The cost of uniqueness in groundwater model calibration

Calibration of a groundwater model requires that hydraulic properties be estimated throughout a model domain. This generally constitutes an underdetermined inverse problem, for which a Solution can only be found when some kind of regularization device is included in the inversion process. Inclusion of regularization in the calibration process can be implicit, for example through the use of zones of constant parameter value, or explicit, for example through solution of a constrained minimization problem in which parameters are made to respect preferred values, or preferred relationships, to the degree necessary for a unique solution to be obtained. The cost of uniqueness is this: no matter which regularization methodology is employed, the inevitable consequence of its use is a loss of detail in the calibrated field. This, ill turn, can lead to erroneous predictions made by a model that is ostensibly well calibrated. Information made available as a by-product of the regularized inversion process allows the reasons for this loss of detail to be better understood. In particular, it is easily demonstrated that the estimated value for an hydraulic property at any point within a model domain is, in fact, a weighted average of the true hydraulic property over a much larger area. This averaging process causes loss of resolution in the estimated field. Where hydraulic conductivity is the hydraulic property being estimated, high averaging weights exist in areas that are strategically disposed with respect to measurement wells, while other areas may contribute very little to the estimated hydraulic conductivity at any point within the model domain, this possibly making the detection of hydraulic conductivity anomalies in these latter areas almost impossible. A study of the post-calibration parameter field covariance matrix allows further insights into the loss of system detail incurred through the calibration process to be gained. A comparison of pre- and post-calibration parameter covariance matrices shows that the latter often possess a much smaller spectral bandwidth than the former. It is also demonstrated that, as all inevitable consequence of the fact that a calibrated model cannot replicate every detail of the true system, model-to-measurement residuals can show a high degree of spatial correlation, a fact which must be taken into account when assessing these residuals either qualitatively, or quantitatively in the exploration of model predictive uncertainty. These principles are demonstrated using a synthetic case in which spatial parameter definition is based oil pilot points, and calibration is Implemented using both zones of piecewise constancy and constrained minimization regularization. (C) 2005 Elsevier Ltd. All rights reserved.

On the absolute calibration of bench-top small-angle X-ray scattering instruments: a comparison of different standard methods

Absolute calibration relates the measured (arbitrary) intensity to the differential scattering cross section of the sample, which contains all of the quantitative information specific to the material. The importance of absolute calibration in small-angle scattering experiments has long been recognized. This work details the absolute calibration procedure of a small-angle X-ray scattering instrument from Bruker AXS. The absolute calibration presented here was achieved by using a number of different types of primary and secondary standards. The samples were: a glassy carbon specimen, which had been independently calibrated from neutron radiation; a range of pure liquids, which can be used as primary standards as their differential scattering cross section is directly related to their isothermal compressibility; and a suspension of monodisperse silica particles for which the differential scattering cross section is obtained from Porod's law. Good agreement was obtained between the different standard samples, provided that care was taken to obtain significant signal averaging and all sources of background scattering were accounted for. The specimen best suited for routine calibration was the glassy carbon sample, due to its relatively intense scattering and stability over time; however, initial calibration from a primary source is necessary. Pure liquids can be used as primary calibration standards, but the measurements take significantly longer and are, therefore, less suited for frequent use.

Point-wise confidence interval estimation by neural networks: A comparative study based on automotive engine calibration.

In developing neural network techniques for real world applications it is still very rare to see estimates of confidence placed on the neural network predictions. This is a major deficiency, especially in safety-critical systems. In this paper we explore three distinct methods of producing point-wise confidence intervals using neural networks. We compare and contrast Bayesian, Gaussian Process and Predictive error bars evaluated on real data. The problem domain is concerned with the calibration of a real automotive engine management system for both air-fuel ratio determination and on-line ignition timing. This problem requires real-time control and is a good candidate for exploring the use of confidence predictions due to its safety-critical nature.

