Performance Prediction Research Articles

Deep learning-based prediction of float model performance in floatplanes: A case study on lift-to-drag coefficient ratio

Developing an engineering design is resource-intensive and time-consuming, particularly for the floats of a floatplane design, due to its complexity and limited testing facilities. Intelligent-based computational design (IBCD) techniques, which integrate computational design techniques and machine learning (ML) algorithms, offer a solution to reduce required testing by providing predictions. This paper proposes a deep learning (DL)-based IBCD method for modeling floats' lift-to-drag coefficient ratio (C<sub>L</sub>/C<sub>D</sub>), where DL is one of the most powerful ML. The proposed method consists of two phases: hyper-parameter optimization and DL model training and evaluation. A genetic algorithm (GA) is employed in the first phase to explore complex hyper-parameter combinations efficiently. Evaluation of the predicted C<sub>L</sub>/C<sub>D</sub> of the floats using the DL model resulted in a satisfactory R-squared of 0.9329 and the lowest mean squared error (MSE) of 0,001536. These results demonstrate the ability of DL model to predict the float's performance accurately and can facilitate further design optimization. Thus, the proposed method can offer a time-efficient and cost-effective solution for predicting float performance, aiding in optimizing floatplane designs and enhancing their functionalities.

Stacking classifier method for prediction of human body performance

<p>A healthy body is the capital of success and supports human activities. To maintain health, humans need to avoid disease. A healthy life is everyone’s dream and should start early. Busy activities often hinder a healthy lifestyle. Nonetheless, it is important for every individual to lead a healthy lifestyle. Human activities determine health and the implementation of a healthy life. One method that can perform classification with machine learning is extreme gradient boosting (XGBoost). XGBoost is one of the techniques in machine learning for regression analysis and classification based on gradient boosting decision tree (GBDT). By using gradient descent to minimize the error when creating a new model, the algorithm is called gradient boosting. In determining a classification starting from determining the model to the results, usually only using one algorithm method, and combining other methods together with the method is an algorithm called random forest classifier. Among these merging methods are, stacking classifier, voting classifier, and bagging classifier. The conclusion obtained from the results of this research is that the test results show that the stacking classifier achieves the highest accuracy of 76.07%, making it the best method in this research. And the stacking classifier has a precision of 76.96%, recall of 75.83%, and F1-score of 75.81%. This shows that the model has a good balance between the ability to provide true positive results and the ability to recover positive data.</p>

Student Performance Prediction

Abstract: Academic integration is now defined by how students are able to access, and effectively interact with, and are supported by engineering faculty and staff; Academic integration, which is a very important construct that contributes to a student’s persistence and ultimate graduation. But, several factors, either numeric or non-numeric can have varying degrees of influence on academic performance of students. Rankings of aforementioned factors, hopefully, will assist the management in order to initiate corrective action with a view to students' academic performance improvement. In an academic environment, teachers are frequently in close contact with students and therefore their belief about the degree of influence of factors on students’ academic performance could be of vital significance. The case study presented in this paper uses fuzzy relational calculus and the Dempster Shafer theory of evidence to examine and suggest a feasible and realistic ranking approach of only chosen factors!

RANSE simulations for unmanned sailboat performance prediction: Coupling aerodynamics and hydrodynamics in self-sailing

During the design phase of sailboats, naval architects typically employ hydrodynamic and aerodynamic models of the sailboat, utilizing a Velocity Prediction Program (VPP) to solve the equations of motion, thus determining the sailboat's speed and performance. Traditional methods for solving sailboat motion equations often employ simplified empirical formulas or costly experimental procedures, which yield less accurate results. This study presents the RANSE-based CFD software STAR-CCM + to comprehensively model the sailboat's hull-sail-rudder assembly. By implementing overset grid technology and the Volume of Fluid (VOF) model, the paper aims to couple the solution of aerodynamic and hydrodynamic forces under wind influence and introduces PID rudder control technology to maintain a balance of forces and convergence of motion state. This approach enables the prediction of the sailboat's speed and performance, such as yaw angle, heel, and rudder angle. Results demonstrate minimal variations in sailboat speed (<0.05 m/s) and heading (<0.1°) in upwind, downwind and reaching sailing conditions, underscoring the method's feasibility and advocating for further research into self-sailing performance.

