Model Input Parameters Research Articles

Comparison of classic time series and artificial intelligence models, various Holt-Winters hybrid models in predicting the monthly flow discharge in Marun dam reservoir

In this study, the data at Idenak hydrometric station were used to predict the inflow to Maroun Dam reservoir. For this purpose, different models such as artificial intelligence, Holt-Winters and hybrid models were used. Partial mutual information algorithm was used to determine the input parameters affecting modeling the monthly inflow by artificial intelligence models. According to the Hempel and Akaike information criterion, we introduced the monthly inflow with a 3-month lag, and the temperature with a 1-month lag, with respect to the lowest values of Akaike and the highest values of Hempel as input parameters of artificial intelligence models. The results showed the weak performance of the Holt-Winters model compared to other models and confirmed the superiority of the Holt-adaptive network-based fuzzy inference system (ANFIS) hybrid model with the root-mean-square error of 54 and the coefficient of determination (R2) of 0.83 in the testing process compared to other mentioned models. In addition, the above hybrid models performed better than other models in the test process.

Global Sensitivity Analysis of a Dam Breaching Model: To Which Extent Is Parameter Sensitivity Case‐Dependent?

AbstractFailure of dams and dikes often leads to devastating consequences in protected areas. Numerical models are instrumental tools to assess flood risk and guide emergency management, but numerous uncertainties affect model outcomes. Identifying uncertain model input parameters that induce high uncertainties in model outputs is essential. This paper focuses on two model outputs: the maximum breach discharge and the time to reach this peak. Using our implementation of a simplified physically based dam breaching model developed by Wu (2016), a global sensitivity analysis was conducted based on Sobol indices of total order. Unlike in most previous studies, many different configurations (twenty‐seven), both at laboratory and field scales, were considered. For each of them, input variables were ranked according to the significance of the contribution of their uncertainty to the output variability, and the dependency between reference configurations and sensitivity analysis results was highlighted. Depending on the considered case study, input parameter uncertainties with the largest impact on output variability were different, and so was the magnitude of output uncertainties. We demonstrate that sensitivity analysis results obtained for a specific configuration cannot be transferred as it to other configurations. Finally, sensitivity and uncertainty quantification results were combined in a decision tree to determine which input parameter uncertainty is the most critical in each configuration and what standard deviation in the output variables is expected. The global sensitivity analysis procedure presented here may apply to a wide variety of models in environmental sciences.

Development of a NOx Calculation Model for Low-Speed Marine Diesel Engines Based on Soft Measurement Technology

With the increasing level of intelligence of marine engines, there is an increasing demand for the online monitoring of engines, and marine NOx emissions have been of great concern. In this paper, a NOx simulation model is developed based on virtual measurement technology, which can calculate and predict NOx emissions based on the current operating state parameters of low-speed two-stroke diesel engines. First, the calibrated 3D simulation model is used to design the experiments to obtain the simulation experimental samples. Based on the NOx generation mechanism and diesel engine work-related parameters, the relevant factors were selected as alternative input parameters for the NOx emission model. The correlation analysis was then performed on the input parameters using the grey relational analysis correlation method and the Pearson correlation coefficient, and the principal component analysis method was used to reduce the dimensionality of the relevant factors by minimizing the loss of important information in reducing the complexity of the whole model. Then, the structure-related parameters of the backpropagation neural network (BPNN) were adaptively optimized using the group method of data handling (GMDH) to improve the accuracy of the NOx soft measurement model. Finally, the developed GMDH–BP model was validated with data and compared with the error evaluation index of BPNN and BPNN optimized by genetic algorithm (GA), and the developed NOx simulation model demonstrated high prediction accuracy under the same hyperparameter settings. The result provides technical support for the subsequent realization of the real-time online monitoring of NOx emissions from low-speed marine diesel engines without NOx sensors.

