(ansys的97单元计算铁磁材料电磁场有缺陷的实例)97单元3D线圈电压励磁（附APDL程序）

jan_2000

ANSYS的97单元电压加载原理大致知道是怎么回事了，相关理论文献见附件。但此程序仍然没调通，继续寻求高人指点！

wqlishijie

34# jan_2000

看了你们长篇的讨论，好想向你们学习哦，我是刚学ANSYS的，做变压器的某些电磁分析，是否可以加两位的QQ？  我的Q是308393778，你们若不方便公布，能否麻烦你们加下我呢？？

jan_2000

顶一下！呵呵

jan_2000

楼主真执着，要想帖子不沉，可以向版主申请置顶或者改个有新意的名字
fossiler 发表于 2010-10-21 13:31

呵呵，好像斑竹不怎么来的。

jan_2000

呵呵，重大发现：97单元作用于铁磁材料计算时有缺陷，慎用！！！
对比如下（主程序部分完全一致，分别用97和117单元进行计算，结果完全不同！！！）。

jan_2000

*SET,b,0.1   
*SET,L1,0.180      
*SET,L2,0.170   
*SET,a,0.03   
*SET,i,1*1.41414  !电流幅值   
*SET,freq,50
*SET,pi,3.1415926
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!   
*SET,TurnNumber,220        
*SET,CoilThickness,0.010     
*SET,CoilHeight,0.085         
*SET,CoilResistivity,1.85e-8   
*SET,AirGap,0.001            
!根据铁心尺寸确定圆筒绕组内径
*IF,a,LT,b,THEN
r= SQRT((a/2)**2+(b/2)**2)+ AirGap
*ELSE
r= SQRT((a/2)**2+(b/2)+ AirGap
*ENDIF
!
*SET, dext, r*0.5            
!
/prep7
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
et,1,97,0
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
et,2,97,1  
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!以下为线圈设立
et,3,97,0
!空气
!!!!!!!!!!!!!!!!!
MP,MURX,1,1
!!!!!!!!!!!!!!!!!
!铁心部分
!!!!!!!!!!!!!!!!!
MP,MURX,2,950  
MP,MURY,2,950  
MP,MURZ,2,20  
MP,RSVX,2,2.1053e-007   !sigma=0.5e7(S/m)
MP,RSVY,2,2.1053e-007
MP,RSVZ,2,6.8400e-004
!!!!!!!!!!!!!!!!!
!线圈
!!!!!!!!!!!!!!!!!
MP,MURX,3,1
MP,RSVX,3,CoilResistivity
!!!!!!!!!!!!!!!!!
!线圈建模
wpoffs,-(L1/2-a/2),0,0
wprot, , 90 ,   
cylind, r, r+ CoilThickness, -CoilHeight/2, CoilHeight/2, 0, 360
wprot, , -90 ,   
wpoffs,(L1/2-a/2),0,0
!铁心建模
block,-L1/2,L1/2,-L2/2,L2/2,-b/2,b/2
block,-L1/2+a,L1/2-a,-L2/2+a,L2/2-a,-b/2,b/2
!空气包
*get,xmin,kp,,mnloc,x        ! Minimum X, Y, or Z keypoint coordinate in the selected set (in the active coordinate system).
*get,xmax,kp,,mxloc,x
*get,ymin,kp,,mnloc,y
*get,ymax,kp,,mxloc,y
*get,zmin,kp,,mnloc,z
*get,zmax,kp,,mxloc,z
block, xmin-dext, xmax+dext, ymin-dext, ymax+dext, zmin-dext, zmax+dext   
vovlap,all
numcmp,all         !重新进行实体编号（->1，2，3...）
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
vsel,s,VOLU,,3
cm, IronCore,volu
allsel, all
vsel,s,VOLU,,4
vsel,a, VOLU,,5
cm, AirBig,volu
allsel, all
vsel,s,VOLU,,1
vsel,a, VOLU,,2
cm, Coil,volu
allsel, all
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
cmsel, s, AirBig
VATT, 1, , 1, 0
cmsel, s, IronCore
VATT, 2, , 2, 0
!!!!!!!!!!!!!!!!!!!

