Transcript for:
Dimensional Analysis Tutorial

going to give you a little tutorial on dimensional  analysis for years i thought dimensional analysis   was this strange art but then i i later realized  wait there's a systematic way to do it and   i was overjoyed for most physical systems  there's a limited number of basic dimensions   and these basic dimensions are mass and  i'll use your capital m to represent mass   there's length right distances can  be written in terms of a length   speeds are length per time so oh  time that's another basic dimension   that's kind of it you can also think of  temperature i guess but other things like force   are just based on these basic dimensions so for  example force is mass length per time squared so   it could be decomposed into these basic dimensions  there's typically three basic dimensions i might   just add that sometimes temperature is included  we're thinking of things that come from science   and physics you might say well what about dollars  for studying a financial system as this forced   example tells us other system properties can be  given in terms of these basic ones like force   and i will use this symbol of square brackets  to denote dimension so for example dimension of   force our m mass times length over time  squared what about dimensions of the velocity   of something that's going to be length per unit  time or the acceleration that's how the velocity   changes with time so it's going to be length over  time squared and so on so for material physical   things there will be these basic dimensions we  haven't yet said anything about units there's two   primary systems of units there's the english one  call it english but we seem to only use it here   like measuring force in pounds mass and slugs  then there's the s i system of units we need to   distinguish units and dimensions i'm just going  to talk about the main system of units to be   distinguished from dimension and that's the system  international s i units we sometimes just sort of   reverse it international system here we measure  mass in kilograms we measure length in meters we   measure time and seconds we measure degrees and  kelvin all the other units are derived from those   derived units that we attach people's  names to sometimes like newtons one newton   equals one kilogram meter second squared  it's a derived unit from these basic three   which match these basic three dimensions an  equation like the equation that we have up above   has this equation up here this equation of motion  needs to be dimensionally homogeneous meaning   all of the terms must have the same dimension  maybe i'll put a equation one next to that   so an equation describing the physical system  must be dimensionally homogeneous for example   equation one up there the equation of motion for  the bead in the hoop so if we would write that   as left-hand side equals a right-hand side if  you take the units of the left-hand side that   must equal the units of the right-hand side this  denotes i keep saying units denotes the dimensions   the length of my car has dimensions of length the  length of the desk in front of me dimensions of   length the mass of that chair back there that's  dimensions of mass usually these are combined and   that's why it's not always totally obvious if we  look up here if you were to analyze the dimensions   all the terms have the same dimension but  we'll get to that i first want to mention   something called the buckingham pi theorem  this buckingham pi theorem it's useful in   trying to non-dimensionalize an equation  if we have an equation like this it's   chock full of dimensions mass and length if  you want to systematically study something   you want to non-dimensionalize that equation  so you can figure out which combinations of   parameters are really the most important so what  you'll end up with is an equation that has no   dimensions to it but its behavior will be  described by a few non-dimensional parameters   as well as non-dimensional variables and a  non-dimensional time so everything becomes   non-dimensionalized so instead of there being  theta and time and five parameters maybe we'll   have something less when we non-dimensionalize  this equation of motion toward that end   we consult the buckingham pi theorem for help with  non-dimensionalizing we consult the buckingham pi   theorem here's the thing if an equation has k  variable and here they use variables differently   from me variables that just means any kind of  number this lumps together what i had called   variables as in things that vary with time as  well as parameters if an equation has k and k   is going to be a integer then it can be reduced  to a relationship among k minus r independent   dimensionless or non-dimensionalized products or  groups or as we might say numbers non-dimensional   number where r it's an integer it's the minimum  number of basic dimensions required to describe   all the variables so there's something interesting  here there's this number k and this number k minus   r let's just rewrite what we have try on the  equation of motion for the bead so that was   equation one i'll just write it m r remember this  is different r equals minus b theta dot minus mg   sine theta plus m r omega squared sine theta  cosine theta if we call these buckingham pi   variables you know what it calls variables   what do we have let me just go through we've  got the five things i called parameters m r b g   omega but then we also have theta and t  according to the buckingham pi theorem   we have k equals seven variables now how many  basic dimensions do we need to describe all of   these to find that out we just sort of find out  the dimensions of everything dimension of mass   is of course m mass dimensions of the radius  of the hoop well that's a length you have to   do some work to find out what dimensions of b  is but because that equation is dimensionally   homogeneous it turns out that that damping term  has dimensions of mass times length over time   dimensions of the gravitational acceleration  has dimensions of acceleration so it's length   over time squared dimensions of the  angular velocity angular velocity is   technically the number of radians which is  dimensionless divided by time so it's rotation   rate is radians for time since radians  are dimensionless it's 1 and time is t   what about dimensions of the angle theta you  might be thinking well isn't that degrees or   something well no for this equation to be true  you've got to write the angle theta in terms of   radians and radians do not have dimension they're  not a mass they're not a length they're not a time   so they're already non-dimensionalized so the way  we represent that is we just sort of write one   and then dimensions of time are of course time  so if you look at how many things do we have here   what basic dimensions i mean we only mentioned  three do we use all three here we got mass we   got length we got time r equals three that's the  minimum number of basic dimensions to describe   all of these variables in the equation what does  the buckingham pi theorem tells it says that this   equation can be reduced to a relationship among  k minus r dimensionless numbers k minus r equals   four that means equation one can be turned into  a i know