The agreement in accuracy between tomograms, resistograms, and the actual condition of the wood from lime trees harvested from cities

Abstract

The internal quality of the wood is one of the main factors affecting the stability of trees, and it has always been of great interest to science and practice. For this reason, the present study aims to compare the results obtained by wood tomograms with those of resistance to drilling and the visual appearance after cutting a slice with a chain-saw, both to evaluate the presence and dimensions of the inside defects, and also to evaluate the irregularities of the wood structure. Round pieces of lime wood harvested from public areas were used for comparison by taking sound tomograms, followed by taking resistograms on two perpendicular directions at the same level. The results showed that internal wood defects are not always the ones that lead to reduced speeds of sound propagation through the wood. In addition, there were instances in which changes in the internal structure of the wood led to improperly colored tomograms, namely the sections characterizing the point of insertion of a thick branch in the trunk, where the tomograms indicated low speeds of sound transfer through the wood in the stem and high speeds in the wood of the branch.

Full Article

The Agreement in Accuracy between Tomograms, Resistograms, and the Actual Condition of the Wood from Lime Trees Harvested from Cities

Elena C. Mușat *

The internal quality of the wood is one of the main factors affecting the stability of trees, and it has always been of great interest to science and practice. For this reason, the present study aims to compare the results obtained by wood tomograms with those of resistance to drilling and the visual appearance after cutting a slice with a chain-saw, both to evaluate the presence and dimensions of the inside defects, and also to evaluate the irregularities of the wood structure. Round pieces of lime wood harvested from public areas were used for comparison by taking sound tomograms, followed by taking resistograms on two perpendicular directions at the same level. The results showed that internal wood defects are not always the ones that lead to reduced speeds of sound propagation through the wood. In addition, there were instances in which changes in the internal structure of the wood led to improperly colored tomograms, namely the sections characterizing the point of insertion of a thick branch in the trunk, where the tomograms indicated low speeds of sound transfer through the wood in the stem and high speeds in the wood of the branch.

DOI: 10.15376/biores.18.1.1757-1779

Keywords: Sound tomograms; Power drill machine; Relative resistance to drilling; Lime trees; Wood quality; Internal defects

Contact information: Department of Forest Engineering, Forest Management Planning and Terrestrial Measurements, Faculty of Silviculture and Forest Engineering, Transilvania University of Brasov, Şirul Beethoven 1, 500123, Romania; Corresponding author: elena.musat@unitbv.ro

INTRODUCTION

Green areas of cities, parks, and public gardens are enjoyed by the community (Kirkpatrick et al. 2012), so those responsible for these areas aim to obtain as many varieties as possible of shrubs and trees, with different shapes and habits (Camacho-Cervantes et al. 2014).

Regardless of the area in which they are located, during their existence, the trees change their internal structure and shape, either due to the natural causes imposed by the stages of development or environmental conditions, or due to anthropogenic causes such as fires or pruning (Musat et al. 2020). The anthropogenic actions mostly occur in trees from cities, whose growth and development are influenced either by pruning of crowns (Seifert et al. 2010) or by reducing the space for root development, particularly in the case of street trees because they are forced to develop their root system among the cables and pipes in the soil, near unsuitable materials (Saebø et al. 2005; Bartens et al. 2010), or soils poor in nutrients (Parascan and Danciu 2001).

As any human intervention in nature does not remain without consequences (Nimară et al. 1964; Suciu 1975), trees also react through their own adaptation processes. Sometimes the anthropogenic actions are so severe that they affect the integrity and even the stability of the trees, especially in the case of trees located in the cities, near the road, which are cut extremely intensively. In contrast to trees from forests managed for production, whose management is aimed at getting high quality wood for various industrial uses (Sandoz and Lorin 1996; Garrett 1997; Wang et al. 2007; Mu et al. 2010; Du et al. 2015; Qu et al. 2020; Lin and Wu 2013; Sandak et al. 2020), for urban trees the main aim is to maintain their vitality, integrity, and stability for as long as possible (Kirkpatrick et al. 2012; Camacho-Cervantes et al. 2014; Sandoz and Lorin 1996; Deflorio et al. 2008; Wang et al. 2009). Initially, the condition of a tree and its maintenance or removal was decided by a single person, who visually evaluated it (van Wassenaer and Richardson 2009). Then, the verification of the internal quality of wood in standing trees became a long-term concern (Roughton 1982; Bucur 1986; Bucur 2003; Alvers et al. 2015). As a result, a series of tree condition assessment devices were developed, some of which are highly invasive, some are less invasive, and the rest are considered non-invasive (Catena 2004; Deflorio et al. 2008; van Wassenaer and Richardson 2009).

