Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “machining”

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 1,747 records · Page 97
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

NSTX-U Machine Core Vacuum Seals Upgrade Design

The vacuum in the National Spherical Torus Experiment Upgrade (NSTX-U) is in the range of 2.66x10⁻⁶ Pa (2x 10⁻⁸ torr). To enhance vacuum seals and lower gas permeation in the NSTX-U and therefore ensuring operation safety and quality, the machine core vacuum seal design has been upgraded using double elastomeric seals and pumped interspaces (guard channels). In NSTX-U machine core, there are six major vacuum seals, including upper and lower vacuum vessel main flanges, upper and lower bellows flanges, and two in the ceramic isolation ring assembly that are mounted to the upper vacuum vessel main flange. The elastomeric seals are commercially available and customized fluorocarbon (Viton) O-rings. To ensure appropriate O-ring compression ratios for the vacuum seal, the O-ring grooves are customized with appropriate tolerances considering specific situations for each seal. In this upgrade, the vacuum in each interspace is designed to reach 13.33 Pa (0.1 torr), thus the overall pressure difference crossing each double O-ring seal is expected to be reduced by at least three orders of magnitude, and accordingly the overall rates of leakage and permeation will be lowered by three orders of magnitude. The corresponding pump-down times for each seal, at room temperature, are estimated at about 60 s. Additionally, the vacuum pumped interspaces will be able to function as a real-time leak monitor.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2018

Dynamic State Estimation for Multi-Machine Power System by Unscented Kalman Filter With Enhanced Numerical Stability

In this paper, in order to enhance the numerical stability of the unscented Kalman filter (UKF) used for power system dynamic state estimation, a new UKF with guaranteed positive semidifinite estimation error covariance (UKFGPS) is proposed and compared with five existing approaches, including UKFschol, UKF-kappa, UKFmodified, UKF-Delta Q, and the squareroot UKF (SRUKF). These methods and the extended Kalman filter (EKF) are tested by performing dynamic state estimation on WSCC 3-machine 9-bus system and NPCC 48-machine 140-bus system. For WSCC system, all methods obtain good estimates. However, for NPCC system, both EKF and the classic UKF fail. It is found that UKFschol, UKF-kappa, and UKF-Delta Q do not work well in some estimations while UKFGPS works well in most cases. UKFmodified and SRUKF can always work well, indicating their better scalability mainly due to the enhanced numerical stability.

dynamic state estimation↗
DOE OSTI2020

Enhancement of microbiome management by machine learning for biological wastewater treatment

Here we propose to develop microbiome-based machine learning models to predict the response of biological wastewater treatment systems to environmental or operational disturbances or to design specific microbiomes to achieve a desired system function. These machine learning models can be used to enhance the stability of microbiome-based biological systems and warn against the failure of these systems.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Community‐Level Metabolic Shifts Following Land Use Change in the Amazon Rainforest Identified by a Supervised Machine Leaning Approach

ABSTRACT The Amazon rainforest has been subjected to high rates of deforestation, mostly for pasturelands, over the last few decades. This change in plant cover is known to alter the soil microbiome and the functions it mediates, but the genomic changes underlying this response are still unresolved. In this study, we used a combination of deep shotgun metagenomics complemented by a supervised machine learning approach to compare the metabolic strategies of tropical soil microbial communities in pristine forests and long‐term established pastures in the Amazon. Machine learning‐derived metagenome analysis indicated that microbial community structures (bacteria, archaea and viruses) and the composition of protein‐coding genes were distinct in each plant cover type environment. Forest and pasture soils had different genomic diversities for the above three taxonomic groups, characterised by their protein‐coding genes. These differences in metagenome profiles in soils under forests and pastures suggest that metabolic strategies related to carbohydrate and energy metabolisms were altered at community level. Changes were also consistent with known modifications to the C and N cycles caused by long‐term shifts in aboveground vegetation and were also associated with several soil physicochemical properties known to change with land use, such as the C/N ratio, soil temperature and exchangeable acidity. In addition, our analysis reveals that these alterations in land use can also result in changes to the composition and diversity of the soil DNA virome. Collectively, our study indicates that soil microbial communities shift their overall metabolic strategies, driven by genomic alterations observed in pristine forests and long‐term established pastures with implications for the C and N cycles.