Compressive strength prediction of hydrothermally solidified clay with different machine learning techniques

Hydrothermal solidification technology can transform waste clay into high-strength building materials with lower energy consumption. Few machine learning methods have been applied to strength prediction of hydrothermally solidified materials, whose mechanical properties differ from those of conventional cement-based materials by virtue of the production method. In this study, six machine learning methods, random forest (RF), gated recurrent unit (GRU), k-nearest neighbors (KNN), back-propagation neuron network (BPNN), extreme gradient boosting (XGBoost) and gaussian process regression (GPR), were used to develop the compressive strength prediction models of hydrothermally solidified clay. These methods can efficiently predict the compressive strength without time-consuming and costly tests and guide the optimization of the mix proportions, thus facilitating the application of hydrothermally solidified clay. For this purpose, 140 experimental data obtained from hydrothermally solidified clay with different Ca(OH)2 content, dry density, moisture content, curing time and curing temperature were used to proposed the prediction models. In the modeling process, these five effective variable parameters were considered as the modeling input parameters. The performance of these models was compared by correlation coefficient (R2), mean absolute error (MAE) and root mean square error (RMSE). The validity of these models was confirmed by the K-fold cross-validation method, and the influence weight of each input parameter on the prediction results was characterized by sensitivity analysis. The results demonstrated that the GPR model provided the best prediction with an R2 value of 0.989. In the sensitivity analysis, curing time had the greatest influence on the predicted results, followed by dry density, Ca(OH)2 content, curing temperature and water content. Comprehensively considering the cost, energy consumption and compressive strength, the applicable mix proportion of hydrothermally solidified clay includes Ca(OH)2 content (not less than 16.64%), moisture content (10%), curing temperature (not less than 87.35 °C), curing time (18 h) and dry density (1.9 g/cm3).

Machine learning implemented exploration of the adsorption mechanism of carbon dioxide onto porous carbons

Adsorption of CO2 on porous carbons has been identified as one of the promising methods for carbon capture, which is essential for meeting the sustainable developmental goal (SDG) with respect to climate action, i.e., SDG 13. This research implemented six supervised machine learning (ML) models (gradient boosting decision tree (GBDT), extreme gradient boosting (XGB), light boost gradient machine (LBGM), random forest (RF), categorical boosting (Catboost), and adaptive boosting (Adaboost)) to understand and predict the CO2 adsorption mechanism and adsorption uptake, respectively. The results recommended that the GBDT outperformed the remaining five ML models for CO2 adsorption. However, XGB, LBGM, RF, and Catboost also represented the prediction in the acceptable range. The GBDT model indicated the accurate prediction of CO2 uptake onto the porous carbons considering adsorbent properties and adsorption conditions as model input parameters. Next, two-factor partial dependence plots revealed a lucid explanation of how the combinations of two input features affect the model prediction. Furthermore, SHapley Additive exPlainations (SHAP), a novel explication approach based on ML models, were employed to understand and elucidate the CO2 adsorption and model prediction. The SHAP explanations, implemented on the GBDT model, revealed the rigorous relationships among the input features and output variables based on the GBDT prediction. Additionally, SHAP provided clear-cut feature importance analysis and individual feature impact on the prediction. SHAP also explained two instances of CO2 adsorption. Along with the data-driven insightful explanation of CO2 adsorption onto porous carbons, this study also provides a promising method to predict the clear-cut performance of porous carbons for CO2 adsorption without performing any experiments and open new avenues for researchers to implement this study in the field of adsorption because a lot of data is being generated.

Stochastic Framework for Optimal Control of Planetary Reentry Trajectories Under Multilevel Uncertainties

We present a novel stochastic optimal control framework that accounts for various types of uncertainties, with application to reentry trajectory planning. The formulation of the optimal trajectory control problem is presented in the context of an indirect method where a functional objective associated with the terminal vehicle speed is to be minimized. Uncertain input parameters in the optimal trajectory control model, including aerodynamic parameters and initial and terminal conditions, are modeled as aleatory random variables, while the statistical parameters of these aleatory distributions are themselves random variables. The parametric and model uncertainties are simultaneously propagated through an extended polynomial chaos expansion (EPCE) formalism. Several metrics are described to evaluate response statistics and presented as insightful tools for robust decision making. Specifically, the response probability density function (PDF) reflecting influence of both epistemic and aleatory uncertainties is obtained. By sampling over the random variables representing model error, an ensemble of response PDFs is generated and the associated failure probability is estimated as a random variable with its own polynomial chaos expansion. Besides, the sensitivity index functions of response PDF with respect to the statistical parameters are evaluated. Coupling parametric and model uncertainties within the EPCE framework leads to a robust and efficient paradigm for multilevel uncertainty propagation and PDF characterization in general optimal control problems.

