Results in Engineering | Data-driven prediction of electroacoustic responses of high-voltage burst-driven piezoelectric transmitter systems

By  李勇    2026-10-19    Visited 10 times

Abstract

High-voltage burst-driven piezoelectric transmitter systems exhibit complex electroacoustic responses governed by drive electronics, load coupling, dynamic impedance, and transient nonlinearities. This study establishes a synchronous measurement system and electroacoustic database using a monopole acoustic logging transducer. Four prediction frameworks are compared: a PCA–MLP model using control parameters, a temporal convolutional network (TCN) using theoretical control waveforms, a two-stage cascade using predicted electrical waveforms, and a TCN using measured excitation voltage and current. The measured-electrical-input TCN achieves the highest accuracy, with a coefficient of determination of 0.9984, a mean waveform correlation coefficient of 0.9988, and a mean peak-to-peak relative error of 1.172%. These results support excitation voltage and current as intermediate transmitter-state variables. However, the comparison does not isolate input-information effects from differences in mapping complexity. Evaluation at 37 held-out frequency settings within 12–30kHz demonstrates interpolation performance, without establishing robustness beyond the measured band or at untrained amplitudes or cycle counts. The framework enables virtual sensing of hydrophone responses and offers a potential basis for transmitter-state evaluation, frequency screening, and excitation optimization within the characterized domain. Application to other high-voltage burst-driven piezoelectric transmitter systems requires additional validation and appropriate retraining.

Paper Information:

Zhang, K., Zhang, H., Su, Y., Tan, B., and Liu, L., 2026. Data-driven prediction of electroacoustic responses of high-voltage burst-driven piezoelectric transmitter systems. Results in Engineering, 32, 113403.  https://doi.org/10.1016/j.rineng.2026.113403