Maximum Absolute Relative Error Research Articles

Validation of satellite-derived solar irradiance datasets: a case study in Saudi Arabia

A robust dataset of Surface Solar Irradiance is essential for secure competitive financing for solar energy projects. Rating agencies and lenders alike require verification of the solar-resource dataset for utilizing each solar energy project, as this can be translated directly into expected electrical energy and revenues. The accuracy of the dataset and the variability of solar radiation, as recorded by historical solar data, play a significant role in estimating the future performance of the project and its budget. The historical observed solar irradiance datasets by local stations are the best and most reliable for a specific site, but they are not always available for long and continuous periods in any location, especially in arid areas. So, the importance of historical solar radiation datasets derived from satellite-based models arises here. This paper validates the historical modeled datasets of the three most famous satellite-based commercial prediction models (SolarGIS, SUNY, and Solcast) against the observed dataset by six ground stations in Saudi Arabia under different climatic zones. The validation method has been implemented using the standard error metrics: Maximum Absolute Error (MAE) and relative Maximum Bias Error (rMBE). The validation process showed that, in the case of GHI, the discrepancy between observed and predicted values is narrow, while in the case of DNI, the discrepancy is wide. Also, the predicted GHI values are more accurate than predicted DNI values, and -in general- the values predicted by the SUNY model are less accurate than those predicted by SolarGIS and Solcast models for both GHI and DNI. The resultant of this validation process could be accepted not for the six locations under study only but, also for deserts and arid areas across Saudi Arabia and might be extended to similar arid areas around the world.

Data-driven method embedded physical knowledge for entire lifecycle degradation monitoring in aircraft engines

The degradation of aircraft engine performance necessitates real-time monitoring. We introduce a real-time performance-degradation monitoring method comprising a baseline state (BLS) model for pristine engine performance prediction and a real-time model for ongoing engine performance prediction. Both models are LSTM-based and integrated with the engine physical topology; however, they vary in their input parameters. To counteract the data distribution disparities across various engine states during model training, a data augmentation method is proposed. In a case study, data from 600 h of engine running were employed, with thrust as the target monitoring parameter. The initial 50 h dataset covers the entire range of working conditions for the subsequent 550 h, serving as the training dataset for both models. The results demonstrated the accuracy of these models, with the BLS model achieving a mean absolute relative error (MARE) below 0.5% and a maximum absolute relative error (Emax) below 8% in the initial 50 h. Simultaneously, the real-time model yielded a MARE below 0.7% and Emax below 10% for the subsequent 550 h. A comparison of the real-time thrust degradation between this study and traditional methods highlights the feasibility and advantages of this approach.

Time-dependent photovoltaic performance assessment on a global scale using artificial neural networks

The integration of Renewable Energy Sources (RESs), particularly solar PhotoVoltaics (PVs) has become an imperative aspect of sustainable energy systems. In this pursuit, accurate and efficient simulation tools play a pivotal role in optimizing the performance of PV systems. Traditional simulation approaches, while effective, are often characterized by computational complexities and time-intensive processes. This paper introduces a groundbreaking paradigm in solar energy modeling by harnessing the power of Artificial Neural Networks (ANNs) to revolutionize the accuracy and reliability of PV system simulations. In this work, an hourly, daily, monthly and yearly comparison of the electrical energy obtained with the 5-parameter model and those obtained with the ANNs was developed. For this purpose, a very wide ensemble of localities around the world and types of PV systems were considered in the training and validation phase. ANNs exhibited a maximum mean absolute relative error of 3.5% during training and consistently maintained hourly relative errors below 5% across diverse localities during validation. Hourly power forecasting remains acceptable also in localities with extreme weather conditions. Monthly errors peak at high negative and positive latitudes in summer months when daylight duration exceeds nighttime. However, in the least accurate locality, yearly energy forecasting yielded a maximum error of 8%. Empirical equations based on the trained ANNs are proposed and a relative input-output importance criterion was applied to detect the impact of air temperature and solar radiation on the performance of each PV module. The proposed ANNs demonstrate significant utility in decision-making and real-time processes, providing a valuable framework for managing energy flows within a network and predicting energy production during specific time intervals. This alternative approach surpasses conventional dynamic simulation methodologies found in existing literature in terms of computational cost with comparable accuracy.