Elman neural network-based temperature prediction and optimization for lithium-ion batteries

Reliable and precise temperature prediction is one of the most crucial challenges for improving battery performance and preventing thermal runaway. This paper uses a highly adaptive Elman neural network (Elman-NN) to construct a temperature prediction model for lithium-ion batteries in a metal foam aluminum thermal management system. Numerical modeling methods obtain experimental data sets for model training and testing. The input parameters of the neural network prediction model are ambient temperature, battery discharge rate, cooling air flow rate, and state of charge; the output parameters are the maximum, minimum, and average battery temperature. However, due to the limitations of the gradient descent algorithm, the training process of the Elman neural network tends to fall into local optimum solutions. To further improve the prediction accuracy, the Elman-NN structure was optimized using the PSO algorithm, and the model performance was tested and validated. Compared with the original Elman-NN, the hybrid PSO-Elman-NN has smaller MSE and MAE values, with a maximum reduction of 43% and 25%, respectively. For the three test conditions, the maximum predicted temperature difference does not exceed 1.5 K, and the temperature difference decreases further as the discharge time increases. Moreover, the hybrid model’s prediction accuracy is significantly improved, with the coefficients of determination ( R2) increasing by 1.736%, 0.706%, and 1.851%, respectively. The PSO-Elman-NN performed well in terms of compatibility and accuracy of the battery temperature prediction.

Research on Application of Neural Network Model in Location Privacy Protection of Wireless Sensor Networks

Due to the characteristics of WSN (Wireless sensor network), attackers can easily eavesdrop on data packets transmitted between single or even multiple communication links, and separate sensitive data from multiple sensor information, which poses a great threat to location privacy. Therefore, it is very important to effectively protect the privacy of training sample data while using WSN. Neural network is an important research hotspot in AI field, and it is a model close to biological neural network in machine learning algorithm. In this paper, an application of neural network model in WSN location privacy protection is given. A CNN (Convolutional Neural Network) location privacy protection prediction protocol based solely on additive homomorphisms is proposed, which can effectively ensure that input features, model parameters, and intermediate values are not leaked during the prediction process. The experimental results show that the proposed method has good robustness and can effectively protect the private location of the source node.

Prediction Models of ≥2 MeV Electron Daily Fluences for 3 Days at GEO Orbit Using a Long Short-Term Memory Network

Geostationary satellites are exposed to harsh space weather conditions, including ≥2 MeV electrons from the Earth’s radiation belts. To predict ≥2 MeV electron daily fluences at 75°W and 135°W at geostationary orbit for the following three days, long short-term memory (LSTM) network models have been developed using various parameter combinations. Based on the prediction efficiency (PE) values, the most suitable time step of inputs and best combinations of two or three input parameters of models for predictions are recommended. The highest PE values for the following three days with three input parameters were 0.801, 0.658 and 0.523 for 75°W from 1995 to August 2010, and 0.819, 0.643 and 0.508 for 135°W from 1999 to 2010. Based on yearly PE values, the performances of the above models show the solar cycle dependence. The yearly PE values are significantly inversely correlated with the sunspot number, and they vary from 0.606 to 0.859 in predicting the following day at 75°W from 1995 to 2010. We have proven that the poor yearly PE is related to relativistic electron enhancement events, and the first day of events is the most difficult to predict. Compared with previous models, our models are comparable to the top performances of previous models for the first day, and significantly improve the performance for second and third days.

