Error analysis of finite element method for Poisson–Nernst–Planck equations
Not Available
SEARCH · Search NASA
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.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Not Available
The following poster presents a couple of approaches for temperature control at heat exchanger’s exit. One of the approaches is based on using n-step ahead neural network while the other approach utilizes a combination of control + learning module for temperature control. Designing robust controllers for typical heat exchangers is very challenging primarily because of complexities associated with the synthesis of dynamics of industrial heat exchangers. Even if a dynamics model is synthesized, the accuracy of it would be called into question as there can be lot of unmodeled system dynamics and random noise in input parameters. In many situations, the degree of uncertainty in the model of system being controlled limits the utility of optimal control design. Controller can be manually tuned on the field but it is very difficult to determine manual adjustments that result in overall improvement. These practical challenges call for learning based approaches for temperature control that would result in optimal heat exchanger performance. We utilize the existing mathematical model for air-water system synthesized by Underwood and Crawford for testing our algorithms. Underwood and Crawford developed a model of heating coil by fitting a second-order non-linear equations to measurements of air and water temperature and flow rates obtained from the actual coil. The uncertainties in the model include air and water inlet temperatures and air-flow rate. These uncertainties are modeled as random variables that are sampled from a uniform distribution. The control variable is the water flow rate which is controlled for regulating the air-outlet temperature. Neural networks are frequently used in various literatures for learning time-series. We propose a modified version of neural networks known as n-step ahead neural network that take the state variables and output at time t as input to successfully predict output at time t+n. The variable n is a hyper-parameter which can be optimized based on performance constraints. The n-step ahead neural network was then trained on training data generated from experimental model to high training accuracies. This trained model performed appreciably in its objective to regulate outlet temperature of a heat exchanger. The second approach uses a combination of control and learning module for temperature regulation. The control module used is a proportional controller while the learning module used is steady state neural network. The steady state neural network is used to learn the controller output as a function of random variables and temperature set-point. The proportional controller is then used to compensate for transient behavior of linearized system around various operating points. The algorithm works effectively for temperature regulation as neural network captures the non-linear input-output behavior of closed loop system leaving proportional controller to compensate for steady state errors of a linearized plant. The proposed algorithm’s performances were evaluated against the performance of a PI controller. Further, the robustness of the algorithms were also studied in great detail under small perturbations around tuned parameters of the proposed algorithms.
Respiratory motion measured by respiratory inductance plethysmography often deviates from the sinusoidal pattern assumed in the traditional Lissajous figure (loop) analysis used to determine thoraco-abdominal asynchrony, or phase angle phi. We investigated six different time-domain methods of measuring phi, using simulated data with sinusoidal and triangular waveforms, phase shifts of 0-135 degrees, and 10% noise. The techniques were then used on data from 11 lightly anesthetized rhesus monkeys (Macaca mulatta; 7.6 +/- 0.8 kg; 5.7 +/- 0.5 years old), instrumented with a respiratory inductive plethysmograph, and subjected to increasing levels of inspiratory resistive loading ranging from 5-1,000 cmH(2)O. L(-1). sec(-1).The best results were obtained from cross-correlation and maximum linear correlation, with errors less than approximately 5 degrees from the actual phase angle in the simulated data. The worst performance was produced by the loop analysis, which in some cases was in error by more than 30 degrees. Compared to correlation, other analysis techniques performed at an intermediate level. Maximum linear correlation and cross-correlation produced similar results on the data collected from monkeys (SD of the difference, 4.1 degrees ) but all other techniques had a high SD of the difference compared to the correlation techniques.We conclude that phase angles are best measured using cross-correlation or maximum linear correlation, techniques that are independent of waveform shape, and robust in the presence of noise. Copyright 2002 Wiley-Liss, Inc.
