Numerical Methods on the FE Exam
Numerical methods appear across several FE exam disciplines as part of the mathematics or engineering sciences content areas. While the depth of coverage varies, understanding basic numerical techniques is valuable both for the exam and for engineering practice where exact analytical solutions are not always available.
Root Finding
Bisection Method
The bisection method finds a root of f(x) = 0 within a bracket [a, b] where f(a) and f(b) have opposite signs. The method repeatedly halves the bracket by evaluating f at the midpoint and keeping the half that contains the sign change. The method is slow but guaranteed to converge.
Newton-Raphson Method
Newton's method uses the tangent line at a current estimate to find a better estimate of the root. The iteration formula is: x_new = x_old minus f(x_old) divided by f'(x_old). This method converges quickly (quadratically) when the initial guess is close to the root and f'(x) is not near zero at the root. However, it can diverge if the initial guess is poor.
Secant Method
The secant method is similar to Newton's method but approximates the derivative using two previous points instead of computing it analytically. This is useful when the derivative is difficult to calculate. Convergence is slightly slower than Newton's method but faster than bisection.
Numerical Integration
Trapezoidal Rule
The trapezoidal rule approximates the integral of a function by dividing the interval into subintervals and approximating the area under the curve in each subinterval as a trapezoid. The formula for a single interval is: integral approximately equals (h/2) times (f(a) + f(b)), where h is the interval width. For multiple equal subintervals, use the composite trapezoidal rule.
Simpson's Rule
Simpson's 1/3 rule uses a parabolic approximation and requires an even number of subintervals. For a single application: integral approximately equals (h/3) times (f(a) + 4f(m) + f(b)), where m is the midpoint. Simpson's rule is more accurate than the trapezoidal rule for the same number of function evaluations and is exact for polynomials up to degree 3.
Simpson's 3/8 rule uses a cubic approximation and requires subintervals in multiples of three. It is slightly more accurate than the 1/3 rule but is used less frequently.
Curve Fitting
Linear Regression
Least squares linear regression fits the best straight line (y = mx + b) to a set of data points by minimizing the sum of squared residuals. The formulas for the slope (m) and intercept (b) are provided in the FE Reference Handbook. Know how to calculate the correlation coefficient (r) and the coefficient of determination (r squared) to assess the quality of the fit.
Polynomial and Nonlinear Fitting
Higher-order polynomial regression and nonlinear curve fitting extend the least squares approach. For nonlinear models that can be linearized (such as exponential or power functions), transform the data first, then apply linear regression to the transformed data.
Interpolation
Linear interpolation estimates a value between two known data points by assuming a straight-line relationship. For more accuracy with multiple data points, use Lagrange interpolation or Newton's divided difference interpolation. Know the difference between interpolation (estimating within the data range) and extrapolation (estimating outside the data range, which is less reliable).
Numerical Solutions to ODEs
Euler's Method
Euler's method is the simplest numerical method for solving first-order ordinary differential equations (dy/dx = f(x,y)) with an initial condition. The iteration is: y_new = y_old + h times f(x_old, y_old), where h is the step size. The method is first-order accurate, meaning the error is proportional to the step size.
Runge-Kutta Methods
The fourth-order Runge-Kutta method (RK4) is much more accurate than Euler's method for the same step size. It evaluates f(x,y) at four points within each step and uses a weighted average to advance the solution. The FE Reference Handbook provides the RK4 formulas.
Error Analysis
Understand the concept of truncation error (error from approximating a continuous process with a finite number of steps) and round-off error (error from finite precision arithmetic). Truncation error decreases with smaller step sizes, while round-off error can increase with too many calculations. The optimal step size balances these two sources of error.
Study Tips
Numerical methods problems on the FE exam typically require you to perform one or two iterations of a method or to select the most appropriate method for a given situation. Practice the mechanics of each method by hand (not just with software) so you can execute the calculations quickly and accurately. The FE Reference Handbook provides the key formulas, but you need to know how to apply them.