Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Monte Carlo Methods”

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.

At least 1,369 records · Page 76
All sourcesNASA NTRSDOE OSTI
DOE OSTI2014

Dose response of selected solid state detectors in applied homogeneous transverse and longitudinal magnetic fields

Purpose: MR-Linac devices under development worldwide will require standard calibration, commissioning, and quality assurance. Solid state radiation detectors are often used for dose profiles and percent depth dose measurements. The dose response of selected solid state detectors is therefore evaluated in varying transverse and longitudinal magnetic fields for this purpose. Methods: The Monte Carlo code PENELOPE was used to model irradiation of a PTW 60003 diamond detector and IBA PFD diode detector in the presence of a magnetic field. The field itself was varied in strength, and oriented both transversely and longitudinally with respect to the incident photon beam. The long axis of the detectors was oriented either parallel or perpendicular to the photon beam. The dose to the active volume of each detector in air was scored, and its ratio to dose with zero magnetic field strength was determined as the “dose response” in magnetic field. Measurements at low fields for both detectors in transverse magnetic fields were taken to evaluate the accuracy of the simulations. Additional simulations were performed in a water phantom to obtain few representative points for beam profile and percent depth dose measurements. Results: Simulations show significant dose response as a function of magnetic field in transverse field geometries. This response can be near 20% at 1.5 T, and it is highly dependent on the detectors’ relative orientation to the magnetic field, the energy of the photon beam, and detector composition. Measurements at low transverse magnetic fields verify the simulations for both detectors in their relative orientations to radiation beam. Longitudinal magnetic fields, in contrast, show little dose response, rising slowly with magnetic field, and reaching 0.5%–1% at 1.5 T regardless of detector orientation. Water tank and in air simulation results were the same within simulation uncertainty where lateral electronic equilibrium is present and expectedly differed at the beam edge in transverse field orientations only. Due to the difference in design, the two detectors behaved differently. Conclusions: When transverse magnetic fields are present, great care must be taken when using diamond or diode detectors. Dose response varies with relative detector orientation, magnetic field strength, and between detectors. This response can be considerable (∼20% for both detectors). Both detectors in longitudinal fields exhibit little to no dose response as a function of magnetic field. Water tank simulations seem to suggest that the diode detector is better suited to general beam commissioning, and each detector must be investigated separately.

07 ISOTOPE AND RADIATION SOURCES↗
DOE OSTI2015

Assessment of radiation dose in nuclear cardiovascular imaging using realistic computational models

Purpose: Nuclear cardiology plays an important role in clinical assessment and has enormous impact on the management of a variety of cardiovascular diseases. Pediatric patients at different age groups are exposed to a spectrum of radiation dose levels and associated cancer risks different from those of adults in diagnostic nuclear medicine procedures. Therefore, comprehensive radiation dosimetry evaluations for commonly used myocardial perfusion imaging (MPI) and viability radiotracers in target population (children and adults) at different age groups are highly desired. Methods: Using Monte Carlo calculations and biological effects of ionizing radiation VII model, we calculate the S-values for a number of radionuclides (Tl-201, Tc-99m, I-123, C-11, N-13, O-15, F-18, and Rb-82) and estimate the absorbed dose and effective dose for 12 MPI radiotracers in computational models including the newborn, 1-, 5-, 10-, 15-yr-old, and adult male and female computational phantoms. Results: For most organs, {sup 201}Tl produces the highest absorbed dose whereas {sup 82}Rb and {sup 15}O-water produce the lowest absorbed dose. For the newborn baby and adult patient, the effective dose of {sup 82}Rb is 48% and 77% lower than that of {sup 99m}Tc-tetrofosmin (rest), respectively. Conclusions: {sup 82}Rb results in lower effective dose in adults compared to {sup 99m}Tc-labeled tracers. However, this advantage is less apparent in children. The produced dosimetric databases for various radiotracers used in cardiovascular imaging, using new generation of computational models, can be used for risk-benefit assessment of a spectrum of patient population in clinical nuclear cardiology practice.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-J-08: Comparison of Unintended Radiation Doses to Organs at Risk Resulting From the Out-Of-Field Therapeutic Beams and From Image-Guidance X-Ray Procedures

