Search

• A novel method was proposed for accurate constitutive description and workability characterization of Mg alloy during hot deformation. • Combining the AT model with PSO-ANN can significantly improve its prediction accuracy and enhance the generalization ability when describing flow behavior. • The activation energy-processing map established using the PSO-ANN-AT model can exhibit superior predictive capability for microstructure evolution. Hot deformation is a commonly employed processing technique to enhance the ductility and workability of Mg alloy. However, the hot deformation of Mg alloy is highly sensitive to factors such as temperature, strain rate, and strain, leading to complex flow behavior and an exceptionally narrow processing window for Mg alloy. To overcome the shortcomings of the conventional Arrhenius-type (AT) model, this study developed machine learning-based Arrhenius-type (ML-AT) models by combining the genetic algorithm (GA), particle swarm optimization (PSO), and artificial neural network (ANN). Results indicated that when describing the flow behavior of the AQ80 alloy, the PSO-ANN-AT model demonstrates the most prominent prediction accuracy and generalization ability among all ML-AT and AT models. Moreover, an activation energy-processing (AEP) map was established using the reconstructed flow stress and activation energy fields based on the PSO-ANN-AT model. Experimental validations revealed that this AEP map exhibits superior predictive capability for microstructure evolution compared to the one established by the traditional interpolation methods, ultimately contributing to the precise determination of the optimum processing window. These findings provide fresh insights into the accurate constitutive description and workability characterization of Mg alloy during hot deformation. [ABSTRACT FROM AUTHOR]

• Fusion of advanced deep-learning models CNN, LSTM, DNN and RNN with GA optimizes GWPZ mapping. • Overlaying GWPZ with drought maps unveils links between drought levels and groundwater potential. • CMIP6 climate models project changing GWPZ under different socioeconomic scenarios. • Workshops and local collaboration empower communities for sustainable water management. This study presents a novel approach to groundwater resource assessment and drought vulnerability, achieved through integrating cutting-edge deep-learning algorithms into a comprehensive framework. In regions susceptible to drought, ensuring groundwater availability is of paramount importance. Addressing this critical need, our research employs an ensemble of advanced algorithms, including long short-term memory, convolutional neural network, deep neural network, and recurrent neural network. These algorithms are further enhanced through optimization using a genetic algorithm to map groundwater potential zones (GWPZ). Leveraging model validation based on the area under the receiver operating characteristic curve (AUCROC), the long short-term memory-genetic algorithm model emerges as the superior algorithm, boasting the highest values of AUCROC: 0.995 for training and 0.996 for testing). Utilizing this optimized model, this study developed the GWPZ map, subsequently overlaying it with established drought maps developed by the Bangladesh Agricultural Research Council across distinct periods—Pre-Kharif, Kharif, and Rabi. The derived baseline GWPZ spatial distribution revealed five categories—very low (34.99%), low (27.67%), moderate (13.26%), high (11.71%), and very high (12.37%)—and their intersection with drought-prone regions, indicative of probabilities of drought occurrence ranging from very severe to low (15.14% to 24.69%). Moreover, employing the best-predicted model, this study projected future GWPZ for 2050 and 2100 using the coupled model intercomparison project 6 general circulation model data under the ambit of three distinct shared socioeconomic pathways (SSPs): SSP1-2.6, SSP2-4.5, and SSP5-8.5. Our findings suggested a contraction in the groundwater potential area by the end of the twenty-first century (2100). This pioneering integration sheds light on the intricate relationship between groundwater availability and drought susceptibility, furnishing invaluable insights for formulating targeted water resource management strategies. By amalgamating advanced computational techniques with geospatial analyses, this research contributes to a more comprehensive grasp of water resource dynamics within the context of escalating climate challenges. Consequently, it offers a foundation for informed decision-making and implementing sustainable water management practices in regions grappling with the dual challenges of groundwater scarcity and recurrent droughts. [ABSTRACT FROM AUTHOR]

To improve the cell productivity of Corynebacterium glutamicum , its initial specific growth rate was improved by medium improvement using deep neural network (DNN)-assisted design with Bayesian optimization (BO) and a genetic algorithm (GA). To obtain training data for the DNN, experimental design with an orthogonal array was set up using a chemically defined basal medium (GC XII). Based on the cultivation results for the training data, specific growth rates were observed between 0.04 and 0.3/h. The resulting DNN model estimated the test data with high accuracy (R2 test ≥ 0.98). According to the validation cultivation, specific growth rates in the optimal media components estimated by DNN-BO and DNN-GA increased from 0.242 to 0.355/h. Using the optimal media (UCB_3), the specific growth rate, along with other parameters, was evaluated in batch culture. The specific growth rate reached 0.371/h from 3 to 12 h, and the dry cell weight was 28.0 g/L at 22.5 h. From the cultivation, the cell yields against glucose, ammonium ion, phosphate ion, sulfate ion, potassium ion, and magnesium ion were calculated. The cell yield calculation was used to estimate the required amounts of each component, and magnesium was found to limit the cell growth. However, in the follow-up fed-batch cultivation, glucose and magnesium addition was required to achieve the high initial specific growth rate, while appropriate feeding of glucose and magnesium during cultivation resulted in maintaining the high specific growth rate, and obtaining a cell yield of 80 g/L ini. • A novel growth medium for Corynebacterium glutamicum was developed using deep neural networks and optimization algorithms. • During the validation cultivation, high specific growth rates were observed in the proposed media. • The limiting factors for high biomass yield were identified in the proposed medium. • To optimize productivity and yield, a fed-batch culture was designed using the proposed medium. [ABSTRACT FROM AUTHOR]

Building integrated photovoltaic thermal (BIPV/T) systems offer a highly effective means of generating clean energy for both electricity and heating purposes in residential buildings. Hence, this article introduces a new BIPV/T system to optimally minimize the energy consumption of a household residential building. The meticulous design of the proposed BIPV/T system is accomplished through MATLAB/Simulink® dynamic modeling. Performance analysis for the BIPV/T system is performed under different seasonal conditions with in-depth techno-economic analyses to estimate the expected enhancement in the thermal, electrical, and economic performance of the system. Moreover, a sensitivity analysis is conducted to explore the impact of various factors on the energetic and economic performances of the proposed BIPV/T system. More so, the two-layer feed-forward back-propagation artificial neural network modeling is developed to accurately predict the hourly solar radiation and ambient temperature for the BIPV/T. Additionally, a multi-objective optimization using the NSGA-II method is also conducted for the minimization of the total BIPV/T plant area and maximization of the total efficiency and net thermal power of the system as well as to estimate the optimized operating conditions for input variables across different seasons within the provided ranges. The sensitivity analysis revealed that higher solar flux levels lead to increased electric output power of the BIPV/T plant, but total efficiency decreases due to higher thermal losses. Moreover, the proposed NSGA-II shows a feasible method to attain a maximum net thermal power and optimal total efficiency of 5320 W and 63% with a minimal total plant area of 32.89 m2 that attained a very low deviation index from the ideal solution. The levelised cost of electricity is obtained as 0.10 $/kWh under the optimal conditions. Thus, these findings offer valuable insights into the potential of BIPV/T systems as a sustainable and efficient energy solution for residential applications. [ABSTRACT FROM AUTHOR]

This article introduces and analyzes an integrated energy system that may generate electricity and potable water at a low cost, speeding up the desired clean transition process. The proposed system uses flame-assisted fuel cells to improve upon the inefficiencies of traditional power generation setups by increasing fuel utilization and reducing waste heat. The product's energy cost and overall efficiency are improved by using fuel gas condensation combined with a multi-effect desalination process. The engineering equation solver software examines the feasibility of the proposed system using energy, exergy, exergoeconomic, environmental, and sustainability analyses. The parametric analysis also compares the effects of changing key design variables and performance metrics. Then, the genetic algorithm integrated with an artificial neural network model is applied to minimize the power cost while enhancing the exergy efficiency. According to the results, the proposed efficient hybrid system has an affordable energy cost of 0.123 $/kWh and an environmental impact of 524 kg/MWh. The parametric study demonstrates that increasing the current density improves potable water production while reducing the efficiency and increasing the unfavorable power costs. Also, it can be shown that the sustainability indices improve significantly when a lower fuel utilization factor and equivalence ratio are chosen. The results further reveal that the optimization achieves a higher exergy efficiency of 6.2% and a reduced power cost of 0.03 $/kWh than the design condition. The scatter dispersal of key variables indicates that the fuel utilization factor lacks effectiveness, and keeping the current density and fuel temperature close to their upper limits is techno-economically favorable. Due to the significant temperature increase, mixing of the fuel and air, and chemical interactions between the reactants, the irreversibility analysis shows that the mixer and the flame-assisted fuel cell have the greatest destruction rate under the optimal conditions. • A new hybrid system is introduced based on the flame-assisted solid oxide fuel cell. • The proposed system's feasibility is assessed from all aspects. • A genetic algorithm optimizer integrated with machine learning is applied. • Optimization achieves higher efficiency of 6.2% and a lower cost of 0.03 $/kWh. • Choosing a lower fuel utilization factor improves sustainability indices. [ABSTRACT FROM AUTHOR]

