Model Predictions Of Response Research Articles

Digital Twin Implementation for Three-Dimensional Rotordynamic Response via Physics-Informed LSTM Neural Networks

Abstract The rotating assemblies of critical machinery are complex dynamical systems and rotordynamic model response prediction inaccuracy risks machinery failure leading to high production losses. Jeffcott, Euler beam, and high-fidelity 3D solid finite element models are frequently utilized for rotordynamic analyses. Even though the 3D rotor has the higher accuracy, beam models are still widely used in industrial applications. To improve prediction accuracy of the lower-fidelity Jeffcott and beam models, a rotordynamics physics-informed neural network (R-PINN) is proposed. This models physics-informed long short-term memory (LSTM) neural networks that utilize partial or limited measured data, by incorporating physical laws. This approach enables the creation of a Digital Twin, which can produce additional data and help remove noise and outliers. In the current study, two R-PINNs are introduced to validate the superior capability of the model for both low- and high-fidelity physics. Random noise of 10% is introduced into the measured data produced by the Digital Twin to replicate real-world noisy measurements. The result shows that both low- and high-fidelity physics R-PINNs can achieve high accuracy even with high noise data, thereby increasing the robustness of the model. The results clearly demonstrate the ability of the proposed R-PINN algorithm to enhance an Euler beam model's predicted response to the level of accuracy of a 3D solid element model's predicted response, the latter acting as a surrogate for test measurements in an actual application.

Application of data-driven surrogate models for active human model response prediction and restraint system optimization

Surrogate models are a must-have in a scenario-based safety simulation framework to design optimally integrated safety systems for new mobility solutions. The objective of this study is the development of surrogate models for active human model responses under consideration of multiple sampling strategies. A Gaussian process regression is chosen for predicting injury values based on the collision scenario, the occupant's seating position after a pre-crash movement and selected restraint system parameters. The trained models are validated and assessed for each sampling method and the best-performing surrogate model is selected for restraint system parameter optimization.

A machine learning model to predict hepatocellular carcinoma response to transcatheter arterial chemoembolization.

Some patients with hepatocellular carcinoma (HCC) are more likely to experience disease progression despite transcatheter arterial chemoembolization (TACE) treatment, and thus would benefit from early switching to other therapeutic regimens. We sought to evaluate a fully automated machine learning algorithm that uses pre-therapeutic quantitative computed tomography (CT) image features and clinical factors to predict HCC response to TACE. Outcome information from 105 patients receiving first-line treatment with TACE was evaluated retrospectively. The primary clinical endpoint was time to progression (TTP) based on follow-up CT radiological criteria (mRECIST). A 14-week cutoff was used to classify patients as TACE-susceptible (TTP ≥14 weeks) or TACE-refractory (TTP <14 weeks). Response to TACE was predicted using a random forest classifier with the Barcelona Clinic Liver Cancer (BCLC) stage and quantitative image features as input as well as the BCLC stage alone as a control. The model's response prediction accuracy rate was 74.2% (95% CI=64%-82%) using a combination of the BCLC stage plus quantitative image features versus 62.9% (95% CI= 52%-72%) using the BCLC stage alone. Shape image features of the tumor and background liver were the dominant features correlated to the TTP as selected by the Boruta method and were used to predict the outcome. This preliminary study demonstrates that quantitative image features obtained prior to therapy can improve the accuracy of predicting response of HCC to TACE. This approach is likely to provide useful information for aiding HCC patient selection for TACE.