Purpose: The unintended radiation dose to organs at risk (OAR) can be contributed from imaging guidance procedures as well as from leakage and scatter of therapeutic beams. This study compares the imaging dose with the unintended out-of-field therapeutic dose to patient sensitive organs. Methods: The Monte Carlo EGSnrc user codes, BEAMnrc and DOSXYZnrc, were used to simulate kV X-ray sources from imaging devices as well as the therapeutic IMRT/VMAT beams and to calculate doses to target and OARs on patient treatment planning CT images. The accuracy of the Monte Carlo simulations was benchmarked against measurements in phantoms. The dose-volume histogram was utilized in analyzing the patient organ doses. Results: The dose resulting from Standard Head kV-CBCT scans to bone and soft tissues ranges from 0.7 to 1.1 cGy and from 0.03 to 0.3 cGy, respectively. The dose resulting from Thorax scans on the chest to bone and soft tissues ranges from 1.1 to 1.8 cGy and from 0.3 to 0.6 cGy, respectively. The dose resulting from Pelvis scans on the abdomen to bone and soft tissues range from 3.2 to 4.2 cGy and from 1.2 to 2.2 cGy, respectively. The out-of-field doses to OAR are sensitive to the distance between the treated target and the OAR. For a typical Head-and-Neck IMRT/VMAT treatment the out-of-field doses to eyes are 1–3% of the target dose, or 2–6 cGy per fraction. Conclusion: The imaging doses to OAR are predictable based on the imaging protocols used when OARs are within the imaged volume and can be estimated and accounted for by using tabulated values. The unintended out-of-field doses are proportional to the target dose, strongly depend on the distance between the treated target and OAR, and are generally higher comparing to the imaging dose. This work was partially supported by Varian research grant VUMC40590.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-J-23: Characteristics of X-Rays From ExacTrac and Patient Dose From Imaging Procedures

Purpose: The purpose of this investigation is 1) provide the beam characteristics of x-rays produced by a Novalis TX ExacTrac system; 2) present a method to commission such beams, 3) present radiation dose to patients resulting from the imaging procedures. Methods: The Monte Carlo simulations were used to obtain the characteristics of kV beams and validated by measurements. The calculated beam HVLs, profiles and depth-dose curves were benchmarked against measurements. Twelve different image acquisition protocols were studied. The x-ray tube voltage ranges from 70 - 145 kV and milliampere-second (mAs) ranges from 8 - 80 mAs depending on the selection of Cranium, Head & Neck, Thorax or Abdomen imaging protocols. The beam output of each image acquisition protocol was determined by using an ionization chamber. The air kerma calibration factors of the ion chamber were obtained from an Accredited Dosimetry Calibration Laboratory for specified HVLs. Results: The agreements between measured and simulated results were within the uncertainties for HVLs, dose profiles and depth-dose curves. When %dd was normalized at 1 cm depth, its values at depth of 5 cm ranged from 45% to 66% of in water for kV beams range from 70 kVp to 145 kVp respectively. For head images, a typical dose to eye resulting from single projection ranged from 0.008 cGy to 0.025 cGy depending on the cranial image protocols selected. For a single pelvic image the skin dose could reach up to 0.1 cGy from an abdominal protocol. Conclusion: Although multiple pairs of x-ray images are commonly acquired during a daily patient treatment, the imaging doses to patients resulting from the sum of these projected x-rays are generally much less than 0.5 cGy. The knowledge obtained from this investigation can be used to estimate the image dose and optimize the used of the system.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-15: A Comparison of COMS and EP917 Eye Plaque Applicators Using Different Radionuclides

