【作者】 张瑜；

【导师】 邱白晶；

【作者基本信息】 江苏大学 ， 农业机械化工程， 2022， 硕士

【摘要】 农用喷嘴的滤网性堵塞是施药过程中常见的喷嘴故障。常用的农用喷雾设备会根据不同的喷雾需求搭载不同类型的喷嘴,且搭载的喷嘴数量较多,目前常用的方法很难进行多喷嘴的故障检测。基于这一点,本研究根据对农用喷嘴的已有研究及市场占有率的调研,选取了具有代表性的农用喷嘴,并提出了一种基于声信号的农用喷嘴滤网堵塞检测方法,探讨利用麦克风阵列的声源识别定位能力来检测双喷嘴中喷嘴的堵塞情况,并通过试验证明了方法的可行性,能够更便捷地检验喷嘴的堵塞情况,为基于声信号检测喷嘴堵塞问题提供有力依据。主要研究结论如下:1、结合农用喷嘴的结构、工作原理和农药属性进行喷嘴不同堵塞类型及产生堵塞的原因原分析,得出喷嘴滤网性堵塞是目前农业喷雾领域中发生频率极高的故障,确定本文研究的堵塞类型为喷嘴的滤网性堵塞。根据实际喷雾作业中农用喷嘴的使用情况,选取了具有足够的代表性的常规扇形喷嘴、空心锥喷嘴和防飘移喷嘴3种类型6种不同型号的农用喷嘴进行堵塞试验。根据堵塞原因分析,确定了喷嘴滤网的不同堵塞模式及不同堵塞程度对喷雾流量影响的模拟试验。对试验结果进行统计分析,结果显示在同一喷雾压力下,喷嘴不同堵塞模式对喷雾流量变化的影响在统计学上无显著差异,而喷嘴堵塞程度对喷嘴流量变化的影响存在显著差异。2、在分析近场阵列声波传递模型的基础上,结合时延估计和波束形成算法,进行阵列布置方式对双声源定位识别能力试验,结果得到了一个恰当的环形麦克风阵列布置方式,达到声源识别定位和提高信号质量的作用。试验确定麦克风阵列的声源识别和定位的合适阵型为三层环形阵列（阵列设置三层麦克风,第一层布置半径为R=36cm,数目M₁=5个;阵列第二层R=29cm,数目M₂=5个;阵列第三层R=22cm,M₃=5个）。通过试验得出了环形麦克风阵列的布置方式对于提高声源定位识别能力的基本准则:（1）麦克风布置半径和布置层数越大,阵列对声源的定位识别能力越强。（2）同一布置半径上,麦克风数目的增加无法提高阵列的声源定位识别能力。将得到的麦克风阵列的布置方案应用于喷嘴声音信号采集并使用波束形成算法进行阵列信号的处理,处理结果通过峰值信噪比的评估方法能够有效的提升目标信号的信噪比,改善声源质量。3、根据喷雾发声的声学传输率变化的可检测性,利用麦克风阵列的声源识别定位能力,提出了基于声信号检测双喷嘴的滤网堵塞及堵塞程度的方法。该方法包括以下几个步骤:（I）信号数据序列的采集;（II）信号预处理;（III）功率谱密度计算;（IV）频带功率计算;（Ⅴ）去除异常值;（Ⅵ）频带功率差异计算;（Ⅶ）频带功率差异求和;（Ⅷ）确定阈值检测喷嘴堵塞。该方法使用6种具有代表性的农用喷嘴和80不锈钢防滴滤网,在0.2-0.4MPa的喷雾压力下,对喷嘴滤网的三种堵塞程度（D1、D2、D3）进行试验。试验结果表明,随着喷雾压力的增大检测方法的检测性能会降低。在喷雾压力为0.2MPa和0.3MPa下,对不同程度堵塞情况下检测的准确率可以达到94%和83%。当喷嘴发生堵塞时,从喷嘴外部测量的声压响应中观察到功率谱密度是呈现下降趋势,而且频带功率差异之和（SAD）随着滤网堵塞程度的增加而减小。同种型号的不同喷嘴对于检测结果的准确性在统计学上没有显著影响。更多还原

