The dynamics of microbial communities involved in anaerobic digestion of mixed organic waste are notoriously complex and difficult to model, yet successful operation of anaerobic digestion is critical to the goals of diverting high-moisture organic waste from landfills. Machine learning (ML) is ideally suited to capturing complex and nonlinear behavior that cannot be modeled mechanistically. This study uses 8 years of data collected from an industrial-scale anaerobic co-digestion (AcoD) operation at a municipal wastewater treatment plant in Oakland, California, combined with a powerful automated ML method, Tree-based Pipeline Optimization Tool, to develop an improved understanding of how different waste inputs and operating conditions impact biogas yield. The model inputs included daily input volumes of 31 waste streams and 5 operating parameters. Because different wastes are broken down at varying rates, the model explored a range of time lags ascribed to each waste input ranging from 0 to 30 days. The results suggest that the waste types (including rendering waste, lactose, poultry waste, and fats, oils, and greases) differ considerably in their impact on biogas yield on both a per-gallon basis and a mass of volatile solids basis, while operating parameters were not good predictors of yield at this facility.

The California Public Utilities Commission (CPUC) sought support for the development of database schema specifications for a bi-directional data portal that would foster cross-jurisdictional collaboration necessary to improve resilient energy infrastructure planning processes. This technical assistance activity involved Berkeley Lab researchers reviewing 34 local/tribal government hazard mitigation plans to assess (1) the natural hazards that communities are most concerned about; (2) the variety of–and terminology used to describe–critical community infrastructure; and (3) the availability of GIS information that could be incorporated into CPUC-mandated "Microgrid Planning Portals". In addition, we develop a common, but generic data taxonomy showing what fields to collect to encourage consolidating and sharing of this information in the future. Finally, we partnered with the Bishop Paiute tribal government to demonstrate the value of combining electricity infrastructure, natural hazard layers, and critical community infrastructure into a series of maps. The project demonstrated that a single system containing both electric utility and community infrastructure data—as well as information about natural hazards—will help the state of California, the IOUs, first responders, and long-term planners better prepare for—and thus lower their exposure to these ongoing and emerging hazards. The project also identified a number of challenges that will need to be overcome before it is possible to stand up a single system to display this information.

Abstract: High-potential molecules derived from biomass sources may suitably replace or supple-ment traditional nonrenewable hydrocarbon fuels to reduce pollution and fuel processing cost. Ex-perimental property testing of these bioproducts is usually conducted years after initial bench-scale experiments, due to high experimental costs and/or high volume requirements. However, neglecting to conduct property testing early in the pathway development cycle can lead to investments spent on scaling-up production of bioproducts and biofuels that do not perform as expected. Instead, machine-learning techniques can be used to develop quantitative structure–property relationships for molecules using a relatively large training set of molecular descriptor data. For this study, we compiled measured properties, IR spectra, and molecular descriptors of bio-based molecules from databases and published studies for training models of bioproduct properties. We trained regres-sion models with molecular descriptors and will compare results of different estimators. This study describes the first steps towards a performance prediction tool for bio-based alternative fuels. Keywords: Machine learning, biofuels, jet fuels, fuel properties

The Renewable Fuel Standard (RFS) initially set ambitious goals for US cellulosic biofuel production and, although the total renewable fuel volume reached 80% of the established target for 2017, the cellulosic fuel volume reached just 5% of the original goal. This shortfall has, in part, been ascribed to the hesitance of farmers to plant the high-yielding, low-input perennial biomass crops identified as otherwise ideal feedstocks. Policy and market uncertainty also hinder investment in capital-intensive new cellulosic biorefineries. This study combines remote sensing land use data, yield predictions, a fine-resolution geospatial modeling framework, and a novel facility siting algorithm to evaluate the potential for near-term scale-up of cellulosic fuel production using a combination of lower-risk annual feedstocks more familiar to US farmers: corn stover and biomass sorghum. Potential strategies include expansion or retrofitting of existing corn ethanol facilities and targeted construction of new facilities in resource-rich areas. The results indicate that, with a maximum 10% conversion of pastureland and cropland to sorghum in suitable regions, more than 80 of the 214 existing corn ethanol biorefineries could be retrofitted or expanded to accept cellulosic feedstocks and an additional 71 new biorefineries could be built. The resulting land conversion for bioenergy sorghum totals to 4.5% of US cropland and 3.7% of pastureland. If this biomass is converted to ethanol, the total increase in annual production could be 17 billion gallons, just over the original RFS 2022 cellulosic biofuel production target and equivalent to 12% of US gasoline consumption.

