Uncertainty Assessment Methodology Research Articles

Reconstruction, analysis and interpretation of posterior probability distributions of PET images, using the posterior bootstrap

The uncertainty of reconstructed PET images remains difficult to assess and to interpret for the use in diagnostic and quantification tasks. Here we provide (1) an easy-to-use methodology for uncertainty assessment for almost any Bayesian model in PET reconstruction from single datasets and (2) a detailed analysis and interpretation of produced posterior image distributions. We apply a recent posterior bootstrap framework to the PET image reconstruction inverse problem and obtain simple parallelizable algorithms based on random weights and on existing maximum a posteriori (MAP) (posterior maximum) optimization-based algorithms. Posterior distributions are produced, analyzed and interpreted for several common Bayesian models. Their relationship with the distribution of the MAP image estimate over multiple dataset realizations is exposed. The coverage properties of posterior distributions are validated. More insight is obtained for the interpretation of posterior distributions in order to open the way for including uncertainty information into diagnostic and quantification tasks.

Oilfield development system optimization under reservoir production uncertainty

An optimization approach is proposed to support the design decisions of floating production systems and their subsea layout, regarding the production and investment planning. The optimization methodology extends a previous approach by including the dynamics and uncertainty of the oil production during the reservoir lifetime, and the associated dynamics of the gas and water production. The new model integrates production forecast with facilities requirements and economic optimization. An uncertainty assessment methodology is provided and applied to the reservoir production including its decline curve. A case study from Brazil waters is presented and discussed to demonstrate the performance of the method applied to a specific production system.

Impact of thermal gradients on the geometry correction of a bridge coordinate measuring machine

If the problems related to the parts and measurement strategy of Coordinate Measuring Machines (CMMs) are not taken into consideration, temperature variations become the main source of measurement uncertainties. Indeed, they may cause variations in geometry as well as reference point drift. The effect of drift is sometimes minimized by CMM users and is not well quantified in general. The aim of this paper is to present a physical method to determine the evolution of CMM geometry and drift which is based directly on CMM temperature variations and construction parameters, i.e. the position of the axes measurement scales and reference points of each axis. The method is applied to a Zeiss CMM Contura G2. The consequences of these CMM evolutions are simulated in the measurement of a sphere generated by a Renishaw Machine Checking Gauge. The proposed method falls within the framework of an uncertainty assessment methodology performed by multi-level Monte Carlo simulation, where the first level corresponds to the characterization of the CMM evolution.

Data quality and uncertainty assessment methodology for pavement LCA

In performing pavement life cycle assessment (LCA), users are facing various reports of energy intensity coefficient (EIC) of pavement materials which differ considerably among data sources and therefore alter the LCA results significantly. Instead of selecting a certain EIC without or of little explanation for the current pavement LCA practices, this study proposed a methodology to build probability density function (PDF) for EIC based on available data-set and their qualities. Each data was first evaluated about the data quality indicator (DQI) through data quality pedigree matrix and converted to PDF in modified Beta distribution form. Three weighting methods, the DQI one, coefficient of variation (COV) one and analytical hierarchy process (AHP) one, were developed to assign weightings for different data. Monte Carlo simulation was run with the weighted PDF of each data as input to obtain the ultimate PDF for EIC. A case study to estimate the bitumen’s EIC with eight data samples were performed using the proposed methodology. It is found (1): the estimates by the proposed methodology is of higher reliability (lower COV) compared to any single data due to utilisation of information of the overall data samples; (2) the AHP weighting method is most favoured despite the results of the three weighting methods are close; (3) the central estimates of bitumen’s EIC are between5.4~5.8 MJ/kg. The proposed methodology is helpful in aiding calculating EICs for pavement materials and capturing uncertainties in LCA results in a statistical sense.

Different Approaches in Uncertainty Evaluation for Measurement of Complex Surfaces Using Coordinate Measuring Machine

Abstract This paper describes a methodology for uncertainty assessment for Coordinate Measuring Machine measurement of complex real work pieces from industry. The study applied two approaches (in scanning mode only) for estimating the measurement uncertainty with the support of Taguchi plan in the experiment containing five factors: scanning speed, sample density, probe configuration, scanning direction, and position of measuring object. In the first approach the uncertainty was estimated by measuring the basic geometric objects (primitives like sphere and torus) representing the decomposition of complex surfaces and in the second one a complex surface was treated as an unknown quantity. Calculated uncertainty Type A for both measurement tasks was in the range from 0.65 μm to 6.47 μm. Evaluation of the uncertainty Type B covered specifications of the machine and standard uncertainties derived from temperature effects. Total uB component was found to be in order of 0.4 μm. Future research will be directed towards the development and application of simulation methods

Simulation of Naturally Fractured Reservoirs. State of the Art

Naturally fractured reservoirs contain a significant amount of the world oil reserves. The production of this type of reservoirs constitutes a challenge for reservoir engineers. Use of reservoir simulators can help reservoir engineers in the understanding of the main physical mechanisms and in the choice of the best recovery process and its optimization. Significant progress has been made since the first publications on the dual-porosity concept in the sixties. This paper and the preceding one (Part 1) present the current techniques of modeling used in industrial simulators. The optimal way to predict matrix-fracture transfers at the simulator cell scale has no definite answer and various methods are implemented in industrial simulators. This paper focuses on the modeling of physical mechanisms driving flows and interactions/ exchanges within and between fracture and matrix media for a better understanding of proposed flow formula and simulation methods. Typical features of fractured reservoir numerical simulations are also described with an overview of the implementation of geomechanics effects, an application of uncertainty assessment methodology to a fractured gas reservoir and finally a presentation of a history matching methodology for fractured reservoirs.

