SEARCH · Engineering Papers

Results for “variance estimation”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 91 records · Page 5

All sources NASA NTRS DOE OSTI

NASA NTRS1989

Stress distribution and topography of Tellus Regio, Venus

The Tellus Regio area of Venus represents a subset of a narrow latitude band where Pioneer Venus Orbiter (PVO) altimetry data, line-of-sight (LOS) gravity data, and Venera 15/16 radar images have all been obtained with good resolution. Tellus Regio also has a wide variety of surface morphologic features, elevations ranging up to 2.5 km, and a relatively low LOS gravity anomaly. This area was therefore chosen in order to examine the theoretical stress distributions resulting from various models of compensation of the observed topography. These surface stress distributions are then compared with the surface morphology revealed in the Venera 15/16 radar images. Conclusions drawn from these comparisons will enable constraints to be put on various tectonic parameters relevant to Tellus Regio. The stress distribution is calculated as a function of the topography, the equipotential anomaly, and the assumed model parameters. The topography data is obtained from the PVO altimetry. The equipotential anomaly is estimated from the PVO LOS gravity data. The PVO LOS gravity represents the spacecraft accelerations due to mass anomalies within the planet. These accelerations are measured at various altitudes and angles to the local vertical and therefore do not lend themselves to a straightforward conversion. A minimum variance estimator of the LOS gravity data is calculated, taking into account the various spacecraft altitudes and LOS angles and using the measured PVO topography as an a priori constraint. This results in an estimated equivalent surface mass distribution, from which the equipotential anomaly is determined.

Williams, David R.↗

NASA NTRS1991

Estimating The Cost Of Developing Software

Software Cost Estimation Model program, SOFTCOST, developed to provide consistent automated resource-and-schedule mathematical model more formalized than guesswork model. Combines several software-cost models found in open literature into one comprehensive set of algorithms compensating for nearly 50 implementation factors relative to size of task, inherited baseline, organizational and system environment, and difficulty of task. Produces mean and variance estimates of software size, implementation productivity, recommended staff level, probable duration, amount of computer resources required, and amount and cost of software documentation. Written in Microsoft BASIC.

Tausworthe, Robert C.↗

DOE OSTI2022

A new multi-model absolute difference-based sensitivity (MMADS) analysis method to screen non-influential processes under process model and parametric uncertainty

Process-based models have been widely used for hydrologic modeling, and it is a common practice to use sensitivity analysis methods for excluding non-influential hydrologic processes from further investigation and/or model improvement. This study develops a new method called multi-model absolute difference-based sensitivity (MMADS) analysis method to screen non-influential system processes and parameters. MMADS is conceptually similar to the Morris method for addressing parametric uncertainty, but has a unique feature to address both process model uncertainty (i.e., a process may be represented by multiple process models) and process model parameter uncertainty (i.e., parameters associated with a process model are random). MMADS first evaluates absolute differences of a quantity of interest (i.e., a system model output) by varying process models and/or process model parameter values, and then calculates the mean and variance of the differences for investigating process influence. The mean measures overall influence of the process on the quantity of interest, and the variance estimates influence of nonlinear effects of the process and/or its interactions with other processes. MMADS is an extension of the Morris method from a parameter space to a joint parameter-model space for explicitly addressing both process model uncertainty and model parameter uncertainty. The performance of MMADS is evaluated by using two numerical experiments. One experiment is based on Sobol’s G*-function with ten product elements, and has analytical solutions of the MMADS mean and variance of absolute differences. The other experiment is for groundwater flow modeling which considers three processes (i.e., recharge, geology, and snowmelt) that interact with each other. Finally, results indicate that MMADS is computationally efficient and can identify non-influential processes of complex hydrological systems.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2006

