*2.2. Research Design*

The study was conducted at the Azienda agraria sperimentale "L. Toniolo" of the Università degli Studi di Padova (Padova, Italy). Since we were interested in controlling as many variables as possible to better compare RULA and REBA, the evaluation of the differences in risk levels between the two methods was carried out by keeping the machine and participant characteristics constant, while changing the log size. To reduce the inter-individual variability in the adoption of the targeted postures, one single participant was involved in the study. He was a not-specialized operator, with no previous history of either occupational accidents or musculoskeletal disorders. Specific measures were defined for the logs to be used in the tests, with regard to log length and diameter, and an acceptable mass, ranging between 2 and 10 kg, to be manually handled by the operator (Figure 2).

**Figure 2.** Example of tasks to be performed during the tests.

Plane-tree wood was used for the tests, preparing both logs of different length and diameters, and branches of more variable dimensions, in order to observe if and how much the size of the wood induced different postures and gestures, which may be associated with different levels of postural risk. The logs were prepared in three different lengths (1 m, 1.5 m, 2 m) and two diameters (65 mm and 135 mm). For each length/diameter combination and for the branches, three elements were prepared to perform three repeated tests. Therefore, six tasks involving logs plus one task regarding branches were performed. The operator performed each task three times to evaluate any intra-individual difference in postures and gestures with respect to the manipulated element. Thus, overall seven tasks were performed for a total of 21 loadings (six tasks for the logs, each repeated three times, and one task for the branch repeated three times). The 21 loadings were randomly performed.

The research protocol was approved by the Research Advisory Group of the Institute for Agricultural and Earthmoving Machines (IMAMOTER) of the National Research Council of Italy (CNR) on 28 February 2018.

### *2.3. Video Recording*

The operator was video recorded and photographed while feeding the machine. One video camera was placed in front of the infeed chute while a second one recorded images from the side of the infeed chute. During the analysis of the recorded videos, every time a log or a branch was loaded, the frame-by-frame vision was carried out until the most critical position was detected. The video analysis was performed with Kinovea software, an open-access video analysis software available

online reproducing the video in slow-motion. Kinovea is a valid and reliable method to perform motion and postural analysis [63]. It allows for detecting body and its districts' angles during the single posture observations: angles were measured according to the reference axes reported in RULA and REBA methods to obtain the corresponding score.

#### *2.4. Assessment Methods of Applied Biomechanical Risk*

The study investigated the human–chipper interaction comparing two observational methods of biomechanical overload postural risk: RULA method [13] and REBA method [14].

In RULA method [13], the body is divided into different parts gathered into two groups, A and B: group A includes arm, forearm and wrist of the right and left limb; group B includes the neck, the trunk and the feet. The method consists in assigning a score to each segment depending on the posture taken and it allows to obtain two distinct scores (Scores A and B) through the use of numerical tables or spreadsheets; these scores represent the level of postural load of the musculoskeletal system, determined by the combination of the postures of the whole body. Muscle use and force scores are then added to Scores A and B to obtain two new scores (Scores C and D) that, through a third table or a spreadsheet, allow for obtaining the final score, or Grand Score.

Based on the appropriate combination of scores, the final score can range between 1 and 7. The final score is related to four levels of action and four levels of risk. Scores, actions and risk levels for RULA are summarized in Table 1:



**Table 1.** RULA and REBA scores with the respective action level.

RULA: Rapid Upper Limb Assessment; REBA: Rapid Entire Body Assessment.

In REBA method [14], the body is divided into different segments divided into two groups: the first one includes neck, torso and legs; the second group is composed by arm, forearm and wrist without distinction from the right or the left one. The method consists in assigning a score to each segment of the body according to the posture taken and, using numerical tables or spreadsheet, it allows for obtaining two different scores that represent the level of postural load of the musculoskeletal system determined from the combination of the whole body postures. The two scores should be respectively added to the grip score and to the load and strength score to get two new scores (Scores A and B). From the use of a third table or spreadsheet, it is possible to obtain Score C, which, added to the activity score, allows for obtaining the final score, or Grand Score. The final score can range between 1 and 15 and it is related to five levels of action and five levels of risk. Table 1 reports scores, actions and risk levels for REBA:

• Action level 0: The risk is negligible so no action is required.


### *2.5. Data Analysis*

For each task, three risk indices were obtained for each of the two methods, RULA and REBA.

