Quantification of Model-Form Uncertainty in Drift-Diffusion Simulation Using Fractional Derivatives

In modeling and simulation, model-form uncertainty arises from the lack of knowledge and simplification during modeling process and numerical treatment for ease of computation. Traditional uncertainty quantification approaches are based on assumptions of stochasticity in real, reciprocal, or functional spaces to make them computationally tractable. This makes the prediction of important quantities of interest such as rare events difficult. In this paper, a new approach to capture model-form uncertainty is proposed. It is based on fractional calculus, and its flexibility allows us to model a family of non-Gaussian processes, which provides a more generic description of the physical world. A generalized fractional Fokker-Planck equation (fFPE) is proposed to describe the drift-diffusion processes under long-range correlations and memory effects. A new model calibration approach based on the maximum accumulative mutual information is also proposed to reduce model-form uncertainty, where an optimization procedure is taken.

Tolerance Specification and Related Issues for Additively Manufactured Products

Additive manufacturing (AM) has gained increased attention in the last decade as a versatile manufacturing process for customized products. AM processes can create complex free-form shapes, introducing features such as internal cavities and lattices. These complex geometries are either not feasible or very costly with traditional manufacturing processes. This creates new challenges in maintaining and communicating dimensional and geometric accuracy of parts produced. In order to manufacture a product that meets functional needs, the specification of those needs through geometry, material and tolerances is necessary. This paper surveys the current state and needs of geometry related accuracy specification mechanisms for AM, including a review of specification standards such as ASME Y14.5 and ISO 1101. Emerging AM-related tolerancing challenges are identified, and a potential plan of action is put forth for addressing those challenges. Various issues highlighted in this paper are classified as (a) AM-driven specification issues and (b) specification issues highlighted by the versatility of AM processes. AM-driven specification issues include build direction, layer thickness, support structure related specification, and scan/track direction. Specification issues highlighted by the versatility of AM processes include, region-based tolerances for complex freeform surfaces, tolerancing internal functional features, tolerancing lattice and infills. Basic methods of solving these specification issues are also highlighted.

Integrating Rule-Based Systems and Data Analytics Tools Using Open Standard PMML

Rule-based expert systems such as CLIPS (C Language Integrated Production System) are 1) based on inductive (if-then) rules to elicit domain knowledge and 2) designed to reason new knowledge based on existing knowledge and given inputs. Recently, data mining techniques have been advocated for discovering knowledge from massive historical or real-time sensor data. Combining top-down expert-driven rule models with bottom-up data-driven prediction models facilitates enrichment and improvement of the predefined knowledge in an expert system with data-driven insights. However, combining is possible only if there is a common and formal representation of these models so that they are capable of being exchanged, reused, and orchestrated among different authoring tools. This paper investigates the open standard PMML (Predictive Model Mockup Language) in integrating rule-based expert systems with data analytics tools, so that a decision maker would have access to powerful tools in dealing with both reasoning-intensive tasks and data-intensive tasks. We present a process planning use case in the manufacturing domain, which is originally implemented as a CLIPS-based expert system. Different paradigms in interpreting expert system facts and rules as PMML models (and vice versa), as well as challenges in representing and composing these models, have been explored. They will be discussed in detail.

Learning Algorithm Based Modeling and Process Parameters Recommendation System for Binder Jetting Additive Manufacturing Process

Binder Jetting (BJ) process is an additive manufacturing process in which powder materials are selectively joined by binder materials. Products can be manufactured layer by layer directly from 3D model data. It is not always easy for manufacturing engineers to choose proper BJ process parameters to meet the end-product quality and fabrication time requirements. This is because the quality properties of the products fabricated by BJ process are significantly affected by the process parameters. And the relationships between process parameters and quality properties are very complicated. In this paper, a process model is developed by Backward Propagation (BP) Neural Network (NN) algorithm based on 16 groups of orthogonal experiment designed by Taguchi Method to express the relationships between 4 key process parameters and 2 key quality properties. Based on the modeling results, an intelligent parameters recommendation system is developed to predict end-product quality properties and printing time, and to recommend process parameters selection based on the process requirements. It can be used as a guideline for selecting the proper printing parameters in BJ to achieve the desired properties and help to reduce the printing time.