Estimation of Stability and Control Derivatives of an F-15

A technique for real-time estimation of stability and control derivatives (derivatives of moment coefficients with respect to control-surface deflection angles) was used to support a flight demonstration of a concept of an indirect-adaptive intelligent flight control system (IFCS). Traditionally, parameter identification, including estimation of stability and control derivatives, is done post-flight. However, for the indirect-adaptive IFCS concept, parameter identification is required during flight so that the system can modify control laws for a damaged aircraft. The flight demonstration was carried out on a highly modified F-15 airplane (see Figure 1). The main objective was to estimate the stability and control derivatives of the airplane in nearly real time. A secondary goal was to develop a system to automatically assess the quality of the results, so as to be able to tell a learning neural network which data to use. Parameter estimation was performed by use of Fourier-transform regression (FTR) a technique developed at NASA Langley Research Center. FTR is an equation- error technique that operates in the frequency domain. Data are put into the frequency domain by use of a recursive Fourier transform for a discrete frequency set. This calculation simplifies many subsequent calculations, removes biases, and automatically filters out data beyond the chosen frequency range. FTR as applied here was tailored to work with pilot inputs, which produce correlated surface positions that prevent accurate parameter estimates, by replacing half the derivatives with predicted values. FTR was also set up to work only on a recent window of data, to accommodate changes in flight condition. A system of confidence measures was developed to identify quality-parameter estimates that a learning neural network could use. This system judged the estimates primarily on the basis of their estimated variances and of the level of aircraft response. The resulting FTR system was implemented in the Simulink software system and auto-coded in the C programming language for use on the Airborne Research Test System (ARTS II) computer installed in the F-15 airplane. The Simulink model was also used in a control room that utilizes the Ring Buffered Network Bus hardware and software, making it possible to evaluate test points during flights. In-flight parameter estimation was done for piloted and automated maneuvers, primarily at three test conditions. Figure 2 shows results for pitching moment due to symmetric stabilator actuations for a series of three pitch doublet maneuvers (in a doublet maneuver, a command to change attitude in a given direction by a given amount is followed immediately by a command to change attitude in the opposite direction by the same amount). A time window of 5 seconds was used. The portions of the curves shown in red are those that passed the confidence tests. The technique showed good convergence for most derivatives for both kinds of maneuvers - typically within a few seconds. The confidence tests were marginally successful, and it would be necessary to refine them for use in an IFCS.

Smith, Mark↗

NASA NTRSDate not supplied

Use of three-cornered hat error estimates in MERRA-2 to guide an improved reanalysis-Part 1

The three-cornered hat (3CH) method estimates the uncertainties of three different co-located model or observational data sets (Anthes and Rieckh, 2018; Sjoberg et al., 2021). Rieckh et al. (2021) used the 3CH method to compare the random error statistics of different global forecast and reanalysis models, as well as radio occultation (RO) and radiosonde observations. That study showed that the MERRA-2 reanalysis, while having smaller errors in the stratosphere than its predecessor MERRA, had larger errors in the troposphere than many of the other data sets analyzed. The MERRA-2 errors were particularly large in the tropics. In a collaborative effort between UCAR’s COSMIC (Constellation Observing System for Meteorology, Ionosphere and Meteorology) program and NASA’s Global Modeling and Assimilation Office (GMAO), we carried out further 3CH error diagnostics to help isolate the causes of these larger errors and help guide the development of an improved reanalysis. This presentation summarizes random error statistics associated with MERRA-2, ECMWF’s ERA5 reanalysis, and COSMIC-2 (C2) RO observations. We compute 3CH error variance estimates of refractivity, as well as temperature and specific humidity using UCAR’s COSMIC Data Analysis and Archive Center (CDAAC) improved 1D-variational (1D-Var) retrieval (wetPf2) over 15 latitude bands from 45S to 45N. The 1D-Var retrievals of specific humidity and temperature for C2 use NCEP’s Global Forecast System (GFS) as the background. Anthes et al. (2021) showed that it gives accurate estimates of temperature and specific humidity in the tropics and subtropics, even in the challenging environment of intense Hurricane Dorian (2019). This presentation confirms the previous results that MERRA-2 has significantly larger errors in the tropics and subtropics than either C2 or ERA5. Its errors are larger between 30S and 30N compared to 30-45 N-S latitudes, and are also larger over land compared to oceans. Most of the MERRA-2 refractivity errors come from specific humidity, except over land below 3 km where temperature errors are large. These results suggest that moist convection and atmospheric boundary layer physics in MERRA-2 may be responsible for a significant part of the higher uncertainties. These results are being used to guide GMAO in developing an improved next-generation reanalysis, as shown in a companion presentation submitted to this conference (El Akkraoui et al., 2021), which extends this study and describes improvements to MERRA-2 leading to the next GMAO reanalysis.

