Investigation of the dynamic behavior of soybean plants during cutting

Shatter losses of soybeans are still a substantial percentage of potential profits despite extensive research of the problem. Previous efforts have tended to be concerned with the design and operating aspects of harvesting machinery rather than the dynamic characteristics of the plant. In contrast, this study focused on the dynamic behavior of soybean plants during cutting. Mathematical models were developed for the dynamics of the components of the soybean plant: the stalk and pod. The Euler-Bernoulli beam appeared to be a satisfactory model for the stalk, predicting modes of vibration with corresponding natural frequencies. The simple pendulum with a torsional spring was adequate for modeling some aspects of pod motion. The coupled stalk-single pod model showed the effect of the pod on the motion of the stalk to be due to: the added mass of the pod, the motion of the pod, and collisions between the pod and stalk. The model also predicted that the pod and stalk tended to move out of phase for vibration at most frequencies, and that the response of the pod to stalk vibration was frequency dependent, the most significant pod response occurring at lower frequencies. The notion of a cutting function was introduced to represent the aspects of cutting that cause motion of the plant. For the type of cutting device used in this study, a multi-tooth circular saw blade, a sequence of pulses appeared to be an adequate mathematical model cutting function. Experimental determination of cutting functions using stalks with pods attached was hindered by the nonlinearities caused by friction between pods and collisions between pods and the stalk. Data collection was accomplished with accelerometers and an analog-to-digital data acquisition system. The presence of the accelerometers and cables affected the response of the plant and the instrumentation would very likely be damaged by severe plant motion; therefore, another method of sensing plant response would be more appropriate. Digital signal analysis using Fast Fourier Transform methods proved to be an effective method of data analysis. The application of this technology to the study of crop dynamics appears as promising as it has been for other areas of vibration research.