Two-layer Model Research Articles

Diminishing Batch Normalization.

In this article, we propose a generalization of the batch normalization (BN) algorithm, diminishing BN (DBN), where we update the BN parameters in a diminishing moving average way. BN is very effective in accelerating the convergence of a neural network training phase that it has become a common practice. Our proposed DBN algorithm retains the overall structure of the original BN algorithm while introducing a weighted averaging update to some trainable parameters. We provide an analysis of the convergence of the DBN algorithm that converges to a stationary point with respect to the trainable parameters. Our analysis can be easily generalized to the original BN algorithm by setting some parameters to constant. To the best of our knowledge, this analysis is the first of its kind for convergence with BN. We analyze a two-layer model with arbitrary activation functions. Common activation functions, such as ReLU and any smooth activation functions, meet our assumptions. In the numerical experiments, we test the proposed algorithm on complex modern CNN models with stochastic gradients (SGs) and ReLU activation on regression, classification, and image reconstruction tasks. We observe that DBN outperforms the original BN algorithm and benchmark layer normalization (LN) on the MNIST, NI, CIFAR-10, CIFAR-100, and Caltech-UCSD Birds-200-2011 datasets with modern complex CNN models such as Resnet-18 and typical FNN models.

Novel Advanced Artificial Neural Network-Based Online Stator and Rotor Resistance Estimator for Vector-Controlled Speed Sensorless Induction Motor Drives

This paper presents a new approach for the online estimation of stator and rotor resistance of induction motors for speed sensorless vector-controlled drives, using feed-forward artificial neural networks with advanced adaptive learning rates. For the rotor resistance estimation, a neural network model based on rotor speed and stator currents is developed. The rotor flux linkages acquired from the voltage model are compared with the neural network model. The feed-forward neural network employs an adaptive learning rate as the function of the obtained error during training for quick convergence with minimal estimation error. A two-layered neural network model based on the stator voltage and current equations is developed for the stator resistance estimation. The d-q axes stator currents obtained from the developed model are compared with the acquired d-q axes stator currents. For the fast convergence with minimal estimation error, an adaptive learning rate as the function of error is adopted during training. Furthermore, the neural network estimates the induction motor’s speed. The simulation and experimental results justify that the developed algorithms track variation in the resistances quickly and precisely along with the speed as compared with the conventional constant learning rate algorithm, leading to reliable operation of the drive.

Crash severity analysis: A data-enhanced double layer stacking model using semantic understanding

The crash severity analysis is of significant importance in traffic crash prevention and emergency resource allocation. A range of innovations offers potential traffic crash severity prediction models to improve road safety. However, the semantic information inherent in traffic crash data, which is crucial in enabling a deeper understanding of its underlying factors and impacts, has yet to be fully utilized. Moreover, traffic crash data are commonly characterized by a small sample size, which leads to sample imbalance problem resulting in prediction performance decline. To tackle these problems, we propose a semantic understanding-based data-enhanced double-layer stacking model, named EnLKtreeGBDT, for crash severity prediction. Specifically, to fully leverage the inherent semantic information within traffic crash data and analyze the factors influencing crashes, we design a semantic enhancement module for multi-dimensional feature extraction. This module aims to enhance the understanding of crash semantics and improve prediction accuracy. Then we introduce a data enhancement module that utilizes data denoising and migration techniques to address the challenge of data imbalance, reducing the prediction model's dependence on large sample crash data. Furthermore, we construct a two-layer stacking model that combines multiple linear and nonlinear classifiers. This model is designed to augment the capability of learning linear and nonlinear mixed relationships, thereby improving the accuracy of predicting the severity of crashes on complex urban roads. Experiments on historical datasets of UK road safety crashes validate the effectiveness of the proposed model, and superior performance of prediction precision is achieved compared with the state-of-the-arts. The ablation experiments on both semantic and data enhancement modules further confirm the indispensability of each module in the proposed model.

