Optimal Model Structure Research Articles

Improving canopy transpiration model performance by considering concurrent hot and dry conditions

CONTEXTThe Jarvis-Stewart model has commonly been used to estimate transpiration. However, its assumption regarding independent effects from different influencing factors has seldom been explored. OBJECTIVEThe aim is to test whether the Jarvis-Stewart model can well estimate transpiration under concurrent hot and dry conditions with strong environmental interactions. A secondary objective was to test was to propose a new coupling method without the environmental-independence assumption for transpiration estimation. METHODSThis study tested and compared 24 Jarvis-Stewart models composed of various constraint functions to determine the optimal model structure for an orange orchard under concurrent hot and dry conditions. Meanwhile, a new coupling equation that considers interactive environmental effects on transpiration was proposed, and a total of 48 configurations were examined to determine the most effective configuration. A comprehensive comparison was made between the best Jarvis-Stewart model and the best coupling model based on a Bayesian framework. RESULTS AND CONCLUSIONSResults showed that with fewer parameters, the best new coupling model significantly improved model performance compared with the best Jarvis-Stewart model. Specifically, estimation accuracy was slightly improved, with the average mean relative error decreased from 11.79% to 11.06%, and the average coefficient of determination increased from 0.67 to 0.72. More importantly, estimation uncertainty was significantly decreased, with the average uncertainty band width reduction of 42%. Moreover, the new model generated some water use information that was much more consistent with measured data and did not suffer from over-fitting problems as the Jarvis-Stewart model suffered from. Furthermore, the new model did not require additional data or computational cost. SIGNIFICANCEThe results of our study emphasize the necessity for studying environmental interaction effects on plant water consumption, particularly under the possibility of concurrent hot and dry conditions that are likely to occur under future climate conditions.

A deep neural network-based method to predict J-integral for surface cracked plates under biaxial loading

As surface cracks may cause potential failure for engineering structures, it is of vital significance to carry out the fracture assessment for cracked structures. The crack driving force in the form of J-integral is an important input parameter in fracture assessment. This paper focuses on the determination of J-integral for surface cracked plates under biaxial loading. A method for predicting the J-integral based on the deep neural network (DNN) is proposed to estimate J-integral efficiently and accurately. Firstly, extensive three-dimensional (3D) elastic–plastic finite element (FE) analysis is conducted to compute the J-integral along the crack front for developing a FE J-integral datasets containing 1600 cases. Various crack aspect ratios, crack depth ratios, strain hardening exponents, ratios of the loading perpendicular to the crack plane to yield stress, and biaxial ratios, are strategically considered. Secondly, DNN models for predicting the J-integral are constructed and trained according to the obtained datasets by FE analysis. Then, the evaluation criteria for DNN models and the grid search method are utilized to determine the optimal DNN model structure. Finally, test cases are employed to verify the feasibility and prediction accuracy of the determined DNN model. Based on validation results, the determined DNN model exhibits sufficiently accurate prediction performance. This research provides an idea for engineering applications, through building software based on continuously optimized DNN models, the J-integral for surface cracked plates under biaxial loading can be accurately and efficiently predicted for performing fracture assessments.

Intelligent extraction of medical entity relationship based on graph neural network and optimization strategy

Machine Learning technologies have obtained breakthrough in various fields, and medical information extraction has also successfully made great progress through deep learning methods. However, entity relationship overlap is a key issue and challenge in the field of medical entity relationship extraction. Hence, we propose an intelligent extraction method for Chinese medicine entity relationship through improving graph neural network and structure optimization strategy. We optimize model structure and introduce global pointer network. The sentence feature information is captured by attention mechanism and Bi-Long Short Term Memory (LSTM). Grammatical relations between entities are parsed using Graph Convolution Network (GCN) layer. Finally, multiple decoders are used for joint extraction of entity relations through the global pointer network. The model learns the graph structure data by GCN, while the challenge of entity relationship overlap is solved through using the pointer network. The experiments demonstrate that the model achieves 83.77% in terms of F1 value, which is better than other baseline models.

