Artificial intelligence-based surrogate modeling for computational cost-effective optimization of hydrogen liquefaction process

Learning What Makes Catalysts Good

The heating, ventilation, and air conditioning (HVAC) system serving the test room of the SENS i-Lab of the Department of Architecture and Industrial Design of the University of Campania Luigi Vanvitelli (Aversa, south of Italy) has been experimentally investigated through a series of tests performed during both summer and winter under both normal and faulty scenarios. In particular, five distinct typical faults have been artificially implemented in the HVAC system and analyzed during transient and steady-state operation. An optimal artificial neural network-based system model has been created in the MATLAB platform and verified by contrasting the experimental data with the predictions of twenty-two different neural network architectures. The selected artificial neural network architecture has been coupled with a dynamic simulation model developed by using the TRaNsient SYStems (TRNSYS) software platform with the main aims of (i) making available an experimental dataset characterized by labeled normal and faulty data covering a wide range of operating and climatic conditions; (ii) providing an accurate simulation tool able to generate operation data for assisting further research in fault detection and diagnosis of HVAC units; and (iii) evaluating the impact of selected faults on occupant indoor thermo-hygrometric comfort, temporal trends of key operating system parameters, and electric energy consumptions.

In the present study, deep learning neural network model has been employed in many engineering problems including heat transfer prediction. The main consideration of this document is to predict the performance of the boiling heat transfer in helical coils under terrestrial gravity conditions and compare with actual experimental data. Total of 877 data sample has been used in the present neural model. Artificial new Neural Network (ANN) model developed in Python environment with Multi-layer Perceptron (MLP) using four parameters (helical coils dimensions, mass flow rate, heating power, inlet temperature) and one parameter (outlet temperature) has been used in the input layer and output layer in order. Levenberg-Marquardt (LM) algorithm using L2 Regularization to find out the optimal model. A typical feed-forward neural network model composed of three layers, with 30 numbers of neurons in each hidden layer, has been found as optimal based on statistical error analysis. The 4-30-30-1 neural model predicts the characteristics of the helical coil with the accuracy of 98.16 percent in the training stage and 96.68 percent in the testing stage. The result indicated that the proposed ANN model successfully predicts the heat transfer performance in helical coils and can be applied for others operation concerned with heat transfer prediction for future works

An approximation of orbit rendezvous is usually used in the global optimization of multi-target rendezvous missions, which can greatly affect the efficiency of optimization process. A fast neural network-based surrogate model is proposed to approximate the optimal velocity increment of perturbed orbit rendezvous in low Earth orbits. According to a dynamic analysis, the initial and target orbits together with the flight time are transformed into a nine-dimensional normalized vector that is used as the input layer of the neural network. An existing approximation method is introduced to quickly generate the training data. In simulations, different numbers of layer nodes and hidden layers are tested to choose the best parameters. The proposed neural network model demonstrates high precision and high efficiency compared with previous approximation methods and neural network models. The mean relative error is less than 1%. Finally, a case of an optimization of a multi-target rendezvous mission is tested to prove the potential application of the neural network model.

An artificial neural network-based model for roping prediction in aluminum alloy sheet

Machine learning (ML) models provide more choices to patients with diabetes mellitus (DM) to more properly manage blood glucose (BG) levels. However, because of numerous types of ML algorithms, choosing an appropriate model is vitally important. In a systematic review and network meta-analysis, this study aimed to comprehensively assess the performance of ML models in predicting BG levels. In addition, we assessed ML models used to detect and predict adverse BG (hypoglycemia) events by calculating pooled estimates of sensitivity and specificity. PubMed, Embase, Web of Science, and Institute of Electrical and Electronics Engineers Explore databases were systematically searched for studies on predicting BG levels and predicting or detecting adverse BG events using ML models, from inception to November 2022. Studies that assessed the performance of different ML models in predicting or detecting BG levels or adverse BG events of patients with DM were included. Studies with no derivation or performance metrics of ML models were excluded. The Quality Assessment of Diagnostic Accuracy Studies tool was applied to assess the quality of included studies. Primary outcomes were the relative ranking of ML models for predicting BG levels in different prediction horizons (PHs) and pooled estimates of the sensitivity and specificity of ML models in detecting or predicting adverse BG events. In total, 46 eligible studies were included for meta-analysis. Regarding ML models for predicting BG levels, the means of the absolute root mean square error (RMSE) in a PH of 15, 30, 45, and 60 minutes were 18.88 (SD 19.71), 21.40 (SD 12.56), 21.27 (SD 5.17), and 30.01 (SD 7.23) mg/dL, respectively. The neural network model (NNM) showed the highest relative performance in different PHs. Furthermore, the pooled estimates of the positive likelihood ratio and the negative likelihood ratio of ML models were 8.3 (95% CI 5.7-12.0) and 0.31 (95% CI 0.22-0.44), respectively, for predicting hypoglycemia and 2.4 (95% CI 1.6-3.7) and 0.37 (95% CI 0.29-0.46), respectively, for detecting hypoglycemia. Statistically significant high heterogeneity was detected in all subgroups, with different sources of heterogeneity. For predicting precise BG levels, the RMSE increases with a rise in the PH, and the NNM shows the highest relative performance among all the ML models. Meanwhile, current ML models have sufficient ability to predict adverse BG events, while their ability to detect adverse BG events needs to be enhanced. PROSPERO CRD42022375250; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=375250.

