Gaussian Process Regression as a Replicable, Streamlined Approach to Inventory and Uncertainty Analysis in Life Cycle Assessment.

Life cycle assessment plays a critical role in quantifying environmental impacts, but its credibility remains challenged when data and uncertainty analysis are lacking. In this study, we propose a data compilation framework to address these two issues. The framework first quantifies the correlations of production activities among existing data in temporal, geographical, and taxonomic dimensions. The framework then introduces covariance functions to convert these correlations to a similarity matrix, and the Gaussian process regression model is adopted to predict new data based on these covariance functions. The associated uncertainty is automatically characterized using the posterior distribution of predictions. The framework is demonstrated on the nitrogen fertilizer application rate for food production─an activity recognized for its environmental burden─with results capable of reflecting temporal and geographical variations. By introducing the concept of phylogenetic distance as a correlation of taxonomy, the framework provides a quantitative basis for predictions in a proxy data usage scenario. The framework can be used in developing temporally and regionally representative life cycle inventories and databases and can facilitate consistent uncertainty quantification in future life cycle assessment methodologies.

Life cycle environmental impacts of food away from home and mitigation strategies-a review.

Food production and consumption are major drivers of global environmental change, endangering the safe operating space of many environmental areas. Globally, there has been a growing trend of dining out, termed food away from home (FAFH) here, but its environmental sustainability has received insufficient attention. In this review, we examine studies quantifying the life-cycle environmental impacts of FAFH and identify mitigation strategies across the food supply chain. Overall, previous life cycle assessment (LCA) studies focused on the composition of FAFH meals and pre-use life cycle stages, especially food production. Greenhouse gas (GHG) emissions of FAFH meals range from 0.134 kg CO2 e/meal to 13.2 kg CO2 e/meal for school canteen meals, and from 0.60 kg CO2 e/meal to 9.6 kg CO2 e/meal for other catering services. Meat ingredients are the dominant source in a variety of environmental impact categories, and the food production stage usually accounts for over half of the total GHG emissions in the FAFH life cycle. Supply side mitigation strategies include advancing farming practices, updating cold transportation technology, and improving building energy efficiency. Demand side mitigation focuses on dietary change towards meals with less meat ingredients, with nudging and sustainable menu-designing as the two primary groups of strategies. Areas of focus for LCA include improving modeling of building energy consumption related to food consumption, advancing uncertainty characterization of life cycle results, and capturing geographical variations in food production.

Thermal conductivity predictions of herringbone graphite nanofibers using molecular dynamics simulations.

Non-equilibrium molecular dynamics (NEMD) simulations are used to investigate the thermal conductivity of herringbone graphite nanofibers (GNFs) at room temperature by breaking down the axial and transverse conductivity values into intralayer and interlayer components. The optimized Tersoff potential is used to account for intralayer carbon-carbon interactions while the Lennard-Jones potential is used to model the interlayer carbon-carbon interactions. The intralayer thermal conductivity of the graphene layers near room temperature is calculated for different crease angles and number of layers using NEMD with a constant applied heat flux. The edge effect on a layer's thermal conductivity is investigated by computing the thermal conductivity values in both zigzag and armchair directions of the heat flow. The interlayer thermal conductivity is also predicted by imposing hot and cold Nosé-Hoover thermostats on two layers. The limiting case of a 90° crease angle is used to compare the results with those of single-layer graphene and few-layer graphene. The axial and transverse thermal conductivities are then calculated using standard trigonometric conversions of the calculated intralayer and interlayer thermal conductivities, along with calculations of few-layer graphene without a crease. The results show a large influence of the crease angle on the intralayer thermal conductivity, and the saturation of thermal conductivity occurs when number of layers is more than three. The axial thermal conductivity, transverse thermal conductivity in the crease direction, and transverse thermal conductivity normal to the crease for the case of a five-layer herringbone GNF with a 45° crease angle are calculated to be 27 W∕m K, 263 W∕m K, and 1500 W∕m K, respectively, where the axial thermal conductivity is in good agreement with experimental measurements.

An LX-10 kinetic model calibrated using simulations of multiple small-scale thermal safety tests.

A new chemical kinetic model for the beta-delta transition and decomposition of LX-10 (95% octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine, 5% Viton A binder) is presented here. This model implements aspects of previous kinetic models but calibrates the model parameters to data sets of three experiments: differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), and one-dimensional time to explosion (ODTX). The calibration procedure contains three stages: one stage uses open-pan DSC and TGA to develop a base reaction for formation of heavy gases, a second stage features closed-pan DSC to ascertain the autocatalytic behavior of reactant gases attacking the solid explosive, and a final stage adjusts the rate for the breakdown of heavy reactant gases using ODTX experimental data. The resultant model presents a large improvement in the agreement between simulated DSC and TGA results and their respective experiments while maintaining the same level of agreement with ODTX, scaled thermal explosion, and laser heating explosion times when compared to previous models.

Application of global kinetic models to HMX beta-delta transition and cookoff processes.

The reduction of the number of reactions in kinetic models for both the HMX (octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine) beta-delta phase transition and thermal cookoff provides an attractive alternative to traditional multi-stage kinetic models due to reduced calibration effort requirements. In this study, we use the LLNL code ALE3D to provide calibrated kinetic parameters for a two-reaction bidirectional beta-delta HMX phase transition model based on Sandia instrumented thermal ignition (SITI) and scaled thermal explosion (STEX) temperature history curves, and a Prout-Tompkins cookoff model based on one-dimensional time to explosion (ODTX) data. Results show that the two-reaction bidirectional beta-delta transition model presented here agrees as well with STEX and SITI temperature history curves as a reversible four-reaction Arrhenius model yet requires an order of magnitude less computational effort. In addition, a single-reaction Prout-Tompkins model calibrated to ODTX data provides better agreement with ODTX data than a traditional multistep Arrhenius model and can contain up to 90% fewer chemistry-limited time steps for low-temperature ODTX simulations. Manual calibration methods for the Prout-Tompkins kinetics provide much better agreement with ODTX experimental data than parameters derived from differential scanning calorimetry (DSC) measurements at atmospheric pressure. The predicted surface temperature at explosion for STEX cookoff simulations is a weak function of the cookoff model used, and a reduction of up to 15% of chemistry-limited time steps can be achieved by neglecting the beta-delta transition for this type of simulation. Finally, the inclusion of the beta-delta transition model in the overall kinetics model can affect the predicted time to explosion by 1% for the traditional multistep Arrhenius approach, and up to 11% using a Prout-Tompkins cookoff model.