The invention relates to a method for determining cracking of wood. The method can be utilized in industrial processes. Internal cracking of massive wood is one of the most important factors to reduce the quality of wood in drying and in heat modi- fication. It would be important to determine the internal cracking using a nondestructive method. It would be important to detect and remove the defected wood from the industrial process without affecting the production of the high quality wood. With the proposed method, it is possible to estimate the number and position of cracks exceeding the critical size. Wood drying is considered to be the most important phase in wood processing to cause cracking of wood. The properties of each timber board and plank are different and thus the schedule used in drying is always a compromise and the cracking of wood during drying is still a significant problem in wood industries today. Cracking that takes place in wood during drying is mainly due to the moisture gradient devel- oping inside the wood that generates stresses within the wood because wood contracts non-uniformly when drying. The dried surface layer of wood shrinks more than the moister inner wood, resulting in cracks in the surface. Towards the end of the drying schedule, the stresses become opposite as there is compression on the surface and tension in the inner part of wood thus possibly causing internal crack- ing. In addition to the stresses caused by moisture gradient, there are stresses because of the anisotropic shrinkage. Additionally, there is always a certain degree of micro-cracking involved when wood dries, which, if uncontrolled, will propagate and develop into macro cracks. As wood is further processed, the other processes may also cause cracking especially if considerable mechanical stress is targeted at the wood. On the other hand, cracks also develop during the storage if relative humidity and temperature vary a lot. This is a problem especially in planks and wooden structures which are in contact with open-air.

Recently, the glued timber industry has increased a lot due to, e.g. an extensive multi-storey building production. Nowadays all glued timber used for structures is strength graded, but still from time to time, collapses of glued wooden structures are reported, in worst cases causing deaths . According to Sanabria's dissertation (Diss. ETH No. 20404, 2012), there was a recent study including 550 failure cases in 428 glulam roof structures constructed in Germany between 1912 and 2006. It was found that 70% of them were due to cracks in the grain direction of bending mem- bers, of which 62% were on the glue-line. Thus, 38%) have been caused by cracks in the wood. It may be assumed that there has been cracks exceeding critical size in the wooden members. When the glued wood members have been under high external forces, the size of the critical cracks has increased and finally there has been a collapse of a wooden structure. The described invention can be used to detect those critical cracks in production. Surface cracks can be detected using machine vision systems. Both surface and internal properties and defects of wood can be determined using combined X-ray and machine vision systems but the high cost and poor sensitivity to narrow cracks are problems. Microwave based detection methods have been developed but their problem is the high sensitivity for moisture and poor sensitivity for small cracks. Many studies show that the ultrasound technique can be used for measuring internal defects of wood. Density, strength and cracking of wooden panel products, e.g. plywood have been determined using on line ultrasound measurement. Acoustic devices using low frequencies have been developed to determine defects of wood. Because of the low frequency, the methods cannot be used for detecting small and very narrow cracks. Typical methods developed for industrial measurements are based on contact ultrasound, which can be realized for example, with rolling sensors. Contact ultrasound equipment has been developed especially for hardwood industry using roller transducers. The method was used successfully to determine cracks of wood in industry. According to the current state of art, the industrial contact ultra- sound measurement is not very reliable because the contact has to be very good and clean for stable measurements. Good contact can be achieved using water but it is not suitable for wood industries.

It has been shown that the determination of density of a wood based boards and lathe checks of veneer sheets is possible using air-coupled through-transmission ul- trasound. Other applications of air-coupled ultrasound for wood include the detection of integrity of a glue joint. Air-coupled ultrasound methods have been developed for detection of the glue voids in plywood. Commercial systems are available, e.g. from Grecon International (USA) and Electronic Wood Systems (Germany). Solutions based on air-coupled ultrasound have been introduced for determination of internal wood defects, e.g. using wide band ultrasound sensors to determine the change in resonance frequency due to a glue defect.

Typically piezoelectric sensors are used for detecting ultrasound signals. The sensors change the mechanical vibration to small electrical signal which is amplified using preamplifier. Then the measurement signal is amplified, filtered and analyzed. Typical analyzed signal parameters are attenuation and the transit time. The fre- quency content of the signal is also often analyzed. In addition, the development of capacitive air-coupled sensors has been rapid lately and industrial capacitive sensors are available nowadays.

