Application Cases

Experiment Name: Nonlinear Vibration Acoustic Modulation for Intelligent Interface Recognition of Debonding Damage

Research Direction: Structural Health Monitoring

Experiment Content: Utilizing nonlinear vibration acoustic modulation combined with intelligent algorithms to intelligently recognize debonding damage at structural interfaces, effectively eliminating the influence of inherent structural nonlinearities on the recognition results.

Testing Equipment: ATA-2021B high-voltage amplifier, NI system (AWG, DIG, etc. cards), oscilloscope, piezoelectric sensors, etc.

Figure 1: Schematic diagram of the experimental setup for intelligent recognition of debonding damage using nonlinear vibration acoustic modulation

Experiment Process:

Figure 2: Flowchart of the experimental process for intelligent recognition of debonding damage using nonlinear vibration acoustic modulation

Nonlinear vibration acoustic modulation is used to excite the structure. Both high-frequency and low-frequency excitation signals are generated by the AWG card (PXI-5413) and amplified by the ATA-2021B high-voltage amplifier 1 () and power amplifier 2 (FalcoWMA-300) from Aigtek, and then input into the SM411 and SM412 piezoelectric sensors. Response data are collected by another SM412 sensor. The response signals are recorded by the oscilloscope DIG card (PXIe-5110).

Test Results:

Figure 3: Nonlinear vibration acoustic modulation spectra before and after damage. It can be seen that due to the influence of inherent classical nonlinearities in the structure, it is difficult to effectively identify interface debonding damage using only the nonlinear vibration acoustic modulation method.

Figure 4: Recognition results of various artificial intelligence models. It can be seen that the nonlinear vibration acoustic modulation method combined with the GA-BiLSTM-Attention model is the most effective in identifying interface debonding damage.

2000 sets of experiments all obtained stable output signals. After learning by intelligent algorithms, 1000 sets of data before and after damage form a database, which is used to intelligently classify newly collected signals, thereby achieving intelligent recognition of structural interface debonding damage and effectively eliminating the influence of inherent structural nonlinearities on the recognition results.

Power Amplifier Recommendation: ATA-2021B High-Voltage Amplifier

Figure: Specifications of the ATA-2021B High-Voltage Amplifier