Two-Layer Optimization Method for Sharing Energy Storage and Energy considering Subjectivity

The high level of integration of distributed generation systems (DGSs), especially distributed wind and solar, significantly affects the flexibility and controllability of the power system. Aggregating local DGSs and shared energy storage systems (ESSs) within an energy community offers an economically and environmentally viable solution. However, the coupling of shared ESSs with the energy community, while considering subjectivity, is often overlooked. Therefore, this study introduces a two-layer optimization framework that enables DGSs to trade energy freely, voluntarily, and independently and to share ESSs within the energy community, considering participants’ subjectivity. The upper layer optimizes the size of shared ESSs, while the lower layer, structured as a two-layer model, simulates participant interactions. The numerical case shows that, compared to DGSs operating individually, the shared ESS case indicates that community self-sufficiency and self-consumption rates increase by 16.22% and 21.98%, respectively. Additionally, the annual operating cost is reduced by approximately 27.10%, and CO2 emissions are decreased by about 33.24%. Considering DGS’ subjectivity, the self-sufficiency and self-consumption rates are 3.04% lower, and the total operating costs and CO2 emissions are 3.26% and 6.86% higher, respectively.

Evaluation and development of surface layer scheme representation of temperature inversions over boreal forests in Arctic wintertime conditions

Abstract. In this study, the Noah land surface model used in conjunction with the Mellor–Yamada–Janjić surface layer scheme (hereafter, Noah-MYJ) and the Noah multiphysics scheme (Noah-MP) from the Weather Research and Forecasting (WRF) 4.5.1 mesoscale model are evaluated with regard to their performance in reproducing positive temperature gradients over forested areas in the Arctic winter. First, simplified versions of the WRF schemes, recoded in Python, are compared with conceptual models of the surface layer in order to gain insight into the dependence of the temperature gradient on the wind speed at the top of the surface layer. It is shown that the WRF schemes place strong limits on the turbulent collapse, leading to lower surface temperature gradient at low wind speeds than in the conceptual models. We implemented modifications to the WRF schemes to correct this effect. The original and modified versions of Noah-MYJ and Noah-MP are then evaluated compared to long-term measurements at the Ameriflux Poker Flat Research Range, a forest site in interior Alaska. Noah-MP is found to perform better than Noah-MYJ because the former is a two-layer model which explicitly takes into account the effect of the forest canopy. Indeed, a non-negligible temperature gradient is maintained below the canopy at high wind speeds, leading to overall larger gradients than in the absence of vegetation. Furthermore, the modified versions are found to perform better than the original versions of each scheme because they better reproduce strong temperature gradients at low wind speeds.

Optimized Battery Capacity Allocation Method for Wind Farms with Dual Operating Conditions

In order to solve the problems of wind power output volatility and wind power participation in frequency regulation, a method for optimizing the capacity allocation of wind farm storage batteries based on the dual grouping strategy and considering the simultaneous execution of the dual conditions of energy storage in fluctuation smoothing and primary frequency regulation is proposed. Firstly, a two-layer model is established to optimize the capacity allocation under dual operating conditions, i.e., the planning layer takes into account the lifetime, cost, and benefit, and the operation layer considers the wind turbine reserve backup and storage control to participate in the primary frequency regulation in a cooperative manner. Then, the dual battery pack operation strategy is embedded with the variational modal decomposition method to determine the charging and discharging operation strategy of energy storage after considering the grid-optimized reference power. An improved particle swarm algorithm with inverse learning pre-optimization combined with variational crossover post-optimization is embedded in the GUROBI computation to obtain the optimal battery storage capacity allocation scheme. Finally, the superiority of the model proposed in this paper in terms of improving energy storage utilization, service life, and economic efficiency as well as reducing wind power load shedding is verified by comparing it with a single execution working condition scenario and traditional battery control strategy.

Multi-frequency spatial frequency domain imaging: a depth-resolved optical scattering model to isolate scattering contrast in thin layers of skin.