A textile architecture-based discrete modeling approach to simulating fabric draping processes

Fabric draping, which is referred to as the process of forming of textile reinforcements over a 3D mold, is a critical stage in composites manufacturing since it determines the fiber orientation that affects subsequent infusion and curing processes and the resulting structural performance. The goal of this study is to predict the fabric deformation during the draping process and develop in-depth understanding of fabric deformation through an architecture-based discrete Finite Element Analysis (FEA). A new, efficient discrete fabric modeling approach is proposed by representing textile architecture using virtual fiber tows modeled as Timoshenko beams and connected by the springs and dashpots at the intersections of the interlaced tows. Both picture frame and cantilever beam bending tests were carried out to characterize input model parameters. The predictive capability of the proposed modeling approach is demonstrated by predicting the deformation and shear angles of a fabric subject to hemisphere draping. Key deformation modes, including bending and shearing, are successfully captured using the proposed model. The development of the virtual fiber tow model provides an efficient method to illustrate individual tow deformation during draping while achieving computational efficiency in large-scale fabric draping simulations. Discrete fabric architecture and the inter-tow interactions are considered in the proposed model, promoting a deep understanding of fiber tow deformation modes and their contribution to the overall fabric deformation responses.

Resonant ultrasound spectroscopy measurement and modeling of additively manufactured octet truss lattice cubes

In this work, the elastic characteristics of additively manufactured Ti-5553 octet truss lattice cubes were examined using resonant ultrasound spectroscopy (RUS) techniques and correlated to X-ray computed tomography measurements of strut continuity. RUS models and measurements identified an outlier lattice possessing lower effective modulus despite larger effective density. Experimental measurements of density and effective Young's modulus were treated as input parameters in a homogenous anisotropic continuum model. The continuum model exhibited good agreement with experimental RUS measurements, thus confirming the potential for a simplified approximation of the octet truss lattice.

A finite element modeling‐based approach to predict vibrations transmitted through different body segments of the operator within the workspace of a small tractor

AbstractAgricultural tractor drivers are subjected to high levels of whole‐body vibrations and hand arm vibrations during most part of the farm activities due to unevenness of field surface, uneasy posture, improper workplace design, moving parts of the tractor, and other unavoidable circumstances. The comfort level of the operator inside a dynamic tractor is dependent on the level of vibration generated inside the different human body segments. In the present study, a finite element modeling was proposed to predict vertical vibrations (Z‐axis) and frequencies at the different body segments of the seated small tractor operator. The forces required for different controls of the tractor were measured to be used as input parameters in the finite element modeling. The maximum mean forces of the brake (172.8 N) and clutch (153.2 N) were used as the input parameters for the simulation study. The simulated results were validated with the field measured values of vertical accelerations at selected body segments of the operator. The simulation could successfully predict vertical vibrations at selected points of interest (i.e., foot, leg, thigh, lower arm, upper arm, back, and head) except the chest of the body, as the buttock of the operator model was fixed (degree of freedom is equal to zero) in the simulation. The obtained results were compared with the international standards ISO 2631‐1 (1985/1997) and ISO 5349‐1 (2001) to assess the vibration characteristics at the different body segments of the operator. The foot, leg, lower arm, and upper arm of the operator were subjected to vertical vibration frequencies from 10 to 200 Hz. Most of the resonance of vertical accelerations occurred in one‐third octave bands of 20–80 Hz frequencies. The thigh, chest, back, and head of the operator were exposed to vibration frequencies below 40 Hz during field operation. At these parts of the body, the vertical acceleration resonated at lower frequencies, between 2 and 8 Hz.

Surrogate modelling of railway pantograph-catenary interaction using deep Long-Short-Term-Memory neural networks

The interaction performance of the pantograph-catenary is of great importance as it directly determines the current collection quality and operational safety of trains. The finite element method (FEM) is dominantly used for simulating pantograph-catenary interaction, which is normally computationally heavy. In this work, addressing the tremendous computational cost of FEM models, a surrogate model for fast simulations of pantograph-catenary interaction is proposed using deep learning. A dataset containing 30,000 cases of pantograph-catenary interaction is generated by a validated FEM model. A Long-Short-Term-Memory (LSTM) neural network is proposed to learn the inherent nonlinearity between the input model parameters and the output pantograph-catenary contact force from data. The resulting prediction performance indicates that contact forces predicted by the surrogate model are consistent with those simulated by FEM, while the computational efforts of the surrogate model are negligible compared with FEM. Prediction performances using different network architectures and configurations are compared to determine the optimal setting for a pantograph-catenary system. The LSTM-based surrogate model shows high efficiency for simulating pantograph-catenary interactions and promising practicability in optimising catenary structural parameters for design or upgrade.

