This is a fantastic course for anyone interested in Computational Electromagnetics. I am a graduate student in EE and I work professionally as an RF Engineer. It is very difficult to find quality and comprehensible resources as an introduction to this topic. I believe this is because it requires a relative knowledge of software development, electronics, and of coure electromagnetics. However, I am extremely grateful for you sharing this course online. I have been taking this course (metaphorically speaking) in parallel with my Microwave Circuit Analysis and Design graduate course.
That is awesome to hear! This specific video is the first I ever recorded. It was many years ago! Over time, I have added a ton of content and revised much of it. I actually recommend accessing the notes, videos, and other resources through the course website because it contains the latest versions of all this. empossible.net/emp5337/ Thank you so much for the recognition!!! I love hearing how the courses help people.
Thanks for the tutorial! ! I am a new scholar to CEM. My master's degree focuses on theoretical optics. Now I want to try to transfer the direction to CEM as PHD direction, such as electromagnetic information direction, etc. I was fascinated by the combination of explicit and clear mathematical calculations by computers to achieve control of fields. Can I get any advice from you about the path I've chosen?
This is great to hear and I do have lots of advice. This is something I have written about somewhat extensively. I recently started a blog where I have written about topics in EM and also some career related stuff. Here is a link to the blog: empossible.net/blog/ For getting started in computational electromagnetics (CEM), I recommend two posts although some others definitely apply: Learning Electromagnetics: Where to Start empossible.net/electromagnetics-getting-started/ Getting Started in Numerical Methods empossible.net/getting-started-in-numerical-methods/ In addition, I recently wrote a book titled "Electromagnetic and Photonic Simulation for the Beginner: Finite-Difference Frequency-Domain in MATLAB." The book is intended for the complete beginner to get started in CEM. It teaches the art of CEM through finite-difference frequency-domain (FDFD), which I think is the best FIRST method to learn in CEM. It is incredibly versatile and is one of the easiest to learn and implement. You can learn almost all of the concepts and best practices in CEM with a very simple method that will help you learn and test other methods if you ever want to do that. Here is the book website: empossible.net/fdfdbook/ I think a good SECOND method to learn is finite-difference time-domain (FDTD). For this, I created online resources. While also intended for the complete beginner, the online courses are not as comprehensive about teaching the art of CEM. Here is a link to those courses: empossible.thinkific.com/collections/FDTD-in-MATLAB With both FDFD and FDTD, there is literally nothing you will not be able to simulate. Any other method will only have the advantage of being faster and/or more memory efficient. However, they will typically be much more difficult to learn and implement. Hope something here helps!
Excellent lectures, Thank you for making these public. After using commercial programs a great deal I'm learning about how the the methods work hoping that I can write some 'simple' solvers for specific problems in either c++ or Octave. What's your opinion on the Mode Matching technique? I know its very quick although less arbitrary to say, FEM or FDTD. However, it's proved difficult to find books or information on the subject.
I agree it is difficult or impossible to find good starter books. For mode matching, you can work through TMM, RCWA, and MoL in this lecture series. MoL is a numerical mode matching technique. It can be speeded up considerably if analytical solutions to the modes are known. Here is a link to the course website that has the latest version of the notes, links to the UA-cam videos, benchmarking documents, and other resources to help you get your codes working. emlab.utep.edu/ee5390cem.htm
CEM Lectures, Thank you! I look forward to seeing these videos and notes. I wish we had a course like this when I was an undergraduate, your students are very lucky to have such good tuition presented in a lucid fashion with good visuals to demonstrate concepts.
Thank you!! Be sure to visit the official course website, which has links to the videos, the latest version of the notes, and other resources to help you get your codes working. emlab.utep.edu/ee5390cem.htm
Dear Dr. Raymond Rumpf, I have a question about the classification by size scale. Why can we analyze the phenomenon using Ray Optics when the structural dimensions are much larger than the wavelength? and why it is not available when the dimensions are on the order of the wavelength? I know the fact that the dimension relationship lets us know what nature of light will dominate.(Wave or Ray, vector field or scalar quantity) But, It is difficult to understand intuitively and physically why we can think like this. Can I have more basic and detailed explanation about this? Please understand my question even it is the too basic concept. I hope to see your reply. Thanks.