carbon and nitrogen cycles↗
DOE OSTI2019

Evaluating Optimization Strategies for Engine Simulations Using Machine Learning Emulators

This work evaluates different optimization algorithms for computational fluid dynamics (CFD) simulations of engine combustion. Due to the computational expense of CFD simulations, emulators built with machine learning algorithms were used as surrogates for the optimizers. Two types of emulators were used: a Gaussian process (GP) and a weighted variety of machine learning methods called SuperLearner (SL). The emulators were trained using a dataset of 2048 CFD simulations that were run concurrently on a supercomputer. The design of experiments (DOE) for the CFD runs was obtained by perturbing nine input parameters using a Monte-Carlo method. The CFD simulations were of a heavy duty engine running with a low octane gasoline-like fuel at a partially premixed compression ignition mode. Ten optimization algorithms were tested, including types typically used in research applications. Each optimizer was allowed 800 function evaluations and was randomly tested 100 times. The optimizers were evaluated for the median, minimum, and maximum merits obtained in the 100 attempts. Some optimizers required more sequential evaluations, thereby resulting in longer wall clock times to reach an optimum. The best performing optimization methods were particle swarm optimization (PSO), differential evolution (DE), GENOUD (an evolutionary algorithm), and micro-genetic algorithm (GA). These methods found a high median optimum as well as a reasonable minimum optimum of the 100 trials. Moreover, all of these methods were able to operate with less than 100 successive iterations, which reduced the wall clock time required in practice. Two methods were found to be effective but required a much larger number of successive iterations: the DIRECT and MALSCHAINS algorithms. A random search method that completed in a single iteration performed poorly in finding optimum designs but was included to illustrate the limitation of highly concurrent search methods. The last three methods, Nelder–Mead, bound optimization by quadratic approximation (BOBYQA), and constrained optimization by linear approximation (COBYLA), did not perform as well.

Probst, Daniel M.↗
DOE OSTI2020

The Value of Information From Horizontal Distributed Acoustic Sensing Compared to Multicomponent Geophones Via Machine Learning

Abstract Faults play an important role in recharging many geothermal reservoirs, and seismic information can image the locations of these faults. The value of information (VOI) metric is used to objectively quantify and compare the value of two types of seismic receiver data via a machine learning approach. The demonstrated VOI methodology is novel by including spatial models from seismic data and obtaining the information statistics from machine learning. Our two-dimensional numerical experiments compare images created from sparsely spaced (80 m), two-component geophone sampling to high spatial resolution (1 m), single-component DAS. We used a three-fold cross validation of a U-Net convolutional neural networks to achieve average classification statistics. The results suggest that when horizontal sources are utilized, geophones and DAS identify reflectors and non-reflectors at roughly the same rate. The average F1 score for horizontal DAS is 0.939 and 0.931 for geophones. For images created from a vertical source, DAS performed marginally better (F1 = 0.919) than geophones (F1 = 0.877). Our transferrable methodology can provide guidance on which acquisition scenarios can improve images of important structures in the subsurface and present an efficient method for obtaining reliability statistics from high-dimensional, spatial data.

Energy & Fuels↗
DOE OSTI2022

Indirect Tool Condition Monitoring Using Ensemble Machine Learning Techniques

Abstract Tool condition monitoring (TCM) has become a research area of interest due to its potential to significantly reduce manufacturing costs while increasing process visibility and efficiency. Machine learning (ML) is one analysis technique which has demonstrated advantages for TCM applications. However, the commonly studied individual ML models lack generalizability to new machining and environmental conditions, as well as robustness to the unbalanced datasets which are common in TCM. Ensemble ML models have demonstrated superior performance in other fields, but have only begun to be evaluated for TCM. As a result, it is not well understood how their TCM performance compares to that of individual models, or how homogeneous and heterogeneous ensemble models’ performances compare to one another. To fill in these research gaps, milling experiments were conducted using various cutting conditions, and the model groups were compared across several performance metrics. Statistical t-tests were also used to evaluate the significance of model performance differences. Through the analysis of four individual ML models and five ensemble models, all based on the processes’ sound, spindle power, and axial load signals, it was found that on average, the ensemble models performed better than the individual models, and that the homogeneous ensembles outperformed the heterogeneous ensembles.