Purpose: To investigate the effect of plaque design and radionuclides on eye plaque dosimetry. Methods: The Monte Carlo N-particle Code version 6 (MCNP6) was used for radiation transport simulations. The 14 mm and 16 mm diameter COMS plaques and the model EP917 plaque were simulated using brachytherapy seeds containing I-125, Pd-103, and Cs-131 radionuclides. The origin was placed at the scleral inner surface. The central axis (CAX) doses of both COMS plaques at −1 mm, 0 mm, 1 mm, 2 mm, 5 mm, 10 mm, 15 mm, 20 mm, and 22.6 mm were compared to the model EP917 plaque. Dose volume histograms (DVHs) were also created for both COMS plaques for the tumor and outer sclera then compared to results for the model EP917 plaque. Results: For all radionuclides, the EP917 plaque delivered higher dose (max 343%) compared to the COMS plaques, except for the 14 mm COMS plaque with Cs-131 at 1 mm and 2 mm depths from outer sclera surface. This could be due to source design. For all radionuclides, the 14 mm COMS plaque delivered higher doses compared to the 16 mm COMS plaque for the depths up to 5 mm. Dose differences were not significant beyond depths of 10 mm due to ocular lateral scatter for the different plaque designs. Tumor DVHs for the 16 mm COMS plaque with Cs-131 provided better dose homogeneity and conformity compared to other COMS plaques with I-125 and Pd-103. Using Pd-103, DVHs for the 16 mm COMS plaque delivered less dose to outer sclera compared to other plaques. Conclusion: This study identified improved tumor homogeneity upon considering radionuclides and plaque designs, and found that scleral dose with the model EP917 plaque was higher than for the 16 mm COMS plaque for all the radionuclides studied.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-146: Beam Energy Spread Estimate Based On Bragg Peak Measurement

Purpose: ProNova is installing and commissioning a two room proton therapy system in Knoxville, TN. Beam energy out of the 230MeV cyclotron was measured on Jan 24, 2015. Cyclotron beam was delivered into a Zebra multi layered IC detector calibrated in terms of penetration range in water. The analysis of the measured Bragg peak determines penetration range in water which can be subsequently converted into proton beam energy. We extended this analysis to obtain an estimate of the beam energy spread out of the cyclotron. Methods: Using Monte Carlo simulations we established the correlation between Bragg peak shape parameters (width at 50% and 80% dose levels, distal falloff) and penetration range for a monoenergetic proton beam. For large uniform field impinging on a small area detector, we observed linear dependence of each Bragg peak parameter on beam penetration range as shown in Figure A. Then we studied how this correlation changes when the shape of Bragg peak is distorted by the beam focusing conditions. As shown in Figure B, small field size or diverging beam cause Bragg peak deformation predominantly in the proximal region. The distal shape of the renormalized Bragg peaks stays nearly constant. This excludes usage of Bragg peak width parameters for energy spread estimates. Results: The measured Bragg peaks had an average distal falloff of 4.86mm, which corresponds to an effective range of 35.5cm for a monoenergetic beam. The 32.7cm measured penetration range is 2.8cm less. Passage of a 230MeV proton beam through a 2.8cm thick slab of water results in a ±0.56MeV energy spread. As a final check, we confirmed agreement between shapes of the measured Bragg peak and one generated by Monte-Carlo code for proton beam with 0.56 MeV energy spread. Conclusion: Proton beam energy spread can be estimated using Bragg peak analysis.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

