袁全春,徐麗明,邢潔潔,段壯壯,馬 帥,于暢暢,陳 晨
?
機施有機肥散體顆粒離散元模型參數(shù)標定
袁全春,徐麗明※,邢潔潔,段壯壯,馬 帥,于暢暢,陳 晨
(中國農(nóng)業(yè)大學(xué)工學(xué)院,北京 100083)
針對有機肥離散元模型接觸參數(shù)和接觸模型參數(shù)難以通過查閱文獻或試驗直接獲得的問題,該文提出一種通過仿真試驗建立回歸模型并結(jié)合物理試驗尋優(yōu)的方法,對有機肥離散元模型參數(shù)進行標定。考慮到有機肥顆粒間的凝聚力,選擇“Hertz-Mindlin with Johnson-Kendall-Roberts”接觸模型。應(yīng)用Plackett-Burman Design對有機肥離散元模型參數(shù)進行篩選,得到對休止角有顯著影響的參數(shù),即有機肥-有機肥滾動摩擦系數(shù)、表面能JKR和有機肥-鋼板滾動摩擦系數(shù);通過最陡爬坡試驗確定了顯著性參數(shù)的最優(yōu)值區(qū)間,應(yīng)用Central Composite Design建立并優(yōu)化了休止角與顯著性參數(shù)的回歸模型,以實際休止角為目標,求解得到顯著性參數(shù)最優(yōu)值,即有機肥-有機肥滾動摩擦系數(shù)為0.10,JKR為0.015,有機肥-鋼板滾動摩擦系數(shù)為0.11。在標定的參數(shù)下進行仿真驗證試驗,仿真休止角與實際休止角的相對誤差為0.42%,無明顯差異,表明標定的參數(shù)準確,可以為有機肥離散元模型參數(shù)的選取提供參考。
模型;肥料;離散元;參數(shù)標定;凝聚力;休止角
隨著現(xiàn)代化種植管理技術(shù)的提高,人們對果園施肥機械的要求也越來越高。基于離散元法[1-2]建立的農(nóng)業(yè)散體顆粒模型具有較高的準確性,應(yīng)用離散元法研究有機肥與施肥機械的作用機理,可以提高研發(fā)效率,改善機械作業(yè)性能,為農(nóng)機具的設(shè)計及優(yōu)化提供理論依據(jù)[3-5]。應(yīng)用離散元法需要在EDEM軟件中建立有機肥模型,并定義模型的參數(shù),模型參數(shù)的準確性將直接影響研究結(jié)果的可靠性。模型參數(shù)包括材料本征參數(shù)(粒徑、泊松比、剪切模量和密度)、接觸參數(shù)(碰撞恢復(fù)系數(shù)、靜摩擦系數(shù)和滾動摩擦系數(shù))和接觸模型參數(shù)(Bonding模型的法/切向模量、Johnson-Kendall-Roberts模型的表面能JKR等)。材料本征參數(shù)正常是固定的,通??梢酝ㄟ^查閱文獻和試驗測定得到;而接觸參數(shù)和接觸模型參數(shù)很難直接獲得,通常需要通過仿真試驗進行標定[6]。目前,在農(nóng)業(yè)工程領(lǐng)域,很多學(xué)者在農(nóng)業(yè)散體顆粒的離散元模型參數(shù)標定方面做了大量的工作,但對于有機肥離散元模型參數(shù)標定的研究較少。
王云霞等[7]基于休止角試驗仿真標定了玉米種子間靜摩擦系數(shù)和滾動摩擦系數(shù);劉凡一等[8]標定了小麥的離散元模型參數(shù);石林榕等[9]標定了西北旱區(qū)農(nóng)田土壤離散元模型參數(shù);王憲良等[10]標定了土壤顆粒間靜摩擦系數(shù)和滾動摩擦系數(shù);張銳等[11]通過仿真標定,得出顆粒外觀形貌對顆粒間靜摩擦系數(shù)影響較大的結(jié)論;武濤等[12]基于休止角試驗標定了粘性土壤的離散元模型參數(shù)。
