Average Error Research Articles

Flow field reconstruction of compressor blade cascade based on deep learning methods

In the design process of compressors, calculating the S1 flow surface involves solving the Navier-Stokes equations, which results in slow convergence and makes determining cascade characteristics time-consuming. However, deep learning offers significant advantages in flow field reconstruction by not only automatically extracting complex flow features and reducing prediction time but also providing high accuracy in reconstruction. This paper implements the rapid reconstruction of the compressor S1 flow surface cascade flow field using two deep learning models: U-Net and 1D-CNN. Using a double-circular arc airfoil as an example, we selected four key design parameters that define the geometry and position of the airfoil, ultimately designing 5,292 sets of cascade geometries. By performing batch meshing and CFD simulations, we built a cascade flow field dataset. The U-Net neural network uses design parameters as input and outputs the aerodynamic distribution of the cascade flow field. After training, it can directly predict the flow field based on the design parameters. Since the U-Net model cannot directly obtain the aerodynamic parameter distribution and flow field aerodynamic coefficients on the airfoil surface, a 1D-CNN model is used as a complementary approach. The 1D-CNN model takes the design parameters as input and outputs the aerodynamic parameter distribution on the airfoil surface and the flow field aerodynamic coefficients. The prediction results show that the U-Net model achieves an average relative error of less than 1% in cascade flow field reconstruction, while the 1D-CNN model achieves an average relative error of less than 1% in predicting the pressure recovery coefficient and less than 2% in predicting the total pressure loss coefficient. This study presents a method for the rapid reconstruction of compressor blade cascade flow fields, which helps improve design efficiency and shorten the design cycle.

Advancing Mix Design Prediction in 3D Printed Concrete: Predicting Anisotropic Compressive Strength and Slump Flow

Introducing 3D-concrete printing has started a revolution in the construction industry, presenting unique opportunities alongside undeniable challenges. Among these, the major challenge is the iterative process associated with mix design formulation, which results in significant material and time consumption. This research uses machine learning (ML) techniques such as Extreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), Decision Tree Regression (DTR), Gaussian Process Regression (GPR), and Artificial Neural Network (ANN) to overcome these challenges. A dataset containing 21 mix constituent features and 4 output properties (cast and printed compressive strength, and slump flow) was extracted from the literature to investigate the relationship between mix design and performance. The models were assessed using a range of evaluation metrics, including Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Squared Error (MSE), and R-squared value. Gaussian Process Regression (GPR) yielded more favorable results. In the case of cast strength, GPR achieved an R2 value of 0.9069, along with RMSE, MSE, and MAE values of 13.04, 170.12, and 9.40, respectively. A similar trend was observed for printed strengths in directions 1, 2, and 3. GPR achieved R-squared values exceeding 0.91 for all directions, accompanied by significantly lower RMSE values (below 4.1). The machine learning models were also validated using four unique mix designs. These mixes were 3D printed and tested for compressive strength and slump flow. GPR's average error was 10.55%, while SVM achieved a slightly lower average error of 9.38%. Overall, this work presents a novel approach for optimizing 3D-printed concrete by enabling the prediction of slump flow and compressive strength directly from the mix design. This approach can facilitate the design and fabrication of 3D-printed concrete structures that fulfill the necessary strength and printability requirements.

AI-powered ultrasonic thermometry for HIFU therapy in deep organ

High-intensity focused ultrasound (HIFU) is considered as an important non-invasive way for tumor ablation in deep organs. However, accurate real-time monitoring of the temperature field within HIFU focal area remains a challenge. Although ultrasound technology, compared with other approaches, is a good choice for noninvasive and real-time monitoring on the temperature distribution, traditional ultrasonic thermometry mainly relies on the backscattered signal, which is difficult for high temperature (>50 °C) measurement. Given that artificial intelligence (AI) shows significant potential for biomedical applications, we propose an AI-powered ultrasonic thermometry using an end-to-end deep neural network termed Breath-guided Multimodal Teacher-Student (BMTS), which possesses the capability to elucidate the interaction between HIFU and complex heterogeneous biological media. It has been demonstrated experimentally that two-dimension temperature distribution within HIFU focal area in deep organ can be accurately reconstructed with an average error and a frame speed of 0.8 °C and 0.37 s, respectively. Most importantly, the maximum measurable temperature for ultrasonic technology has been successfully expanded to a record value of 67 °C. This breakthrough indicates that the development of AI-powered ultrasonic thermometry is beneficial for precise HIFU therapy planning in the future.