Study on the Optimization of Heat Transfer Coefficient of a Rare Earth Wrought Magnesium Alloy in Residual Stress Analysis

To investigate the heat transfer coefficient (HTC) of a newly developed rare-earth wrought magnesium alloy under different cooling rates, the experiment of solution treatment followed by water quenching or air cooling process was carried out for calculation by lumped capacitance method (LCM) and optimized by inverse heat transfer method (IHTM), and cooling temperature curves were simulated afterward. In water quenching, the larger the temperature difference between the sample and water, the larger the maximum HTC, and the earlier it reached the maximum value, and in air cooling the HTC became larger with the airflow speeds increased. In LCM, the peak values of the HTC were 2840 W/(m2·°C) in water quenching and 54 W/(m2·°C) in air cooling. The corresponding HTC was 2388 W/(m2·°C) in IHTM. The maximum absolute average relative error (AARE) of temperature simulation in water quenching decreased from 8.46% in LCM to 2.45% in IHTM. The residual stress(RS) of a large conical component was simulated using both non-optimized and optimized HTC, the RS in the IHTM was ~30 MPa smaller than that in the ILCM, because the corresponding HTC was smaller, and the comparison of the simulation results with the measurements revealed that the RS using HTC in the IHTM is more accurate.

Microstructure Evolution and Constitutive Modelling of Deformation Behavior for Al-Mg-Si-Cu-Sc-Zr Alloy Processed with Isothermal Multidirectional Forging

This research is devoted to the microstructure evolution and deformation behavior of the Al-1.2Mg-0.7Si-1.0Cu-0.1Sc-0.2Zr alloy during the isothermal multidirectional forging (MDF) in a large cumulative strain and temperature range. The structure investigation of the studied alloy revealed several phases precipitated during solidification, among which θ(Al2Cu), Q(Al5Cu2Mg8Si6), Mg2Si, Sc-bearing W(AlScCu) and V(AlSi2Sc2) phases were observed. The MDF at 150–350 °C and a maximum cumulative strain of 14.4 significantly refined grain structure providing a mean grain size of 1.2–2.1 µm. The L12 structured Al3(Sc,Zr) dispersoids with a mean size of 10 ± 1 nm were formed during two-step homogenization annealing. Due to Zener pinning of the nanoscale dispersoids and fine-grained structure, the alloy exhibited near-superplastic behavior in a temperature range of 460–500 °C and strain rate range of 2 × 10−3–1 × 10−2 s−1 with the maximum elongation to failure of ~300%. After a strengthening heat treatment, the forged alloy exhibited the yield strength of 326 ± 5 MPa, ultimate tensile strength of 366 ± 5 MPa, and elongation of 10 ± 3%. The hot deformation behavior was described using the Arrhenius type model. The developed model demonstrated high predictability accuracy with a maximum average absolute relative error of 6.6%.

Computation of flow rates in rarefied gas flow through circular tubes via machine learning techniques

