This paper argues that there has not been enough discussion in the field of applications of Gaussian Process for the fast moving consumer goods industry. Yet, this technique can be important as it e.g., can provide automatic feature relevance determination and the posterior mean can unlock insights on the data. Significant challenges are the large size and high dimensionality of commercial data at a point of sale. The study reviews approaches in the Gaussian Processes modeling for large data sets, evaluates their performance on commercial sales and shows value of this type of models as a decision-making tool for management.
<p>The efficient and precise monitoring (detection, localization, and quantification) of fugitive methane (CH<sub>4</sub>) emissions is essential in preventing and mitigating greenhouse gas (GHG) emissions from oil and gas industrial facilities and landfills. Various strategies of mole fraction sampling within or in the vicinity of the sites and of atmospheric inversions have been developed for such a monitoring. Many studies have ensured the constant improvement of instrumentation, of measurement strategies and atmospheric inversion techniques.</p><p>In this context, we participated in two controlled-release experiments at the TOTAL Anomaly Detection Initiatives (TADI) test site (Lacq, France) in October 2018 and 2019, dedicated to evaluate the ability of different local-scale atmospheric measurement and inverse modeling systems to localize and quantify point sources. We also conducted a series of 18 campaigns to regularly quantify methane emissions from the active &#8220;Butte-Bellot&#8221; landfill (about 35 km south-east of Paris) since 2018. We developed and applied different inversion approaches to process mobile or fixed-point measurements, which, in both cases, rely on a Gaussian dispersion model to simulate the atmospheric plume from the potential source location or mole fraction sensitivity at the measurement time and location to emissions at the potential source locations.</p><p>The series of CH<sub>4</sub> and carbon dioxide (CO<sub>2</sub>) controlled releases in TADI covered a wide range of release rates (~0.1 to 200 gCH<sub>4</sub>/s and 0.2 to 200 gCO<sub>2</sub>/s) and durations from 4 to 8 minutes (brief) to 25 to 75 minutes (longer). During the corresponding campaigns, we conducted both near-surface mobile and fixed-point (~2-4 m height) in situ atmospheric measurements based on Picarro CRDS, LGR (MGGA and UGGA), and LI-COR (LI-7810) instruments. Both inversions based on mobile measurements and those based on the fixed-point measurements provide estimates with a 20-30% average error for the CH<sub>4</sub> and CO<sub>2</sub> release rates, whatever the duration of the releases. The use of fixed-point measurements during long releases allow for a more precise localization of sources with an average location error of ~8m.</p><p>The analysis of the CH<sub>4 </sub>mobile measurements at the &#8220;Butte-Bellot&#8221; landfill reveals the difficulties in exploiting measurements close to such a site with diffuse emissions whose spatial distribution is difficult to characterize, heterogeneous and highly variable in time. The series of estimates of the total CH<sub>4</sub> emissions from the site based on remote mobile plume cross-sections, based on atmospheric inversions, are discussed.</p><p>This presentation will highlight positive perspectives opened by the proposed inversion approaches and by our results and discuss options for further improvements when processing both types of measurements.</p>
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.