Quality Control in Plywood Manufacturing: Physical Properties of Commercial Plywood of Pinus sp

Wood of genus Pinus have been widely used by plywood Brazilian industry, motivated by their good physical and mechanical properties as well as reforestation policies. For p lywood companies, quality control for its commercialization presents itself as essential. In th is sense, this paper aims to evaluate, by means of physical properties, homogeneity of plywood manufacturing process using Pinus sp. So, we selected four panels, and extracted samples to obtain their physical properties. Analysis of variance (ANOVA) was used to investigate manufacture homogeneity of the panels. ANOVA results, for all physical properties investigated, had presented statistical equivalence, showing consistency of product manufacture, independent of the supplier company.


Introduction
In Brazil, a g reater investment in wood-based industry is enabling the use of wood with more rationality, given the circu mstances that involve wood applications. In addition, knowledge of wood physical and mechanical properties is essential for its compatible utilization [1][2].
Among the techniques currently employed, using wood fro m planted forests is highlighted in the country. Reforestation of Pinus, Eucalyptus and Corymbia genus favor material for emp loying in construction, furniture industry and reconstituted wood panels [3].
Plywood is still an important wood-based product, made with veneers, ju xtaposed with change between fiber direction, bonded with proper adhesive, under certain conditions of temperature and pressure [4]. Ply wood can be used in various applications, such as in furniture industry and building construction [5][6][7].
Quality control in plywood manufacture is comp lex, once there are many elements that affect panel performance, including: options for hardwood or softwood [8][9][10]; pressing and temperature cycles applied [11]; and the kind of adhesive used [12][13][14].
Use of Pinus lu mber in plywood manufacture is lin ked to refo res t at io n o f th is s p ecies . So ut hern o f Brazil is responsible for 83% of Pinus reforestation in the country, followed by the Southeast, with 14.3% [15], regions where a expressive number of co mpanies are located.
Relying on the importance of quality control of panels produced industrially, this study aimed, with the aid of physical properties, evaluated the homogeneity of the industrial manufacturing process of plywood made with Pinus sp.

Material and Methods
For this research, four plywood panels obtained from different work shifts of a line production, in a co mpany localized in São Carlos (SP-Brazil), were considered. Panels were manufactured with Pinus sp veneers, dimensions of 1.25 m × 2.50 m × 18 mm, and using a phenolic resin to gluing.
Fro m the panels, specimens were prepared to perform determination of the physical properties of the commercial plywood, according to Brazilian normat ive documents.
Tests were performed in water absorption (24 hours) recovery in thickness; swelling more recovery thickness swelling; and swelling.

Water Absorption after 24 Hours
Samples were weighed and then immersed in water fo r 24 hours and, subsequently, removed and weighed again. Absorption (A) in water after 24 hours (Equation 1) was Properties of Commercial Plywood of Pinus sp obtained following assumptions and calculation methods of ABNT NBR 9486 [19], using 6 specimens per panel.
In Equation 1, A denotes the amount of water absorbed (%), m f is the specimen final mass (g) and m i is the specimen initial mass (g).

Recovey in Thickness; Swelling more Recovery Thickness Swelling; and S welling
To determine recovery in thickness; swelling more recovery thickness swelling; and swelling in ply wood, after 24 hours of water absorption ABNT NRB 9535 recommendations were used [20].
Fro m each panel, six specimens were prepared, dimensions 75 × 25 mm, packed in a climat ic chamber (25 ± 2)°C. These samples were separated in two sets (one of them considered as a control).
After first thickness measurement, made in the center of each specimen, those used as control were dried at (103 ± 2)° C for 24 hours, taken to the desiccator and measured again.
The other specimens set was immersed in d istilled water at (20 ± 2)°C during 24 hours. After this period, specimens were measured, dried at (103 ± 2)°C for 24 hours, taken to the desiccator and measured again. Thickness recovery; swelling more thickness; swelling in water after 24 hours were calculated using Equations 2, 3 and 4 respectively.
In Equations 2, 3 and 4, e 1 denotes the sum of specimens thickness (mm) o f control set (air-conditioned); e 2 is the sum of the thicknesses of the control set specimens (oven dried); e 3 is the sum of specimens thicknesses before immersion in water; e 4 is the sum of the specimens thicknesses after immersion in water; and e 5 is the sum of the specimens thicknesses after immersion in water (oven dried).

Statistical Analysis
To evaluate possible differences between physical properties of plywood panels manufactured with Pinus sp., analysis of variance (ANOVA) was used, in which the panel fro m where samples were extracted is the factor investigated, associated to four levels (number of selected panels). ANOVA was considered at level of significance (α) 5%; equivalence between mean values for physical properties as null hypothesis (H 0 ); and non-equivalence as alternative hypothesis (H 1 ). P-value less than significance level imp lies rejecting H 0 , accepting it otherwise.
For ANOVA validation, were evaluated: normality (Anderson-Darling), in distribution of responses, and homogeneity between variances by treatment (Bart lett and Levene). Both tests were considered at the 5% significance level. Anderson-Darling test considered normal distributions to null hypothesis, and non-normality as alternative hypothesis. P-value greater than significance level involves accepting H 0 , rejecting it otherwise. Bartlett and Levene tests have equivalence of variances among treatments as null hypothesis and alternative hypothesis as non-equivalence. P-value g reater than significance level involves accepting H0, rejecting it otherwise.    Figures 1 and 2 show results of normality test and variance homogeneity evaluation test, respectively. As P-values are both higher than significance level for each physical properties investigated, distributions can be considered as normal and variances between treatments equivalent, validating ANOVA model.     Tables 5-8 show ANOVA results (factor: panel) for each response variable of interest. As P-values found are h igher than level of significance level (5%), H 0 is accepted, imply ing that properties of all selected panels are equivalent, confirming homogeneity of manufacturing process of the company.

Conclusions
ANOVA results on physical properties revealed statistical equivalence among plywood panels, showing the good standard and consistency of manufacturing panels of the considered company, whose production is geared both for import and for export.