Jeremiah Sjoberg↗

DOE OSTI2020

Accounting for Training Data Error in Machine Learning Applied to Earth Observations

Remote sensing, or Earth Observation (EO), is increasingly used to understand Earth system dynamics and create continuous and categorical maps of biophysical properties and land cover, especially based on recent advances in machine learning (ML). ML models typically require large, spatially explicit training datasets to make accurate predictions. Training data (TD) are typically generated by digitizing polygons on high spatial-resolution imagery, by collecting in situ data, or by using pre-existing datasets. TD are often assumed to accurately represent the truth, but in practice almost always have error, stemming from (1) sample design, and (2) sample collection errors. The latter is particularly relevant for image-interpreted TD, an increasingly commonly used method due to its practicality and the increasing training sample size requirements of modern ML algorithms. TD errors can cause substantial errors in the maps created using ML algorithms, which may impact map use and interpretation. Despite these potential errors and their real-world consequences for map-based decisions, TD error is often not accounted for or reported in EO research. Here we review the current practices for collecting and handling TD. We identify the sources of TD error, and illustrate their impacts using several case studies representing different EO applications (infrastructure mapping, global surface flux estimates, and agricultural monitoring), and provide guidelines for minimizing and accounting for TD errors. To harmonize terminology, we distinguish TD from three other classes of data that should be used to create and assess ML models: training reference data, used to assess the quality of TD during data generation; validation data, used to iteratively improve models; and map reference data, used only for final accuracy assessment. We focus primarily on TD, but our advice is generally applicable to all four classes, and we ground our review in established best practices for map accuracy assessment literature. EO researchers should start by determining the tolerable levels of map error and appropriate error metrics. Next, TD error should be minimized during sample design by choosing a representative spatio-temporal collection strategy, by using spatially and temporally relevant imagery and ancillary data sources during TD creation, and by selecting a set of legend definitions supported by the data. Furthermore, TD error can be minimized during the collection of individual samples by using consensus-based collection strategies, by directly comparing interpreted training observations against expert-generated training reference data to derive TD error metrics, and by providing image interpreters with thorough application-specific training. We strongly advise that TD error is incorporated in model outputs, either directly in bias and variance estimates or, at a minimum, by documenting the sources and implications of error. TD should be fully documented and made available via an open TD repository, allowing others to replicate and assess its use. To guide researchers in this process, we propose three tiers of TD error accounting standards. Finally, we advise researchers to clearly communicate the magnitude and impacts of TD error on map outputs, with specific consideration given to the likely map audience.

58 GEOSCIENCES↗

DOE OSTI2020

Uncertainty amplification due to density/refractive index gradients in background-oriented schlieren experiments

Here, we theoretically analyze the effect of density/refractive index gradients on the measurement precision of background-oriented schlieren (BOS) experiments by deriving the Cramer–Rao lower bound (CRLB) for the 2D centroid estimation process. A model is derived for the diffraction limited image of a dot viewed through a medium containing density gradients that includes the effect of the experimental parameters such as the magnification and f-number. It is shown using the model that nonlinearities in the density gradient field lead to blurring of the dot image. This blurring amplifies the effect of image noise on the centroid estimation process, leading to an increase in the CRLB and a decrease in the measurement precision. The ratio of position uncertainties of a dot in the reference and gradient images is shown to be a function of the ratio of the dot diameters and dot intensities. We termed this parameter the amplification ratio (A F ), and a methodology for reporting position uncertainties in tracking-based BOS measurements is proposed. The theoretical predictions of the dot position estimation variance from the CRLB are compared to ray tracing simulations, and agreement is obtained. The uncertainty amplification is also demonstrated on experimental BOS images of flow induced by a spark discharge, where it is seen that regions of high amplification ratio correspond to regions of density gradients. This analysis elucidates the dependence of the position uncertainty on density and refractive index gradient-induced distortion parameters, provides a methodology for accounting its effect on uncertainty quantification and provides a framework for optimizing experiment design.

