Method for Simultaneously Determining Parameters of at Least One Resin Layer Applied to at Least One Carrier Material
20230145539 · 2023-05-11
Inventors
Cpc classification
International classification
Abstract
Provided is a method for the simultaneous determination of parameters, in particular of at least two, three or four parameters, of at least one resin layer applied to at least one carrier material by recording and evaluating at least one NIR spectrum in a wavelength range between 500 nm and 2500 nm, preferably between 700 nm and 2000 nm, more preferably between 900 nm and 1700 nm, and particularly advantageously between 1450 nm and 1550 nm, using at least one NIR measuring head, in particular at least one NIR multimeter head.
Claims
1. A method for the simultaneous determination of parameters, in particular of at least two, three or four parameters, of at least one resin layer applied to at least one carrier material by recording and evaluating at least one NIR spectrum in a wavelength range between 500 nm and 2500 nm, preferably between 700 nm and 2000 nm, in particular preferably between 900 nm and 1700 nm, and particularly advantageously between 1450 nm and 1550 nm, using at least one NIR measuring head, in particular at least one NIR multimeter head.
2. The method according to claim 1, wherein the parameters are selected from a group comprising the amount of applied resin layer, the degree of curing and degree of cross-linking of the applied resin layer, the moisture content of the applied resin layer, the abrasion resistance and, if applicable, the amount of abrasion-resistant particles or further solids scattered onto the resin layer.
3. The method according to claim 1, further comprising the following steps: recording of at least one NIR spectrum of several reference samples, each with different values of the desired parameters, using at least one NIR measuring head, in particular NIR multi-measuring head in a wavelength range between 500 nm and 2500 nm, preferably between 700 nm and 2000 nm, in particular preferably between 900 nm and 1700 nm, and particularly advantageously between 1450 nm and 1550 nm; determination of the desired parameters of the said reference samples by non-spectroscopic methods; assignment of the determined parameters to the recorded NIR spectra of the said reference samples; creation of a calibration model for the correlation between the spectral data of the NIR spectra and the associated parameter values using multivariate data analysis; applying at least one resin layer to at least one side of the carrier material; recording of at least one NIR spectrum of the resin layer applied to the carrier material using the at least one NIR measuring head, in particular NIR multimeter head, in a wavelength range between 500 nm and 2500 nm, preferably between 700 nm and 2000 nm, more preferably between 900 nm and 1700 nm and particularly advantageously between 1450 nm and 1550 nm; determine the desired parameters of the resin layer applied to the carrier material by comparing the NIR spectrum recorded for the resin layer with the calibration model created.
4. The method according to claim 1, wherein spectral data from the entire recorded spectral range are used to create the calibration model.
5. The method according to claim 1, wherein spectral data from the NIR spectral range between 1450 nm and 1550 nm are used for the creation of the calibration model, which are pre-treated by means of suitable mathematical methods and are subsequently fed to the multivariate data analysis.
6. The method according to claim 1, wherein the resin layer consists of a formaldehyde-containing resin, preferably a melamine-formaldehyde resin, a urea-formaldehyde resin or mixtures of both.
7. The method according to claim 1, wherein the solids content of the resin layer is between 30 and 80 wt. %, preferably between 50 and 65 wt. %.
8. The method according to claim 1, wherein the liquid resin layer comprises abrasion-resistant particles, natural and/or synthetic fibres and further additives.
9. The method according to claim 1, wherein the at least one carrier material is a paper layer, preferably a decorative paper layer or an overlay paper layer, or a wood-based board, preferably a medium-density fibre (MDF), high-density fibre (HDF) or rough particle board (OSB), a plywood board or a wood-plastic composite board (WPC) or a stone-plastic composite board (SPC).
10. The method according to claim 1, wherein the determination of the parameters of the resin layer is carried out continuously and online.
11. A production line for carrying out a method according to claim 1 comprising at least one NIR multimeter head and at least one control system for controlling the production line, wherein the control system of the production line comprises at least one computer-aided evaluation unit and a database.
12. The production line according to claim 11, wherein the comparison of the NIR spectrum measured for the product in the form of the carrier material with applied resin layer with the calibration models created for the respective individual parameters takes place in the evaluation unit and the parameter data determined in this way are stored in the database.