A number of tools have been developed to evaluate the integrity of wood in standing trees based on the principle of wave propagation in solid environments (Wang et al. 2007; Lin et al. 2011; Lin and Wu 2013; Wang 2013; Li et al. 2014; Alves et al. 2015; Bouchet and Danneau 2017), namely those using sounds (Deflorio et al. 2008; Rohanova 2009; Brancheriau et al. 2012) and ultrasounds (Tomikawa et al.1986; Sandoz and Lorin 1996; Garrett 1997; Martinis et al. 2004; Alves et al. 2015). By such determinations, one can get images of the analyzed sections (Liang and Fu 2012; Feng et al. 2014; Li et al. 2014; Alves et al. 2015; Du et al.2015), either by using various computational algorithms (Sandoz and Lorin 1996; Du et al. 2015) or software provided by the tool’s producers (Rinn 2014).

Sound is a wave that propagates through compression and expansion of the environment in which it develops (Beldeanu 2008; Bouchet and Danneau 2017). Knowing the speed of sound through a material is important because it can provide clues about the nature and purity of the material (Bouchet and Danneau 2017). In the case of wood, as a solid material, this principle is used in sound analysis with the aim of detecting hidden irregularities located inside the wood (Garrett 1997; van Wassenaer and Richardson 2009; Ellis 2014; Alves et al. 2015; Du et al. 2015).

Acoustic wood quality assessment methods can be applied both to standing trees (Garrett 1997; Martinis et al. 2004; Lin et al. 2008; Lindström et al. 2009; Feng et al. 2014) and harvested round wood (Sandoz and Lorin 1996; Rohanova 2009). The main difference in the application of methods would be that of different direction in which the sound is propagated (Fu 2005; Beldeanu 2008; Kazemi et al. 2009). As such, in trees the acoustic method can be used only at the level of some cross-sections (Martinis et al. 2004; Wang et al. 2007; van Wassenaer 2010; Feng et al. 2014; Li et al. 2014; Musat et al. 2014; Rinn 2014) located on the stem, branches, or even roots (Sandoz and Lorin 1996; Malinovski et al. 2016), while for harvested stems or logs, the method can be used also parallel to the fibers, in the longitudinal plane (Lear 2005; Rohanova 2009; Wang and Carter 2015).

The speed of sound propagation on the direction of the fibers is species variant, being 3 to 5 times faster than the speed of propagation perpendicular to the fibers (Beldeanu 2008), and it depends on the angle between the emitter-receiver sensor pairs (Feng et al. 2014; Li et al. 2014; Du et al. 2015). This is because the sound wave must cross all annual rings, earlywood and latewood, wider or narrower rings (Sandoz and Lorin 1996; Beaulieu and Dutilleul 2019). In addition, there are variations in propagation speed occurring in the same species. This is common in trees of the same species that have developed in contrasting environments, which caused an impact on the internal structure of the wood (Beldeanu 2008; Lindström et al. 2009; Dinulica et al. 2020). In particular, such changes affect the density of wood, which is known to affect the speed of sound propagation (Tarasiuk et al. 2007; Wang et al. 2007; Beldeanu 2008; Deflorio et al. 2008; Liang and Fu 2012; Dinulica et al. 2016; Bouchet and Danneau 2017), and which increases proportionally to the wood density. In addition, the anisotropy of wood, as one of its main characteristics (Beldeanu 2008; Feng et al. 2014; Du et al. 2015), produces variation in characteristics such as the mechanical and physical behavior across its mass (Lunguleasa 2004; Leboucher2014; Beaulieu and Dutilleul 2019).

When some internal defects are present, they affect the speed of sound propagation in wood (Sandoz and Lorin 1996; Ross et al. 1998; Ross and De Groot 1998; Martinis et al. 2004; Wang 2013; Wu et al. 2018; Moravcki et al. 2021), providing clues about the internal structure. However, the speed of sound propagation is influenced by a lot of factors, of which not all could be seen as defects, i.e., factors that do not affect the integrity of wood.