Traditional BP neural networks frequently encounter local optima during carbon price forecasts. This study adopts a hybrid approach, combining a genetic algorithm and particle swarm optimization (GA-PSO) to improve the BP neural network, resulting in the creation of a GA-PSO-BP neural network model. Seven critical factors were identified affecting carbon prices, and we utilized data on carbon emission trading prices from China for the analysis. Compared to traditional BP neural network models, including GA-BP neural network models optimized solely with genetic algorithms and PSO-BP neural network models enhanced through particle swarm optimization, the findings reveal that the GA-PSO-BP neural network model demonstrates superior performance in terms of precision and robustness. Furthermore, it demonstrates advancements across various error evaluation metrics, thus delivering more accurate forecasts. Offering precise carbon price predictions, the enhanced GA-PSO-BP neural network model proves to be a valuable tool for analyzing the market and making decisions in the carbon pricing sector.

An inverse design framework incorporating a physics-based surrogate model and global optimization is proposed to assist in the design of low warpage ultra-thin packages by adjusting the metal densities over substrate subsections and layers. The surrogate model is derived from two finite element analysis (FEA) models. The first one describes the relationship between the metal density in the substrate layer to the coefficient of thermal expansion (CTE) while the second one describes the relationship between in-plane CTE variation of the substrate to the warpage profile. Results from these two FEA models are used to train separate artificial neural networks (ANN). When these ANNs are run sequentially, the surrogate model can accurately determine the warpage profile for any set of metal densities. Three global optimization algorithms, Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Cross Entropy (CE) were then evaluated using this surrogate model. Three case studies consisting of different warpage profiles (original and 20% reduced warpage) and constraints to the optimization search space (±20% or ±50% change to metal density) were then evaluated using these algorithms. For all three cases, the three algorithms converged to similar solutions, indicating that indeed the global minimum has been attained and determined. However, GA took a significantly longer time to converge than PSO and CE. Based on these results, PSO and CE are recommended to be suitable algorithms to carry out inverse design for this type of problem.

Predicting uniaxial compressive strength from drilling variables aided by hybrid machine learning

This paper proposes the use of Artificial Neural Networks (ANNs), specifically Physics-Informed Neural Networks (PINNs), for dynamic surrogate modelling of wind turbines. PINNs offer the flexibility to model complex relationships while incorporating physics-based constraints, enabling accurate representation of wind turbine dynamics. In this paper, a PINN-based surrogate model is developed for the Blade Element Momentum (BEM) aerodynamic model used in state-of-the-art numerical wind turbine simulations. The PINN model replaces the time-consuming root-finding process in BEM with high-dimensional regression, significantly improving computational efficiency. The PINN model is trained using data generated from a numerical model of the IEA-15MW reference wind turbine, and its performance is compared against conventional data-driven Neural Network (DDNN) models. The proposed surrogate model provides more efficient and accurate evaluations of wind turbine responses compared with traditional surrogate modelling approaches. A significant computational advantage is obtained by using the developed surrogate models with a forty-fold speedup demonstrated compared to the BEM model. Replacing the BEM model with the PINN-based surrogate model for load computation in the numerical model used for dynamic analysis results in an overall reduction of 35% in computational time for a complete dynamic simulation. This is a substantial improvement in efficiency without sacrificing accuracy — the maximum Mean Absolute Error (MAE) values for the surrogate models are of the order of 10−2, which shows that the surrogate models can predict the angle of attack at any blade node with a discrepancy of less than 0.5°. The surrogate models significantly reduce computational time while maintaining high accuracy, making them a promising approach for simulating wind turbine dynamics, especially in fields such as reliability analysis or fatigue estimation where many simulations are necessary.

