Basis Function Support Vector Machine Research Articles

Fault Diagnosis for Wind Turbines Based on ReliefF and eXtreme Gradient Boosting

In order to improve the accuracy of fault diagnosis on wind turbines, this paper presents a method of wind turbine fault diagnosis based on ReliefF algorithm and eXtreme Gradient Boosting (XGBoost) algorithm by using the data in supervisory control and data acquisition (SCADA) system. The algorithm consists of the following two parts: The first part is the ReliefF multi-classification feature selection algorithm. According to the SCADA history data and the wind turbines fault record, the ReliefF algorithm is used to select feature parameters that are highly correlated with common faults. The second part is the XGBoost fault recognition algorithm. First of all, we use the historical data records as the input, and use the ReliefF algorithm to select the SCADA system observation features with high correlation with the fault classification, then use these feature data to build the XGBoost multi classification fault identification model, and finally we input the monitoring data generated by the actual running wind turbine into the XGBoost model to get the operation status of the wind turbine. We compared the algorithm proposed in this paper with other algorithms, such as radial basis function-Support Vector Machine (rbf-SVM) and Adaptive Boosting (AdaBoost) classification algorithms, and the results showed that the classification accuracy using “ReliefF + XGBoost” algorithm was higher than other algorithms.

Accurate extraction of offshore raft aquaculture areas based on a 3D-CNN model

ABSTRACT Offshore aquaculture plays an important role in China’s marine fishery economy. The research on the extraction of offshore aquaculture areas based on remote sensing technology is of great significance for the regulation of offshore fishery resources and the protection of the marine ecological environment. This paper uses the Gaofen-2 series multispectral remote sensing image to extract the offshore aquaculture areas of Lianyungang City. We use the optimum index factor to extract the spectral features of the aquaculture areas and the grey-level co-occurrence matrix to extract their texture features. We use the Bhattacharyya distance to select the spatial and spectrum features and construct the characteristic data set of the aquaculture areas. In this paper, we propose a method to construct a uniform distributed disturbance term to optimize the cross entropy loss function. We employ it in the three-dimensional convolutional neural network (3D-CNN) model, extract the extended feature data set of aquaculture areas, and input it into the radial basis function support vector machine (RBF-SVM) classifier for classification. Within the study area of 150 km2, the experimental results show that the extraction model has high extraction accuracy and strong spatial migration despite complex water backgrounds. The F 1-score values in the training area and the four random test areas were 0.939 or above for 2017 data. In addition, the extraction model also has stable time migration. We used the extraction model on remote sensing data for the same study area in 2018 and 2019. The F 1-scores for all test areas are 0.866 or higher. Therefore, the model proposed in this paper is suitable for the extraction of large-scale and multi-temporal offshore raft aquaculture areas from remote sensing images.

PERBAIKAN KONTRAS CITRA MAMMOGRAM PADA KLASIFIKASI KANKER PAYUDARA BERDASARKAN FITUR GRAY-LEVEL CO-OCCURRENCE MATRIX

In this era to recognize breast tumors can be based on mammogram images. This method will expedite the process of recognition and classification of breast cancer. This research was conducted classification techniques of breast cancer using mammogram images. The proposed model targets classification studies for cases of malignant, and benign cancer. The research consisted of five main stages, preprocessing, histogram equalization, convolution, feature extraction, and classification. For preprocessing cropping the image using region of interest (ROI), for convolution, median filter and histogram equalization are used to improve image quality. Feature extraction using Gray-Level Co-Occurrence Matrix (GLCM) with 5 features, entropy, correlation, contrast, homogeneity, and variance. The final step is the classification using Radial Basis Function Neural Network (RBFNN) and Support Vector Machine (SVM). Based on the hypotheses that have been tested and discussed, the accuracy for RBFNN is 86.27%, while the accuracy for SVM is 84.31%. This shows that the RBFNN method is better than SVM in distinguishing types of breast cancer. These results prove the process of improving image construction using histogram equalization and the median filter is useful in the classification process.

Bio-inspired dimensionality reduction for Parkinson's disease (PD) classification.