Because acoustic analysis does not indicate the type of defect or its exact extent (Deflorio et al. 2008; Feng et al. 2014), and by doing so, it either overestimates (Wang et al. 2009) or underestimates (Martinis et al. 2004; Liang and Fu 2012), it is necessary to increase the number of used sensors (Divos and Divos 2005; Wang et al. 2007; Wunder et al. 2013; Du et al. 2015) or to carry on additional analyses (Tarasiuk et al. 2007; Siegert 2013; Feng et al. 2014) with the aim to determine the type of defect and its extent, and to evaluate its impact on the stability of the trees (Siegert 2013). For instance, it is accepted that the risks become important in case of defects or degradations that affect more than 60% of the analyzed diameter (Sandoz and Lorin 1996).

Due to the structural changes that some natural irregularities of the wood (forking, knots – Balleux 2004; Budakci and Cinar 2004; Alves et al. 2015) or defects (Lunguleasa 2004; Beldeanu 2008; Feng et al. 2014; Du et al. 2015; Du et al. 2018) have on the stability of trees, and the important role that trees are playing in public areas (Saebø et al. 2005; Kirkpatrick et al. 2012; Troxelet al. 2013; Camacho-Cervantes et al. 2014), it remains particularly important to periodically evaluate the internal quality of standing trees (Proto et al. 2020).

The goal of this work was to evaluate the agreement between sound tomograms and the true status of the wood, in the case of lime trees, to evaluate both the inside defects and the irregularities of the wood structure. The following objectives were set for this study: i) to compare the agreement between the sound tomograms and the true status of the wood; ii) to compare the sound tomograms with the diagrams with the relative resistance to drilling; and iii) to check whether the acoustic tomograph could identify the irregularities inside the wood.

EXPERIMENTAL

Field sampling was carried out at one of the teaching facilities of the Faculty of Silviculture and Forest Engineering of the Transilvania University of Brasov, in the spring. The determinations were made on lime round wood, at natural moisture, because the trees were harvested in the same week, a few days earlier. The choice of this species was based on the statements from the literature (Saebø et al. 2005; David 2011; Musat et al. 2014), according to which Tilia species are very common in the cities, and the characteristics of the wood are different compared to those of hardwood species.

As specialists (Tarasiuk et al. 2007; Feng et al. 2014) recommend the use of different methods for the correct identification of defects and their extension, the working methodology first involved performing analyzes by the means of Arbotom® Sonic Tomograph (Rinntech – Fig. 1a), followed by checking the relative resistance of wood to drilling (Fig. 1b) on two perpendicular directions, which was done by the IML Resi F-500S PowerDrill® and, finally, by extracting wood samples in the form of discs at each analyzed level by the use of a motor chain-saw (Fig. 1c). To compare the results with the true status of the wood, each newly created surface was photographed for further analysis at the office.

Fig. 1. The field sampling: a) the sensors were placed around the wood piece for acoustic analyses; b) the measurements were made by a wood drilling machine; c) the cross-cut for evaluating the true status of the wood

An Arbotom®–Rinntech Sonic Tomograph was used to measure the sound propagation. Measurements were done at the bottom/end, and then spaced at 50 cm levels, in directions perpendicular to the wood fibers. Measurements involved fixing the sensors on the circumference of the logs, with the help of special steel nails. The first piece was intended to capture the effect that defects may have on the speed of sound propagation. Taking into account the recommendations from the literature (Divos and Divos 2005; Karlinasari et al. 2011; Li et al. 2014; Rinn 2014), the number of sensors was chosen according to the diameter and complexity of the cross sections at the analyzed levels, being used between 6 and 18 sensors.

After placing the sensors, creating and verifying the connection between the sensors and the Arbotom® soft installed on a laptop, the diameter and the position of each sensor on the circumference were entered in the program (Fig. 2a), including the deviations from the circular shape, when necessary. Measurements were done by inducing sound pulses by successive actuation of the sensors. During the measurements, each sensor acted as a transmitter and receiver (Alves et al. 2015; Proto et al. 2020; Morovcik et al. 2021). Thus, starting with sensor number 1, each sensor was hit with a metal hammer 7 times to generate sound waves that propagated to all other sensors, which acted as receivers. The number of pulses was chosen according to the ambient noise (Tarasiuk et al. 2007; Musat et al. 2020) and based on the recommendations of the manufacturer (http://rinntech.de, 5 to 10 pulses for each transmitter, and the number of required pulses increasing with the noise level in the area). When the wave from the transmitter reaches a receiver, the tomograph program automatically calculates the speed of sound propagation through the wood. During the measurements, the value of the transmission errors was permanently monitored so as to be less than 3%, as specified in the literature (Wang et al. 2007).