A Comparative Study of Immediate Annuities and Ordinary Annuities in establishing the Phantom of Zero Liability Under the Trusteeship Pension Valuation Structure

Background: The liabilities of a pension scheme define the financial value to be paid at a definite period in the future. The underlying goal of pension plans is to provide retirees with sufficient stream of income to enable them to live a decent financially independent life post-employment period. The regulatory framework for occupational pension schemes necessitates the services of trustees as administrators who assume legal administrative responsibilities on the scheme and saddled to oversee actuarial valuations of the scheme's liabilities at definite points in time. Objectives: The objective of this paper is (i) to empirically examine the drivers of pension liability and how they are evaluated by the trustee’s model. (ii) Specifically, the study intends to use input parameters of the trustee model to establish the conditions for which the value of liability is zero under trusteeship annuity factor. Methods: This study applies trustees’ valuation model, the present values together with infinitesimal calculus. Salary data as well as demographic data were obtained from an agricultural production services company located in Jos-South, Nigeria. Results: Computational evidence from our results proves that the total service liability under the conditions of the current model is vanishingly zero. However, when the annuity factor is replaced by life table annuity, the service liability does not vanish. Conclusion: The total service liability obtained as zero therefore initiates inquiry as to whether this current valuation framework causes potential uncertainties for the pension trustees who are responsibly saddled with both administration and core decision-making responsibilities of the system.

Calibrating a Physics-Based Corrosion Model with Field-Based Bridge Condition Data

This paper presents a novel method for calibrating a physics-based corrosion prediction model using a deterioration curve derived from field-based bridge condition assessments. The corrosion model used in this study considers inputs such as corrosion rate to predict outputs such as concrete crack width. Deterioration curves that present bridge element condition versus time are converted into crack width versus time using inspection guidance from transportation agencies. Input parameters of the corrosion model are calibrated by matching the crack widths predicted by the corrosion model with those predicted by the deterioration curves. The calibration method is demonstrated for a reinforced concrete bridge column in this article. The corrosion rate before cracking, the pitting factor, and the ratio of corrosion rates at critical crack width and before cracking significantly influenced the model results and were calibrated. Postcalibration input parameter ranges were narrower than but within the ranges reported in the literature. A physics-based corrosion model thus calibrated using field inspection data can potentially improve the predictions of future structural health and aid in asset management.

Surface Mounted Production Line Capacity Prediction Based on Data Driven

Capacity prediction of surface mount production lines is important for improving productivity, dynamic production scheduling, and optimizing production structure. An integrated mutual information-cuckoo search-extreme learning machine (MI-CS-ELM) prediction model is proposed for the impact of nonlinear and high-dimensional features of surface-mounted production lines on capacity. Firstly, the features affecting the production capacity are ranked based on the mutual information method, and the predictive input data set is reconstructed. Secondly, the input parameters of the prediction model are adjusted by the cuckoo search algorithm and the capacity prediction model is built based on an extreme limit learning machine. Finally, the rationality of the proposed method is verified by an example to guide capacity planning and production scheduling optimization.

Predictive Evaluation of Pavement Deterioration in Ibadan, Oyo State, Nigeria

This work presents a mathematical model for road pavement deterioration condition in Ibadan. The effective deformation model shows the response of the pavement to the axial loads with respect the age, t of the pavement. The correction factor in the model compensates for the other secondary factors that may influence pavement deterioration. Throughout this study, the most common factors influencing pavement distress and the behaviour of distress have been identified and used as input parameters in a predictive mathematical model. The model shows that the pavement will deform along horizontal plane and also deform vertically. The increase in the volume of the failing section implies that there will be increase in the international roughness index (IRI) of the pavement. Also, the results show that the higher the value of the load on a pavement, the higher the IRI increased becomes.