The goal was to analyze the highest level of risk calculated, considering the variability of the execution of the gesture, evaluated three times separately. Among the results obtained, the highest value, among the three for each task, and then the highest between the right and left side was selected.

The results of the two scores calculated for each task with RULA and REBA were compared to highlight any tendency of any of the two methods to over or underestimate the risk. This was done both on the basis of risks category, considering the action level—color code (Table 1), and the values of normalized indices.

In order to compare the differences between REBA and RULA scores, it was necessary to normalize the absolute values as the two methods are based on different scales: RULA has four risk levels that categorize scores from 1 to 7, while REBA is based on five levels of risk that categorize scores from 1 to 15. The use of standardized scores to compare different methods for postural risk assessment has been adopted previously in similar analyses [64,65], thus we normalized the score values ranging from 0 to 1, applying the following formula to each worst value obtained for each task by the two methods:

normalized value = [score obs − score value min]/[score value max − score value min].

#### **3. Results**

The risk assessment carried out by applying RULA and REBA (Table 2) showed a medium-high level of risk for all the tasks. None of the tasks reported a score referable to a negligible or low risk as well as a very high risk level that would require an immediate intervention.



<sup>1</sup> Note. Tasks 1–6 represent the loading of logs with three different lengths (1 m, 1.5 m, 2 m) and 2 diameters (65 mm and 135 mm); Task 7 represents the loading of branches.

Regarding the differences between the risk indices of each task, it can be seen that, based on the dimensional differences of the handled logs, the risk level is limited in the case of logs with longer length and smaller diameter. This could be due to the posture induced by small length logs which could cause a higher postural overload because of the greater proximity to the machine that could determine worse joint angles. The highest indices were observed in the manipulation of short and larger diameter logs (Task 2), while the lower levels of risk indices occurred in the manipulation of the branches (Task 7). The comparison between the results obtained with RULA and REBA, based on the resulting action level, highlighted a good overlapping between the two methods, with the exception of Task 3, related to wooden elements of an intermediate size (1.5 m) and small diameter (65 mm). In this case, the RULA carried out a more precautionary assessment as it gave an index that fell into a higher action level compared to the corresponding level calculated with REBA. In the comparison of the normalized numerical indices, all of the RULA scores were higher than the corresponding REBA scores for all of the tasks (Table 3). The application of RULA showed to be more precautionary than REBA for all the tasks, as it returned a higher score value in each loading.

**Table 3.** Comparison between the normalized scores of RULA and REBA indices for the seven tasks.


<sup>1</sup> Note. Tasks 1–6 represent the loading of logs with different lengths and diameters; Task 7 represents the loading of branches.

#### **4. Discussion**

The present study showed that RULA and REBA are two effective methods for assessing the postural overload determined by incongruous postures adopted in the act of manually loading the wood-chipper with logs of various diameters and lengths. The two methods appeared to be both effective and suitable for the identification and quantification of the level of postural risk using a manually fed wood-chipper, as they highlighted similar levels of urgency of intervention, analysis and modification actions. The high congruence between RULA and REBA evaluations confirmed similar results from other investigations in the industrial context [66]. However, the present study showed that RULA tended to be more precautionary, giving a higher risk index, for all the tasks, consistent with previous studies comparing the two methods in the industrial sector [18]. This underestimation by REBA was also found in KOSHA's research [67] in different sectors like ship building, automotive, electronics, general manufacturing, and service industries.

Comparing the action level in which the calculated indices fall, and, therefore, the corresponding severity and urgency of intervention, RULA and REBA suggested an almost identical level and, in only one case (Task 3, intermediate size and small diameter wooden elements), RULA returned a higher action level than REBA (Table 2). Task 3 is also the one which reported the highest difference between normalized RULA and REBA scores. Comparing normalized scores, RULA always resulted in being more precautionary as it returned higher values for all the tasks. The evaluation of the branches' manipulation task presented the most contained difference between the two normalized values, but also in this case the risk estimated by RULA, even if slightly, prevailed. RULA appeared to be constantly more precautionary in the risk assessment for all of the tasks considered, both in medium or high risk conditions. RULA also seemed to be more precise as it better highlighted the differences in the level of risk exposure between the manipulation of logs and branches. This could be due to the presence of different postural issues of the upper limbs, such as radial or ulnar deviation of the wrist, rotation of the wrist and movements performed across the body or out to the side, which play an important role in the postural risk assessment of these tasks and are not considered by REBA. About the branches loading, RULA normalized score returned a value of 0.33, which was the lowest of all tasks. REBA did not seem to be influenced by these differences among logs and branches manipulation. The minimum normalized REBA score was 0.29, and it was the same value for both the manipulation of the branches (Task 7) and for the manipulation of logs of Tasks 3 and 5.

