Sensing miscanthus swath volume for maximizing baler throughput rate

The cost of harvesting is one of the constraints in improving biofuel economics. Adjusting the ground speed of harvesting machines to maximize their throughput rate in relation to yield level is one of the ways to reduce harvesting cost. To control the ground speed of a baler, a real-time sensor that can predict the swathed-biomass yield in front of the baler would be an invaluable tool. It was hypothesized that the swathed-biomass volume is a reliable predictor of the swathed-biomass yield. A light detection and ranging (LIDAR) based sensing system was developed to sense swathed-biomass volume and was field tested. To correlate the sensed swathed-biomass volume with the Miscanthus yield, a new bale-specific method was developed. A linear correlation (R² = 0.76) was found between Miscanthus yield and sensed swathed-biomass volume at the time of mowing-conditioning. Yield maps were generated using the developed correlation model. The model was also used to predict the harvested plot biomass. The average percent error in predicting the harvested plot biomass was 12.4% for training plots and 10.1% for validation plots. The analysis revealed that a proper control strategy to maximize the baler throughput rate would be to either operate the large square baler at the maximum feasible ground speed (6 km h^-1) or at the maximum achievable throughput rate (35 Mg h^-1). For the specific case analyzed, the proposed control strategy would result in the following advantages: 38.0% higher field capacity, 35.6% higher throughput rate, and 30.4% lower baling cost. Further studies and field evaluation of the proposed control strategy are needed. Studies are also needed to extend the swathed-biomass volume sensing concept to corn stover, other energy grasses, and forage crops.