Cross-phase automated structural design of modular buildings using a unified matrix method and multi-objective optimization

Traditional structural design of modular buildings involves multi-phase work, which is a labor-intensive process with low efficiency. The tasks in each stage are independent and fragmented. To address such problem, a cross-phase automated structural design of modular buildings is proposed. From the computer-aided design plain drawing, the building elements are extracted to conduct the sematic segmentation of zones using the layer classification method and graph structure. Based on a unified matrix method, the categorized modules are arranged to meet the architectural functions and the structural analysis model is established. Then a multi-objective optimization method is applied to perform the structural design considering the safety demand and economic benefit of the structure. Lastly, the data file is generated to conduct the structure-to-digital model delivery. The proposed method is shown to be effective and fully automated through two practical cases. The proposed framework integrates module layout, parametric modeling, structural optimization and digital model generation as a continues procedure, which also realizes the efficient transfer of different data items including architectural plan, structural model and digital model. It provides a guideline to automated structural design of modular buildings.

Optimization path of China’s energy industry structure under low carbon economy situation

The study discusses the main aspects of optimizing the structure of China’s energy industry in a situation with a low-carbon economy. To build a model for forecasting electricity demand, the method of partial least squares regression is used. The basic scenario and the scenario with restrictions are set taking into account the peculiarities of the development of a new normal economy. Based on the baseline scenario and the restricted scenario, the total energy demand, energy consumption structure and CO2 emissions in China are projected. Taking into account energy, economic and environmental factors, a multi-purpose optimization model of the energy consumption structure was built and the structure of China’s energy consumption and the corresponding CO2 emissions under optimization scenarios were obtained. This research describes revise the energy consumption structure in China, should reducing energy consumption and carbon dioxide emissions is very helpful.

Multi-objective structural optimization and degradation model of magnesium alloy ureteral stent

BackgroundMg alloys have attractive properties, including biocompatibility, biodegradability, and ideal mechanical properties. Moreover, Mg alloys are regarded as one of the promising candidates for manufacturing ureteral stents. This study proposed a multi-objective optimization method based on the Kriging surrogate model, NSGA-Ⅲ, and finite element analysis to improve the degradation performance of Mg alloy ureteral stents. MethodsThe finite element model for the degradation of Mg alloy ureteral stents has been established to compare the degradation performance of the stents under different parameters. Latin hypercube sampling was adopted to generate train sample points in the design space. Meanwhile, the Kriging surrogate model was constructed between strut parameters and stent degradation behavior. The NSGA-Ⅲ was utilized to determine the optimal solution in the global design space. ResultsThe optimized stent achieved 5.52 ​× ​degradation uniformity (M), 10 ​× ​degradation time (DT), and 4 ​× ​work time (FT). The errors between the Kriging surrogate model and the finite element calculation results were less than 6%. ConclusionThe optimized stent achieved better degradation performance. The degradation behavior of stents was dependent on the design parameters. The multi-objective optimization method based on the Kriging surrogate model and finite element analysis was effective in stent design optimization problems.

Accurate detection and intelligent classification of solar cells defects based on photoluminescence images: A novel study on the optimized YOLOv5 model

In the production process of solar cells, inevitable faults such as cracks, dirt, dark spots, and scratches may occur, which could potentially impact the lifespan and power generation efficiency of solar cells. Addressing this issue, this paper combines neural networks with photoluminescence detection technology and proposes a novel neural network model for the classification and grading of defects in solar cells. Firstly, the YOLOv5 model is optimized and adjusted for algorithm and network structure. The optimization process is divided into three parts: global optimization of the network structure, optimization of the neck network structure, and optimization of the head structure, each addressing specific issues in recognition, detection, and classification. The impact of the optimized network model on recognition and detection speed is analyzed, and solutions are proposed to address any observed effects. Additionally, an iterative update of neural network hyperparameter combinations is performed for solar cell defect identification. Finally, using the ultimately optimized model structure in conjunction with the optimal hyperparameter combination, comparative experiments are conducted on neural networks for different target identification using the photoluminescence characteristics dataset of solar cells. The recognition improvement of the optimized model and its differences from other models are analyzed.

