4. Earthquakes, Records, and Nonlinear Responses

The estimation of MDOF nonlinear structural response given an earth‐quake of magnitude M at distance R is studied with respect to issues such as the benefits and harms of (1) first scaling the records, (2) selecting records from the “wrong” magnitude, (3) alternative choices for how to scale the records, and (4) scaling records to a significantly higher or lower intensity, etc. We find that properly chosen scaling can reduce the necessity of the number of nonlinear analyses by a factor of about four, and that proper scaling does not introduce any bias. Several global and local nonlinear damage measures are considered. A five‐DOF model of a steel structure is used; other cases are under study. The paper finishes with a demonstration of the use of such results in the estimation of the annual probability of exceeding a specified interstory ductility (drift) or other damage measures.