Flexible Variant Research Articles

Systematic analysis of geometric inaccuracy and its contributing factors in roboforming

Incremental sheet metal forming is a highly versatile die-less forming process for manufacturing complex sheet metal components. Robot-assisted incremental sheet forming, or roboforming, allows a wider range of tool motion, providing the capability to shape intricate components. This makes roboforming the most flexible variant of the incremental forming method. However, the serial arrangement of links and joints in a robotic manipulator results in low positional accuracy under forming loads due to insufficient structural stiffness. The stiffness of the machine frame and tool directly impacts the accuracy of the final formed profile. The impact of machine compliance on component shape in incremental sheet forming is substantial in roboforming. This work presents a methodology for systematic analysis of the factors contributing to the errors in the geometric shape of robot-based forming. Experimental and numerical methods are used to estimate the material springback, tool/tool holder deflections, and errors due to machine compliance. The CNC machine frame is relatively stiffer than the industrial robots, such that material springback is estimated based on the experimental trials on CNC for cone and variable wall angle cone profiles. Tool and tool holder deflections are estimated using finite element simulations. The analytical method using the Virtual Joint Model is used to model the joint stiffness, and consequently, the robot Cartesian stiffness is estimated to predict path deviation contributing to geometric shape errors. The proportional contribution of each factor in the overall deviation in the Roboforming is also quantified.

PatchSkip: A lightweight technique for effectively alleviating over-smoothing in vision transformers

Recently vision transformers (ViTs) have encountered the over-smoothing problem, which reduces their capacity by mapping input patches into a similar latent representation. Existing methods introduce regularization terms to alleviate over-smoothing but often increase computational costs. To address this, this paper proposes PatchSkip, a novel and flexible dropout variant, alleviating the over-smoothing problem of ViTs in a lightweight manner. Specifically, PatchSkip draws inspiration from the fact that a similar over-smoothing problem in GNNs is primarily caused by static adjacent matrices leading to solitary message passing mode between nodes. PatchSkip constructs graphs with patch embeddings and analyzes the adjacent matrices in ViTs. By randomly selecting specific patch embeddings to bypass transformer blocks, PatchSkip is proved to generate various adjacent matrices and acts as a multi-mode message passing engine, providing diverse modes of message passing between patches. The effectiveness of PatchSkip in preventing over-smoothing is demonstrated through theoretical proofs and empirical visualizations. Furthermore, PatchSkip is evaluated on various datasets and backbones, showing significant performance improvements while reducing computational costs. For example, when trained on Tiny-ImageNet from scratch, PatchSkip improves the performance of the vanilla CrossViT by 3.85% while reducing computational costs by over 20%.

Optimizing Text Clustering Efficiency through Flexible Latent Dirichlet Allocation Method: Exploring the Impact of Data Features and Threshold Modification

A parallel corpus comprising Croatian EU legislative documents automatically translated into English spans 28 years and is enriched with metadata, including creation year and hierarchical classifier tags denoting descriptors, document types, and fields. However, nearly two-thirds of the approximately 1.5 thousand texts lack complete metadata, necessitating labor intensive manual efforts that pose challenges for human administration. This incompleteness issue can be observed in the case of official legal sites functioning as regular service provisioning databases. In response, this paper introduces an artificial cognitive and multilabel classification approach to expedite the tagging process with only a fraction of the manual effort. Leveraging the Latent Dirichlet Allocation (LDA) algorithm, our method assigns field values or tags to incompletely labeled documents. We implement a Flexible LDA variant, incorporating the influence of topics close to the most probable topic, regulated by a relative probability threshold (RPT). We evaluate the LDA prediction's dependence on document prefiltering and RPT values. Furthermore, we investigate the dependence of quantitative linguistic properties on the type and speciality of pre-processing tasks. Our algorithm, built on error-correcting optimizing codes, succesfully predicts a mixture of topic probabilities for these legal texts. This prediction is achieved by calculating the Hamming distance of binary feature vectors created using the legal fields of the EUROVOC multilingual thesaurus.

Facile synthesis of 5α-functionalized pyrroles via multicomponent reactions of vicinal tricarbonyl compounds, enamines, and nucleophiles

A simple and general strategy was developed for the efficient synthesis of diverse flexible variant α-functionalized pyrrole derivatives.