wpoffs,-(L1/2-a/2),0,0
wprot,,90
CSWPLA, 11,1,,
cmsel, s, Coil
VATT, 3, 1, 3, 11
!!!!!!!!!!!!!!!!!!!
!网格化
!!!!!!!!!!!!!!!!!!!
allsel,all
LESIZE,all,b/6, , , , , , ,1
cmsel,s,Coil
cmsel,a, IronCore
vsweep,all
allsel,all
cmsel,s, AirBig
MSHAPE,1,3D    !Mesh with 3-D tetrahedral-shaped elements
MSHKEY,0       !Use free meshing
vmesh,all
allsel,all
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
/SOL
!*  
!!!!!!!!!!!!!!!!!!!
!定义问题类型：HARMONIC
!!!!!!!!!!!!!!!!!!!
ANTYPE,3
!!!!!!!!!!!!!!!!!!!!!
harfrq,freq
allsel,all
ASEL,S,EXT
DA,all,ASYM
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!*  
cmsel,s, Coil
eslv,s
BFE,all,JS,,0,i*TurnNumber /( CoilThickness* CoilHeight),0
allsel,all
SOLVE
/POST1  
set,last
POWERH

jan_2000

！117
*SET,b,0.1   
*SET,L1,0.180      
*SET,L2,0.170   
*SET,a,0.03   
*SET,i,1*1.41414  !电流幅值   
*SET,freq,50
*SET,pi,3.1415926
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!   
*SET,TurnNumber,220         
*SET,CoilThickness,0.010     
*SET,CoilHeight,0.085         
*SET,CoilResistivity,1.85e-8   
*SET,AirGap,0.001           
!根据铁心尺寸确定圆筒绕组内径
*IF,a,LT,b,THEN
r= SQRT((a/2)**2+(b/2)**2)+ AirGap
*ELSE
r= SQRT((a/2)**2+(b/2)+ AirGap
*ENDIF
!
*SET, dext, r*0.5            
!
/prep7
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
et,1,117,0
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!以下为铁心区设立
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
et,2,117,1  
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!以下为线圈设立
et,3,117,0
!空气
!!!!!!!!!!!!!!!!!
MP,MURX,1,1
!!!!!!!!!!!!!!!!!
!铁心部分
!!!!!!!!!!!!!!!!!
MP,MURX,2,950  
MP,MURY,2,950  
MP,MURZ,2,20  
MP,RSVX,2,2.1053e-007   !sigma=0.5e7(S/m)
MP,RSVY,2,2.1053e-007
MP,RSVZ,2,6.8400e-004
!!!!!!!!!!!!!!!!!
!线圈
!!!!!!!!!!!!!!!!!
MP,MURX,3,1
MP,RSVX,3,CoilResistivity
!!!!!!!!!!!!!!!!!
!线圈建模
wpoffs,-(L1/2-a/2),0,0
wprot, , 90 ,   
cylind, r, r+ CoilThickness, -CoilHeight/2, CoilHeight/2, 0, 360
wprot, , -90 ,   
wpoffs,(L1/2-a/2),0,0
!铁心建模
block,-L1/2,L1/2,-L2/2,L2/2,-b/2,b/2
block,-L1/2+a,L1/2-a,-L2/2+a,L2/2-a,-b/2,b/2
!空气包
*get,xmin,kp,,mnloc,x        ! Minimum X, Y, or Z keypoint coordinate in the selected set (in the active coordinate system).
*get,xmax,kp,,mxloc,x
*get,ymin,kp,,mnloc,y
*get,ymax,kp,,mxloc,y
*get,zmin,kp,,mnloc,z
*get,zmax,kp,,mxloc,z
block, xmin-dext, xmax+dext, ymin-dext, ymax+dext, zmin-dext, zmax+dext   
vovlap,all
numcmp,all         !重新进行实体编号（->1，2，3...）
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
vsel,s,VOLU,,3
cm, IronCore,volu
allsel, all
vsel,s,VOLU,,4
vsel,a, VOLU,,5
cm, AirBig,volu
allsel, all
vsel,s,VOLU,,1
vsel,a, VOLU,,2
cm, Coil,volu
allsel, all
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
cmsel, s, AirBig
VATT, 1, , 1, 0
cmsel, s, IronCore
VATT, 2, , 2, 0
!!!!!!!!!!!!!!!!!!!
wpoffs,-(L1/2-a/2),0,0
wprot,,90
CSWPLA, 11,1,,
cmsel, s, Coil
VATT, 3, 1, 3, 11
!!!!!!!!!!!!!!!!!!!
!网格化
!!!!!!!!!!!!!!!!!!!
allsel,all
LESIZE,all,b/6, , , , , , ,1
cmsel,s,Coil
cmsel,a, IronCore
vsweep,all
allsel,all
cmsel,s, AirBig
MSHAPE,1,3D    !Mesh with 3-D tetrahedral-shaped elements
MSHKEY,0       !Use free meshing
vmesh,all
allsel,all
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
/SOL
!*  
!!!!!!!!!!!!!!!!!!!
!定义问题类型：HARMONIC
!!!!!!!!!!!!!!!!!!!
ANTYPE,3
!!!!!!!!!!!!!!!!!!!!!
harfrq,freq
allsel,all
ASEL,S,EXT
DA,ALL,AZ,0
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!施加电压激励
!*  
cmsel,s, Coil
eslv,s
BFE,all,JS,,0,i*TurnNumber /( CoilThickness* CoilHeight),0
allsel,all
SOLVE
/POST1  
set,last
POWERH