42 ENGINEERING↗

DOE OSTI2019

Inverse regression for ridge recovery: a data-driven approach for parameter reduction in computer experiments

Parameter reduction can enable otherwise infeasible design and uncertainty studies with modern computational science models that contain several input parameters. In statistical regression, techniques for sufficient dimension reduction (SDR) use data to reduce the predictor dimension of a regression problem. A computational scientist hoping to use SDR for parameter reduction encounters a problem: a computer prediction is best represented by a deterministic function of the inputs, so data comprised of computer simulation queries fail to satisfy the SDR assumptions. To address this problem, here we interpret SDR methods sliced inverse regression (SIR) and sliced average variance estimation (SAVE) as estimating the directions of a ridge function, which is a composition of a low-dimensional linear transformation with a nonlinear function. Within this interpretation, SIR and SAVE estimate matrices of integrals whose column spaces are contained in the ridge directions’ span; we analyze and numerically verify convergence of these column spaces as the number of computer model queries increases. Moreover, we show example functions that are not ridge functions but whose inverse conditional moment matrices are low-rank. Consequently, the computational scientist should beware when using SIR and SAVE for parameter reduction, since SIR and SAVE may mistakenly suggest that truly important directions are unimportant.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Adaptive Algebraic Derivative Estimation for Battery Electric Buses Energy Consumption Forecasting

The limited service life of onboard batteries for EVs is a challenge, underscoring the need for real-time battery usage prediction. This paper proposes an adaptive Algebraic Derivative Estimation (ADE) approach for forecasting the energy consumption of battery electric buses. By dynamically adjusting the sliding window length, the adaptive ADE retains the fixed-length ADE’s key advantage—namely, operating online without reliance on extensive historical datasets—while substantially bolstering forecast accuracy by actively trading estimation bias off estimation variance. Comparative experiments against both the conventional ADE with a fixed length and a representative machine learning algorithm, XGBoost, were conducted, with performance evaluated via root mean square error, mean absolute error, and the coefficient of determination. The results demonstrate that the proposed approach significantly outperforms baseline methods.

Cui, Tianyang [The University of Texas at Dallas]↗

DOE OSTI2021

A panel analysis of groundwater use in California

We report groundwater is a relevant source of drinking and agricultural water in many regions of the world, but many aquifers have been unsustainably over-drafted and polluted. This has significant environmental, health, and economic implications. We rely on panel analysis, with small-sample corrections for cluster-robust variance estimation and hypothesis testing, to investigate the dynamics of groundwater extraction. We focus on California, yet our approach could be helpful to analyze the dynamics of groundwater extraction in other groundwater-reliant regions of the world. In California, over-reliance on groundwater has led to significant overdraft, affecting long-term water supply reliability and groundwater pumping costs. It further caused subsidence and infrastructure damage, harmed groundwater-dependent ecosystems, and threatened the sustainability of groundwater resources in the state. We use panel data of the 56 California Water Plan planning areas over the 1998–2015 period. We concentrate on agricultural and urban water use and the major water projects in the state, to provide a better understanding of the relationships between groundwater extraction and water use and supply. Results suggest that reducing agricultural water in Central California and urban water in Southern California could reduce groundwater extraction in these regions by approximately the same amount of the reduced water. Other opportunities to reduce the stress on the groundwater resources in the state are available for other regions, yet with lower benefits. Results also suggest that a decrease in deliveries from the Central Valley Project to the southern part of the Central Valley would increase groundwater extraction by approximately the same proportion. Changes in deliveries from the major water projects in the state, as well as from other sources of surface water, would also have some, yet lower impacts on groundwater extraction in the Central and Southern California.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2018

A Test of the Tropical 200- to 300-hPa Warming Rate in Climate Models

Overall climate sensitivity to CO 2 doubling in a general circulation model results from a complex system of parameterizations in combination with the underlying model structure. We refer to this as the model’s “major hypothesis” and we assume it to be testable. We explain four criteria a valid test should meet: measurability, specificity, independence and uniqueness. We argue that temperature change in the tropical 200-300 hPa layer meets these criteria. Comparing modeled to observed trends over the past 60 years using a persistence-robust variance estimator shows that all models warm more rapidly than observations and in the majority of individual cases the discrepancy is statistically significant. We argue that this provides informative evidence against the major hypothesis in most current climate models.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2023