本文擬針對樣品有機肥,應(yīng)用EDEM軟件,選擇“Hertz-Mindlin with Johnson-Kendall-Roberts”凝聚力接觸模型,進行Plackett-Burman Design、最陡爬坡試驗和Central Composite Design等有機肥休止角仿真試驗,并結(jié)合有機肥休止角物理試驗對有機肥離散元模型參數(shù)進行標定。以期獲得較為準確的有機肥離散元模型參數(shù),為有機肥離散元模型參數(shù)的選取提供參考。同時,提供一種通過測定易于測定的參數(shù)(休止角)來推導(dǎo)難以測定的有機肥離散元模型參數(shù)的方法。
本文所用有機肥為蘭州花海鎮(zhèn)紫龍珠葡萄合作社的有機肥,由純羊糞發(fā)酵得到。為滿足施肥均勻性的要求,對結(jié)塊有機肥進行粉碎處理。有機肥的基本參數(shù)如表1所示。
本文采用物理試驗與仿真試驗相結(jié)合的方法[13-14]對有機肥離散元模型參數(shù)進行標定。首先進行物理試驗,采用圓筒提升的方法獲得有機肥顆粒堆,并測量有機肥實際休止角。然后利用軟件EDEM2.6進行仿真試驗,先通過Plackett-Burman Design對有機肥離散元模型參數(shù)進行篩選,得到對休止角有顯著影響的參數(shù);再通過最陡爬坡試驗確定顯著性參數(shù)的最優(yōu)值區(qū)間,通過Central Composite Design響應(yīng)面分析方法建立并優(yōu)化有機肥休止角與顯著性參數(shù)的回歸模型,得到回歸方程,以實際休止角為目標值對回歸方程求解尋優(yōu),得到顯著性參數(shù)最優(yōu)值。最后在標定的參數(shù)下進行仿真試驗,對比有機肥仿真休止角和實際休止角,驗證標定的有機肥離散元模型參數(shù)的準確性。不同的物料對休止角有顯著影響的離散元模型參數(shù)是有差異的,先進行篩選可以針對有機肥的特性進行標定。與正交試驗相比,采用響應(yīng)面分析方法建立的回歸模型是連續(xù)的,尋優(yōu)得到的最優(yōu)值更準確。
表1 有機肥基本性質(zhì)
本文采用圓筒提升的方法進行試驗[15],如圖1所示。試驗時,首先使鋼質(zhì)圓筒(內(nèi)徑40 mm、高80 mm)底面與鋼板(長400 mm、寬200 mm)接觸,然后向鋼質(zhì)圓筒內(nèi)填充有機肥直至填滿。使用REGER萬能試驗機將鋼質(zhì)圓筒以20 mm/s的速度向上提升,從而使有機肥形成一個顆粒堆,最后使用三量181-101型數(shù)顯傾角儀(分辨率:0.05°,精度:±0.2°)測量有機肥休止角。該試驗重復(fù)10次,取平均值,最終得到有機肥實際休止角為38.15°。
圖1 有機肥休止角物理試驗
1.4.1 接觸模型
有機肥顆粒間受水分和化學(xué)物質(zhì)的作用存在粘附現(xiàn)象,普通的接觸模型難以準確地模擬施肥機械作業(yè)過程中有機肥的力學(xué)行為。本文選擇“Hertz-Mindlin with Johnson-Kendall-Roberts”接觸模型[16],該模型是一個凝聚力接觸模型,可以考慮在接觸區(qū)域中范德華力的影響和允許用戶模擬強黏性的系統(tǒng),如干燥的粉末或濕顆粒。在這個模型中,切向彈性力、法向耗散力和切向耗散力均與Hertz-Mindlin(no slip)接觸模型中的計算方法一致,但JKR法向彈性力的實現(xiàn)基于Johson-Kendall-Roberts理論,取決于重疊量、相互作用參數(shù)和表面能
式中JKR為JKR法向彈性力,N;為相互接觸的兩個顆粒的接觸圓半徑,m;為重疊量,m為表面能,N/m;E為當(dāng)量彈性模量,Pa;R為當(dāng)量半徑,m。當(dāng)量彈性模量和當(dāng)量半徑定義為