MO-FG-BRA-06: Material Optimisation for Nanoparticle-Sensitized Radiotherapy

Purpose: To evaluate the potential applications of nanoparticles of different elemental compositions as X-ray radiosensitizers. Methods: Geant4 Monte Carlo models were used to model the interactions of 10 nm radius nanoparticles of elements ranging from silicon to mercury with either monoenergetic keV X-rays or a 6 MV Linac spectrum. Secondary electron spectra emitted from nanoparticles and resulting radial doses were calculated. Potential biological impacts were evaluated through the use of the Local Effect Model. Results: For keV radiation, fluoresence photons and energetic photoelectrons deposit the majority of energy emitted by nanoparticles, over long ranges. However, short-range energy deposition is dominated by Auger electrons. While total energy deposition increases with atomic number, short-range energy deposition has a more complex behaviour, with local maxima around germanium (Z=32) and gadolinium (Z=64). For megavoltage radiation, Compton electrons deposit the majority of energy, but short- range depositions are dominated by interactions with secondary electrons. As a Result, energy deposition following ionising events following Linac irradiation is largely independent of atomic number.When analysed through the Local Effect Model, these effects translate into significant variations in predicted cell killing with atomic number for keV exposures, but comparatively limited effects for megavoltage exposures. Conclusions: Although heavier elements typically lead to greater increases in dose within target volumes, on the nanometre scale effects are typically dominated by low-energy secondary electrons, whose distribution has a complex relationship with both nanoparticle composition and beam energy.These results suggest that maximum radiosensitization may not necessarily be achieved through maximizing atomic number, but rather a balance between overall interaction cross-section and local dose distribution. The different behaviours seen when comparing keV and MV behaviours may also provide future mechanisms for separating physical and biological mechanisms of nanoparticle sensitisation, a key step in their development as a potential therapeutic agent.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

TU-F-CAMPUS-J-02: Developing a Phenomenological Model of the Proton Trajectory Within a Heterogeneous Medium Required for Proton Imaging

Purpose: To develop an accurate phenomenological model of the cubic spline trajectory (CST) estimate of the proton path, accounting for the initial proton energy and water equivalent thickness (WET) traversed. Methods: Geant4 Monte Carlo (MC) simulations were used to calculate the path of protons crossing various slabs (5–30 cm WET) of different material (LN300, water and CB2–50% CaCO3) for a range of initial energies (150–330MeV). For each MC trajectory, CST was constructed based on the proton entrance and exit information and compared with the MC using the root mean square (RMS) metric. The CST path is dependent on the direction vector magnitudes (|P0,1|). First, |P0,1| is set to the proton path length. Then, a factor Λ is introduced to modify |P0,1|. The factor is varied to minimize the RMS with MC paths for every configuration. Finally, a set of Λopt factors that minimizes the RMS is presented. These are dependent on the ratio between WET and water equivalent path length (WEPL). The resolution along the path is investigated with a set of slabs. MTF analysis is performed on proton radiographs of a line-pair phantom reconstructed using the CST trajectories (Λopt and Λ1). Results: Λopt was fitted to the ratio of WET/WEPL using a power function (Y=1-AXB where A=0.36, B=4.07). The RMS deviation calculated along the path, between the CST and the MC path, increases with the WET. The increase is larger when using Λ1 than Λopt (difference of 5.0% with WET/WEPL=0.86). For 230(330) MeV protons, the MTF10% was found to increase by 40%(6%) respectively for a thick phantom (30cm) and by 25%(1%) for thinner phantom (25cm) when using the Λopt model compared to the Λ1 model. Conclusion: Based on these results, using CST with the Λopt factor reduces the RMS deviation and increases the spatial resolution when reconstructing proton trajectories.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2016

Therapeutic analysis of high-dose-rate {sup 192}Ir vaginal cuff brachytherapy for endometrial cancer using a cylindrical target volume model and varied cancer cell distributions

