Saturday, July 31, 2010

PHY 103 N Week 1.


Lecture 1. ( PLEASE ALSO READ COMMENTS ON THE FIRST WELCOME POST).
In this week we first learnt about scalars, vectors and their transformation properties. Typically a quantity is a vector under a set of coordinate transformations provided its components transforms in a specific way under that set of transformations.The components in the transformed coordinate system are functions of the components in the old coordinate system. Like in rotations and inversions. We will be mostly concerned with vectors under rotations.
*Actually the word "direction" is dependent on which coordinate system one uses. The transformation law is a more accurate specification of the 'direction" of a vector which is independent of coordinate system in use.
* Tensors are objects with more components and may be described as products of vectors. Like rank 2 tensor is like a product of two vectors having 9 components in three dimensions all of which may not be independent. Scalars are rank 0 and vectors are rank 1 tensors. Tensors transform like products of vectors. So a rank 2 tensor under rotations will transform as the product of two rotation matrices under rotations.
* A pesudovector transforms just like a vector under rotations but with a relative minus sign to vectors under inversions. Simmilarly for pesduscalars and pseudo tensors.
Lecture 2
-----------
Introduced scalar fields and vector fields as scalar and vector functions of the co-ordinates. A scalar transforms as φ'(x', y', z')=φ ( x, y, z). Note that the function on RHS is different from the function on the LHS but the old function evaluated in the old coordinates is equal to the numerical value of the new function in the new co-ordnate system.A vector field has three components ( in 3 dimensions) each of which transforms like a scalar field.
The ∇ or the gradient operator has been introduced. This is a vector operator
such that under rotations its components which are ∇ = i ∂ / ∂ x + j ∂ / ∂ y + k ∂ / ∂ z which transform like a vector . Note that although it has no magnitude or direction its components transform like a vector under rotations.
The gradient of a scalar field ∇ φ ( x,y,z) is a vector field. It provides the direction of maximal change in φ. Constant φ surfaces are called level surfaces and the direction of grad φ is always normal to the level surface.
Lecture 3.
----------
Introduced the divergence of a vector field ∇ . v ( x, y, z) given by the dot product of the grad operator with a vector field. The divergence of a vector field provides a measure of the "spread" or relative increase or decrease in "magnitude" around that point. If v is a fluid flow velocity vector then div v positive signifies a "tap" and negative signifies a "sink" and zero signifies a "laminar flow".
Introduced the Curl of a vector field ∇ x v( x, y, z) given by the cross product of the grad operator with a vector field v. The result is another vector field. The curl of a vector field at a point provides a measure of how much the direction of the vector field curls or wraps around the point in question. Curl can be non zero also due to the relative magnitudes being different in different
points around the point in question. For a fluid flow the non zero curl is the regions where a "whirlpool" or a "vortex" appears.
We also learnt about other products and vector identities with the grad operator. Please NOTE that some of the vector identities are very different for the grad operator. This is because the grad apart from being a vector is also a differential operator.
We also introduced the scalar operator ∇ 2 which is the sum of the second partial derivaties with respect to x, y, z. The Laplacian can act both on scalar and vector fields. On a vector field v the Laplacian acts on each component of the vector field. ( remember each component of a vector field is a scalar field itself. so it is the sum of the Laplacian acting on the 3 scalar fields which are the components of the vector field v.

25 comments:

  1. Some people had a little problem with inversion transformation and cross product of the unit vectors
    i, j , k. The solution to the problem is as follows.
    There are two ways in writing -vector A one by writing the components as - the original components
    or inverting the unit vectors. Doing both of them is not allowed. So best way is to use - of components and keep unit vectors as i , j and k
    NOT -i, -j, -k.

    ReplyDelete
  2. sir could u also give questions other than that in griffiths in ipsa and dipa?

    ReplyDelete
  3. You have to be patient because this is the first problem set. Also Griffiths is the standard text for the course so all the problems will be either
    from Griffith or roughly at the level of Griffiths
    which is the level of First Year Freshman Electromagnetic Theory course. You also need to focus on the concepts because those are your key to solving the problems.

    ReplyDelete
  4. I FIND IT DIFFICULT TO RELATE THE MATRICES WITH THE TRASFORMATION IT SEEMS TO BE VERY ABSTRACT . COULD YOU PLZ ELABORATE THAT PART.

    ReplyDelete
  5. sir, this is wrt to inversion....

    when we invert a system of co-ordinate axes, then are we not changing our system to "left handed" co-ordinate system??

    is that the reason why (-i x -j) turns out to be -k (i,j,k are unit vectors in the right handed co-ordinate system)....

    ReplyDelete
  6. sir,
    could u please make available soft copies of ipsa and dipa.

    ReplyDelete
  7. For the 2 dim rotation I have derived this in the class and its also worked out in Griffiths. In the Matrix form each vector is specified by a column
    of components called a column vector. This is also a representation of a vector by an "ordered triple".
    The components in the new co-ordinate system is given by the matrix multiplication of the transformation matrix with the components in the old co-ordinate system written in the "column vetor" form.

    Each such transformation is specified by a matrix
    where theta specifies the amount of rotation. For 3 dimensional rotation will have to be specified by 3 such angles.