jan_2000

看来，很遗憾，利用97单元给线圈加载电压，从而对铁芯励磁，看来只能是幻影一场啊！ANSYS难道就没有别的单元能做类似的事情？别的软件能吗，有待高手解疑！

自我解答：
1. ANSYS作3D电磁计算的只有97单元能进行载压（包括直接加和场路耦合）；
2. 似乎ANSOFT能实现，可以试试看。

jan_2000

顺便给出一个117单元励磁、铁芯叠片各向异性仿真的涡流和磁密计算结果，请鉴赏。

jan_2000

改了一下标题名，再顶一下，呵呵。

sugars1985

有个疑问，用solid97载压线圈，进行谐波分析，最终用powerh算出来的powerloss 只计算了实部，即 PAVG= Ireal^2*R   (Ireal为电流实部幅值除以sqrt(2))
我用bfe 和circu124单元进行加载，结果都出现相同的问题。
而如果人工提取电流实部虚部，再求出I有效值，则I^2*R值不变，
而用Powerh算，只要电压加载的相角改变，则PAVG就会跟着变，而且严格符合PAVG= Ireal^2*R
求高手解答

jan_2000

有个疑问，用solid97载压线圈，进行谐波分析，最终用powerh算出来的powerloss 只计算了实部，即 PAVG= Ireal^2*R   (Ireal为电流实部幅值除以sqrt(2))
我用bfe 和circu124单元进行加载，结果都出现相同的问题。
而 ...
sugars1985 发表于 2011-3-18 16:25

你的描述有点让人很难理解。
能否将你的源程序贴出来？

sugars1985

/PREP7  

CYLIND,0.01,0.02,0.1,0.13   !建立一个圆筒作为线圈
BLOCK,-0.04,0.04,-0.04,0.04,0.05,0.15
VOVLAP,ALL  
NUMCMP,VOLU
ET,1,97   !除线圈外的其它实体单元类型
KEYOPT,1,1,1
ET,2,97   !solid 定义电路耦合线圈
KEYOPT,2,1,3  !AX, AY, AZ, CURR, EMF degrees of freedom: circuit-coupled stranded coil
KEYOPT,2,2,0
KEYOPT,2,5,0
ET,3,111
KEYOPT,3,1,1  !无限远单元
MPTEMP,,,,,,,,   !空气
MPTEMP,1,0  
MPDATA,MURX,1,,1
MPTEMP,,,,,,,,  
MPTEMP,1,0  
MPDATA,RSVX,1,,1.75
MPTEMP,,,,,,,,   !线圈
MPTEMP,1,0  
MPDATA,MURX,2,,1
MPTEMP,,,,,,,,  
MPTEMP,1,0  
MPDATA,RSVX,2,,1.75e-8