Ultrasound has been used for quality control, e.g. in food, plastic and metal indus- tries. Ultrasound has been used in wood processing industries, e.g. for strength determination and detecting glue voids in plywood production.

The object of the present invention is to provide a new method for the determination of cracking in wood to overcome a number of drawbacks associated with the current production methods. More specifically, the object of the invention is to provide a method that allows reliable, easy, quick and efficient method for monitoring and controlling the production of wood products. The processes comprise, e.g. planing, gluing or thermal modification.

The object of the invention is achieved with a method characterized by which is presented in the claims. With the method in accordance with the invention, ultrasound is used to measure a cross section of wood using air-coupled ultrasound sensors at opposite sides of the object. The sensors are aligned at an angle of 0.2 - 25 degrees with respect to sample surface normal at the same side of the cross section plane. Two ultrasound sensors are used to send ultrasound bursts into wood and the cracking of wood affects the signal when it propagates through the wood and is then received using receiver ultrasound sensors. The direction, quality and the number of cracks are determined by measuring amplitude, velocity, phase and shape of the signal using measurements at least in two different directions and different frequencies. The essential feature of the method is that the size, shape, frequency and angle of the sensors are chosen so that the wave interface does not significantly affect the measurement e.g. the wavelength of the propagating signal is much higher than the phase difference in the sensor caused by the used angle.

The propagation of ultrasound wave is affected by the other properties of wood including density and moisture content. Their effect can be efficiently eliminated by determining reflections which are not sensitive to cracking. On the other hand, the basic properties of wood can be determined using measurements from the regions which are not cracked. This requires that the wood is not thoroughly cracked which can be easily observed with ultrasound measurements. The effect of knots on the propagation of the ultrasound signal is very different compared with the effect of cracking and their effect can be eliminated using multi-parameter signal analyses. The invention can be used for quality control in industrial line and as a separate online test equipment.

The air-coupled measurement enables the contactless determination and thus the resins, dirt, dust and other possible materials on wood surface do not destroy the sensors and their effect is by far lower than using contact measurement. By using the air-coupled measurement, the sensors do not affect the production and tight alignment system is not necessary.

In a preferred embodiment of the invention, two sensor pairs using two frequencies are used to determine cracking through the face of sample using both vertical and horizontal directions. The sensors are placed at the opposite sites of the wood and aligned at an angle of 0.2 - 25 degrees with respect to the surface normal vector at the same side of the cross section plane. The direction of the crack has an effect on the ultrasound response. If the direction of a crack is parallel with the propagating ultrasound wave, the effect of a crack is negligible or sometimes a crack may even amplify the signal. Typically the effect of a crack is the highest when the crack is transversally oriented compared with the ultrasound propagation direction. By comparing the signals measured at two different directions from the same position of wood it is possible to estimate the angle of a crack. By choosing the preferred fre- quencies the wave interference of the signals can be minimized. On the other hand, it is possible to conduct measurements at different positions in lengthwise direction on a lengthwise wood processing line and by synchronizing the measurements they can be combined as they were conducted from the same cross section of wood.

In a preferred embodiment of the invention, air-coupled ultrasound is used to de- termine cracking of wood. By monitoring the wood by measuring cross sections using sensors that are placed at the opposite sides of the wood and aligned at an angle of 0.2 - 25 degrees at the same side of the cross section plane using two frequencies, it is possible to determine cracking with higher accuracy and speed. Thus it is possible to speed up and optimize the wood processing formulas for different types and sizes of wood to produce high quality wood without cracking.

In the following, the invention is presented in greater detail with reference to the attached drawings where

Figure 1 is a schematic of one embodiment of the method in accordance with the invention; and Figure 2 is a schematic of another embodiment of the method in accordance with the invention.

Figure 1 shows a cross section of wood 5 which is moving on a lengthwise production line. Air-coupled ultrasound sensors are positioned above, below and on the side positions (1-4). Sensors 1 and 2 are used as transmitters and sensors 3 and 4 as receivers. The number of the sensors may vary according to the application and wood size and they can be positioned as is practical and appropriate.