A New Modification of Rani Distribution with More Flexibility in Application

In this article, the Rani distribution has been modified to a more flexible variant. The new model is called "the XRani distribution" and it is also a one-parameter distribution just as the parent distribution-Rani. The properties of the XRani distribution were derived and an analysis of the behaviour was conducted. The parameter was estimated using the maximum likelihood procedure. Four lifetime data sets were used to illustrate the importance of the proposed model. From the results, the proposed model competes favorably among the members of the Lindley class of distributions.

Improved Performance in the Control of DC-DC Three-Phase Power Electronic Converter Using Fractional-Order SMC and Synergetic Controllers and RL-TD3 Agent

In this article, starting from a benchmark represented by a Direct Current-to-Direct Current (DC-DC) three-phase power electronic converter used as an interface and interconnection between the grid and a DC microgrid, we compare the performances of a series of control structures—starting with the classical proportional integrator (PI) type and continuing with more advanced ones, such as sliding mode control (SMC), integer-order synergetic, and fractional-order (FO) controllers—in terms of maintaining the constant DC voltage of the DC microgrid. We present the topology and the mathematical modeling using differential equations and transfer functions of the DC-DC three-phase power electronic converter that provides the interface between the grid and a DC microgrid. The main task of the presented control systems is to maintain the DC voltage supplied to the microgrid at an imposed constant value, regardless of the total value of the current absorbed by the consumers connected to the DC microgrid. We present the elements of fractional calculus that were used to synthesize a first set of FO PI, FO tilt-integral-derivative (TID), and FO lead-lag controllers with Matlab R2021b and the Fractional-order Modeling and Control (FOMCON) toolbox, and these controllers significantly improved the control system performance of the DC-DC three-phase power electronic converter compared to classical PI controllers. The next set of proposed and synthesized controllers were based on SMC, together with its more general and flexible synergetic control variant, and both integer-order and FO controllers were developed. The proposed control structures are cascade control structures combining the SMC properties of robustness and control over nonlinear systems for the outer voltage control loop with the use of properly tuned synergetic controllers to obtain faster response time for the inner current control loop. To achieve superior performance, this type of cascade control also used a properly trained reinforcement learning-twin delayed deep deterministic policy gradient (RL-TD3) agent, which provides correction signals overlapping with the command signals of the current and voltage controllers. We present the Matlab/Simulink R2021b implementations of the synthesized controllers and the RL-TD3 agent, along with the results of numerical simulations performed for the comparison of the performance of the control structures.

The effect of information level of digital worker guidance systems on assembly performance, user experience and strain

Worker guidance systems provide product-specific and digital assembly information and can make an important contribution to increasing productivity and quality and relieving employees, especially in flexible variant assembly. However, a critical factor here is the successful design of the systems. While much of the research focuses on aspects of the hardware, this article is dedicated to the question of how the configuration of the level of information, i.e. the amount and depth of information in worker guidance systems, is affected. For this purpose, two studies, a laboratory study at the Technical University of Darmstadt (N = 53) and a field study at a company (N = 30), were conducted and two variants of the information level of a worker guidance system were compared. The study results show that assembly quality in particular is significantly influenced by the level of information; to the advantage of detailed information. No significant differences were found for production times. For the subjective evaluation of the system as well as for the objective and subjectively perceived strain, it can be stated that the high level of information also performs better. The results thus help to design worker guidance systems more precisely so that they fulfill the information needs of the employees as well as possible.

Engineering of a protein probe with multiple inputs and multiple outputs for evaluation of alpha synuclein aggregation states

The aggregation of α-synuclein (αS) into oligomers and fibrils is implicated in the pathology of Parkinson’s Disease (PD). While a molecular probe for rapid and comprehensive evaluation of αS aggregation states is critical for a better understanding of PD pathology, identification of therapeutic candidates, and the development of early diagnostic strategies, no such probe has yet to be developed. A structurally flexible αS variant, PG65, was previously developed as a target binding-driven, conformation-switching molecular probe for rapid αS oligomer detection. Though informative, detection using PG65 provides no comprehensive assessment of the αS aggregation states. In the present study, we report engineering of a molecular probe, PG65-MIMO (a PG65 variant with Multiple-Inputs and Multiple-Outputs), that rapidly (within 2 hr) produces comprehensive information on αS aggregation states. PG65-MIMO generates distinct fluorescence responses to the three major αS conformers (monomers, oligomers, and fibrils). PG65-MIMO also displays unique fluorescent signals for αS oligomers, depending on the tris(2-carboxyethyl)phosphine (TCEP) concentration. Our results suggest that the TCEP dependent signaling of PG65-MIMO may be associated with its conformational states. Overall, our study illustrates engineering of an αS variant to create a molecular probe for handling multiple inputs and multiple outputs, addressing the technological gap in αS detection.