RULA highlighted how biomechanical workload in the interaction with the wood-chipper is limited to the upper part of the body and it determines an important involvement especially of the wrist and forearm, and on the rachis due to flexion of the chest and postural asymmetries. REBA scores, obtained evaluating the lower limb biomechanical risk in greater detail, had no higher values than RULA in all the tasks, confirming the absence of serious criticalities for the lower limbs in wood-chipper manual feeding.

According to the results of the present study—even though they were obtained by observing one single operator and should not therefore be considered as conclusive—RULA is more precautionary, probably because of the more precise workload evaluation of the upper limbs and trunk. RULA considers in a much more limited way the postural and workload conditions of the lower limbs compared to REBA, but, nevertheless, in the present investigation, REBA did not give higher risk estimation values, which means that, during the loading of the wood-chipper, there was not any evidence of risk to the lower limbs (otherwise, REBA would have highlighted it with a higher risk index). Therefore, RULA may be a more adequate method for the assessment of postural risk of interaction with the machinery when exposure to risk of the lower limbs is less relevant. This consideration confirms results from previous studies which reported the effectiveness of RULA in evaluating the interaction with agricultural machines [44].

The analysis pointed out different RULA scores due to the different lengths and diameters of logs, with shorter ones determining incongruous postures. This consideration opens new perspectives for the analysis of human–wood-chipper interaction stressing the importance of avoiding short logs during the cutting operations, when they are intended to be chipped. In this way, a postural disadvantage occurs even if the mass is reduced. Future investigation should be addressed at investigating a larger number of operators regarding how much log mass, besides the length, determines incongruous postures (high RULA score) and physical effort (applied force and perceived fatigue). The results of such studies could provide suggestions about an "optimal" cutting length for logs, determining a better postural conditions for operators of manually fed wood-chippers. Indeed, in ergonomics, it is not enough to concentrate on the design of the machine, but it is important to intervene in the different components of human–machine interaction through a holistic approach, as widely documented in the literature and recommended by international standards [68–71].

Besides the attention paid to safety and technical characteristics of the forestry machine, this study highlighted the importance to inform and train the operators to perform gestures, postures and activities in accordance with the ergonomic principles. In this way, in addition to optimize the human–machine interaction, it can be possible to intervene on the organization and management of the whole activity. This integrated approach may contribute to risk reduction and enhance the system productivity and operator's wellbeing.

#### **5. Conclusions**

Agricultural and forestry operators interact with a variety of tasks and machinery, which can require ad hoc methods for the assessment of postural risk. The present investigation showed that the RULA method is suitable for the evaluation of postural overload in the human–machine interaction related to a manually fed wood-chipper, more than the REBA method, because it showed indices that corresponded to a higher level of risk for all the tasks observed, independently of the shape, size, mass of the wooden material and, therefore, it would be a more precautionary method to protect operator's health. Further considerations about the postural risk of manually-loaded wood-chippers should focus not only on the posture of the operators, but also on the safety standards imposed by the safety regulations (for instance, the former EN 13,525:2005 + A2:2009 [62]). The extended application of RULA in real and differentiated agro-forestry conditions will allow for assessing postural risks and physical effort to improve human–machine interaction and operators' wellbeing in the wood-chipping activity. **Author Contributions:** M.M.C., A.G., F.C., A.C., E.C. and S.G. conceived and designed the study; A.C. and S.G. collected the data; M.M.C. and A.G. analyzed the data; M.M.C., A.G., F.C., A.C. and S.G. interpreted the data; M.M.C., A.G. and F.C. wrote the paper; A.C., E.C. and S.G. critically reviewed the paper; M.M.C., A.G., F.C., A.C., E.C. and S.G. gave final approval.

**Funding:** This study was supported by the "Protection of agricultural machinery operators from crush, entanglement, shearing" (PROMOSIC) project, funded by INAIL, the Italian National Institute for Insurance against Accidents at Work.

**Acknowledgments:** This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sector.

**Conflicts of Interest:** The authors declare no conflict of interest.
