Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing Productivity and Cost of Processing and Top-skidding Long Logs * Bruce McMorland Senior Researcher, Harvesting Operations, Forest Engineering Research Institute of Canada Vancouver, BC Email: bruce-m@vcr.feric.ca Abstract The Forest Engineering Research Institute of Canada (FERIC) measured the productivity of two harvesters and a rubber tired grapple skidder operating in the southeastern Interior of British Columbia. The harvesters produced long logs at the stump and the skidder top-skidded the logs to roadside landings. FERIC assessed this long-log/top-skid harvest technique and compared its costs with those for tree-length/butt-skid systems that are more typically used. Keywords: costs, productivity, harvesters, grapple skidders, top skidding, machine interaction 1. BACKGROUND Tembec Industries Inc. in Cranbrook, BC operates several sawmills in the area as well as a pulp mill at Skookumchuck. Tembec s sawmills require long logs as their primary fibre source, and utilize short logs cut to a maximum 6.25 m (20 6 ) to supplement mill production. Much of Tembec s wood supply is located on lower slopes of the Rocky Mountains, and frequently on slopes with grades from 20 to 60%. Historically, Tembec s contractors have dealt with the narrow valleys and steep slopes by using small landings often oversized road turnouts built as part of the road system. Tembec has employed various systems on these sites including feller-bunchers with steep-slope capability, grapple skidders operating on skid trail networks, and processors stationed at roadside or on the small landings. To try and improve overall production and reduce costs, Tembec has experimented with a variety of equipment configurations. On some sites, cut-to-length harvesters sometimes produced long logs that would then be handled by grapple skidders to roadside, and sometimes the harvester produced short logs that would then be handled by forwarders. In 2002, one trial involved grapple skidders to transport stems from a feller buncher to a dangle-head processor located on bladed skid trails. This system produced both long and short logs, which were then re-skidded downhill to landings for loading by a front-end loader. Tembec tested top-skidding and butt-skidding techniques with several of the equipment configurations. Based on these experiences, Tembec staff drew several conclusions: Feller-processing long lengths was more productive than producing short lengths, since fewer cuts were made. Skidding was more productive than forwarding. Top skidding of long logs was more productive than butt-skidding because the skidder could collect a larger turn volume. * The 29 th Council on Forest Engineering Conference. Coeur d Alene, Idaho, July 30-August 2,. W. Chung and H.S. Han, editors. pp. 107-114.
Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing Removing a dedicated processor from the conventional operation improved safety at the landing because only 3 machines (skidder, loader, truck) interacted instead of four. In 2003, Tembec worked in partnership with one of its contractors to commit to a system that used a stump-area harvester to produce both long and short logs. A grapple skidder would then skid the piles of processed logs, by the top, downhill to the landing. Tembec requested that FERIC assess this long-log/top skid system and compare its performance to that of the treelength/butt skid system. 2. OBJECTIVES FERIC agreed on a project with the following objectives: Quantify harvester productivity for feller-processing long logs and compare the results to studies conducted for the INTERFACE 1 model. Assess length-measurement accuracy of the feller-processors. Quantify rubber-tired grapple skidder productivity when skidding bunches of logs by their tops. Assess machine interaction at the landing for long-log and tree-length systems. Assess and contrast productivity and cost for long-log and tree-length systems. 3. STUDY METHODS FERIC conducted detailed-timing studies to assess productivity of the long-log system during the fall of 2003 and January 2004. The studies were on the felling-processing and skidding phases, and used the procedures FERIC employs to collect data for the INTERFACE model. During January and February 2004, FERIC also performed shift-level monitoring of the long-log contractor and one of Tembec s conventional tree-length contractors. Harvesting machines were instrumented with Servis recorders and the operators provided daily charts and shift reports. The machines were monitored while each contractor harvested two small blocks. Production, obtained from Tembec s weigh scale records, totalled about 5100 m³ for the long-log system and 6600 m³ for the tree-length system. Working times were obtained from the charts and shift reports, and volumes were obtained by totalling the log loads delivered from each block as recorded in Tembec s weigh scale records. 3.1 Equipment and system descriptions The equipment employed in the study is shown in Table 1.
Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing Table 1. Equipment complement for each system Felling/feller-processor equipment Skidders Long-log/top-skid system Caterpillar TK 722 harvester John Deere 2054 harvester each with Waratah HTH 20 head John Deere 648 G III rubber-tired grapple skidder Caterpillar 517 tracked grapple skidder Tree-length/butt-skid system John Deere 753 GL feller-buncher with Gilbert 22 head Case 9020 with Hultdins directional falling head Ranger H67 rubber-tired grapple skidder John Deere JD 650 G tracked line skidder Processor Linkbelt 2800 LX with Denharco 550 DH Loader John Deere 644 G front-end loader None trucks were self loaders 1 INTERFACE is a model developed by FERIC to predict the costs for harvesting various harvest prescriptions using a variety of harvesting equipment. The software contains predefined rates of production as determined by previous and ongoing FERIC studies. The harvesters normally produced four sorts. Products were separated according to whether they were redwood (Douglas fir and western larch) or whitewood (pine and spruce) species, and then further separated into short and long logs. The target length for the short log group was 6.25 m (20 6 ) and shorter lengths in 61 cm (2-foot) multiples were allowed. Long logs were defined as longer than 6.25 m and primarily comprised lengths from 11.4 m to 16.9 m in 61 cm multiples (37 6 and 55 6 in 2-foot multiples.) The four product sorts were stacked in separate piles in the stump area prior to skidding. For the top-skid system, the long-log contractor used two skidders in a two-stage skidding process. First, the tracked Caterpillar 517 operated in the stump area in a pre-skid bunching phase in which the skidder operator first gathered and stacked the harvester log piles into larger bunches (Figure 1). It then forwarded the piles to a position on the primary skid trail. From there, the John Deere 648 top-skidded the large bunches the rest of the way to the landing (Figure 2). Figure 1. Tracked skidder performing Figure 2. Rubber-tired skidder * The 29 th Council on Forest Engineering Conference. Coeur d Alene, Idaho, July 30-August 2,. W. Chung and H.S. Han, editors. pp. 115-125.
Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing pre-skid bunching arriving at landing 4. RESULTS AND DISCUSSION 4.1 Harvester productivity FERIC studied the two harvesters for the long-log contractor in the fall and in the winter. Both operators were experienced. Table 2 shows some of the site conditions and results from the productivity studies. The table shows the block-average tree size from the cruise, but FERIC also measured the gross tree volume at the study sites within the blocks so that these studies could be compared with data from other FERIC studies. Figure 3 shows harvester productivity from studies conducted by FERIC in western Canada for the INTERFACE model. The graph shows results for 37 studies of harvesters that produced short logs (<6.25 m in length) compared to five harvesters that produced long logs (McMorland 2003). The graph includes the four studies shown in Table 2. The long-log group comprised the three studies from Table 2 and two earlier studies on Tembec operations (McMorland 2003). Table 2. Harvester productivity Fall 2003 January 2004 Block 052 Block 152 Block 37 JD 2054 Cat 722 JD 2054 Cat 722 Cruise tree size, m³ (block average) No cruise data 0.57 0.25 0.62 Silvicultural prescription Reserve Douglas-fir, western larch (100 stems/ha) Reserve Douglas-fir, western larch (30 stems/ha) Reserve aspen & Douglas-fir, western larch (max 30 stems/ha) Average slope at study site, % 21 15 38 7 No. of products 1 2 4 4 4 Proportion of long logs, % All short logs 90 60 86 Average gross stem volume in 0.34 0.28 0.51 0.40 study, m³ Average stems / cycle (no.) 1.0 1.0 1.0 1.0 Average logs / stem (no.) 2.2 1.2 1.8 1.6 Average stems / PMH 2 (no.) 102 91 80 70 Average gross volume / PMH, m³ 34.7 25.6 40.8 28.0 1 Products were based on redwood and whitewood species species (Douglas fir and western larch vs. pine and spruce) and then length (short logs vs. long logs). 2 PMH = Productive Machine Hours the amount of time the machine was doing work related to its prime function; excludes delays >10 min. The placement of the trend line for the long-log results suggests that harvester productivity was higher when the machines manufactured long logs. This may be true, but it is also visually apparent that the two sets of data overlap. The five observations for long logs may be too small a sample to represent a different data group.