Enhancement of accuracy in double sided incremental forming by compensating tool path for machine tool errors

Incremental Sheet Forming (ISF) is a dieless forming process and has capability to form complex 3D geometries. Double Sided Incremental Forming (DSIF) is the most flexible variant of ISF as it uses two tools, one on either side of the sheet (one tool forms the sheet and other acts as support). However, forming the components with acceptable accuracy is one of the challenges as it depends on many factors such as sheet and tool(s) deflections as well as machine tool errors. If these deflections/deviations are not compensated in tool path, accuracy of the component gets affected and support tool may lose contact or squeeze the sheet while forming components. In the present work, grid-based method is used to measure the machine tool error at discrete points of the work envelope. To enhance the accuracy and support tool contact condition, machine tool errors are compensated in tool path along with tool and sheet deflection compensations developed earlier. To demonstrate the effectiveness of developed methodology on the accuracy, components having different geometric characteristics as well as opening sizes are formed with and without compensating for machine tool errors in tool path. Results indicate that deviation in symmetry within the same contour reduced from 850 μm to less than 40 μm. In addition, support tool has maintained contact throughout support surface with maximum force of 50 N when machine tool errors are compensated.

Fixed point semantics for stream reasoning

Reasoning over streams of input data is an essential part of human intelligence. During the last decade stream reasoning has emerged as a research area within the AI-community with many potential applications. In fact, the increased availability of streaming data via services like Google and Facebook has raised the need for reasoning engines coping with data that changes at high rate. Recently, the rule-based formalism LARS for non-monotonic stream reasoning under the answer set semantics has been introduced. Syntactically, LARS programs are logic programs with negation incorporating operators for temporal reasoning, most notably window operators for selecting relevant time points. Unfortunately, by preselecting fixed intervals for the semantic evaluation of programs, the rigid semantics of LARS programs is not flexible enough to constructively cope with rapidly changing data dependencies. Moreover, we show that defining the answer set semantics of LARS in terms of FLP reducts leads to undesirable circular justifications similar to other ASP extensions. This paper fixes all of the aforementioned shortcomings of LARS. More precisely, we contribute to the foundations of stream reasoning by providing an operational fixed point semantics for a fully flexible variant of LARS and we show that our semantics is sound and constructive in the sense that answer sets are derivable bottom-up and free of circular justifications.

A circular economy life cycle costing model (CE-LCC) for building components

The building industry is responsible for the highest resource use, amount of waste and emissions of all industries. The principles of the Circular Economy (CE) could offer an approach to create a more sustainable built environment. For a transition towards a circular built environment, a comprehensive assessment method is needed to support the development of circular building products. As a step towards such a method, we developed an economic assessment in the form of a Circular Economy Life Cycle Cost (CE-LCC) model. It is based on existing Life Cycle Cost techniques and adapted to meet the requirements of CE products. The model is developed to (1) consider products as a composite of components and parts with different and multiple use cycles, (2) include processes that take place after the end of use, (3) provide practical and usable information to all stakeholders, and (4) facilitate alignment of the functional unit and system boundaries with LCA. To test the model, it has been applied to the case of the Circular Kitchen (CIK). Three variants of the CIK were compared to each other and the ‘business-as-usual’ case to determine which variant is the most economically competitive on the long term. The model indicates that the most flexible variant of the CIK has the lowest LCC outcome, even when considering multiple interest, lifespan and remanufacturing and recycling scenarios. Although, the model could benefit from further research and application, it can support the transition towards a more sustainable (building) industry.

Efficient variants of the CMRH method for solving a sequence of multi-shifted non-Hermitian linear systems simultaneously

Multi-shifted linear systems with non-Hermitian coefficient matrices arise in numerical solutions of time-dependent partial/fractional differential equations (PDEs/FDEs), in control theory, PageRank problems, and other research fields. We derive efficient variants of the restarted Changing Minimal Residual method based on the cost-effective Hessenberg procedure (CMRH) for this problem class. Then, we introduce a flexible variant of the algorithm that allows to use variable preconditioning at each iteration to further accelerate the convergence of shifted CMRH. We analyse the performance of the new class of methods in the numerical solution of PDEs and FDEs, also against other multi-shifted Krylov subspace methods.

