EVALUATION OF THE STRUCTURAL NOISE SOURCES OF A FOUR-PIECE GRAVITY DROP FORGE HAMMER BY RESIDUAL SPECTRUM TECHNIQUES.

Properly implemented residual spectrum modeling can yield useful quantitative source identification. A three-input model applied to a tightly controlled production hammer accounted for 82% of the energy radiated past the operator's position; the ram appeared to be the major contributor with the columns and anvil contributing lesser, roughly equal, amounts. However, a five-input model applied to a less strictly controlled hammer accounted for only 64% of the energy. This reduced effectiveness is attributed to poor blow control. To improve a model's effectiveness at higher frequencies, it may be feasible to improve an element's characterization by means of a linear combination of transducers distributed over its surface.