6926475912

w

data can be specified in cartesian. cylindrical. or spherical coordinates. For structures włth repetitive geometry a pattern-entry capability can significantly reduce the data input time. Modal pa r a met er s are computed and stored on a floppy disk. This Information can be displayed on a graphics screen or plotted on paper depending on the specific need.

Virtually all commercially available modal analysis Systems i^clude an IEEE 488 parał lei interface bus as a standard or an option. It provide$ a convenient way for fast communication with a variety of digital devices. For example, the overall operation can be coordinated, and further Processing can be carried out by a host Computer. A schematic of this con-figuration i$ shown in Figurę 3. Notę that the modal analysis processor. graphics terminal, and storage device can be replaced by a commercial desk-top Computer with an associated reduction in overall cost of the system. An alternative system configura-tion that is particularly useful in data transfer and communication from remote test sites is a voice-grade telephone linę and a modem coupler to interface the FFT analyzer to the modal processor.

Figurę 3. Typical Structure of a Modal Analysis System under Control of a Host Computer

Functionał requirements. The first step in selecting a modal analysis system for a particular application is understanding the specific needs of that application. For industrial applications of modal testing the following requirements are typically adequate.

Acceptance of a wide rangę of measured sig-nałs having a variety of transient and frequency band characteristics

ii.    Capability to handle up to 300 degrees of freedom of measured data in a single analysis

iii.    FFT with frequency resolution of at least 400 spectral lines per 512

iv.    Zoom analysis capability

v.    Capability to perform    statistical    error-band

analysis

vi.    Static display and plot of mode-shape extremes

vii.    Animated (dynamie) display of modę shapes

viii.    Color graphics

ix.    Hidden-line display

x.    Multiple-pen color plotting    and    improved    linę

resolution

xi.    Capability to generate an accurate time-domain model (mass, stiffness. and damping matrices).

Comparative study. Four representative modal analysis systems are selected for a comparative study. We shall cali them System A, System B, System C. and System D for convenience. This group is not intended to be exhaustive; there are other comparable systems. The availability of data has been the primary reason for these particular choices.

Several generał comments can be madę with respect to the functional requirements listed earlier. All four systems meet most of these requirements. The capability to accept a variety of measured signals depends primarily on the weighting window (1,2) -options available in the FFT analyzer. For example, the System D analyzer provides the conventional Hanning and box-car Windows. In addition, it pro-vides a flat-top window for accurate amplitudę results and a half-Hanning window suitable for such transient data as impact-test data and for lightly damped responses having extensive end oscillations.

When the sample period (the time between two adjacent digital data simples) is AT, the maximum frequency that has any significance in digital Fourier analysis results is the Nyquist frequency (cut-off frequency) f_c (1)

1

f * -

c 2AT

This relationship is known as the sampling theorem. Furthermore. if the record length used in a single FFT computation is T, it will require an input mem-

**\

V»

SI