Striatal dopamine contributions to skilled motor learning.

Coordinated multi-joint limb and digit movements - "manual dexterity" - underlie both specialized skills (e.g., playing the piano) and more mundane tasks (e.g., tying shoelaces). Impairments in dexterous skill cause significant disability, as occurs with motor cortical injury, Parkinson's Disease, and a range of other pathologies. Clinical observations, as well as basic investigations, suggest that cortico-striatal circuits play a critical role in learning and performing dexterous skills. Furthermore, dopaminergic signaling in these regions is implicated in synaptic plasticity and motor learning. Nonetheless, the role of striatal dopamine signaling in skilled motor learning remains poorly understood. Here, we use fiber photometry paired with a genetically encoded dopamine sensor to investigate striatal dopamine release in both male and female mice as they learn and perform a skilled reaching task. Dopamine rapidly increases during a skilled reach and peaks near pellet consumption. In dorsolateral striatum, dopamine dynamics are faster than in dorsomedial and ventral striatum. Across training, as reaching performance improves, dopamine signaling shifts from pellet consumption to cues that predict pellet availability, particularly in medial and ventral areas of striatum. Furthermore, performance prediction errors are present across the striatum, with reduced dopamine release after an unsuccessful reach. These findings show that dopamine dynamics during skilled motor behaviors change with learning and are differentially regulated across striatal subregions.Significance Statement Dexterity is central to everyday life but impaired in Parkinson's Disease, which is characterized by midbrain dopamine neuron degeneration. These neurons project to the striatum, where dopamine release plays a role in motor learning and execution. Endogenous dopamine dynamics in striatal subregions and how they relate to dexterous skill remain unclear. We used a fluorescence-based dopamine biosensor to elucidate the pattern of dopamine release in striatal subregions as mice learn and execute single pellet skilled reaching. Dopamine release varies across dorsolateral, dorsomedial, and ventral striatum as mice learn the task, and demonstrates 'performance prediction errors' when mice fail to retrieve the pellet. These results suggest that phasic dopamine dynamics are important for learning and executing dexterous skills.

Bayesian dynamic modelling for probabilistic prediction of pavement condition

Significant funds have been allocated to maintain road networks each year in developed countries. Performance prediction is crucial for pavement management systems to adjust working plans and budget allocation. As a Bayesian nonparametric method, Gaussian process regression (GPR) is powerful in predicting nonlinear time series and quantifying uncertainty. However, it remains computationally intensive and fails to adapt to the time-varying characteristics. To address such issues, a dynamic GPR model is proposed for probabilistic prediction of the International Roughness Index (IRI) for flexible pavements. A moving window strategy is developed to substantially shrink the size of training data, which effectively alleviates computational cost and thus leads to a dynamic GPR. A genetic algorithm is then adopted to determine the optimal window size by considering the trade-off between computational efficiency and accuracy. A dataset acquired from Long-Term Pavement Performance (LTPP) is used to demonstrate the feasibility of the dynamic GPR. Its performance is compared to traditional GPR as well as dynamic and static Bayesian linear regression (BLR) models. The comparison results indicate that the proposed dynamic GPR can increase the accuracy by 0.86, 1.52, and 2.27 times for dynamic BLR, static GPR, and static BLR, respectively. It exhibits the best results in terms of accuracy and uncertainty metrics due to its nonlinear modelling and time-varying ability.

A STUDY ON COMPETITOR ANALYTICS AS A DECISION MAKING CATALYST IN A LEATHER MANUFACTURING COMPANY

This study investigates the impact of competitor analytics on decision-making at a leather manufacturing company, using machine learning algorithms such as Random Forest, Support Vector Regression (SVR), and XGBoost. Trained on a dataset of 5,871 samples, the models were evaluated with Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE). SVR emerged as the most efficient model. Python libraries NumPy and Pandas in Jupyter Notebook facilitated the data analysis. The SVR model provided valuable insights for strategic decision-making and competitor performance predictions. This study underscores the importance of competitor analytics and machine learning in enhancing manufacturing strategies. Keywords: competitor analytics, machine learning algorithms, strategic decision making, predictive modeling, python, data science

Influence of corrosion on seismic performance of steel beam-to-column connections