Current methods for wound healing assessment rely on visual inspection, which gives qualitative information. Optical methods allow for quantitative non-invasive measurements of optical properties relevant to wound healing. Spatial frequency domain imaging (SFDI) measures the absorption and reduced scattering coefficients of tissue. Typically, SFDI assumes homogeneous tissue; however, layered structures are present in skin. We evaluate a multi-frequency approach to process SFDI data that estimates depth-specific scattering over differing penetration depths. Multi-layer phantoms were manufactured to mimic wound healing scattering contrast in depth. An SFDI device imaged these phantoms and data were processed according to our multi-frequency approach. The depth sensitive data were then compared with a two-layer scattering model based on light fluence. The measured scattering from the phantoms changed with spatial frequency as our two-layer model predicted. The performance of two models solutions for SFDI was consistently better than the standard diffusion approximation. We presented an approach to process SFDI data that returns depth-resolved scattering contrast. This method allows for the implementation of layered optical models that more accurately represent physiologic parameters in thin tissue structures as in wound healing.

Optimization of Berth-Tug Co-Scheduling in Container Terminals under Dual-Carbon Contexts

In order to address the dynamic changes in vessel preferences for berth lines caused by the deployment of shore-based power equipment in major ports and the collaborative scheduling problem of berthing and towing assistance, this paper quantifies the environmental costs of pollutants from the main engines of tugs and auxiliary engines of container ships using an environmental tax. Additionally, considering the economic costs such as vessel delay and shore power cable connection, a two-layer mixed-integer linear programming model is constructed using the task sequence mapping method. This model integrates the allocation of continuous berths at container terminals with coordinated towing scheduling for shore power selection. A solution approach is designed by combining the commercial solver (CPLEX) and the immune particle swarm optimization algorithm (IAPSO). The proposed scheme is validated using the example of the Nansha Phase IV Terminal at the Port of Guangzhou. The results show that compared to the traditional first-come-first-served and adjacent scheduling schemes, the collaborative scheduling scheme proposed in this paper reduces the total cost by 21.73%. By effectively utilizing berth resources and shore power equipment while densely arranging collaborative tasks and appropriately increasing the number of tugs, the port can convert the economic cost of leasing a small number of tugs (increased by 10.63%) into environmental benefits (decreased by 33.88%). This approach provides a reference for addressing nearshore pollution emissions in ports.

Assessing the risk and effect of climate change with two-layer fuzzy logic-SWARA: A comparative practice in Turkiye

Many factors such as population growth, development of industry/technology, and increase in production-consumption disrupt the ecological balance and cause climate change, which is a global problem. Determining the criteria that cause climate change is very important in finding effective solutions to the problem. In the study, the criteria were determined, weighted with a new method, Step-wise Weight Assessment Ratio Analysis (SWARA), and ranked according to their priorities with two-layer fuzzy logic model. The Fuzzy SWARA method allows the evaluation process, which becomes complicated due to the difficulties and factors experienced in decision-making, to be carried out more effectively and realistically. The risk and effect of climate change in Turkiye were evaluated regionally. However, the developed model also has a wide application area. Research findings revealed that the highest risk/effect of climate change have the Marmara and Central Anatolia regions. The lowest risk region is the Eastern Anatolia. Air pollution, population growth and deforestation have the highest weights. Important suggestions have presented especially for priority criteria. In this way, the factors that should be prioritized in climate change environmental problem solutions have been revealed and will make it easier for researchers and managers to provide more effective management.