A new base of wind turbine noise measurement data and its application for a systematic validation of sound propagation models

Abstract. Extensive measurements in the area of wind turbines were performed in order to validate a sound propagation model which is based on the Crank–Nicolson parabolic equation method. The measurements were carried out over a flat grass-covered landscape and under various environmental conditions. During the measurements, meteorological and wind turbine performance data were acquired and acoustical data sets were recorded at distances of 178, 535 and 845 m from the wind turbine. By processing and analysing the measurement data, validation cases and input parameters for the sound propagation model were derived. The validation includes five groups that are characterised by different sound propagation directions, i.e. downwind, crosswind and upwind conditions in varying strength. In strong upwind situations, the sound pressure levels at larger distances are overestimated because turbulence is not considered in the modelling. In the other directions, the model reproduces the measured sound propagation losses well in the overall sound pressure level and in the third octave band spectra. As in the recorded measurements, frequency-dependent maxima and minima are identified, and losses generally increase with increasing distance and frequency. The agreement between measured and modelled sound propagation losses decreases with distance. The data sets used in the validation are freely accessible for further research.

A methodology for data-driven modeling and prediction of the drag losses of wet clutches

AbstractIn wet clutches, load-independent drag losses occur in the disengaged state and under differential speed due to fluid shearing. The drag torque of a wet clutch can be determined accurately and reliably by means of costly and time-consuming measurements. As an alternative, the drag losses can already be precisely calculated in the early development phase using computing-intensive CFD models. In contrast, simple analytical calculation models allow a rough but non-time-consuming estimation. Therefore, the aim of this study was to develop a methodology that can be used to build a data-driven model for the prediction of the drag losses of wet clutches with low computational effort and, at the same time, sufficient accuracy under consideration of a high number of influencing parameters. For building the model, we use supervised machine learning algorithms. The methodology covers all relevant steps, from data generation to the validated prediction model as well as its usage. The methodology comprises six main steps. In Step 1, the data is generated on a suitable test rig. In Step 2, characteristic values of each measurement are evaluated to quantify the drag loss behavior. The characteristic values serve as target values to train the model. In Step 3, the structure and quality of the dataset are analyzed and, subsequently, the model input parameters are defined. In Step 4, the relationships between the investigated influencing parameters (model input) and the characteristic values (model output) are determined. Symbolic regression and Gaussian process regression have both been proven to be suitable for this task. Lastly, the model is used in Step 5 to predict the characteristic values. Based on the predictions, the drag torque can be predicted as a function of differential speed in Step 6, using an approximation function. The model allows a user-oriented prediction of the drag torque even for a high number of parameters with low computational effort and sufficient accuracy at the same time.

Stability of asymmetric cell division: A deformable cell model of cytokinesis applied to C. elegans

Cell division during early embryogenesis is linked to key morphogenic events such as embryo symmetry breaking and tissue patterning. It is thought that the physical surrounding of cells together with cell intrinsic cues act as a mechanical “mold,” guiding cell division to ensure these events are robust. To quantify how cell division is affected by the mechanical and geometrical environment, we present a novel computational mechanical model of cytokinesis, the final phase of cell division. Simulations with the model reproduced experimentally observed furrow dynamics and describe the volume ratio of daughter cells in asymmetric cell divisions, based on the position and orientation of the mitotic spindle. For dividing cells in geometrically confined environments, we show how the orientation of confinement relative to the division axis modulates the volume ratio in asymmetric cell division. Further, we quantified how cortex viscosity and surface tension determine the shape of a dividing cell and govern bubble-instabilities in asymmetric cell division. Finally, we simulated the formation of the three body axes via sequential (a)symmetric divisions up until the six-cell stage of early C. elegans development, which proceeds within the confines of an eggshell. We demonstrate how model input parameters spindle position and orientation provide sufficient information to reliably predict the volume ratio of daughter cells during the cleavage phase of development. However, for egg geometries perturbed by compression, the model predicts that a change in confinement alone is insufficient to explain experimentally observed differences in cell volume. This points to an effect of the compression on the spindle positioning mechanism. Additionally, the model predicts that confinement stabilizes asymmetric cell divisions against bubble-instabilities.