Nonlinear Finite Element Modeling of Reinforced Concrete Walls with Varying Aspect Ratios

ABSTRACT A relatively simple finite element modeling approach – named the Fixed-Strut-Angle Finite Element (FSAFE) Model – is presented in this paper for simulating the hysteretic lateral load behavior of reinforced concrete structural walls with varying levels of flexural and shear deformation contributions on wall response. The behavioral characteristics of the constitutive panel elements incorporated in the model formulation are based on a fixed-crack-angle modeling methodology that also incorporates simple behavioral models for shear–aggregate interlock in concrete and dowel action on reinforcing bars, constituting the shear stress transfer mechanisms across the cracks. Model response predictions are compared with experimentally measured responses of selected wall specimens with varying geometry and reinforcement characteristics; including relatively slender (aspect ratio of 3.0) flexure-controlled walls with rectangular and T-shaped cross-sections, medium-rise walls (aspect ratios of 1.5 and 2.0) with coupled shear-flexural responses, and squat walls (shear span-to-depth ratios of 0.44 and 1.0) with shear-dominant responses. The FSAFE model demonstrates a reasonable level of accuracy in predicting the nonlinear hysteretic response of the wall specimens investigated. Accurate predictions are obtained for the experimentally measured lateral load vs. displacement response characteristics of the walls, including their lateral load capacity, sftiffness, and ductility, as well as their hysteretic response attributes. The model also provides reasonably accurate estimates of the relative contributions of nonlinear flexural and shear deformations to wall lateral displacements, as well as local deformation characteristics including the spreading of nonlinear flexural and shear deformations along wall height, the distribution of transverse normal strains along wall height, and the longitudinal strain profiles across the wall cross-section. It is identified that consideration of instability related wall failures, including buckling of reinforcement and out-of-plane instability of the wall boundary region, is necessary for further improvement of model accuracy.

Micro-lens fabrication by a long-stroke precision stage with switching control based on model response prediction

This paper develops an automatic control switching mechanism for a long-stroke precision stage intended to improve micro-lens fabrication. The stage consists of a piezoelectric transducer (PZT) stage and a motor stage that achieves centimeter-length travel distances with nanometer precision. We obtained the PZT stage model by experiments and then applied robust loop-shaping techniques to design two controllers: one gave fast responses and the other gave smooth responses. The two controllers were further simplified as robust proportional-integral-derivative (PID) controllers for hardware implementation. We then developed a control switching mechanism, based on model response prediction, to combine the merits of these two controllers. For the motor stage, we derived the stage model by experiments and designed a gain-scheduling controller that regulated the stage to move at the fastest speed for long travel distances until it was within the stroke of the PZT stage. The two stages were then combined with a double-loop control structure to provide a travel range of 10 cm with a precision of 1.23 nm. We implemented the designed control mechanisms and structures for experimental verification. The combined stage was also integrated with a two-photon polymerization (TPP) system to fabricate micro-lenses. The optical properties of the micro-lenses were evaluated to demonstrate the effectiveness of the proposed control in precision positioning for long travel distances.

Fast and Reliable Modeling of Piezoelectric Transducers for Energy Harvesting Applications

This paper presents a fast and reliable model identification technique for piezoelectric transducers based on an equivalent electromechanical circuit easily implementable on SPICE-like simulation tools. Model parameter extraction is simple and requires just standard and inexpensive laboratory equipment. The equivalent circuits of four cantilevered Q220-A4-303YB transducers from Piezo Systems with different tip masses were identified and several simulations were performed with different electric loading conditions; measurement results agree with model response predictions in all the experiments. Indeed, the equivalent circuit representation permits the evaluation of the response of a real harvester system, where the electronic load is a synchronized switching converter which usually causes a significant feedback on the mechanical part of the system during energy extraction. Both simulation and measurements show that the damping effect is particularly important near resonance, where the adopted model is able to fit the experimental data and provides a more realistic description of the behavior of the system.

