Gas sensors based on semiconductor metal oxide (SMO) demonstrated commendable efficiency in gas detection, but susceptibility to environmental fluctuations remains a noteworthy concern. Given the inherent intricacies of the gas detection process, it is imperative to undertake a comprehensive investigation into the multifaceted implications of environmental variations on sensor performance. In this research, we selected a laboratory-fabricated porous ZnO nanosphere gas sensor as the subject of study, which showcased a remarkable response (148) to 100 ppm ethanol, a response/recovery times of 29.5/141 s, along with a detection limit of 50 ppb.To elucidate the influence of environmental changes on sensor performance, we harnessed the response surface method (RSM) as an analysis tool. Furthermore, we expanded the utility of RSM to forecast the gas-sensing response, response time, and recovery time for both laboratory-fabricated and commercial sensors. The average errors associated with the laboratory sensors were determined to be 6.00%, 5.84%, and 6.75%, respectively, while those of the commercial sensors were 3.62%, 4.96%, and 5.09%. These disparities comfortably fall within established acceptable limits, underscoring RSM's practicality and precision in predicting the gas properties of SMO-based gas sensors.
14 May 2024 in IEEE Transactions on Industrial Electronics
Acetone and butanone, key ketone organic solvents in industrial electronics manufacturing, with their flammable and explosive properties pose risks at specific concentrations upon evaporation. Accordingly, accurate detection of ketone gases using stable semiconductor gas sensors holds critical significance in electronics manufacturing. However, inaccurate ketone gas detection through contemporary intelligent recognition methods is hindered by the limited differentiation in semiconductor gas sensor signals for ketone gases, especially in mixed gas environments. To address this, we propose a novel data-driven solution—the multiscale bidirectional temporal convolutional networks (Ms-BiTCN) framework, integrated with time–frequency analysis. Dynamic temperature modulation and variational mode decomposition Hilbert–Huang transform (VMD-HHT) are employed to transform gas sensor signals into the time–frequency domain. Subsignal frequency–domain data determines gas types, while subsignal time–domain data quantifies gas concentration. Experimental validation, including acetone, butanone, and acetone–butanone mixtures, demonstrates Ms-BiTCN’s superior performance over state-of-the-art methods, achieving a ketone organic recognition precision of 96.3% and a concentration quantification root mean square error (RMSE) of 2.09 ppm. This innovative framework has the potential to greatly enhance safety and efficiency in integrated circuit lithography by offering accurate gas level tracking and analysis.
10 August 2022 in IEEE Transactions on Instrumentation and Measurement
The precursor chemical (PC) plays a crucial role in the drug manufacture. In recent years, PC intelligent recognition technology based on the analysis of semiconductor gas sensor signals has become a research hotspot due to PC’s volatility. However, in practice, the weak discrimination for commercial semiconductor gas sensor signals of PC gases is frequent, making it difficult to accurately recognize PC gases using the existing intelligent recognition methods. To tackle this problem, this article proposes a data-driven model based on ensemble empirical mode decomposition (EEMD) and gate recurrent unit (GRU) union convolutional neural network XGBoost (GUCNN-XGBoost) methods for PC gases recognition. First, the signals are collected by the commercial semiconductor gas sensor (MQ2) under the square wave temperature modulation, and then processed by the EEMD to obtain the intrinsic mode functions (IMF1–IMF7). Second, the GUCNN-XGBoost model is constructed to automatically extracts the binary morphological characteristics (BMC) and the time sequence characteristics (TSC) of the IMF1–IMF6; the XGBoost is utilized as the final classifier instead of common softmax to realize the high precision of 97.54%. Subsequently, we employ the IMF7-XGBoost regression model to achieve the accurate estimation of PC gases concentration, which can result in an average mean absolute error (MAE) of 4.98. Finally, a typical comparative experimental analysis validates the effectiveness of the proposed approach.
12 September 2024 in IEEE Transactions on Industrial Informatics
Ensuring safety in hazardous chemical production relies heavily on robust gas monitoring systems for industrial reliability. Integrating dynamic temperature modulation signal processing and artificial intelligence with semiconductor gas sensors advances gas sensing capabilities significantly. Challenges include manual signal parameterization for various gases and complex network deployment on resource-limited devices. To tackle these, we propose the lightweight and precise dynamic signal adaptive reconstruction multitask knowledge distillation network teacher–student (DSAR-MTKDNet $\mathbb {T}-\mathbb{S}$ . DSAR enhances response signals from individual gas sensors through adaptive reconstruction, while multitask knowledge distillation compresses an long short term memory networks (LSTM)-based teacher model (MTKDNet $\mathbb {T}$ ) into a streamlined 1D-CNN-based student model (MTKDNet $\mathbb {S}$ ), boosting the student model's prediction accuracy. Experimental findings on datasets including toluene, acetone, and acetic anhydride demonstrate DSAR-MTKDNet $\mathbb {T}-\mathbb{S}$ ’s superior performance over existing methods. This demonstrates the superior practicality and versatility of the approach in enhancing gas monitoring for hazardous chemical production environments.
6 September 2023 in IEEE Transactions on Instrumentation and Measurement
Precursor chemical mixtures gases (PMG) analysis plays a significant role in combatting drug-related crimes. However, existing intelligent methods for analyzing PMG components are constrained by weak signals, as well as signal coupling issues related to semiconductor gas sensors. To address this issue, a variational mode decomposition (VMD) adversarial autoencoder joint multi-task convolutional neural network (VAE-MTCNN) algorithm is proposed. The proposed algorithm uses a semiconductor gas sensor MQ4, to collect signals in square wave temperature modulation, which are then processed by VMD to minimize signal coupling. The sub-signals are then augmented using the adversarial autoencoder through adversarial training. The construction of RGB digital portraits (RDP) is used to enhance the validity of characteristic expressions by converting sub-signals to 2D images. Finally, the MTCNN is designed to achieve PMG component recognition and quantification using two parallel task channels. The performance of the proposed method is validated using PMG samples, including acetone, ether and acetone & ether. Experimental results show that the VAE-MTCNN approach outperforms current state-of-the-art methods, with a PMG component recognition precision of 96.3% and a concentration quantification root mean square error (RMSE) of 3.50 ppm. Therefore, this method has great potential to improve drug-related crime detection.
1 January 2026 in IEEE Transactions on Industrial Informatics