Using acoustic emissions to monitor crop throughput of a large square baler

Abstract

This paper demonstrates that acoustic signals emitted by a large square baler during field operation may be analysed to discriminate between high and low levels of machine loading (crop mass-flow rate at constant bale density). Thus there may be potential for utilising acoustic signals in an automatic closed-loop control system designed to optimise machine field performance. Acoustic data and other independent measures of crop throughput were collected from a large square baler operating in three different fields and forming a total of 57 bales. After signal processing, throughput-related parameters were extracted from the acoustic data and subsequently used to discriminate between high and low mass-flow rate operation cycles within a particular field, with accuracy of up to 95%. The physical origin of each sound used in the discrimination was determined. Although the mean values of the parameters for the same throughput level were field (operating condition)-specific, the same crop throughput-related trends were seen for all operational conditions.