CM,_Y,VOLU     !赋予材料属性
VSEL, , , ,       1
CM,_Y1,VOLU
CMSEL,S,_Y  
!*  
CMSEL,S,_Y1
VATT,       2, ,   2,       0   
CMSEL,S,_Y  
CMDELE,_Y   
CMDELE,_Y1  
!*  
CM,_Y,VOLU  
VSEL, , , ,       2
CM,_Y1,VOLU
CMSEL,S,_Y  
!*  
CMSEL,S,_Y1
VATT,       1, ,   1,       0   
CMSEL,S,_Y  
CMDELE,_Y   
CMDELE,_Y1  
ALLSEL,ALL
MSHAPE,1,3D
MSHKEY,0
!*  
FLST,5,2,6,ORDE,2   
FITEM,5,1   
FITEM,5,-2  
CM,_Y,VOLU  
VSEL, , , ,P51X
CM,_Y1,VOLU
CHKMSH,'VOLU'   
CMSEL,S,_Y  
!*  
VMESH,_Y1   
!*  
CMDELE,_Y   
CMDELE,_Y1  
CMDELE,_Y2  

asel,s,ext   !无限远单元
cm,a_air,area
cmsel,s,a_air
extopt,esize,1
type,3
mat,1
FLST,2,6,5,ORDE,2   
FITEM,2,7  
FITEM,2,-12
VEXT,P51X, , ,0.001,0.001,0.001,,,,
numcmp,all
aslv
cmsel,u,a_air
nsla,s,1
sf,all,inf

ALLSEL,ALL

ESEL,S,MAT,,2    !定义线圈组件
CM,COIL,ELEM
ALLSEL,ALL  

WPAVE,0,0,0
WPCSYS,-1,0
CSYS,0  
LOCAL,16,1,0,0,0.1+(0.03/2), , , ,1,1,  

VSEL,S,MAT,,2   
VSUM, DEFAULT   
*GET,v,VOLU, ,VOLU

ALLSEL,ALL  
R,1,(0.02-0.01)*0.03,500,v,0,1,0,  
RMORE,1,0.5,  
ESEL,S,MAT,,2     !为线圈单元赋予实常数
EMODIF,ALL,REAL,1,
EMODIF,ALL,ESYS,16,
numcmp,all
!----------------------耦合线圈单元curr和emf自由度
ESEL,S,MAT,,2    !耦合线圈单元结点的电流自由度
NSLE,S

CP,1,CURR,ALL  
CP,2,EMF,ALL
*Get,NodeIK,Node,0,Num,Max !用于124线圈单元的IK节点

!-----------------------建立电路---------------------
Allsel,ALL
CSYS,0
ET,4,CIRCU124,5   !Stranded Coil(To FEM domain);!线圈单元
ET,5,CIRCU124,4   !独立电压源
R,4,1             !2号实常数Symmetry factor for stranded coil线圈对称系数
R,5,240,0,        !3号实常数 电压源240v相角0度
N                 !建立3个点，用于电路节点
N  
N              
*Get,Nmax,Node,0,Num,Max !获得刚建立的两个点的节点号
Type,5
Real,5
E,Nmax,Nmax-1,Nmax-2     !建立独立电压源;
Type,4
Real,4
E,Nmax,Nmax-1,NodeIK   !建立124单元类型的线圈并关联到solid97线圈单元上的一个点
D,Nmax-1,volt,0
FINISH
/SOL
ALLSEL,ALL
ANTYPE,3!谐波分析
HARFRQ,50,
SOLVE   
FINISH
!ANTYPE,TRANS
!TIME,0.02
!DELTIM,0.001
!Allsel,ALL
!OUTRES,ALL,ALL,  
!OUTPR,BASIC,ALL,
!SOLVE
!FINISH
/POST1
ESEL,S,MAT,,2     
NSLE,S
*Get,Nmax,Node,0,Num,Max  
ETABLE,R,NMISC, 16  
ETABLE,L,NMISC, 17  
SSUM
*GET,R,SSUM, ,ITEM,R
*GET,L,SSUM, ,ITEM,L
SET,1,LAST,1,0, , ,
VSEL,S,MAT,,6   
*GET,v,VOLU, ,VOLU
*GET,cureal,NODE,Nmax,CURR,
SET,1,LAST,1,1, , ,
*GET,curimg,NODE,Nmax,CURR,
I=(cureal**2+curimg**2)**(1/2)/sqrt(2)!电流有效值
Pr=I**2*R
ESEL,S,MAT,,2
Powerh
!这个是我建立的小模型，谢谢
46# jan_2000