式中E,ν,R和E,ν,R分別為相互接觸的2個顆粒的彈性模量、泊松比和半徑。
當(dāng)=0時,力變成Hertz-Mindlin法向力
即使顆粒并不是直接接觸,該模型也提供吸引凝聚力,顆粒間具有非零凝聚力的最大間隙為
式中δ為顆粒間具有非零凝聚力時的法向最大間隙,m;α為2個顆粒的接觸圓半徑,m。
當(dāng)顆粒并非實際接觸并且間隙小于δ時,凝聚力達到最大值
摩擦力的計算和Hertz-Mindlin(no slip)接觸模型不同,不同在于它是取決于JKR法向力的正向排斥部分。因此,該模型在接觸力的凝聚力分量更大時提供一個更大的摩擦力。
1.4.2 有機肥和結(jié)構(gòu)模型
在進行仿真前必須要先建立有機肥和結(jié)構(gòu)的模型,有機肥顆粒近似球形,故以基本球體作為有機肥模型。仿真試驗表明:相同條件下,僅建立半徑1 mm的有機肥模型形成的休止角與根據(jù)粒徑分布建立多種半徑的有機肥模型形成的休止角的相對誤差為2.22%??紤]到顆粒半徑太小會影響計算速度,本文僅建立半徑1 mm(占比最大)的有機肥模型。在SolidWorks中建立鋼質(zhì)圓筒和鋼板的三維模型,并導(dǎo)入EDEM2.6中作為結(jié)構(gòu)模型。通過試驗和查閱相關(guān)文獻[17-22],得到有機肥和鋼板的本征參數(shù),如表2所示。
1.4.3 仿真參數(shù)的設(shè)置
本文在進行仿真試驗[23]時,設(shè)置鋼質(zhì)圓筒的提升速度20 mm/s(此速度下可以較好地形成肥堆),生成顆??倲?shù)量10 000個,生成速率2 000個/s,固定時間步長是Rayleigh時間步長的22%,數(shù)據(jù)保存時間間隔為0.01 s,網(wǎng)格大小為2 mm。
表2 有機肥、鋼板本征參數(shù)
1.5.1 Plackett-Burman Design篩選顯著性參數(shù)
接觸參數(shù)和接觸模型參數(shù)中并不是所有參數(shù)都對休止角有顯著影響[24-26],沒有顯著影響的參數(shù)并不能基于休止角來標定,否則標定出的參數(shù)是不準確的。所以本文應(yīng)用軟件Design Expert 8.0.6[27-28]進行Plackett-Burman Design[29],對接觸參數(shù)(有機肥-有機肥恢復(fù)系數(shù)、有機肥-有機肥靜摩擦系數(shù)、有機肥-有機肥滾動摩擦系數(shù)、有機肥-鋼板恢復(fù)系數(shù)、有機肥-鋼板靜摩擦系數(shù)和有機肥-鋼板滾動摩擦系數(shù))和接觸模型參數(shù)(JKR)進行篩選,篩選出對有機肥休止角有顯著影響的參數(shù)。
參考文獻[17-21]中,有機肥顆粒間的恢復(fù)系數(shù)范圍為0.1~0.6,靜摩擦系數(shù)范圍為0.3~1,滾動摩擦系數(shù)范圍為0.1~0.5;有機肥顆粒與鋼板間的恢復(fù)系數(shù)范圍為0.048~0.6,靜摩擦系數(shù)范圍為0.2~0.86,滾動摩擦系數(shù)范圍為0.1~0.5。通過綜合對比參考文獻中肥料和本文中有機肥特性的差異,確定了接觸參數(shù)的取值范圍,并通過仿真預(yù)試驗確定了JKR的取值范圍。Plackett-Burman Design的因素水平如表3所示,共7個因素,各取1個高水平、1個低水平,并選擇1個中心點;試驗方案及結(jié)果如表4所示,設(shè)置4個空白列用于誤差分析,共進行13次試驗。
表3 Plackett-Burman Design因素水平表
注:O為有機肥;S為鋼板;JKR為表面能
Note: O is organic fertilizer, S is steel plate and JKR is surface energy.
