A yield mapping system for sugar cane chopper harvesters

[Abstract]: Yield maps provide essential information for the spatial analysis and evaluation of crop production management at a within field level. Technology has been developed to conduct yield mapping in various crops including grain, potatoes and forage, but as yet no technology exists for yield mapping sugar cane. The chopper harvester is the most common form of mechanical harvester for sugar cane. Therefore, the goal of this research is to develop a yield mapping system for the chopper type sugar cane harvester. After a review, it is proposed that a suitable accuracy goal for the sugar cane mass flow sensor would be ‘less than 5% cumulative measurement error, 95% of the time (2 standard deviations), measured over a 100m2 harvest area’. Existing mass flow sensors for other crops are reviewed. Based on this review four potential techniques are proposed to measure the mass flow rate of sugar cane. These were defined as the chopper power, elevator power and feed roller separation and weigh pad. These were tested simultaneously by placing various sensors on a single harvester and comparing the sensor outputs with the mass flow rate as measured by a weigh truck. In this trial, all techniques offered potential but none produced results close to the accuracy goal. A weighing technique, known as the ‘weigh pad’, offered the most potential for improvement and potential to accurately measure the mass flow rate with a single calibration under all conditions. The weigh pad technique suffered from very small load cell sensitivity to flow rate, drift in baseline readings and susceptibility to mechanical noise/acceleration dynamics. An opportunity arose to install a complete yield mapping system on a harvester within a commercial operation. This opportunity was accepted to assess the potential for applying yield maps to the agronomic management of sugar cane. Because the weigh pad sensor required further development at this stage, chopper and elevator power were used as a measure of mass flow rate. A full yield mapping system was developed. Yield mapping, directed soil sampling and variable rate gypsum application was conducted on a case study field. Economic analysis shows a clear economic benefit when compared with standard management. Analysis is conducted on the weigh pad sensor examining its susceptibility to mechanical noise/acceleration dynamics. Theory is developed to mathematically model the effects of acceleration dynamics on the accuracy of weigh pad sensor. Laboratory bench testing supported the mathematical model. From the theoretical and experimental analysis a number of conclusions are drawn: · The weigh pad should be made as light as possible to minimise the error due to dynamic conditions. · Electronic analogue filters should be used to reduce the noise due to external acceleration. · The weigh pad should be as rigid as possible to maximise its natural frequency. A new weigh pad sensor was designed based on these conclusions. Field trials indicated the effects of external accelerations dynamics were significantly reduced. Baseline drift was then found as the next major factor limiting accuracy. The baseline drift was principally caused by the secondary extractor fan of the harvester inducing a negative pressure on the weigh pad. A rubber curtain placed between the weigh pad and the secondary extractor fan reduced the negative force on the weigh pad due to the secondary extractor fan by 74% (from 17 N to 4.4 N). Therefore it is recommended the curtain be used to minimise the impact of the secondary extractor fan on the baseline drift of the weigh pad. A yield mapping system has been developed for the sugar cane chopper harvester incorporating the weigh pad sensor, a ground speed sensor, a DGPS receiver, a yield display/monitor and data logger. Three identical systems have been constructed and installed on three harvesters for the 1998 cane harvest season. The results show sugar cane could be yield mapped using standard yield mapping principles. The level of accuracy being achieved by the yield mapping system is less than 16% error, with 95% confidence, over a measurement area of approximately 1400 m2. Although the accuracy achieved is not to the desired research goal, yield maps were produced with satisfactory detail to make agronomic management decisions. The reliability of the sugar cane yield mapping system under field condition in a commercial operation was satisfactory. However, two techniques are proposed (“auto-zeroing” and “batch weighing” techniques) to improve the accuracy and reliability of the weigh pad readings during wet or adverse harvesting conditions. After note: At the time of writing the NCEA along with Case Austoft (CNH) were continuing to conduct research and development on the system and are intending to make the yield mapping system available as a standard item on new harvesters and a retrofit unit on existing harvesters in the near future (C. Barret, per. comm. 2001). The proposed “autozeroing” and “batch weighing” techniques are being tested.