A modified Johnson-Cook constitutive model of Inconel 690 weld overlay taking into account the strain rate softening effect

Precise material constitutive models are essential for finite element (FE) cutting simulation and analytical calculations of cutting mechanics. In this work, a split Hopkinson pressure bar (SHPB) equipment was employed to investigate the dynamic mechanical behavior of Inconel 690 weld overlay with a wide range of temperatures (20℃, 200℃, 400℃, 600℃, and 800℃) and strain rates (5000s-1, 7500s-1, 10,000s-1, and 12,500s-1). Based on the material flow characteristics at different loading conditions, a modified Johnson-Cook (JC) constitutive model was established, which takes into account the strain rate softening effect at high strain rates and the softening behavior induced by the competition between work hardening and dynamic recovery at large strains. Compared with the original JC constitutive model, the modified constitutive model can provide a closer correspondence between the predicted flow stresses and experimental data. The modified material constitutive model was introduced into FE simulation for orthogonal cutting of Inconel 690 weld overlay. The results indicate that the average prediction errors for the main cutting force and thrust force are 6.38% and 11.38%, respectively, validating the applicability and reliability of the modified JC constitutive model for cutting simulation applications. This study can provide a mechanical foundation for FE analysis in high-speed cutting of Inconel 690 weld overlay, which is conducive to optimizing cutting parameters, reducing machining costs, and enhancing production efficiency in practical industrial manufacturing.

Co-firing characteristic prediction of solid waste and coal for supercritical CO2 power cycle based on CFD simulation and machine learning algorithm

The co-firing technology of combustible solid waste (CSW) and coal in the supercritical CO2 (S-CO2) circulating fluidized bed (CFB) can effectively deal with domestic waste, promote social and environmental benefits, improve the coal conversion rate, and reduce pollutant emission. This study focuses on the co-firing characteristics of CSW and coal under S-CO2 power cycle, and simulations are conducted by employing Multiphase Particle-in-cell (MP-PIC) method integrated with the comprehensive chemical reaction models in a 300 MW S-CO2 CFB boiler. Effects of operating parameters including fuel mixture proportion and first stage stoichiometry on the gas emission characteristics are further analyzed. Based on training and testing database based on the simulation results, a novel Improved Whale Optimization Algorithm and Bi-dictionary Long Short-Term Memory (IWOA-BiLSTM) algorithm model is established to predict CFB temperature, NOx emission concentration, and SO2 emission concentration, respectively. Results show that CO and SO2 decrease with the coal mass ratio of the fuel mixture increasing, while NOx increases. With the increase of first stage stoichiometry, CO increases, NOx declines, and the change of SO2 is not obvious. Compared with two other basic algorithm models, the prediction error of the proposed algorithm model for the three targets is minimal with the average relative error of 0.032 %, 0.231 %, and 0.157 %, respectively, which can meet the prediction requirements with acceptable accuracy.

Prediction of vertical transport of microplastics: Shape- and aging-dependent drag models

The prediction of vertical transport of microplastics (MPs) is essential for understanding their multidimensional transport, fate, and environmental risks, but drag models applicable to aging MPs are currently understudied. In this study, pristine and UV-aged polyethylene terephthalate (PET) and polystyrene (PS) MPs were used for settling experiments. Combining physicochemical properties and transport data, a shape-dependent drag model based on the Corey shape factor was optimized with average errors of 9.73% and 10.42% and coefficients of determination of 0.6878 and 0.8359 for predicting the settling terminal velocities (ut) for PET and PS MPs, respectively. Meanwhile, aging-dependent drag models were constructed by incorporating the carbonyl index as functional forms of the newly defined aging index, which can be used to differentiate the effects of shape and aging characteristics on the vertical transport of MPs. These aging-dependent models showed better predictive abilities with average errors of 3.97% and 4.56% in predicting ut for PET MPs, and of 5.89% and 6.91% for PS MPs. Additionally, the drag models in this study improved applicability to predict vertical transport of environmentally-collected weathered MPs. With the continuous improvement of the transport database of diverse MPs, this study is expected to provide scientific support for predicting the environmental behaviors of MPs and formulating targeted pollution control strategies.