Steel structures are not only subjected to the threaten of single hazard (e.g., earthquake), but also suffers from the coupling environmental effect such as corrosion throughout the entire service life cycle. This study explores the seismic performance of steel beam-to-column connections after corrosion. Accelerated corrosion tests were conducted on a series of practical steel beam-to-column connections, and the seismic performances of the specimens before and after corrosion were compared. A modeling method of corroded steel members that considers distribution randomness of corrosion depth was proposed. Finally, the proposed method is applied to conduct time-dependent seismic performance prediction of corroded beam-to-column connections. The results show that corrosion mainly influences the buckling and post buckling behaviors of steel beam-to-column connections. Due to earlier appearance of local buckling, more evident stiffness degradation is observed for the corroded specimen. Corrosion also induces performance degradation of the panel zone. For the corroded specimen with cover plate, multiple small cracks initiate at the junction between the cover plate and beam flange, and fracture damage is more likely to happen for the beam-to-column connections after corrosion. Both yield and ultimate load-carrying capacities of the connections decrease linearly with the increase of accelerated corrosion time, and the load-carrying capacity deterioration rate of the cover plate strengthened connection is quicker than that of the un-strengthened connection. The initial stiffness approximately degrades linearly with the increase of corrosion time. Corrosion degree also influences the energy dissipation capability of the beam-to-column connections, especially after the occurrence of local buckling.

Performance prediction for network-level flexible pavements using an optimised ANN approach

ABSTRACT Considering that the road network shares traffic flow as a system, this study aimed to predict the network-level flexible pavement performance, using an optimised artificial neural network (ANN) approach to predict the pavement performance of each individual road, and the Ford-Fulkerson algorithm to determine the weight coefficients of roads. ANNs were developed to predict distress conditions, functional conditions and structural performance using the inputs of the pavement age and structural, traffic and climatic conditions. ANNs were trained with Long-Term Pavement Performance (LTPP) program data, and almost all the coefficient of determination values between ANN outputs and measured results are larger than 0.9. This ANN approach was optimised by establishing grey models to provide predictive short-term performance data for training ANNs, and employing Kalman Filter to modify the long-term performance prediction. Applications in typical LTPP sections validated the effectiveness of the optimisation method. This study used the reduction in road network capacity due to the assumed unavailability of a certain road to quantify its role (weight coefficient), which was calculated using the Ford-Fulkerson algorithm. This network-level performance prediction approach was applied in a 25-expressways network in Shanghai, which validated the good generalisation capabilities of ANNs and feasibility of the Ford-Fulkerson algorithm.

PerfTop: Towards performance prediction of distributed learning over general topology

Distributed learning with multiple GPUs has been widely adopted to accelerate the training process of large-scale deep neural networks. However, misconfiguration of the GPU clusters with various communication primitives and topologies could potentially diminish the gains in parallel computation and lead to significant degradation in training efficiency. Predicting the performance of distributed learning enables service providers to identify potential bottlenecks beforehand. In this work, we propose a Performance prediction framework over General Topologies, called PerfTop, for accurate estimation of per-iteration execution time. The main strategy is to integrate computation time prediction with an analytical model to map the nonlinearity in communication and fine-grained computation-communication patterns. This enables accurate prediction of a variety of neural network models over general topologies, such as tree, hierarchical, and exponential. Our extensive experiments show that PerfTop outperforms existing methods in estimating both computation and communication time, particularly for communication, surpassing the existing methods by over 45%. Meanwhile, it achieves an accuracy of above 85% in predicting the execution time over general topologies compared to simple topologies such as star and ring from the previous works.

A Visual Survey of Tunnel Boring Machine (TBM) Performance in Tunneling Excavation: Mainstream Direction, Brief Review and Future Prospects

