Measurement data handling in identification of mechanical systems

Abstract

Integration algorithms for data handling in identification of linear stationary systems from sampled transient data are proposed. The problem of identification from several simultaneously measured phase coordinates is detailed. The identification procedure is generalized to the case of irregularly sampled data. The efficiency of the algorithms is demonstrated by examples

Identification of linear time-invariant systems

“…The following coefficients were assumed in (6.1): g, g g g 1 1 1 10 : , = = . Note that the data in columns I-III agree with those in Table 2 [8]. As is seen, the algorithm described in Sec.…”

“…6 produces the most accurate estimates. Example 2 [8]. Let us consider the integration procedure for the identification of the parameters of the vibration-isolation system shown in Fig.…”

“…In what follows, we will focus on the second part of the problem, as in [1,2,8], i.e., we will estimate m j and b j , given the order of the model. The determination of the model order is an independent problem [3,9].…”

“…Algorithms for estimating m j and b j were detailed for both regularly [1] and irregularly [2,8] sampled data. These algorithms involve the decomposition of the original problem into two problems: (i) estimation of the parameters m j and (ii) estimation of the parameters b j .…”

“…6, November, 2011 objective function proposed in [2]. It was shown in [1,5,8] by way of examples that this approach increases the accuracy of identification.…”