Technoeconomic analysis (TEA) is an approach for conducting process design and simulation, informed by empirical data, to estimate capital costs, operating costs, mass balances, and energy balances for a commercial scale biorefinery. TEA serves as a useful method to screen potential research priorities, identify cost bottlenecks at the earliest stages of research, and provide the mass and energy data needed to conduct life-cycle environmental assessments. Recent studies have produced new tools and methods to enable faster iteration on potential designs, more robust uncertainty analysis, and greater accessibility through the use of open-source platforms. There is also a trend toward more expansive system boundaries to incorporate the impact of policy incentives, use-phase performance differences, and potential impacts on global market supply.

Technoeconomic analysis and life-cycle assessment are critical to guiding and prioritizing bench-scale experiments and to evaluating economic and environmental performance of biofuel or biochemical production processes at scale. Traditionally, commercial process simulation tools have been used to develop detailed models for these purposes. However, developing and running such models can be costly and computationally intensive, which limits the degree to which they can be shared and reproduced in the broader research community. This study evaluates the potential of an automated machine learning approach to develop surrogate models based on conventional process simulation models. The analysis focuses on several high-value biofuels and bioproducts for which pathways of production from biomass feedstocks have been well-established. The results demonstrate that surrogate models can be an accurate and effective tool for approximating the cost, mass and energy balance outputs of more complex process simulations at a fraction of the computational expense.

Machine learning has proven to be a powerful tool for accelerating biofuel development. Although numerous models are available to predict a range of properties using chemical descriptors, there is a trade-off between interpretability and performance. Neural networks provide predictive models with high accuracy at the expense of some interpretability, while simpler models such as linear regression often lack in accuracy. In addition to model architecture, feature selection is also critical for developing interpretable and accurate predictive models. We present a method for systematically selecting molecular descriptor features and developing interpretable machine learning models without sacrificing accuracy. Our method simplifies the process of selecting features by reducing feature multicollinearity and enables discoveries of new relationships between global properties and molecular descriptors. To demonstrate our approach, we developed models for predicting melting point, boiling point, flash point, yield sooting index, and net heat of combustion with the help of the Tree-based Pipeline Optimization Tool (TPOT). For training, we used publicly available experimental data for up to 8351 molecules. Our models accurately predict various molecular properties for organic molecules (mean absolute percent error (MAPE) ranges from 3.3% to 10.5%) and provide a set of features that are well-correlated to the property. This method enables researchers to explore sets of features that significantly contribute to the prediction of the property, offering new scientific insights. To help accelerate early stage biofuel research and development, we also integrated the data and models into a open-source, interactive web tool.

In 2019, the United States consumed over 57 million metric tons (MMT) of plastic with less than 7% recovered for reuse. This study provides an updated material flow analysis at national and regional scales for all durable and single-use plastics in the United States. From this material flow analysis, we develop a series of alternative future national plastic flow scenarios that envision a scale-up of recycling technologies, incorporating technical limitations and sorting infrastructure constraints. The results suggest that a maximum of 68% (24 MMT) of plastic waste could be diverted from landfills by scaling up existing commercial recycling technologies. Based on the current technological landscape, reaching near-zero waste is only possible if processes that are operating at pilot and laboratory scales can be effectively scaled and coupled with improved sorting infrastructure. Through these scenarios with increased recycling, the availability of postconsumer resin stocks could increase by 22–43 MMT.

Sorghum is an attractive feedstock for biobased fuel and chemical production because it is familiar to farmers, naturally drought tolerant, and versatile as a food, feed, and fuel crop. Although sorghum is a promising feedstock, particularly in regions that experience drought stress, little is known about how drought conditions impact the ease of conversion of sorghum to fuels and products. This study combines agronomic field trials with a high-throughput experimental pipeline to explore the field performance and liquid biofuel (bisabolene) yields resulting from three sorghum types (photosensitive forage sorghum, optimized grain sorghum, and drought-resistant grain sorghum) grown under pre- and postflowering water limitations in two different California locations. Multiple drought treatments are compared to the control, as the timing (preflowering versus postflowering) of drought stress elicits different survival strategies and corresponding impacts on yield and composition. Forage-type sorghum maintained the highest biomass yields across all irrigation conditions and locations. Glucose and xylose yields resulting from ionic liquid pretreatment and enzymatic saccharification were not significantly impacted by irrigation treatments but differed by location and genotype. However, Rhodosporidium toruloides grown on the resulting plant hydrolysates unexpectedly produced higher titers of bisabolene for drought-stressed sorghum samples regardless of genotype.