In this paper

structural system identification without input information is studied. To shear-typed building under a special kind of unknown input

a new method is proposed to inverse load process and estimate structural parameters. In this method

the unknown inputs acting on different floors are supposed to be proportional and the computing process can be divided into two steps. By defining a basal unknown input

this problem is changed to obtain the basal unknown input

the structural parameters and the proportion coefficients of the rest unknown inputs relative to basal unknown input. In the first step

a two-floor sub-structure including the basal unknown input is used

applying Genetic Algorithm

the unknown parameters of the sub-structure

the basal unknown input and one proportion coefficient can be solved respectively. In the second step

using the basal unknown input obtained before

the least-square method is applied to solve the inversion problem of the rest unknown structural parameters and the rest proportion coefficients of the whole structure. Due to not needing the proportional coefficients beforehand

this method have a wide range of application background. At last

a numerical example is used and the results show that this method can effectively identify the structural parameters and the unknown inputs.