13. The production line according to claim 10, wherein the production line is networked with at least one further production line.
14. The method according to claim 1, wherein the parameters are used to control at least one production line.
15. The method according to claim 14, wherein the parameters are fed to a “self-learning” AI-based evaluation system (machine learning) for further optimisation of the control of the at least one production line and/or for prediction of the function of the at least one production line and/or for optimisation of the start-up process of the at least one production line.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0127] The proposed solution is explained in more detail below with reference to the figures in the drawing using an example of an embodiment.
[0128]
[0129]
DESCRIPTION OF THE INVENTION
[0130] The production line shown schematically in
[0131] Downstream of the first applicator roller 1, a first scattering device 10 is also provided for uniformly scattering the abrasion-resistant material, such as corundum, onto the first resin layer on the upper surface of the HDF board. Drying of the first resin layer is then performed in the first convection dryer 1a. After the first convection dryer 1a, a first NIR multimeter head is provided.
[0132] This is followed by a second double applicator 2 for applying a second resin layer and a second convection dryer 2a for drying the second resin layer.
[0133] Downstream of the third double applicator 3 for applying the third resin layer, there may be another scattering device 20 for applying glass beads to the third resin layer followed by a third convection dryer 3a for drying the third resin layer. The scattering device 20 for the glass beads is optional. The glass beads may also be applied together with the third resin layer. After the third convection dryer 3a, a second NIR multimeter head is provided.
[0134] After application of the fourth resin layer, which in the case of the fourth resin layer on the top side may contain e.g. cellulose fibres, in a fourth double applicator 4 and drying in a fourth convection dryer 4a, the layered structure is pressed in a short-cycle press 5. The pressed boards are cooled and stored.
[0135] It is generally possible that further NIR measuring heads are provided in the production line. For example, after application and drying of the last resin layer before pressing, another NIR measuring head can be provided in the production line (not shown). It is also possible that one or more NIR measuring heads are additionally provided in the production line below the material board being passed by, so that the underside of the material boards is illuminated (not shown).
[0136] The two NIR multimeter heads are connected to a control system with an evaluation unit and database for processing and storing the determined NIR data. In the event of deviations from the target values, the production conditions are automatically adjusted by the system control or regulation.
[0137]
[0138] Here, too, the NIR multimeter head is connected to a control system with an evaluation unit and database for processing and storing the determined NIR data. In the event of deviations from the target values, the production conditions are automatically adjusted by the system control or regulation.
Example 1: Determining the Amount of Resin Layer Applied to a Wood-Based Board Provided with a Decor
[0139] Create a Reference Sample and Calibrate:
[0140] The calibration of the NIR measuring system for the on-line determination of the application quantity is carried out as follows.
[0141] Several 50 cm×50 cm samples are made from an HDF board. The back and edges of the samples are sealed with aluminium tape to prevent moisture loss during drying. A sample is coated with a resin solution from above using an applicator roller. The amount applied is determined by weighing out the samples before and after coating. The sample is then dried in a convection dryer at approx. 190° C. for approx. 15 seconds and then measured with the NIR spectrometer. The calibration samples are measured with the NIR measuring system DA 7400.
[0142] The measurement is started manually and takes approx. 10 s. During this process, the reference boards are moved by hand in a circle under the measuring head. By varying the application quantity, calibration is achieved over a wide range of application quantities from 30 g/m.sup.2 to 145 g/m.sup.2. In this way, several reference samples are measured.
[0143] A calibration model is created from the reference spectra, which is used to determine the application amount of an unknown sample. The calibration model is created using multivariate data analysis. This is done with suitable analysis software, e.g. with the previously mentioned analysis software The Unscrambler from the company CAMO.
[0144] To create the calibration model, spectral data from the NIR spectral range of the liquid resin between 1450 and 1550 nm are used, which are first pre-treated using suitable mathematical methods and then fed to the multivariate data analysis.
[0145] The software programme used enables special pre-treatment techniques of the spectral data to minimise various interfering factors on the measurement such as surface condition of the samples, infrared inactive fillers in the coating or different colours of the samples and others.
