Inside a small sawmill: where a manual bandsaw operator's time, hearing and back go

Introduction

Small-scale sawmills are the backbone of wood processing in many rural regions of Europe. They convert local logs into construction timber and boards with simple equipment, often a manually driven horizontal bandsaw, and they rely heavily on the physical effort of one or two workers. Despite their economic importance, such operations are rarely studied in detail. Mill managers know roughly how many cubic metres they produce, but not how working time is divided between productive and non-productive tasks, how productivity depends on log size, or how much strain the work places on the operator.

A 2021 case study in the journal Forests, by Stelian Alexandru Borz, Maryam Oghnoum, Marina Viorela Marcu, Arpad Lorincz and Andrea Rosario Proto, set out to fill part of this gap. The team monitored a small Romanian sawmill that runs a manually driven bandsaw and measured four things together: time consumption by work element, productivity and recovery, noise exposure, and working posture. The combination is useful because it shows what a typical day of such work costs in both output and operator health.

Research methodology

Site and equipment. The study was carried out in a hangar at a small company in Capalnita, Harghita County, Romania. The firm was founded in 1993 and processes mainly Norway spruce and silver fir. Three people work there, but usually one worker operates the bandsaw. The machine is a horizontal bandsaw with a 10 m rolling rail, a maximum sawing length of about 8.5 m and a maximum sawing diameter of 74 cm. It has a 7.5 kW motor and water-cooled blades (4585 mm long, 38 mm wide, 1.1 mm thick, 2.2 mm pitch). The work is largely manual: logs are moved, loaded and clamped by hand, the sawing height is adjusted, the blade is fed through the log, products are removed and rotated, and the frame is retracted after each cut.

Sample. Data were collected over seven operating days. In that time the company processed 86 logs, about 26 m3 over bark, of which 73 were spruce and 13 were fir. Daily log counts were 11, 14, 11, 11, 13, 14 and 12. Average log diameters ranged from 12 to 56.5 cm, lengths from 200 to 700 cm and volumes from 0.057 to 1.516 m3.

Time study. Each day was video-recorded. Logs were measured and labelled with paint so that they could be identified later. The elemental tasks were identified from the video, and times were recorded in seconds using snap-back chronometry. From these data the authors derived delay-free and total cycle times, gross and net productivity rates, gross and net efficiency rates and the recovery rate (the share of log volume that becomes lumber).

Noise. A Class 2 noise datalogger was worn on the operator's helmet and logged about 186,000 observations at 1 Hz. One of the seven days was lost to a battery failure, leaving six days of noise data. Results were normalised to an 8-hour working day (LEX,8h) following European practice. The authors note that a circular saw was rarely used, so the readings mainly reflect the bandsaw, though other noise sources cannot be fully separated.

Posture. The Ovako Working Posture Analysis System (OWAS) was applied to frames taken from the videos. Three days were selected, and frames were sampled with pseudorandom numbers from 28, 25 and 29 media files of 20 minutes each. From 8,200 extracted images, 2,999 valid observations were analysed. Force exertion was assumed to be low and was not measured. The largest margin of error at 95% confidence was 1.73% (for action category 1).

Statistics. Normality was tested with the Shapiro-Wilk test. Correlation analysis (threshold 0.5) guided the choice of variables, so highly correlated predictors were not used together. Simple regression models (linear, polynomial, exponential, power) were fitted, with the best model chosen by R-squared. Significance was set at p of 0.05 or less, and the analysis used Excel with the Real Statistics add-in.

Key findings

Output. Over the study period the mill produced 695 pieces of lumber, about 18 m3 of product, together with about 8 m3 of residues. The recovery rate varied from 38.8% to 95% and averaged around 69%. Net productivity was about 0.5 m3 per hour and gross productivity about 0.4 m3 per hour, equivalent to efficiency of 2.045 h/m3 (net) and 2.539 h/m3 (gross). Total observed time was about 51 hours. The shares of delay-free time quoted in the paper's different sections (89.2% in one place, about 65% in another) are not easily reconciled in the text we reviewed, so readers should check the definitions in the original before reusing either number.

Where the time goes. For the delay-free work cycle, the approximate shares were:

Effective sawing is therefore only about a quarter of the productive cycle; the rest is handling, setting up and repositioning.

Models. The regression models were statistically significant for blade adjustment, effective sawing, saw return, unloading, and loading and fixing lumber. Typical fits (R-squared) were 0.76 to 0.89 for blade adjustment, 0.89 to 0.94 for effective sawing, 0.71 to 0.75 for unloading lumber and 0.73 to 0.81 for loading and fixing lumber. Log loading was predicted poorly from volume (R-squared 0.49). Delay-free cycle time was explained well by log diameter (83%), volume (88%) and number of pieces (87%). Productivity itself was only moderately explained by log size (about 47% by diameter with an exponential function and 44% by volume with a polynomial), with a hint that productivity may peak at some volume and then decline. The number of pieces obtained rose linearly with diameter and polynomially with volume, and the production-to-residue ratio stayed below 4 in 95% of cases.