Figure 2 shows a schematic viewed from up to down and wood (6) is moving lengthwise on a production line. Air-coupled ultrasound sensors are positioned above, below and on the side positions and one sensor pair including transmitter (1) and one receiver (3) is shown. The number of the sensors may vary according to the application and wood size and they can be positioned as is practical and appropriate

The following section provides a description of the invention by means of examples. Example 1. Determination of cracking from wood moving on a lengthwise production line

The following section provides a description of the first embodiment of the invention with reference to Figure 1 in which cracking is determined from a plank moving on a lengthwise production line. In this example of an embodiment of the inven- tion, the sensors measuring through the edge are GMP (Gas Matrix Piezoelectric) sensors at 116 kHz and the sensors measuring through the face are GMP sensors at 210 kHz. The sensors are attached into the body of the production line and they are positioned in a manner that the transmitting and the receiving sensors are on the opposite sides of the wood being measured, as advantageously at 50 mm distance from the wood sides. The sensors are advantageously at opposite directions in an angle of 10 degrees from the normal vector of the wood surface. The sensors are acoustically shielded by attaching them with rubber insulated securing bolts. Before the beginning of the measurements, the ultrasound measurement is calibrated measuring without the wood between the sensors and the reflection from the receiving sensor can also be utilized. The sensors are connected to the ultrasound measurement unit, which comprises of pre-amplifier, amplifier and low-pass and high-pass filters. The signal from the measurement unit is lead to signal analyzer and to control unit of the measurement which can be used to control the measurement and to implement the analysis of the measurement. The measurement values from the output of the analyzer are used for the determination of the cracking of the wood.

In this example, the control and analyses unit is a computer with analog/digital converter which converts the measured signals suitable for the control and analysis software. There is a software in the computer which continuously measures the channels used for the ultrasound method, analyses the results and sends the measurement data through wire or wirelessly to monitoring computer. The classification of the planks is controlled with the same computer, which can mark or remove the planks or parts of the planks that are too cracked. Information on the size and posi- tion of the cracking is gained from each plank. Certain critical cracking thresholds, which must not be exceeded, have been determined for each timber product type. The thresholds have been determined based on theoretical calculations which have been experimentally verified for example by testing the planks destructively.

Example 2. Application of multiple variable methods In this example of an embodiment of the invention, the method is similar to example 1 or examples 3 through 32, with reference made to Figure 1 and 2, except that the method used for analysing the measurements is a multiple variable method. The multiple variable method may, for example, involve the use of multiple variable regression, principal component regression (PCR) or partial least squares (PLS). By making measurements on the calibration samples, it is possible to formulate a multiple variable matrix that shows, in separate columns, the actual crack information measured from each calibration sample as well as the values for the specified ultrasound parameters. The number, size and orientation of cracks can be determined by means of a standard destructive test method and visual analyses. By using one of said multiple variable methods, the error between fitting of the crack function and the actual cracks can be minimized. This yields a multiple variable matrix that contains the factors for the parameters as well as the constant factors. The cracks are determined by measuring the ultrasound parameters from the sample and using the crack functions based on the specified matrix factors. Both linear and non-linear multiple variable methods can be used. . The non-linear methods include Bayes-classifier, neural networks, genetic algorithms and nearest neighbour classifiers. The Bayes classifier was used in the current example. At first a training set is used for the model by using determined ultrasound parameters from cracked and non-cracked wood. After the ultrasound measurements, the wood sam- pies are analysed using destructive method to analyse the cracking from cut cross sections as used with the above multi-regression models. After training and testing the model, the model can be used to predict the class of cracking using air-coupled ultrasound measurement.

Example 3. Determination of cracking from plank moving on a transversal produc- tion line

In this example of an embodiment of the invention, the method is similar to examples 1 through 2 or examples 4 through 32, with reference made to Figure 1 and 2. In this embodiment, the cracking is determined from a plank moving transversally. Thus, the comprehensive measurement can be realised by using, for example, ma- trix sensors, multiple sensors and/or by moving and/or scanning the sensor or sensor pairs to measure the plank with appropriate comprehensiveness.

Example 4. Inclined measurement.

In this example of an embodiment of the invention, the method is similar to examples 1 through 3 or examples 5 through 32, with reference made to Figure 1 and 2. In this embodiment, the oppositely connected ultrasound sensors are not perpendicular to each other but the angle between the transmitting sensor and the receiver is adjusted to enable the movement of the wood between, above or below and/or from the side of the sensors.