This study employs scientometric analysis to investigate the current trajectory of research on tunnel boring machine (TBM) performance and collaborative efforts. Utilizing software tools like Pajek 5.16 and VOSviewer 1.6.18, it scrutinizes literature from 2000 to 2021 sourced from the Web of Science (WOS). The findings illuminate TBM research as an interdisciplinary and intersectoral field attracting increasing national and institutional attention. Notable contributions from China, Iran, the United States, Turkey, and Australia underscore the global significance of TBM research. The recent upsurge in annual publications, primarily driven by Chinese research initiatives, reflects a renewed vigor in TBM exploration. Additionally, the paper presents a succinct evaluation of TBM advantages and drawbacks compared to conventional drill and blast methods, discussing key considerations in excavation methodology selection. Moreover, the study comprehensively reviews TBM performance prediction models, categorizing them into theoretical, empirical, and artificial intelligence-driven approaches. Finally, rooted in metaverse theory, the discourse delves into the immersive learning model and the architecture of a TBM metaverse. In the future, the immersive training and learning model diagram can be employed in scenarios such as employee training and the promotion of safety knowledge. Additionally, the TBM metaverse architecture can simulate, monitor, diagnose, predict, and control the organization, management, and service processes and behaviors of TBMs. This will enhance efficient collaboration across various aspects of the project production cycle. This forward-looking perspective anticipates future trends in TBM technology, emphasizing societal impact and enhancement of economic benefits.

Performance Prediction of Power Beacon-Aided Wireless Sensor-Powered Non-Orthogonal Multiple-Access Internet-of-Things Networks Under Imperfect Channel State Information

In this paper, we investigate a novel power beacon (PB)-aided wireless sensor-powered non-orthogonal multiple-access (NOMA) Internet-of-Things (IoT) network under imperfect channel state information (CSI). Furthermore, the exact expression outage probability (OP) of two IoT users is derived to analyze the performance of the considered network. To give further insight, the expression asymptotic OP and diversity order are also expressed when the transmit power at the PB goes to infinity. Furthermore, a deep neural network (DNN) framework is proposed to concurrently forecast IoT users’ OP in relation to real-time setups for IoT users. Additionally, when compared to the traditional analysis, our created DNN shows the shortest run-time prediction, and the outcomes predicted by the DNN model almost match those of the simulation. In addition, numerical results validate our analysis, simulation, and prediction through a Monte Carlo Simulation. Furthermore, the results show the impact of the main parameter on our proposed system. Finally, these findings show that NOMA performs better than the conventional orthogonal multiple-access (OMA) techniques.

A performance prediction method for on-site chillers based on dynamic graph convolutional network enhanced by association rules

A performance prediction method for on-site chillers based on dynamic graph convolutional network enhanced by association rules

A particle-level discrete element model for simulating the performance of MRFs

The particle dynamics model of MRFs (MRFs) is based on the discrete element technique, which takes into account essential MRF variables such as magnetic field, particle parameters, particle volume fraction, carrier fluid properties, and other parameters. Firstly, to digitally characterize the microstructure of MRFs, the particle coordination number, as a new perspective is creatively proposed. The validity and predictability of the created model are then tested in conjunction with the rheological performance studies by comparing it to the known continuous medium model. Finally, based on the proposed model, the influence of each influencing factor on the magnetorheological effect is simulated and regularity analysis is given from the perspective of particles. Linking macroscopic qualities and MRF microstructure, which offers a theoretical basis for the preparation of MRFs with improved performance and prediction of performance under various working situations.

The dynamic of photovoltaic resources on its performance predictability, based on two new approaches

The dynamic of photovoltaic resources on its performance predictability, based on two new approaches

Improving part-load performance of combined-cycle gas turbines by optimizing variable geometry control strategy for compressor and power turbine combined adjustment

The variable geometry methods currently used in combined-cycle gas turbines are compressor variable inlet guide vanes (VIGV) or power turbine variable area nozzles (VAN). On this basis, this study presents the optimal variable geometry control strategy for compressor and power turbine combined adjustment ([Formula: see text]) using the Differential Evolutionary Algorithm with the LM2500+ gas turbine. The aim is to further improve the part-load performance of the combined-cycle gas turbine. Firstly, a part-load performance prediction model for variable geometry gas turbines is established based on the component method. Subsequently, a variable geometry gas turbine part-load performance optimization model is developed by combining the Differential Evolution Algorithm. Finally, the optimum combination of stagger angles for the compressor inlet vane and power turbine nozzle is calculated at each part-load condition. Compared to the VIGV and VAN control strategies, the [Formula: see text] control strategy proposed in this paper shows a higher stability margin and better economy. The [Formula: see text] control strategy maintains a constant exhaust temperature within a part load range from 20% to 100% with the stability margin exceeding 14%. In comparison with the VAN control strategy, the fuel flow rate decreases by 1.152% at 45% relative load power and by 3.435% at 20.0% relative load power with the [Formula: see text] control strategy.

