INSTANT DETERMINATION OF MOISTURE AND BULK DENSITY OF WOODCHIPS IN CELLULOSE INDUSTRY

Purpose: This research aims to achieve the instantaneous determination of moisture content and bulk density of wood chips for energy generation and pulp production in the pulp and paper industry. Theoretical framework: The moisture content and bulk density of woodchips influence the quality of biomass for energy generation and cellulose production in various ways. These variables affect the calorific value, mass and volume of logistical transport, the quantity of reagents in the reactor for wood delignification, among other factors. For this reason, a quick and accurate evaluation of these factors in the field is of fundamental importance to improve industrial quality and productivity. Method: To achieve this goal, a pilot test was conducted using a capacitive sensor and a portable scale to determine the moisture and bulk density of wood chips. Calibration equations were developed in the laboratory for the sensor, correlating the readings with the actual moisture content of seven types of wood chips, including Eucalyptus spp. and Pinus spp. The results of field measurements using these equations were compared with the gravimetric method, as well as bulk density measurements. Results and conclusion: The moisture calibration equations showed a high correlation between the capacitive sensor and the gravimetric method. The field density results were similar to those obtained in the laboratory. It was concluded that the investigated methods show promise for use in the pulp and paper industry. Research implications: The validation and accuracy of portable instruments for the instantaneous determination of moisture content and bulk density of wood chips are of utmost importance for manufacturers of portable meters and for industries that use this raw material. Originality/value: The instantaneous determination of moisture content and bulk density can contribute to improving the energy and production efficiency of boilers that generate heat for pulp digesters. Moreover, it optimizes logistical transportation and quality control against the attack of xylophagous organisms, preventing waste of raw materials, which is essential for sustainable development.


INTRODUCTION
The bulk density is a wood variable that presents one of the largest correlations with the performance in the industrial process of pulp production (Souza et al., 2002) and the woodchips' moisture influences several steps of cellulose industry (Klock & Andrade, 2013).The first of them would be about the calorific power of the logs barks that are used as byproduct for energy generation (Lopes, 2012).The second one, according to the authors, is the chipping, that requires homogeneous moisture.And finally, in the pulping stage, as higher the woodchips.moisture, better.Therefore, it is desirable that obtaining this data happens in a functional and quick way.
Among the density informationbasic density and bulk density -, the bulk density is the information that is most easily determined and can be obtained instantaneously and from large samples, since it is the quotient between the total mass of a given material stored in a certain volume by its volume (Klock & Andrade, 2013).
For the moisture measurements, the electric meters provide an immediate response and use a non-destructive and inexpensive method, which facilitates data collection, especially in industry (Calonego et al., 2006).They can be classified into two groups: resistive (conductance) and capacitive (dielectric) (Simpson & Service, 1991).
The dielectric method, which uses the capacitive principle, is based on changes in an electric field.The changes happen due to the dielectric properties of the material matrix under alternating current radiation (Jensen et al., 2006).However, the resultant of the dielectric properties of wet wood is not the simple combination of the dielectric properties of dry wood and water (Paz, 2010).Water has different thermodynamic properties in solid materials.For this reason, many authors claim that moisture meters that use the dielectric method provide an estimate moisture information just below 30% (wet basis), that is approximately the fiber saturation point (James, 1988).However, other researches with capacitive principle moisture meters of different brands, obtaining determination coefficients between instrument results and gravimetric method close to 90% for moisture ranges that reached up to 50% (wet basis) (Jensen et al., 2006).
The present work is part of a project whose objective is to propose and validate a method for the instantaneous and online measurement of woodchips moisture and density, that is applicable in the digestor and power boiler feed lines.The project aims at early responses to pulp mills controls through operational stability and higher quality in its products.
For the moisture determination, the dielectric method, capacitive principle was chosen.Here is reported the pilot phase of the project, which used handle equipment that did not require mechanical and electrical installations at plants in the industries where the project began.Once satisfactory results were obtained at the end of the pilot test, the second phase is the identification of opportunities for equipment installation in Brazilian pulp industries and for online and continuous measurement and to obtain instantaneous data for the process.

MATERIALS AND METHODS
In the pilot phase the data collection was made with the material listet in Table 1, identified from A to G.For the moisture determination, the application of the capacitive principle was investigated and aiming to get the information about bulk density, results from the simple quotient between a sample mass, measured in the field and its known volume was analyzed.In each industry where the pilot project was developed, the moisture and the bulk density were determined using the equipment MUG M75-D (Marrari Automação Industrial LTDA, Brazil).
First, to measure the moisture, a calibration equation was programmed in the equipment for all woodchips types.The simple regression was used, where the independent variable bits is the capacitive sensor reading and the dependent variable is the moisture.The primary method that was considered the reference for the elaboration of the calibration equation was the gravimetric method, described by NBR 14929 (ABNT, 2003).After that, moisture readings were made using the created regression model and were compared to the gravimetric method results through analysis of variance with one way.
Simultaneously to the comparison between the moisture results, in the field, the portable weighing machine that is coupled with the equipment was used to determine the bulk density of the chips A, B, C, and E, since the volume of the container where the chips were deposited for the exposition to the capacitive sensor was known.This result was graphically compared to the results of the application of the Brazilian standard for woodchips bulk density determination NBR 14984 (ABNT, 2003), using the equipment Mecatécnica (Mecatécnica, Brazil).
All the density results in this work related are presented in wet basis.For the conversion of bulk density in wet basis obtained in the laboratory to dry basis, the moisture result from the gravimetric method was used.For the conversion of the density information in wet basis obtained in the field to dry basis, the moisture result from the capacitive sensor was used.
For the creation of the regression model, the comparison between the results collected in the field and the results of the application of the Brazilian standards in the laboratory 15 samples were used.Figure 1 shows the sequence of the data collection and the quantity of material used to obtain each information.The regression analysis for the calibration curves and the variance analysis were made using the 17th Version of the software Minitab®.For the generation of the comparison graphics for density, the 10th Version of the software Excel® was used.