Example 5. Measurement deviating from the direct line In this example of an embodiment of the invention, the method is similar to examples 1 through 4 or examples 6 through 32, with reference made to Figure 1 and 2. In this embodiment the ultrasound sensors in the opposite sides are not directly facing each other.

Example 6. Determination of cracking as comparative measurement In this example of an embodiment of the invention, the method is similar to examples 1 through 5 or examples 7 through 32, with reference made to Figure 1 and 2. In this embodiment, a plank is advantageously measured before and after processing (e.g. drying or thermal treatment), whereupon the changes in the plank taking place during the processing can be very accurately determined and the accuracy for the crack determination can be enhanced by comparing the results from the same plank with the earlier measurements.

Example 7. Determination of cracking as differential measurement In this example of an embodiment of the invention, the method is similar to examples 1 through 6 or examples 8 through 32, with reference made to Figure 1 and 2. In this embodiment, planks are advantageously scanned and several sensors or matrix sensors are used and the differences between spatially adjacent measurements are determined. Thus, considerably smaller differences in the measurement parameters can be observed compared with the direct measurement values. By using the speed and amplitude of the change in the ultrasound parameters as well as the spatial dimension, the cracking analysis can be further defined.

Example 8. Determination of cracking in strength grading In this example of an embodiment of the invention, the method is similar to examples 1 through 7 or examples 9 through 32, with reference made to Figure 1 and 2. In this embodiment a plank is measured before and after strength grading in which small mechanical stress is applied into the wood. Thus, the changes in the plank can be very accurately determined and the accuracy of the strength grading can be en- hanced as the results can be compared with the earlier measurements from the same plank. Thus, especially the cracks arising and/or propagating/propagated during the stress can be detected.

Example 9. Monitoring method

In this example of an embodiment of the invention, the method is similar to exam- pies 1 through 8 or examples 11 through 32, with reference made to Figure 1 and 2, except that in this application the link to the control system is lacking. Thus, the method, as presented in this example, only serves the purpose of monitoring the cracking and outputting readings to the monitoring device. This application may be used, for instance, when the drying process is to be optimized, the drying process is to be monitored or the optimum drying of wood species other than those already calibrated is to be studied.

Example 10. Control method

In this example of an embodiment of the invention, the method is similar to examples 1 through 8 or examples 11 through 32, with reference made to Figure 1 and 2, except that no monitoring system is provided in this application. As a result, the method only serves the purpose of classifier. However, this embodiment of the invention can be used for routine processes or when the quality of the wood is to be optimized, instead. Example 11. Compress drying

In this example of an embodiment of the invention, the method is similar to examples 1 through 10 or examples 13 through 32, and reference is made to Figure 1 and 2. In this embodiment, the method is configured to determine the cracking due compress drying, which is affected not only by the drying process but also by the compressive force applied. The method can be used for determining cracking due to the process by measuring the wood before and after the process.

Example 12. Heat modification

In this example of an embodiment of the invention, the method is similar to exam- pies 1 through 10 or examples 14 through 32, and reference is made to Figure 1 and 2. In this embodiment, the method is configured to determine the cracking due to heat modification, which is affected not only by the drying process but also by the level of heat modification. The method can be used for determining cracking due to the process by measuring the wood before and after the process. Example 13. Other wood modifications

In this example of an embodiment of the invention, the method is similar to examples 1 through 10 or examples 15 through 32, and reference is made to Figure 1 and 2. In this embodiment, the method is configured to determine the cracking due to wood modification, which can be e.g. chemical treatment, biological treatment, im- pregnation, oil treatment or surface modification or any combination of these modifications. The ultrasound method can be used for determining cracking due to the process by measuring the wood before and after the process.