Numerical analyses of error covariance matrix in linear filtering problem, computational errors in Kalman filtering, and coasting flight equations
Purpose: The aim of this work was to simulate the effect of dose distribution changes on detector array calibrations and to explore compensatory methods that are used during calibration measurements. Methods: The array calibration technique that was investigated is known as wide field (WF) calibration. Using this method, a linear array [y-axis (65 detectors) of the IC PROFILER (Sun Nuclear Corporation, Melbourne, FL)] is calibrated with three measurements ({alpha}, {theta}, and {lambda}); each measurement uses the same radiation field, which is larger than the array. For measurement configuration {theta}, the array is rotated by 180 deg. from its position in {alpha}; for {lambda}, the array is shifted by one detector from its position in {theta}. The relative detector sensitivities are then determined through ratios of detector readings at the same field locations (using {theta} and {lambda}). This method results in error propagation that is proportional to the number of detectors in the array. During the procedure, the calibration protocol operates under three postulates, which state that (a) the beam shape does not change between measurements; (b) the relative sensitivities of the detectors do not change; and (c) the scatter to the array does not change as the array is moved. The WF calibration's sensitivity to a postulate (a) violation was quantified by applying a sine shaped perturbation (of up to 0.1%) to {alpha}, {theta}, or {lambda}, and then determining the change relative to a baseline calibration. Postulate (a) violations were minimized by using a continuous beam and mechanized array movement during {theta} and {lambda}. A continuously on beam demonstrated more stable beam symmetry as compared to cycling the beam on and off between measurements. Additional side-scatter was also used to satisfy postulate (c). Results: Simulated symmetry perturbations of 0.1% to {theta} or {lambda} resulted in calibration errors of up to 2%; {alpha} was relatively immune to perturbation (<0.1% error). Wide field calibration error on a linear accelerator with similar symmetry variations was {+-}1.6%. Using a continuous beam during {theta} and {lambda} with additional side-scatter reduced the calibration error from {+-}1.6% to {+-}0.48%. Conclusions: This work increased the reproducibility of WF calibrations by limiting the effect of measurement perturbations primarily due to linear accelerator symmetry variations. The same technique would work for any array using WF calibration.
An error analysis is performed for a global positioning and navigation system which relies on an orthogonal array of interferometry baselines on board a geostationary satellite. The error analysis involves linearization of a mathematical model which is based on the relationship between the measured phase differences, the known transmitter positions, and other systematic error parameters. According to the error analysis, position accuracy is critically dependent on the baseline length and the magnitude of the random component of the measuring errors. It is shown that a satellite interferometry system with baselines of about 50 m can yield position accuracies on the order of 20 m.
Purpose: To investigate the applicability of 3-dimensional (3D) surface imaging for image guidance in deep-inspiration breath-hold radiation therapy (DIBH-RT) for patients with left-sided breast cancer. For this purpose, setup data based on captured 3D surfaces was compared with setup data based on cone beam computed tomography (CBCT). Methods and Materials: Twenty patients treated with DIBH-RT after breast-conserving surgery (BCS) were included. Before the start of treatment, each patient underwent a breath-hold CT scan for planning purposes. During treatment, dose delivery was preceded by setup verification using CBCT of the left breast. 3D surfaces were captured by a surface imaging system concurrently with the CBCT scan. Retrospectively, surface registrations were performed for CBCT to CT and for a captured 3D surface to CT. The resulting setup errors were compared with linear regression analysis. For the differences between setup errors, group mean, systematic error, random error, and 95% limits of agreement were calculated. Furthermore, receiver operating characteristic (ROC) analysis was performed. Results: Good correlation between setup errors was found: R{sup 2}=0.70, 0.90, 0.82 in left-right, craniocaudal, and anterior-posterior directions, respectively. Systematic errors were {<=}0.17 cm in all directions. Random errors were {<=}0.15 cm. The limits of agreement were -0.34-0.48, -0.42-0.39, and -0.52-0.23 cm in left-right, craniocaudal, and anterior-posterior directions, respectively. ROC analysis showed that a threshold between 0.4 and 0.8 cm corresponds to promising true positive rates (0.78-0.95) and false positive rates (0.12-0.28). Conclusions: The results support the application of 3D surface imaging for image guidance in DIBH-RT after BCS.
Purpose: The aim of this study was to quantify interfractional pancreatic position variation using fiducial markers visible on daily cone beam computed tomography (CBCT) scans. In addition, we analyzed possible migration of the markers to investigate their suitability for tumor localization. Methods and Materials: For 13 pancreatic cancer patients with implanted Visicoil markers, CBCT scans were obtained before 17 to 25 fractions (300 CBCTs in total). Image registration with the reference CT was used to determine the displacement of the 2 to 3 markers relative to bony anatomy and to each other. We analyzed the distance between marker pairs as a function of time to identify marker registration error (SD of linear fit residuals) and possible marker migration. For each patient, we determined the mean displacement of markers relative to the reference CT (systematic position error) and the spread in displacements (random position error). From this, we calculated the group systematic error, Σ, and group random error, σ. Results: Marker pair distances showed slight trends with time (range, −0.14 to 0.14 mm/day), possibly due to tissue deformation, but no shifts that would indicate marker migration. The mean SD of the fit residuals was 0.8 mm. We found large interfractional position variations, with for 116 of 300 (39%) fractions a 3-dimensional vector displacement of >10 mm. The spread in displacement varied significantly (P<.01) between patients, from a vector range of 9.1 mm to one of 24.6 mm. For the patient group, Σ was 3.8, 6.6, and 3.5 mm; and σ was 3.6, 4.7 and 2.5 mm, in left–right, superior–inferior, and anterior–posterior directions, respectively. Conclusions: We found large systematic displacements of the fiducial markers relative to bony anatomy, in addition to wide distributions of displacement. These results for interfractional position variation confirm the potential benefit of using fiducial markers rather than bony anatomy for daily online position verification for pancreatic cancer patients.