Kinetic theory and modeling have been proven extremely suitable in computing the flow rates in rarefied gas pipe flows, but they are computationally expensive and more importantly not practical in design and optimization of micro- and vacuum systems. In an effort to reduce the computational cost and improve accessibility when dealing with such systems, two efficient methods are employed by leveraging machine learning (ML). More specifically, random forest regression (RFR) and symbolic regression (SR) have been adopted, suggesting a framework capable of extracting numerical predictions and analytical equations, respectively, exclusively derived from data. The database of the reduced flow rates W used in the current ML framework has been obtained using kinetic modeling and it refers to nonlinear flows through circular tubes (tube length over radius l∈[0,5]\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$l \\in [0,5]$$\\end{document} and downstream over upstream pressure p∈[0,0.9]\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$p \\in [0,0.9]$$\\end{document}) in a very wide range of the gas rarefaction parameter δ∈[0,103]\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\delta \\in [0,10^3]$$\\end{document}. The accuracy of both RFR and SR models is assessed using statistical metrics, as well as the relative error between the ML predictions and the kinetic database. The predictions obtained by RFR show very good fit on the simulation data, having a maximum absolute relative error of less than 12.5%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$12.5\\%$$\\end{document}. Various expressions of the form of W=W(p,l,δ)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$W=W(p,l,\\delta )$$\\end{document} with different accuracy and complexity are acquired from SR. The proposed equation, valid in the whole range of the relevant parameters, exhibits a maximum absolute relative error less than 17%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$17\\%$$\\end{document}. To further improve the accuracy, the dataset is divided into three subsets in terms of δ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\delta$$\\end{document} and one SR-based closed-form expression of each subset is proposed, achieving a maximum absolute relative error smaller than 9%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$9\\%$$\\end{document}. Very good performance of all proposed equations is observed, as indicated by the obtained accuracy measures. Overall, the present ML-predicted data may be very useful in gaseous microfluidics and vacuum technology for engineering purposes.

Design optimization of soft robotic fingers biologically inspired by the fin ray effect with intrinsic force sensing

Soft grippers display adaptability, dexterity, and safety, garnering increased interest in their applications. However, accurately modeling their nonlinear deformation and sensing the grasping force remain challenging. Here, a generalized closed-form kinetostatic model of a biomimetic FRE soft finger is established based on the chained-beam-constraint model. A methodology combining a geometry-based contact constraint technique is presented, utilizing the minimum strain energy principle for sensorless, precise, and rapid estimation of the deformed contours and contact force when the FRE finger grasps an arbitrary object with a continuous surface. Method verification using FEA demonstrates a maximum error of no more than 1.28% and 2.44% for the contour and contact force estimation, respectively. Subsequently, a multiobjective genetic algorithm is used to optimize the FRE finger design for better adaptability and greater grasping force in a special scenario. Finally, the finger's capability to accurately perceive grasping force and deformed contours was demonstrated by four groups of grasping tests using cylinder and cube objects, with maximum average absolute error and relative error of 0.083 N, 9.84%, and 1.335 mm, 10.06%, respectively.

English

With the rapid development of economy, the performance appraisal of human resource management in agricultural enterprises has gained more attention. Based on the convolutional neural network theory, this paper constructs the benefit evaluation system of human resource management in agricultural enterprises, determines the weight to obtain the comprehensive benefit result, and then obtains the evaluation level. According to the main factors affecting the human resource benefit of agricultural enterprises, the model designed the input and output indexes of the efficiency evaluation of single well of production capacity construction, and solved the problem of measuring the informatization ability of human resource management in agricultural enterprises. In the simulation process, the convolutional neural network designed the evaluation index of the index system. According to the different contribution and importance degree to the system security, the difference between the evaluation index can be expressed by assigning different weight values. Secondly, for items with similar loads on multiple factors or with low loads on a single factor, it was adopted to obtain a scale consisting of 54 items, including 15 items in interpersonal skills sub-scale, 16 items in learning development sub-scale and 23 items in growth driver sub-scale. The experimental results show that the convolutional neural network training can obtain the reasonable scale value of the test sample management area. The evaluation index of the model prediction effect shows that the absolute relative error, maximum absolute relative error and average absolute relative error are all within 5%, and the equality coefficient is 0.9845, which is greater than 0.9, indicating that the reasonable scale predicted value has a high degree of fitting with the expected value. The prediction results of the model are ideal, which effectively improves the accuracy of benefit evaluation

Experimental and numerical study on cooling performance of a novel earth-to-air heat exchanger system with an inlet plenum chamber

