Monitoring of spectral characteristics dynamics

One of the key factors in the course of technological products manufacturing is the testing of products resistance to mechanical impacts. These tests are performed in order to reveal structural and manufacturing defects and to evaluate performance characteristics of the product in a complex operational environment. Tests complexity for mechanical factors impact is caused by the possibility of partial or complete tested specimen destruction. In the case of tested sample destruction, the information of the tests performance may be useful for revealing the reason of the specimen destruction, hence, the more comprehensive the information is, the better. In the case, if destruction of the tested specimen is inadmissible, there should be determined a threshold for the test performance suspension.

In the course of the tests for classical shock impact, the tested specimen is exposed to a considerable short-term force impact. The main portion of the energy transmitted to the tested object comes to the displacement of shaker stand together with the attached specimen – the remaining portion of energy is consumed by the sample and leads to deformation of the specimen. The amount of the energy consumed by the sample determines the degree of the tested object deformation. The purpose is to determine the tested object deformation without using the fault detector (tested object deformation estimation shall be performed based on system response on external impact).

Figure 1 – sensors positioning

In this case, a modeling clay piece (total weight – 100gr) is used as a vibration damper – each fragment has a cube shape with dimensions 41×41×41 mm (see Figure 1). Modeling clay piece is placed in the middle of the shaker table.The first sensor was placed near the modeling clay cube at the expansion platform. The second one is placed onto the modeling clay cube. The tests were performed by means of ZETLAB shaker controller – a software used for shaker control based on ZET 017 FFT spectrum analyzer. Vibration impact tests consisted of half-sine shape classical shock with an amplitude of 1 g and 4 ms duration. An input signal is forwarded to the shaker system so that the control accelerometer could make a diagram of the pulse having corresponding shape. In this case, “BK” control accelerometer is used, necessary pulse shape is a half-sine signal.

Figure 2 depicts “Multichannel oscilloscope” window with a “valley value” at the signal peak received from the second sensor. This “valley value” is caused by damping properties of the modeling clay. The energy used for shaker platform movement if partially absorbed by the modeling clay and deforms it. The Software “Modal analysis” uses sensors signals for calculation of measured system parameters including its performance – “Integral F*s”. In accordance with the measurements, the modeling clay cube has consumed the following amount of energy:

ΔЕ=Е₁-Е₂=0,0938-0,0510=0,0428 J

This very amount of energy was spent for modeling clay cube deformation and has partially been dissipated as heat. Modeling clay cube deformation leads to a change of FR characteristics. FR value dynamics is well seen in the narrowband spectrum.

Figure 2 — ZETLAB Virtual laboratory – tests performance – “Classical shock” Software application (top left section), “Modal analysis” (at the bottom), “Multichannel oscilloscope” (at the right). 

Figure 3 shows instant spectrum diagram at the end of tests procedure, averaged spectrum diagram for the first 100 sec of the test and a comparison chart of the first two sensors. In Figure 3 one can see that the biggest difference between the spectrum diagrams lays within the range of 110-150 Hz. Since the impact signal is a half-sine wave with 4 ms duration, the full cycle has 8 ms duration, the frequency value is 125 Hz. So, the biggest FR value change in the modeling clay cube occurred in the impact frequency area (which confirms the test results validity).

Figure 3 — shock spectra: instant, average and their difference

Instant spectrum RMS is  0,001316 g, average spectrum RMS – 0,001293 g, spectrum difference RMS – 0,000075 g, relative change value –  5.8%.

In the case of continuous tests performance, the deformation value of the tested specimen gradually increases. In order to determine the deformation value, we shall use the method of comparing the shock pulse spectrum with that of the sample spectrum. As a sample spectrum value, we shall use an arithmetical average of the first shock impacts pulses. In the course of vibration tests, the sensor placed onto the modeling clay cube has declined from the vertical position, which definitely has affected the transfer characteristics.