m 3 / PMH Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing Harvester Productivity Treelength - 11 Long logs - 5 Short logs -37 70.0 60.0 50.0 y = 88.221x 0.6799 R 2 = 0.5274 y = 66.98x 0.7738 R 2 = 0.7996 40.0 30.0 20.0 y = 56.092x 0.6862 R 2 = 0.8218 10.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 m 3 / stem Figure 3. Harvester productivity for tree- length logs, long logs and short logs. Figure 3 includes the results for 11 additional harvesters that produced tree-length logs (stems that were topped at maximum length after being measured). Harvesters in those studies operated similarly to stump-area delimbers, except that they made an accurate length measurement prior to topping. The trendline and data grouping for the tree-length results are markedly higher than the long-log and short-log results. The following conclusions can be drawn from Figure 3. Harvesters that manufactured tree-length logs had higher hourly productivity than harvesters that produced shorter lengths. Harvesters that manufactured long logs (lengths between 12.2 and 18.3 m, or 40- and 60 feet) had slightly higher productivity than harvesters that produced logs less than or equal to 6.25 m long. In the tree-size range between 0.25 and 0.40 m³/stem, hourly productivity can be similar for all three product groups. 4.2 Harvester length measurement accuracy Table 3 summarizes the length assessment. Almost 80% of the logs were manufactured within the length tolerances allowed by Tembec. About 8% of logs were over length by an average of 3 cm, and 13% were under length by an average of nearly 10 cm. The under-length category was strongly influenced by one log that was 61 cm under the minimum and which accounted for almost half of the category length value. If this log was excluded from the assessment, the average under-length measure for the other 13 logs would have been 5.9 cm.
Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing Table 3. Length measurement assessment Short logs (<6.26m) no. measured % average length under/over, cm Within tolerance Under length Over length 15 3 0 83 17 0 23.1 Long logs (>6.25) no. measured % average length under/over, cm All logs no. measured % average length under/over, cm 72 78 87 79 11 12 6.2 14 13 9.8 9 10 3.1 9 8 3.1 4.3 Grapple skidder productivity Examples of large turn volumes are shown in Figures 4 and 5. Turn volumes were consistently between 9.5 and 12.0 m 3 per turn, and averaged 11.4 m 3. The number of pieces per turn ranged from 10 to 53, and averaged 30. The large turn volumes resulted in very high hourly volume production for the rubbertired skidder. Figure 6 shows data from three top-skidding studies and 54 butt-skidding studies conducted by FERIC. A few of the studies were on adverse skids, but most of the machines skidded down favourable slopes. Each data point is the average hourly productivity of one study. The top-skid data point at the 200-m class is from a Tembec contractor studied earlier as part of the INTERFACE program. This contractor also performed top skidding with a 115-kW class machine, but only conducted additional bunching on some of the turns. Figure 4. End view of tops of a large top-skidded turn Figure 5. Large-volume grapple skidder turn The two data points at the 400-m distance class are the top-skidding studies from the John Deere 648 (115-kW size class) described in this report. 2 These two studies show hourly volume production of approximately 70 m 3 per Productive Machine Hour (PMH). These results are higher than the productivities found in other skidder studies at this distance, likely because bunches were made larger by additional bunching The butt-skidding studies were conducted on rubber-tired machines in the 115 kw, 125 kw and 135 kw engine-size classes. Roughly, these groups correspond to 160, 170 and 180 hp
m 3 / PMH Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing engine-size classes. About half of the studies were with 115 kw machines, one-third were with 125 kw machines and the remainder were with the 135 kw class. 180 160 140 120 100 80 60 40 20 0 Rubber-tired grapple skidders Butt skidding - 54 studies Long-log, top skid- 3 studies y = 364.24x -0.3909 R 2 = 0.2644 0 100 200 300 400 500 600 Distance, m Figure 6. Skidder productivity. 2 Other common models in the in the 115-kW size-class include the Timberjack 460D, Ranger 667, and Cat 525B. Top skidding is often considered inappropriate because of unacceptable levels of wood breakage. FERIC did not observe any breakage as a result of top skidding during these studies. Figures 2, 4 and 5 show that tops of top-skidded bunched logs are packed together and tightly grasped. The logs tend to support each other, resulting in turns that are less susceptible to bending than turns that are top-skidded by the comparatively flimsy stem leader. 4.4 Machine interaction at roadside for long-log and tree-length systems Table 4 shows the results of the machine interaction studies. The table lists the percentage of time that different numbers of machines were active on the landing. FERIC studied a tree-length contractor in this assessment (System B in Table 4) in order to include a conventional contractor with a front-end loader. Observations were recorded at 5-minute intervals for one shift at each operation, starting around 7:00 am and finishing between 3:30 and 4:00 pm.
Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing Table 4. Percentage of time machine interaction occurred Equipment type Long-log / top skid 2 skidders 1 front-end loader System A 2 skidders 1 processor Tree-length / butt skid System B 2 skidders ½ time front-end loader ½ time processor Landing empty, % 2 0 2 1 harvest machine, % 74 57 65 2 harvest machines, % 22 29 31 3 harvest machines, % 2 14 2 Trucks present on landing, % 58 20 33 The long-log / top skid system resulted in less harvest machine interaction at the landing. Only one harvest machine was active at the landing for 74% of the time, compared to 57% and 65% of the time for the tree-length systems. Although reducing the amount of interaction does not guarantee that there will be fewer safety incidents, certainly the risk of an incident is reduced. 4.5 Productivity and cost for long-log and tree-length systems Productivity FERIC collected shift-level data at the long-log and tree-length contractors operations as each harvested two small blocks. Table 5 presents the block descriptions. Table 5. Block descriptions for winter blocks Harvest system Prescription Gross volume (m3/ha) Long-log / top-skid Block 152 Block 037 Tree-length / butt skid Block 12 Block 13 Reserve D-fir/larch (30 stems/ha) Reserve aspen, D-fir (max 30 stems/ha) Reserve larch >25 cm (10 stems/ha) Reserve non-lodgepole pine >18 cm 303 372 288 279 Gross merchantable tree size, m3 0.57 0.62 0.21 0.32 Species Distribution Lodgepole Western Douglas- Spruce pine % larch % fir % % 77 80 92 92 2 7 6 6 15 3 2 2 6 10 The long-log blocks had a greater volume/ha and larger average stem size than the treelength areas. The topography was generally similar in that the harvest blocks were on 20-30% slopes, but the treelength block had some trails with 55-m segments of 15% adverse grades. Occasional short pitches of adverse skidding in the long-log blocks ranged from 8% to 27% and accounted for about 5% of skid distance. Average skid distance was approximately 170 m at the tree-length operation and 380 m in the long-log blocks. Costs The two harvesters used by the contractor in the long-log block did not have identical scheduling or productivity. The on-board computers could not be used to apportion production between the machines because one of the computers was not operating correctly.
Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing For costing purposes, the harvesters results have been combined to derive a weighted average hourly cost. Table 6 shows the production and working hours for both harvesters, their hourly costs, and the weighted combined cost. Table 6. Weighted hourly cost for harvesters in long-log blocks a Harvester 1 Caterpillar TK 722 Harvester 2 John Deere 2054 Combined Weighted Average, $ / SMH Number of shifts 9.5 8.5 18.0 Productive machine hours (PMH) 91.8 80.4 172.2 Delay, h 10.8 8.9 19.7 Scheduled machine hours (SMH) 102.67 89.25 191.9 Machine cost, $ / SMH 197.33 160.64 Cost, $ 20,259 14,337 34,596 180.27 a Note that long-log is a descriptor used to differentiate from tree-length operations. The long-log harvesting system produces a small percentage of short logs. Table 7 shows the cost comparison between the two systems. The weighted hourly volume for the harvesters was 26.4 m 3 /SMH, meaning that each harvester produced roughly at that rate. The long-log/top-skid combination was estimated to be $2.92/m 3 less expensive to employ than the more conventional tree-length/butt skid system. Most of the savings for the long-log system occurred because the harvester had a onestep-to-product cost of $6.82/m 3 compared to the two-step-to-product cost of $10.53/m 3 for the buncher/processor system. Some of the savings is a result of productivity differences between the two systems due to differences in average stem size (larger at the long-log operation). However, the skidding phase cost for the long-log system was also lower even though two skidding machines were utilized and skid distances were considerably longer. The lower skidding cost was the result of the large top-skidded turn volumes. As noted, the long-log/top skid operation had a larger average stem size compared to the butt-skid operation. A smaller stem size probably would not alter direct skidding productivity at this operation turn volumes would be unlikely