DASS Good: Explainable Data Mining of Spatial Cohort Data

Developing applicable clinical machine learning models is a difficult task when the data includes spatial information, for example, radiation dose distributions across adjacent organs at risk. We describe the co-design of a modeling system, DASS, to support the hybrid human-machine development and validation of predictive models for estimating long-term toxicities related to radiotherapy doses in head and neck cancer patients. Developed in collaboration with domain experts in oncology and data mining, DASS incorporates human-in-the-loop visual steering, spatial data, and explainable AI to augment domain knowledge with automatic data mining. We demonstrate DASS with the development of two practical clinical stratification models and report feedback from domain experts. Finally, we describe the design lessons learned from this collaborative experience.     

Determination of the physical and mechanical properties of iron-based powder materials produced by microwave sintering

Microwave energy is highly efficient for heating and processing different materials. In recent years, this type of heat transfer has been used in sintering process. Rapid and highly efficient heating, time and energy saving, and improved properties of sintered materials are advantages of microwave sintering. In this paper, Fe and Fe-Cu powder compact samples (cylindrical and bone shapes) are sintered both in microwave and electrical tube furnaces. The microwave generator has 2.45 GHz frequency and 1 KW power. Times are selected in the range of 5–25 min for microwave sintering and 5–40 min for electrical heating. The sintering temperature is set at 1120°C. Samples are sintered in the reducing atmosphere of 95% N₂ + 5% H₂ mixture. The density, hardness, and tensile strength of the samples are measured. The results are compared. The results show that the microwave-sintered materials have a finer microstructure. The microwave-sintered materials have 6–8% higher density, 5–10 HV5 higher hardness, and about 10% higher tensile strength than conventionally sintered materials.     

Uniform modeling of parameter dependent nonlinear systems

This paper addresses the problem of approximating parameter dependent nonlinear systems in a unified framework. This modeling has been presented for the first time in the form of parameter dependent piecewise affine systems. In this model, the matrices and vectors defining piecewise affine systems are affine functions of parameters. Modeling of the system is done based on distinct spaces of state and parameter, and the operating regions are partitioned into the sections that we call ’multiplied simplices’. It is proven that this method of partitioning leads to less complexity of the approximated model compared with the few existing methods for modeling of parameter dependent nonlinear systems. It is also proven that the approximation is continuous for continuous functions and can be arbitrarily close to the original one. Next, the approximation error is calculated for a special class of parameter dependent nonlinear systems. For this class of systems, by solving an optimization problem, the operating regions can be partitioned into the minimum number of hyper-rectangles such that the modeling error does not exceed a specified value. This modeling method can be the first step towards analyzing the parameter dependent nonlinear systems with a uniform method.     

Model updating of multistory shear buildings for simultaneous identification of mass,stiffness and damping matrices using two different soft-computing methods

In this research, two novel methods for simultaneous identification of mass–damping–stiffness of shear buildings are proposed. The first method presents a procedure to estimate the natural frequencies, modal damping ratios, and modal shapes of shear buildings from their forced vibration responses. To estimate the coefficient matrices of a state-space model, an auto-regressive exogenous excitation (ARX) model cooperating with a neural network concept is employed. The modal parameters of the structure are then evaluated from the eigenparameters of the coefficient matrix of the model. Finally, modal parameters are used to identify the physical/structural (i.e., mass, damping, and stiffness) matrices of the structure. In the second method, a direct strategy of physical/structural identification is developed from the dynamic responses of the structure without any eigenvalue analysis or optimization processes that are usually necessary in inverse problems. This method modifies the governing equations of motion based on relative responses of consecutive stories such that the new set of equations can be implemented in a cluster of artificial neural networks. The number of neural networks is equal to the number of degree-of-freedom of the structure. It is shown the noise effects may partially be eliminated by using high-order finite impulse response (FIR) filters in both methods. Finally, the feasibility and accuracy of the presented model updating methods are examined through numerical studies on multistory shear buildings using the simulated records with various noise levels. The excellent agreement of the obtained results with those of the finite element models shows the feasibility of the proposed methods.     