Fig. 2. The software of Arbotom® tomograph: a) Rectifying the circumference of the transverse section of the round wood: with the red line the initial shape, and with blue line the rectified profile; b) The lines drawn by the specific soft of the sonic tomograph between the transmitters sensors (1 and 2) and the receivers (other sensors); c) The connections between the pairs transmitter-receiver sensors; d) The tomogram reconstructed based on the average speed of sound propagation through the wood

Based on several propagation speeds recorded between the transmitter and the receivers, the tomograph program calculates an average speed according to which it draws a link line between each transmitter-receiver pair (Fig. 2b). As all the sensors play the role of transmitters and receivers, for a given section the program builds a set of connecting lines between the sensors (Fig. 2c). Based on these speeds, the program constructs a colored tomographic image (Wu et al. 2018), which indicates/segments by a color palette. The healthy wood or areas without internal irregularities are where the wave can be transmitted faster (http://au.ictinternational.com/casestudies/example-arbotom-raport), which is in contrast to areas having rot, degradation, or mechanically damaged wood (http://ictinternational.com/casestudies/detecting-fungal-decay-in-palm-stems-by-resistance-drilling), (Fig. 2d).

Measurement of the drilling resistances was done using an IML Resi F-500S PowerDrill®, which enables the penetration of the drill into the wood at a constant pace, making it possible to get the variation in resistance as a function of penetration depth. The device was equipped with a drill of 50 cm in length and 3 mm in diameter (https://www.iml-service.com/product/iml-powerdrill/), which allowed the penetration of the entire section.

For each section at which a tomogram was taken, two measurements were done with the wood driller machine. The directions of measurement were always north-south and east-west facing, where the north direction corresponded to the position of the first sensor placed on the trunk to measure the speed.

Regarding the values of the relative resistance at drilling, it was assumed that the wood was healthy if on the diagram the values of resistance were uniform, without significant sudden oscillations (Rinn 1994; Proto et al. 2020) or if the resistances increased progressively from the periphery of the stem towards its center. In contrast, areas with rot are commonly identified by a sudden decrease in resistance, which tends to 0% (Wu et al. 2018), a behavior which is characteristic of parts with internal holes (hollows). In the same way, areas with wood in various stages of degradation or areas with structural irregularities, characterized by sudden and short-lived oscillations of relative resistance compared to those of the surrounding wood, can also be detected.

Following the analyses regarding the relative resistance at drilling, the round wood pieces were cross-cut by using a Husqvarna chain-saw, which was handled by a qualified operator. Each newly created surface was photographed using a photo camera Sony, model DSLR-A200k with lens SAL 18…70 mm. The present measurements also attempted to check whether the acoustic method can recognize the small defects inside the wood. Even if the defects smaller the 1 cm can be seen by the photo camera and by visual evaluation, these small defects cannot influence the stability of the entire tree.

The images were intended to reflect the true status of the wood inside the stem and were saved in relation to the number of wood samples and level analyzed, so that comparisons between tomograms, resistograms, and photographs could be made later. Probes and photographs of the sections were taken immediately after the measurements done with the sonic tomograph and the wood drill machine, so as to avoid the mistakes of association that might affect the interpretation of the results.

RESULTS AND DISCUSSION

Speed of Propagation and Resistance to Penetration

The measurements done by the tomograph resulted in a total of 31 tomograms and 62 resistograms, which were compared to the real condition of the wood, visible from the sections made with the mechanical chain-saw at each level.

The sounds were not always transmitted between all pairs of sensors (transmitter – receiver), so that the total number of formed links was less than the number of possible links. Such situations were identified only at the second (at level of 410 cm, between sensors pair 5-6 and 6-5) and the third piece of wood, at the level of 10 cm (sensors pair 5-6 and 6-5) and at 210 cm, between the sensors 3-4 and 4-3. This problem was observed also by other researchers (Du et al. 2015; Du et al. 2018), who mentioned that the accuracy of the tomograms near the sensors is significantly lower than that inside the trunk.

The results indicated that the highest share (73 to 94%) was of speeds between 1001 and 1500 m/s. At a first glance, this does not point out special problems since the literature sets a reference speed of 1400 m/s for lime wood (Sandoz and Lorin 1996). In dried healthy lime wood, the speed in the longitudinal direction is 3700 m/s (Beldeanu 1999; Beldeanu 2008); the same sources (Beldeanu 1999; Beldeanu 2008) also claim that the speed of sound perpendicular to the fibers is reduced by 3 to 5 times compared to that along the fibers.

