Control of gaps in technical structures during ground vibration testing

OBRABOTKAMETALLOV MATERIAL SCIENCE Том 20 № 3 2018 EQUIPMEN . INSTRUM TS Vol. 3 No. 2 2021 Introduction A fair number of technical structures contain gaps (backlashes) which can be contingently divided into two types. The first type includes gaps in connections between substructures which are introduced so that the connections may operate correctly. Sizes of such gaps are usually normalized. Another type is the backlashes, which occur during operation. Due to the normalized gaps usually expand while operating, both of the types may lead to increased loading and wear of mechanical parts, an alteration in dynamical characteristics and a deterioration in a technical state of mechanical structures. It explains the necessity to control the gaps. When ground vibration testing (GVT) of the structures is performed, it seems expedient to use the tests to detect such gaps. Vibrodiagnostics of machinery is widely used to evaluate a technical condition of mechanical transmissions, couplings and bearings [1–5]. These rotating parts of machines tend to generate oscillations when rotating parts are imbalanced, contain the backlashes or a shaft misalignment has occurred. Oscillations, which are diagnosed as vibrations of chassis, contain information about dynamic processes in the working machine. One has to select a specific piece of data in order to detect structural damage and make decisions on structural health [6–8]. Methods of vibration-based damage detection can be divided into three categories. The first one consists of methods to detect mechanical damage by changes in modal properties [9–15]. It is important to notice that the modal properties do not alter drastically, even though the mechanical structure is damaged considerably. Additionally, the conclusive damage identification is limited due to the fact that modal properties are integral characteristics, and the location and size of the defect are differential ones [16]. The second category is represented by the methods of the wave-based damage detection [17–21]. However, the application of these methods is considered to be challenging when the structures have discontinuities such as holes and notches. When design parameters of a mechanical structure are close to ones of a linear dynamical system, portraits of oscillations distort and vibrational subharmonic and superharmonic resonances occur. The methods for the damage identification based on these features may be combined into the third group [22–30]. The paper [29] shows that one can determine and evaluate the gaps in the control wiring of the deflectable surfaces by using nonlinear distortions of portraits of oscillations. In this way, results of the GVT are utilized to calculate the distortions. The current paper is aimed at creating a vibration-based technique to monitor the gaps by means of the distortions of the portraits of oscillations. So, the computer program has been developed and integrated into the vibration testing software to compute and plot the aforementioned distortions. Also, to detect the structural damage sequentially, the novel approach has been proposed. This allows not only to identify gaps, but also to estimate its size. Research Technique The identification of the gaps in the mechanical structures by the means of the portraits of oscillations is similar to the vibration-based cracks detection [30]. The steady-state forced oscillations of the technical structures, which have been measured by acceleration sensors, were excited by means of shakers. The sensors are located near moving connections and attachment points of mechanical parts and apparatus. The sensor signals are represented as the portraits: the vertical scanning is proportional to the signal and the horizontal scanning – to its first harmonic with the phase shift of π/2. In case of a linear system, the portraits are circles. The presence of the gaps distorts the portraits of oscillations specifically (Fig. 1, n – the sensor signal, n 1 – its first harmonic with the phase shift of π/2). To estimate the distortions numerically, the first harmonic was subtracted from the Fourier series of the portrait of oscillations; the absolute maximum of the residue was calculated over the oscillation period and used subsequently as the distortion parameter Ψ. The value of the parameter Ψ was normalized and denoted as ξ. The ξ distributions were plotted on controlled objects. The locations of the gaps were determined through the positions of the local maxima of the distortions.