Engineering↗
DOE OSTI2011

Implementing RapidArc into clinical routine: A comprehensive program from machine QA to TPS validation and patient QA

Purpose: With the increased commercial availability of intensity modulated arc therapy (IMAT) comes the need for comprehensive QA programs, covering the different aspects of this newly available technology. This manuscript proposes such a program for the RapidArc (RA) (Varian Medical Systems, Palo Alto) IMAT solution. Methods: The program was developed and tested out for a Millennium120 MLC on iX Clinacs and a HighDefinition MLC on a Novalis TX, using a variety of measurement equipment including Gafchromic film, 2D ion chamber arrays (Seven29 and StarCheck, PTW, Freiburg, Germany) with inclinometer and Octavius phantom, the Delta4 systam (ScandiDos, Uppsala, Sweden) and the portal imager (EPID). First, a number of complementary machine QA tests were developed to monitor the correct interplay between the accelerating/decelerating gantry, the variable dose rate and the MLC position, straining the delivery to the maximum allowed limits. Second, a systematic approach to the validation of the dose calculation for RA was adopted, starting with static gantry and RA specific static MLC shapes and gradually moving to dynamic gantry, dynamic MLC shapes. RA plans were then optimized on a series of artificial structures created within the homogeneous Octavius phantom and within a heterogeneous lung phantom. These served the double purpose of testing the behavior of the optimization algorithm (PRO) as well as the precision of the forward dose calculation. Finally, patient QA on a series of clinical cases was performed with different methods. In addition to the well established in-phantom QA, we evaluated the portal dosimetry solution within the Varian approach. Results: For routine machine QA, the ''Snooker Cue'' test on the EPID proved to be the most sensitive to overall problem detection. It is also the most practical one. The ''Twinkle'' and ''Sunrise'' tests were useful to obtain well differentiated information on the individual treatment delivery components. The AAA8.9 dose calculations showed excellent agreement with all corresponding measurements, except in areas where the 2.5 mm fixed fluence resolution was insufficient to accurately model the tongue and groove effect or the dose through nearly closed opposing leafs. Such cases benefited from the increased fluence resolution in AAA10.0. In the clinical RA fields, these effects were smeared out spatially and the impact of the fluence resolution was considerably less pronounced. The RA plans on the artificial structure sets demonstrated some interesting characteristics of the PRO8.9 optimizer, such as a sometimes unexpected dependence on the collimator rotation and a suboptimal coverage of targets within lung tissue. Although the portal dosimetry was successfully validated, we are reluctant to use it as a sole means of patient QA as long as no gantry angle information is embedded. Conclusions: The all-in validation program allows a systematic approach in monitoring the different levels of RA treatments. With the systematic approach comes a better understanding of both the capabilities and the limits of the used solution. The program can be useful for implementation, but also for the validation of major upgrades.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2012

Correction factors for A1SL ionization chamber dosimetry in TomoTherapy: Machine-specific, plan-class, and clinical fields