Purpose: To evaluate high-dose-rate (HDR) vaginal cuff brachytherapy (VCBT) in the treatment of endometrial cancer in a cylindrical target volume with either a varied or a constant cancer cell distributions using the linear quadratic (LQ) model. Methods: A Monte Carlo (MC) technique was used to calculate the 3D dose distribution of HDR VCBT over a variety of cylinder diameters and treatment lengths. A treatment planning system (TPS) was used to make plans for the various cylinder diameters, treatment lengths, and prescriptions using the clinical protocol. The dwell times obtained from the TPS were fed into MC. The LQ model was used to evaluate the therapeutic outcome of two brachytherapy regimens prescribed either at 0.5 cm depth (5.5 Gy × 4 fractions) or at the vaginal mucosal surface (8.8 Gy × 4 fractions) for the treatment of endometrial cancer. An experimentally determined endometrial cancer cell distribution, which showed a varied and resembled a half-Gaussian distribution, was used in radiobiology modeling. The equivalent uniform dose (EUD) to cancer cells was calculated for each treatment scenario. The therapeutic ratio (TR) was defined by comparing VCBT with a uniform dose radiotherapy plan in term of normal cell survival at the same level of cancer cell killing. Calculations of clinical impact were run twice assuming two different types of cancer cell density distributions in the cylindrical target volume: (1) a half-Gaussian or (2) a uniform distribution. Results: EUDs were weakly dependent on cylinder size, treatment length, and the prescription depth, but strongly dependent on the cancer cell distribution. TRs were strongly dependent on the cylinder size, treatment length, types of the cancer cell distributions, and the sensitivity of normal tissue. With a half-Gaussian distribution of cancer cells which populated at the vaginal mucosa the most, the EUDs were between 6.9 Gy × 4 and 7.8 Gy × 4, the TRs were in the range from (5.0){sup 4} to (13.4){sup 4} for the radiosensitive normal tissue depending on the cylinder size, treatment lengths, prescription depth, and dose as well. However, for a uniform cancer cell distribution, the EUDs were between 6.3 Gy × 4 and 7.1 Gy × 4, and the TRs were found to be between (1.4){sup 4} and (1.7){sup 4}. For the uniformly interspersed cancer and radio-resistant normal cells, the TRs were less than 1. The two VCBT prescription regimens were found to be equivalent in terms of EUDs and TRs. Conclusions: HDR VCBT strongly favors cylindrical target volume with the cancer cell distribution following its dosimetric trend. Assuming a half-Gaussian distribution of cancer cells, the HDR VCBT provides a considerable radiobiological advantage over the external beam radiotherapy (EBRT) in terms of sparing more normal tissues while maintaining the same level of cancer cell killing. But for the uniform cancer cell distribution and radio-resistant normal tissue, the radiobiology outcome of the HDR VCBT does not show an advantage over the EBRT. This study strongly suggests that radiation therapy design should consider the cancer cell distribution inside the target volume in addition to the shape of target.

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2016

Radiation dose and cancer risk to out-of-field and partially in-field organs from radiotherapy for symptomatic vertebral hemangiomas

Purpose: Vertebral hemangiomas (VHs) are the most common benign tumors of the spine that may cause bone resorption. Megavoltage irradiation is usually the treatment of choice for the management of symptomatic VHs. The current study was conducted to estimate the risk for carcinogenesis from radiotherapy of this benign disease on the basis of the calculated radiation doses to healthy organs. Methods: The Monte Carlo N-particle transport code was employed to simulate the irradiation with 6 MV x-rays of a VH presented in the cervical, upper thoracic, lower thoracic, and lumbar spine. The average radiation dose (D{sub av}) received by each critical organ located outside the primarily irradiated area was calculated. Three-dimensional treatment plans were also generated for the VHs occurring at the four different sites of the spinal cord based on patients’ computed tomography data. The organ equivalent dose (OED) to each radiosensitive structure, which was partly encompassed by the applied treatment fields, was calculated with the aid of differential dose–volume histograms. The D{sub av} and the OED values were combined with a linear-no-threshold model and a nonlinear mechanistic model, respectively, to estimate the organ-, age-, and gender-specific lifetime attributable risks (LARs) for cancer development. The estimated risks were compared with the respective nominal lifetime intrinsic risks (LIRs) for the unexposed population. Results: For a standard target dose of 34 Gy, the OED varied from 0.39–5.15 Gy by the organ of interest and the irradiation site. The D{sub av} range for the out-of-field organs was 4.9 × 10{sup −4} to 0.56 Gy. The LAR for the appearance of malignancies in the partially in-field organs after radiotherapy of male and female patients was (0.08%–1.8%) and (0.09%–1.9%), respectively. These risk values were 1.5–15.5 times lower when compared to the respective LIRs. The lifetime probability for out-of-field cancer induction in irradiated males and females was (2.5 × 10{sup −4} to 7.7 × 10{sup −2})% and (1.4 × 10{sup −4} to 2.6 × 10{sup −1})%, respectively. The above risks were one to four orders of magnitude lower than the LIRs. Conclusions: The probability for the development of out-of-field malignancies due to radiotherapy for VHs is trivial with respect to the nominal risk for unexposed population. The respective cancer risks to partially in-field organs are smaller than the nominal probabilities but they should not be considered as inconsiderable. These risks may be taken into account during the follow-up of patients treated for a symptomatic VH.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2016