Example 14. Other wood product applications

In this example of an embodiment of the invention, the method is similar to exam- pies 1 through 13 or examples 15 through 32, and reference is made to Figure 1 and 2. In this embodiment, the object of the measurement is not a plank but the method can also be applied for other type wood products or wood-based products comprising glued wood products, beams, round wood, wood cut to different shapes, wood- plastic composites and sandwich-structures. Example 15. Glued wood products

In this example of an embodiment of the invention, the method is similar to examples 1 through 14 or examples 16 through 32, and reference is made to Figure 1 and 2. In this embodiment, the object of the measurement is not a plank but the object comprises a glue line. In this embodiment, the method can be applied for distinguishing the internal crack from glue line defects, whereupon both the glue line defects and internal cracking can be determined more accurately than if applying a known method for detection of glue line defects.

Example 16. Wood component production line measurement

In this example of an embodiment of the invention, the method is similar to examples 1 through 15 or examples 17 through 32, and reference is made to Figure 1 and 2. In this embodiment, the method is applied to determine cracking in a component production line, where components are produced for various products. The product can be for example a glue-jointed timber, whereupon the internally cracked components or their parts can be discarded automatically on the production line in a similar manner as knots are removed.

Example 17. Static measurement In this example of an embodiment of the invention, the method is similar to examples 1 through 15 or examples 17 through 32, and reference is made to Figure 1 and 2. In this embodiment, the method is applied to determine cracking in a stationary wood object, whereupon the method can be applied to determine the cracking in the position of the measurement. By moving the sensors, for example, the size of the crack can be estimated in longitudinal and width direction and the total cracking of the object can be determined by manual scanning and/or by using a robot technique.

Example 18. Measurement of wooden structures

In this example of an embodiment of the invention, the method is similar to examples 1 through 17 or examples 19 through 32, and reference is made to Figure 1 and 2. In this embodiment, the method is applied to determine cracking in a stationary wooden structure, whereupon the method can be applied to determine the cracking in the position of the measurement. By moving the sensors, for example, the size of the crack can be estimated in longitudinal and width direction and the total cracking of the object can be determined by manual scanning and/or using a robot technique. Example 19. Measurement of glulam beams

In this example of an embodiment of the invention, the method is similar to examples 1 through 18 or examples 20 through 32, and reference is made to Figure 1 and 2. In this embodiment, the method is applied to determine cracking in a glulam beam structure, whereupon the method can be applied to determine the cracking in the position of the measurement. By moving the sensors, for example, the size of the crack can be estimated in longitudinal and width direction and the total cracking of the object can be determined by manual scanning and/or using a robot technique.

Example 20. Multiple sensor application

In this example of an embodiment of the invention, the method is similar to examples 1 through 19 or examples 21 through 32, and reference is made to Figure 1 and 2. In this embodiment, the method is combined with, for example, electrical imped- ance, radio wave, microwave, optical, IR, NIR, X-ray or gamma-ray measurement, which is used to determine other properties of the plank comprising the moisture content, density, knottiness, angular grain, heartwood/sapwood ratio and/or juvenile wood fraction. Thus, the method can be improved because the applied parallel method is not sensitive to cracking and thus the erroneous measurements caused by another defect than cracking, can be effectively eliminated.

Example 21. Measurement of dimension

In this example of an embodiment of the invention, the method is similar to examples 1 through 20 or examples 22 through 32, and reference is made to Figure 1 and 2. In this embodiment, as the cracking is measured, the method can be applied to determine the dimensions of the measured object using reflection measurements. The reflected pulse can be determined both with the transmitting sensors and in the through-transmission measurement, with the receiving sensors, whereupon the distance can be determined from all four sides. For the transmitting sensor, the measurement is a normal reflection measurement, whereas in the through-transmission measurement the receiving sensor can be used to determine the other burst which is another reflection of the through-transmitted signal from the plank surface. For an accurate dimension measurement, a separate air temperature and humidity measurement is required if the conditions of the measurement location vary considerably.

Example 22. Measurement of plank shape distortions In this example of an embodiment of the invention, the method is similar to examples 1 through 21 or examples 23 through 32, and reference is made to Figure 1 and 2. In this embodiment, as the cracking is measured and similarly to the dimension measurement, the method can be applied to determine the shape distortions comprising wane, twist, bow, crook and cup using reflection measurements. If the shape distortion is considerable, the distortion measurement can be realised as in the example 21.