Stochastic gradient descent for optimization for nuclear systems

The use of gradient descent methods for optimizing k-eigenvalue nuclear systems has been shown to be useful in the past, but the use of k-eigenvalue gradients have proved computationally challenging due to their stochastic nature. ADAM is a gradient descent method that accounts for gradients with a stochastic nature. This analysis uses challenge problems constructed to verify if ADAM is a suitable tool to optimize k-eigenvalue nuclear systems. ADAM is able to successfully optimize nuclear systems using the gradients of k-eigenvalue problems despite their stochastic nature and uncertainty. Furthermore, it is clearly demonstrated that low-compute time, high-variance estimates of the gradient lead to better performance in the optimization challenge problems tested here.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2026

A variational framework for residual-based adaptivity in neural PDE solvers and operator learning

Residual-based adaptive strategies are widely used in scientific machine learning yet remain largely heuristic. We introduce a variational framework that formalizes these methods through convex transformations of the residual, where different transformations correspond to distinct objective functionals. For instance, exponential weights target uniform error minimization, while linear weights recover quadratic error minimization. This perspective reveals adaptive weighting as a means of selecting sampling distributions that optimize a primal objective, directly linking discretization choices to error metrics. This principled approach yields three key benefits: it enables systematic design of adaptive schemes, reduces discretization error by lowering estimator variance, and enhances learning dynamics by improving gradient signal-to-noise ratio. Extending the framework to operator learning, we demonstrate substantial performance gains across diverse optimizers and architectures. Our results provide a theoretical perspective for residual-based adaptivity and establish a foundation for principled discretization and training.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Planck 2018 results - VIII. Gravitational lensing

We present measurements of the cosmic microwave background (CMB) lensing potential using the final Planck 2018 temperature and polarization data. Using polarization maps filtered to account for the noise anisotropy, we increase the significance of the detection of lensing in the polarization maps from 5σ to 9σ. Combined with temperature, lensing is detected at 40σ. We present an extensive set of tests of the robustness of the lensing-potential power spectrum, and construct a minimum-variance estimator likelihood over lensing multipoles 8 ≤ L ≤ 400 (extending the range to lower L compared to 2015), which we use to constrain cosmological parameters. We find good consistency between lensing constraints and the results from the Planck CMB power spectra within the ΛCDM model. Combined with baryon density and other weak priors, the lensing analysis alone constrains σ 8 Ωm 0.25 = 0.589 ± 0.020 (1σ errors). Also combining with baryon acoustic oscillation data, we find tight individual parameter constraints, σ8 = 0.811 ± 0.019, H0 = 67.9-1.3+1.2 km s -1 Mpc -1 , and Ω m = 0.303 -0.018 +0.016 . Combining with Planck CMB power spectrum data, we measure σ 8 to better than 1% precision, finding σ 8 = 0.811 ± 0.006. CMB lensing reconstruction data are complementary to galaxy lensing data at lower redshift, having a different degeneracy direction in σ 8 - Ω m space; we find consistency with the lensing results from the Dark Energy Survey, and give combined lensing-only parameter constraints that are tighter than joint results using galaxy clustering. Using the Planck cosmic infrared background (CIB) maps as an additional tracer of high-redshift matter, we make a combined Planck-only estimate of the lensing potential over 60% of the sky with considerably more small-scale signal. We additionally demonstrate delensing of the Planck power spectra using the joint and individual lensing potential estimates, detecting a maximum removal of 40% of the lensing-induced power in all spectra. The improvement in the sharpening of the acoustic peaks by including both CIB and the quadratic lensing reconstruction is detected at high significance.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2023

Recipes for when physics fails: recovering robust learning of physics informed neural networks

Abstract Physics-informed neural networks (PINNs) have been shown to be effective in solving partial differential equations by capturing the physics induced constraints as a part of the training loss function. This paper shows that a PINN can be sensitive to errors in training data and overfit itself in dynamically propagating these errors over the domain of the solution of the PDE. It also shows how physical regularizations based on continuity criteria and conservation laws fail to address this issue and rather introduce problems of their own causing the deep network to converge to a physics-obeying local minimum instead of the global minimum. We introduce Gaussian process (GP) based smoothing that recovers the performance of a PINN and promises a robust architecture against noise/errors in measurements. Additionally, we illustrate an inexpensive method of quantifying the evolution of uncertainty based on the variance estimation of GPs on boundary data. Robust PINN performance is also shown to be achievable by choice of sparse sets of inducing points based on sparsely induced GPs. We demonstrate the performance of our proposed methods and compare the results from existing benchmark models in literature for time-dependent Schrödinger and Burgers’ equations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