Technical Note: Dosimetry of Leipzig and Valencia applicators without the plastic cap

Purpose: High dose rate (HDR) brachytherapy for treatment of small skin lesions using the Leipzig and Valencia applicators is a widely used technique. These applicators are equipped with an attachable plastic cap to be placed during fraction delivery to ensure electronic equilibrium and to prevent secondary electrons from reaching the skin surface. The purpose of this study is to report on the dosimetric impact of the cap being absent during HDR fraction delivery, which has not been explored previously in the literature. Methods: GEANT4 Monte Carlo simulations (version 10.0) have been performed for the Leipzig and Valencia applicators with and without the plastic cap. In order to validate the Monte Carlo simulations, experimental measurements using radiochromic films have been done. Results: Dose absorbed within 1 mm of the skin surface increases by a factor of 1500% for the Leipzig applicators and of 180% for the Valencia applicators. Deeper than 1 mm, the overdosage flattens up to a 10% increase. Conclusions: Differences of treating with or without the plastic cap are significant. Users must check always that the plastic cap is in place before any treatment in order to avoid overdosage of the skin. Prior to skin HDR fraction delivery, the timeout checklist should include verification of the cap placement.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2016

SU-F-T-126: Microdosimetic Evaluation of Proton Energy Distributions in the Vicinity of Metal Implants

Purpose: To evaluate the impact of titanium and surgical stainless steel implants on the microscopic dose distribution in proton treatment plans Methods: Geant4 Monte Carlo simulations were used to analyze the microdosimetric distribution of proton radiation in the vicinity of 3.1 mm thick CP Grade 4 titanium (Ti) or 316 stainless steel (SS316) plates in a water phantom. Additional simulations were performed using either water, or water with a density equivalent to the respective metals (Tiwater, SS316water) (to reflect common practice in treatment planning). Implants were placed at the COM of SOBPs of 157 MeV (range of ∼15 cm in water) protons with 30 or 60 mm modulation. Primary and secondary particle dose and fluence, frequency-weighted and dose-weighted average lineal energy, average radiation quality factor, dose equivalent and energy deposition histograms in the plate vicinity were compared. Results: Preliminary results show frequency-weighted (yf) and dose-weighted lineal energy (yd) was increased downstream of the Ti plate (yf = 3.1 keV/µm; yd = 5.5 keV/µm) and Tiwater (yf = 4.1 keV/µm; yd = 6.8 keV/µm) compared to that of water (ie, the absence of a plate) (yf = 2.5 keV/µm; yd = 4.5 keV/µm). In addition, downstream proton dose deposition was also elevated due to the presence of the Ti plate or Tiwater. The additional dose deposited at higher lineal energy implies that tissues downstream of the plate will receive a higher dose equivalent. Detailed analyses of the Ti, Tiwater, SS316, and SS316 water simulations will be presented. Conclusion: The presence of high-density materials introduces changes in the spatial distribution of radiation in the vicinity of an implant. Further work quantifying these effects could be incorporated into future treatment planning systems resulting in more accurate treatment plans. This project was sponsored with funding from the Department of Defense (DOD # W81XWH-10-2-0192).

