In a linear differential equation, resonance in a differential equation occurs when the forcing belongs to a mode annihilated by the homogeneous operator, so an ordinary trial of the same shape cannot be inverted. For with , a polynomial factor in supplies a resonant exponential particular solution. The root multiplicity determines the required power.
For , is a particular solution. Direct differentiation gives . By contrast, the resonant equation has , which still decays: resonance enhances a mode relative to its natural exponential, without always causing an unbounded physical response. Under logarithmic coordinates in a Cauchy-Euler equation, polynomial factors in transformed time become logarithms of .
A resonance condition specifies a match between forcing parameters and a homogeneous mode. For the constant-coefficient system , an exponential trial requiresIf is not an eigenvalue, this is invertible. If it is an eigenvalue, solvability depends on the direction , not merely on the parameter match. For a diagonalizable matrix, every left eigenvector with eigenvalue requires for a same-shape trial; a nonzero projection drives a polynomial-exponential response.
For and , forcing gives a term, whereas forcing has the ordinary particular solution . Thus a spectral match is a potential resonance, and coupling to that mode determines whether the resonant response actually appears.
Resonant forcing couples an input to a natural homogeneous mode. For the undamped harmonic oscillator equation , , a particular response isIts linearly growing envelope follows by direct substitution and reflects coherent energy input each cycle. Initial conditions add bounded homogeneous sine and cosine terms, which cannot remove that secular growth.
Positive damping changes the conclusion. For with , the steady harmonic amplitude isIt is finite at , where it equals . A finite resonance peak in a damped system is therefore distinct from the unbounded undamped resonant solution. The damping and forcing conventions must be specified before asserting linear amplitude growth.
Articles by others on the same topic
There are currently no matching articles.