Solving the simultaneous equation for 5 masses

n.b. This is "for interest”.

Suppose in the 5-mass case we need to find both angles: we simply know that they are A, -B, 0, B, -A.  The two equations to be solved are

1 + 2 cosB + 2 cosA = 0

2(sinB - 2sinA) = 0

The other two equations (remember we have 4 - the two force components and two moment components) are automatically satisfied because of the symmetry of the A, -B, 0, B, A configuration.

We could solve this algebraically:

  • square both equations
  • substitute one in the other
  • this gives a quadratic for one angle
  • substitute to find the other angle


There are easier ways, though, if you as an engineer just want a quick solution.  One is to use an iterative method:

 beta = 360/5;  % 72 degrees, initial guess

for i = 1:5, 

    alpha = 180 - asind(sind(beta)/2); 

    beta = acosd(-0.5-cosd(alpha)); 

    disp([alpha beta]), 

end

  151.6062   67.6849

  152.4479   67.2575

  152.5402   67.2112

  152.5503   67.2062

  152.5514   67.2056

Try doing this on your calculator.  Check it's in degrees mode, then:

   72 =
   180 - sin-1(sin(ANS) / 2)  =
   cos-1(-0.5 - cos(ANS)) =


Press “=" a few times, using the UP arrow each time to go back to the previous equation so you alternate between the “ 180 - …"  and "cos-1(…” lines - watch it converge, giving you better and better estimates for the A and B values.


We could also use an algebraic solver such as the symbolic toolbox in Matlab.  If you have the toolbox installed (type “help symbolic" to find out) you can simply type:

>> syms A B

>> eqn1 = 1 + 2*cos(B) + 2*cos(A) ==0; % note “=="

>> eqn2 = sin(B) - 2*sin(A) ==0;

>> AB = solve(eqn1, eqn2)

AB = 

    A: [4x1 sym]

    B: [4x1 sym]

% AB is a “structure” with a field for each coefficient.

>> AB.A

ans =

 -2*atan((11*(- (2*10^(1/2))/3 - 5/3)^(1/2))/2 + (3*(- (2*10^(1/2))/3 - 5/3)^(3/2))/2)

  2*atan((11*(- (2*10^(1/2))/3 - 5/3)^(1/2))/2 + (3*(- (2*10^(1/2))/3 - 5/3)^(3/2))/2)

     -2*atan((11*((2*10^(1/2))/3 - 5/3)^(1/2))/2 + (3*((2*10^(1/2))/3 - 5/3)^(3/2))/2)

      2*atan((11*((2*10^(1/2))/3 - 5/3)^(1/2))/2 + (3*((2*10^(1/2))/3 - 5/3)^(3/2))/2)

 

>> eval(AB.A)

ans =

   0.0000 + 0.6528i

   0.0000 - 0.6528i

  -2.6625 + 0.0000i

   2.6625 + 0.0000i

% There are 4 solutions. We aren’t interested in the imaginary ones.  The two real terms are the radian equivalents of the +/-152.6 degrees (A) and +/-67.2 degrees (B).

>> eval(AB.B)

ans =

   3.1416 - 1.1382i

  -3.1416 + 1.1382i

  -1.1730 + 0.0000i

   1.1730 + 0.0000i


Footnote:

Back in 1884 this kind of calculation was the basis for the Yarrow-Schlick-Tweedy system for engine balancing (we will see the relevance of shaft balancing to engines with pistons later, in balancing_app).  Triple expansion marine steam engines could not have moment balance (only 3 cylinders): the solution was to convert the large LP (‘low pressure”cylider into two smaller ones, positioned at each end of the engine.   The crankshaft for this 4 cylinder engine was rather like the above 5-mass shaft, with the central mass removed.

Nowadays it is mostly of academic interest.  5 cylinder car engines have 72 degree angles because this gives an even firing interval (the engine feels rough is the interval is uneven - this is acceptable on a motorcycle but less so for a luxury car).  

Large marine engines (up to 18 cylinders) have crank angles carefully chosen to achieve both inertial balance and also to avoid exciting any torsional vibration modes.



© Roger Moss 2015