Given the demand for developing the efficient Machine Learning (ML) classification models for healthcare data, and the potentiality of Bio-Inspired Optimization (BIO) algorithms to tackle the problem of high dimensional data, we investigate the range of ML classification models trained with the optimal subset of features of PD data set for efficient PD classification. We used two BIO algorithms, Genetic Algorithm (GA) and Binary Particle Swarm Optimization (BPSO), to determine the optimal subset of features of PD data set. The data set chosen for investigation comprises 756 observations (rows or records) taken over 755 attributes (columns or dimensions or features) from 252 PD patients. We employed MaxAbsolute feature scaling method to normalize the data and one hold cross-validation method to avoid biased results. Accordingly, the data is split in to training and testing set in the ratio of 70% and 30%. Subsequently, we employed GA and BPSO algorithms separately on 11 ML classifiers (Logistic Regression (LR), linear Support Vector Machine (lSVM), radial basis function Support Vector Machine (rSVM), Gaussian Naïve Bayes (GNB), Gaussian Process Classifier (GPC), k-Nearest Neighbor (kNN), Decision Tree (DT), Random Forest (RF), Multilayer Perceptron (MLP), Ada Boost (AB) and Quadratic Discriminant Analysis (QDA)), to determine the optimal subset of features (reduction of dimensionality) contributing to the highest classification accuracy. Among all the bio-inspired ML classifiers employed: GA-inspired MLP produced the maximum dimensionality reduction of 52.32% by selecting only 359 features and delivering 85.1% of the classification accuracy; GA-inspired AB delivered the maximum classification accuracy of 90.7% producing the dimensionality reduction of 41.43% by selecting only 441 features; And, BPSO-inspired GNB produced the maximum dimensionality reduction of 47.14% by selecting 396 features and delivering the classification accuracy of 79.3%; BPSOMLP delivered the maximum classification accuracy of 89% and produced 46.48% of the dimensionality reduction by selecting only 403 features.

RBF-SVM 기반 고관절 보행 보조 외골격 로봇의 보행 모드 판단 알고리즘 개발

The purpose of this study was to suggest the method for automated locomotion modes (Level Walking, Stair Ascent, Stair Descent) detection based on the Radial Basis Function Support Vector Machine (RBF-SVM) for the hip gait assist robot. The universal hip gait assist robot had a limit in detection of the walking intention of users because of the limited sensors’ quantity. Through the offline training, using MATLAB, we trained the collected gait data of users wearing the hip gait assist robot and obtained the parameter of the RBF-SVM model. In the online test, using LabVIEW, we developed the algorithm for the locomotion modes decision of individuals using the optimized parameter of the RBF-SVM. Finally, we executed the gait test for three terrains through the walking environment’s test platform. As a result, the locomotion modes decision rate for three terrains was 98.5%, 99%, and 98% respectively. And the decision delay time of algorithm was 0.03 s, 0.03 s, and 0.06 s respectively.

Two approaches to compilation of bilingual multi-word terminology lists from lexical resources

AbstractIn this paper, we present two approaches and the implemented system for bilingual terminology extraction that rely on an aligned bilingual domain corpus, a terminology extractor for a target language, and a tool for chunk alignment. The two approaches differ in the way terminology for the source language is obtained: the first relies on an existing domain terminology lexicon, while the second one uses a term extraction tool. For both approaches, four experiments were performed with two parameters being varied. In the experiments presented in this paper, the source language was English, and the target language Serbian, and a selected domain was Library and Information Science, for which an aligned corpus exists, as well as a bilingual terminological dictionary. For term extraction, we used the FlexiTerm tool for the source language and a shallow parser for the target language, while for word alignment we used GIZA++. The evaluation results show that for the first approach the F1 score varies from 29.43% to 51.15%, while for the second it varies from 61.03% to 71.03%. On the basis of the evaluation results, we developed a binary classifier that decides whether a candidate pair, composed of aligned source and target terms, is valid. We trained and evaluated different classifiers on a list of manually labeled candidate pairs obtained after the implementation of our extraction system. The best results in a fivefold cross-validation setting were achieved with the Radial Basis Function Support Vector Machine classifier, giving a F1 score of 82.09% and accuracy of 78.49%.