Transient gas path fault diagnosis of aero-engine based on domain adaptive offline reinforcement learning

Real-time measurement parameters are crucial for diagnosing faults in aero-engine gas path performance, ensuring engine reliability, and mitigating potential economic losses. Traditional aero-engines performance diagnosis was mainly based on the measurements of steady-state condition and lacked the utilization of data under transient conditions. Gas path diagnosis of aero-engines under transient conditions is crucial for early fault detection and safety of flight within the envelope. The challenge lies in the inconsistent distribution of performance deviations caused by variable operating conditions, especially with complex fault types, which can undermine diagnostic credibility. To improve reliability of gas path diagnosis under transient conditions, an offline reinforcement learning fault diagnosis framework based on a transient aero-engine performance model is proposed. To address the issue of variable operating conditions during transient states, a domain adaptive approach is utilized to reconstruct the measurement baseline and facilitate the transfer of different performance deviation distributions. Additionally, by adding spool acceleration as a measurement parameter, the multi-component fault coupling is solved. Finally, validation with actual operating data simulates fault cases, demonstrating the proposed method's efficacy in quantitatively detecting gradual, sudden, and multiple component faults under transient conditions with high accuracy and efficiency. The method proposed in this study achieves a computational speed improvement by 64% compared to the conventional method, achieving a time of 0.13 seconds, with an average error of less than 0.00389%. Additionally, it demonstrates strong robustness in the presence of noise, with an average error of less than 0.03125%. This proposed method improves real-time fault detection under transient conditions for its higher accuracy and efficiency, and therefore significantly enhance gas path health monitoring and diagnosis capability.

A Novel Fitting Polynomial Approach for An Accurate Soc Estimation In Li-Ion Batteries Considering Temperature Hysteresis

Lithium-ion batteries are essential to modern technology, requiring accurate estimation of the state of charge (SOC) for optimal performance. Traditional methods such as Coulomb Counting (CC) are ineffective in the face of temperature variations, leading to inaccuracies in SOC estimation, which in turn cause obvious deformation of hysteresis curves. To address this, this paper introduces a novel method called Polynomial Fit State of Charge (FPSOC), for effective SOC estimation. This method incorporates a fifth-degree polynomial fitting that accounts for a wide range of temperature variations (from -10°C to +80°C), a feature that, according to the authors, has not been offered by all previously published methods. A series of simulation tests using the MATLAB/Simulink tool are conducted under various temperature profiles to evaluate the effectiveness of the FPSOC method. The results demonstrate the notable superiority of the FPSOC model compared to the CC method, with a significantly reduced RMSE of only 0.93% compared to 6.77% of the CC model. Particularly effective at low SOC levels (30%), the FPSOC model demonstrates precision up to six times higher compared to the CC model. Additionally, when evaluated against other recent SOC estimation techniques such as CM, RLSF, EKF, DST, BBDST, ASMO, LPM_H, LSTM-SA Group A and B, and baseline ECM-ID, The FPSOC method proves extremely accurate, with the lowest average error under different temperature conditions.

Experimental investigation on in-situ laser-assisted mechanical ruling of single crystal silicon

Abstract In this paper, response surface methodology (RSM) was employed as a robust and convenient predictive tool to establish the correlation between process parameters of in situ laser-assisted mechanical ruling and the ductile-to-brittle transition of single-crystal silicon. The interaction effects among three factors laser power, ruling speed, and negative rake angle on the ductile-to-brittle transition of single-crystal silicon were investigated. The optimal combination of process parameters was determined to be a laser power of 30W, a ruling speed of 0.25 mm s−1, and a negative rake angle of 58°. Utilizing a self-assembled setup and the optimal process parameters, multiple processing experiments were conducted. The average error between the experimental and predicted values was found to be 2.8%.