This paper aims to study the feasibility of municipal sewage sludge utilization as an energy source in a polygeneration system. This system offers distinctive benefits such as contribution to the principled removal of sewage sludge, simultaneous utilization of raw and digested sludge in different parts of the system, and production of renewable hydrogen from bio-waste. 4E (energy, exergy, exergoeconomic, and environmental) analyses, are performed to understand the system performance comprehensively. Then, parametric studies are examined the impact of changing the values of main parameters on the system operation. Afterward, a multi-objective optimization based on a genetic algorithm is carried out to achieve optimal values, considering a trade-off between the exergy efficiency and the total cost rate. Meanwhile, this work harnesses the potential of artificial neural networks to expedite complex and time-consuming optimization processes. According to the results, the gasifier exhibits the highest rate of exergy destruction, and the primary cost of consumption is attributed to its heat supply. The multi-objective optimization findings show that the optimum point has an exergy efficiency of 38.26 % and a total cost rate of 58.17 M$/year. The hydrogen production rate, energy efficiency, and net power generation rate for the optimal case are determined as 1692 kg/h, 35.24 %, and 4269 kW, respectively. Also, the unit cost of hydrogen in the optimal case is obtained 1.49 $/kg which offers a cost-effective solution for hydrogen production. [Display omitted] • 4E analysis and optimization performed on a bio-waste driven polygeneration system. • Plant optimized using artificial neural network and multi-objective genetic algorithm. • The feasibility of integrating a membrane for hydrogen separation investigated. • The optimum exergy efficiency and total cost rate obtained 38.26 % and 58.17 M$/year. • Hydrogen and electricity cost of 1.49 $/kg and 0.0145 $/kWh achieved for optimal case. [ABSTRACT FROM AUTHOR]

Hydrogen purification from steam methane reforming (SMR) by pressure swing adsorption (PSA) technology is a common method to obtain high-purity hydrogen. The breakthrough time can well reflect the adsorption dynamics and help PSA cycle design. In order to avoid time-consuming and labor-intensive breakthrough curve experiments, it is very necessary to develop a fast and accurate surrogate model. In this study, a genetic algorithm (GA) and artificial neural network (ANN) are combined to predict and optimize breakthrough times of adsorbates in activated carbon/zeolite layered beds. Using the Latin hypercube sampling strategy, training data sets of GA-optimized ANN (GA-ANN) obtains from the physical model of adsorption, heat and mass transfer model. The genetic algorithm (GA) optimizes the weights and biases of ANN for better performance. The ANN topology has 4 input variables (superficial velocity, activated carbon height, adsorption pressure and feed temperature) and 3 output variables (breakthrough times of CH 4 , CO and CO 2). The number of neurons in the hidden layer for the GA-ANN model was optimized as 7 to predict and optimize the maximum breakthrough time of SMR with a four-component (H 2 /CH 4 /CO/CO 2) system. The sensitivity analysis displays that relative importance to the breakthrough time is in the order of superficial velocity (40.88%) > adsorption pressure (24.55%) > activated carbon height (21.04%) > feed temperature (13.53%). To obtain high-purity hydrogen, the GA-ANN model combined with multi-objective GA optimization is used to maximize the breakthrough times of CH 4 and CO. The GA-ANN surrogate model proposed in this paper can not only accurately and quickly achieve the purpose of predicting the system breakthrough time with a high correlation coefficient (R = 0.9937) but also obtains the optimal operating conditions of PSA hydrogen purification. • A rigorous heat and mass transfer model of multi-component adsorption is established. • Artificial neural network (ANN) is well-trained by Latin hypercube sampling method. • Genetic algorithm (GA) optimizes the weights and biases of ANN for better performance. • Superficial velocity has the highest relative importance value on breakthrough time. • GA-optimized ANN model and GA combined for maximal breakthrough time of impurities. [ABSTRACT FROM AUTHOR]

Accurate isotherm models contribute to predicting the amount of adsorbate adsorbed in the adsorbent within a certain range of pressure and temperature. Experimental adsorption isotherms of five components (CO 2 , CO, N 2 , CH 4 and H 2) on pelletized zeolite 13X from publication data were correlated with common-used fundamental isotherm models, including Langmuir, Sips, dual-site Langmuir (DSL), and temperature-dependent (TD) isotherm models, including TD Langmuir, TD Sips and TD DSL. Overall, the average relative errors (AREs) of TD models are higher than those of common-used fundamental isotherm models, but TD models have better applicability. In addition, three surrogate artificial neural networks (ANNs) optimized by genetic algorithm (GA) of single- and multi-component gas adsorption isotherm models are proposed. Three GA-ANN models set the adsorption amount as the output and take one, two or three potential inputs, such as adsorption pressure, temperature and adsorbate components. Three GA-ANN models generally have lower AREs and better applicability than fundamental isotherm models and TD isotherm models. The proposed GA-ANN3 model, which has three inputs, makes a new attempt at adsorption isotherms and can fit the multi-component gas adsorption at different temperatures with the change of adsorption pressure. The adsorber dynamics for multi-component adsorption (CO 2 : CO: N 2 : H 2 = 19.9: 0.1: 44.6: 35.4 mol%) in zeolite 13X bed are compared by the extended TD Langmuir model and extended TD DSL model. Since the AREs of the TD DSL model are smaller than that of the TD Langmuir, the breakthrough curves of multi-component gas simulated by the TD DSL model are better than that of TD Langmuir. These results can contribute to relevant researchers choosing suitable adsorption isotherm models for numerical simulations of adsorption processes. • Fitted dual-site Langmuir model performs better in breakthrough curve simulation. • Artificial neural network (ANN) enables construction of adsorption isotherm models. • Genetic algorithm (GA) optimizes the weights and biases of ANN for better performance. • Three GA-ANN models have lower average relative errors and better applicability. • GA-ANN3 attempts to express adsorption isotherms for multi-component mixtures. [ABSTRACT FROM AUTHOR]

• An optimal path planning is proposed for spacecraft fleet to asteroids detumbling. • Optimal low-thrust trajectories for the spacecraft fleets in a 3D high perturbed asteroid environment is designed through a combined GA-Collocation method. • A deep neural network is used to rapidly approximate the gravitational acceleration of the asteroid. • The optimal problem utilized optimization method, as GA for guessing systematically the optimal initial costates, with high speed and accuracy be able to it yield consistent and reliable results. Following space discoveries, asteroids were found as rich sources of minerals and organic matter that can be exploited. In this paper, we present a precise, fast, and robust energy-optimal soft landing with a combined GA-Collocation method for a spacecraft fleet aimed at an irregular asteroid detumbling mission. It is assumed that the spacecraft fleet carried to the asteroid's equilibrium points introduced as the start locations of the mission. Here, the 433 Eros asteroid has been considered as the stony target known for its irregular elongated shape and nearly large dimension. We modeled the gravitational potential field by polyhedron as the most accurate one for such asymmetric objects. Hence, to decline the high burden computational of related equations in gravitational acceleration calculation, a Deep Neural Network with seven hidden layers and a relatively large dataset covering 6000 pairs has been developed. We utilized a Genetic Algorithm to guess systematically more optimal and reliable initial costates of some key parameters to raise the speed and accuracy of estimation. The acceptable efficiency and accuracy in optimal path prediction of the presented approach for a fleet of spacecraft approved throughout precise simulations. Eventually, the feasibility of the proposed approach is demonstrated through the corresponding results. [ABSTRACT FROM AUTHOR]

Partial differential equations (PDEs) are essential mathematical models for describing a wide range of physical phenomena. Numerically, Physic-Informed Neural Networks (PINNs), a variant of artificial neural networks, present a promising method for solving PDEs. However, due to limitation in accuracy and stability, various adaptive PINN variants have been proposed. We have designed a novel approach that adopted self-adaptive PINN (SA-PINN) with two optimization techniques: the genetic algorithm (GA) and the limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) algorithm. Self-adaptive PINN modifies the weights in the loss function to be fully trainable, enabling the ANN to learn and stabilize the PINN in approximating the difficult regions of the solution. GA initializes the population of ANN trainable parameters to optimize the training process with less number of iterations, while L-BFGS is used to find the best solution accurately. Our proposed approach, named SA-PINN-GA-LBFGS, is tested on solving several benchmark PDE problems including elliptic, parabolic, and hyperbolic types. We compare our results with state-of-the-art methods, demonstrating that SA-PINN-GA-LBFGS provides higher accuracy and greater efficiency. [ABSTRACT FROM AUTHOR]

Distracting while driving is a serious issue that causes serious direct and indirect harm to the society. To avoid these problems, detecting dangerous drivers' behaviour is very important.This research focuses on detecting driver behaviour with a combination of artificial deep learning and machine learning models with genetic algorithm. Most of the previous works have focused on using convolutional neural network as deep learning model or support vector machine as machine learning model for actions detection of drivers from input images. The proposed structure will use genetic algorithms to first choose the structure of feature extractor from famous CNN models such as VGG19, ResNet50, and DenseNet121. After mentioning feature extractor, proposed framework contains two layer of dense layer for classification as a deep learning model. On the machine learning side K nearest neighbor, random forest, support vector machine, and extreme boost algorithms have been used as classifiers. Genetic algorithms will specify number of neurons and activation functions of these layers for deep learning and hyperparameters such as number of estimators for machine learning models. Proposed model has been developed with the use of state farm dataset that contains information of 1 safe driving class and 9 dangerous behaviours such as texting while driving, talking with passengers, drinking, etc. Experimental results indicate 99.80% accuracy for classification of the state farm distracted driver detection with combination of genetic algorithms and deep neural networks. Compared to similar research, the proposed approach has shown superior results for classification of state farm distracted driver detection. Proposed approach chooses the feature extraction model and hyperparameters of the classification layer automatically. Thus it can be used for driving behaviour classification with seeing new situation too. Proposed framework can be used as a real time driver's distraction detection to decrease car traffic accidents and alleviate corresponding damages to the drivers. [ABSTRACT FROM AUTHOR]

