Performance of artificial neural networks and linear models to quantify 4 trace gas species in an oil and gas production region with low-cost sensors

SENSORS AND ACTUATORS B-CHEMICAL

卷:283

Article

Casey, Joanna Gordon¹  Collier-Oxandale, Ashley²  Hannigan, Michael¹ 

[1] Univ Colorado, Dept Mech Engn, Engn Ctr, ECME 114,1111 Engn Dr, Boulder, CO 80309 USA

[2] Univ Colorado, Dept Environm Engn, 4001 Discovery Dr, Boulder, CO 80303 USA

关键词: Low-cost sensors;    Field calibration;    Artificial neural networks;    Linear regression models;    Air quality;    Oil and gas production;   

DOI  :  10.1016/j.snb.2018.12.049

来源: Elsevier

【 摘 要 】

We tested the performance of regression via inverse linear models (LMs), direct LMs, and artificial neural networks (ANNs) towards field calibration of low-cost gas sensors in an area influenced by oil and gas production activities to quantify O-3, CO, CO2, and CH4 in ambient air. Sensors were co-located with reference measurements in Greeley, Colorado. We selected a three-month period of data in the spring of 2017 to conduct our analysis. Approximately two months of measurements book ending the middle test month were used for model training. We found that ANNs generally outperformed LMs and that direct LMs generally outperformed inverted LMs. An analysis of model residuals during the test period revealed that ANNs were better able to mitigate curvature and linear trends relative to direct LMs with the same set of inputs.

【 授权许可】

Free