Employee Turnover Prediction Based on Ensemble Learning DGNK Model

Employee turnover is a problem that can have significant negative impacts on an enterprise. It not only results in the loss of valuable talent and knowledge but also incurs substantial costs in terms of hiring, onboarding, and training new employees. Therefore, predicting employee intent to quit can be crucial for organizations to take proactive measures to prevent it from happening. Early detection of employee turnover intention will help enterprise develop and enhance core competitiveness. This study aims to predict the employee intension to quit. In the present study, more than 1,400 samples containing 31 features of a company’s employees were collected from Kaggle website as data sets. A two-layer DGNK model was designed with decision tree, gradient boosting, naive bayes and k-nearest neighbor model as the primary classifier and gradient boosting as the secondary classifier to build the predictive model of employee turnover intention. The experimental outcomes show that DGNK model based on two-layer ensemble learning has the best outcome, while naive bayes model has the worst outcome. In conclusion, this study highlights the importance of predicting employee turnover intention as an effective strategy to enhance organizational performance and competitive advantage. Furthermore, the success achieved in the study suggests that machine learning models like DGNK can play a crucial role in achieving this goal.

Multiparameter shallow-seismic waveform inversion based on the Jensen–Shannon divergence

SUMMARY Seismic full-waveform inversion (FWI) or waveform inversion (WI) has gained extensive attention as a cutting-edge imaging method, which is expected to reveal the high-resolution images of complex geological structures. In this paper, we regard each 1-D signal in the inversion system as a 1-D probability distribution, then use the Jensen–Shannon divergence from information theory to measure the discrepancy between the predicted and observed signals, and finally implement a novel 2-D multiparameter shallow-seismic WI (MSWI). Essentially, the novel approach achieves an implicit weighting along the time-axis for each 1-D adjoint source defined by the classical WI (CWI), thus enhancing the extra illumination for a deeper medium compared with the CWI. By evaluating the inversion results of the two-layer model and fault model, the reconstruction accuracy for S-wave velocity and density of the new method is increased by about 30 and 20 per cent compared with that of the CWI under the same conditions, respectively. The reconstruction performance for P-wave velocity of these two methods is almost equal. In addition, the new 2-D MSWI is also resilient to white Gaussian noise in the data. Numerically, the inversion system has almost the strongest sensitivities to the S-wave velocity and density, performing the poorest sensitivity to the P-wave velocity. Finally, we test the novel method with a detection case for a power tunnel.

Robust engine slipping start control of hybrid electric vehicles with uncertainty in clutch slipping torque and change in driver demand torque

The engine slipping start (ESS) benefits parallel hybrid electric vehicles from stable ignition and emission reduction. However, inappropriate coordination between the traction motor torque and clutch slipping torque during the ESS will lead to poor smoothness of the vehicle and failed start of the engine. Uncertainty in clutch slipping torque and change in driver demand torque bring tough challenges with sluggish convergence and intensive vehicle jerk in practice. To deal with this problem, a novel two-layer model reference adaptive controller (MRAC) which contains two parallel reference models is proposed to improve robustness and convergence rate simultaneously. On one hand, uncertainties of clutch slipping torque are divided into a low-frequency part and a high-frequency part, and adaptive laws based on the output feedback are designed contrapuntally to enhance robustness. On the other hand, two parallel reference models are designed to accelerate the tracking error convergence rate without changing the reference profiles, which is generated according to the driver demand torque in real time. To test the robustness and convergence rate, the proposed two-layer MRAC is compared with the classical MRAC and proportional–integral controller under the driving scenario with uncertain clutch slipping torque and abrupt change in driver demand torque. The sensitivity with different adaptive gains and low-frequency and high-frequency uncertainties in clutch slipping torque are examined. Finally, hardware-in-the-loop experiments are performed to verify the effectiveness of the proposed two-layer MRAC.