Purpose: Recently, an international working group on nonstandard fields presented a new formalism for ionization chamber reference dosimetry of small and nonstandard fields [Alfonso et al., Med. Phys. 35, 5179-5186 (2008)] which has been adopted by AAPM TG-148. This work presents an experimental determination of the correction factors for reference dosimetry with an Exradin A1SL thimble ionization chamber in a TomoTherapy unit, focusing on: (i) machine-specific reference field, (ii) plan-class-specific reference field, and (iii) two clinical treatments. Methods: Ionization chamber measurements were performed in the TomoTherapy unit for intermediate (machine-specific and plan-class-specific) calibration fields, based on the reference conditions defined by AAPM TG-148, and two clinical treatments (lung and head-and-neck). Alanine reference dosimetry was employed to determine absorbed dose to water at the point of interest for the fields under investigation. The corresponding chamber correction factors were calculated from alanine to ionization chamber measurements ratios. Results: Two different methods of determining the beam quality correction factor k{sub Q,Q{sub 0}} for the A1SL ionization chamber in this TomoTherapy unit, where reference conditions for conventional beam quality determination cannot be met, result in consistent values. The observed values of overall correction factors obtained for intermediate and clinical fields are consistently around 0.98 with a typical expanded relative uncertainty of 2% (k = 2), which when considered make such correction factors compatible with unity. However, all of them are systematically lower than unity, which is shown to be significant when a hypothesis test assuming a t-student distribution is performed (p=1.8x10{sup -2}). Correction factors k{sub Q{sub c{sub l{sub i{sub n,Q{sub p{sub c{sub s{sub r}{sup f{sub c}{sub l}{sub i}{sub n},f{sub p}{sub c}{sub s}{sub r}}}}}}}}}} and k{sub Q{sub c{sub l{sub i{sub n,Q{sub m{sub s{sub r}{sup f{sub c}{sub l}{sub i}{sub n},f{sub m}{sub s}{sub r}}}}}}}}}, which are needed for the computation of field factors for relative dosimetry of clinical beams, have been found to be very close to unity for two clinical treatments. Conclusions: The results indicate that the helical field deliveries in this study (including two clinical fields) do not introduce changes on the ion chamber correction factors for dosimetry. For those two specific clinical cases, ratios of chamber readings accurately represent field output factors. The values observed here for intermediate calibration fields are in agreement with previously published data based on alanine dosimetry but differ from values recently reported obtained via radiochromic dosimetry.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2012

Systematic measurements of whole-body dose distributions for various treatment machines and delivery techniques in radiation therapy

Purpose: Contemporary radiotherapy treatment techniques, such as intensity-modulated radiation therapy and volumetric modulated arc therapy, could increase the radiation-induced malignancies because of the increased beam-on time, i.e., number of monitor units needed to deliver the same dose to the target and the larger volume irradiated with low doses. In this study, whole-body dose distributions from typical radiotherapy patient plans using different treatment techniques and therapy machines were measured using the same measurement setup and irradiation intention. Methods: Individually calibrated thermoluminescent dosimeters were used to measure absorbed dose in an anthropomorphic phantom at 184 locations. The dose distributions from 6 MV beams were compared in terms of treatment technique (3D-conformal, intensity-modulated radiation therapy, volumetric modulated arc therapy, helical TomoTherapy, stereotactic radiotherapy, hard wedges, and flattening filter-free radiotherapy) and therapy machine (Elekta, Siemens and Varian linear accelerators, Accuray CyberKnife and TomoTherapy). Results: Close to the target, the doses from intensity-modulated treatments (including flattening filter-free) were below the dose from a static treatment plan, whereas the CyberKnife showed a larger dose by a factor of two. Far away from the treatment field, the dose from intensity-modulated treatments showed an increase in dose from stray radiation of about 50% compared to the 3D-conformal treatment. For the flattening filter-free photon beams, the dose from stray radiation far away from the target was slightly lower than the dose from a static treatment. The CyberKnife irradiation and the treatment using hard wedges increased the dose from stray radiation by nearly a factor of three compared to the 3D-conformal treatment. Conclusions: This study showed that the dose outside of the treated volume is influenced by several sources. Therefore, when comparing different treatment techniques, the dose ratios vary with distance to the isocenter. The effective dose outside the treated volume of intensity-modulated treatments with or without flattening filter was 10%-30% larger when compared to 3D-conformal radiotherapy. This dose increase is much lower than the monitor unit scaled effective dose from a static treatment.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-J-191: Motion Prediction Using Extreme Learning Machine in Image Guided Radiotherapy

Purpose: Real-time motion tracking is a critical issue in image guided radiotherapy due to the time latency caused by image processing and system response. It is of great necessity to fast and accurately predict the future position of the respiratory motion and the tumor location. Methods: The prediction of respiratory position was done based on the positioning and tracking module in ARTS-IGRT system which was developed by FDS Team (www.fds.org.cn). An approach involving with the extreme learning machine (ELM) was adopted to predict the future respiratory position as well as the tumor’s location by training the past trajectories. For the training process, a feed-forward neural network with one single hidden layer was used for the learning. First, the number of hidden nodes was figured out for the single layered feed forward network (SLFN). Then the input weights and hidden layer biases of the SLFN were randomly assigned to calculate the hidden neuron output matrix. Finally, the predicted movement were obtained by applying the output weights and compared with the actual movement. Breathing movement acquired from the external infrared markers was used to test the prediction accuracy. And the implanted marker movement for the prostate cancer was used to test the implementation of the tumor motion prediction. Results: The accuracy of the predicted motion and the actual motion was tested. Five volunteers with different breathing patterns were tested. The average prediction time was 0.281s. And the standard deviation of prediction accuracy was 0.002 for the respiratory motion and 0.001 for the tumor motion. Conclusion: The extreme learning machine method can provide an accurate and fast prediction of the respiratory motion and the tumor location and therefore can meet the requirements of real-time tumor-tracking in image guided radiotherapy.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-225: Correction Matrix for PinPoint Ionization Chamber for Dosimetric Measurements in the Newly Released Incise™ Multileaf Collimator Shaped Small Field for CyberKnife M6™ Machine