Response Surface Model Predictions of Emergence and Response to Pain in the Recovery Room

Sevoflurane-remifentanil interaction models that predict responsiveness and response to painful stimuli have been evaluated in patients undergoing elective surgery. Preliminary evaluations of model predictions were found to be consistent with observations in patients anesthetized with sevoflurane, remifentanil, and fentanyl. This study explored the feasibility of adapting the predictions of sevoflurane-remifentanil interaction models to an isoflurane-fentanyl anesthetic. We hypothesized that model predictions adapted for isoflurane and fentanyl are consistent with observed patient responses and are similar to the predictions observed in our previous work with sevoflurane-remifentanil/fentanyl anesthetics. Twenty-five patients scheduled for elective surgery received a fentanyl-isoflurane anesthetic. Model predictions of unresponsiveness were recorded at emergence, and predictions of a response to noxious stimulus were recorded when patients first required analgesics in the recovery room. Model predictions were compared with observations with graphical and temporal analyses. Results were also compared with our previous predictions after the administration of a sevoflurane-remifentanil/fentanyl anesthetic. Although patients were anesthetized, model predictions indicated a high likelihood that patients would be unresponsive (> or = 99%). After the termination of the anesthetic, model predictions of responsiveness well described the actual fraction of patients observed to be responsive during emergence. Half of the patients woke within 2 min of the 50% model-predicted probability of unresponsiveness; 70% woke within 4 min. Similarly, predictions of a response to a noxious stimulus were consistent with the number of patients who required fentanyl in the recovery room. Model predictions after the administration of an isoflurane-fentanyl anesthetic were similar to model predictions after a sevoflurane-remifentanil/fentanyl anesthetic. The results confirmed our study hypothesis; model predictions for unresponsiveness and no response to painful stimuli, adapted to isoflurane-fentanyl were consistent with observations. These results were similar to our previous study comparing model predictions and patient observations after a sevoflurane-remifentanil/fentanyl anesthetic.

Representation of the vowel /ε/ in normal and impaired auditory nerve fibers: Model predictions of responses in cats

The temporal response of auditory-nerve (AN) fibers to a steady-state vowel is investigated using a computational auditory-periphery model. The model predictions are validated against a wide range of physiological data for both normal and impaired fibers in cats. The model incorporates two parallel filter paths, component 1 (C1) and component 2 (C2), which correspond to the active and passive modes of basilar membrane vibration, respectively, in the cochlea. The outputs of the two filters are subsequently transduced by two separate functions, added together, and then low-pass filtered by the inner hair cell (IHC) membrane, which is followed by the IHC-AN synapse and discharge generator. The C1 response dominates at low and moderate levels and is responsible for synchrony capture and multiformant responses seen in the vowel responses. The C2 response dominates at high levels and contributes to the loss of synchrony capture observed in normal and impaired fibers. The interaction between C1 and C2 responses explains the behavior of AN fibers in the transition region, which is characterized by two important observations in the vowel responses: First, all components of the vowel undergo the C1/C2 transition simultaneously, and second, the responses to the nonformant components of the vowel become substantial.

Coupled Nonlinear Barge Motions, Part II: Stochastic Models and Stability Analysis

A computationally efficient quasi-two-degree-of-freedom (Q2DOF) stochastic model and a stability analysis of barges in random seas are presented in this paper. Based on the deterministic 2DOF coupled roll-heave model with high-degree polynomial approximation of restoring forces and moments developed in Part I, an attempt is made to further reduce the DOF of the model for efficient stochastic stability analysis by decoupling the heave effects on roll motion, resulting in a one-degree-of-freedom (1DOF) roll-only model. Using the Markov assumption, stochastic differential equations governing the evolution of probability densities of roll-heave and roll responses for the two low-DOF models are derived via the Fokker-Planck formulation. Numerical results of roll responses for the 2DOF and 1DOF models, using direct simulation in the time domain and the path integral solution technique in the probability domain, are compared to determine the effects of neglecting the influence of heave on roll motion and assess the relative computational efforts required. It is observed that the 1DOF model is computationally very efficient and the 2DOF model response predictions are quite accurate. However, the nonlinear roll-heave coupling is found to be significant and needs to be directly taken into account, rendering the 1DOF roll-only model inadequate for practical use. The 2DOF model is impractical for long-duration real-time response computation due to the insurmountable computational effort required. By taking advantage of the observed strong correlation between measured heave and wave elevation in the experimental results, an accurate and efficient Q2DOF model is developed by expressing the heave response in the 2DOF model as a function of wave elevation, thus reducing the effective DOF to unity. This Q2DOF model is essential as it reduces the computational effort by a factor of 10−5 compared to that of the 2DOF model, thus making practical stochastic analysis possible. A stochastic stability analysis of the barge under operational and survival sea states specified by the U.S. Navy is presented using the Q2DOF model based on first passage time formulation.