Noise. The equivalent level during the observed period was 83.52 dB(A). Normalised to 8 hours, the exposure level (LEX,8h) was 91.62 dB(A), which the abstract rounds to about 92 dB(A). Every day exceeded the 80 dB(A) lower action level. According to the paper's discussion, the normalised level also exceeds the 87 dB(A) exposure limit value, while individual daily levels stayed below it. The authors point out that a good part of the observed time was delays, so higher productive time would probably mean higher noise exposure.

Posture. In the OWAS analysis the back was in the worst posture code (code four) about half of the time. The arms were mostly in the unproblematic code one, whereas legs fell into codes three to six in almost 70% of observations. Action category 1 dominated, followed by categories 2 and 4. The postural risk index was about 224, which the authors interpret as indicating a need for corrective action. The poorest postures occurred during manual log handling (moving, loading, fixing, rotating and removing logs) and when unloading lumber, tasks that the paper estimates at somewhere above 20% to about 30% of workplace time, depending on the section. The authors did not break the OWAS results down by task in the text.

Context. The paper compares recovery with other studies, noting a 58% log conversion ratio in 52 small sawmills in Nigeria and efficiency estimates of 61 to 66% reported elsewhere, so the 69% average recovery looks acceptable. It also recalls that sawing noise is commonly reported at 80 to 120 dB, and mentions a study in British Columbia with average exposures of 92.0 dB(A).

Practical applications for sawmills

From the paper.

• Handling dominates the cycle. Effective sawing is about 27% of delay-free time, while preparation, fixing, loading and unloading together take most of the rest.
• The authors recommend adding mechanical log-handling devices to the bandsaw, modernising machinery to reduce manual handling, reducing noise at the source and maintaining blades and equipment.
• Hearing protection should continue to be used even if noise controls are introduced, and personnel training is described as essential.
• Cycle times can be predicted from log diameter or volume with good accuracy, which can support scheduling and costing.

Portal interpretation (not claims of the paper).

• Because handling dominates, a modest investment in a log turner, loader or lumber-unloading aid is likely to improve both productivity and posture more than a faster saw would. A mill can test this by timing its own cycle with a smartphone video, as the study did.
• The models offer a rough way to quote processing time per log. They were built on one machine and one operator, so they should be recalibrated locally.
• A mill with a similar set-up should measure noise at the operator position, since the reported normalised value is above the 87 dB(A) limit cited in the paper.
• Sorting logs by size before sawing could smooth the workflow, given the strong dependence of cycle time on diameter and volume.

Limitations

• Single case study. One mill and effectively one operator, with mostly softwood logs (spruce and fir). Results cannot be generalised to hardwood, to other machines or to different skill levels.
• Force not measured. Force exertion was assumed to be low (below 10 kg), the most favourable case, which likely understates the load of log handling and the real postural risk.
• Posture sample. Posture was analysed on three selected days and 2,999 valid observations out of 8,200 frames; the text does not give OWAS results by task.
• Noise data. One day of noise data was lost. The effect of log size on noise was not evaluated, nor were softwood-hardwood differences. Noise from other sources, such as the rarely used circular saw, could not be fully separated.
• Ambiguities in the paper. Some figures appear in more than one form (for example, the share of time spent in poor-posture tasks, and the delay-free share of time, where we saw 89.2% and about 65%). We report them as given and suggest checking the original.
• Missing details. The daily noise table was not available to us in the text we reviewed.
• Model fit. Productivity models explained less than half of the variation (about 44 to 47%), so log size alone is not a good predictor of hourly output.

Conclusions

The study shows that a small bandsaw operation can achieve acceptable recovery (about 69% on average) with limited waste, but at a human cost: normalised noise exposure of about 92 dB(A) and awkward postures during log and lumber handling. Time consumption can be modelled well from log size, which is useful for planning, while productivity is influenced by other factors that the study did not capture. The authors' message is that improving safety and efficiency in small mills requires modernised equipment that reduces manual handling and noise, together with training and hearing protection. As a single-site, single-operator case, it is best treated as a well-documented example and a method template for other small mills.

Original research

• Authors: Stelian Alexandru Borz, Maryam Oghnoum, Marina Viorela Marcu, Arpad Lorincz, Andrea Rosario Proto
• Title: Performance of Small-Scale Sawmilling Operations: A Case Study on Time Consumption, Productivity and Main Ergonomics for a Manually Driven Bandsaw
• Journal: Forests (MDPI), volume 12, issue 6, article 810
• Year: 2021
• DOI: 10.3390/f12060810
• Link: https://www.mdpi.com/1999-4907/12/6/810
• Licence: Creative Commons Attribution (CC BY); the MDPI page metadata lists CC BY 3.0

This article is an independent summary prepared by the portal; the original paper should be consulted for full details and exact figures.