Purpose: For small field dosimetry, such as measurements of output factors for cones or MLC-shaped irregular small fields, ion chambers often Result in an underestimation of the dose, due to both the volume averaging effect and the lack of lateral charged particle equilibrium. This work presents a mathematical model for correction matrix for a PTW PinPoint ionization chamber for dosimetric measurements made in the newly released Incise™ Multileaf collimator fields of the CyberKnife M6™ machine. Methods: A correction matrix for a PTW 0.015cc PinPoint ionization chamber was developed by modeling its 3D dose response in twelve cone-shaped circular fields created using the 5mm, 7.5mm, 10mm, 12.5mm, 15mm, 20mm, 25mm, 30mm, 35mm, 40mm, 50mm, 60mm cones in a CyberKnife M6™ machine. For each field size, hundreds of readings were recorded for every 2mm chamber shift in the horizontal plane. The contribution of each dose pixel to a measurement point depended on the radial distance and the angle to the chamber axis. These readings were then compared with the theoretical dose as obtained with Monte Carlo calculation. A penalized least-square optimization algorithm was developed to generate the correction matrix. After the parameter fitting, the mathematical model was validated for MLC-shaped irregular fields. Results: The optimization algorithm used for parameter fitting was stable and the resulted response factors were smooth in spatial domain. After correction with the mathematical model, the chamber reading matched with the calculation for all the tested fields to within 2%. Conclusion: A novel mathematical model has been developed for PinPoint chamber for dosimetric measurements in small MLC-shaped irregular fields. The correction matrix is dependent on detector, treatment unit and the geometry of setup. The model can be applied to non-standard composite fields and provides an access to IMRT point dose validation.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-255: Development of a Michigan Quality Assurance (MQA) Database for Clinical Machine Operations

Purpose: A unified database system was developed to allow accumulation, review and analysis of quality assurance (QA) data for measurement, treatment, imaging and simulation equipment in our department. Recording these data in a database allows a unified and structured approach to review and analysis of data gathered using commercial database tools. Methods: A clinical database was developed to track records of quality assurance operations on linear accelerators, a computed tomography (CT) scanner, high dose rate (HDR) afterloader and imaging systems such as on-board imaging (OBI) and Calypso in our department. The database was developed using Microsoft Access database and visual basic for applications (VBA) programming interface. Separate modules were written for accumulation, review and analysis of daily, monthly and annual QA data. All modules were designed to use structured query language (SQL) as the basis of data accumulation and review. The SQL strings are dynamically re-written at run time. The database also features embedded documentation, storage of documents produced during QA activities and the ability to annotate all data within the database. Tests are defined in a set of tables that define test type, specific value, and schedule. Results: Daily, Monthly and Annual QA data has been taken in parallel with established procedures to test MQA. The database has been used to aggregate data across machines to examine the consistency of machine parameters and operations within the clinic for several months. Conclusion: The MQA application has been developed as an interface to a commercially available SQL engine (JET 5.0) and a standard database back-end. The MQA system has been used for several months for routine data collection.. The system is robust, relatively simple to extend and can be migrated to a commercial SQL server.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