Forecasting Grain Production in Hubei Province Based on Markov Modeling

Abstract. Grain production plays a crucial role in economic and social development. Accurate prediction of grain production can help agricultural decision-makers allocate resources rationally and do a good job at the next stage of the arrangements for the deployment of the work.This article used the production data, which was from 2023 Hubei Provincial Statistical Yearbook,to simulate and predict the grain production for several years,with the method of Markov prediction. It can be seen from the results that the average fitting error of the model is 3.47%. Using the text dataset of grain production from 2020 to 2022, the results show that the model can effectively improve the prediction accuracy of agricultural grain procuction. Finally, the model is used to predict the agricultural output value structure of China from 2023 to 2027. Markov prediction can improve the predictability and initiative of the work and provide scientific basis for the future decision-making of agricultural managers, and has guiding significance for the work of grain storage and distribution.

Positioning Sensor Data Mining and Correlation Analysis Based on Multi-mode Radio Frequency Identification

Real-time positioning and sensor data mining are critical for various Internet of things (IoT) applications, but current approaches face challenges such as limited accuracy, high cost, and energy inefficiency. This paper proposes a novel fusion intelligent structure that combines radio frequency identification (RFID) technology and wireless sensor networks (WSN) to enable accurate and efficient positioning and data mining. The proposed method consists of two key components: an energy heterogeneity strategy for optimizing network energy consumption, and a data association feature analysis technique for mining sensor data. The results demonstrated that the proposed approach achieved superior positioning accuracy compared to other algorithms, with an average error of less than 0.300 under various conditions. The data detection accuracy was also high, with a maximum false detection rate of 3%. Furthermore, the proposed energy heterogeneity strategy significantly improved network energy utilization and stability. Real-world scenario testing validated the practicality of the proposed approach, with an average positioning error of less than 0.500 meters. These results validate the effectiveness of the proposed RFID-WSN fusion structure for real-time positioning and sensor data mining, providing a promising solution for IoT applications.

A novel algorithm for generating random fiber distributions for fiber reinforced composites based on lattice partition

AbstractThis paper propose a random fiber distribution generation algorithm based on lattice partition, which addresses the issue of limited increase in fiber volume fraction caused by the random movement of fibers in previous methods. By partitioning the representative volume element (RVE) window into lattices and moving fibers directionally, the proposed algorithm significantly increases the number of fibers within the RVE window. The generated fiber distribution achieves a fiber volume fraction of up to 80%, with an average execution time of 243.4 s. The influence of algorithm parameters on the randomness of fiber distribution is examined through statistical analysis of the generated distributions. Finite element models with various parameters are established to predict the elastic properties of the composites. The average relative error between the predicted values and the experimental results is 8.6%, demonstrating the effectiveness of the proposed algorithm. This algorithm offers a simple and effective approach for generating fiber distributions in the numerical study of the micromechanics of composites.Highlights A new lattice‐partition method is proposed to generate fiber distributions. The fiber volume fraction of generated fiber distribution is up to 80%. The randomness of fiber distribution is analyzed through statistical analysis. The algorithm predicts elastic properties with an average relative error of 8.6%.

Simulation study on magneto-acoustic concentration tomography of magnetic nanoparticles based on vectorial acoustic source and BICGSTAB method

Abstract To overcome the vulnerability to noise of the reconstructed image and simplify the cumbersome iteration process of algorithm in the magneto-acoustic concentration tomography with magnetic induction (MACT-MI) for magnetic nanoparticles (MNPs), we established the matrix relationship between the concentration of MNPs and the first-order derivative of sound pressure based on the reconstruction method of vectorial acoustic source, and proposed the application of BICGSTAB method in solving the concentration distribution. Firstly, a simulation model was established in COMSOL Multiphysics. Secondly, the obtained data were substituted into the derived formula for imaging reconstruction. Finally, the quality of the reconstructed image was analyzed. The effects of MNP radius, shape, asymptotic concentration, and SNR on the reconstruction results were studied. Simulation results show that under the same noise condition, compared with the reconstruction method based on the LSQR-trapezoidal method, the average correlation coefficient increased by 32.9%, the average relative error decreased by 48.5%, the average structural similarity increased by 48.2%, and the average iterations decreased by 58.5%. The proposed method shows superior imaging quality and noise immunity. The research provides a theoretical basis for subsequent experimental research.