利用軟件Design Expert 8.0.6對試驗結(jié)果進行方差分析,得到7個參數(shù)對休止角的影響效果和顯著性,如表5所示。由表5可知,7個參數(shù)對休止角的效應(yīng)均是正效應(yīng),即休止角隨著參數(shù)的增大而增大;其中對休止角影響顯著(<0.05)的參數(shù)包括:有機肥-有機肥滾動摩擦系數(shù)()、JKR()和有機肥-鋼板滾動摩擦系數(shù)(),而其余參數(shù)影響不顯著。分析得出,恢復(fù)系數(shù)對休止角影響不顯著的原因在于基于休止角的試驗中不存在明顯的彈性碰撞,而靜摩擦系數(shù)沒有顯著影響的原因在于有機肥顆粒的形狀,有機肥顆粒是球形顆粒,相比于玉米等非球形顆粒,滾動摩擦系數(shù)對休止角的影響要比靜摩擦系數(shù)顯著。
表4 Plackett-Burman Design方案及結(jié)果
注:,,,為空白列。
Note:,,andindicates blank column.
表5 Plackett-Burman Design試驗結(jié)果方差分析
1.5.2 最陡爬坡試驗確定顯著性參數(shù)最優(yōu)值區(qū)間
應(yīng)用響應(yīng)面分析方法建立回歸模型求解最優(yōu)值的前提是因素的最優(yōu)值在所選高低水平范圍內(nèi),最陡爬坡試驗可以較快地確定因素最優(yōu)值所在區(qū)間。根據(jù)Plackett-Burman Design的試驗結(jié)果,本文只將3個顯著性參數(shù)(有機肥-有機肥滾動摩擦系數(shù)、JKR和有機肥-鋼板滾動摩擦系數(shù))按照選定步長逐步增加(參數(shù)影響效應(yīng)是正效應(yīng)),其余參數(shù)選擇中間水平(有機肥-有機肥恢復(fù)系數(shù)0.6、有機肥-有機肥靜摩擦系數(shù)0.65、有機肥-鋼板恢復(fù)摩擦系數(shù)0.6、有機肥-鋼板靜摩擦系數(shù)0.7)進行最陡爬坡試驗,并計算有機肥仿真休止角與實際休止角的相對誤差,試驗方案及結(jié)果如表6所示。
由表6可知,隨著有機肥-有機肥滾動摩擦系數(shù)()、JKR()和有機肥-鋼板滾動摩擦系數(shù)()的增大,有機肥仿真休止角不斷增大,仿真休止角與實際休止角的相對誤差先減小后增大。第2組試驗的相對誤差最小,所以3個顯著性參數(shù)的最優(yōu)區(qū)間在第2組試驗所選的水平附近,因此選擇第1、2、3組試驗所選的水平進行響應(yīng)面分析試驗,建立回歸模型求解顯著性參數(shù)的最優(yōu)值。
表6 最陡爬坡試驗方案及結(jié)果
1.5.3 響應(yīng)面分析試驗建立回歸模型求解顯著性參數(shù)最優(yōu)值
根據(jù)最陡爬坡試驗結(jié)果,進行Central Composite Design響應(yīng)面[30]分析試驗。試驗中,非顯著性參數(shù)同樣選擇中間水平(有機肥-有機肥恢復(fù)系數(shù)0.6、有機肥-有機肥靜摩擦系數(shù)0.65、有機肥-鋼板恢復(fù)摩擦系數(shù)0.6、有機肥-鋼板靜摩擦系數(shù)0.7),顯著性參數(shù)水平如表7所示,各取5個水平。選擇6個中心點進行誤差估計,試驗方案及結(jié)果如表8所示,共進行20次試驗,包括6次中心點的重復(fù)試驗。
表7 Central Composite Design因素水平表
表8 Central Composite Design方案及結(jié)果
應(yīng)用軟件Design Expert 8.0.6對試驗結(jié)果進行分析,得到二次回歸模型。該二次回歸模型的方差分析如表9所示,有機肥-有機肥滾動摩擦系數(shù)()和JKR()對有機肥休止角的影響十分顯著(<0.01);該線性回歸模型<0.01,說明休止角與所得回歸方程關(guān)系是極顯著的;失擬項=0.051 9>0.05,說明所得回歸方程與實際擬合中非正常誤差所占比例小,擬合性較好。本試驗的變異系數(shù)CV=2.06%,故本試驗可靠性較好;決定系數(shù)2=0.965 5和校正決定系數(shù)2adj=0.934 5,接近1,故所得回歸方程可靠度較高;精確度Adeq Precision=20.458,故該回歸模型精確度良好。