In recent years, Deep Learning models have shown a great performance in complex optimization problems. They generally require large training datasets, which is a limitation in most practical cases. Transfer learning allows importing the first layers of a pre-trained architecture and connecting them to fully-connected layers to adapt them to a new problem. Consequently, the configuration of the these layers becomes crucial for the performance of the model. Unfortunately, the optimization of these models is usually a computationally demanding task. One strategy to optimize Deep Learning models is the pruning scheme. Pruning methods are focused on reducing the complexity of the network, assuming an expected performance penalty of the model once pruned. However, the pruning could potentially be used to improve the performance, using an optimization algorithm to identify and eventually remove unnecessary connections among neurons. This work proposes EvoPruneDeepTL, an evolutionary pruning model for Transfer Learning based Deep Neural Networks which replaces the last fully-connected layers with sparse layers optimized by a genetic algorithm. Depending on its solution encoding strategy, our proposed model can either perform optimized pruning or feature selection over the densely connected part of the neural network. We carry out different experiments with several datasets to assess the benefits of our proposal. Results show the contribution of EvoPruneDeepTL and feature selection to the overall computational efficiency of the network as a result of the optimization process. In particular, the accuracy is improved, reducing at the same time the number of active neurons in the final layers. [ABSTRACT FROM AUTHOR]

The synergistic effects of two or more carrier species significantly intensify the performance of liquid membranes. However, this process is difficult to model as different carrier species make different complexes with permeating species. Artificial Neural Network (ANN) based modeling is helpful in this case. First number of neurons in the hidden layer was optimized to reduce mean squared error (MSE). Minimum MSE (10.96) was achieved with ten neurons. Simulations showed that neodymium transport initially increased with increasing carrier species TOPO concentration but later decreased due to increased viscosity of liquid membrane. Similarly, higher neodymium transport was found at a moderate concentration of H 2 SO 4 in receiving phase and a low concentration of HCl in the feed phase. Trained ANN was coupled with the Genetic Algorithm to determine the optimum operating parameters in their given range. 0.12 M TOPO concentration in the liquid membrane, 4.2 M of H 2 SO 4 concentration in the receiving phase, and 0.5 M of HCl concentration in the feed phase were optimum values of the operating parameters to achieve 100% neodymium transport through the liquid membrane in 360 min operation. [Display omitted] • Liquid Membrane was modeled using Artificial Neural Network. • Artificial Neural Network was coupled with Genetic Algorithm for optimization. • 0.12 M TOPO was found optimum carrier species concentration. • 0.42 M H 2 SO 4 was found optimum concentration of stipend agent. • 100% Neodymium transport was achieved after 360 min of operation. [ABSTRACT FROM AUTHOR]

The article considers the appliance possibility of mathematical apparatus of artificial neural networks for controlling processes of intellectual synthesis of large discrete systems models with a specified behavior. The analysis of the influence of different methods and its classification has carried out, which allows defining the moving trajectory of population in a space of decisions: diffusion to search new extrema and compact study the field in order to increase the speed of population convergence. To solve the problem, using the neural network approach, the two graphical presentations of population conditions have been done: the first, the value of fitness functions of each individual and, the second, some individuals with the same value of adaptation functions. The article defines the structure of the neural network in the form of the network without feed-back and number of neurons in the hidden layer. The training of neural networks has been done and the examples of management decision making based on graphical presentation of the population condition have been provided. The unification of evolutionary procedures by one mathematical apparatus of the Petri net was proposed, which allows using the technology of parallel calculation under programmed realization. Using parallel calculation, combined with proposed models and methods should increase the processing speed of decision-making support systems based on genetic algorithms. [ABSTRACT FROM AUTHOR]

During operation Pelton turbines suffer from erosion and cavitation wear, which decreases the hydraulic efficiency. Welding repair is a cost-effective method for recovering the original bucket geometry but requires templates to ensure proper results. However, those templates may be unavailable for Pelton units with decades of operation. This is the case of one of the 2.8 MW Pelton units at Nima II, a small hydropower plant in Colombia that began operation in 1942. This paper presents the redesign and optimization of the bucket. First, the bucket is rebuilt following both, technical recommendations, and Design of Experiments (DOE) to verify the effects of the basic dimensions on the hydraulic efficiency. Experimental measurements on the eroded turbine validate the two-phase computational fluid dynamics (CFD) modelling to be applied in the optimization process. Second, the best inner surface is found by using a multi-objective optimization based on Artificial Neural Networks trained by the previous DOE and a genetic algorithm. At the end, the CFD results suggest a total hydraulic efficiency recovery of 8 %. Further experimental tests on the optimized runner showed a difference of 0.7 % from the predicted value. [ABSTRACT FROM AUTHOR]

The operation control plays a great role in the running performance of an industrial process. To achieve a successful control, the key point is searching for the target value of the control parameters. The relationship between operation parameters and performance parameters is usually nonlinear and complex, so it is hard to control an industrial process excellently with workers' experience. Based on a large amount of actual operation data, a bridge between control parameters and performance parameters might be built by the neural network. Prediction model is established by using neural network, genetic algorithm was then used to refine these parameters, the optimal condition can be further achieved by combining the relation bridge and genetic algorithm (GA). Paying attention to a cement clinker production process, this article established an optimizing approach based on the neural network and genetic algorithm and one-year operation data (615 sets). The feeding mass rate of raw meal and coal at the precalciner and the feeding coal mass rate at the rotary kiln are selected as the main independent variable and control parameters, and the specific standard coal consumption is determined as the key performance parameter and the optimization objective function. The mean square error and the correlation coefficient of the established neural network model are 12.84 and 0.89, respectively. The relative errors of approximately all prediction data (92.52 %) are within ±5 %. The optimal values for the raw meal feeding rate, the coal feeding rate into precalciner, and the coal feeding rate into rotary kiln are 259.57 t/h, 7.84 t/h, and 7.40 t/h, respectively. In that optimal condition, the specific standard coal consumption reaches 80.00 kg standard coal /t clinker (This is a relative value, as there is a slight drift in zero point of the factory's instruments). A highly accurate neural network model is developed, which can significantly reduce standard coal consumption and improve industrial energy efficiency. • Operating parameters of a cement plant were optimized by BP neural network and GA. • Feeding quantity of raw meal, coal at kiln inlet and outlet were the main objects. • The three optimal independent variables are 259.6, 7.8 and 7.4 t/h, respectively. [ABSTRACT FROM AUTHOR]

• Enhanced cellulo-xylanolytic consortium production from Hypocrea lixii GGRK4 using MOGA-ANN is reported. • The production of CMCase (9.43 fold) and xylanase (4.40 fold) higher than un-optimized process. • The improved deinking efficiency and brightness is reported for photocopier paper and newspaper. • The physical strength of the waste papers were enhanced whereas double fold property was decreased proving its reusability. • A significant fungal enzyme consortium preparation for improved waste paper deinking achieved. This study reports the cellulo-xylanolytic cocktail production from Hypocrea lixii GGRK4 using multi-objective genetic algorithm-artificial neural network tool, resulting in 8.32 ± 1.07 IU/mL, 51.53 ± 3.78 IU/mL activity of CMCase and xylanase, respectively with more than 85 % residual activity at 60 °C and pH 6.0. Interestingly, metal ions viz. K+ and Ca2+ stimulated the enzyme activity, whereas Fe2+ and Cu2+ reduced the activity. Significant amounts of hydrophobic compounds, chromophores, and phenolics were released after wastepapers deinking. The deinking efficiency of 73.60 ± 2.45 % and 38.60 ± 1.34 % was obtained for photocopier paper and newspaper, respectively, whereas brightness of 89.90 ± 2.10 % ISO and 44.90 ± 1.63 % ISO was reported for both types of waste papers. The physical strength of deinked photocopier paper and newspapers, i.e., tensile index (3.10 and 0.50 %), tearing index (7.10 and 4.83 %), and burst factor (8.61) were enhanced whereas double fold property was decreased proving wastepaper reusability. This consortium showed effective and significant enzymatic deinking efficiency for recycled wastepapers. [ABSTRACT FROM AUTHOR]

[Display omitted] • Machine learning model (ANN) was trained to estimate the performance of Fischer–Tropsch synthesis perfectly. • The effect of different parameters such as temperature, pressure, H 2 /CO, and space velocity on CO conversion and products selectivity was studied. • Genetic Algorithm optimization technique was applied to maximize the selectivity of liquid fuel by manipulating the operating parameters. This study aimed to increase liquid fuel (C 5+) product selectivity from CO and H 2 conversion over a Cu/Fe-based catalyst and followed by artificial neural network simulation to enable a profound visualization strategy for Fischer–Tropsch (FT) process. The experiments were performed in a fixed-bed reactor over temperatures (T) of 483–553 K, pressures (P) of 14.5–203 psig, space velocities (SV) of 1000–2900 hr−1, and H₂/CO feed ratios of 1–6.2. Mathematical modeling (MM) and artificial neural network (ANN) were utilized to simulate the nonlinear CO conversion and product selectivity of the FT process. Nine simultaneous reaction pathway was developed to evaluate the reactant and product behaviors under varying FT conditions. In addition, a multilayered parameter processed by a neural network was used to predict the performance of the CO hydrogenation process. A comparison between these two approaches demonstrated that ANN more closely mimicked the nonlinear conversion and selectivity of the FT process than MM, and the correlation coefficients for the predicted mole fractions and CO conversion were 0.991 and 0.986, respectively. A hybrid Genetic Algorithm–ANN procedure was constructed to predict the optimum operating conditions to increase liquid fuel over the catalyst. The optimum C 5+ selectivity (75.1 %), which meets the FT process objective, was achieved under operating conditions of T = 483 K, P = 203 psig, SV = 1303 hr−1, and H₂/CO feed ratio = 1.01 for process design optimization. [ABSTRACT FROM AUTHOR]