In this study, the particle swarm optimization (PSO) and back propagation neural network (BPNN) surrogate model in combination with a multi-objective genetic algorithm are developed for the design optimization of a bionic liquid cooling plate with a spider-web channel structure. The single-factor sensitivity analysis is first conducted based on the numerical simulation approach, identifying three key factors as design variables for optimizing design objectives such as maximum temperature (Tmax), maximum temperature difference (ΔTmax), and pressure drop (ΔP). Subsequently, the PSO algorithm is used to optimize the parameters of the BPNN structure, thereby constructing the PSO-BPNN surrogate model. Next, the non-dominated sorting genetic algorithm II (NSGA-II) is employed to obtain the Pareto optimal set, and the TOPSIS with the entropy weight method is used to determine the optimal solution, eliminating subjective preferences in decision-making. The results show that the PSO-BPNN model outperforms the traditional BPNN in prediction accuracy for all three objectives. Compared to the initial structure, the Tmax and ΔTmax are reduced by 1.09 °C and 0.41 °C in the optimized structure, respectively, with an increase in ΔP by 21.24 Pa.

HKUST Electronic Theses Artificial neural network-based surrogate modelling methods in accelerating topology optimization process for large-scale designs by Tan Ren Kai thesis 2022 1 online resource (xvi, 100 pages) : illustrations (some color) Topology design which optimizes the materials or…Read more ›

This paper presents the development and evaluation of Artificial Neural Networks (ANN) based models and optimally selected surrogate models to provide the day-ahead forecast of the hourly-averaged energy load of buildings, by relating it to eight weather parameters as well as the hour of the day. Although ANN and other surrogate models have been used to predict building energy loads in the past, there is a limited understanding of what type of model prescriptions impact their performance as well as how un-recorded impact factors (e.g., human behavior and building repair work) should be accounted for. Here, the recorded energy data of three university buildings, from 9/2013–12/2015, is cleaned and synchronized with the local weather data. The data is then classified into eight classes; weekends and weekdays of Fall/Winter/Spring/Summer semesters. Both Multi-Layer Perceptron (MLP) and Radial Basis Function (RBF) NNs are explored. Differing number of hidden layers and transfer function choices are also explored, leading to the choice of the hyperbolic-tangent-sigmoid transfer function and 60 hidden layers. Similarly, an automated surrogate modeling framework is used to select the best models from among a pool of Kriging, RBF, and SVR models. A baseline concept, that uses energy information from the previous day as an added input to the ANN, helps to account for otherwise unrecorded recent changes in the building behavior, leading to improvement in fidelity of up to 30%.

In this paper, the drilling behavior of a new class of composite materials has been experimentally investigated. Thecomposite laminates have been manufactured using glass fibers, epoxy resin, and filler materials. The abundantly availableagro-based waste materials (coconut coir, rice husk, and wheat husk) have been used as filler materials. The drillingexperiments have been performed at several levels of feed (0.03 to 0.3 mm/rev.) and speed (90 to 2800 RPM) using differenttypes of drill bits. The effect of these parameters on the drilling forces (axial thrust and torque) has been analyzed for alltypes of laminates under investigation. The artificial neural network-based models have also been proposed to compute thedrilling forces. The fitness of the models has been measured in terms of mean percentage error between the predicted andactual values. From the investigation, it has been found that the drilling forces computed by the neural network models werequite close to the experimental values.

Two-wheeled wheelchairs have been used as alternatives for the elderly and disabled people to perform physical activities due to their restriction of movement. Significant challenges posed by two-wheeled wheelchairs control due to their inherent instability, resembling that inverted pendulum system. This research addresses these challenges by developing a dynamic non-linear model and stability control using computational algorithms. A neural network-based Nonlinear Autoregressive model with Exogenous Inputs (NARX) was developed, capturing and behaving similar to the complex dynamics of the wheelchair system with the identification on the experimental input–output data. Ensuring stable and responsive control, design of optimized PD-type and PID-type fuzzy logic controllers using Particle Swarm Optimization (PSO) were established and were tested under a simulation environment. The performance was evaluated across various metrics, including Integral Squared Error (ISE), Integral Absolute Error (IAE), Mean Squared (MSE), and Integral Time Absolute Error (ITAE). The result demonstrates that the PSO Optimized PID-type fuzzy logic controller with scaling factor from MSE index performance come out as best overall, significantly outperforms PD-type fuzzy logic controller, reducing its settling time by 12.5% to 35 s, minimizing overshoot to 0.81°, and achieving a negligible steady-state error of 0.046%. These results highlight the significant of integrating fuzzy logic control and PSO to the neural network model in enhancing the stability and performance of two-wheeled wheelchair systems, offering user safety and comfort.