A comparative investigation into the influence of the constitutive model on the prediction of in-plane formability for Nakazima and Marciniak tests

The present study investigates the use of both Marciniak and Nakazima tests to generate the forming limit curves (FLCs) for a power-law hardening dual phase steel, DP980, and an aluminum-magnesium alloy, AA5182, that exhibits dynamic hardening and saturation-type behavior. The recently proposed methodology of Min et al. (2016) to compensate for the so-called process effects of non-linear strain paths (NLSP) and contact pressure was evaluated and applied to the Nakazima FLCs to enable comparison with the Marciniak FLCs and the formability predictions of the Modified Maximum Force Criterion (MMFC) of Hora et al. (1994, 2013) for in-plane stretching. An emphasis was placed upon the experimental determination of the constitutive model to plastic strains in excess of 0.5 using tensile and shear tests. A flexible variant of the Hockett-Sherby (1975) hardening model for large strain levels along with constraints upon the hardening model calibration are proposed. It is demonstrated that the choice of hardening model, test data, and the calibration procedure can have a marked influence on the Nakazima process corrections and the predicted FLC using the MMFC model. An extension to the Linear Best Fit (LBF) limit strain detection methodology by Volk and Hora (2011) was developed to account for bend severity and the material hardening response and is compared with the ISO 12004-2 limit strains. If the hardening model is accurately calibrated to large strain levels, the analytical MMFC predictions were in excellent agreement with the process-corrected Nakazima and Marciniak FLCs for DP980. The results for the AA5182 were found to be strongly dependent upon the choice of hardening model that influences both the MMFC and the contact pressure correction.

Exploring individual and group differences in latent brain networks using cross-validated simultaneous component analysis

Component models such as PCA and ICA are often used to reduce neuroimaging data into a smaller number of components, which are thought to reflect latent brain networks. When data from multiple subjects are available, the components are typically estimated simultaneously (i.e., for all subjects combined) using either tensor ICA or group ICA. As we demonstrate in this paper, neither of these approaches is ideal if one hopes to find latent brain networks that cross-validate to new samples of data. Specifically, we note that the tensor ICA model is too rigid to capture real-world heterogeneity in the component time courses, whereas the group ICA approach is too flexible to uniquely identify latent brain networks. For multi-subject component analysis, we recommend comparing a hierarchy of simultaneous component analysis (SCA) models. Our proposed model hierarchy includes a flexible variant of the SCA framework (the Parafac2 model), which is able to both (i) model heterogeneity in the component time courses, and (ii) uniquely identify latent brain networks. Furthermore, we propose cross-validation methods to tune the relevant model parameters, which reduces the potential of over-fitting the observed data. Using simulated and real data examples, we demonstrate the benefits of the proposed approach for finding credible components that reveal interpretable individual and group differences in latent brain networks.

Human online adaptation to changes in prior probability.

Optimal sensory decision-making requires the combination of uncertain sensory signals with prior expectations. The effect of prior probability is often described as a shift in the decision criterion. Can observers track sudden changes in probability? To answer this question, we used a change-point detection paradigm that is frequently used to examine behavior in changing environments. In a pair of orientation-categorization tasks, we investigated the effects of changing probabilities on decision-making. In both tasks, category probability was updated using a sample-and-hold procedure: probability was held constant for a period of time before jumping to another probability state that was randomly selected from a predetermined set of probability states. We developed an ideal Bayesian change-point detection model in which the observer marginalizes over both the current run length (i.e., time since last change) and the current category probability. We compared this model to various alternative models that correspond to different strategies—from approximately Bayesian to simple heuristics—that the observers may have adopted to update their beliefs about probabilities. While a number of models provided decent fits to the data, model comparison favored a model in which probability is estimated following an exponential averaging model with a bias towards equal priors, consistent with a conservative bias, and a flexible variant of the Bayesian change-point detection model with incorrect beliefs. We interpret the former as a simpler, more biologically plausible explanation suggesting that the mechanism underlying change of decision criterion is a combination of on-line estimation of prior probability and a stable, long-term equal-probability prior, thus operating at two very different timescales.

Calixarene-mediated assembly of a small antifungal protein.