Previous studies have focused on the horizontal buried pipe section regarding the performance improvement of the earth-to-air heat exchanger, showing limited attention paid to the vertical air inlet section. In this study, we proposed a novel earth-to-air heat exchange system utilising an inlet plenum chamber (IPC-EAHE) to pre-treat the outdoor air in the inlet section. To evaluate its cooling performance, a numerical model was developed and validated based on a field experiment. The results showed a good agreement between simulated and experimental data, with a maximum absolute relative error of 4.43 % for the outlet air temperature. Then comparative studies between the IPC-EAHE and traditional EAHE are conducted under three airflow rates. Results indicated that the IPC-EAHE has a lower and more stable outlet air temperature than traditional EAHE. Compared to the traditional EAHE, the IPC-EAHE reduces the fluctuation of outlet air temperature by 53.13 %, 29.84 %, and 26.60 % under the airflow rates of 190 m3/h, 380 m3/h, and 570 m3/h, respectively. This improves the free cooling potential of EAHE for buildings. Moreover, the IPC-EAHE can reduce the buried pipe length by around 60.30 %, 53.0 %, and 50.0 % for three airflow rates, respectively. The friction resistance loss of the IPC section is at least 20 % less than that of the traditional EAHE vertical inlet section. The above results demonstrate that the IPC-EAHE offers greater renewable energy utilization efficiency and less buried pipe length than the traditional EAHE.

Investigation of pressure wave propagation and attenuation characteristics in managed pressure drilling by fast switching throttle valve

Pressure waves possess many significant applications in the oil and gas drilling engineering field, such as mud pulse telemetry (MPT) and measurement while drilling (MWD). The focus of this research is to study the pressure wave propagation and attenuation characteristics of wellbore liquid-phase flow in managed pressure drilling (MPD) by fast switching throttle valve (FSTV). First, a mathematical model of transient pressure wave propagation along the wellbore in both upstream and downstream directions is proposed in MPD by FSTV based on the one-dimensional transient flow theory. The model considering the frictional shear effect between the pipe walls is solved by utilizing the method of characteristics. Meanwhile, boundary conditions at the drill string inlet and annulus outlet, at the throttle valve, at the junction of drill bit, and at the reducer are adequately taken into account according to the actual situation of fluid flow. Second, a laboratory experiment of excited pressure waves in a vertical wellbore is conducted to measure the variation of the pressure fluctuation with different pump rates by FSTV. Comparing with the measurement result, the calculation result is discovered that the overall change is consistent, where the maximum absolute relative error at the peak of the pressure wave is only 4.5%. Finally, it further analyzes influential factors affecting the propagation and attenuation behaviors of wellbore pressure waves in liquid-phase flow based on the model. The results indicate that pressure waveforms present sinusoidal wave propagation, and pump rates, pressure wave speed, excitation time, fluid type, mud density, fluid viscosity, and borehole size exert varying degrees of influence on downhole pressure fluctuations. The proposed model achieves accurate quantitative interpretation and analysis of downhole pressure wave in MPD by FSTV, which has important significance for the realization of safe and efficient drilling.

Online Model Parameter and State of Charge Estimation of Li-Ion Battery Using Unscented Kalman Filter Considering Effects of Temperatures and C-Rates