Example 23. Determination of cracking in a shape distorted plank

In this example of an embodiment of the invention, the method is similar to exam- pies 1 through 22 or examples 24 through 32, and reference is made to Figure 1 and 2. If the shape distortion is considerable, the determination of cracking can be challenging, because the surfaces of the plank are not in the right position in relation to the transmitting and receiving sensors. In this embodiment, the plank can be guided, for example, by four rollers (two above and below before and after the measurement location) to fulfil the requirement of position of the surfaces in relation to the sensors in the measurement location.

Example 24. Calibration method

In this example of an embodiment of the invention, the method is similar to examples 1 through 23 or examples 25 through 32, and reference is made to Figure 1 and 2. The method can be continuously calibrated by measuring only air while the object of the measurement is not in the location of the sensors. Thus the reference values for each ultrasound parameter can be determined and the variation in the conditions during long-time measurements can be taken into account.

Example 25. Automatic measurement for various dimensions In this example of an embodiment of the invention, the method is similar to examples 1 through 24 or examples 26 through 32, and reference is made to Figure 1 and 2. The method can be applied to determine the location and dimension of the plank continuously, as in example 21 (Measurement of dimension). Thus, the method for determination of cracking can continuously utilize earlier determinations of suitable analyses algorithm for each dimension.

Example 26. Single-sensor reflection and through-transmission measurement

In this example of an embodiment of the invention, the method is similar to examples 1 through 25 or examples 27 through 32, and reference is made to Figure 1 and 2. The method can be applied with a reflector in the place of the receiving sensor re- fleeting the through-transmitted signal into a preferred direction, where the signal can be reflected as air-coupled back to the transmitting sensor with reflec- tor/reflectors. Thus a single sensor can be applied for reflection and through- transmission measurement.

Example 27. Application with capacitive ultrasound sensors

In this example of an embodiment of the invention, the method is similar to exam- pies 1 through 26 or examples 28 through 32, and reference is made to Figure 1 and 2. The method can be applied with a capacitive sensor as transmitting and/or receiving sensor.

Example 28. Application with matrix ultrasound sensors

In this example of an embodiment of the invention, the method is similar to exam- pies 1 through 27 or examples 29 through 32, and reference is made to Figure 1 and 2. The method can be applied with a matrix sensor as transmitting and/or receiving sensor/sensors. With the matrix sensor, multiple measurements can be conducted from various positions of the studied object. Matrix sensor can be realized, for example, based on the capacitive sensor technique. Example 29. Exchange sensor application

In this example of an embodiment of the invention, the method is similar to examples 1 through 28 or examples 30 through 32, and reference is made to Figure 1 and 2. The method can be applied with changing the function of a transmitting and/or receiving sensor/sensors, i.e., the transmitting sensors become receiving sensors and receiving sensors become transmitting sensors and the measurements are conducted on the same or a near-by position. The propagation of ultrasound burst is analysed in both directions, and thus the measurement can be improved.

Example 30. Multiple receiver application

In this example of an embodiment of the invention, the method is similar to exam- pies 1 through 29 or examples 31 through 32, and reference is made to Figure 1 and 2. The method can be applied using two or more sensors instead of one transmitting and/or receiving sensor. By setting the sensors in preferred angle and relation to each other, the measurement can be further improved and speeded up, because the measurement can be made in two or many positions in the wood. This is advanta- geous especially in the fast on line -measurement.

Example 31. Strength measurement In this example of an embodiment of the invention, the method is similar to examples 1 through 30 or 32, and reference is made to Figure 1 and 2. The method can be applied in strength measurement by analysing the elastic properties of the wood material in the positions without cracking or other internal defects. By combining the measurements from the cracked positions with this information, the strength of the wood can be estimated more accurately than before.

Example 32. Multi-sensor application for strength measurement

In this example of an embodiment of the invention, the method is similar to examples 1 through 31 and reference is made to Figure 1 and 2. In this embodiment, the method is combined with, for example, electrical impedance, radio wave, microwave, optical, IR, NIR, X-ray or gamma-ray measurement, which is used to determine other properties of the plank comprising the moisture content, density, knotti- ness, angular grain, heartwood/sapwood ratio and/or juvenile wood fraction. Thus, the strength measurement method (Example 31) can be improved because, for ex- ample the moisture content of the wood can be determined very accurately.

The practical applications of the invention are not limited to the above examples, instead, the invention and its embodiments may be varied within the scope of protection provided by the claims.