A Design and Fabrication Framework for Periodic Lattice-Based Cellular Structures in Additive Manufacturing

A design framework that incorporates a size optimization algorithm is proposed for periodic lattice-based cellular structures fabricated by additive manufacturing. A 3D modeling process for the lattice-based cellular structures is integrated into the design framework for non-linear finite element analysis (FEA) and production. Material properties for the 3D printed parts are determined for the finite element study using reverse engineering of actual measured data. The lattice layout that will be used in the optimization is selected and the size of the cross sections is optimized using in-house optimization approach for both yield and local buckling criteria. The 3D model for the optimized lattice structure is built and non-linear finite element study is conducted to predict the performance. The approach is demonstrated on a compression block with periodic lattice-based unit cells. Physical parts are 3D printed and tested to compare with the simulations.

Plane Recognition on the Basis of Cloud of Points Determined by Photogrammetry

The article presents the results of research on developing a mathematical model allowing to identify the planes in a measured object, based on a cloud of points located on its walls. A basic mathematical description of surfaces is presented. An algorithm for determining the general equation of the plane on the basis of three points described in the Cartesian coordinate system has been developed. The algorithm has then been used to determine all possible planes in a given set of measured points, which were then subjected to the process of elimination and normalization. Filtered equations of the planes were grouped in order to finally determine the set of the sought-after equations. The designed algorithms have been implemented in a C++ computer program and their effect has been verified on a sample object with 3 methods of measurement: a contact measurement, a structured-light scanner method and with photogrammetry. The calculated equations have been compared with equations developed by a referenced commercial software. The results of the comparison have proved the correctness of the tested algorithms. The mathematical model will be used for rapid assessment of the geometrical parameters of castings and of the size of the machining allowances, as well as in the automatic settings of the casting in the machining space. The research was supported by the National Centre for Research and Development, Poland within the ongoing project No. LIDER / 07/76 / L-3/11 / NCBR / 2012.

Graph Visualization Styles for Use in Configuration Management: A User Study

The purpose of this research is to conduct a user study in order to determine the effect of numerous variables for data representation on the ability to answer questions about the system being represented. This research will be used in the development of a computer-based visualization tool to support configuration change management. The researchers hypothesized that the graph geometry and order of the questions being asked would not affect the results, while the color of the graph and the information being represented would affect the number of correct responses. The results showed an increase in the response accuracy for the answerable questions when the amount of information displayed in the data representation was minimized. On the other hand, none of the other factors showed to have a significant effect on the accuracy of the responses. The most significant limitation in this study was the possibility for different users putting different levels of effort into answering the questions.

Solid Mechanics Based Design and Optimization for Support Structure Generation in Stereolithography Based Additive Manufacturing

Support structures are typically required to hold parts in place in various additive manufacturing processes. Design of support structure includes identifying both anchor locations and geometries. Extensive work has been done to optimize the anchor locations to reliably keep part in position, and minimize the contacting area as well as the total volume of the support structures. However, relatively few studies have been focused on the mechanical property analysis of the structure. In this paper, we proposed a novel design optimization method to identify the anchor geometry based on solid mechanics theory. Finite element analysis method is utilized to study the stress distribution on both the support structure and main part. Particle Swarm Optimization (PSO) algorithm with a novel constraining handling strategy is employed to optimize the design model. A gradient descent local search algorithm is utilized to quickly locate the global solution in the vicinity explored by PSO. The developed optimization framework is deployed on a bottom-up projection based Stereolithography process. The experimental results show that the optimized design can efficiently reduce the material used on support structure and marks left on the part.

Identifying Possible Patient Slips and Falls Using Motion Capture Experiments

The Centers for Disease and Control report that falls are the most common cause of injury in older adults. Moderate to severe fall-related injuries significantly interfere with independent living and reduce quality of life, and it is necessary to prevent these falls whenever possible. The present study seeks to identify factors within a hospital bedroom and bathroom setting that may lead to falls. A motion capture experiment was conducted in a laboratory setting on thirty subjects over the age of seventy using one bedroom and two bathroom mockups designed to match the dimensions and layout of a representative room drawn from the archives of a large healthcare design firm. Data were post processed using Cortex and Visual3D software. A potential fall was defined as a period of time during which the jerk trajectory of the upper body’s center of mass remained consistently high. Preliminary results suggest that falls are more likely to occur when a patient is reaching, taking backwards steps, or turning. Future work includes locating each potential fall in a video recording to be analyzed by healthcare professionals including healthcare designers, clinicians, and a kinesiology expert. Identifying potential falls may lead to safer designs for hospital bedrooms and bathrooms and improved education for elderly adults about how to prevent falls.