[0146] The colour influence in the case of determining the amount of liquid resin on decor layers on the NIR measurement can additionally be solved by forming decor groups that have a similar colour division. For this purpose, all decors used in the calibration are divided into three groups according to their colour shade. Group 1 contains light decors and group 2 medium decors. Calibration samples of dark decors are grouped in group 3. By dividing the entire model into three group models according to the colour of the decors produced, the systematic (decor-dependent) deviation is reduced to <5%, which leads to an increase in the accuracy of the online measurement. At the same time, not every decor produced on the line has to be recorded in the DCS model. For the NIR measurement of the application quantity of the resin on the HDF boards with different decors, it is sufficient to develop a PLS regression model from a few representatives of each decor group, which can be applied to the entire group.
[0147] Determining the Application Amount of Liquid Resin on a Decorative Layer:
[0148] Calibration is carried out by recording the NIR spectra of several reference samples (weighing boards 50 cm×50 cm) printed with different decors and with a known amount of resin. The samples are coated from above with a resin solution (application quantity between 10 g/m.sup.2 and 150 g/m.sup.2) using an applicator roller. From the calibration spectra obtained, a PLS regression model is created for each deco group, which is then used for the on-line measurement at the plant.
[0149] For the on-line measurement of the application amount in the process, models for each decor group are installed in the NIR measuring system.
Example 2: Determining the Moisture Content of the Resin Layer Applied to a Wood-Based Board
[0150] Melamine-formaldehyde resin layers are applied to wood-based boards. These serve as reference samples. NIR spectra are then recorded in a wavelength range between 900 and 1700 nm. For the recording of the NIR spectra, a NIR measuring device from Perten, for example, is used. The measuring head bears the designation DA 7400.
[0151] Suitable calibration models are created for further evaluation of the NIR spectra. Thus, a first calibration model is created for the NIR spectra of the reference samples (without the Darr sample), which was determined using Partial Least Squares (PLS) regression.
[0152] This model is used to determine the residual moisture in the Darr sample. Using the first calibration model, a moisture content for the Darr-dried sample is calculated, for example, by the previously mentioned analysis programme SIMCA-P from Umetrics AB. For this purpose, the Partial Least Square (PLS) regression is used to create a calibration function that describes a dependency between spectrum and moisture content. For the moisture content of the Darr-dried sample, a moisture content is calculated by the analysis programme using the calibration function created.
[0153] Subsequently, the amount of moisture content for the Darr-dried sample is added to all used moisture values of the calibration samples or reference samples and the moisture content of the Darr-dried sample is set equal to zero. From these new calibration values of the moisture and the measured spectra, a second calibration model is created with the help of Partial Least Squares (PLS) regression, which is now suitable for creating a relationship between the measured NIR spectra of a resin layer to be measured on a carrier board and the NIR spectra of reference samples with known moisture content.
[0154] Using the second PLS calibration model of linear regression, the NIR spectrum determined for a sample is then assigned to a specific moisture content.
Example 3: Determining the Degree of Curing of the Resin Layer Applied to a Wood-Based Board
[0155] Create a Reference Sample and Calibrate:
[0156] Calibration is performed by taking an NIR spectrum of a cured sample, which is then tested for curing using acid tests, and is carried out as follows.
[0157] HDF boards (207 cm×280 cm) printed with various decors are coated from above with liquid melamine-formaldehyde (MF) resin or with an overlay paper impregnated with MF resin on the coating line by means of an application roller and then pressed in a short-cycle press at 190-210° C. and approx. 40 bar for 8 to 36 seconds. During this process, the protective layer is cured. By varying the pressing time and pressing temperature, samples with differently cured protective layers are obtained.
[0158] When calibrating for an online measurement, the NIR spectra are recorded a few seconds after the pressing process directly on the production line. Subsequently, the board is tested for curing by means of an acid test at the points where the NIR spectra were recorded. An average value is calculated from the results of the acid test for one board and assigned to the spectrum of this board.
[0159] In this way, several reference spectra of differently cured boards with different colour decors (for calibration of the online measurement) are recorded.
[0160] The acid test is carried out as follows. Three drops of 4 M HCl are added to a board that has cooled down to room temperature. After 25 minutes of exposure time, the acid is rinsed off with water. Based on the visual and haptic assessment of the surface at the point of application, a statement is made about the quality of the curing.