Classifying electrical activity of the brain during imaginary movements of untrained subjects using artificial neural networks

Introduction: Developing new classification methods for human brain electrical activity patterns corresponding to actual movements or motor imagery is an essential interdisciplinary problem in brain-computer interface research. One of the most promising approaches is the development of methods based on artificial neural networks. Purpose: The development of ANN-based methods for classifying electroencephalographic patterns associated with motor imagery in untrained subjects. Methods: Classifiers based on linear neural networks, multi-layer perceptrons, radial basis function networks and support vector machines. Results: The authors selected the optimal type, topology, learning algorithms and parameters of an artificial neural network in order to provide the most accurate and fast classification of lower limb motor imagery EEG signals. It has been studied how the number of the analyzed channels of a multichannel EEG and their choice affect the quality of motor imagery patterns classification. Optimal configurations were obtained for the electrode arrangements. The influence of EEG pre-processing on the accuracy of motor imagery recognition was analyzed. A computational experiment showed the accuracy of 90-95% in untrained subjects. Radial basis function network demonstrated the best performance. Besides, the dataset dimensionality has been significantly reduced down to 6–12 channels without any classification accuracy loss. Practical relevance: The obtained results can be useful for the developers of motor imagery EEG classification algorithms used in brain-computer interfaces.

Ансамбль нейромереж GRNN на підставі зміщених поверхонь відгуку для задач електронної комерції

Розв'язок задач електронної комерції, які здебільшого характеризуються нелінійними поверхнями відгуку, є важливим завданням. Застосування сучасних засобів обчислювального інтелекту не завжди є доречним зважаючи на складність реалізації процедур навчання і налагодження. Неітеративні засоби та нейронні мережі без навчання також не забезпечують задовільної точності результату. З огляду на це у роботі описано новий ансамбль на підставі нейронних мереж узагальненої регресії. Основна ідея розробленого ансамблю полягає в лінеаризації поверхні відгуку, що задається даними наявної вибірки. Для цього отримана за допомогою мережі GRNN поверхня подається на вхід лінійної нейроподібної структури. Така комбінація забезпечує підвищення точності роботи ансамблю під час розв'язання поставленої задачі. Описаний ансамбль застосовано для розв'язання задачі прогнозування ціни на вживані автомобілі. Експериментальним способом підібрано оптимальні параметри його роботи. Шляхом порівняння із відомими методами встановлено найвищу точність його роботи. Результати експериментальних досліджень порівняно з теоретичними оцінками на підставі висновків теореми Кондорсе про журі присяжних. Розроблений ансамбль нейронних мереж узагальненої регресії на підставі зміщення поверхонь відгуку та з додатковим використанням нейроподібних структур Моделі послідовних геометричних перетворень варто застосовувати під час розв'язання різноманітних задач електронної комерції підвищеної точності.

A Comparative Study of Linear and Non-Linear Classifiers in Sensory Motor Imagery Based Brain Computer Interface

Over the period of 4–5 decades of the field electrophysiology on neural signal events in single cell recording, it can be concluded that there is production of far field potentials containing near synchronous field patterns might reach at the scalp electrodes. These neural signals then can be analyzed and converted into the control signals for computers and other electronic devices. Electroencephalography (EEG) is a method to acquire these neural signals from the scalp of human brain. EEG signals are simple, economical and have high temporal resolution properties. These properties make it advantageous to use widely in the medical as well as non-medical applications. The event related synchronization and desynchronization (ERS/ERD) pattern present in EEG during sensory motor imagery (SMI) process over the cortical area is an important feature to take BCI towards realistic approach. The accurate classification of these ERS/ERD pattern present in EEG signal is dependent on classification accuracy of different classifiers. So, the objective of the paper is to analyze the classification accuracy of linear and non-linear classifiers used for BCI system design. This paper also presents the comparative study of linear (Linear discriminant analysis and Support Vector Machine) and non-linear classifiers (Bayesian and Radial Basis Function-Support Vector Machine) using BCI competition IV dataset 2a. The result concluded that linear classifiers (LDA and SVM) have outperformed the non-linear classifiers on EEG data with the average performance across subjects 75.26±12.23, 72.42±11.12.