Initially a separate orthogonal expansion for the low-k components and FEM for high-k components was considered, or a multi-scale FEM method, but a more attractive hybrid FDM/FEM model for parallel computations has been devised with minimal impact to formerly developed code. This system involves using Lagrange polynomials along magnetic field lines to establish a FDM discretization of the toroidal contributions to the perpendicular Laplacian, which is the operator of importance in gyrokinetic solution of the vector potential. The method was found analytically to be stable, due to some dissipation from varying field-line lengths. The accuracy of this method has been evaluated using the method of manufactured solutions (with Bei Wang of Princeton University), whereby we find that for a Poisson problem the maximum error roughly diminishes linearly as a function of the grid resolution (uniformly refined), whereas the L2 norm as square root of the resolution improvement. Most of the error appears to be contributed near boundaries and where the FEM grid has anisotropy, so we are trying to improve this analysis using improved tessellations obtained from Mark Shepard, and publish the results as soon as the problems with the grid have been eliminated.
A new approach for determination of machine-tool settings for spiral bevel gears is proposed. The proposed settings provide a predesigned parabolic function of transmission errors and the desired location and orientation of the bearing contact. The predesigned parabolic function of transmission errors is able to absorb piece-wise linear functions of transmission errors that are caused by the gear misalignment and reduce gear noise. The gears are face-milled by head cutters with conical surfaces or surfaces of revolution. A computer program for simulation of meshing, bearing contact and determination of transmission errors for misaligned gear has been developed.
This study will demonstrate the importance of application of computer numerically controlled (CNC) machines in generation of gear tooth surfaces with new topology. This topology decreases gear vibration and will extend the gear capacity and service life. A preliminary investigation by a tooth contact analysis (TCA) program has shown that gear tooth surfaces in line contact (for instance, involute helical gears with parallel axes, worm gear drives with cylindrical worms, etc.) are very sensitive to angular errors of misalignment that cause edge contact and an unfavorable shape of transmission errors and vibration. The new topology of gear tooth surfaces is based on the localization of bearing contact, and the synthesis of a predesigned parabolic function of transmission errors that is able to absorb a piecewise linear function of transmission errors caused by gear misalignment. The report will describe the following topics: description of kinematics of CNC machines with six degrees of freedom that can be applied for generation of gear tooth surfaces with new topology. A new method for grinding of gear tooth surfaces by a cone surface or surface of revolution based on application of CNC machines is described. This method provides an optimal approximation of the ground surface to the given one. This method is especially beneficial when undeveloped ruled surfaces are to be ground. Execution of motions of the CNC machine is also described. The solution to this problem can be applied as well for the transfer of machine tool settings from a conventional generator to the CNC machine. The developed theory required the derivation of a modified equation of meshing based on application of the concept of space curves, space curves represented on surfaces, geodesic curvature, surface torsion, etc. Condensed information on these topics of differential geometry is provided as well.
An approach for design and generation of low-noise helical gears with localized bearing contact is proposed. The approach is applied to double circular arc helical gears and modified involute helical gears. The reduction of noise and vibration is achieved by application of a predesigned parabolic function of transmission errors that is able to absorb a discontinuous linear function of transmission errors caused by misalignment. The localization of the bearing contact is achieved by the mismatch of pinion-gear tooth surfaces. Computerized simulation of meshing and contact of the designed gears demonstrated that the proposed approach will produce a pair of gears that has a parabolic transmission error function even when misalignment is present. Numerical examples for illustration of the developed approach are given.
The design and generation of modified involute helical gears that have a localized and stable bearing contact, and reduced noise and vibration characteristics are described. The localization of the bearing contact is achieved by the mismatch of the two generating surfaces that are used for generation of the pinion and the gear. The reduction of noise and vibration will be achieved by application of a parabolic function of transmission errors that is able to absorb the almost linear function of transmission errors caused by gear misalignment. The meshing and contact of misaligned gear drives can be analyzed by application of computer programs that have been developed. The computations confirmed the effectiveness of the proposed modification of the gear geometry. A numerical example that illustrates the developed theory is provided.