A Bi-Level Optimal Scheduling Strategy for Microgrids for Temperature-Controlled Capacity and Time-Shifted Capacity, Considering Customer Satisfaction

Since microgrids can effectively integrate renewable energy, energy storage devices, and controllable loads, this advantage promotes the rapid development and application of microgrid technology. However, with the high proportion of renewable energy access, only considering how energy is optimally distributed in microgrids can no longer meet the actual demand. How to aggregate user-side controllable loads to form regulation resources has become a research hotspot, and the users, as a passive party in the load scheduling process, should also be an important consideration in their perception of the use of electricity. First, a control model for temperature-controlled loads and a time-shift model for time-shiftable loads are developed. Then, the comprehensive electricity satisfaction model of users is established, and the two-layer optimal scheduling model of microgrids considering users’ satisfaction is proposed, with users as the upper layer and microgrids as the lower layer, and the two-layer model is transformed into a single-layer model according to the KKT condition for solving. Finally, the effect of the weighting factor for satisfaction on the economy is discussed through the analysis of examples, which verifies the effectiveness of the two-layer model.

Optimised operation of integrated community energy system considering integrated energy pricing strategy: A two-layer Stackelberg game approach

Aiming at the problems of new energy absorption and inter-subject interest conflict in the integrated community energy system, this study proposed an integrated energy pricing strategy that considered the interconversion of various energy sources and built a two-layer optimisation model of the electric-gas-heat‑hydrogen interconnection integrated community energy system. The upper-layer operator model maximises the revenue by setting energy price and selling energy to users. The lower-layer users model minimises the energy cost by actively adjusting energy usage strategy according to energy price. The two sides rationally pursue the maximisation of their respective interests, and the process of bargaining can be described by Stackelberg game. The model is also considered for combining the characteristics of elastic load and vehicle-to-grid on the demand side and operates jointly with the bus battery swapping station and two-stage power-to-gas for optimizing system operation. Through theoretical derivation, it is proved that there is a unique Stackelberg equilibrium solution for the proposed game model. Further, the two-layer optimisation model is transformed into a mixed integer quadratic programming problem by applying the Karush-Kuhn-Tucher optimal condition, linear relaxation technique, and duality theorem. Finally, the global optimal energy price is obtained by solving the converted model. The optimisation results in different scenarios show that the proposed model and method not only protect the interests of the operator and users but also improve the wind power absorption capacity of the community and reduce the load fluctuation.

Developing a smart system for infant incubators using the internet of things and artificial intelligence

This research is developing an incubator system that integrates the internet of things and artificial intelligence to improve care for premature babies. The system workflow starts with sensors that collect data from the incubator. Then, the data is sent in real-time to the internet of things (IoT) broker eclipse mosquito using the message queue telemetry transport (MQTT) protocol version 5.0. After that, the data is stored in a database for analysis using the long short-term memory network (LSTM) method and displayed in a web application using an application programming interface (API) service. Furthermore, the experimental results produce as many as 2,880 rows of data stored in the database. The correlation coefficient between the target attribute and other attributes ranges from 0.23 to 0.48. Next, several experiments were conducted to evaluate the model-predicted value on the test data. The best results are obtained using a two-layer LSTM configuration model, each with 60 neurons and a lookback setting 6. This model produces an R<sup>2</sup> value of 0.934, with a root mean square error (RMSE) value of 0.015 and a mean absolute error (MAE) of 0.008. In addition, the R<sup>2</sup> value was also evaluated for each attribute used as input, with a result of values between 0.590 and 0.845.

A multilayer network model of interaction between rumor propagation and media influence.

Rumors spread among the crowd have an impact on media influence, while media influence also has an impact on rumor dissemination. This article constructs a two-layer rumor media interaction network model, in which the rumors spread in the crowd are described using the susceptibility-apathy-propagation-recovery model, and the media influence is described using the corresponding flow model. The rationality of the model is studied, and then a detailed analysis of the model is conducted. In the simulation section, we undertake a sensitivity analysis of the crucial parameters within our model, focusing particularly on their impact on the basic reproduction number. According to data simulation analysis, the following conclusion can be drawn: First, when the media unilaterally influences the crowd and does not accept feedback from the crowd, the influence of the media will decrease to zero over time, which has a negative effect on the spread of rumors among the crowd (the degree of rumor dissemination decreases). Second, when the media does not affect the audience and accepts feedback from the audience, this state is similar to the media collecting information stage, which is to accept rumors from the audience but temporarily not disclose their thoughts. At this time, both the media influence and the spread of rumors in the audience will decrease. Finally, the model is validated using an actual dataset of rumors. The simulation results show an R-squared value of 0.9606, indicating that the proposed model can accurately simulate rumor propagation in real social networks.