Local shell-side coefficients in shell and tube exchanges the use of a mass transfer technique

A diffusion-controlled electrochemical mass transfer technique has been employed in making local measurements of shell-side coefficients in segmentally baffled shell and tube heat exchangers. Corresponding heat transfer data are predicted through the Chilton and Colburn heat and mass transfer analogy. Mass transfer coefficients were measured for baffle spacing lengths of individual tubes in an internal baffle compartment. Shell-side pressure measurements were also made. Baffle compartment average coefficients derived from individual tube coefficients are shown to be in good agreement with reported experimental bundle average heat transfer data for a heat exchanger model of similar geometry. Mass transfer coefficients of individual tubes compare favourably with those obtained previously by another mass transfer technique. Experimental data are reported for a variety of segmental baffle configurations over the shell-side Reynolds number range 100 to 42 000. Baffles with zero clearances were studied at three baffle cuts and two baffle spacings. Baffle geometry is shown to have a large effect on the distribution of tube coefficients within the baffle compartment. Fluid "jetting" is identified with some baffle configurations. No simple characteristic velocity is found to correlate zonal or baffle compartment average mass transfer data for the effect of both baffle cut and baffle spacing. Experiments with baffle clearances typical of commercial heat exchangers are also reported. The effect of leakage streams associated with these baffles is identified. Investigations were extended to double segmental baffles for which no data had previously been published. The similarity in the shell-side characteristics of this baffle arrangement and two parallel single segmental baffle arrangements is demonstrated. A general relationship between the shell-side mass transfer performance and pressure drop was indicated by the data for all the baffle configurations examined.

Statnote 15: the non-parametric correlation coefficients

If in a correlation test, one or both variables are small whole numbers, scores based on a limited scale, or percentages, a non-parametric correlation coefficient should be considered as an alternative to Pearson’s ‘r’. Kendall’s t and Spearman’s rs are similar tests but the former should be considered if the analysis is to be extended to include partial correlations. If the data contain many tied values, then gamma should be considered as a suitable test.

Statnote 17: using a regression line for prediction and calibration

Two types of prediction problem can be solved using a regression line viz., prediction of the ‘population’ regression line at the point ‘x’ and prediction of an ‘individual’ new member of the population ‘y1’ for which ‘x1’ has been measured. The second problem is probably the most commonly encountered and the most relevant to calibration studies. A regression line is likely to be most useful for calibration if the range of values of the X variable is large, if there is a good representation of the ‘x,y’ values across the range of X, and if several estimates of ‘y’ are made at each ‘x’. It is poor statistical practice to use a regression line for calibration or prediction beyond the limits of the data.

Transfer coefficients over inclined tube banks

The thesis presents experimental results for shell-side transfer coefficients and pressure drops across four different tube banks, using small-scale models, with yawed tubes, as found in many types of heat exchangers, boilers and nuclear reactors. The tube banks investigated have a staggered tube layout on a rotated square pitch, with a 1.25 pitch-to-diameter ratio. The angle of attack was varied between 45o and 90o. An extensive range of Reynolds number, i.e. 0.5. to 12,600, covering so-called laminar, transition and turbulent flows, was investigated. A diffusion-controlled electrochemical mass transfer technique has been employed to measure mass transfer coefficients. The heat transfer coefficients may be then readily obtained from the mass transfer values by applying the well-established Chilton-Colburn analogy. The results for the normal tube bank, which forms the base case for the study on inclined tube banks, show close agreement with previous work. The transfer coefficients and pressure drops of the inclined tube banks are compared with results from the ideal normal tube bank to examine the effect of inclination angle on heat transfer and pressure drop variations. The variation of the transfer coefficients row-by-row and the entrance and exit effects have also been investigated. An auxilary investigation has been carried out on the role of natural convection. A preliminary correlation of transfer coefficients and pressure drops against the variation in the yaw angle has been attempted. The results are discussed in the light of the few existing theoretical treatments and experimental data for these situations, and recommendations made for future work.