Treatment of Unidentifiability in Structural Model Updating

The present study addresses the issues of non-uniqueness and unidentifiability arising in structural model updating. A Bayesian probabilistic framework is used for model updating which properly handles the uncertainties due to model error and measurement noise associated with model updating. Uncertainties in the model parameters are quantified by probability density functions (PDF) specifying the relative plausibilities of the possible values of the parameters. The Bayesian formulation is well-suited for updating the PDF of the uncertain model parameters taking into account engineering experience and measured dynamic data. Methods are presented for approximating this updated PDF for the general unidentifiable case for which the region of significant probabilities is concentrated in the neighborhood of a manifold of lower dimension than the original parameter space. This PDF is useful for both model updating and structural damage predictions. Asymptotic approximations are also developed for computing the uncertainties in the model response predictions. It is demonstrated that unidentifiable cases are not treatable by existing results valid only for identifiable cases for which the dimension of the manifold is exactly zero. Two examples involving simulated model error and measurement noise are presented to demonstrate the advantages of the new proposed method in effectively addressing unidentifiability issues.

Modelling of reinforced concrete structural walls

The reinforced concrete structural wall is an important lateral load resisting element. It is increasingly used by designers in new structures as well the rehabilitation of existing ones. There are models that represent the flexural behaviour of walls to various degrees of accuracy. However, an efficient model is needed for accurate representation of the flexural and shear behaviour of these walls. A macro model that represents the behaviour of structural walls is developed. The model consists of nonlinear springs connected by linear beam elements. The defined models of these springs are capable of describing the hysteretic characteristics of reinforced concrete structures with sufficient accuracy. Moreover, the model is capable of idealizing both shear and flexural behaviours of reinforced concrete structural walls. The model response predictions are verified to be in close agreement with experimental measurements.

A Simulation-Based Approach For Designing Effective Field-Sampling Programs To Evaluate Contamination Risk Of Groundwater Supplies

An approach to model discrimination and network design for evaluation of groundwater contamination risk is proposed and demonstrated by application to a site in a glaciofluvial aquifer in Sweden. The approach consists of first hypothesizing alternative conceptual models of hydrogeology at the site on the basis of both quantitative data and qualitative information. The conceptual models are then expressed as two-dimensional numerical models of groundwater flow and solute transport, and model attributes controlling risk to the water supply are determined by simulation. Model predictions of response to a specific field test are made with each model that affects risk. Regions for effective measurement networks are then identified. Effective networks are those that capture sufficient information to determine which of the hypothesized models best describes the system with a minimum of measurement points. For the example site in Sweden, the network is designed such that important system parameters may be accurately estimated at the same time as model discrimination is carried out.

Ribcage compressibility in living subjects

The purpose of this study was to examine the responses of normal living subjects to the application of anteroposterior forces to the ribcage. Seventeen subjects aged between 25 and 37 years were tested during slow oscillatory loading while breath-holding at the end of a normal expiration. The mean stiffness coefficient was found to be 9.4 N mm −1 ( SD 2.9) and the mean gradient of the force-strain relation was 1888 N ( SD 646). Comparison with previously published cadaver data indicates that the embalmed cadaver ribcage stiffness is in the order of three times stiffer than living subjects, while fresh cadavers showed comparable stiffness to living subjects. A number of studies have used models to predict and understand the behaviour of the thoracic spine. Validation of the behaviour of models which include the thoracic spine and ribcage depends on comparison of model response predictions with observed responses of human subjects. The present study provides data on the anteroposterior compressibility of the ribcage of living subjects which may be suitable for use in model validation studies.