Scholarlink

<p>Humidity sensors play a critical role in healthcare, environmental monitoring, industrial automation, and energy systems. Achieving a balance between high sensitivity, fast response, and long-term predictive reliability remains a challenge. In this work, a CNTs/GO/PANI ternary nanocomposite-based humidity sensor was synthesized via an in situ polymerization method to address this gap. The morphology, crystal structure, and chemical bonding of the nanocomposite were characterized using scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM), X-ray diffraction (XRD), and Fourier transform infrared microscopy (FT-IR-IM). The performance of the synthesized nanocomposite was evaluated over a broad relative humidity range (11–97% RH) and frequency range (50&#xa0;Hz–100&#xa0;kHz). The investigated sensor demonstrated good sensitivity and good short-term stability with a low hysteresis of 2.1%. To extend beyond standard short-term evaluation, a long short-term memory (LSTM) neural network was trained using a sliding window approach and leave-one-out cross-validation (LOO-CV) to predict the sensor’s long-term response behavior, and benchmarked against a naïve baseline, a simple recurrent neural network (RNN), and a gated recurrent unit (GRU) network. The LSTM model achieved the lowest average error of 0.4102 among all evaluated models with Diebold–Mariano tests confirming the improvement was statistically significant in each case. These results demonstrate that the CNTs/GO/PANI nanocomposite exhibits strong humidity-sensing performance across a wide RH range, and that the LSTM framework offers a reliable and statistically validated approach for predicting long-term sensor stability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Synthesis, characterization, and LSTM-based modeling of the long-term stability of CNTs/GO/PANI humidity sensor

  • 期刊论文

  • 发表日期：

    2026年7月27日

查看全文
  • Mohamed Morsy,
  • Mohamed K. Hassanien,
  • Mohamed Farouk

摘要

Humidity sensors play a critical role in healthcare, environmental monitoring, industrial automation, and energy systems. Achieving a balance between high sensitivity, fast response, and long-term predictive reliability remains a challenge. In this work, a CNTs/GO/PANI ternary nanocomposite-based humidity sensor was synthesized via an in situ polymerization method to address this gap. The morphology, crystal structure, and chemical bonding of the nanocomposite were characterized using scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM), X-ray diffraction (XRD), and Fourier transform infrared microscopy (FT-IR-IM). The performance of the synthesized nanocomposite was evaluated over a broad relative humidity range (11–97% RH) and frequency range (50 Hz–100 kHz). The investigated sensor demonstrated good sensitivity and good short-term stability with a low hysteresis of 2.1%. To extend beyond standard short-term evaluation, a long short-term memory (LSTM) neural network was trained using a sliding window approach and leave-one-out cross-validation (LOO-CV) to predict the sensor’s long-term response behavior, and benchmarked against a naïve baseline, a simple recurrent neural network (RNN), and a gated recurrent unit (GRU) network. The LSTM model achieved the lowest average error of 0.4102 among all evaluated models with Diebold–Mariano tests confirming the improvement was statistically significant in each case. These results demonstrate that the CNTs/GO/PANI nanocomposite exhibits strong humidity-sensing performance across a wide RH range, and that the LSTM framework offers a reliable and statistically validated approach for predicting long-term sensor stability.