Diabetic peripheral neuropathy is the most common complication of diabetes, generating physical consequences and limitations in diabetic population. Early diagnosis of this condition based on posture-related tests is still a research issue in the human body movement field. Currently, there is a need for affordable and reliable tools to identify and monitor the progression of this condition. The main goal of this work is to provide a method based on posture feature extraction from the center of pressure time series data, data analysis and machine learning models, in order to find patterns that allow the discrimination among healthy, diabetic and neuropathic people. Time series data from participants are captured using a low cost Wii Balance Board platform, following an experimental protocol. Following a data processing process, that includes reduction of noise, generation of new features, and data augmentation using a method inspired in genetic algorithms to overcome the small number of data. A hierarchical approached is applied in order to deal with class values overlapping. Machine learning algorithms are applied to generate classification models. Results show a range of 89%–97% F1-score, according the case being considered, of the feed-forward neural network models, after the application of the hierarchical approach, being superior to the other classifiers. The high predictive score achieved provides empirical evidence that the proposed approach, based on data augmentation, hierarchical classification and classical machine learning models, can reliably support the detection of diabetic neuropathy condition, also suggesting a simplification of the current procedure. • A method is proposed to identify neuropathies in people using time series analysis. • Data come from 32 adult people under postural tests, producing 433 time series. • Data augmentation based on genetic algorithms is used to deal with class imbalance. • A hierarchical classification method is used to improve accuracy of classifiers. • The proposed methods achieved a maximum accuracy of 97% on real test data. [ABSTRACT FROM AUTHOR]

[Display omitted] • The syndicated optimization for engine-ORC combined system is conducted. • The power, bsfc, and BSNOx of combined system are improved by coupling ANN with NSGA. • The optimum injection timing, valve phases, pump and expander speeds are obtained. • The optimized BSNOx reduces 37.16%, without the deterioration of bsfc and power. Organic Rankine cycle (ORC) could compensate for the brake power loss caused by in-engine purification by recovering exhaust energy, thus achieving lower emissions and higher power output for the engine. Aimed at identifying optimum parameters that could maximize the performance of a diesel engine-ORC combined system, this study conducts a multi-objective optimization of dynamic performance, fuel consumption, and NOx emission of the combined system, considering the engine's injection timing, intake, and exhaust phases, ORC's working fluid pump speed, and expander speed as decision variables. Initially, a validated simulation model is used to study the effect of these parameters on the performance of the combined system, indicating the necessity of optimization. Subsequently, initial datasets are obtained, with the data of decision variables obtained using the D-optimum Latin hypercube sampling method and the performance indexes calculated by the simulation model. Following these initial datasets, artificial neural network (ANN) models predicting the performance of the combined system are constructed. And the power output, brake-specific fuel consumption (bsfc), and NOx emission are optimized by coupling ANN with non-dominated sorting genetic algorithm-III. Multi-objective optimization results indicate that overly reducing NOx emissions can degrade the dynamic performance and fuel efficiency of the combined system, with the degradation of 1.85% in power output and 1.89% in bsfc as a consequence of a maximum reduction of 43.79% in Brake-specific NOx (BSNOx). Prioritizing maximum power could substantially reduce fuel consumption but weaken the effectiveness of NOx reduction, with a 4.03% improvement in power output and a reduction of 3.87% and 16.48% in bsfc and BSNOx, respectively. A balanced approach considering both dynamic performance and NOx emission is the most suitable for enhancing performance with emission control in mind. This could ensure a 37.16% reduction in BSNOx, maintaining comparable power and fuel consumption. This study offers valuable insights for future advancements in engine performance. [ABSTRACT FROM AUTHOR]

The injection molding process of non-pneumatic tires is a crucial step in ensuring high-quality product and performance. To improve the quality of non-pneumatic tire processing, this study focuses on honeycomb non-pneumatic tire. Based on the measured material's viscoelastic and dissipative heating characteristics, a co-simulation model and calculation method were established by combining injection molding process optimization with tire's mechanical properties. Compared with the usual simulation of tire rolling performance, this study considered the quality defects caused by injection molding when simulating the tire rolling performance, which was more in line with reality and had higher simulation accuracy. The effectiveness of such co-simulation method was verified by experiments. Based on this co-simulation method, a dataset of tires rolling resistance and dynamic temperature rise was obtained, and a neural network model between the non-pneumatic tire injection molding process and the rolling resistance and temperature rise was established. The genetic algorithm was used to optimize the injection molding process parameters. This method not only saves the calculation time and cost, but also avoids the error caused by the simulation separation of the molding process and the tire rolling process, providing theoretical support and reference for the design and manufacturing of similar tires. [ABSTRACT FROM AUTHOR]

• Fluid flow data of past iterations in inverse design were leveraged through data-driven algorithms. • Machine learning techniques were integrated into the ESA inverse design method to remove the dependency on the user's input. • Deep learning models learned the features of the pressure distribution in the training set to avoid unrealistic and fishtail-geometry-corresponding pressure distributions. • Deep learning models were utilized as a surrogate model for the genetic algorithm to achieve optimal target parameter distribution for an inverse design algorithm. Conventional aerodynamic inverse design (AID) methods have major limitations in terms of optimality and actuality of target parameter distribution. In this research, the target pressure distribution (TPD) of the FX63–137 airfoil was manually corrected to enhance its lift-to-drag ratio from a design perspective, and the corresponding geometry was obtained by the elastic surface algorithm (ESA) inverse design method. An artificial neural network (ANN) was incorporated into the ESA to optimize the TPD and filter out pressure distributions that are not achievable or correspond to fishtail geometries. For this purpose, a genetic algorithm–deep learning (GA–DL) model was developed to optimize the TPD in which the DL model was used to approximate the cost function as a surrogate model. The data generated from CFD simulations during shape modifications in the inverse design were used to train the DL model to correlate PDs with aerodynamic coefficients. Using this approach, the TPD of the FX63–137 airfoil was optimized at a determined angle of attack to reach the maximum lift-to-drag ratio. The unrealistic PDs corresponding to fishtail geometries were eliminated automatically by the ANN during the online process of the inverse design. The results showed that the lift-to-drag ratio of the FX63–137 airfoil increased by more than 18 %. Contrary to traditional methods, the GA–DL model produced several greatly different improved geometries, which were more robust over a wide range of angles of attack. The proposed method addresses the greatest weakness of AID methods in determining the optimal target parameter distribution regardless of the user's input and initial design and without imposing further computational cost for acquiring an external dataset by exploiting the data obtained in the early stages of the design procedure. [Display omitted] [ABSTRACT FROM AUTHOR]

The objective of this paper is to optimize the energy consumption and performance of fuel cell electric bicycles (FCEBs) under specific key input parameters. The paper applied an integrated method that includes an artificial neural network (ANN) and genetic algorithm (GA) to forecast and identify an optimal performance and energy consumption of FCEBs. The simulation model of FCEBs is established and simulated in MATLAB-Simulink environment to generate 1000 data points, that are used for training, validating, testing artificial neural network, the ANN architecture containing five input neurons, two hidden neurons and two output neurons, respectively. Furthermore, the GA is integrated to find the maximum performance and energy consumption once the ANN has exactly been trained. The study found that the FCEB configuration can achieve an effective performance at 30.3 km/h with required power of 210.4 W under speed level_5, radius of wheel 0.39 m, frontal area 0.423 m2, slope grade 0%. In order to validate and verify the simulated results, the experimental approach method was conducted in the same condition. The experimental results fit well with the simulated results in the same initial input parameters. • A method integrating the artificial neural network with genetic algorithm is applied. • The consumed power and fuel cell voltage are optimized by combining artificial neural network and genetic algorithm (ANN-GA). • The effective performance area of fuel cell electric bicycle under various a structure and operating parameters is identified. • Deep-learning results reflect major physical mechanism of fuel cell electric bicycle operation. [ABSTRACT FROM AUTHOR]

• A handheld STN detector based on phase-locked amplification technology. • A novel ring-shaped optical path structure is designed to improve accuracy. • Eight characteristic wavelengths for STN detection is efficiently extracted. Soil total nitrogen (STN) serves as a critical indicator for agricultural productivity. Rapid STN detection is crucial for nutrient monitoring in farmland and for guiding precision fertilization. Traditional near-infrared spectroscopy-based STN detectors necessitate darkrooms or subterranean settings to negate ambient light interference, restricting their practical use. This study develops a handheld rapid STN detector employing phase-locked amplification technology to address this issue. The device comprises optical, circuit, and control units. The optical unit features a ring-shaped split fiber structure to mitigate soil surface irregularities' effects on the signal, with a central InGaAs photodiode for signal reception. Eight wavelengths (970 nm, 1050 nm, 1085 nm, 1200 nm, 1300 nm, 1450 nm, 1550 nm, and 1600 nm) are selected for the light sources through an information theory optimized genetic algorithm. The circuit unit employs direct digital synthesis and operational amplifiers for light signal generation and modulation. Reflected soil light signals are demodulated using an InGaAs photodiode and a phase-locked amplifier, effectively filtering ambient light noise. The control unit integrates an STM32F103C8T6 microcontroller and a Raspberry Pi 4B. The STM32 microcontroller manages signal generation, channel selection, frequency control, A/D conversion, and display, while the Raspberry Pi handles data processing, uploading, and transmission. A BP neural network model optimized by genetic algorithms is embedded for real-time STN data acquisition. Cloud-based data storage and visualization are facilitated through the Alibaba Cloud IoT platform. Performance testing reveals the device's stability and interference resistance, with a standard deviation of electrical signal under varying light conditions below 3 mV. Compared to commercial spectrometers, the correlation coefficient (r) for reflectance data at eight wavebands exceeds 0.9. Accuracy testing shows an R2 above 0.8 and an RMSE of 0.30 g/kg when comparing detector-obtained STN values with standard chemical methods. These results demonstrate the detector's efficacy in ambient light isolation, stability, and accuracy in agricultural settings, offering technical support for variable fertilization in precision farming. [ABSTRACT FROM AUTHOR]