Simulation Analysis of Frog-Inspired Take-Off Performance Based on Different Structural Models.

The frog-inspired jumping robot is an interesting topic in the field of biomechanics and bionics. However, due to the frog's explosive movement and large range of joint motion, it is very difficult to make their structure completely bionic. To obtain the optimal jumping motion model, the musculoskeletal structure, jumping movement mechanism, and characteristics of frogs are first systematically analyzed, and the corresponding structural and kinematic parameters are obtained. Based on biological characteristics, a model of the articular bone structure is created, which can fully describe the features of frog movement. According to the various factors affecting the frog's jumping movement, mass and constraints are added, and the complex biological joint structure is simplified into four different jumping structure models. The jumping ground reaction force, velocity, and displacement of the center of mass, joint torque, and other motion information of these four models are obtained through ADAMS simulation to reveal the jumping movement mechanism and the influencing factors of frogs. Finally, various motion features are analyzed and compared to determine the optimal structural model of the comprehensive index, which provides a theoretical basis for the design of the frog-inspired jumping robot.

Why Do We Not Wear Masks Anymore during the COVID-19 Wave? Vaccination Precludes the Adoption of Personal Non-Pharmaceutical Interventions: A Quantitative Study of Taiwanese Residents.

Background and Objectives: This study examined whether the decline in people's adoption of personal NPIs (e.g., mask wearing) results from the preclusion by vaccination. This study also incorporates the concepts of risk perception and the risk-as-feelings model to elucidate the possible mechanisms behind this preclusion. Materials and Methods: Two cross-sectional surveys (N = 462 in Survey 1 and N = 505 in Survey 2) were administered before and during the first outbreak of COVID-19 in Taiwan. The survey items were designed to measure participants' perceived severity of COVID-19, worry about COVID-19, intention to adopt personal NPIs, and attitudes toward COVID-19 vaccines. Utilizing the risk perception framework, we conducted multigroup SEM (Structural Equation Modeling) to construct the optimal structural model for both samples. Results and Conclusions: The multigroup SEM results showed that worry (i.e., the emotional component of risk perception) fully mediates the influence of the perceived severity of COVID-19 (i.e., the cognitive component of risk perception) on the intention to adopt NPIs in both surveys [z = 4.03, p < 0.001 for Survey 1 and z = 2.49, p < 0.050 for Survey 2]. Before the outbreak (i.e., Survey 1), people's attitudes toward COVID-19 vaccines showed no significant association with their worry about COVID-19 [z = 0.66, p = 0.508]. However, in Survey 2, following the real outbreak of COVID-19, people's attitudes toward COVID-19 vaccines negatively predicts their worry about COVID-19 [z = -4.31, p < 0.001], indirectly resulting in a negative effect on their intention to adopt personal NPIs. This suggests the occurrence of the Peltzman effect. That is, vaccination fosters a sense of safety, subsequently diminishing alertness to COVID-19, and thus reducing the intention to adopt personal NPIs.

Review: Recent advances for the diffusion model

As the generative model technology becomes more and more popular, more and more people have invested in the research of the current State-of-the-art (SOTA) generative model-diffusion model. This paper reviews all SOTA generation models using the diffusion model for text-to-image generation since the emergence of the diffusion model, including the denoising diffusion probabilistic model (DDPM), DALL·E model, imagen model, stable diffusion model, and diffusion transformer architecture (DiT) model. In the theoretical section, the basic principles behind the diffusion model are reviewed in detail in the way of mathematical calculation, including the training process of the model and the mathematical principles behind the sampling process. Moreover, this paper focuses on the technical characteristics of these models and various improvements made after model iteration, such as model structure optimization, more efficient and accurate training methods, and the application of other optimization techniques widely used in the field of deep learning to diffusion models. In the end, the technical route of the development of the diffusion model is summarized, and some predictions are made.