SU-E-T-568: Neutron Dose Survey of a Compact Single Room Proton Machine

Purpose: To ensure acceptable radiation limits are maintained for those working at and near the machine during its operation, a comprehensive radiation survey was performed prior to the clinical release of Mevion S250 compact proton machine at Stephenson Oklahoma Cancer Center. Methods: The Mevion S250 proton therapy system consists of the following: a superconducting cyclotron to accelerate the proton particles, a passive double scattering system for beam shaping, and paired orthogonal x-ray imaging systems for patient setup and verification via a 6D robotic couch. All equipment is housed within a single vault of compact design. Two beam delivery applicators are available for patient treatment, offering field sizes of as great as 14 cm and 25 cm in diameter, respectively. Typical clinical dose rates are between 1 and 2 Gy/min with a fixed beam energy of 250 MeV. The large applicator (25 cm in diameter) was used in conjunction with a custom cut brass aperture to create a 20 cm x 20 cm field size at beam isocenter. A 30 cm − 30 cm − 35 cm high density plastic phantom was placed in the beam path to mimic the conditions creating patient scatter. Measurements integrated-ambient-neutron-dose-equivalence were made with a SWENDII detector. Gantry angles of 0, 90 and 180 degrees, with a maximum dose rate of 150 MU/min (for large applicator) and beam configuration of option 1 (range 25 cm and 20 cm modulation), were selected as testing conditions. At each point of interest, the highest reading was recorded at 30 cm from the barrier surface. Results: The highest neutron dose was estimated to be 0.085 mSv/year at the console area. Conclusion: All controlled areas are under 5 mSv/year and the uncontrolled areas are under 1 mSv/year. The radiation protection provided by the proton vault is of sufficient quality.

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2015

SU-E-T-636: ProteusONE Machine QA Procedure and Stabiity Study: Half Year Clinical Operation

Purpose: The objective of this study is to evaluate the stability of ProteusOne, the 1st commercial PBS proton system, throughout the daily QA and monthly over 6 month clinical operation. Method: Daily QA test includes IGRT position/repositioning, output in the middle of SOBP, beam flatness, symmetry, inplane and crossplane dimensions as well as energy range check. Daily range shifter QA consist of output, symmetry and field size checks to make sure its integrity. In 30 mins Daily QA test, all the measurements are performed using the MatriXXPT (IBA dosimetry). The data from these measurement was collected and compare over the first 6 month of clinical operation. In addition to the items check in daily QA, the summary also includes the monthly QA gantry star shots, absolute position check using a novel device, XRV-100. Results: Average machine output at the center of the spread out bragg peak was 197.5±.8 cGy and was within 1%of the baseline at 198.4 cGy. Beam flatness was within 1% cross plane with an average of 0.67±0.12% and 2% in-plane with an average of 1.08±0.17% compared to baseline measurements of 0.6 and 1.03, respectively. In all cases the radiation isocenter shift was less than or equal to 1mm. Output for the range shifter was within 2% for each individual measurement and averaged 34.4±.2cGy compare to a baseline reading of 34.5cGy. The average range shifter in and cross plane field size measurements were 19.8±0.5cm and 20.5±0.4cm compared with baseline values of 20.19cm and 20.79cm, respectively. Range shifter field symmetry had an average of less 1% for both in-plane and cross plane measurements. Conclusion: All machine metrics over the past 6 months have proved to be stable. Although, some averages are outside the baseline measurement they are within 1% tolerance and the deviation across all measurements is minimal.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2015

SU-E-T-660: Quantitative Fault Testing for Commissioning of Proton Therapy Machines

Purpose: To ensure proper fault testing for the first single room proton therapy machine by establishing a common set of acceptance testing and commissioning parameters with the manufacturer. The following work details the parameters tested and associated results. Methods: Dose rates in service mode were varied to ensure that when the threshold for maximum or minimum MU/min was met, the beam promptly shut off. The flatness parameter was tested by purposely assigning an incorrect secondary scatter, to ensure the beam shut off when detecting a heterogeneous profile. The beam symmetry parameter was tested by altering the steering coil up to 3.0A, thereby forcing the beam to be asymmetric and shut off. Lastly, the quench system was tested by ramping down the magnet to 5% capacity, whereby the quench button was engaged to bring down the magnet current to a safe level. Results: A dose rate increase or decrease in excess of 10% shut the beam off within 5 seconds as observed by the current on a Matrixx ionization chamber array (IBA Dosimetry, Bartlett, TN) A 3.0A change in the beam steering coil introduced a 2% change in the flatness and symmetry profiles with respect to baseline measurements resulting in the beam shutting off within 5 seconds. An incorrect 2nd scatterer introduced a flatness of 4.1% and symmetry of 6.4% which immediately triggered a beam shut off. Finally, the quench system worked as expected during the ramp down procedure. Conclusion: A fault testing plan to check dosimetric faults and the quench system was performed for the first single room proton therapy system. All dosimetric parameters and machine conditions were met to our satisfaction. We propose that the same type of fault testing should be applied to any proton system during commissioning, including scanning beam systems.