07 ISOTOPE AND RADIATION SOURCES↗
DOE OSTI2016

SU-F-T-239: Evaluation of Plastic Scintillator Light Output for Various Lead Doping Concentrations: Towards LET Detection

Purpose: This work seeks to develop a beam quality meter using multiple differently doped plastic scintillators that are thus intrinsically beam-quality dependent. Plastic scintillators spontaneously emit visible light upon irradiation; the amount of light produced is dependent on stopping power (closely related to LET) according to Birks’ law. Doping plastic scintillators can be used to tune their sensitivity to specific LET ranges. Methods: GEANT4.10.1 Monte Carlo (MC) was used to evaluate the response of various scintillator dopant combinations. MC radiation transport and scintillator light response were validated against previously published literature. Current work involves evaluating detector response experimentally; to that end, a detector prototype with interchangeable scintillator housing was constructed. Measurement set-up guides light emitted by the scintillator to a photomultiplier tube via a glass taper junction coupled to an optical fiber. The resulting signal is measured by an electrometer, and normalized to dose readout from a diode. Measurements have been done using clinical electron and orthovoltage beams. MC response (simulated scintillator light normalized to dose scored inside the scintillating volume) was evaluated for four different LET radiations for an undoped and 1%Pb doped scintillator (σ=0.85%). Simulated incident electrons included: 0.05, 0.1, 0.2, 6, 12, and 18 MeV; these energies correspond to a range of stopping power (related to LET) values ranging from 1.824 to 11.09 MeVcm{sup 2}g{sup −1} (SCOL from NIST-ESTAR). Results: Initial MC results show a distinct divergence in scintillator response as LET increases. The response for undoped plastic scintillator indicated a 35.0% increase in signal when going from 18 MeV (low LET) to 0.05 MeV (high LET) while 1%-Pb doped scintillator indicated a 100.9% increase. Conclusion: After validating MC against measurement, simulations will be used to test various concentrations (2%, 4%, 6%) of different high-Z material dopants (W, Mo) to optimize the scintillator types for the beam quality meter. NSERC Discovery Grant RGPIN-435608.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2018

Toward a predictive theory of correlated materials

Correlated electron materials display a rich variety of notable properties ranging from unconventional superconductivity to metal-insulator transitions. These properties are of interest from the point of view of applications but are hard to treat theoretically, as they result from multiple competing energy scales. Although possible in more weakly correlated materials, theoretical design and spectroscopy of strongly correlated electron materials have been a difficult challenge for many years. In this paper, by treating all the relevant energy scales with sufficient accuracy, complementary advances in Green’s functions and quantum Monte Carlo methods open a path to first-principles computational property predictions in this class of materials.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Ensemble approximate control variate estimators: Applications to multi-fidelity importance sampling.

The recent growth in multifidelity uncertainty quantification has given rise to a large set of variance reduction techniques that leverage information from model ensembles to provide variance reduction for estimates of the statistics of a high-fidelity model. In this paper we provide two contributions: (1) we utilize an ensemble estimator to account for uncertainties in the optimal weights of approximate control variate (ACV) approaches and derive lower bounds on the number of samples required to guarantee variance reduction; and (2) we extend an existing multifidelity importance sampling (MFIS) scheme to leverage control variates. Our approach directly addresses a limitation of many multifidelity sampling strategies that require the usage of pilot samples to estimate covariances. As such we make significant progress towards both increasing the practicality of approximate control variates—for instance, by accounting for the effect of pilot samples—and using multifidelity approaches more effectively for estimating low-probability events. The numerical results indicate our hybrid MFIS-ACV estimator achieves up to 50% improvement in variance reduction over the existing state-of-the-art MFIS estimator, which had already shown an outstanding convergence rate compared to the Monte Carlo method, on several problems of computational mechanics.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Bias-Variance Trade-Off in Physics-Informed Neural Networks with Randomized Smoothing for High-Dimensional PDEs