This paper proposed a technique for online modeling and state of charge (SoC) estimation of Li-ion batteries considering the effects of operating temperatures and C-rates. The proposed battery model, which is updated in real-time, is incorporated with the unscented Kalman filter (UKF) for accurate and online SoC estimation. In the proposed technique, Thevenin-based equivalent circuit model of Li-ion battery is used. Each of the RC networks of the Thevenin model is represented as a first-order linear time-invariant (LTI) system. The offline RC parameters are extracted as the parameters of the first-order LTI system using the final value theorem. Then the offline RC parameters are converted to the online model parameters compensating the effects of operating temperatures and C-rates. Finally, the UKF technique is augmented with the proposed online battery model for higher accurate and online SoC estimation. An experimental setup has been developed in the LabVIEW platform to validate the proposed battery model and apply it for online SoC estimation. The proposed online battery model and UKF are developed in Matlab and then integrated with the LabVIEW program using the Matlab-LabVIEW interface called Math Script. The proposed technique has been rigorously validated in the laboratory under hybrid pulse power characterization (HPPC) and federal urban driving schedule (FUDS) tests under wide range of varying operating temperatures and C-rates. In the HPPC test, the SoC estimated using the proposed technique has a maximum absolute relative percentage error (ARPE) of less than 2%, whereas the battery model that does not account for temperature and C-rate effects has a maximum ARPE of 8.79 percent, which is 4.30 times higher than the proposed technique. The proposed approach yielded mean ARPE of 0.30%, 1.02%, and 0.81% in the FUDS test at 25 °C, 0 °C, and 40 °C, respectively, which is performed in dynamic identical C-rates. Furthermore, higher accuracy and effectiveness of the proposed technique have been validated under the presence of white and coloured measurement noises.

Numerical analysis of liquid flow measurement using multipath ultrasonic phased array flowmeter

In this paper, multipath ultrasonic flowmeter using phased array transducer is introduced to measure the flowrate. In the proposed flowmeter, the path velocities are achieved by dynamic beam steering via phased array sensor then numerical integration is used to calculate the flowrate. Turbulent flow profile through a straight pipe, flow field downstream of single bend, and double bend are studied in this paper. For each case, the influence of the number of paths, beam angle, and installation angle on the performance of the proposed flowmeter is discussed. Because of symmetric velocity profile for turbulent flow through the straight pipe, relative errors are not dependent on the installation angle and its minimum value is obtained 0.03 % under [±56.6°, 0] path arrangement. Relative errors after single and double bend depend on the installation angle due to the asymmetric flow field. In 3-path configuration, for velocity profile after single bend and double bend, the minimum error range is 1.35 % and 2.38 % under [±55o, 0] and [±45o, 0] path arrangement, respectively. In 5-path arrangement, after single bend and double bend, the minimum error range is achieved by 0.14 % and 1.03 % under [±70o, ±35o, 0] and [±65o, ±25o, 0], respectively. The validity of multipath ultrasonic phased array flowmeter after the single bend in Re = 2 × 105 to 106 has been examined with the computational fluid dynamics simulation. The maximum absolute relative error under [±55o,0] and [±70o,±35o,0] was achieved at about 3 % and 1.8 % respectively.

Estimating FAO Blaney-Criddle b-Factor Using Soft Computing Models

FAO Blaney-Criddle has been generally an accepted method for estimating reference crop evapotranspiration. In this regard, it is inevitable to estimate the b-factor provided by the Food and Agriculture Organization (FAO) of the United Nations Irrigation and Drainage Paper number 24. In this study, five soft computing methods, namely random forest (RF), M5 model tree (M5), support vector regression with the polynomial function (SVR-poly), support vector regression with radial basis function kernel (SVR-rbf), and random tree (RT), were adapted to estimate the b-factor. And Their performances were also compared. The suitable hyper-parameters for each soft computing method were investigated. Five statistical indices were deployed to evaluate their performance, i.e., the coefficient of determination (r2), the mean absolute relative error (MARE), the maximum absolute relative error (MXARE), the standard deviation of the absolute relative error (DEV), and the number of samples with an error greater than 2% (NE > 2%). Findings reveal that SVR-rbf gave the highest performance among five soft computing models, followed by the M5, RF, SVR-poly, and RT. The M5 also derived a new explicit equation for b estimation. SVR-rbf provided a bit lower efficacy than the radial basis function network but outperformed the regression equations. Models’ Applicability for estimating monthly reference evapotranspiration (ETo) was demonstrated.