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2015

MO-F-CAMPUS-J-02: Automatic Recognition of Patient Treatment Site in Portal Images Using Machine Learning

Purpose: To investigate the method to automatically recognize the treatment site in the X-Ray portal images. It could be useful to detect potential treatment errors, and to provide guidance to sequential tasks, e.g. automatically verify the patient daily setup. Methods: The portal images were exported from MOSAIQ as DICOM files, and were 1) processed with a threshold based intensity transformation algorithm to enhance contrast, and 2) where then down-sampled (from 1024×768 to 128×96) by using bi-cubic interpolation algorithm. An appearance-based vector space model (VSM) was used to rearrange the images into vectors. A principal component analysis (PCA) method was used to reduce the vector dimensions. A multi-class support vector machine (SVM), with radial basis function kernel, was used to build the treatment site recognition models. These models were then used to recognize the treatment sites in the portal image. Portal images of 120 patients were included in the study. The images were selected to cover six treatment sites: brain, head and neck, breast, lung, abdomen and pelvis. Each site had images of the twenty patients. Cross-validation experiments were performed to evaluate the performance. Results: MATLAB image processing Toolbox and scikit-learn (a machine learning library in python) were used to implement the proposed method. The average accuracies using the AP and RT images separately were 95% and 94% respectively. The average accuracy using AP and RT images together was 98%. Computation time was ∼0.16 seconds per patient with AP or RT image, ∼0.33 seconds per patient with both of AP and RT images. Conclusion: The proposed method of treatment site recognition is efficient and accurate. It is not sensitive to the differences of image intensity, size and positions of patients in the portal images. It could be useful for the patient safety assurance. The work was partially supported by a research grant from Varian Medical System.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2016

SU-C-BRA-04: Automated Segmentation of Head-And-Neck CT Images for Radiotherapy Treatment Planning Via Multi-Atlas Machine Learning (MAML)

Purpose: Accurate image segmentation is a crucial step during image guided radiation therapy. This work proposes multi-atlas machine learning (MAML) algorithm for automated segmentation of head-and-neck CT images. Methods: As the first step, the algorithm utilizes normalized mutual information as similarity metric, affine registration combined with multiresolution B-Spline registration, and then fuses together using the label fusion strategy via Plastimatch. As the second step, the following feature selection strategy is proposed to extract five feature components from reference or atlas images: intensity (I), distance map (D), box (B), center of gravity (C) and stable point (S). The box feature B is novel. It describes a relative position from each point to minimum inscribed rectangle of ROI. The center-of-gravity feature C is the 3D Euclidean distance from a sample point to the ROI center of gravity, and then S is the distance of the sample point to the landmarks. Then, we adopt random forest (RF) in Scikit-learn, a Python module integrating a wide range of state-of-the-art machine learning algorithms as classifier. Different feature and atlas strategies are used for different ROIs for improved performance, such as multi-atlas strategy with reference box for brainstem, and single-atlas strategy with reference landmark for optic chiasm. Results: The algorithm was validated on a set of 33 CT images with manual contours using a leave-one-out cross-validation strategy. Dice similarity coefficients between manual contours and automated contours were calculated: the proposed MAML method had an improvement from 0.79 to 0.83 for brainstem and 0.11 to 0.52 for optic chiasm with respect to multi-atlas segmentation method (MA). Conclusion: A MAML method has been proposed for automated segmentation of head-and-neck CT images with improved performance. It provides the comparable result in brainstem and the improved result in optic chiasm compared with MA. Xuhua Ren and Hao Gao were partially supported by the NSFC (#11405105), the 973 Program (#2015CB856000), and the Shanghai Pujiang Talent Program (#14PJ1404500).

60 APPLIED LIFE SCIENCES↗