Rather than overall device size, make the classification based on the smallest feature size of a device. For example, if we put some crazy nanostructures onto the surface of a lens, that lens will have a polarization response and traditional ray tracing cannot be used. As just a lens, ray tracing would work fine. When feature sizes are much larger than a wavelength, we are using refraction which is wavelength independent and polarization independent, thus ray tracing can be used.
+Nikhil Dhingra See Lectures 26 and 27 here: emlab.utep.edu/ee5390cem.htm Unfortunately, FEM needs to be a semester course, like finite-difference time-domain. This course is on my radar to create some day, but I have not done it yet. Very sorry!
(1) basic programming (2) basic knowledge of MATLAB (3) basic knowledge of electromagnetics (4) basic knowledge of differential equations, vector calculus, and linear algebra. I would not worry about having an expert level of knowledge of any of these. Before taking this course, I recommend working through Topics 6 and 7 in my Computational Methods course to learn the finite-difference method. empossible.net/academics/emp4301_5301/ You may also be interested in a book I just published to help beginners get started in CEM: empossible.net/fdfdbook/ I also have an online course to teach finite-difference time-domain to the complete beginner: empossible.thinkific.com/collections/FDTD-in-MATLAB I hope something here helps! Good luck and have fun!
I usually handle this in class when I work through some codes with the students. I know I do have a slide or two somewhere, but I cannot seem to find it. The short story on convergence is that you keep increasing the number of points on your grid until your answer does not change much any more. You can plot your answer as a function of grid resolution and you will see it follow a random-ish looking exponential curve that approaches some asymptote. In principle, you can make your calculation as precise as you want, but at the cost of simulation time. That really is the trade off. We want fast simulations and we also want converged simulations and the two constantly fight each other.
You got it! I posted them to the course website along with other resources to help you. The electronic notes have a lot of corrections, revisions, and a content additions since they were last recorded so expect some differences between the notes and videos. emlab.utep.edu/ee5390cem.htm
Unfortunately, there really isn't one. I could recommend a book maybe depending on the numerical method you are interested in. Methods like finite-difference time-domain are covered well in textbooks. Other methods I don't think are covered well anywhere. What method are you interested in learning more?
The lectures are no longer named this way. I am not sure what 17 has become. I recommend accessing the videos through the course website because they link to the latest versions of the notes and videos and also have other learning resources for you. Here is the website: empossible.net/emp5337/ Let me know if you find what you are looking for.
There is very little information on this channel about signal processing. The focus is more on electromagnetics and computation. There are definitely many overlapping concepts, however.
There is quite a bit on UA-cam for non-electromagnetic problems. Otherwise, seek textbooks. I wish I had a better answer for you! Some day I plan on developing an entire course on variational methods like FEM, method of moments, spectral domain, etc. I doubt that will happen any time soon.
This is a fantastic course for anyone interested in Computational Electromagnetics. I am a graduate student in EE and I work professionally as an RF Engineer. It is very difficult to find quality and comprehensible resources as an introduction to this topic. I believe this is because it requires a relative knowledge of software development, electronics, and of coure electromagnetics. However, I am extremely grateful for you sharing this course online. I have been taking this course (metaphorically speaking) in parallel with my Microwave Circuit Analysis and Design graduate course.
That is awesome to hear! This specific video is the first I ever recorded. It was many years ago! Over time, I have added a ton of content and revised much of it. I actually recommend accessing the notes, videos, and other resources through the course website because it contains the latest versions of all this.
empossible.net/emp5337/
Thank you so much for the recognition!!! I love hearing how the courses help people.
Thank for SOOOOOO much for sharing these great education resources!!!!!!!!!!!! I really love your teaching style!
