Department of Process Engineering and Chemical Technology, Faculty of Chemistry, Gdansk University of Technology, 11/12 G, Narutowicza Str., 80-233 Gdansk, Poland.
Molecules. 2022 Jun 29;27(13):4180. doi: 10.3390/molecules27134180.
This article presents a new way to determine odor nuisance based on the proposed odor air quality index (OAQII), using an instrumental method. This indicator relates the most important odor features, such as intensity, hedonic tone and odor concentration. The research was conducted at the compost screening yard of the municipal treatment plant in Central Poland, on which a self-constructed gas sensor array was placed. It consisted of five commercially available gas sensors: three metal oxide semiconductor (MOS) chemical sensors and two electrochemical ones. To calibrate and validate the matrix, odor concentrations were determined within the composting yard using the field olfactometry technique. Five mathematical models (e.g., multiple linear regression and principal component regression) were used as calibration methods. Two methods were used to extract signals from the matrix: maximum signal values from individual sensors and the logarithm of the ratio of the maximum signal to the sensor baseline. The developed models were used to determine the predicted odor concentrations. The selection of the optimal model was based on the compatibility with olfactometric measurements, taking the mean square error as a criterion and their accordance with the proposed OAQII. For the first method of extracting signals from the matrix, the best model was characterized by RMSE equal to 8.092 and consistency in indices at the level of 0.85. In the case of the logarithmic approach, these values were 4.220 and 0.98, respectively. The obtained results allow to conclude that gas sensor arrays can be successfully used for air quality monitoring; however, the key issues are data processing and the selection of an appropriate mathematical model.
本文提出了一种基于所提出的气味空气质量指数（OAQII）的仪器方法来确定气味滋扰的新方法。该指标涉及到强度、愉悦度和气味浓度等最重要的气味特征。该研究在波兰中部城市处理厂的堆肥筛选场进行，在该场地上放置了一个自制的气体传感器阵列。它由五个市售的气体传感器组成：三个金属氧化物半导体（MOS）化学传感器和两个电化学传感器。为了校准和验证矩阵，使用现场嗅闻技术在堆肥场内确定了气味浓度。使用了五种数学模型（例如，多元线性回归和主成分回归）作为校准方法。从矩阵中提取信号有两种方法：单个传感器的最大信号值和最大信号与传感器基线的对数之比。所开发的模型用于确定预测的气味浓度。最优模型的选择基于与嗅闻测量的兼容性，以均方误差作为标准，并与所提出的 OAQII 一致。对于从矩阵中提取信号的第一种方法，最佳模型的 RMSE 等于 8.092，并且在指数上与指标的一致性为 0.85。在对数方法的情况下，这些值分别为 4.220 和 0.98。所得结果表明，气体传感器阵列可成功用于空气质量监测；然而，关键问题是数据处理和选择适当的数学模型。