Analysis of Microphone Coupling and Zero-Point Calibration Model for Non-Resonant Photoacoustic Spectroscopy

Dissolved Gas Analysis (DGA) of transformer insulation oil is an effective method for monitoring transformer operating conditions and diagnosing faults. Photoacoustic spectroscopy (PAS) is widely employed for online dissolved gas analysis in transformer oil. The positioning of the pressure equalization hole within the sound field significantly impacts the microphone’s sensitivity when coupled with the photoacoustic cell in non-resonant photoacoustic spectroscopy. The static temperature and humidity characteristics inside the photoacoustic cell also have a notable influence on the microphone’s output signal. Consequently, an analysis is performed to explore the relationship between the sensitivity of the photoacoustic spectrometer and temperature/humidity, leading to the optimization of the system configuration based on the microphone parameters. Moreover, the photoacoustic zero-point signal is influenced by the temperature and humidity levels within the photoacoustic cell. To address this problem, air mixtures at different humilities are measured, resulting in the establishment of a zero-point model for the background signal in Photoacoustic Spectroscopy (PAS). Subsequently, the experimental detection of acetylene under various humidity mixtures is conducted, successfully extracting effective signals for the measurement of 5 ppm acetylene gas. These results serve as evidence showcasing the efficacy of both the proposed structural optimization and zero-point model put forth in this study.

VGG16 based on dilated convolution for face recognition

Face recognition has wide applications in fields such as security systems, biometrics, and human-computer interaction. However, traditional face recognition methods face challenges in capturing details and reducing model complexity. To address these issues, this paper proposes a new method based on VGG16, which improves recognition accuracy and reduces parameter quantity by introducing dilated convolution and parameter pruning. First, the hole convolution is introduced to expand the Receptive field and capture more details to enhance the ability of the model in distinguishing facial features. Next, parameter pruning is applied to reduce redundant parameters, optimize model structure, and improve computational efficiency. This article conducted experimental evaluation on the classic face recognition dataset CK+ dataset. The results show that the proposed method is significantly superior to the traditional VGG16 model in terms of recognition accuracy. At the same time, the use of pruning technology significantly reduces the number of parameters in the model and improves computational efficiency. The experimental outcomes conclusively validate the effectiveness and feasibility of the proposed method.

An efficient and accurate deep learning method for tree species classification that integrates depthwise separable convolution and dilated convolution using hyperspectral data

ABSTRACT Addressing accuracy and computational complexity challenges in hyperspectral image classification for small sample and multi-species scenarios, we developed DSC-DC, a lightweight convolutional neural network. This model is based on Depthwise Separable Convolution and Dilated Convolution and was trained using the Teakettle Experimental Forest dataset (USA). In this study, DSC-DC achieved an overall accuracy (OA) of 99.83%, average accuracy (AA) of 99.64%, and Kappa coefficient of 0.9996. Compared to Support Vector Machine and K-Nearest Neighbors, it demonstrated markedly higher OA (3.88% to 7.55%) and AA (30.71% to 34.09%). Compared to Inception-V3, ResNet50, and MSR-3DCNN, DSC-DC marginally outperformed in accuracy (OA: 0.06% to 0.31%; AA: 0.32% to 3.64%) while reducing the training time by 3.5, 5, and 35 times, and the prediction time by 2, 3, and 17 times, respectively. Moreover, DSC-DC exhibits slightly superior accuracy and efficiency compared to a 5-layer optimal structure of the 3D-CNN model. The application of the DSC-DC model to the hyperspectral dataset from the Jiepai branch of the Gaofeng State Owned Forest Farm in the Guangxi province, China, further demonstrated the reliability, versatility, and practical potential of this model. This study provides a reliable and efficient reference solution for small-sample and multi-tree classification tasks.