to change because turns are composed of logs, not stems but the pre-skid cost may increase to reflect more accumulation time necessary for large turn volumes. The front-end loader was not instrumented with a recording device so net productive time was not measured. The loader was required on site earlier each day than the other harvesting machines in order to load early trucks, and it stayed until the end of the working day with the other machines to organize the landing and log decks for the next morning. This resulted in more scheduled time/day for the loader than for the other machines. Table 7. Productivity and cost comparison Long-log/top-skid b Harvesters Pre-skid Skidder Shifts (no) 18 8 8 Productive Machine Hours (PMH) 172.2 79.4 76.5 Delays (h) 19.7 8.6 10.5. Scheduled Machine Hours (SMH) 191.9 88.0 87.0 Utilization (%) 90 90 88 Machine Cost a ($/SMH) 180.27 140.09 113.04 Volume (m3/sm H) 26.4 57.7 58.3 Cost ($/m3 ) 6.82 2.43 1.94
Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing Front-end loader Total 8 104.0 89.00 48.8 1.82 13.01 Tree-length/butt-skid c Buncher Skidder Processor Miscellaneous d Total 24 32 28 180.6 261.7 203.9 28.3 24.8 58.0 208.9 286.5 261.9 43.1 86 91 78 146.06 112.82 150.34 31.8 23.2 25.4 4.60 4.87 5.92 0.54 15.94 a These costs are based on FERIC s standard costing methodology for determining machine and operating costs. These costs do not include supervision, profit, or overhead and are not the actual costs incurred by the contractor or company. b Total volume produced was 5074 m3 c Total volume produced was 6640 m3 d Occasional tracked cable skidder and feller-director 5. CONCLUSIONS / IMPLEMENTATION Harvesters that manufactured tree-length logs had higher hourly productivity than harvesters that produced shorter lengths. Harvesters that manufactured long logs between 12.2 and 18.3 m (40 and 60 feet) had slightly higher productivity than harvesters that produced logs less than or equal to 6.25 m long. In the tree-size range between 0.25 and 0.40 m³/stem, hourly productivity can be similar for all three product groups. Most of the logs were manufactured within the length tolerances allowed by Tembec. Skidder turn volumes when top-skidded were consistently between 9.5 and 12.0 m 3 per turn because bunches were made larger by additional gathering and bunching at the stump area. This resulted in a very high hourly volume production of approximately 70 m 3 /PMH at an average skid distance of about 400 m. These results were higher than those found in other skidder studies, regardless of machine size. FERIC did not observe any breakage as a result of top skidding during these studies. The long-log / top skid system is potentially safer because it resulted in less harvest machine interaction at the landing. As well, the long-log/top-skid combination was estimated to be $2.92 less expensive per m 3 to employ than the more conventional tree-length/butt skid system. Most of the savings for the long-log system occurred because the harvester had a onestep-to-product cost of $6.82/m 3 compared to the two-step-to-product cost of $10.53/m 3 for the buncher/processor system. Skidding cost for the long-log system was also lower, even though two skidding machines were utilized and the skid distances were considerably longer. The lower cost was the result of the large top-skidded turn volumes. Although outside the scope of this study, there are additional benefits from using a longlog/top skid system: reduced regeneration costs / higher survival rates (increased moisture retention and less grass at the stump area) savings in landing construction reduction or elimination of slash disposal safety hazard avoidance / reduced cost of accidents
Council on Forest Engineering (COFE) Conference Proceedings: Working Globally Sharing 6. ACKNOWLEDGEMENTS The author thanks the staff at Tembec Industries Inc., Norm Roberts Logging and Crabbe Logging for their help in setting up and conducting this study. In addition, thanks to FERIC staff Joanne Lennerton and Pat Forrester for help with field work and analysis, and Marv Clark, Tony Sauder, Yvonne Chu and Shelley Kerr for valuable input to the report. 7. LITERATURE CITED McMorland, B. 2003. INTERFACE productivity data from western Canada: harvesters II. FERIC, Vancouver, BC. Progress Report No. 8. 6 pp.