RESULTS AND DISCUSSIONS
Figures 2, 3 and 4 present the calibration equations and the correlation between the sensor Reading, in bits, and the moisture results for each analyzed material, in this order: Eucalyptus spp.woodchips, Pinus spp.woodchips, both destinated for the cellulosis production process and mix of biomass, destinated for energy generation in the power boiler.Despite the statement of James (1998) on the impossibility of measuring wood with moisture values above 30%; the results found here show that exept chips C, the other materials presented determination coefficient R² greater than 90% for a moisture range covering chips with moisture up to 55%.That is, the reading of the capacitive sensor explains more than 90% of the moisture information of the chips analysed.For both conifers and hardwoods the R² values found by Jensen (2006) were also greater than 90%: 96% using the Pandis meter (Exotek, Germany) and 92% using the Schaller meter (Schaller, Austria).However, in a norrower measurement range than in the present research, of up to 50% moisture on wet basis.
The lowest R² value found for chips C and for chips E can probably be explained by the small moisture range in which the materials were analyzed, which hindered the creation of a regression model.While for the other materials the moisture values were between 10 and 55%; for chips collected in the line that feeds the sieve, the moisture was between 40 and 55%, precluding the creation and the trend analysis of a broader model.
The figures 5, 6, 7, 8, and 9 demonstrate the graphs of the moisture measurement range obtained by the one-way analysis of variance.The first bar displays the range of the results of the first repetition obtained in the laboratory (lab1).The second one shows the results of the second repetition (lab2) and the third one shows the range of the results obtained with the capacitive sensor (cap).Figures 5, 6, 7, 8, and 9 demonstrate that no statistical difference was observed, with 95% significance, between laboratories 1 and 2, as well as in the capacitive method, validating the use of this tool.
The comparative analyses for density measurements were made graphically for the the 15 samples analyzed.Figures 10, 11, 12 and 13 present the laboratory (lab) results and the equipment (MUG M75) results for bulk density (DA) in dry basis in ascending order for chips A, B, C and E.  It is possible to identify that the differences found for chips A, B, and C, of Eucalyptus spp.are similar to the differences found for chips E, of Pinus spp.This fact draws attention because many authors highlight the high variability between trees and within trees of the various Eucalyptus species (Ferreira et al., 2006;Souza et al., 2002).As the chips analyzed are composed of different species harvested from different fields, it was expected more difficulty to determine the bulk density of Eucalyptus spp., compared to the determination of Pinus spp.' chips density.This situation may have resulted from efficient chip homogenization that occurs during their transport on the conveyor belts of the industrial plant where they were collected.
Moreover, when comparing chips A and B, are collected in the Province of Espírito Santo and Mato Grosso do Sul (Center of Brazil) and chips C and D, are collected in a factory in the Province of Paraná (South of Brazil); bigger errors between measurements are found for chips C and D. It happened due to the differences in the industrial processes that use this material.While in the mills that produce chips A and B there is a cellulose digester that is fed with more homogeneous woodchips, woodchips C and D were collected before the conveyor that feeds the woodchips in the factory's patio, so, before crossing the sieve and before the homogenization of the granulometry.The heterogeneity of chips C and D is further enhanced, because of the high storage time after chipping causes greater variability in moisture and consequent greater diversity of material analyzed within the same sample.

CONCLUSIONS
The creation of the calibration curves allows to conclude that the correlation between the woodchips moisture and the capacitive sensor reading is applicable, even for chips with moisture above the fiber point saturation and that, therefore, the moisture determination in a instantaneous way and in the field is possible.
The analysis of variance for moisture alows to conclude that the results from the moisture readings made in the field with the capacitive sensor and the results from the application of the gravimetric method are similar.The comparative plots for bulk density show similar results for field determinations and determinations following the Brazilian standard with Mecatécnica equipment.However, a deeper analysis is required for these results, as a variance analysis, which requires more than one repetition per sample in future data collections.
Moisture measurement can be applied with the dielectric method and capacitive principle and the bulk density can be estimated by the simply quotient between mass measured in the filed and an known volume of the sample.Future works are recommended applying these methods in industrial lines: installing equipments in the conveyor before the digestor feeding in cellulosis industries and in the conveyor before the power boiler feeding.

Figure 1 :
Figure 1: A Brief Scheme from the Data Collection.Source: The authors (2023)

Figure 2 :Figure 3 :Figure 4 :
Figure 2: Correlation between the sensor Reading and the moisture for Eucalyptus spp.woodchips.Source: The authors (2023)

Figure 5 :
Figure 5: Comparison between moisture results for chips A. Source: The authors (2023)

Figure 6 :
Figure 6: Comparison between moisture results for chips B. Source: The authors (2023)

Figure 7 :
Figure 7: Comparison between moisture results for chips C. Source: The authors (2023)

Figure 8 :
Figure 8: Comparison between moisture results for chips D. Source: The authors (2023)

Figure 10 :
Figure 10: Comparison between results from laboratory (lab) and equipment (MUG M75) of bulk density (DA) for chips A. Source: The authors (2023)

Table 1 .
Materials Used for Data Collection.The material "mix of biomass" is an heterogeneous composition of forest residue, woodchips provided by sawmills, woodchips collected in the pile and losses from the cellulosis production.