Updated Chronic Copper Bioavailability Models for Invertebrates and Algae.

Chronic copper (Cu) bioavailability models have been successfully implemented in European risk assessment frameworks and compliance evaluations. However, they were developed almost two decades ago, which calls for an update. In the study, we present updated chronic Cu bioavailability models for invertebrates and algae. They consider recent ecotoxicity data sets and use the more recent speciation model Windermere Humic Aqueous Model (WHAM) VII and an optimized model structure (i.e., a generalized bioavailability model [gBAM]). Contrary to the classic biotic ligand model, a gBAM models the effect of pH on Cu2+ toxicity via a log-linear relationship parametrized through the pH slope SpH . The recalibrated SpH parameters are -0.208 for invertebrates (Daphnia magna, two clones) and -0.975 for algae (Raphidocelis subcapitata and Chlorella vulgaris). The updated models predict 80% to 100% of the observed effect levels for eight different species within a factor of 2. The only exception was one of the two data sets considering subchronic 7-day mortality to Hyalella azteca: the prediction performance of the updated invertebrate model at pH ≥ 8.3 was poor because the effect of pH on Cu2+ toxicity appeared to be dependent on the pH itself (with a steeper pH slope compared with the updated invertebrate model at pH ≥ 8.1). The prediction performance of the updated Cu bioavailability models was similar to or better than that of the models used for regulatory application in Europe until now, with one exception (i.e., H. azteca). Together with the recently published fish bioavailability model, the models developed in the present study constitute a complete, updated, and consistent bioavailability model set. Overall, the updated chronic Cu bioavailability model set is robust and can be used in regulatory applications. The updated bioavailability model set is currently used under the European Union Registration, Evaluation, Authorisation, and Restriction of Chemicals framework regulation to guide the safe use of Cu. Environ Toxicol Chem 2024;43:450-467. © 2023 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

基于线性规划与熵权TOPSIS法的农林牧渔结构最优化模型研究——以山东省为例

基于线性规划与熵权TOPSIS法的农林牧渔结构最优化模型研究——以山东省为例

Internet Smart Technology Drives High Quality Development of Agricultural Economy

Abstract This paper builds a mechanism for the integration of the digital economy and agricultural economy and analyzes the scientific path of agricultural economic development under the mode of “Internet+”. Secondly, on the basis of Internet technology, for the multi-objective optimization problem of agricultural circular economy, the multi-objective optimization problem is solved by selecting decision variables of agricultural circular economy and determining the objective function of agricultural circular economy. Then, from the actual situation of the regional agricultural economy, we determine the constraints, decision variable parameters, and the adaptability function so as to construct the structural optimization model of the agricultural circular economy and carry out an example analysis of the high-quality development of the agricultural economy. The results show that on the model optimization analysis, the optimized planting area of corn is 15,956 hectares, and the optimized numbers of pigs, cattle and poultry are 302, 144, 25, 543 and 579, 1332, respectively. On the model application prediction analysis, it is recommended to increase the support for agriculture to 102.5% (2022=100%) while increasing the support for forestry to 105.1% (2022=100%). This study provides strong support and guidance for the optimization of the agricultural and industrial structure and provides a more scientific decision-making basis for the high-quality development of the agricultural economy.

Elastodynamic modeling and structural error compensation for a 5-DOF parallel mechanism with subclosed loops