New Probe of Cosmic Birefringence Using Galaxy Polarization and Shapes

We propose a novel statistical method to measure cosmic birefringence and demonstrate its power in probing parity violation due to axions. Exploiting an empirical correlation between the integrated radio polarization direction of a spiral galaxy and its apparent shape, we devise an unbiased minimum-variance estimator for the rotation angle, which should achieve an uncertainty of 5°–15° per galaxy. In conclusion, large galaxy samples from the forthcoming SKA continuum surveys, together with optical shape catalogs, promise a comparable or even lower noise power spectrum for the rotation angle than in the CMB Stage-IV (CMB-S4) experiment, with different systematics.

Axion-like particles↗

DOE OSTI2022

Key environmental and production factors for understanding variation in switchgrass chemical attributes

Abstract Switchgrass ( Panicum virgatum L.) is a promising feedstock for bioenergy and bioproducts; however, its inherent variability in chemical attributes creates challenges for uniform conversion efficiencies and product quality. It is necessary to understand the range of variation and factors (i.e., field management, environmental) influencing chemical attributes for process improvement and risk assessment. The objectives of this study were to (1) examine the impact of nitrogen fertilizer application rate, year, and location on switchgrass chemical attributes, (2) examine the relationships among chemical attributes, weather and soil data, and (3) develop models to predict chemical attributes using environmental factors. Switchgrass samples from a field study spanning four locations including upland cultivars, one location including a lowland cultivar, and between three and six harvest years were assessed for glucan, xylan, lignin, volatiles, carbon, nitrogen, and ash concentrations. Using variance estimation, location/cultivar, nitrogen application rate, and year explained 65%–96% of the variation for switchgrass chemical attributes. Location/cultivar × year interaction was a significant factor for all chemical attributes indicating environmental‐based influences. Nitrogen rate was less influential. Production variables and environmental conditions occurring during the switchgrass field trials were used to successfully predict chemical attributes using linear regression models. Upland switchgrass results highlight the complexity in plant responses to growing conditions because all production and environmental variables had strong relationships with one or more chemical attributes. Lowland switchgrass was limited to observations of year‐to‐year environmental variability and nitrogen application rate. All explanatory variable categories were important for lowland switchgrass models but stand age and precipitation relationships were particularly strong. The relationships found in this study can be used to understand spatial and temporal variation in switchgrass chemical attributes. The ability to predict chemical attributes critical for conversion processes in a geospatial/temporal manner would provide state‐of‐the‐art knowledge for risk assessment in the bioenergy and bioproducts industry.

09 BIOMASS FUELS↗

DOE OSTI2022

Neural Network-Based Electric Vehicle Range Prediction for Smart Charging Optimization

Range prediction is a standard feature in most modern road vehicles, allowing drivers to make informed decisions about when to refuel. Most vehicles make range predictions through data- or model-driven means, monitoring the average fuel consumption rate or using a tuned vehicle model to predict fuel consumption. The uncertainty of future driving conditions makes the range prediction problem challenging, particularly for less pervasive battery electric vehicles (BEV). Most contemporary machine learning-based methods attempt to forecast the battery SOC discharge profile to predict vehicle range. In this work, we propose a novel approach using two recurrent neural networks (RNNs) to predict the remaining range of BEVs and the minimum charge required to safely complete a trip. Each RNN has two outputs that can be used for statistical analysis to account for uncertainties; the first loss function leads to mean and variance estimation (MVE), while the second results in bounded interval estimation (BIE). These outputs of the proposed RNNs are then used to predict the probability of a vehicle completing a given trip without charging, or if charging is needed, the remaining range and minimum charging required to finish the trip with high probability. Training data was generated using a low-order physics model to estimate vehicle energy consumption from historical drive cycle data collected from medium-duty last-mile delivery vehicles. Here, the proposed method demonstrated high accuracy in the presence of day-to-day route variability, with the root-mean-square error (RMSE) below 6% for both RNN models.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