Smart Insole Based on Flexi Force and Flex Sensor for Monitoring Different Body Postures.

This study aims to fabricate smart insoles using wireless Flexi force and bend sensing technology. Polyvinyl chloride (PVC) film was chosen as the substrate to hold all the sensors. The developed smart insole has a three-layer structure (insole-PVC layer-fabric layer) and is calibrated in an isolation laboratory to evaluate its measurement performance. One male volunteer subject exhibited four different body postures, namely tree pose, forward-leaning, squatting, and forward folding pose. Changes in pressure distribution were considered to be similar for the forward, squat, and forward-folded positions. When subjects performed a full squat, the flex sensor exhibited maximum flexion during the squat position, and the flex sensor response against the squat pose was found to be higher by about 18.18% than in the forward lean, respectively. The tree pose has the highest error rate at the first metatarsal, about 18.6%, of which the maximum absolute relative error of the sensor is less than 5%. Plantar pressure distribution and body posture measurements were successfully validated using Flexi force and flex sensors embedded in the smart insole. The smart insole proposed in this research work has broader prospects for clinical application.

Hybrid Models for Solving the Colebrook–White Equation Using Artificial Neural Networks

This study proposes hybrid models to solve the Colebrook–White equation by combining explicit equations available in the literature to solve the Colebrook–White equation with an error function. The hybrid model is in the form of fH=fo−eA. fH is the friction factor value f predicted by the hybrid model, fo is the value of f calculated using several explicit formulas for the Colebrook–White equation, and eA is the error function determined using the neural network procedures. The hybrid equation consists of a series of hyperbolic tangent functions whose number corresponds to the number of neurons in the hidden layer. The simulation results showed that the hybrid models using five hyperbolic tangent functions could produce reasonable predictions of friction factors, with the maximum absolute relative error (MAXRE) around one tenth, or ten times lower than that produced by the corresponding existing formula. The simplified hybrid models are also given using four and three tangent hyperbolic functions. These simplified models still provide accurate results with MAXRE of less than 0.1%.

Optimizing the Sampling Area across an Old-Growth Forest via UAV-Borne Laser Scanning, GNSS, and Radial Surveying

Aboveground biomass, volume, and basal area are among the most important structural attributes in forestry. Direct measurements are cost-intensive and time-consuming, especially for old-growth forests exhibiting a complex structure over a rugged topography. We defined a methodology to optimize the plot size and the (total) sampling area, allowing for structural attributes with a tolerable error to be estimated. The plot size was assessed by analyzing the semivariogram of a CHM model derived via UAV laser scanning, while the sampling area was based on the calculation of the absolute relative error as a function of allometric relationships. The allometric relationships allowed the structural attributes from trees’ height to be derived. The validation was based on the positioning of a number of trees via total station and GNSS surveys. Since high trees occlude the GNSS signal transmission, a strategy to facilitate the positioning was to fix the solution using the GLONASS constellation alone (showing the highest visibility during the survey), and then using the GPS constellation to increase the position accuracy (up to PDOP~5−10). The tree heights estimated via UAV laser scanning were strongly correlated (r2 = 0.98, RMSE = 2.80 m) with those measured in situ. Assuming a maximum absolute relative error in the estimation of the structural attribute (20% within this work), the proposed methodology allowed the portion of the forest surface (≤60%) to be sampled to be quantified to obtain a low average error in the calculation of the above mentioned structural attributes (≤13%).

Anomaly Detection of Disconnects Using SSTDR and Variational Autoencoders

This article utilizes variational autoencoder (VAE) and spread spectrum time domain reflectometry (SSTDR) to detect, isolate, and characterize anomalous data (or faults) in a photovoltaic (PV) array. The goal is to learn the distribution of non-faulty input signals, inspect the reconstruction error of test signals, flag anomalies, and then locate or characterize the anomalous data using a predicted baseline rather than a fixed baseline that might be too rigid. The use of VAE handles imbalanced data better than other methods used for classification of PV faults because of its unsupervised nature. We consider only disconnects in this work, and our results show an overall accuracy of 96% for detecting true negatives (non-faulty data), a 99% true positive rate of detecting anomalies, 0.997 area under the ROC curve, 0.99 area under the precision-recall curve, and a maximum percentage absolute relative error of 0.40% in locating the faults on a 5-panel setup with a 59.13 m leader cable.