Oil thermal annealed nano‐structured indium tin oxide thin films for display applications

Indium tin oxide‐coated thin films (200 nm) are deposited on glass substrates by using R.f. sputtering technique. Here, we investigate the influence of new technique of treatment, which is called as “oil thermal annealing” on the nano‐structured indium tin oxide thin films at fixed temperature (150 °C) which improves adhesion strength, electrical conductivity and optical properties (transmittance) of the films. Oil thermal annealing is used to reduce inherent defects that may be introduced during the prepared thin film and cooling processes. Proposed technique is highly suitable for liquid crystal displays, solar cells and organic light emitting diodes, and many other display‐related applications.     

Uncertainty evaluation of calculated and measured kinetics parameters of Tehran Research Reactor

Effective delayed neutron fraction β_eff and neutron generation time Λ are important factors in reactor physics calculation and transient analysis. In the first stage of this research, these kinetics parameters have been calculated for two states of Tehran Research Reactor (TRR), i.e. cold (fuel, clad and coolant temperature 20 °C) and hot (fuel, clad and coolant temperature 65, 49 and 44 °C, respectively) states using MTR_PC computer code. The ratio of (β_eff)_i/(β_eff)_core plays an important role in reactivity accident analysis codes. This parameter and its contribution to effective delayed neutron fraction from each nucleus have been calculated in cold and hot reactor states. Uncertainty of effective delayed neutron fraction is evaluated in terms of following four quantities; basic delayed neutron constants, delayed neutron spectra, energy dependence of delayed neutron yield (ν_d) and fission cross-section of ²³⁵U and ²³⁸U. In the second stage, these parameters have been measured with an experimental method based on Inhour equation. The calculated and measured values are in good agreement. Relative Percent Errors (RPEs) are 2.8% for β_eff and 5.7% for Λ in the cold state.     

Development of a 2-D 2-group neutron noise simulator for hexagonal geometries

In this paper, the development of a neutron noise simulator for hexagonal-structured reactor cores using both the forward and the adjoint methods is reported. The spatial discretisation of both 2-D 2-group static and dynamic equations is based on a developed box-scheme finite difference method for hexagonal mesh boxes. Using the power iteration method for the static calculations, the 2-group neutron flux and its adjoint with the corresponding eigenvalues are obtained by the developed static simulator. The results are then benchmarked against the well-known CITATION computer code. The dynamic calculations are performed in the frequency domain which leads to discarding of the time discretisation. Then, the developed 2-D 2-group neutron noise simulator calculates both the discretised forward and the adjoint reactor transfer function between a point source and its induced neutron noise, by assuming the neutron noise source as an “absorber of variable strength” type. The neutron noise induced by a “vibrating absorber” type of noise source may also be modeled using the calculated transfer function. The viability of the simulator is verified for different benchmark cases.     

Robust Torque Control of Switched-Reluctance Motors Without a Shaft-Position Sensor

This paper describes a technique for stabilizing the operation of variable-reluctance stepping-motor drives operating without a shaft-position sensor. In such systems there is a trade-off between the system efficiency and the torque margin (or pull-out torque) which depends on the width of the phase conduction pulse width. The scheme described in the paper permits the motor to run in the steady state with both narrow conduction pulse widths and high efficiency. Under transient or overload conditions the conduction pulse width is increased in response to a change in the dc link current, providing an increase in available torque. Tests on a small motor drive have produced a steady-state torque margin of over 300 percent and of 200 percent under step-change conditions.     

Effects of composition and sintering time on liquid phase sintered Co-Cu samples in microgravity

Twelve Co-Cu powder compact samples with different liquid volume fractions were processed during microgravity liquid phase sintering on a suborbital sounding rocket and three Space Shuttle missions. The processing times ranged from 2.5 minutes to 66 minutes. The samples exhibited dimension stability after sintering. Microstructural evolutions such as densification, dihedral angle, contact per grain and grain growth rates, indicated a dependency on Cu composition and sintering time. Grain growth analysis showed a diffusion-controlled grain growth mechanism. The diffusional layer was found in a microgravity processed 70vol%Co-Cu sample. A mechanism that explains the transient nature of this diffusion layer is proposed and used to explain the results at other processing times. Agglomeration and coalescence of particles were observed in this study, and the grain size distributions were in agreement with LSEM model, which incorporates the effect of higher solid volume fraction and particle coalescence. Analysis of the samples also revealed considerable pore formation and metamorphosis. Unlike the Fe-Cu samples, in which pore breakup was found, pore filling and coarsening dominate in all Co-Cu samples. The evolution of these parameters has been used to enhance the understanding of driving forces that contribute to the pore metamorphosis during liquid phase sintering in the Co-Cu system under microgravity.