    If its still not clear please see me.

    ReplyDelete
  8. For inversion transformation please see my first comment. Indeed the co-ordinate system will be left handed. But its easier to just invert the components. So that you still work with the old co-ordinate system i, j, k.

    ReplyDelete
  9. Soft copies of IPSA and DIPA will be available from the next problem set.

    ReplyDelete
  10. sir i can't get any physical significance of tensors

    ReplyDelete
  11. Mathematically it behaves like the product of vectors ( its called an outer product, this you will
    know later). Physically for you its a little hard to grasp because you have not used it in Physics yet. You will soon know in your next semester PHY 102 N that Moment of Inertia is a rank 2 tensor.
    In Physics usually tensors come in when physical quantities become direction dependent. Electrical conductivity for certain medium is a rank two tensor. Stress in Elasticity is a tensor, strain is also a tensor. For the time being its best to understand it mathematically in terms of its transformation.

    ReplyDelete
  12. SIR
    IN CASE OF INVERSION WE SAY THAT THE VECTOR A TRANSFORMS TO -A AND B TO -B, BUT IF C = A * B ,TAKING CROSS PRODUCT GIVES US C WHICH IS ALSO A VECTOR FROM OUR BASIC VECTOR ALGEBRA , SO C SHOULD ALSO TRANSFORM TO -C UNDER INVERSION , BUT IT DOES NOT HAPPEN AND UNDER INVERSION C REMAINS C AS -A * -B = A*B , AND WE SAY THAT C IS A PSEUDO VECTOR BUT ITS NOT CONVINCING FOR ME , ARE THE THINGS RELATED TO VECTORS WHICH WE STUDIED EARLIER INCORRECT , PLEASE ELABORATE

    ReplyDelete
  13. sir i am feeling great problem while using opertaros. what should i do?

    ReplyDelete
  14. No nothing that you had studied was incorrect. Its just that the whole story was not told to you. So you did not know the difference between a vector and pseudovector earlier. This did not create a problem because pseduvectors behave just like vectors under rotations and satisfies all the properties of vectors. Its only under Inversion which is completely different class of transformations than rotations that the pseuduvectors behave differently.

    If you are still not convinced please talk to me
    after the lectures.

    ReplyDelete
  15. Dont worry Operators are new objects and simply you are not used to them. There are problems in DIPA and IPSA and you should also practice from Griffiths.

    All you have to remember when using operators is that they are differential operators and satisfy
    the rules of differential operators like d/dx when they operate on sums and products of functions.

    Apart from this operators also have a transformation characteristic ( under rotations etc.) they may transform as a Scalar like the Laplacian operator or like a vector example the Grad operator. ( There are also Tensor Operators
    but its beyond our scope for this course.)

    ReplyDelete
  16. There were two other comments which somehow I have not been able to publish on the blog. Some error whilst publishing in the blog software.

    Both students one who wrote about Freshers Intro. and one who wrote about difficulties with transformations are advised to see me to discuss
    after the Lectures or DIPA/IPSA Sessions. In DIPA I
    am in L3 and in IPSA i am in TB 105 with Section F5

    ReplyDelete
  17. OK i have found both the missing comments they are in the first post on Welcome to the Blog. I have replied to them also.

    ReplyDelete
  18. Sir, from where iI can check the answers of the questions I solved from GRIFFITH.

    ReplyDelete
  19. Giriffiths solution set are possibly available either online or in the students network DC++.

    ReplyDelete
  20. at the dipa session we were told that if grad(f) is zero it could be a saddle point..that is it could be constant in one direction and changing in another. I couldn't actually get that meaning of saddle because if grad(f) is zero that means the change should not be there. i tried to understaqnd it by inflection point but couldn't actually get it

    ReplyDelete
  21. in case of electric field due to a point charge the field shows zero divergence except at origin. i tried to figure out the reason that though the field is diverging its magnitude is decreasing hence effect of each other is cancelled. is this correct?

    ReplyDelete
  22. No it is not correct. The entire contribution to the divergence comes from the origin but it is actually infinite. When I will introduce the Dirac delta function this would take care of this problem.
    So please wait a little for me to introduce that.

    ReplyDelete
  23. Yes it is correct about the grad f being zero at a
    saddle point. For one dimension its just a point of inflection eg y=x cubed. For 2 and 3 dimensions
    the number of derivatives are more like del f del x,
    del f del y and so on. A saddle point is like the generalization of the point of inflection in 1 dimension. But in more than 1 dimension the derivatives are direction dependent as the function may be constant somewhere but either increasing or decreasing in some dierction.

    Check these links for visualization

    http://en.wikipedia.org/wiki/Saddle_point

    http://mathworld.wolfram.com/SaddlePoint.html

    http://www.ucl.ac.uk/Mathematics/geomath/level2/pdiff/pd9.html

    ReplyDelete
  24. Sir i wanted to know the solution of the question in griffith that asks to show that the divergence of a vector v in two dimension acts like a scalar

    ReplyDelete
  25. Divergence is always a scalar under transformations.
    To prove this you have to show that the scalar field
    resulting from Div of a vector field V transforms like a scalar field. (This is not important for the exams and its a little tricky to show).

    ReplyDelete