Numerical Study of the Cavitation Performance of an Ice-Blocked Propeller Considering the Free Surface Effect

Propeller cavitation performance can be predicted based on model tests or simulations. However, the cavitation performance of an ice-blocked propeller near the free surface differs from that of a propeller in the cavitation tunnel. Therefore, research on the cavitation performance simulation of propellers near the free surface holds crucial scientific significance. In this study, a coupled model was established using Computational Fluid Dynamics (CFD) and the Volume of Fluid (VOF) coupling method. The CFD-VOF model weighted the overlapping grids and simulated the cavitation performance of an ice-blocked propeller using various immersion depths, cavitation numbers, and advance coefficients. The propeller inflow ahead of the propeller and the wake field behind it were controlled to accurately obtain the propeller cavitation performance. Moreover, a comparison was conducted between the cavitation tunnel test results and the numerical simulation results at various immersion depths. When the immersion depth was at a distance of 1D, the effect of the free surface on the propeller cavitation performance became significant. When the immersion depth was at a distance of 9D, the average errors between the numerical simulation and the model test data were within 10%. This study analyzed the cavitation performance of ice-blocked propellers near the free surface and provided valuable insights for the design of ice-class propellers.

Robust pollination for tomato farming using deep learning and visual servoing

Abstract In this work, we propose a novel approach for tomato pollination that utilizes visual servo control. The objective is to meet the growing demand for automated robotic pollinators to overcome the decline in bee populations. Our approach focuses on addressing this challenge by leveraging visual servo control to guide the pollination process. The proposed method leverages deep learning to estimate the orientations and depth of detected flower, incorporating CAD-based synthetic images to ensure dataset diversity. By utilizing a 3D camera, the system accurately estimates flower depth information for visual servoing. The robustness of the approach is validated through experiments conducted in a laboratory environment with a 3D printed tomato flower plant. The results demonstrate a high detection rate, with a mean average precision of 91.2 %. Furthermore, the average depth error for accurately localizing the pollination target is impressively minimal, measuring only 1.1 cm. This research presents a promising solution for tomato pollination, showcasing the effectiveness of visual-guided servo control and its potential to address the challenges posed by diminishing bee populations in greenhouses.

Assessing Project Resilience Through Reference Class Forecasting and Radial Basis Function Neural Network

A project needs to be able to anticipate potential threats, respond effectively to adverse events, and adapt to environmental changes. This overall capability is known as project resilience. In order to make efficient project decisions when the project is subjected to disruption, such as adjusting the project budget, reformulating the work plan, and rationalizing the allocation of resources, it is necessary to quantitatively understand the level of project resilience. Therefore, this paper develops a novel approach for forecasting project performance, illustrating the changes in performance levels during the disruption and recovery phases of a project and thus quantitatively assessing project resilience. While there are several methods for assessing project resilience in existing research, the majority of assessment approaches originate from within projects and are highly subjective, which makes it difficult to objectively reflect the level of project resilience. Moreover, the availability of project samples is limited, which makes it difficult to forecast the level of project performance. In view of the fact that the Reference Class Forecasting (RCF) technique avoids subjectivity and the Radial Basis Function (RBF) neural network is known to be better at forecasting small sample datasets, this paper therefore combines the RCF technique and the RBF neural network to construct a model that forecasts the project performance of the current project after experiencing a disruption, further assessing the level of the project resilience. Specifically, this paper first presents a conceptual model of project resilience assessment; subsequently, an RBF neural network model that takes into account project budget, duration, risk level of disruption, and performance before disruption based on the RCF technique is developed to forecast project performance after experiencing disruption; and finally, the level of project resilience is assessed through calculating the ratio of recovery to loss of project performance. The model is trained and validated using 64 completed construction projects with disruptions as the datasets. The results show that the average relative errors between the forecast results of schedule performance index (SPI) and the real values are less than 5%, and the R2 of the training set and the testing set is 0.991 and 0.964, respectively, and the discrepancy between the forecasted and real values of project resilience is less than 10%. These illustrate that the model performs well and is feasible for quantifying the level of project resilience, clarifying its impact on project disruption and recovery situations, and facilitating the decision-makers of the project to make reasonable decisions.