A thrust stand calibration technique for use in testing repetitively-pulsed electric thrusters for in-space propulsion has been developed and tested using a modified hanging pendulum thrust stand. In the implementation of this technique, current pulses are applied to a solenoidal coil to produce a pulsed magnetic field that acts against the magnetic field produced by a permanent magnet mounted to the thrust stand pendulum arm. The force on the magnet is applied in this non-contact manner, with the entire pulsed force transferred to the pendulum arm through a piezoelectric force transducer to provide a time-accurate force measurement. Modeling of the pendulum arm dynamics reveals that after an initial transient in thrust stand motion the quasisteady average deflection of the thrust stand arm away from the unforced or zero position can be related to the average applied force through a simple linear Hooke s law relationship. Modeling demonstrates that this technique is universally applicable except when the pulsing period is increased to the point where it approaches the period of natural thrust stand motion. Calibration data were obtained using a modified hanging pendulum thrust stand previously used for steady-state thrust measurements. Data were obtained for varying impulse bit at constant pulse frequency and for varying pulse frequency. The two data sets exhibit excellent quantitative agreement with each other as the constant relating average deflection and average thrust match within the errors on the linear regression curve fit of the data. Quantitatively, the error on the calibration coefficient is roughly 1% of the coefficient value.
An electronic system adjusts the frequency of a tunable laser, eventually locking the frequency to a peak in the optical absorption spectrum of a gas (or of a Fabry-Perot cavity that has an absorption peak like that of a gas). This system was developed to enable precise locking of the frequency of a laser used in differential absorption LIDAR measurements of trace atmospheric gases. This system also has great commercial potential as a prototype of means for precise control of frequencies of lasers in future dense wavelength-division-multiplexing optical communications systems. The operation of this system is completely automatic: Unlike in the operation of some prior laser-frequency-locking systems, there is ordinarily no need for a human operator to adjust the frequency manually to an initial value close enough to the peak to enable automatic locking to take over. Instead, this system also automatically performs the initial adjustment. The system (see Figure 1) is based on a concept of (1) initially modulating the laser frequency to sweep it through a spectral range that includes the desired absorption peak, (2) determining the derivative of the absorption peak with respect to the laser frequency for use as an error signal, (3) identifying the desired frequency [at the very top (which is also the middle) of the peak] as the frequency where the derivative goes to zero, and (4) thereafter keeping the frequency within a locking range and adjusting the frequency as needed to keep the derivative (the error signal) as close as possible to zero. More specifically, the system utilizes the fact that in addition to a zero crossing at the top of the absorption peak, the error signal also closely approximates a straight line in the vicinity of the zero crossing (see Figure 2). This vicinity is the locking range because the linearity of the error signal in this range makes it useful as a source of feedback for a proportional + integral + derivative control scheme that constantly adjusts the frequency in an effort to drive the error to zero. When the laser frequency deviates from the midpeak value but remains within the locking range, the magnitude and sign of the error signal indicate the amount of detuning and the control circuitry adjusts the frequency by what it estimates to be the negative of this amount in an effort to bring the error to zero.
It has been previously shown that the temporal integration of hyperbolic partial differential equations (PDE's) may, because of boundary conditions, lead to deterioration of accuracy of the solution. A procedure for removal of this error in the linear case has been established previously. In the present paper we consider hyperbolic (PDE's) (linear and non-linear) whose boundary treatment is done via the SAT-procedure. A methodology is present for recovery of the full order of accuracy, and has been applied to the case of a 4th order explicit finite difference scheme.
Real sun-tracking solar concentrators imperfectly follow the solar disk, operationally sustaining both transverse and axial misalignments. This paper describes an analysis of the solar concentration performance of a line-focusing flat-base Fresnel lens in the presence of small transverse tracking errors. Simple optics and ray-tracing techniques are used to evaluate the lens solar transmittance and focal-plane imaging characteristics. Computer-generated example data for an f/1.0 lens indicate that less than a 1% transmittance degradation occurs for transverse errors up to 2.5 deg. In this range, solar-image profiles shift laterally in the focal plane, the peak concentration ratio drops, and profile asymmetry increases with tracking error. With profile shift as the primary factor, the ninety-percent target-intercept width increases rapidly for small misalignments, e.g., almost threefold for a 1-deg error. The analytical model and computational results provide a design base for tracking and absorber systems for the linear-Fresnel-lens solar concentrator.
Probability distribution errors effect in linear and planar impulse orbital transfers, considering angular orientation and impulsive velocity