[0161] For calibration, the result of the test is correlated with spectral data. The calibration model is created using multivariate data analysis. This is done with suitable analysis software, e.g. The Unscrambler from CAMO. This programme makes it possible, among other things, to minimise various interfering factors on the measurement, such as surface properties of the samples, infrared-inactive fillers in the coating or different colours of the samples, by means of special pre-treatment techniques of the spectral data. The colour influence on the NIR measurement can, as described before, be solved additionally by the formation of deco groups, which have a similar colour division.
[0162] A calibration model is created from the reference spectra, which can be used to determine (predict) the curing of an unknown sample.
[0163] Online NIR measurement then directly predicts the result of the acid test or quality of the cure. Online measurement of the degree of cure of a resin coating:
[0164] With the online determination of the curing, the NIR measurement or recording of the NIR spectra is carried out directly at the production plant, immediately after the curing of the resin coating.
[0165] After creating a calibration model, this is installed in the measuring device. As the samples pass under the measuring head, several NIR spectra are recorded from the coating. The calibration model is used to calculate an average cure (acid test) of the coating from the recorded spectra. In this way, each board is tested for the quality of the curing during production.
Example 4: Determining the Degree of Cross-Linking of Impregnates
[0166] Create a Reference Sample and Calibrate:
[0167] In the case of impregnated papers, impregnates in different drying states are first measured with the NIR measuring head for the calibration of the NIR measurement. For this purpose, impregnates are taken from an impregnation channel that are produced at different speeds. The speeds of the impregnation channel are varied upwards and downwards around the optimum known per se. Then the samples are measured by NIR measuring head and then the VC value is determined in the oven. After five minutes of drying, the cured final state (C-state) is reached. After drying in the oven, the samples are measured a second time using the NIR measuring head. This measurement provides the end point of the drying.
[0168] After correction of the baseline shift, the spectra show change in absorption intensity of the NIR bands at about 1450 nm (water) and at about 1490 nm (N—H groups).
[0169] One can then create a correlation between the VC values and the NIR spectra. The creation of the calibration model describing the correlation between the NIR spectra and the corresponding VC values is done with the multivariate regression methods (e.g. with MLR, PCR or PLS regression methods, etc.). The evaluation is carried out over the entire spectrum. The longer the sample is dried in the oven, the higher the degree of cross-linking. Samples can also be measured that have a higher VC value than usually aimed for. By testing these under-dried samples, improved control of the impregnation channel (dry channel) can also be achieved. The under-dried samples have a methylol group content that is above the optimum, while the over-dried samples have a methylol group content that is below the optimum. Thus, by recording the NIR spectra in combination with visualisation, it is possible to control the channel based on the methylol group content. With these data, after calibration for the different resin and paper systems, a basic prediction can be made about the state of the curing.
[0170] The tests on the curing state can be carried out directly at the outlet of an impregnation system. The NIR measuring head delivers several hundred measured values per minute, making continuous process monitoring possible. For quality tests in the warehouse, a mobile measuring device can be used.
[0171] The NIR measurement of the cross-linking can be carried out directly at the production line (online measurement) with a mobile device (e.g. with the previously mentioned DA 7400 device from Perten) or in the laboratory (off-line measurement) with a stationary device (e.g. with the DA 7250 device from Perten).
[0172] Measurement of the Degree of Cross-Linking:
[0173] In an impregnation channel, an overlay (paper weight: 25 g/m.sup.2) is impregnated with a mixture of melamine resin and corundum. The total application quantity is approx. 75 g/m.sup.2, of which approx. 55 g is melamine resin. The impregnation channel is run at different speeds. Samples are taken from the individual speed variants in such a test and measured with the help of the NIR measuring head. They were then examined for their VC value. VC values of 3.0 to 8% were found. All samples were measured a second time with the NIR measuring head after drying.
[0174] A correlation was then created from the spectra and the VC values, which allowed a prediction of the VC value on other overlay patterns.
