Joseph F. DeCarolis

The case for repeatable analysis with energy economy optimization models

Energy economy optimization (EEO) models employ formal search techniques to explore the future decision space over several decades in order to deliver policy-relevant insights. EEO models are a critical tool for decision-makers who must make near-term decisions with long-term effects in the face of large future uncertainties. While the number of model-based analyses proliferates, insufficient attention is paid to transparency in model development and application. Given the complex, data-intensive nature of EEO models and the general lack of access to source code and data, many of the assumptions underlying model-based analysis are hidden from external observers. This paper discusses the simplifications and subjective judgments involved in the model building process, which cannot be fully articulated in journal papers, reports, or model documentation. In addition, we argue that for all practical purposes, EEO model-based insights cannot be validated through comparison to real world outcomes. As a result, modelers are left without credible metrics to assess a models ability to deliver reliable insight. We assert that EEO models should be discoverable through interrogation of publicly available source code and data. In addition, third parties should be able to run a specific model instance in order to independently verify published results. Yet a review of twelve EEO models suggests that in most cases, replication of model results is currently impossible. We provide several recommendations to help develop and sustain a software framework for repeatable model analysis.

Measuring radio frequency properties of materials in pulsed magnetic fields with a tunnel diode oscillator

Tunnel diode oscillators have been used in many types of experiments that measure the properties of materials. We present the details of an apparatus that extend these tunnel diode techniques to measure the properties of materials in pulsed magnetic fields. In the most common version of this method, a sample is placed in the inductor of a small rf tank circuit powered by a tunnel diode and the conductivity, magnetization, or penetration depth is measured. We explain in this article how the sample and configuration of the radio frequency fields determine which property is measured. Our major innovations are to stabilize the tunnel diode oscillator during a magnet pulse by using compensated coils in the tank circuit and the development of two methods, one digital and one analog, to measure the frequency and amplitude shifts in the oscillator during the short (10 s of ms) magnet pulse. We illustrate the power of this new measurement method by showing preliminary results of the superconducting transition and ...

Using observed data to improve estimated methane collection from select U.S. landfills.

The anaerobic decomposition of solid waste in a landfill produces methane, a potent greenhouse gas, and if recovered, a valuable energy commodity. Methane generation from U.S. landfills is usually estimated using the U.S. EPAs Landfill Gas Emissions Model (LandGEM). Default values for the two key parameters within LandGEM, the first-order decay rate (k) and the methane production potential (L0) are based on data collected in the 1990s. In this study, observed methane collection data from 11 U.S. landfills and estimates of gas collection efficiencies developed from site-specific gas well installation data were included in a reformulated LandGEM equation. Formal search techniques were employed to optimize k for each landfill to find the minimum sum of squared errors (SSE) between the LandGEM prediction and the observed collection data. Across nearly all landfills, the optimal k was found to be higher than the default AP-42 of 0.04 yr(-1) and the weighted average decay for the 11 landfills was 0.09 - 0.12 yr(-1). The results suggest that the default k value assumed in LandGEM is likely too low, which implies that more methane is produced in the early years following waste burial when gas collection efficiencies tend to be lower.

How Much Do Electric Drive Vehicles Matter to Future U.S. Emissions

Hybrid, plug-in hybrid, and battery electric vehicles--known collectively as electric drive vehicles (EDVs)--may represent a clean and affordable option to meet growing U.S. light duty vehicle (LDV) demand. The goal of this study is 2-fold: identify the conditions under which EDVs achieve high LDV market penetration in the U.S. and quantify the associated change in CO2, SO2, and NOX emissions through midcentury. We employ the Integrated MARKAL-EFOM System (TIMES), a bottom-up energy system model, along with a U.S. data set developed for this analysis. To characterize EDV deployment through 2050, varying assumptions related to crude oil and natural gas prices, a CO2 policy, a federal renewable portfolio standard, and vehicle battery cost were combined to form 108 different scenarios. Across these scenarios, oil prices and battery cost have the biggest effect on EDV deployment. The model results do not demonstrate a clear and consistent trend toward lower system-wide emissions as EDV deployment increases. In addition to the trade-off between lower tailpipe and higher electric sector emissions associated with plug-in vehicles, the scenarios produce system-wide emissions effects that often mask the effect of EDV deployment.