Physics-Informed Neural Networks (PINNs) have triggered a paradigm shift in scientific computing, leveraging mesh-free properties and robust approximation capabilities. While proving effective for low-dimensional partial differential equations (PDEs), the computational cost of PINNs remains a hurdle in high-dimensional scenarios. This is particularly pronounced when computing high-order and high-dimensional derivatives in the physics-informed loss. Randomized Smoothing PINN (RS-PINN) introduces Gaussian noise for stochastic smoothing of the original neural net model, enabling the use of Monte Carlo methods for derivative approximation, which eliminates the need for costly automatic differentiation. Despite its computational efficiency, especially in the approximation of high-dimensional derivatives, RS-PINN introduces biases in both loss and gradients, negatively impacting convergence, especially when coupled with stochastic gradient descent (SGD) algorithms. We present a comprehensive analysis of biases in RS-PINN, attributing them to the nonlinearity of the Mean Squared Error (MSE) loss as well as the intrinsic nonlinearity of the PDE itself. We propose tailored bias correction techniques, delineating their application based on the order of PDE nonlinearity. The derivation of an unbiased RS-PINN allows for a detailed examination of its advantages and disadvantages compared to the biased version. Specifically, the biased version has a lower variance and runs faster than the unbiased version, but it is less accurate due to the bias. To optimize the bias-variance trade-off, we combine the two approaches in a hybrid method that balances the rapid convergence of the biased version with the high accuracy of the unbiased version. In addition to methodological contributions, we present an enhanced implementation of RS-PINN. Extensive experiments on diverse high-dimensional PDEs, including Fokker-Planck, Hamilton-Jacobi-Bellman (HJB), viscous Burgers’, Allen-Cahn, and Sine-Gordon equations, illustrate the bias-variance trade-off and highlight the effectiveness of the hybrid RS-PINN. Empirical guidelines are provided for selecting biased, unbiased, or hybrid versions, depending on the dimensionality and nonlinearity of the specific PDE problem.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Supersymmetric Wilson loops on the lattice in the large N limit

We propose additional tests of holography by studying supersymmetric Wilson loops in p+1-dimensional maximally supersymmetric Yang?Mills (SYM) theories on the lattice in the large N limit. In the dual gravity description, this computation involves calculating the area of a fundamental string worldsheet in certain Type II supergravity backgrounds. Though thermodynamic observables have been computed on the lattice using Monte Carlo methods and agree with the supergravity results in various dimensions, not much has been done for the gauge-invariant operators such as the Wilson loop. We provide analytical predictions for these loops for various non-conformal Dp-brane background cases with ? ? 2 p?2 in the large N limit and comment on how these can be computed on non-orthogonal lattices in various models.

Jha, Raghav G. (ORCID:0000000329330102)↗
DOE OSTI2018

Random center vortex model

This work presents a model of center vortices, expressed by random lines in three-dimensional space–time. Monte Carlo methods were used to ensemble these closed random lines. The physical space wherein the vortex lines are defined is a cuboid including periodic boundary conditions. Reconnections are allowed aside from moving, growing, and shrinking of the vortex configuration. Therefore, the ensemble contains a variable number of closed vortex clusters, which are considered significant in realizing the deconfining phase transition. The average vortex density and length are studied as a function of vortex density [Formula: see text] by using this model. At different temperatures, potential [Formula: see text] between quark and antiquark is studied as a function of distance [Formula: see text]. This model is able to emphasize the confinement physics qualitative properties in [Formula: see text] Yang–Mills theory.

Physics↗