Performance Prediction of the Elastic Support Structure of a Wind Turbine Based on Multi-Task Learning

Performance Prediction of the Elastic Support Structure of a Wind Turbine Based on Multi-Task Learning

An experimental validation of diffusion-based devolatilization models in extruders using post-industrial and post-consumer plastic waste

Abstract Extrusion is a key process in mechanical recycling. In a degassing step, volatile components, including all impurities and moisture, are removed from a polymer melt to ensure consistently high quality of the recyclates. Predicting devolatilization performance is therefore of interest in the design of degassing screws; in the plastics industry, it also plays an important role in the transition from a linear to a circular economy. Using two different devolatilization models, we first modelled the degassing process of a lab-scale twin-screw extruder and an industrial-scale recycling single-screw extruder. We then predicted the devolatilization performance of both machines, validated the results with experimental data obtained from emissions tests carried out with post-industrial and post-consumer polypropylene materials and performed linear regression analysis to compare our two models in terms of predictive quality. Our results showed that both models are equally suitable for reliable prediction of the devolatilization performance.

PVDF membranes for membrane distillation prepared by VIPS – A combined study of artificial water channel effectiveness and membrane performance prediction

Background A combined VIPS-NIPS technique was used to investigate the potential of self-supporting flat-sheet polyvinylidene fluoride (PVDF) membranes containing amphiphilic I-quartet Artificial Water Channels (AWC) for Direct Contact Membrane Distillation (DCMD). The AWC are formed in situ by self-assembly of HC6 molecules upon contact with water during membrane formation. Methods A Design of Experiment (DoE) was used to investigate the influence of various experimental parameters and their interactions on membrane performance, namely polymer and AWC concentrations in the dope solution, relative humidity and duration of the VIPS treatment. Pure water permeability (PWP), Liquid Entry Pressure (LEP) and DCMD flux were used to characterize membrane performance. A Response Surface Methodology (RSM) was used to evaluate the DoE results and a second order model was fitted. Based on the model predictions several multiple response-optimized (MRO) membranes were prepared. Results The performance improvement of the MRO membranes was 7.0 times for PWP (M-T4) and 3.3 times for MD flux (M-T3) over the average performance of the DoE membranes, while maintaining LEP of 1.9 bar and Salt rejection of 99.9%. In addition, hybrid PVDF/HC6 MRO membranes improved PWP by 3.2 times and MD flux by 1.6 times over pure PVDF MRO membranes. Conclusions DoE in conjunction with RSM enabled the prediction of membrane PWP, LEP and MD flux performance by varying the levels of the experimental parameters within predefined limits. The HC6 additive improves PWP and MD flux without loss of LEP or selectivity. HC6 promotes the formation of porous skin and asymmetric cross-sections with finger-like structures. Probably by affecting diffusion during phase separation due to its amphiphilic properties.

Prediction of residual performance of corroded RC beams from observed surface cracks by 3D RBSM with MPC

Corrosion of reinforcing steel in concrete structures can significantly compromise their performance. Accurately estimating this deterioration for in-service structures is challenging due to uncertainties surrounding the distribution of internal corrosion and cracking patterns, compounded by the limitations of destructive testing. This research proposes a novel approach to predict the remaining load-bearing capacity of corroded reinforced concrete (RC) structures based on the pattern of surface cracks caused by corrosion. The method integrates a model predictive control (MPC) algorithm with a three-dimensional Rigid Body Spring Model (3D RBSM) to estimate internal corrosion levels from observed cracks and a normal 3D RBSM simulation to predict the residual performance. The proposed system is validated through analysis of four corroded beam specimens based on experimental work reported in the literature. In use, the MPC algorithm first estimates a corrosion distribution in a RC beam and generates simultaneous corrosion-induced crack and stress predictions. This information is then used to predict residual load-bearing capacity. The system accurately forecasts the non-uniform corrosion distribution, flexural capacity, and load-displacement curves of a specimen, offering insight into the performance differences among the specimens and demonstrating its potential to assess the serviceability of existing RC structures with corrosion issues.