This study aimed at using α-L-arabinofuranosidase CcABF to improve the clarity and active substances in fermented ginkgo kernel juice by artificial neural network (ANN) modeling and genetic algorithm (GA) optimization. A credible three-layer feedforward ANN model was established to predict the optimal parameters for CcABF clarification. The experiments proved the highest transmittance of 89.40% for fermented ginkgo kernel juice with this understanding, which exhibited a 25.56% increase over the unclarified group. With the clarification of CcABF, the antioxidant capacity in juice was enhanced with the increase of total phenolic and flavone contents, and the maximum DPPH and hydroxyl radical scavenging rates were increased by 89.71% and 26.65%, respectively. The contents of toxic ginkgolic acids declined markedly, while the active ingredients of ginkgetin and ginkgolide B showed a modest increase. Moreover, changes in free amino acids and volatile compounds improved the nutritive value and flavor of clarified fermented ginkgo kernel juice. • The fermented ginkgo kernel juice was clarified by α-L-arabinofuranosidase CcABF. • The transmittance of ginkgo kernel juice reached 89.40% by ANN and GA optimization. • Beneficial substances of clarified ginkgo kernel juice increased to varying degrees. • The antioxidation function of clarified ginkgo kernel juice was enhanced obviously. • The nutritive values and flavor of the clarified ginkgo kernel juice were optimized. [ABSTRACT FROM AUTHOR]

Owing to the inherent characteristics of ground beef, adulteration presents a substantial risk for suppliers and consumers alike. This study developed a robust and novel method for identifying replacement fraud in ground beef with beef liver, beef heart, and pork using Near Infrared-Hyperspectral Imaging (NIR-HSI) coupled with chemometric and other statistical methods. More specifically, NIR-HSI provided an efficient and accurate means of identifying each type of adulteration using the classification model Genetic Algorithm (GA) - Backpropagation Artificial Neural Network (BPANN), showing perfect sensitivity and specificity (a value of 1.00) for the calibration and the validation sets for all types of adulteration. As an alternative to chemometric analysis, Hyperspectral Imaging-Root Mean Square (HSI-RMS) value, based on the RMS cut-off calculation, was determined to discriminate types of adulterations without the need of resource-intensive modelling. This HSI-RMS approach provides a simple-to-use method that avoids the complexity of HSI data processing and aims to directly understand the similarity between different spectra of one sample in the pixel level. Different types of adulteration show noticeable differences reflected in the HSI-RMS value (varying from 55 to 1439), which demonstrate the potential of HSI-RMS concept as a novel and valuable alternative for assessing the HSI data and facilitating the identification of adulterants. [Display omitted] • Endogenous and exogenous adulteration were detected simultaneously using NIR-HSI. • Sensitivity and specificity exhibit a value of 1.00 for validation set using raw average spectra from HSI data. • HSI-RMS was first developed for discovering the internal differences between pixels/spectra within the HSI image. • HSI-RMS value can be a valuable parameter for rapidly identifying meat adulteration without the need of modelling. [ABSTRACT FROM AUTHOR]

Indirect evaporative cooling has garnered considerable attention and research owing to its notable advantages, especially high cooling efficiency and low energy consumption. However, the complexity and diversity of parameters affecting indirect evaporative cooler (IEC) performance necessitate the development of a fast and efficient performance prediction model, which remains a critical research objective. To address the particular challenge, this paper introduces a novel approach that utilizes the principles of neural network theory to construct a predictive model based on backpropagation artificial neural networks (BP-ANN). The developed model is employed to investigate the impact of six main operating parameters of the water spray system on the IEC performance. Wet-bulb efficiency and pressure drop of secondary air channel are utilized as the performance evaluation indexes. The model is rigorously validated by relevant experimental data, enabling a comprehensive analysis of the influence of each parameter on the operational performance of the whole system. The significance of the influential parameters is evaluated through grey relational analysis. Furthermore, a multi-objective optimization approach employing a genetic algorithm (GA) is introduced to effectively determine the optimal operating parameters of the IEC system. The results demonstrate that the whole system can efficiently operate with low energy consumption when the wet-bulb efficiency and pressure drop weights are balanced. The optimal operating parameters include a nozzle pressure of 0.15 MPa, spacing of 85 mm, aperture of 6 mm, count of 11, spray density of 0.85 kg/(s ∙ m2) and air velocity of 2 m/s, at which time the wet-bulb efficiency was 0.85, the pressure drop was 210 Pa, and the coefficient of performance (COP) was 12.4, which demonstrated an efficiency increase of 23.2% in comparison with the base system. The findings of this research significantly contribute to a comprehensive understanding of the influence of water-spray operating conditions on IEC performance, and provide valuable insights for the design and operation of high-efficient IEC systems in practical applications. • The datasets were generated by a laboratory-scale prototype of the IEC system. • The ANN model for IEC considering spray operation was trained and validated. • Effects of nozzle and operation characteristics parameters were investigated. • The significance of the influential parameters was evaluated. • The optimal parameters for IEC were determined based on genetic algorithms. [ABSTRACT FROM AUTHOR]

• Time and temperature were most influential during MW irradiation extraction. • The optimum conditions by RSM were 620 W MW power at 74 °C for 18 min. • The optimum conditions by ANN-GA were 685 W MW power at 75 °C for 24 min. • Peleg's model showed best fitting for explaining kinetics of bioactives extraction. • The extract showed remarkable bioactive yields and antioxidant activity. Pomegranate (Punica granatum L.) peel is an appealing agri-food byproduct due to its high content of health-beneficial phytochemicals. However, the extraction processes may alter their bioactivity. A microwave irradiation treatment was applied to enhance the extraction efficiency. Extraction time, microwave power, and temperature were examined and optimized numerically using polynomial model and genetic algorithm. Time and temperature were the most influential factors in intensifying bioactives extraction. The optimal variables were 620 W microwave power at 74 °C for 18 min (RSM), and 685 W power at 75 °C for 24 min (ANN-GA). The obtained yields of total polyphenols, flavonoids, punicalagin, ellagic acid, and total hydrolyzable tannins contents were 165.53 mg GAE/g, 40.78 mg QE/g, 91.05 mg/g, 4.28 mg/g, and 130.82 mg TAE/g of pomegranate peel, respectively. The extract showed significant antioxidant capacities, with 65.74 % DPPH inhibition and 71.26 % ABTS inhibition. Several kinetic models were studied to define the extraction kinetics, with Peleg's model being the best fitting with the highest R2 (0.963–0.987) and lowest RMSE (0.95–5.08). LCMS analysis confirms the presence of 15 hydrolyzable tannins, with punicalagin and ellagic acid being the most abundant. These findings demonstrated the significant potential of microwave irradiation for extracting bioactives from fruit byproducts. [Display omitted] [ABSTRACT FROM AUTHOR]

This work presents a novel hybrid (physics- and data-driven) model for short-term (intra-day and day-ahead, 3h-24h) wind power forecasting (STWPF). Traditionally, STWPF predictors admitted very few meteorological variables only from the grid points closest to the turbines. Here, with the aim to further capture the underlying atmospheric processes ruling the wind variability in the wind farm, the approach relies on drastically expanding the predictor space, composed of numerous meteorological variables throughout a large geographical domain, retrieved from a weather forecasting model (COSMO-1). An ad-hoc genetic algorithm that optimizes the selection of predictors is designed and combined with feed-forward artificial neural networks for its cost function evaluation. The introduced model is compared to multiple benchmark models in a 16-turbine wind farm in the Swiss Jura mountains. For +12h and +24h lead times, the new approach shows a root-mean squared error normalized to the installed wind farm capacity of 11% and 11.6%, respectively. These values entail ∼ 16% higher forecasting skill compared to state-of-the-art predictor frameworks. Results highlight the ability of the presented approach to systematically select as predictors different variables with a well-known impact on the wind farm performance, such as the turbulent kinetic energy or the vertical wind shear. Clustering the data according to the wind direction provides substantial benefit. In addition, it provides a better understanding of the attained improvement: largest performances occur in those wind directions affected by highly complex terrain. This indicates that the proposed model can be especially suitable for wind farms in complex terrain. • A new physics- and data-driven model for 3h-24h wind power prediction is presented. • It relies on massively expanding the predictor space, set by weather model outputs. • It shows capable to select variables with well-known impact on the wind power signal. • It shows 16% higher forecasting skill compared to state-of-the-art predictor spaces. • The proposed model can be especially suitable for wind farms in complex terrain. [ABSTRACT FROM AUTHOR]

Microchannel heat sinks provide the solution to the ever-increasing heat flux generated from micro-electric components. In this study, performance optimization of a microchannel heat sink with delta winglet vortex generators was carried out based on the data obtained from numerical CFD simulations. A total of 192 design points were generated by altering the fluid velocity in terms of Reynolds number (Re), winglet width (W d), length (L d), and the angle of attachment (β) of the winglet. The Artificial Neural Network (ANN) model coupled with Non-dominated Sorting Genetic Algorithm NSGA-II was used simultaneously to minimize the friction factor and increase the Nusselt number. The ANN model predicted the output values within the error limit of 10%. The Pareto optimal front generated by the algorithm contains the input parameters in the range 982 < Re < 988, 177 μ m < L d < 233 μ m , 10 μ m < W d < 25 μ m , 57° < β < 64°. Decision-making methods TOPSIS, LINMAP, and Shannon entropy were employed to calculate the optimal solution from the data set and the obtained points showcased 70%, 120% and 158% surge in Nusselt number while an increase of friction factor by 35%, 109%, and 140% respectively is reported. The Performance Evaluation Criteria (PEC) values obtained from the best solutions were 1.52, 1.72, and 1.92, respectively. Furthermore, the accuracy of the optimal solutions was verified numerically. The flow and thermal field of the microchannel are also analyzed, and results showed that the angle of attachment and width of the winglet played a crucial role in the overall performance. [Display omitted] • Investigated the hydrothermal performance and optimization of the microchannel heat sink with delta winglets. • Analyzed 192 design points obtained by controlling four input variables. • Optimized solution points generated by the combination of ANN and NSGA-II. • Employed decision-making criteria to select the best solution in terms of overall performance. • PEC values of 1.52, 1.72, and 1.92 were achieved from the optimized best solution set. [ABSTRACT FROM AUTHOR]