[0175] This process can of course also be used for systems where the pre-cured resins are applied directly to printed or non-printed wood-based boards. The same problems arise with these boards as with impregnated papers. Here, too, it is important to know the curing state of the resin on the board. This applies both to a linked production line, in which the resin is applied and dried before further processing in a short-cycle (KT) or continuous press (Conti press), and to a non-linked production line, in which the resin is applied and dried before the intermediate product is processed further at a later time. In the case described, the determination of the cross-linking state is all the more important because, in contrast to impregnated papers, a quality determination of the parameters usually to be determined is otherwise hardly possible. By means of a measurement carried out directly on the production line, an adjustment of the production parameters can be made immediately in case of deviations from the target state. These can be, for example, the feed rate and the dryer setting (air temperature, air speed and air flow). Finally, it is possible to control the dryers via the curing state of the synthetic resin.
Example 5: Determining the Abrasion Resistance
[0176] Create a Reference Sample and Calibrate:
[0177] a) Calibration in the case of an already cured wear layer is carried out by recording an NIR spectrum of a carrier board provided with an already cured wear layer as a reference sample in analogy to the procedure described under b) directly below.
[0178] b) Calibration in the case of a not yet hardened wear layer is carried out by recording an NIR spectrum of a carrier board provided with a wear layer but not yet pressed as a reference sample, which is tested for abrasion resistance after the pressing process.
[0179] For this purpose, a printed HDF board is evenly coated with liquid melamine-formaldehyde resin with glass and corundum particles from above on a coating machine by means of an applicator roller via several roller applicators with intermediate drying. The amount of solid particles in the total coating varies depending on the abrasion class produced and ranges from 10 to 50 g/m.sup.2. The solid particles used have a diameter between 10 and 100 μm.
[0180] Before the pressing process in the KT press, an NIR spectrum is recorded from the coated carrier board in a predetermined section of the carrier board.
[0181] The board is then pressed in a short-cycle press at 200° C. and 40 bar for approx. 8 seconds. During this process, the protective layer is completely hardened. After the board has cooled down, several (especially four) 10 cm×10 cm samples (P1-P4) are taken for the abrasion resistance test. The samples for the abrasion resistance test are taken in the area of the board where the NIR spectrum was recorded.
[0182] The abrasion values are determined according to the procedure in DIN EN 15468:2006 (directly coated laminate flooring without overlay) with reference to DIN EN 13329:2017 and an average value is formed from the abrasion values and assigned to the measured NIR spectrum. In this way, several reference spectra of coated boards with different colour decors are recorded. A calibration model is created from the reference spectra, which can be used to determine or predict the abrasion resistance of an unknown sample. The calibration model is created using multivariate data analysis. This is done with suitable analysis software, e.g. with the previously mentioned analysis software The Unscrambler from the company CAMO.
[0183] The NIR spectrum was recorded in a wavelength range between 900 and 1700 nm. For the recording of the NIR spectra, the aforementioned NIR measuring device from Perten was used, whose measuring head is called DA7400.
[0184] Online Measurement of a Resin Coating with Wear Particles:
[0185] The measurement is carried out by recording NIR spectra of a synthetic resin layer (melamine resin) that is pre-dried but not yet post-cured in a short-cycle press on a carrier board (e.g. an HDF board), which is tested for its behaviour to abrasion stress after the pressing process. By measuring a large number of samples both spectroscopically and according to the standard for determining abrasion resistance, a dependency was previously determined via a calibration model.
[0186] NIR spectra of three samples with the same amount of resin applied but without corundum as wear particles or with different amounts of corundum are measured. These show different results when testing the behaviour against abrasion stress. The samples were tested in accordance with DIN15468 and DIN EN 13329: 2017—Laminate flooring—Elements with a top layer based on aminoplastic thermosetting resins, Annex E. For sample 1 (120 μm resin layer without corundum, upper dashed curve in the DIN standard), a wear class lower than AC2 was determined in the test of the behaviour against abrasion stress, for sample 2 (120 μm resin layer with 20 g corundum/m.sup.2, lower continuous curve in the DIN standard) a wear class AC 2, and for sample 3 (120 μm resin layer with 40 g corundum/m.sup.2, middle dot-dash-shaped curve in the DIN standard) a wear class AC 3. Samples 2 and 3 differ in the amount of wear-inhibiting particles.