【Abstract】 Screen clogging of agricultural nozzles is a common nozzle failure during application.Commonly used agricultural sprayers are equipped with different types of nozzles according to different spray requirements,and there are a large number of nozzles,so it is difficult to detect the malfunction of multiple nozzles by the current commonly used methods.Based on this phenomenon,this study selected the representative agricultural nozzle on the basis of the existing research of agricultural nozzle and the market share of the research,and a farm nozzle filter clogging detection method based on acoustic signal was brought up,using the sound source identification ability of microphone array to detect the nozzle clogging situation in double nozzle,through the experiments the method was proved feasible,it can check nozzle clogging more conveniently and provide strong basis for nozzle clogging detection based on acoustic signal.The main conclusions are as follows:1.Combined with the structure,working principle and pesticide properties of spray nozzle,different types of nozzle clogging and the causes of clogging were analyzed.It was concluded that the filter screen clogging of nozzle was a very frequent fault in the field of agricultural spray at present,and the type of clogging studied in this paper was the filter screen clogging of nozzle.According to the use of agricultural nozzles in actual spraying operations,three types of agricultural nozzles with sufficient representativeness,including conventional fan nozzles,hollow cone nozzles and antidrift nozzles,were selected for clogging tests.According to the analysis of the cause of the clogging,the simulation experiment of the influence of different clogging modes of the nozzle screen and different clogging degrees on the spray flow was determined.Statistical analysis of the test results shows that under the same spray pressure,there is no statistically significant difference in the influence of different nozzle clogging modes on the change of spray flow,while the influence of nozzle clogging degree on the change of nozzle flow is significantly different.2.Based on the analysis of the acoustic transmission model of the near field array,combined with the delay estimation and beamforming algorithm,the identification ability of the array arrangement to the dual sound source is tested.The results show that an appropriate arrangement of the ring microphone array achieves the identification and location of the sound source and improves the signal quality.It is determined that the appropriate formation of microphone array for sound source recognition and location is three-layer ring array(the array is equipped with three-layer microphone,the radius of the first layer is R=36cm,and the number is M1=5;The second layer of array R=29cm,number M2=5;The third layer of array R=22cm,M3=5).The basic criteria for improving the location and identification ability of the sound source by the arrangement of the ring microphone array are obtained:(1)The larger the radius and the number of layers of the microphone array are,the stronger the location and identification ability of the sound source array is.(2)In the same layout radius,the increase of the number of microphones cannot improve the array’s sound source location and recognition ability.The obtained microphone array layout scheme is applied to the nozzle sound signal acquisition and beamforming algorithm is used to process the array signal.The processing results can effectively improve the SNR of target signal and improve the quality of sound source through the evaluation method of peak SNR.3.According to the detectability of acoustic transmission rate variation of spray sound and the ability of sound source identification and location of microphone array,a method is proposed to detect filter clogging of nozzle and clogging degree of double nozzles based on acoustic signals.The method includes the following steps:(I)Acquisition of signal data sequence;(II)Signal preprocessing;(III)Power spectral density calculation;(IV)Frequency band power calculation;(V)Remove outliers;(VI)Calculation of frequency band power difference;(Ⅶ)Sum of frequency band power difference;(VIII)Determine the threshold for detecting nozzle clogging.In this method,6 typical agricultural nozzles and 80 mesh stainless steel anti-drip filter screen were used to test three clogging degrees(D1,D2,D3)of the nozzle filter screen at a spray pressure of 0.2-0.4 MPa.Test results show that the detection performance of the method decreases with the increase of spray pressure.When the spray pressure is 0.2 and 0.3MPa,the detection accuracy of different degree of clogging can reach 94% and 83%.When the nozzle blocks,the power spectral density measured from the sound pressure response outside the nozzle is observed to decrease,and the sum of frequency band power difference(SAD)decreases with the increase of the clogging degree of the filter.Different nozzles of the same type had no statistically significant effect on the accuracy of the test results.更多还原