sugars1985

有个疑问，用solid97载压线圈，进行谐波分析，最终用powerh算出来的powerloss 只计算了实部，即 PAVG= Ireal^2*R   (Ireal为电流实部幅值除以sqrt(2))
我用bfe 和circu124单元进行加载，结果都出现相同的问题。
而如果人工提取电流实部虚部，再求出I有效值，则I^2*R值不变，
而用Powerh算，只要电压加载的相角改变，则PAVG就会跟着变，而且严格符合PAVG= Ireal^2*R
46# jan_2000

sugars1985

理论上，电压的相角应该不影响线圈的损耗吧。

jan_2000

理论上，电压的相角应该不影响线圈的损耗吧。
sugars1985 发表于 2011-3-19 00:02

电压或者励磁电流的相角当然不影响线圈的损耗！我觉得是你的程序有问题，我算了一下，你的程序里，一个厚度为1厘米、高度为3厘米、平均半径1.5厘米的线圈，通240V电压后，居然铜耗高达7935W，这是绝对不可能的事情。请再反复核查你的程序！

jan_2000

/PREP7  

CYLIND,0.01,0.02,0.1,0.13   !建立一个圆筒作为线圈
BLOCK,-0.04,0.04,-0.04,0.04,0.05,0.15
VOVLAP,ALL  
NUMCMP,VOLU
ET,1,97   !除线圈外的其它实体单元类型
KEYOPT,1,1,1
ET,2,97   !soli ...
sugars1985 发表于 2011-3-18 23:58

我刚才又算了一下，大致明白了ANSYS中POWERH计算铜耗的原理：
是利用，计算得到的电流有效值*电压有效值*COS(PHI)，其中COS(PHI)为功率因数。也不对，这样算出来结果也和程序计算结果不一致。ANSYS电阻和电感的计算原理，我尚不得知，看来还得探究一下其计算原理。

jan_2000

47# sugars1985

但令人困惑的是，你的例子中这么简单的情况，线圈的电感是不应该变化的，因为线圈匝数和磁场磁阻未变。但我算了电压相位为0和60度角的两种情况，得到的电感分别为 1.16mH 和 2.08mH，这也是不正常的。这样一来，电感变化，功率因数也将改变，所以POWERH计算的铜耗也在变化。因此，我判断，要么是你的程序有瑕疵，要么ANSYS的97单元在这方面基于的有关理论有问题。

jan_2000

这个是ansys软件中powerh宏文件内部所存在的bug，只要在POWERH宏文件里边加一条语句就可以解决你的问题。但是这样可能会破坏POWERH宏文件解决其它一般类问题。不过从你的回复，可以看出你已经找到问题所在了。
49# ...
fossiler 发表于 2011-3-19 21:25

高手就是厉害，一语道破天机。

sugars1985

非常感谢，我只是个初学者，不敢轻易怀疑ansys是否纯真bug，听您这么一说，我觉得的确有可能是powerh本身的问题，不知道能不能说说，应该怎样改这个宏呢？ 53# fossiler