However, there were large variations in the minimum values recorded, starting from 283 m/s (section from 56 cm – the first piece of wood), continuing to 300 m/s (section from 10 cm of the second piece), and reaching 1136 m/s (the 110 cm section of the third piece). Comparing the minimum values with the tomograms and the true status of the wood, only some of these values can be attributed to serious defects located inside the wood. In the second piece located at the level of 10 cm from the thick end of the stem, two defects were present, namely a hollow and a knot, and the minimum speed recorded on the direction of sensors 8-3 is justified by the presence of a rotten area in various stages of development (Fig. 3).

Fig. 3. Extreme values of the speed of sound propagation through wood: a) location of the sensors at the analyzed section; b) directions of sounds propagation

If the wood is healthy, then the wave can pass in a straight line from the transmitter to the receiver (Feng et al. 2014; Rinn 2014), while if the tree has rot at the analyzed level, the sound wave must bypass the affected area (Garrett 1997; Lin and Wu 2013). Even if the path of the wave is not clear in degraded wood, the speed of propagation of the sound is much slower than in wood without defects (Lin and Wu 2013; Wang 2013; Rinn 2014). Compared to all the other low values of the speeds, sometimes they have nothing to do with internal defects, as the wood is healthy. However, in the tangential direction, the propagation velocities are lower than those on the radial direction (Beldeanu 2008; Lin et al. 2008; Kazemi et al. 2009; Liang et al. 2010; Feng et al. 2014). In addition, the sections in which these values were recorded have an oval shape, which further supports the claims that the propagation velocities are closely related to the anatomical structure of the wood (Feng et al. 2014; Alves et al. 2015) and that an uneven width of the annual rings influences the density of the wood (Filipovici 1964; Sandoz and Lorin 1996; Beldeanu 2008).

Comparison of Tomograms with the Real State of the Wood at the Analyzed Levels

By comparing the tomograms with the newly created sections at the analyzed levels, it was found that in some cases the reconstructed image correctly illustrated the real condition of the wood (Figs. 4 through 7). This happened when the wood at the level of the analyzed section was healthy and did not show structural unevenness, which was observed also from the diagrams with the relative resistance to drill.

Fig. 4. The results from the level of 56 cm of the first piece of lime

Fig. 5. The results from the level of 136 cm of the first piece of lime

Fig. 6. The results from the level of 110 cm of the third piece of lime

Fig. 7. The results from the level of 210 cm of the third piece of lime

In some cases (Figs. 8 and 9), tomograms illustrated lower speeds of sound transfer through wood on the tomographic images, even if the actual condition of the wood indicated healthy wood. These two figures support, once again, the influence of the structural characteristics of wood on the speed of sound propagation (Sandoz and Lorin 1996; Lindström et al. 2009) and the fact that the tomograph cannot distinguish between the wood with defects and healthy wood, but with structural irregularities. The areas characterized by lower speeds were located either in the central area of the stem (Fig. 8) or in its lateral part (Fig. 9). The presence of wider annual rings was noticed in some sections, corresponding either to more favorable climatic conditions in the development of the tree (Rinn 1988; Beldeanu 1999; Beldeanu 2008), or to the local conditions of tree growth. These wider annual rings suppose a different proportion of early and latewood (Filipovici 1964; Beldeanu 1999; Beldeanu 2008), which influences the wood density in the area (Nicolotti et al. 2003; Lin et al. 2008; Liang and Fu 2012; Feng et al. 2014; Li et al. 2014) and, finally, the speed of sound propagation (Sandoz and Lorin 1996; Wang et al. 2007; Leboucher 2014).

Fig. 8. The influence of wood density from the center of the trunk on the sound speeds

Fig. 9. Speeds of sound propagation through a portion with wide annual rings

The comparative analysis of the tomograms with the visual appearance after cutting a slice with a chain-saw showed that some small defects located inside the stem were not evident in the tomogram (Figs. 10 through 14), which is consistent with the results of Martinis et al. (2004), who stated that gaps of 1 to 2 cm in diameter are difficult to detect through an acoustic method. This is somewhat supported, on the one hand, by the small size of the defects, but also by the fact that, due to their size, these defects can be situated in between the paths of speed propagation or can be traversed only in one direction by the waves (Proto et al. 2020), which does not significantly influence the propagation speeds constructed by the tomogram. In this regard, Wang et al. (2007) point out that an unidirectional wave can only detect inner rot if it occupies more than 20% of the total area covered by that wave.