表9 Central Composite Design二次回歸模型方差分析
通過剔除影響非常不顯著的項,并嘗試添加三次項對二次回歸模型進行優(yōu)化。優(yōu)化后的回歸模型方差分析如表10所示,失擬項=0.405 6、變異系數(shù)CV=1.30%、決定系數(shù)2=0.980 7、校正決定系數(shù)2adj=0.973 8、精確度Adeq Precision=45.591。較優(yōu)化前,所得回歸方程在擬合性、可靠性和精確性等方面都有很大改善。優(yōu)化后的回歸方程為
=23.605 72+215.056 53+198.853 17
+9.059-1586.300 932+4906.266 463(9)
表10 Central Composite Design優(yōu)化回歸模型方差分析
應(yīng)用軟件Design Expert 8.0.6以有機肥實際休止角為目標對優(yōu)化后的回歸方程求解尋優(yōu),得到3個顯著性參數(shù)的最優(yōu)值,有機肥-有機肥滾動摩擦系數(shù)為0.10,JKR為0.015,有機肥-鋼板滾動摩擦系數(shù)為0.11。
為了驗證所得3個顯著性參數(shù)最優(yōu)值的準確性,本文將3個顯著性參數(shù)設(shè)置為最優(yōu)值(有機肥-有機肥滾動摩擦系數(shù)0.10、JKR 0.015、有機肥-鋼板滾動摩擦系數(shù)0.11),非顯著性參數(shù)同樣選擇中間水平(有機肥-有機肥恢復(fù)系數(shù)0.6、有機肥-有機肥靜摩擦系數(shù)0.65、有機肥-鋼板恢復(fù)摩擦系數(shù)0.6、有機肥-鋼板靜摩擦系數(shù)0.7),其他設(shè)置不變,應(yīng)用軟件EDEM 2.6進行有機肥休止角仿真試驗,測得有機肥仿真休止角為37.99°。與有機肥實際休止角38.15°的相對誤差為0.42%,無顯著性差異,說明所得3個顯著性參數(shù)的最優(yōu)值準確可靠。仿真試驗與物理試驗的對比如圖2所示,兩者有機肥顆粒堆輪廓十分接近。
圖2 仿真試驗與物理試驗對比
1)通過Plackett-Burman Design篩選出對有機肥休止角有顯著影響的接觸參數(shù)和模型參數(shù)包括:有機肥-有機肥滾動摩擦系數(shù)、表面能JKR和有機肥-鋼板滾動摩擦系數(shù)。
2)通過最陡爬坡試驗確定了3個顯著性參數(shù)的最優(yōu)值區(qū)間,通過Central Composite Design響應(yīng)面分析試驗建立并優(yōu)化了有機肥休止角與顯著性參數(shù)之間的回歸模型,對優(yōu)化后的回歸模型方差分析發(fā)現(xiàn),除了3個顯著性參數(shù)對有機肥休止角影響顯著外,有機肥-有機肥滾動摩擦系數(shù)的二次項和三次項對有機肥休止角也有顯著影響。
3)以有機肥實際休止角為目標值,對優(yōu)化后的回歸方程進行尋優(yōu)求解,得到3顯著性參數(shù)的最優(yōu)值,有機肥-有機肥滾動摩擦系數(shù)為0.10,表面能JKR為0.015,有機肥-鋼板滾動摩擦系數(shù)為0.11。對比驗證試驗結(jié)果表明,有機肥仿真休止角與實際休止角無顯著差異,所標定的有機肥離散元模型參數(shù)準確可靠。
[1] Cundall P A, Strack O D L. A discrete numerical method for granular assemblis[J]. Geptechnique, 1979, 29(1): 47-65.
[2] 胡國明. 顆粒系統(tǒng)的離散元素法分析仿真—離散元素法的工業(yè)應(yīng)用與EDEM軟件簡介[M]. 武漢:武漢理工大學(xué)出版社,2010.