[0187] In the NIR spectra determined, the chemical information of the absorption is superimposed by the scattering of the NIR light by the solid particles. In addition to the slight baseline shift, a slight change in the shape of the spectra can be seen, which is due to the scattering by the solid particles. With increased solid content, the scattering increases, especially at shorter wavelengths.
[0188] When creating a regression model, in addition to chemical information on the absorption, the scattering of the NIR radiation on the solid particles is also used to determine the behaviour in relation to abrasion stress. Accordingly, when creating the regression model, the spectroscopic data are related to the values obtained when testing the behaviour against abrasion stress.
[0189] Since the scattering of the NIR light by the solid particles makes a significant contribution to determining the behaviour in relation to abrasion stress, in addition to the main factors that explain the chemical variance of the samples, other main factors are also taken into account that describe, among other things, the morphology of the coating. The main factors here are the peaks in the spectrum, the scattering and the baseline shift.
Example 6
[0190] In an impregnation channel, a 30 g/m.sup.2 overlay is impregnated with a melamine resin (solids content: 55 wt %) in a first step in an impregnation bath. The working width of the dryer is 2070 mm and the feed rate is approx. 50 m/min. The melamine resin contains the usual additives such as hardeners, wetting agents, release agents, etc. Behind the impregnation tray there is a breathing section and a squeegee system or a pair of squeegee rollers with which excess resin is removed. The target resin coverage is approx. 300%. A scatterer is placed in front of a float dryer, with which approx. 20 g corundum/m.sup.2 is scattered into the still damp melamine resin. The corundum has a grain size of F220 according to the FEPA standard. The web is then dried to a residual moisture content of approx. 6% in the flotation dryer. Behind the float dryer is a chill roll system that cools the impregnate to room temperature. Behind the cooling rollers is an NIR multimeter head that moves on a traverse over the web and monitors the resin application, cross-linking, moisture and the amount of corundum applied. In the event of deviations from the target specifications, an automatic control or regulation system changes the squeegee/crush rollers, the spreading quantity and/or the temperature control in the dryer.
Example 7
[0191] In a coating plant, printed HDF boards (format: 2800×2070×7 mm) are first coated with a layer of melamine resin. The melamine resin is applied via a roller application unit to the print, which was previously covered with a pre-dried melamine coating of approx. 20 g melamine resin solid/m.sup.2. The application quantity is approx. 100 g melamine resin liquid/m.sup.2. The solids content of the resin is approx. 55% by weight and the resin contains the usual auxiliaries (hardener, wetting agent, defoamer, etc.). Approx. 30 g corundum/m.sup.2 are scattered into the melamine resin with the help of a scattering unit. The corundum has a grain size of F 220 according to the FEPA standard mentioned above. The boards are transported through an NIR dryer and dried by a transport system. Behind the dryer is a first NIR multimeter head mounted on a crosshead that determines the amount of resin applied, the amount of corundum applied and the moisture content of the board. In case of deviation from the target values, the NIR multimeter head, which is connected to the plant control system, adjusts the plant parameters (resin application, spreading quantity and dryer temperature). In further roller application units, approx. 30 g melamine resin liquid/m.sup.2 (solids content: approx. 55 wt. % with the usual auxiliary materials) is applied twice with subsequent intermediate drying (circulating air or IR drying). Afterwards, approx. 60 g melamine resin liquid/m.sup.2 (solids content: approx. 55 wt. % with the usual auxiliary materials) is applied in another roller applicator. Afterwards, approx. 20 g glass beads/m.sup.2 (Potters company, glass bead type 065-90) are scattered on with a scatterer. The resin is dried in an NIR dryer. Afterwards, the application quantity of the resin and the glass beads as well as the moisture and the degree of cross-linking of the resin are checked with another NIR multimeter head. In case of deviations from the target values, an automatic adjustment is made by the system control or regulation, which is connected to the further NIR multimeter head. In a final application unit, approx. 30 g melamine resin liquid/m.sup.2 (solids content: approx. 55 wt % with the usual auxiliary materials) is applied again. This resin is dried in an IR dryer. Afterwards, the board is pressed with an impregnated backing paper in a short-cycle press.