Ivar KJELBERG
                                                                                                                                                    COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
                                                         
                            
                         
                                                
    
        Please login with a confirmed email address before reporting spam
     
    
 
                                                Posted:
                            
                                2 decades ago                            
                            
                                Jul 13, 2009, 2:03 a.m. EDT                            
                        
                        
                                                    Hello
Well your 2D object has a certain "thickness" d  see  the Subdomain Settings, and the boundary a ceratain length, for me this is your "Area". You have to keep your physics consistent.
If you do not know your Boundary length, use COMSOL to calculate it by selecting the boundary and a "Options - Integration Coupling Variable - Boundary Variables" and give a name such as "Length_1" and integrate the value "1" along ds, this gives you the total length of the boundary you have selected.
Then you can define your current as a Con stant as  "I_1"   "2[mA]" and you use "I_1/(Length_1*d_emdc)" as your current denisty entry. now if you are using another physics adapt the name of it "d_emc" or "d_emes" or "d_emqa" ...  You may also give directly the current denisty when you know it, but do not forget to change it too if you change the thickness or boundaries of your model.
Now usig the "Port" approach, I understand that COMSOL looks after some of these variable and surface calculations for you. I havnt used them enough myself to really understand your question. And I have no "current denisty on my "ports". "Ports" was introduced one or a few versions ago, to speed up and minimise the formulas you need to type.
I suggest you run a few of the examples, re-read the doc, attend a training course, and always try a few simple cases you know and can analyse by hand, and if not sufficient then be more explicit in your questions. 
Good luck                                                
                                                
                            Hello
Well your 2D object has a certain "thickness" d  see  the Subdomain Settings, and the boundary a ceratain length, for me this is your "Area". You have to keep your physics consistent.
If you do not know your Boundary length, use COMSOL to calculate it by selecting the boundary and a "Options - Integration Coupling Variable - Boundary Variables" and give a name such as "Length_1" and integrate the value "1" along ds, this gives you the total length of the boundary you have selected.
Then you can define your current as a Con stant as  "I_1"   "2[mA]" and you use "I_1/(Length_1*d_emdc)" as your current denisty entry. now if you are using another physics adapt the name of it "d_emc" or "d_emes" or "d_emqa" ...  You may also give directly the current denisty when you know it, but do not forget to change it too if you change the thickness or boundaries of your model.
Now usig the "Port" approach, I understand that COMSOL looks after some of these variable and surface calculations for you. I havnt used them enough myself to really understand your question. And I have no "current denisty on my "ports". "Ports" was introduced one or a few versions ago, to speed up and minimise the formulas you need to type.
I suggest you run a few of the examples, re-read the doc, attend a training course, and always try a few simple cases you know and can analyse by hand, and if not sufficient then be more explicit in your questions. 
Good luck                        
                                                
                                                                                                            
                                             
                        
                        
                                                
    
        Please login with a confirmed email address before reporting spam
     
    
 
                                                Posted:
                            
                                2 decades ago                            
                            
                                Jul 19, 2009, 10:14 p.m. EDT                            
                        
                        
                                                    Thanks a lot for the comments                                                
                                                
                            Thanks a lot for the comments