In order to reduce the adverse effects of structural errors on a novel 5-degrees of freedom (DOF) parallel kinematic mechanism (PKM) with subclosed loops, studies on the elastodynamic modeling and structural error compensation are conducted in this paper. First, considering that introduction of subclosed loops significantly increased the complexity and difficulty of elastodynamic modeling, an equivalent unit modeling method for the branch chain with subclosed loop is proposed based on force analysis and deformation compatibility condition. Then, elastodynamic model of the whole 5-DOF PKM is established with the matrix structural analysis based method. Second, by defining a six-dimensional error (three translational errors and three rotational errors) for each node, structural error model of the PKM is constructed based on closed-loop vector method. To consider the coupling effects of static and dynamic structural errors, the dynamic structural errors derived by above elastodynamic model would be substituted into the structural error model along with given static structural errors. Finally, structural error compensation optimization model is established and structural errors are compensated by adjusting the driven parameters based on the improved particle swarm optimization (PSO) algorithm. Numerical simulations are conducted to verify the proposed elastodynamic modeling and structural error compensation methods. After compensation, the position errors along x, y, and z directions are respectively decreased by 68.036%, 81.409%, and 90.625%, and the angle errors around x and y directions are respectively decreased by 91.426% and 62.042%.

Thermal performance estimation of a v-corrugated solar air heater using ann techniques

A solar air heater (SAH) is a unique form of solar thermal collector that utilizes solar energy emitted from the sun to produce heated air. Various experimental and theoretical investigations have been undertaken to improve the poor thermal performance of SAHs. The difficulties related to these studies drew attention toward a reliable soft computing technique exemplified by the Artificial Neural Network (ANN) technique. The current work applied actual meteorological data from Miskolc City, Hungary, to an ANN model with the structure of a Multi-layer Perceptron (MLP) to forecast the energy performance of a V-corrugated solar-powered air heater. Seven input parameters and one output parameter make up the ANN structure, with a single hidden layer. For the purpose of selecting the most effective network for predicting output parameters, ten neurons have been assessed. The suggested ANN model was trained with 336 data sets using the Levenberg-Marquardt (LM) learning technique. The comparison of anticipated and real thermal performance values shows a very good agreement. The statistical error analysis showed that the optimal ANN model structure of 7-8-1 can reliably and accurately predict SAH’s thermal performance and thus it can save both time and cost.

Structural optimization model of confined polyhedral composite subsea pipelines under pressure and thermal fields

This paper investigates the structural optimization of a polyhedral composite subsea pipeline (cylinder) under pressure and thermal fields. The pipeline is confined tightly and deforms inward when it is subjected to external loadings. The interface is frictionless between the pipeline and its surrounding medium. Based on the above assumptions, thin-walled shell principles, and an admissible displacement function, the potential energy of a pipeline per unit length is obtained explicitly by simplifying the radius and bending rigidity. After taking the first derivative of the potential energy, two equilibrium equations are obtained. By combining these two equations, the critical buckling pressure of the polyhedral pipeline is expressed analytically with the inclusion of the temperature effects. Then, the present analytical study is compared with other numerical and experimental results, and excellent agreements are reached. A configuration factor is defined as the buckling pressure between the polyhedral and circular pipeline. Finally, parametric studies show the configuration factor decreases with the increase of thickness-to-radius ratio, the increase of the number of sides, and the increase of the temperature variation, respectively. Therefore, a polyhedral pipeline with a low thickness-to-radius ratio is recommended in engineering practices since it may reduce the material cost.

Toward optimized operation of freshwater generator using computer vision, and its economic and environmental benefits

Sufficient quantities of fresh water (FW) must be maintained for the operation of ships and fresh water generators (FWGs), used to produce FW on ships tend to be suboptimal because of work schedules and low-skilled marine crew. Therefore, a system that optimally and automatically operates the discharge valve of an FWG distillate pump by recognizing the FW condition through a sight glass using computer vision technology was developed. Three deep-learning models were trained using videos of FW flowing and filling the sight glass. An optimal model structure was selected in the pre-test phase, which consisted of (i) a test using a camera and a monitor showing the validation videos, and (ii) a test based on data. In the real-test phase, an electrical valve and stepper valve were used as optimization valves, and the flow rates of the fuel oil for the generator engine (G/E) and FW for the FWG were recorded for 10 experimental conditions. The stepper valve operated in a more stable manner than the electrical valve, and its prediction performance with CNN-LSTM was superior. Finally, the economic and environmental impacts were evaluated based on the flow rates of fuel oil for the G/E and FW for the FWG.