Spherical pressure-resistant shells, as a universal structural component of deep-sea submersibles, provide a safe and normal operating environment for personnel and internal equipment. In the paper it presented and optimized the BP neural network model based on a genetic algorithm (GA) accordingly, and the method and accuracy are validated through by a beam model. Simultaneously focusing on steel spherical shells, the study proposed a dataset that captures the influence of the primary dimension of the shell (radius-to-thickness ratio, R / t) on the critical pressure response. The genetic algorithm is employed to optimize the back propagation (BP) neural network model for predicting critical pressure. The structural reliability is adopted as a design criterion to determinate and optimize the geometric parameters and critical pressure of the spherical shell structure. Finally, an ultra-high-strength steel spherical model is designed, constructed and meanwhile collapse pressure tests are accomplished to verify the accuracy of the presented improved BP neural network model based on the computational reliability method. The results reveal that the machine learning optimization design method proposed in this paper can effectively enhance the accuracy of critical pressure predictions and the precision of reliability assessments for deep-sea spherical shells. • An optimized BP neural network model based on a genetic algorithm accordingly is proposed, and its accuracy is validated by a beam model. • An improved BP neural network model based on the reliability method (CV-GA-BP-MCS) is presented and confirmed accordingly. • A deep-sea ultra-strength steel spherical model designed by CV-GA-BP-MCS has been designed, tested and verified respectively. [ABSTRACT FROM AUTHOR]

This study proposes a systematic methodology for improving PEMFC's performance combining computational fluid dynamic (CFD), artificial neural network (ANN), and intelligent optimization algorithms. Firstly, a three-dimensional (3-D) multiphase PEMFC CFD model with 3-D fine-mesh flow field is developed. Then the key structural features of the fine-mesh flow field are extracted as optimization decision variables, and the sampling points are selected by using the Latin hypercube sampling (LHS) experimental method. The power density and oxygen uniformity index of sampling points are calculated by CFD modeling to form the database, which is used to train the artificial neural network (ANN) surrogate model. Finally, the single-objective optimization (SOO) and multi-objective optimization (MOO) are implemented by using genetic algorithm (GA) and non-dominated sorting genetic algorithm (NSGA-II), respectively. It was found that using trained ANN surrogate models can get a high prediction precision. The maximum power density of SOO is increased by 7.546% than that of base case and is 0.562% larger than that of MOO case. However, the overall pressure drop in cathode flow field of SOO case is greater than that of MOO case and the base case. Furthermore, the oxygen concentration, the oxygen uniformity index and the water removal capacity of MOO case are better than that of SOO case. It is recommended that the improved flow field structure optimized by MOO is more beneficial to improve the overall performance of PEMFC. [Display omitted] • A novel three-dimensional fine-mesh flow field of PEMFC is designed. • Proposing a co-simulation model for improving the novel PEMFC's performance. • The single-objective and multi-objective optimizations are performed using genetic algorithm. [ABSTRACT FROM AUTHOR]

Hydrogen production from the supercritical water gasification (SCWG) of sewage sludge (SS) is a sustainable and efficient process. However, the challenging and intricate task for the experimental technique is to find out the correlation between proximate, ultimate analysis and gasification conditions with H 2 production. This process is complicated, expensive and requires many experimental techniques. To accurately predict and analyze the effect of input parameters on SCWG of SS process economically, an efficient model must be developed. The novelty of this study includes the consideration of four different machine learning (ML) methods integrated with Genetic Algorithm for the prediction, analysis, and evaluation of Hydrogen yield from the supercritical water gasification of sewage sludge. The ML methods included Support Vector Machine, Ensembled Tree, Gaussian Process Regression, and Artificial Neural Network. The results suggests that GPR is favored for predicting Hydrogen yield (Coefficient of determination (R2) = 0.997, Root Mean Square Error (RMSE) = 0.093, and is highly recommended for dealing with complex variable-target correlation. On the other hand, the performance of Support Vector Machine (SVM) was poor with R2 = 0.761 and RMSE = 2.479. The R2 and RMSE for Ensembled Tree (ET) and Artificial Neural Network (ANN) was 0.994, 0.560 and 0.943, 1.521 respectively. The partial dependence plot shows that temperature, moisture content and pressure are among the effective parameters of SCWG. Furthermore, optimization techniques such as genetic algorithms are incorporated to optimize H 2 production by tuning the ML hyperparameters. Additionally, a Graphical User Interface was developed by utilizing the optimized GPR model for ease in computing H 2 yield.The optimum ML method integrated with GA will be beneficial for researcher to predict the H 2 yield for the experimental work. [Display omitted] • Machine learning methods integrated with Genetic Algorithm are analyzed for prediction, analysis & evaluation of H 2 yield. • Partial dependence plot shows that temperature, moisture content and pressure are among the effective parameters. • Graphical User Interface was developed by utilizing the optimized GPR model for ease in computing H 2 yield. [ABSTRACT FROM AUTHOR]

The study aims to optimize the geothermal and solar-assisted sustainable energy and hydrogen production system by considering the genetic algorithm. The study will be useful by integrating hydrogen as an energy storage unit to bring sustainability to smart grid systems. Using the Artificial Neural Network (ANN) based Genetic Algorithm (GA) optimization technique in the study will ensure that the system is constantly studied in the most suitable under different climatic and operating conditions, including unit product cost and the plant's power output. The water temperature of the Afyon Geothermal Power Plant varies between 70 and 130 °C, and its mass flow rate varies between 70 and 150 kg/s. In addition, the solar radiation varies between 300 and 1000 W/m2 for different periods. The net power generated from the region's geothermal and solar energy-supported system is calculated as 2900 kW. If all of this produced power is used for hydrogen production in the electrolysis unit, 0.0185 kg/s hydrogen can be produced. The results indicated that the overall energy and exergy efficiencies of the integrated system are 4.97% and 16.0%, respectively. The cost of electricity generated in the combined geothermal and solar power plant is 0.027 $/kWh if the electricity is directly supplied to the grid and used. The optimized cost of hydrogen produced using the electricity produced in geothermal and solar power plants in the electrolysis unit is calculated as 1.576 $/kg H 2. The optimized unit cost of electricity produced due to hydrogen in the fuel cell is calculated as 0.091 $/kWh. • An integrated system consists of geothermal and solar energy, an electrolyzer, and a fuel cell unit. • Thermoeconomic analysis and the ANN-based GA optimization are evaluated. • The overall system energy and exergy efficiencies are calculated at 4.97% and 16.0%. • The optimum cost of produced hydrogen and electricity are calculated as 1.576 $/kg H 2 and 0.091 $/kWh, respectively. [ABSTRACT FROM AUTHOR]

Gravity assist is crucial for tour exploration of Jupiter's Galilean moons. To achieve high accuracy and low cost gravity-assisted transfer trajectories, the circular, restricted, three-body problem (CR3BP) model is adopted. Since there is no analytical solution for CR3BP, the orbit propagation must be calculated by numerical integration, which requires considerable time for the design of complex Jupiter exploration missions. Artificial neural networks (ANNs) that can fully approximate complex nonlinear relationships are used to model accurate and computationally efficient maps between the orbit states before and after gravity assist. Thereafter, a new framework based on the ANN model is developed for the fast design of Jovian-moon gravity-assisted transfers, by which optimal trajectories are obtained using a genetic algorithm (GA). Finally, the accuracy and performance of the proposed method are verified by two representative test cases of the Jupiter-Europa and Jupiter-Ganymede systems. The simulation results show that using the trained ANN model can significantly improve the convergence speed of the genetic algorithm, and high-precision, optimal solutions can be obtained rapidly. [ABSTRACT FROM AUTHOR]

Current practices lack reliable models for continuously assessing compaction quality from limited data, which hinders effective compaction quality control and the acceptance of earthwork. Using material source parameters as inputs for assessment models makes it infeasible to achieve a real-time assessment of the compaction quality in earthwork. These issues have hindered the popularization of emerging intelligent compaction (IC) and intelligent rolling compaction (IRC) technologies. Therefore, this study addresses the need for small-sample training by introducing an uncertainty-based multi-population genetic algorithm fused with a backpropagation neural network model (UB-MPGA-BP), enabling the real-time intelligent assessment of compaction quality based on few-shot learning. In case studies involving high- and low-liquid limit silts, existing models trained on large datasets were notably inadequate for small-sample evaluation tasks. However, the proposed method not only performed well in silt cases but also exhibited high performance in rockfill materials. The analysis revealed that in small-sample scenarios without reliance on material source parameters, the proposed method has the advantages of real-time accuracy, generalization, few-shot data-driven, and simple input parameters, which enable the rapid assessment of compaction quality. The model can potentially facilitate IC/IRC applications in other engineering practices. [ABSTRACT FROM AUTHOR]