谢谢！
Hi Wade!
Thanks for the tutorial! ! I am a new scholar to CEM. My master's degree focuses on theoretical optics. Now I want to try to transfer the direction to CEM as PHD direction, such as electromagnetic information direction, etc. I was fascinated by the combination of explicit and clear mathematical calculations by computers to achieve control of fields. Can I get any advice from you about the path I've chosen?
This is great to hear and I do have lots of advice. This is something I have written about somewhat extensively. I recently started a blog where I have written about topics in EM and also some career related stuff. Here is a link to the blog:
empossible.net/blog/
For getting started in computational electromagnetics (CEM), I recommend two posts although some others definitely apply:
Learning Electromagnetics: Where to Start
empossible.net/electromagnetics-getting-started/
Getting Started in Numerical Methods
empossible.net/getting-started-in-numerical-methods/
In addition, I recently wrote a book titled "Electromagnetic and Photonic Simulation for the Beginner: Finite-Difference Frequency-Domain in MATLAB." The book is intended for the complete beginner to get started in CEM. It teaches the art of CEM through finite-difference frequency-domain (FDFD), which I think is the best FIRST method to learn in CEM. It is incredibly versatile and is one of the easiest to learn and implement. You can learn almost all of the concepts and best practices in CEM with a very simple method that will help you learn and test other methods if you ever want to do that. Here is the book website:
empossible.net/fdfdbook/
I think a good SECOND method to learn is finite-difference time-domain (FDTD). For this, I created online resources. While also intended for the complete beginner, the online courses are not as comprehensive about teaching the art of CEM. Here is a link to those courses:
empossible.thinkific.com/collections/FDTD-in-MATLAB
With both FDFD and FDTD, there is literally nothing you will not be able to simulate. Any other method will only have the advantage of being faster and/or more memory efficient. However, they will typically be much more difficult to learn and implement.
Hope something here helps!
Very kind advices and these really helps me! Thank you@@empossible1577
Excellent lectures, Thank you for making these public. After using commercial programs a great deal I'm learning about how the the methods work hoping that I can write some 'simple' solvers for specific problems in either c++ or Octave. What's your opinion on the Mode Matching technique? I know its very quick although less arbitrary to say, FEM or FDTD. However, it's proved difficult to find books or information on the subject.
I agree it is difficult or impossible to find good starter books. For mode matching, you can work through TMM, RCWA, and MoL in this lecture series. MoL is a numerical mode matching technique. It can be speeded up considerably if analytical solutions to the modes are known. Here is a link to the course website that has the latest version of the notes, links to the UA-cam videos, benchmarking documents, and other resources to help you get your codes working.
emlab.utep.edu/ee5390cem.htm
CEM Lectures, Thank you! I look forward to seeing these videos and notes. I wish we had a course like this when I was an undergraduate, your students are very lucky to have such good tuition presented in a lucid fashion with good visuals to demonstrate concepts.
Thank you!! Be sure to visit the official course website, which has links to the videos, the latest version of the notes, and other resources to help you get your codes working.
emlab.utep.edu/ee5390cem.htm
Thanks so much !
Dear Dr. Raymond Rumpf,
I have a question about the classification by size scale.
Why can we analyze the phenomenon using Ray Optics when the structural dimensions are much larger than the wavelength? and why it is not available when the dimensions are on the order of the wavelength?
I know the fact that the dimension relationship lets us know what nature of light will dominate.(Wave or Ray, vector field or scalar quantity)
But, It is difficult to understand intuitively and physically why we can think like this.
Can I have more basic and detailed explanation about this?
Please understand my question even it is the too basic concept.
I hope to see your reply.
Thanks.
Rather than overall device size, make the classification based on the smallest feature size of a device. For example, if we put some crazy nanostructures onto the surface of a lens, that lens will have a polarization response and traditional ray tracing cannot be used. As just a lens, ray tracing would work fine. When feature sizes are much larger than a wavelength, we are using refraction which is wavelength independent and polarization independent, thus ray tracing can be used.