Collaborative hierarchical scheduling model of interconnected multi-microgrid and ADN considering DR with different strategies

The demand response (DR) resources of multiple entities have different operation strategies, which makes it difficult to realize the co-optimization. This paper proposes a cooperative hierarchical scheduling optimization model of interconnected multi-microgrid and an active distribution network (ADN) considering DR with different operation strategies. First, the collaborative hierarchical scheduling framework is proposed and the interaction characteristics and cooperative scheduling mode of multi-interconnected microgrid and ADN are analyzed. Second, the two-layer cooperative hierarchical scheduling model is established, considering DR with different strategies. The upper layer model takes the minimum operating cost of the ADN as the objective to optimize the trading tariff between the ADN and microgrids, and its own DR, while the lower layer model takes the operating cost of multi-microgrid as the objective to optimize the purchasing and selling of electrical energy for the ADN, the DR, and the interaction power. Finally, the case studies with three microgrids and an ADN are used to demonstrate the validity and effectiveness of the proposed model.

Train-earth-grid coupling DC magnetic bias model for stray current diffusion

The DC magnetic bias phenomenon in transformers can result in increased noise and harmonic components. This study focuses on the DC magnetic bias generated by stray currents in the DC traction power supply system of neutral grounding transformers under multiple train operation conditions. A two-layer soil model is utilized to calculate geodetic potential, with rail potentials from both mainline and depot serving as sources for stray current within the model. Two paths of stray current intrusion into the transformer via the grounding network and cable armor layer from the Main substation have been modeled and analyzed. A simulation test has been designed to verify that the maximum error between experimental data and model calculation data is 8.87%. The implementation of a blocking connection device between the depot and mainline, as well as a change in grounding strategy for the metal protective layer, can effectively reduce DC bias magnetic current in the substation.

Asphaltene Adsorption on Solid Surfaces Investigated Using Quartz Crystal Microbalance with Dissipation under Flow Conditions.

Asphaltenes can cause operational challenges in petroleum production facilities and adversely affect production by adsorption on mineral surfaces and alteration of the oil wettability of reservoirs. Therefore, understanding asphaltene adsorption mechanisms and their effects is crucial to improving the efficiency of oil production and reducing costs. In this study, we focus on understanding the impact of asphaltene concentration and the depositing environment of asphaltene adsorption on solid surfaces using the quartz crystal microbalance with dissipation (QCM-D) technique. The initial and long-term kinetics of adsorption at different concentrations were examined on three different solid surfaces including silicon dioxide to represent quartz mineral, stainless steel, and gold. The frequency-dissipation data showed evidence of monolayer adsorption initially, followed by multilayer formation. At short times, the adsorbed mass increased linearly with time, suggesting that the process was kinetically controlled rather than diffusion-controlled. The results were reproducible and did not depend on convection velocity but did depend on the surface material. At later stages, the monolayer development appeared to follow the random sequential adsorption (RSA) theory. Once multilayer adsorption commenced, the rates agreed well with the two-layer model of Zhu and Gu, 1990. The impact of asphaltene adsorption on the wettability of the surface was examined using contact angle studies, which showed decreasing water wettability with an increase in the adsorbed mass. The contact angle of water after 12 h of adsorption leveled off at around 100° on all three surfaces. Contact angle measurements were also used to evaluate if brine salinity causes the wettability alteration of surfaces with the adsorbed asphaltene. The results indicate that at 3% NaCl solution, the contact angle decreased only slightly by less than 2°.

Energy and economic analysis of building integrated photovoltaic thermal system: Seasonal dynamic modeling assisted with machine learning-aided method and multi-objective genetic optimization

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.