During different growth periods,canopy size and density in orchards are variable,which need application conditions(flow rate and air flow)to be adjusted to match the canopy’s characteristics.In order to improve orcha...During different growth periods,canopy size and density in orchards are variable,which need application conditions(flow rate and air flow)to be adjusted to match the canopy’s characteristics.In order to improve orchard sprayer’s automatic operating performance,an automatic variable-rate orchard sprayer(VARS)fixed with 40 electromagnetic valves and 8 brushless fans was developed based on the canopy’s spatial dimensions.Each solenoid valve and brushless motor can be individually adjusted in real-time through pulse width modulation(PWM)signals emitted by a control system to adjust each nozzle’s spout and fan rotation speed.A high-precision laser scanning sensor(light detecting and ranging,LIDAR)was adopted as the detector to measure the canopy volume using the variable rate algorithm principle.Field experiments were conducted in an apple orchard,and conventional air blast sprayer(CABS)and directed air-jet sprayer(DAJS)were tested as a comparison.Results showed that on average,46%less spraying solution was applied compared to conventional applications,while penetration rate was similar to DAJS.Normalized deposition in the canopy with variable application was higher than that of conventional applications,indicating that electronic sprayers are more efficient than conventional sprayers.It was also observed that VARS could significantly reduce off-target loss.The field experiment showed that the newly developed variable-rate sprayer can greatly reduce pesticide use and protect the environment for the orchard fruit production,and also provide a reference for design and performance optimization for plant protection machinery.展开更多
将减法聚类与伪逆法相结合建立了模糊RBF(Radial Basis Function neural network,径向基神经网络)模型。通过正交试验,获得不同激光功率、扫描速度、预热温度和切片厚度参数条件下SLS试件的尺寸误差,在此基础上获得训练与预测样本数据,...将减法聚类与伪逆法相结合建立了模糊RBF(Radial Basis Function neural network,径向基神经网络)模型。通过正交试验,获得不同激光功率、扫描速度、预热温度和切片厚度参数条件下SLS试件的尺寸误差,在此基础上获得训练与预测样本数据,并对该模型进行仿真。结果显示预测平均总误差为2.16%,表明该模型具有建模快、模型简单、训练速度快、预测精度高,泛化能力强的优点,可根据不同烧结工艺参数对SLS制件尺寸精度进行较准确地预测,以便指导实际生产。展开更多
基金The authors acknowledge that this work was financially supported by Special Fund for Agro-scientific Research in Public Interest(No.201503130)Beijing Science and technology plan projects(No.D171100002317003)National Natural Science Foundation of China(31470099).
文摘During different growth periods,canopy size and density in orchards are variable,which need application conditions(flow rate and air flow)to be adjusted to match the canopy’s characteristics.In order to improve orchard sprayer’s automatic operating performance,an automatic variable-rate orchard sprayer(VARS)fixed with 40 electromagnetic valves and 8 brushless fans was developed based on the canopy’s spatial dimensions.Each solenoid valve and brushless motor can be individually adjusted in real-time through pulse width modulation(PWM)signals emitted by a control system to adjust each nozzle’s spout and fan rotation speed.A high-precision laser scanning sensor(light detecting and ranging,LIDAR)was adopted as the detector to measure the canopy volume using the variable rate algorithm principle.Field experiments were conducted in an apple orchard,and conventional air blast sprayer(CABS)and directed air-jet sprayer(DAJS)were tested as a comparison.Results showed that on average,46%less spraying solution was applied compared to conventional applications,while penetration rate was similar to DAJS.Normalized deposition in the canopy with variable application was higher than that of conventional applications,indicating that electronic sprayers are more efficient than conventional sprayers.It was also observed that VARS could significantly reduce off-target loss.The field experiment showed that the newly developed variable-rate sprayer can greatly reduce pesticide use and protect the environment for the orchard fruit production,and also provide a reference for design and performance optimization for plant protection machinery.
文摘将减法聚类与伪逆法相结合建立了模糊RBF(Radial Basis Function neural network,径向基神经网络)模型。通过正交试验,获得不同激光功率、扫描速度、预热温度和切片厚度参数条件下SLS试件的尺寸误差,在此基础上获得训练与预测样本数据,并对该模型进行仿真。结果显示预测平均总误差为2.16%,表明该模型具有建模快、模型简单、训练速度快、预测精度高,泛化能力强的优点,可根据不同烧结工艺参数对SLS制件尺寸精度进行较准确地预测,以便指导实际生产。
