in the smallest ellipse in the contour diagram. Elliptical contours are obtained when there is a good interaction between the independent variables. The green lines indicate the values of the variables as well as the areas in red present the variation of each response, i.e. the properties of the cover bunch.

Figure 1. Nonwoven bagging product.

Figure 2. Surface plots of tear strength TS.

Figure 3. Surface plots of tensile strength T.

Figure 4. Surface plots of weight W.

Figure 5. Surface plots of air permeability A.

Figure 2 shows the effects of hydrophobic process (pressure, T concentration, cylinder speed) on Tear strength. It is confirmed that pressure and Tubicoat concentration have least effect on Tensile strength. It does not vary much with change in both parameters.

From Figure 3, it was observed that Tear strength increases linearly with increase in the cylinder speed and Tubicoat hp concentration. It is also evident that Tear strength has a maximum when cylinder pressure value is low. The cylinder speed has a major effect on the Tear strength.

Minimum weight was induced when the cylinder speed and tubicoat concentration were high. Figure 4 shows the interaction of Tubicoat concentration, speed and cylinder pressure on the weight of the nonwoven cover. It is observed that the desirable value of weight was induced for the lowest tubicoat concentration values. Figure 5 shows the interaction of cylinder speed; tubicoat concentration and cylinder pressure on induced the air permeability of nonwoven cover. It is observed that, for the increase in cylinder velocity, the increase in air permeability is observed. It is also evident that the cylinder pressure has no effect on the air permeability of the cover. The similar result was obtained for the Tubicoat concentration.

Figure 6 shows that the increase in cylinder pressure of the hydrophobic procedure increases the resistance to water penetration and also increase in cylinder velocity increases the resistance to water penetration. The small amount of Tubicoat concentration required to water resistance the parts.

Figure 6. Surface plots of resistance to water penetration R.

Conclusion

The below conclusions were drawn from the experimental investigation:

• Nonwoven cover bunch behavior was analyzed successfully through tensile strength, tear strength, weight, resistance to water penetration and air permeability by using response surface methodology.

• An empirical model was developed to predict cover bunch quality.

• A maximum tear strength of 23,86 N obtained under at Pressure is 1 bar, cylinder speed is 20 trs/min and Tubicoat hp concentration is 35 g/l.

• A maximum Tensile strength of 345 N obtained under at Pressure is 2 bar, cylinder speed is 030 trs/min and Tubicoat hp concentration is 50 g/l.

• A maximum Air permeability of 1856 mm/s obtained under at Pressure is 1 bar, cylinder speed is 30 trs/min and Tubicoat hp concentration is 25 g/l.

• A target Weight of 54,13 g/m2 obtained under at Pressure is 1,5 bar, cylinder speed is 10trs/min and Tubicoat hp concentration is 50 g/l.

• A maximum resistance to water penetration of 68 mm H2O obtained under at Pressure is 2 mm, cylinder speed is 30 trs//min and tubicoat hp concentration is 20 g/l.

• The error values between analytical and experimental properties of nonwoven cover < 0.05 show the effectiveness of the theoretical model in prediction.

Highlights

• The cover quality is designed based on a mathematical model and further optimized using MINITAB.

• The error values between experimental and analytical properties of date bunch cover show the effectiveness of the theoretical model in prediction.

• The analyses obtained when applying the ANOVA method with 95% of confidence level.

• Values of significance p value in all the models have to be less than 0.05, indicates that models are adequate and the terms have a significant effect on responses for the nonwoven products.

• The cover date bunch remains influenced after harvesting period by the resistance to water penetration properties and tear strength as well.

Acknowledgments

The author thankfully acknowledges the assistance of Dr Samah Maatoug from the University of Tabuk, who reviewed and provided comments that greatly improved the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.