A comprehensive modeling approach for determining the role and nature of interfacial morphology in mixed matrix membranes

In this study, a new algorithm was developed to mathematically insight into the interfacial morphology in mixed matrix membranes (MMMs) by only using related experimental permeation data to a maximum loading of particles used in MMM preparation. According to this algorithm, the maximum possible interphase thickness is determined based on the size and the maximum used loading of particles in membrane fabrication. The benefits of this approach are determining the optimal characteristic parameters of the surrounded particles including interphase thickness, effective permeability of interphase, and porous particles (if applicable) affected by the blockage, with no arbitrary assumption to them. The validity of the proposed modeling approach was evaluated using 154 data points related to experimental permeability data of 13 series MMMs. Comparing the results obtained by using related experimental data to maximum used particle loading and all loading levels revealed that the sensitivity of the modified Maxwell model to the number of applied experimental data points is negligible. The predictions indicate that the maximum average absolute relative error (AARE) is 9%. Using modified GPG and modified Higuchi models instead of the modified Maxwell model in the proposed modeling approach showed that this algorithm is not sensitive to the type of model.

Intelligent computing of positive switching impulse breakdown voltage of rod-plane air gap based on extremely randomized trees algorithm

The insulating strength of an air gap is an important basis for external insulation design of power transmission and transformation project, and both gap structure and meteorological conditions have a great influence on the breakdown voltage of air gap. The temperature, humidity and other meteorological conditions in the actual operating environment of power transmission and transformation projects may change greatly with time, so presently, there is no better method to calculate the breakdown voltage of air gaps in various meteorological conditions. In this paper, a prediction model for positive switching impulse breakdown voltage of rod-plane air gap based on extremely randomized trees is proposed. All samples are divided into training set and test set by Self-Organizing Mapping neural network and stratified sampling method. In the test set, the predicted breakdown voltage results of rod-plane air gap based on decision trees, random forest, extremely randomized trees and support vector machine are compared, with reference to the rectification results of atmospheric correction method. The result shows that the maximum absolute relative error of the model is 4.3% and the mean absolute percentage error is 1.6%. Then, the importance of features of import variables is calculated with the trained extremely randomized trees model. Finally, the breakdown voltage of rod-plane air gap in correspondence to different gap distances in four extreme meteorological conditions is calculated with the model. The method can provide reference for calculation of the breakdown voltage of air gap in various meteorological conditions.

A prediction method of missing vehicle position information based on least square support vector machine

The continuous development of VANET has accelerated the development of V2X communication. In the DSRC communication mode of VANET, the location information of the vehicles is interfered by factors such as high-density broadcasting and electromagnetic radiation, which can lead to the loss of the original vehicle information data collected by GPS easily. To solve it, this paper proposed the Least Squared SVM based Beacon Data Complete Algorithm. Unlike previous studies that historical trends of vehicle operation were mainly used to predict vehicle location., this method attempts to find a function, which is used to establish the relationship between the lost value and the past value of the vehicle. On this basis, a nonlinear function approximation strategy is used to predict the position of the missing vehicle. Part of the original data was lost artificially to complete checking calculation and to verify the effectiveness of it. The results show that the average relative error between the complemented vehicle position data and the real data is 0.45% and the maximum absolute relative error is 8.25%. This method has the advantage of not needing to extract historical trend data and high calculation accuracy compared with the methods such as PWHOG algorithm, difference matrix, and moving average data preprocessing. It is suitable for real-time acquisition of vehicle position of VANET and can reduce the complexity of detection time.