[3] 賀一鳴,吳明亮,向偉,等. 離散元法在農(nóng)業(yè)工程領(lǐng)域的應(yīng)用進展[J]. 中國農(nóng)學(xué)通報,2017,33(20):133-137. He Yiming, Wu Mingliang, Xiang Wei, et al. Application progress of discrete element method in agricultural engineering[J]. Chinese Agricultural Science Bulletin, 2017, 33(20): 133-137. (in Chinese with English abstract)
[4] 于建群,付宏,李紅,等. 離散元法及其在農(nóng)業(yè)機械工作部件研究與設(shè)計中的應(yīng)用[J]. 農(nóng)業(yè)工程學(xué)報,2005,21(5):1-6. Yu Jianqun, Fu Hong, Li Hong, et al. Application of discrete element method to research and design of working parts of agricultural machines[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2005, 21(5): 1-6. (in Chinese with English abstract)
[5] 馬征,李耀明,徐立章. 農(nóng)業(yè)工程領(lǐng)域顆粒運動研究綜述[J]. 農(nóng)業(yè)機械學(xué)報,2013,44(2):22-29. Ma Zheng, Li Yaoming, Xu Lizhang. Summarize of particle movements research in agricultural engineering realm[J]. Transactions of the Chinese Society of Agricultural Machinery, 2013, 44(2): 22-29. (in Chinese with English abstract)
[6] 龔明. 材料物料特性參數(shù)標定[R]. 北京:EDEM用戶大會報告,2013.
[7] 王云霞,梁志杰,張東興,等. 基于離散元的玉米種子顆粒模型種間接觸參數(shù)標定[J]. 農(nóng)業(yè)工程學(xué)報,2016,32(22):36-42. Wang Yunxia, Liang Zhijie, Zhang Dongxing, et al. Calibration method of contact characteristic parameters for corn seeds based on EDEM[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2016, 32(22): 36-42. (in Chinese with English abstract)
[8] 劉凡一,張艦,李博,等. 基于堆積試驗的小麥離散元參數(shù)分析及標定[J]. 農(nóng)業(yè)工程學(xué)報,2016,32(12):247-253. Liu Fanyi, Zhang Jian, Li Bo, et al. Calibration of parameters of wheat required in discrete element method simulation based on repose angle of particle heap[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2016, 32(12): 247-253. (in Chinese with English abstract)
[9] 石林榕,趙武云,孫偉. 基于離散元的西北旱區(qū)農(nóng)田土壤顆粒接觸模型和參數(shù)標定[J]. 農(nóng)業(yè)工程學(xué)報,2017,33(21):181-187. Shi Linrong, Zhao Wuyun, Sun Wei. Parameter calibration of soil particles contact model of farmland soil in northwest arid region based on discrete element method[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(21): 181-187. (in Chinese with English abstract)
[10] 王憲良,胡紅,王慶杰,等. 基于離散元的土壤模型參數(shù)標定方法[J]. 農(nóng)業(yè)機械學(xué)報,2017,48(12):78-85. Wang Xianliang, Hu Hong, Wang Qingjie, et al. Calibration method of soil contact characteristic parameters based on DEM theory[J]. Transactions of the Chinese Society of Agricultural Machinery, 2017, 48(12): 78-85. (in Chinese with English abstract)
[11] 張銳,韓佃雷,吉巧麗,等. 離散元模擬中沙土參數(shù)標定方法研究[J]. 農(nóng)業(yè)機械學(xué)報,2017,48(3):49-56. Zhang Rui, Han Dianlei, Ji Qiaoli, et al. Calibration methods of sandy soil parameters in simulation of discrete element method[J]. Transactions of the Chinese Society of Agricultural Machinery, 2017, 48(3): 49-56. (in Chinese with English abstract)
[12] 武濤,黃偉鳳,陳學(xué)深,等. 考慮顆粒間粘結(jié)力的黏性土壤離散元模型參數(shù)標定[J]. 華南農(nóng)業(yè)大學(xué)學(xué)報,2017,38(3):93-98. Wu Tao, Huang Weifeng, Chen Xueshen, et al. Calibration of discrete element model parameters for cohesive soil considering the cohesion between particles[J]. Journal of South China Agricultural University, 2017, 38(3): 93-98. (in Chinese with English abstract)
[13] Michael R, Kevin J H. A methodical calibration procedure for discrete element models[J]. Powder Technology, 2017, 307: 73-83.
[14] Coetzee C J. Calibration of the discrete element method[J]. Powder Technology, 2017, 310: 104-142.