Synthetic macrocycles such as calixarenes and cucurbiturils are increasingly applied as mediators of protein assembly and crystallization. The macrocycle can facilitate assembly by providing a surface on which two or more proteins bind simultaneously. This work explores the capacity of the sulfonato-calix[n]arene (sclx n ) series to effect crystallization of PAF, a small, cationic antifungal protein. Co-crystallization with sclx4, sclx6 or sclx8 led to high-resolution crystal structures. In the absence of sclx n , diffraction-quality crystals of PAF were not obtained. Interestingly, all three sclx n were bound to a similar patch on PAF. The largest and most flexible variant, sclx8, yielded a dimer of PAF. Complex formation was evident in solution via NMR and ITC experiments, showing more pronounced effects with increasing macrocycle size. In agreement with the crystal structure, the ITC data suggested that sclx8 acts as a bidentate ligand. The contributions of calixarene size/conformation to protein recognition and assembly are discussed. Finally, it is suggested that the conserved binding site for anionic calixarenes implicates this region of PAF in membrane binding, which is a prerequisite for antifungal activity.

In Situ 13C NMR Spectroscopy Study of CO2/CH4 Mixture Adsorption by Metal-Organic Frameworks: Does Flexibility Influence Selectivity?

Metal-organic frameworks are promising candidates for selective separation processes such as CO2 removal from methane (natural gas sweetening). Framework flexibility, that is, the ability of a MOF lattice to change its structure as a function of parameters like pressure, temperature, and type of adsorbed molecules, is only observed for some special compounds. The main question of our present work is: does framework flexibility influence the adsorption selectivity? As a direct quantitative method to monitor the adsorption of both, carbon dioxide and methane, we make use of high-pressure in situ 13C NMR spectroscopy of 13CO2/13CH4 gas mixtures. This method allows to distinguish between the two gases as well as between adsorbed molecules and the interparticle gas phase. Gas mixture adsorption is studied under isothermal conditions. The selectivity factor for CO2 adsorption from CO2/CH4 mixtures is measured as a function of total gas pressure. The flexible material SNU-9 as well as the flexible and the nonflexible variant of DUT-8(Ni) are compared. Maximum selectivity factors for CO2 are observed for the flexible variant of DUT-8(Ni) in its open, large-pore state. In contrast, the rigid variant of DUT-8(Ni) and SNU-9 especially in its intermediate state exhibits lower adsorption selectivity factors. This observation indicates significant influence of the framework elasticity on the adsorption selectivity.

Double Sided Incremental Forming: Capabilities and Challenges

Incremental Sheet Forming (ISF) is gaining importance because of its flexibility to form customized/low volume sheet metal parts. Out of different variants of ISF, Double Sided Incremental Forming (DSIF) is most flexible variant which uses two tools, one to form the geometry and the other to provide local support. Complex geometries with features on both sides of sheet can be formed in single setup by changing the roles of forming and supporting tools based on the geometrical characteristics. This article presents the evolution of DSIF process, machines, methodologies and strategies to form complex geometries and their prediction. A multi-stage methodology is attempted to enhance the accuracy of large components is presented. Results show that the accuracy of a 250 mm × 250 mm size conical component is enhanced by 50% (maximum error reduced from 2.9 mm to 1.4 mm) using two-stage forming.

Nonparametric Bayesian label prediction on a graph

An implementation of a nonparametric Bayesian approach to solving binary classification problems on graphs is described. A hierarchical Bayesian approach with a randomly scaled Gaussian prior is considered. The prior uses the graph Laplacian to take into account the underlying geometry of the graph. A method based on a theoretically optimal prior and a more flexible variant using partial conjugacy are proposed. Two simulated data examples and two examples using real data are used in order to illustrate the proposed methods.

The Bhaduri–Marglin post-Kaleckian model in the history of distribution and growth theories: an assessment by means of model closures*

Starting from a review of the main strands of orthodox and heterodox distribution and growth models and their distinguishing features, with the post-Kaleckian Bhaduri and Marglin (1990) (and Kurz 1990) model as a specific but highly flexible variant of heterodox distribution and growth theories, we develop a simple modelling framework in which we can treat these different theories as different variants of model closure. In a simple closed private one-good economy model, each theory is presented drawing on the relationship between the rate of profit and the rate of growth, as well as on the consideration of one major adjusting variable allowing for the convergence of the endogenous variables of the model to their equilibrium values. This allows for a systematic comparison of exogenous and endogenous variables, of the ‘logic’ or the chain of causalities in each of the approaches, and of the generation of the long-run equilibrium positions of the system. It is finally shown that the post-Kaleckian model is able to cover many but not all of the results generated by the old neoclassical growth model, new neoclassical growth theories, classical–Marxian distribution and growth approaches, and post-Keynesian Kaldor–Robinson and Kalecki–Steindl distribution and growth theories.