High Entropy Alloys (HEAs) are a novel class of multi-component alloys with compositional flexibility, presenting a promising alternative to traditional alloys. This research aims to enhance the strength of HEAs while achieving adequate ductility, a challenging objective due to their conflicting nature. We applied evolutionary data-driven models and Bi-Objective Genetic Programming (BioGP) with genetic algorithms (GA) to accurately predict and optimize yield strength and ductility. The predictions were validated through experimental methods, including casting by vacuum arc melting and comprehensive mechanical and microstructural characterization. Our integrated approach successfully developed an HEA exhibiting a strength of 1795 ± 21 MPa and a ductility of 31.45 %. This study highlights the effectiveness of combining data-driven models with experimental validation to advance the development of high-performance materials. • In the present work, HEAs with high strength and ductility, comparable to Ni-based superalloys, were designed. • Advanced data-driven techniques like EvoNN, BioGP, and EvoDN2 were used for efficient alloy composition optimization. • Strength and elongation were modeled with 32 parameters, achieving over 94% prediction accuracy using EvoDN2. • Developed an HEA with 1795 ± 21 MPa yield strength and 31.45% ductility, balancing strength and ductility. • Model predictions were validated through vacuum arc melting and advanced mechanical characterization. [ABSTRACT FROM AUTHOR]

Fluid-elastic instability (FEI) poses a particularly significant challenge to the safety of steam generators, given its potential to cause substantial damage in short periods. The tube bundle system model is inherently complex, resulting in expensive costs for experiments and numerical simulations during the design process. In this study, a substantial volume of FEI experimental data, conducted by various scholars, was compiled to create FEI surrogate model. The accuracy and feasibility of this surrogate model were demonstrated and compared to experimental results. This work represents the pioneering reference on the optimization of tube array configurations, achieved by combining FEI deep neural network (DNN) surrogate model with genetic algorithm (GA). The design variables for optimization encompassed tube array configurations and structural parameters. The optimization objective is to achieve a tubular array structure exhibiting a higher critical velocity, U r , enabling safe operation of the heat exchanger tube bundle across a wider range of flow rates without any FEI accident. The developed approach holds promise for data-driven FEI analysis and optimized tube bundle design, resulting in substantial time and cost savings. We publicly share all code implementations, and we believe that our efforts open a door for the surrogate-model-assisted structural optimization of tube arrays. • The FEI surrogate model is established by training extensive experimental data. • The tube array configurations are first set as input variables in the FEI-DNN model. • The first refernce on tube bundle FEI optimization combined DNN with GA (GA-DNN). [ABSTRACT FROM AUTHOR]

Panax vietnamensis (PV), a member of the Araliaceae family, is noted for its saponin-rich roots and rhizomes. Recent pharmacological studies have shown its anti-inflammatory, anti-cancer, anti-myocardial ischemia, and hypoglycemic effects, establishing PV as a highly valuable variety of ginseng worldwide. Herein, we used an ultrasound-assisted method to optimize the extraction of total saponins from PV, and developed a genetic algorithm (GA) to optimize the backpropagation (BP) neural network model and predict the best process conditions. The best process with an ethanol concentration of 65.8 %, extraction times of 4.0, ultrasonic temperature of 40.7°C, and an extraction rate of 4.4103 % was achieved through response surface optimization. Conversely, the best process optimized by the GA-BP neural network model had an ethanol concentration of 70.0 %, extraction times of 4.0, ultrasonic temperature of 41.0°C, and an extraction rate of 4.5492 %. Moreover, 254 terpenoids, consisting of 182 triterpenes, 33 monoterpenes, 19 sesquiterpenes, and 20 diterpenes, along with various types of saponins, were identified through ultra-high performance liquid chromatography-MS/MS (UPLC-MS/MS). Specifically, 87 dammarane-type saponins were detected, including 46 protopanaxadiol types, 41 protopanaxatriol types, 5 octilonol types, and 18 oleanolic acid types. The total saponin extraction rate achieved through the GA-BP neural network model surpassed that of the response surface optimization process, indicating the superiority and stability of the GA-BP neural network optimized process and thus revealing the optimal method for PV total saponin extraction. The comprehensive analysis of PV saponins using UPLC-MS/MS enhanced our understanding of total saponins in PV, laying the foundations for the strategic development and utilization of PV plant resources. • Introduces the novel use of the Response Surface Analysis and Backpropagation Neural Network-Genetic Algorithm method to optimize the extraction of saponins from the root of Panax vietnamensis. • The saponin component in Panax vietnamensis was detected and analyzed for the first time using UPLC-MS/MS. Explore further insights into the composition and content of saponins in Panax vietnamensis. [ABSTRACT FROM AUTHOR]

• A theoretical model was developed for optimizing spatial arrangement of reactor-scale cascade dual-catalyst system. • Defined a new descriptor (Delta Switch) to spatially quantify the contributions of optimal dual-catalyst distribution. • The effect of side reaction from each catalyst on the optimized arrangement were isolated and analyzed by variation of Delta Switch value. • A simplified and practicable arrangement optimization framework was built based on patterns of Delta Switch value and isolated side-reaction analysis. The performance of cascade dual-catalyst systems is highly related to the catalysts spatial arrangement within the bed layers. However, optimization tools for the cascade dual-catalyst layer arrangements are still in need of development. In this study, an optimization framework based on genetic algorithms (GA) and artificial neural networks (ANN) is developed for optimization while explore effects of side reaction on optimized arrangement. Further, a new descriptor as known as delta switch (DS) is introduced. Isolated analysis of the effects of the two types of side reactions shows that the solely side reaction on S0 results in a repeated S0-S1-S0 structure, while the introduce of side reactions on S1 disrupts this structural regularity and reduces the optimal length of the bed layer. Based on DS patterns, a simplified cascade dual-catalysts spatial optimization framework (≤3 layers) is employed, achieving ≥ 95 % performance of original GA. [ABSTRACT FROM AUTHOR]

This study addresses the challenge of multi-objective optimization of a composite shell structure while adhering to constraints on the number of calls to a pseudo-experimental model, simulating real experiments. Two considered objective functions are defined to determine the investigated structure's dynamic properties and material costs; the optimization involves genetic algorithms, neural surrogate model and multi-fidelity finite-element models. The results of multi-objective optimization were presented as Pareto fronts. A new strategy for preliminary result verification is proposed, significantly reducing the need for a computationally intensive complete verification that requires complex models or experimental investigations. Two different indicators are applied to assess the quality of the obtained Pareto fronts; one is a new one proposed in the paper. Moreover, a multi-fidelity approach is discussed, and three finite element models with different mesh densities are employed, together with a pseudo-experimental model constructed using high-fidelity results and incorporating a nonlinear transformation. However, challenges arise due to the arbitrarily constrained number of pseudo-experiments, limiting future experiments is crucial. The study highlights the need for further analysis of Pareto front indicators and statistical analysis of applied tools like deep neural networks and genetic algorithms. Future research directions include exploring ensemble learning in surrogate models for potential optimization benefits. • Multi-Fidelity Modeling: Demonstrated the effectiveness of a multi-fidelity approach. • Deep Network Surrogate Models: Showed the ability to capture complex data structures. • Multi-Objective Optimization: Determined dynamic properties and material costs. • Curriculum Learning: An innovative strategy for surrogate model improvement. • Pareto Front Quality: Evaluated using relevant indicators. [ABSTRACT FROM AUTHOR]

This study proposes a novel approach combining Machine Learning (ML) techniques and Genetic Algorithms (GA) for predicting High-Performance Computing (HPC) job run times. The objective is to create a prediction method universally applicable to any HPC system, irrespective of workload characteristics, application specific parameters, user behavior, or hardware architecture. Since user-supplied run time estimates are often inaccurate, we aim to categorize job runtimes into several classes, allowing users to select appropriate classes for their jobs. A Genetic Algorithm is developed to optimally define these runtime classes, determining both the number of classes and the time intervals represented by them. Four Machine Learning algorithms (K Nearest Neighbours, Support Vector Regression, Extreme Gradient Boosting and Deep Neural Networks) are implemented for run time prediction. A unique set of features extracted from historical job data serves as input to the Machine Learning models. The generalized nature of our method is demonstrated by validating its performance on data from six clusters with distinct configurations, applications and runtime distributions. Our results illustrate the superior performance of Machine Learning models incorporating GA-defined runtime classes for all datasets. Across all six datasets, our method achieves R2 scores exceeding 0.8, and accuracy greater than 0.7. [Display omitted] • Proposed an innovative method for predicting HPC job run times combining Machine Learning (ML) and Genetic Algorithms (GA). • Replaced run time estimations by users with runtime classes defined by GA, enabling users to select from predefined classes. • Implemented and compared four ML models for run time predictions: KNN, SVR, XGBoost and DNN. • The proposed method demonstrated superior performance across data from six diverse clusters. [ABSTRACT FROM AUTHOR]

The bionic fins inspired by the shark are proposed to enhance the thermal-hydraulic performance of the intercoolers in this work. The geometric parameters of the fins are investigated to examine the effects on thermal-hydraulic performance. The height and width of protrusions are chosen as optimization parameters based on the results of the geometric parameter analysis. Non-dominated genetic algorithms are combined with artificial neural networks to obtain the optimal solution for the geometry parameters. The artificial neural networks used in this study exhibit a reliable accuracy with errors below 10 % (with mostly remaining below 5 %). The optimal values for the height and width of the protrusions are determined to be 1.01 mm and 0.82 mm, which results in a Colburn factor of 0.01923 and a comprehensive index of 0.04305. Compared with the original design, the optimized structure yields 17.69 % and 4.77 % increments for the Colburn factor and comprehensive index. Compared with the original model, the intercooler with the optimal structure exhibits a 9.9 % increase in entropy generation and a 20.2 % enhancement in the exergy difference between the inlet and outlet. • Bionic fins are designed to enhance the thermal performance of intercoolers. • Globally optimal results are obtained by non-dominated genetic algorithms. • Optimal fin is proposed with a protrusion height of 1.01 mm and width of 0.82 mm. • A 17.69 % increment in Colburn factor is obtained for the optimized intercooler. • Comprehensive index of the intercooler increases by 4.77 % with the optimized fin. [ABSTRACT FROM AUTHOR]