During tests, it is necessary to have control over deviations of the sample spectrum in real-time mode (there is no need to have the deviation value only upon tests completion). Control software operating principle is quite simple: during the specified time period at the beginning of tests performance spectrum measurements results data are obtained for the purpose of further sample spectrum parameters calculation, which are further used for referencing and deviation diagram construction. If the parameter exceeds the set threshold value, the operator will receive a corresponding notification message.

This Software application has been implemented in SCADA-system ZETVIEW. Figure 4 shows the project allowing to implement the previously described algorithm. Top left section of the scheme has “input channel” element – the signal from it (blue circles and connecting lines) is forwarded to the input of “narrow-band spectrum”. At the output of “narrow-band” spectrum, there are formed spectra arrays and frequency bands (violet circles and connecting lines) – they are depicted in “spectrum diagram”. Spectra arrays are also forwarded to the element “array stacking”, which stepwise summarizes spectra arrays with the results of previous stacking within first 100 sec. The number of seconds is prescribed in the element “number of averagings” – the reverse count is performed by “increment” level (bottom left part). Upon completion of the specified time period the “multiplexer” will be switched over to the first channel used for transmission of zero array from “zero array” element having the same parameters with the spectrum. Upon the stepwise division of the array by the number of averagings specified in the “averaged spectrum” element, there is formed an averaged array for 100 sec of spectrum duration which is further forwarded to “spectrum diagram” and is deducted from the instant spectrum. From the array at the output of “spectra deduction” element RMS value is calculated (green circles and connection lines), which is then compared  with the “threshold value”, forwarded to the “RMS difference” indicator and recorded by “diagram timer” signal (red circles and connection lines) at the end of “deviation” array. Simultaneously, by a signal from “diagram timer” in the “increment” element (in the right section) a time from the test beginning is calculated and is then forwarded to “measurements” counter element and is recorded at the end of “time” array.  “Deviation” arrays are depicted on”RMS difference diagram” at the axis set by “time” array.

Figure 4 — SCADA-project ZETView – average spectrum deviation recording 

SCADA-project ZETView consists of two parts – project view (Figure 4) and operator view (Figure 5). The operator can see all necessary information relating to the project performance – diagrams, indicators and project control keys. The project is activated by “On” key (the key name will be changed for “Activated”), key “Reset” is used to restart measurements. Upon completion of the specified “accumulation time,” the color indication switches over to red color. Digital indicators below depict current “RMS deviation” of an instant spectrum from average one as well as the current test time. Upon exceeding the “threshold deviation value” the indicator turns red. The upper diagram depicts averaged and instant spectra. The lower diagram shows the deviation of instant spectrum from average one during the tests performance duration.

The lower diagram clearly shows that the difference between averaged spectrum at the beginning of the test and instant spectrum eventually increases. External accidental impacts sometimes lead to “splashes” in the diagram, however, minimum difference value constantly increases. The constant growth of minimum deviation values is a result of irreversible deformations occurring inside of the modeling clay cube. In the case of continuous and intensive tests, modeling clay deformations may become visible.

Since the modeling clay is a lamellated material, the growth of deformations number will not lead to its destruction. Products made of more fragile materials (like tempered steel, glass, etc.) may be partially or completely split into pieces as the deformation impacts achieve a certain degree. Hence, control of FR characteristics change allows to determine tested specimen destruction moment. As sufficient statistical spectral characteristics and NDT data is obtained, it becomes possible to determine exact deformation degree using spectral characteristics change information.

Figure 5 — SCADA-project ZETView – average spectrum deviation recording  – operator’s interface view

SCADA-system ZETView enables fast and simple (the above-described project has been prepared within an hour) implementation of algorithms of any complexity as well as results representation in a format, convenient for further analysis. Signals processing operations are represented in a clear view depicting operations sequence. SCADA-system ZETView is simple and user-friendly. An ordinary user without any programming skills can make a complex program for data acquisition and objects control with neat and convenient human-machine interface.

Spectral characteristics control is necessary not only in the case of vibration tests performance but also in many other scientific and technological spheres. Similar tasks also arise in the case of buildings structural control (intrinsic frequency oscillation changes), machinery defects detection, and so on. The above algorithm and its implementation in SCADA-system ZETView can be easily used for solving similar tasks.