[15] 田曉紅,李光濤,張淑麗. 谷物自然休止角測量方法的探究[J]. 糧食加工,2010,35(1):68-71. Tian Xiaohong, Li Guangtao, Zhang Shuli. Determination of angle of repose[J]. Grain Processing, 2010, 35(1): 68-71. (in Chinese with English abstract)
[16] DEM Solutions. EDEM 2.6 theory reference guide[M]. Edinburgh: DEM Solutions, 2014.
[17] 張曉明. 有機肥顆粒熱風(fēng)干燥工藝及裝備研究[D]. 北京:中國農(nóng)業(yè)大學(xué),2017. Zhang Xiaoming. Study on the Drying Process and Equipment for Organic Fertilizer Rellets[D].Beijing: China Agricultural University, 2017.
[18] 羅帥,袁巧霞,GOUDA Shaban,等. 基于JKR粘結(jié)模型的蚯蚓糞機質(zhì)離散元法參數(shù)標定[J]. 農(nóng)業(yè)機械學(xué)報,2018,49(4):343-350.Luo Shuai, Yuan Qiaoxia, GOUDA Shaban, et al. Parameters calibration of vermicomposting nursery substrate with discrete element method based on JKR contact model[J]. ransactions of the Chinese Society of Agricultural Machinery, 2018, 49(4): 343-350. (in Chinese with English abstract)
[19] Landry H, Lagu? C, Roberge M. Discrete element representation of manure products[J]. Computers and electronics in agriculture, 2006, 51: 17-34.
[20] Landry H, Thirion F, Lagu? C, et al. Numerical modeling of the flow of organic fertilizers in land application equipment[J]. Computers and electronics in agriculture, 2006, 51: 35-53.
[21] Landry H, Lagu? C, Roberge M. Physical and rheological properties of manure products[J]. Applied Engineering in Agriculture, 2004, 20(3): 277-288.
[22] 聞邦椿. 機械設(shè)計手冊(第5版)[M]. 北京:機械工業(yè)出版社,2010.
[23] 王國強. 離散單元法及其在EDEM上的實踐[M]. 西安:西北工業(yè)大學(xué)出版社,2010.
[24] 夏鵬,李郁,楊公波. 散粒物料堆積角離散元仿真研究[J]. 起重運輸機械,2015(2);107-110. Xia Peng, Li Yu, Yang Gongbo. Study on the discrete element simulation of scattered materials[J]. Hoisting and Conveying Machinery, 2015(2):107-110. (in Chinese with English abstract)
[25] 韓燕龍,賈富國,唐玉榮,等. 顆粒滾動摩擦系數(shù)對堆積特性的影響[J]. 物理學(xué)報,2014,63(17):4501-4507. Han Yanlong, Jia Fuguo, Tang Yurong, et al. Influence of granular coefficient of rolling friction on accumulation characteristics[J]. Acta Physica Sinica, 2014, 63(17): 4501-4507. (in Chinese with English abstract)
[26] 賈旭光,陳曦,李鑫. 不同粒度和堆載形態(tài)下散體瞬時自然安息角的實驗研究[J]. 現(xiàn)代礦業(yè),2015,549(1):25-27. Jia Xuguang, Chen Xi, Li Xin. Experiment research on instantaneous natural repose angle of granular slope under different granularity and stack form[J]. Modern Mining, 2015, 549(1): 25-27. (in Chinese with English abstract)
[27] 徐向宏,何明珠. 試驗設(shè)計與Design-Expert、SPSS應(yīng)用[M]. 北京:科學(xué)出版社,2010.
[28] 葛宜元. 試驗設(shè)計方法與Design-Expert軟件應(yīng)用[M]. 哈爾濱:哈爾濱工業(yè)大學(xué)出版社,2015.
[29] Ledi T, Pat S, Terri M, et al. Plackett-Burman designs[EB/OL]. 2013-10-30[2013-10-30]. http://www.itl. nist.gov/div898/handbook/pri/section3/pri335.html.
[30] Ledi T, Pat S, Terri M, et al. Response surface designs[EB/OL]. 2013-10-30[2013-10-30]. http://www. itl.nist.gov/div898/handbook/pri/section3/pri336.html.