Please add some lectures on Finite Element Method with emphasis on Electromagnetics.
+Nikhil Dhingra See Lectures 26 and 27 here:
emlab.utep.edu/ee5390cem.htm
Unfortunately, FEM needs to be a semester course, like finite-difference time-domain. This course is on my radar to create some day, but I have not done it yet. Very sorry!
What you have drawn on white/black board at slide 38
Thanks Professor
Does anyone know where to find the book: computational methods for electromagnetics/ Peterson?
What are the prerequisites for this course sir
(1) basic programming
(2) basic knowledge of MATLAB
(3) basic knowledge of electromagnetics
(4) basic knowledge of differential equations, vector calculus, and linear algebra.
I would not worry about having an expert level of knowledge of any of these. Before taking this course, I recommend working through Topics 6 and 7 in my Computational Methods course to learn the finite-difference method.
empossible.net/academics/emp4301_5301/
You may also be interested in a book I just published to help beginners get started in CEM:
empossible.net/fdfdbook/
I also have an online course to teach finite-difference time-domain to the complete beginner:
empossible.thinkific.com/collections/FDTD-in-MATLAB
I hope something here helps! Good luck and have fun!
@@empossible1577 Thank you very much, dear sir
you mentioned in lecture 13 you will illustrate convergence in latter lecture which number is it ,thanks
I usually handle this in class when I work through some codes with the students. I know I do have a slide or two somewhere, but I cannot seem to find it.
The short story on convergence is that you keep increasing the number of points on your grid until your answer does not change much any more. You can plot your answer as a function of grid resolution and you will see it follow a random-ish looking exponential curve that approaches some asymptote. In principle, you can make your calculation as precise as you want, but at the cost of simulation time. That really is the trade off. We want fast simulations and we also want converged simulations and the two constantly fight each other.
Would you mind to post your slides as a PDF ?
You got it! I posted them to the course website along with other resources to help you. The electronic notes have a lot of corrections, revisions, and a content additions since they were last recorded so expect some differences between the notes and videos.
emlab.utep.edu/ee5390cem.htm
+CEM Lectures Thank you very much !!! :) :)
Hi.. Dorian!!!!
Could I get pdf copy for this course as well if you still have it?
I have started this course. I would be thankful to you.
Email Id- paragparashar03@gmail.com
what are the reference books that you recommend for this course
Unfortunately, there really isn't one. I could recommend a book maybe depending on the numerical method you are interested in. Methods like finite-difference time-domain are covered well in textbooks. Other methods I don't think are covered well anywhere. What method are you interested in learning more?
Why is there no Lecture 17?
The lectures are no longer named this way. I am not sure what 17 has become. I recommend accessing the videos through the course website because they link to the latest versions of the notes and videos and also have other learning resources for you. Here is the website:
empossible.net/emp5337/
Let me know if you find what you are looking for.
Does CEM drive to signal processing
There is very little information on this channel about signal processing. The focus is more on electromagnetics and computation. There are definitely many overlapping concepts, however.
CEM Lectures thanx
👍4️⃣ 2️⃣0️⃣
HI WADE
"Letcure 27" is not working in the link
emlab.utep.edu/ee5390cem.htm
This lecture actually has not been recorded yet. This is mostly because the lecture notes themselves are still under development. Sorry!!!!
Actually, I need lectures on FEM. So, can you suggest any material for that?
There is quite a bit on UA-cam for non-electromagnetic problems. Otherwise, seek textbooks. I wish I had a better answer for you! Some day I plan on developing an entire course on variational methods like FEM, method of moments, spectral domain, etc. I doubt that will happen any time soon.
Deer sir give me those slides
You got it...here are the slides to the entire course:
empossible.net/emp5337/
HI WADEEEE
Very poor Audio quality :(
Very sorry. This was the first-ever recording I did. The audio gets better in other videos.
Nice man you are. :)
Hi Wade!