Flattening behaviour of weft-knitted spacer fabrics

AbstractWeft-knitted spacer fabrics are thick 3D knitted structures prized for their cushioning properties which have gathered increasing attention in the last decade. The thickness of a spacer fabric is one of its most influential parameters and strongly impacts its cushioning properties, wearability, thermal insulation or permeability. However, the fabric's natural undulation and high deformability make its thickness measurement uneasy. The current standard measurement methods require to measure the fabric thickness after compressing it until a fixed threshold stress value is reached to flatten it. The diversity of these threshold values is confusing, and each of them is unsuitable to variety of fabric rigidity. In this article, a standard for thickness evaluation was proposed and used to measure the thickness of 20 samples knitted with 5 independent parameters. The measured thickness was compared to the thickness measured at a threshold value of 1 kPa and to a theoretical thickness. The proposed measurement standard was proved reproducible and efficient for all fabrics when the threshold measures showed large errors on the softer and stiffer samples. The flattening stress of the fabrics ranged from 86 to 5262 Pa and could not be approximated by a single standard value. The theoretical thickness was more accurate, predicting the thickness with an average error of 3.8%.

A Monocular Ranging Method for Ship Targets Based on Unmanned Surface Vessels in a Shaking Environment

Aiming to address errors in the estimation of the position and attitude of an unmanned vessel, especially during vibration, where the rapid loss of feature point information hinders continuous attitude estimation and global trajectory mapping, this paper improves the monocular ORB-SLAM framework based on the characteristics of the marine environment. In general, we extract the location area of the artificial sea target in the video, build a virtual feature set for it, and filter the background features. When shaking occurs, GNSS information is combined and the target feature set is used to complete the map reconstruction task. Specifically, firstly, the sea target area of interest is detected by YOLOv5, and the feature extraction and matching method is optimized in the front-end tracking stage to adapt to the sea environment. In the key frame selection and local map optimization stage, the characteristics of the feature set are improved to further improve the positioning accuracy, to provide more accurate position and attitude information about the unmanned platform. We use GNSS information to provide the scale and world coordinates for the map. Finally, the target distance is measured by the beam ranging method. In this paper, marine unmanned platform data, GNSS, and AIS position data are autonomously collected, and experiments are carried out using the proposed marine ranging system. Experimental results show that the maximum measurement error of this method is 9.2%, and the average error is 4.7%.

Predicting Natural Gas Production in Various Nations Using a Fractional Grey Bernoulli Approach

Accurate forecasting of natural gas production is crucial for economic stability, environmental sustainability, and market investment. This study presents an advanced forecasting method using the fractional grey Bernoulli model, which combines fractional accumulation and Bernoulli processes to enhance the predictive accuracy for nonlinear datasets. The model’s versatility and flexibility allow it to adapt to various data characteristics and complexities, thereby outperforming traditional grey models in forecasting performance. To optimize the model parameters, this study employs the Particle Swarm Optimization (PSO) algorithm, further improving the model’s effectiveness. Empirical analysis of natural gas production data from Brazil, Italy, and Qatar demonstrates that the model exhibits significant advantages in both fitting and forecasting capabilities. The findings indicate that the fractional grey Bernoulli model achieves high accuracy and reliability in predicting natural gas production in these countries, providing a robust framework for strategic energy planning and investment decision-making. With average forecast errors of 1.9113%, 4.0353%, and 1.8902% for natural gas production in Brazil, Italy, and Qatar respectively, this study underscores the model’s effectiveness in enhancing forecast reliability and minimizing risk, providing valuable insights for sustainable energy development.

Enhancing the performance of a resonance-based sensor network for soft robots using binary excitation

Embedded, flexible, multi-sensor sensing networks have shown the potential to provide soft robots with reliable feedback while navigating unstructured environments. Time delay associated with extracting information from these sensing networks and the complexity of constructing them are significant obstacles to their development. This paper presents a novel enhancement to an existing class of embedded sensor network with the potential to overcome these challenges. In its original version, this sensor network extracts information from multiple reactive sensors on a two-wire electrical circuit simultaneously. This paper proposes to change the excitation signal applied to this sensor network to a binary signal. This change offers two key advantages: it provides the ability to employ small, inexpensive microcontrollers and results in a faster data extraction process. The potential of this enhanced system is demonstrated here with a proof of concept implementation. The self-inductance of all inductance-based sensors within this proof of concept sensor network can be measured at a rate of over 5000 times per second with an average measurement error of less than 2%.