Parameter calibration of discrete element model of organic fertilizer particles for mechanical fertilization
Yuan Quanchun, Xu Liming※, Xing Jiejie, Duan Zhuangzhuang, Ma Shuai, Yu Changchang, Chen Chen
(,,100083,)
As the parameters of discrete element model for organic fertilizer are difficult to obtain directly from literatures or experiments, this study proposed a method to calibrate the parameters based on tests of the angle of repose. Firstly, a physical test was carried out. The diameters of organic fertilizer particles were between 0.25 and 2 mm, the fertilizer moisture was 13.7%, and the organic fertilizer was crushed to meet the requirements of uniformity of fertilization. In the test, in order to obtain the organic fertilizer heap, the cylinder was lifted up by a universal testing instrument at a speed of 20 mm/s. Then, the angle of repose of the fertilizer was measured with inclinometer and the test was repeated for 10 times, as the mean angle of repose was 38.15°. Secondly, the simulation test was carried out. Considering the cohesion characteristics between particles of organic fertilizer, the contact model of Hertz-Mindlin with Johnson-Kendall-Roberts was used in simulation modeling. Since the organic fertilizer particles were near-spherical, the basic model of the software default was adopted to simulate organic fertilizer granules. In the simulation test, the relative error was 2.22% by comparing the angle of repose formed by a variety of particle sizes according to actual particle size distribution and the angle of repose formed by the same particle size (1 mm). Therefore, a basic sphere model with a radius of 1 mm was used and the three-dimension geometry model built by SolidWorks was imported into the simulation software, with the cylinder lifting speed was 20 mm/s, the total number of particles was 10 000, the generation rate was 2 000/s, the fixed time step was 22% of Rayleigh time step, the target save interval was 0.01 s, and the cell size was 2 mm. Not all discrete element model parameters of organic fertilizer had significant impact on the angle of repose, so a screening through Plackett-Burman Design was made. The results showed that rolling friction coefficient between organic fertilizer, surface energy (JKR) and rolling friction coefficient between organic fertilizer and steel plate had significant impact on angle of repose. The optimal value interval of the there significant parameters was determined by the steepest ascent test, and built the linear regression model of angle of repose through Central Composite Design. The fitting of the linear regression model was optimized by adding the quadratic term and cubic term of rolling friction coefficient between organic fertilizer with the fitting of the regression model had a great improvement, and thevalue of the lack of fit was improved from 0.091 to 0.405 6. Thirdly, the optimal value of the significant parameters was obtained with the actual angle of repose as a goal, with the rolling friction coefficient between organic fertilizer was 0.10, the JKR was 0.015, and the rolling friction coefficient between organic fertilizer and steel plate was 0.11. Finally, in order to verify the accuracy of the calibrated parameters, the simulation verification test of the calibration parameters was carried out, and the relative error between the simulated angle of repose and the actual angle of repose is 0.42%, which meant that the calibration parameters were accurate. This study can provide reference for the selection of discrete element model parameters of organic fertilizer.
models; fertilizers; discrete element method; calibration of parameters; cohesion; angle of repose
10.11975/j.issn.1002-6819.2018.18.003
S220.1
A
1002-6819(2018)-18-0021-07
2018-02-01
2018-06-12
現(xiàn)代農(nóng)業(yè)產(chǎn)業(yè)技術(shù)體系建設(shè)專項資金資助(CARS-29)
袁全春,博士生,主要從事生物生產(chǎn)自動化研究。 Email:yqcmail@qq.com
徐麗明,教授,博士生導(dǎo)師,主要從事生物生產(chǎn)自動化研究。Email:xlmoffice@126.com
袁全春,徐麗明,邢潔潔,段壯壯,馬 帥,于暢暢,陳 晨. 機施有機肥散體顆粒離散元模型參數(shù)標定[J]. 農(nóng)業(yè)工程學(xué)報,2018,34(18):21-27. doi:10.11975/j.issn.1002-6819.2018.18.003 http://www.tcsae.org
Yuan Quanchun, Xu Liming, Xing Jiejie, Duan Zhuangzhuang, Ma Shuai, Yu Changchang, Chen Chen. Parameter calibration of discrete element model of organic fertilizer particles for mechanical fertilization[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34(18): 21-27. (in Chinese with English abstract)
doi: 10.11975/j.issn.1002- 6819. 2018. 18.003 http://www.tcsae.org