it's vague sounding like a relationship   among four dimension less numbers or products  groups so that means we need four dimensionless   numbers to non-dimensionalize this equation okay  we're going to go through this exercise you can   see and you actually have choices in what these  dimensionless groups are this is an equation of   motion for theta let's put a little square  one theta the main thing that we care about   the main thing that varies with time is already  dimensionless that's good maybe we should stick   with that theta is one of them so now there's  only three left what else do we have we could   look at this equation up here we see that mg this  has units of force forces mass length time squared   m r omega squared is also dimensions of  force we can take the ratio of these two   the ratio of a force divided by a force  is going to be a non-dimensional number   if we take the ratio and we can kind of pick  which one is the numerator which one's the   denominator i'll do this m r omega squared over  m g the m's cancel out we've got r omega squared   over g that's a non-dimensional number we'll call  this gamma that's how we'll define gamma gamma   is our second non-dimensional number and notice  something about gamma because r is positive g is   positive omega squared is positive so gamma is  greater than or equal to zero it could be zero   if we have a non-rotating hoop number three that's  why there's only two more we can find or introduce   a characteristic time scale this is standard  thing when you have a differential equation where   there's derivatives with respect to time once we  have a characteristic time scale hopefully this   won't lead to confusion but i'll call it capital  t then we can introduce a dimensionless time let's   call it tau it'll be normal time measured in  seconds divided by capital t which will also be   measured in seconds but then this dimensionless  time will be our third dimensionless number the   way that we can find a characteristic time  scale is we just we suppose there is one   and we rewrite the derivatives instead of being  d theta d little t we write them as d theta d tau   and plug them into the equation  what do i mean by that well theta   you could think of it as it's a function  of time but you could also think of it   it's a function of this new thing we've  introduced called the dimensionless time   which is a function of normal time in the equation  up above we've got theta dot which means what   that's just shorthand for d theta d t using the  chain rule we could write this as d theta d tau   d tau d t and what is d tau d t well from this  formula over here this is one over capital t one   over the time scale so that means wherever we have  the derivative with respect to the actual time we   could substitute in this one over t d theta detail  let's look up here we have a first derivative we   also have a second derivative over here if you  do this twice here's what you'll get theta double   dot equals 1 over d squared second derivative of  theta with respect to the non-dimensional time tau and so this d theta detail since theta and tau  are both dimensionless this is dimensionless   so it's second derivative  of theta with respect to tau   plug these two equations back into the  original ode so we've got r 1 over t squared   d squared theta d tau squared that was the  left hand side well i guess we had an m   and then minus b one over t d theta  d tau minus m g sine theta plus m r   omega squared sine theta cosine theta let's  divide everything by m so just divide through by m i want you to look at how we defined gamma up here  it's r omega squared divided by g and it looks   like we're close to getting that in this formula  if we divide everything by g so let's divide by g   and this goes away g r over g we put parentheses  around this and say oh this is gamma everything   is non-dimensional this is a non-dimensional  term this is a non-dimensional term this whole   thing is non-dimensional and because d theta  d tau is non-dimensional of this minus sign   we've got that this is  non-dimensional and so is this and now we have a choice we're going to pick a  characteristic time scale to make one of these   non-dimensional parameters equal to one this  problem's interesting in that there's a choice   sometimes in other equations one thing just pops  out here we have a choice of what to pick for   the characteristic time scale we want to make  one of these two terms order one but you means   we'll just set it equal to one if one of them is  set equal to one then the other one will be some   number it'll be some other non-dimensional number  that's not necessarily one what we're going to do   is pick a time scale that's related to the damping  parameter the friction b just because we have to   pick something and actually looking ahead i  know that that makes the problem interesting   we will choose b over mg t equal to one so that  the characteristic time scale is b over m g   that sets what the non-dimensional time  will be what did we pick we picked this to   be non-dimensional by a choice of t so that's in  some sense number three that was the third thing   the last one just sort of given to  us by this first term here in purple because for that choice of t what does  that become t get what we get r over g   t squared this ends up being r over  g mg over b squared it's got to be   non-dimensional and it is we'll call  it epsilon so epsilon is m squared g r   over b squared so this is our fourth  and final non-dimensional number   so now our equation of motion can be written in  its non-dimensionalized form i think we call that   equation one we did make a choice a choice of time  scale we might not say this is the non-dimensional   form it's just a non-dimensional form we've got  epsilon times this second order derivative term   equals uh because of our choice of time scale the  coefficient of the damping term is just one and   then there's also coefficient one in front of  this gravity term and then gamma um oh i guess   another thing to know is epsilon is greater than  or equal to zero as well gamma sine theta cosine   theta what are we left with we have theta changing  with the non-dimensional time tau so i'd say we   have one dependent variable theta one independent  variable tau and two parameters just two epsilon   and gamma and we have interpretations i think  epsilon is an inertial term epsilon gives the   relative importance of inertia to damping and  gamma is related to the relative importance of   relative importance of the rotational  effect to gravity so in this case these have interpretations before we go on and analyze this  there any other cases of non-dimensional numbers   that you've come across any famous ones reynolds  number yeah yeah the reynolds number like inertia   to viscosity fluids has lots of non-dimensional  numbers reynolds is the big one in fact i guess   you could think of reynolds as a bifurcation  parameter for low values of reynolds number it's   laminar flow and then at some point as you turn  the knob of reynolds number there's turbulence   and so that's one of the big and salt things  is well okay exactly how is the reynolds number   a bifurcation parameter and what's really going  on in the infinite dimensional space of the fluid