Analysis of material recovery facilities for use in life-cycle assessment

Insights derived from life-cycle assessment of solid waste management strategies depend critically on assumptions, data, and modeling at the unit process level. Based on new primary data, a process model was developed to estimate the cost and energy use associated with material recovery facilities (MRFs), which are responsible for sorting recyclables into saleable streams and as such represent a key piece of recycling infrastructure. The model includes four modules, each with a different process flow, for separation of single-stream, dual-stream, pre-sorted recyclables, and mixed-waste. Each MRF type has a distinct combination of equipment and default input waste composition. Model results for total amortized costs from each MRF type ranged from

Characterization of Uncertainty in Estimation of Methane Collection from Select U.S. Landfills

19.8 to

Systematic exploration of efficient strategies to manage solid waste in U.S. municipalities: perspectives from the solid waste optimization life-cycle framework (SWOLF).

24.9 per Mg (1Mg=1 metric ton) of waste input. Electricity use ranged from 4.7 to 7.8kWh per Mg of waste input. In a single-stream MRF, equipment required for glass separation consumes 28% of total facility electricity consumption, while all other pieces of material recovery equipment consume less than 10% of total electricity. The dual-stream and mixed-waste MRFs have similar electricity consumption to a single-stream MRF. Glass separation contributes a much larger fraction of electricity consumption in a pre-sorted MRF, due to lower overall facility electricity consumption. Parametric analysis revealed that reducing separation efficiency for each piece of equipment by 25% altered total facility electricity consumption by less than 4% in each case. When model results were compared with actual data for an existing single-stream MRF, the model estimated the facilitys electricity consumption within 2%. The results from this study can be integrated into LCAs of solid waste management with system boundaries that extend from the curb through final disposal.

Evaluating the US Mid-Century Strategy for Deep Decarbonization amidst early century uncertainty

Methane is a potent greenhouse gas generated from the anaerobic decomposition of waste in landfills. If captured, methane can be beneficially used to generate electricity. To inventory emissions and assist the landfill industry with energy recovery projects, the U.S. EPA developed the Landfill Gas Emissions Model (LandGEM) that includes two key parameters: the first-order decay rate (k) and methane production potential (L0). By using data from 11 U.S. landfills, Monte Carlo simulations were performed to quantify the effect of uncertainty in gas collection efficiency and municipal solid waste fraction on optimal k values and collectable methane. A dual-phase model and associated parameters were also developed to evaluate its performance relative to a single-phase model (SPM) similar to LandGEM. The SPM is shown to give lower error in estimating methane collection, with site-specific best-fit k values. Most of the optimal k values are notably greater than the U.S. EPAs default of 0.04 yr(-1), which implies that the gas generation decreases more rapidly than predicted at the current default. We translated the uncertainty in collectable methane into uncertainty in engine requirements and potential economic losses to demonstrate the practical significance to landfill operators. The results indicate that landfill operators could overpay for engine capacity by

Systematic Evaluation of Industrial, Commercial, and Institutional Food Waste Management Strategies in the United States

30,000-780,000 based on overestimates of collectable methane.

US Energy-Related Greenhouse Gas Emissions in the Absence of Federal Climate Policy

Solid waste management (SWM) systems must proactively adapt to changing policy requirements, waste composition, and an evolving energy system to sustainably manage future solid waste. This study represents the first application of an optimizable dynamic life-cycle assessment framework capable of considering these future changes. The framework was used to draw insights by analyzing the SWM system of a hypothetical suburban U.S. city of 100 000 people over 30 years while considering changes to population, waste generation, and energy mix and costs. The SWM system included 3 waste generation sectors, 30 types of waste materials, and 9 processes for waste separation, treatment, and disposal. A business-as-usual scenario (BAU) was compared to three optimization scenarios that (1) minimized cost (Min Cost), (2) maximized diversion (Max Diversion), and (3) minimized greenhouse gas (GHG) emissions (Min GHG) from the system. The Min Cost scenario saved