2.2. Laboratory Experiments

Seven VIG type plates were investigated in this study. The construction of the tested panels is typical. They consist of two sheets of tempered glass hermetically sealed around the edges with a seal made of steel. The thicknesses of the panes are different, and they define the thickness of the whole composite element. The width of the edge seal is 9 mm around the perimeter of the panel. There is a vacuum between the glass panes, and the possible pressure deviation inside the panel is up to 0.1 Pa. The thickness of the vacuum is 0.3 mm. A system of steel support pillars is placed between the glass panes. The diameter of a single pillar is 0.6 mm, and their spacing is 55 mm.

Plates vary in thickness and dimensions in plan view. Table 1 presents the variable geometric parameters of the tested plates and defines the plate numbering, which will be used in this paper. Table 2 presents the fixed parameters of the tested elements.

Variable geometric parameters of the tested VIG plates.

Constant parameters of the tested VIG plates.

In order to test the vacuum windows, a special test stand in the shape of a cube with a 2 m side has been created. The structure of the stand is made of 40 mm × 40 mm aluminium profile sections. The composite panels were suspended from the frame structure using fabric elastic bands. It was assumed that the tape would wrap around the slab along all four edges. The tapes used were made of polyester with a width of 25 m. Figure 3 illustrates the method used to suspend the panels.

Board suspension method using elastic bands.

The basic properties of the elastic tape were determined via additional testing. Four tensile tests were performed on a tape with an initial length of 15 cm using a Zwick Roell Z050 (ZwickRoell, Ulm, Germany) static testing machine. The results of the test are described in Section 2.3. The Young’s modulus of the tape was determined as 4 GPa.

Non-destructive measurement technology based on the PULSE vibration test method from Brüel & Kjær (Nærum, Denmark) were used to test the panels. The basic element in this method is a LAN-XI type 3160 (Brüel & Kjær, Nærum, Denmark) measuring cassette, which is the data acquisition system and the central unit of the entire measurement system. This element has several input and output positions, to which the vibration excitation device, the measuring apparatus, and the computer are connected.

A specialised Brüel & Kjær impact hammer type 8206 with a sensitivity of 22.7 mV/N was used as the excitation device. It allows for generating impulse signals and, contrary to electrodynamic actuators, does not cause additional load of the tested object. The hammer has three interchangeable tips—rubber, plastic, and aluminium. Ultimately, the plastic tip was used because it allowed for accurate reading of results with little risk of overload.

The recording of the dynamic response of the composite structures was measured using Brüel & Kjær type 4399 accelerometers. The accelerometer is used to measure linear and angular acceleration. Three uniaxial accelerometers will be used in this study to study accelerations in the vertical direction. In order not to disturb the structure of the composite plates and the vacuum action, the accelerometers were attached to the surface of the test pieces with beeswax. Two accelerometers were attached to the vacuum glazing-one on the top surface of the plate and one on the bottom surface. Two measurement points were chosen at distances from the edge of the plate equal to 0.2 and 0.5 of the plate’s edge length to minimize the possibility of a situation where the accelerometer is located at a place with zero vibration amplitude. Additionally, a control measurement was made by placing one accelerometer on the aluminium frame. Figure 4, Figure 5 and Figure 6 show the location and mounting of the accelerometers.

Mounting method and location of accelerometer on aluminium frame.

Mounting method of accelerometers on the VIG panel.

Accelerometer locations on the VIG panel.

A computer with installed software compatible with the measurement equipment and the impact hammer is also connected to the data acquisition system via LAN cable. Figure 7 shows the complete test stand.

Scheme of the test stand with its elements.

Brüel & Kjær software was used to observe and record measurements of the dynamic response of the VIG plates. The Front-end Pulse communicates with the computer via a LAN network cable. However, the computer must know the IP address that the Front-end Pulse uses. The process of setting up the computer to communicate with the Front-end was conducted using the program PULSE Front-end Setup.

PULSE Labshop can transform the measured data into a frequency domain and determine the FRF (Frequency Response Function) and thus modal parameters such as the natural frequency using Fast Fourier Transform (FFT). The Fast Fourier Transform is a powerful algorithm for calculating the Discrete Fourier Transform (DFT). The DFT is a discrete equivalent of the continuous Fourier transform, which is used because of the discrete nature of the measured data. Performing a change of the domain of the function from the domain x^o to the domain s^o for the function f(x), the Fourier transform can be described by Equation (1):

in which i is an imaginary unit (i2=−1) [6].

The discrete Fourier transform for a sequence x_n consisting of N elements is given by the formula:

where x0,…,xN−1 are the signal samples, and i the imaginary unit (i2=−1) [6].

Reducing the number of necessary arithmetic operations in FFT relative to DFT is carried out by dividing the sequence into smaller strings. One of the most popular methods of determining the FFT is the classical Cooley–Turkey algorithm, in which the sequences of signal values have a basis equal to 2. It is assumed that N is a power of two and there are N samples of the discrete signal. This sequence is first decomposed into strings consisting of samples with even x_e(n) and odd x_o(n) indices. The Fourier transform of the strings is performed: X_e(n), X_o(n) and sums the transforms of the whole sequence of samples according to formula:

for 0≤k≤N2 [13].

The signals connected to the data acquisition system were added in PULSE Labshop. The first three signals were accelerometers, located on the top and bottom surfaces of the VIG plate and aluminium frame, respectively, acc_top, acc_bottom, and acc_frame. The fourth and last input added was a vibration forcing device, the impact hammer: hammer. A signal group was also created to which all transducers were added. This group was connected to the FFT analyser. Then, cross-spectrum functions were added:

After all settings had been applied, measurements were made using the Start button. Five combinations of the VIG plate test method were considered, differing in the position of the accelerometers on the plate and where the loads were applied. Two accelerometer location positions were assumed, location No. 1 and location No. 2 (Figure 6). The forcing will be applied at 2 selected points and at random. Three trials were performed for each combination. Table 3 shows the description of the combinations and their numbering, which will be used later in the paper.

Combinations of VIG plate test method.