EFFECT OF VARIOUS FACTORS ON BUNKER SILAGE DENSITY

Previous models to predict density in bunker silos were developed from empirical data and a limited number of variables such as crop moisture, packing tractor weight, time of compaction, and layer thickness. To develop a more general relationship between bunker silo density and these and other variables, a laboratory apparatus was developed to simulate pressure, time of compaction and layer thickness as applied in a bunker silo. Chopped alfalfa, grass and wholeplant corn were placed in layers of 0.15, 0.30, 0.45 and 0.60 m in a rectangular container 482 mm by 584 mm simulating the foot print of a tractor tire. Pressure between 20 and 80 kPa was applied to the forage by a platen. The most frequently used pressure of 40 kPa corresponded to the weight of a 4600 kg tractor spread over four tires. The total time of compaction varied between 1 and 10 s; the most frequently used time of 5 s was equivalent to two tires passing four times at a speed of 3.4 km/h. A total of 23 tests were conducted in alfalfa and grass, and 25 tests in whole-plant corn. Only results of the hay crops are presented here. The pre-compressed density of the first layer (0.30 m high) averaged 72 and 55 kg DM/m3 for alfalfa and grass, respectively. The highest compressed density ranged between 138 and 339 kg DM/m3 with an average of 220 kg DM/m3. After releasing pressure, the relaxed density of the first uppermost layer ranged between 81 and 152 kg DM/m3 with an average of 127 kg DM/m3. After 6 layers, the average relaxed density became 181 kg DM/m3. As successive layers were added, the cumulative relaxed DM density increased according to a logarithmic model. Within the experimental range, parameters of the logarithmic model were not significantly affected by layer thickness or time of compaction even though short compaction times and thick layers reduced density in trials. These parameters were significantly affected by pressure, dry matter content, crop species and chop length. More laboratory data are needed to understand interactions between the variables while field validation is necessary to extrapolate results for deep bunker silos.