Lignite Quality Uncertainty Estimation for the Assessment of CO2 Emissions

In this study, models were developed for the estimation of CO2 emissions from lignite- fired power plants based on routinely measured quality parameters of lignite and the power-plant efficiency. These models could be incorporated into the production scheduling optimization plan of the lignite mining. A probabilistic methodology for uncertainty assessment of CO2 emissions and assigning this uncertainty to different sources (lignite quality, activity, and operational data of power plants) was performed. A Monte Carlo uncertainty and sensitivity analyses method was found appropriate because of the strong correlations between input quality parameters. This method was applied to the largest power plant in Greece for the estimation of the uncertainties in annual, monthly, and daily CO2 emissions. Sensitivity analysis revealed that the variance of CO2 emissions was mainly determined by the uncertainty of the quality parameters of lignite, with carbon content being the most dominant parameter.

A statistical methodology for quantification of uncertainty in best estimate code physical models

A novel uncertainty assessment methodology, based on a statistical non-parametric approach, is presented in this paper. It achieves quantification of code physical model uncertainty by making use of model performance information obtained from studies of appropriate separate-effect tests. Uncertainties are quantified in the form of estimated probability density functions (pdf’s), calculated with a newly developed non-parametric estimator. The new estimator objectively predicts the probability distribution of the model’s ‘error’ (its uncertainty) from databases reflecting the model’s accuracy on the basis of available experiments. The methodology is completed by applying a novel multi-dimensional clustering technique based on the comparison of model error samples with the Kruskall–Wallis test. This takes into account the fact that a model’s uncertainty depends on system conditions, since a best estimate code can give predictions for which the accuracy is affected by the regions of the physical space in which the experiments occur. The final result is an objective, rigorous and accurate manner of assigning uncertainty to coded models, i.e. the input information needed by code uncertainty propagation methodologies used for assessing the accuracy of best estimate codes in nuclear systems analysis. The new methodology has been applied to the quantification of the uncertainty in the RETRAN-3D void model and then used in the analysis of an independent separate-effect experiment. This has clearly demonstrated the basic feasibility of the approach, as well as its advantages in yielding narrower uncertainty bands in quantifying the code’s accuracy for void fraction predictions.

Uncertainty Assessment for Towing Tank Tests With Example for Surface Combatant DTMB Model 5415

Uncertainty assessment methodology, procedures, and results are presented for most typical towing tank tests using a 3.048 m geosym of naval combatant DTMB model 5415, which is an established benchmark for computational fluid dynamics validation. The tests include resistance, sinkage and trim, wave profile, wave elevations, and nominal wake. The procedures are summarized and follow International Towing Tank Conference Quality Manual Procedures. The facility and measurement systems are briefly described, and detailed uncertainty assessment examples for each test are provided with descriptions of bias and precision limits and total uncertainties.

Calibrating and Validating Deterministic Traffic Models

Calibration and validation of traffic models are processes that depend on field data that are often limited but are essential for determination of inputs to the model and assessment of its reliability. Quantification and systematization of the calibration and validation process expose statistical issues inherent in the use of such data. Formalization of the calibration and validation process naturally leads to the use of Bayesian methodology for assessment of uncertainties in model predictions that arise from a multiplicity of sources, especially statistical variability in estimation and calibration of the input parameters and model discrepancy. The general problem was elucidated in an earlier paper; this paper carries out the full calibration and validation process in the context of a widely used deterministic traffic model, namely, the Highway Capacity Manual model for control delay at signalized intersection approaches. In particular, the reliability of the model was assessed through quantification of the uncertainty in the estimation of model parameters, predictions of model delay, and predictions obtained by adjusting the data used in the model. While the methods are described in a specific context, they can be used generally but are inhibited at times by computational burdens that must be overcome.

Calibrating and Validating Deterministic Traffic Models: Application to Highway Capacity Manual Control Delay at Signalized Intersections

Calibration and validation of traffic models are processes that depend on field data that are often limited but are essential for determination of inputs to the model and assessment of its reliability. Quantification and systematization of the calibration and validation process expose statistical issues inherent in the use of such data. Formalization of the calibration and validation process naturally leads to the use of Bayesian methodology for assessment of uncertainties in model predictions that arise from a multiplicity of sources, especially statistical variability in estimation and calibration of the input parameters and model discrepancy. The general problem was elucidated in an earlier paper; this paper carries out the full calibration and validation process in the context of a widely used deterministic traffic model, namely, the Highway Capacity Manual model for control delay at signalized intersection approaches. In particular, the reliability of the model was assessed through quantification of ...