• A dataset generation method in large depth of field is proposed. • A 3D displacement measuring device is applied to measure the expansion of boiler. • The MOGA can improve the efficiency and precision of ANN simultaneously. • The PSF can reduce pixel coordinate errors by half compared to one that does not include it. In this paper, an acronym for stereo vision calibration-point spread function (SVC-PSF), is designed to accurately represent image degradation in large-depth-of-field (LDOF). The error of the dataset generated by the SVC-PSF is 50 % lower than that of Stereo Vision Calibration (SVC) method in LDOF. A three-dimensional point coordinate reconstruction artificial neural network (ANN) is improved through a multi-objective genetic algorithm (MOGA) to satisfy both computational accuracy and real-time performance. The MOGA improved ANN can reconstruct 1250 points within 0.069 s on an embedded platform. Finally, SVC-PSF is used to generate a dataset, and the MOGA-ANN improved model is trained based on this dataset. An embedded device, stereo displacement measurement device (SDMD) is designed to run SVC-PSF-MOGA-ANN algorithm, and this device is applied in thermal power plant boiler expansion monitor. The real-time monitor results shows that when the boiler power increases, more working medium is needed to produce steam, so the boiler's own weight increases, and the boiler expands downward under the action of gravity. At low power, the required working medium is reduced, and the corresponding expansion amount is also reduced. [ABSTRACT FROM AUTHOR]

The Artificial neural networks (ANN) model established in this study can accurately predict the rolling force of Titanium-Aluminium (TiAl) alloy during isothermal rolling process, and can achieve online monitoring and rolling plan optimization. TiAl alloys are widely used in the aerospace field due to their high strength, low density, good corrosion resistance, and high temperature stability. However, the wrinkling and cracking problems of TiAl alloys are still difficult to solve. The important way to solve this problem is isothermal rolling, and the precise control of rolling force during the rolling process provides a strong guarantee for solving the problem. This study established a rolling force prediction model for TiAl alloy isothermal rolling based on genetic algorithm optimized Back Propagation (BP) neural network. The database used for training and testing is from a controlled two-roll rolling mill. The input parameters of the network are inlet temperature, roller temperature, rolling speed, reduction ratio and friction coefficient between roller and strip. The superiority of genetic algorithm is verified by comparing genetic algorithm with BP neural network optimized by fuzzy theory. The results show that when the population size is 40, the crossover probability is 0.2, the mutation probability is 0.1, the evolutionary algebra is 10, the transfer functions of the hidden layer and the output layer are tansig, and the structure of the neural network is 5-5-1, optimization works best. The maximum relative error is only 5.3%, the average relative error is only 1.6%, and the correlation coefficient is 99.78%. • The GANN model established in this study has significantly improved prediction accuracy compared to previous models. • The GANN model provides a powerful tool for solving the problem of narrow processing windows for TiAl alloys. • The prediction problem of rolling force with strong coupling characteristics was solved using numerical simulation method. [ABSTRACT FROM AUTHOR]

[Display omitted] • A GP-based design framework is proposed for compound distillation sequences (CDS). • Exhaustive method and elitist GA are used to optimize the SDC and DWC respectively. • ANN models are used to predict the steady state characteristics of the SDC and DWC. • N-butane/n-pentane/benzene/toluene/o-xylene/2,4-methyl-benzene separation is studied. • The CDSs achieve smaller TAC than the direct and indirect SDC sequences. For the separation of multi-component mixtures, compound distillation sequences (CDSs) comprising simple distillation columns (SDCs) and dividing-wall columns (DWCs) offer the advantage of significant energy-saving potential and relatively straightforward industrial implementation. This work proposes a design framework to facilitate the development of optimal CDSs. In this framework, genetic programming is employed to explore the optimal sequence; simultaneously, the exhaustive search method and the elitist genetic algorithm are separately employed to search for the optimal design of, respectively, SDCs and DWCs involved. Furthermore, to efficiently estimate the total annual cost (TAC) and both the flow rates and compositions of products of the SDCs and DWCs during the optimization process, artificial neural network (ANN)-based performance estimators were developed for them. An illustrative case study on n-butane/n-pentane/benzene/toluene/o-xylene/2,4-methyl-benzene mixtures is conducted. Analysis of the optimization results demonstrates a significant improvement in economic performance for the resulting CDS compared to the direct and indirect SDC sequences. [ABSTRACT FROM AUTHOR]

• Filling model of zero-dimensional gas and one-dimensional wall is built and verified. • Modified formula and artificial neural networks (ANN) models for MC are proposed. • Latin hypercube sampling, Genetic algorithm and Sobol sensitivity analysis are used. • Modified formula and ANN models for MC parameter improve MC's calculation accuracy. • ANN model of MC extends applicability to situations of filling time less than 30s. The MC method was officially released in SAE J2601 protocol in 2016 to refuel light-duty hydrogen fuel cell vehicles quickly and safely. For the non-communication MC method, the pressure target for controlling the end of filling is calculated based on the formula of the final hydrogen temperature, which contains a key parameter, MC. SAE has proposed two formulas to estimate the MC parameter. This article proposes a modified formula and artificial neural networks (ANN) model for MC parameter. To generate simulated data for fitting the modified formula and training the ANN model, a hydrogen filling model (0D1D) with a zero-dimensional gas and a one-dimensional tank wall is established and thoroughly verified. In establishing the ANN model, Latin hypercube sampling (LHS) is used to sample independent variables randomly, and genetic algorithms (GA) are used to optimize the initial weights and biases of the ANN. The Sobol sensitivity analysis is used to determine the importance of various initial conditions on the MC parameter. Research shows that the final fueling time and initial pressure have a greater impact on the MC parameter , while the ambient and precooling temperatures have little impact. The modified formula of the MC parameter improves R2 from the original 0.76073 to 0.98132 when fitting the MC simulated data. In situations of short filling time, the modified formula and ANN models of the MC parameter reduce the relative error between the 0D1D model by 13.8 % − 58.8 % and 16.6 % − 82.3 %, respectively, thus improving the calculation accuracy of the final hydrogen temperature and ending pressure target. Moreover, the ANN model can predict the MC parameter during the whole fueling process, while the MC formula utilizes the fixed value when the fueling time is less than 30 s. This work has positive significance for improving refuelling protocol. [ABSTRACT FROM AUTHOR]

Failure assessment is one of the fundamental tasks for optimization of composite pressure vessels (CPVs). However, the extensive design space of composites usually leads to costly and repetitive work of failure assessment that hinders the design and optimization of CPVs. In this work, a combined FEM simulation and machine learning approach is used to predict the failure factors R of CPVs for further optimization. CPVs with various design parameters are automatically generated and analyzed in ABAQUS to obtain failure factors under various external loadings. Then, an assembled deep neural network (DNN) namely Multi-DNN is trained to map relations between design parameters and R of CPVs. Further, transfer learning (TL) is introduced to improve the extensibility of our approach by fine-tuning the pre-trained Multi-DNN using scarce data of new design space. Such a TL-Multi-DNN model can precisely predict R of CPVs with new design parameters. Additionally, combining the trained models and genetic algorithm, the optimization of CPVs is carried out with high efficiency and low computational cost. As a result, R of CPV with various given and new design parameters can be estimated directly by the trained models and the optimal layer sequence of CPV can be obtained efficiently. [ABSTRACT FROM AUTHOR]

The ejectors are crucial in enhancing the efficiency of the CO 2 -based refrigeration cycle. Although the ejector can be optimized by experimental and computational fluid dynamic simulation to improve its performance, these methods are time-consuming and complicated. This research aims to present a fast and systematic approach for optimizing CO 2 ejector by combining variance analysis, surrogate model, and non-dominated sorting genetic algorithm. Firstly, the key geometric structures affecting the performance of ejector were determined by variance analysis. Secondly, the influence of key geometric structures and boundary conditions on ejector efficiency and entrainment ratio was analyzed by using computational fluid dynamic simulation, and the database was generated. Next, the artificial neural network surrogate model was established to forecast the ejector performance. Finally, the non-dominated sorting genetic algorithm was involved to optimize the ejector for maximizing ejector efficiency and entrainment ratio. The results show that the efficiencies of the optimized ejectors are above 42%, and the entrainment ratios exceed 0.7. Compared with the basic model, the average efficiencies have increased about 13.20%. When the pressure lift is 0.7 MPa, the entrainment ratios of the optimized ejectors are 0.696 (−20 °C ≤ T e < −15 °C), 0.82(-15 °C ≤ T e < −5 °C), 0.60 (−5 °C ≤ T e < 5 °C) and 0.90 (5 °C ≤ T e < 15 °C), respectively. The above research demonstrates that the success of this approach in addressing time-consuming multi-optimization problems. It offers a fast and systematic approach for the multi-objective optimization of CO 2 ejector as a valuable reference for engineering applications. • Homogeneous relaxation model is established to simulate the CO 2 ejector. • The key geometric structures are determined by variance analysis method. • A fast and systematic multi-objective optimization framework is proposed. • The ejector efficiency and entrainment ratio are optimized simultaneously. • The boundary conditions of the ejector are considered as independent variable. [ABSTRACT FROM AUTHOR]