Human action recognition using descriptor based on selective finite element analysis

Abstract This paper proposes a novel local descriptor evaluated from the Finite Element Analysis for human action recognition. This local descriptor represents the distinctive human poses in the form of the stiffness matrix. This stiffness matrix gives the information of motion as well as shape change of the human body while performing an action. Initially, the human body is represented in the silhouette form. Most prominent points of the silhouette are then selected. This silhouette is discretized into several finite small triangle faces (elements) where the prominent points of the boundaries are the vertices of the triangles. The stiffness matrix of each triangle is then calculated. The feature vector representing the action video frame is constructed by combining all stiffness matrices of all possible triangles. These feature vectors are given to the Radial Basis Function-Support Vector Machine (RBF-SVM) classifier. The proposed method shows its superiority over other existing state-of-the-art methods on the challenging datasets Weizmann, KTH, Ballet, and IXMAS.

Adaptive data-driven nonlinear synchro squeezed transform with single class radial basis function kernel support vector machine applied to wind turbine planetary gearbox fault diagnostics

Planetary stage gears operated at low rotational speed and varying wind speed result variation in load. Variable speed and variable load induce nonstationary operating conditions. Vibration signal measured from Wind power gear transmission systems are embedded with multiple sources of vibration and attenuated considerably as it travels from source of vibration to measuring point. Efficacious multi-component decomposition without mode mixing ensures the accurate fault signature recognition. Synchro squeezing transform is the promising tool that represents the ridges with high resolution in time as well as in frequency axis. An efficient vibration analysis technique, short windowed Fourier synchro squeezing transform with nonlinear radial basis function kernel support vector machine is proposed to detect the mechanical faults in low speed planetary stage of wind turbines. Raw vibration is modeled in time–frequency plane to extract fault pattern signatures effectively with high resolution by adapting an empirical nonlinear tool synchro squeezing transforms. Amplitude modulation and frequency modulation parameters are sculpted from instantaneous amplitude and instantaneous phase, frequency. Hybrid feature space with signal attributes, statistical moments, and randomness measures are extricated from amplitude modulation-frequency modulation components. Single class radial basis function support vector machine is trained with hybrid features. The fault detection accuracy of the proposed method is compared with the standard variants of empirical mode decomposition. The proposed short windowed Fourier synchro squeezing transform-radial basis function kernel support vector machine shows 98.2% accuracy, 98% sensitivity, and 98% specificity.

Diagnosis of Parkinson’s Disease at an Early Stage Using Volume Rendering SPECT Image Slices

The non-degenerative variant called scans without evidence of dopaminergic deficit (SWEDD) is clinically analyzed and wrongly understood as Parkinson’s disease (PD) that results ineffective diagnosis of PD in the early stage. The present work is designed to improve the diagnostic accuracy at the early stage of PD from SWEDD and healthy control (HC). The volume rendering image slices are used as a novel method to achieve better diagnostic accuracy. These image slices are chosen from the single-photon emission computed tomography (SPECT) images based on their striatal uptake region, which contributes appreciated information on the shape of the striatum. Features related to surface and the shape of the striatum are calculated from the segmented region of the chosen image slices to illustrate the good amount of variations among early PD, SWEDD, and HC. Among these feature sets, the most optimized feature is selected using genetic algorithm. The performance of the classifiers like linear, radial basis function-support vector machine (RBF-SVM), extreme learning machine (ELM) activation functions, and RBF-ELM are investigated and compared based on the most optimized feature. It is noted that the RBF-ELM offers better performance with an accuracy of 98.23% than the other classifiers. This also proves that the present work is better than the previous studies. Hence, the proposed approach could act as an aid in the detection of early stage of PD.

Gaussian process regression-based forecasting model of dam deformation

The displacement at various measurement points is a critical indicator that can intuitively reflect the operational properties of a dam. It is important to analyse displacement monitoring data in a timely manner and make reliable predictions of dam safety. This paper proposes a GPR-based model for dam displacement forecasting. The input variables of the monitoring model consider hydraulic factors, thermal factors and irreversible factors, and the output variables are the observed displacements of the dam. An example analysis based on the proposed method is performed on a prototype gravity dam, and the performance of different simple/combined covariance functions is investigated to obtain the optimal choice. Compared to multiple linear regression, radial basis function network (RBFN) and support vector machine (SVM) methods, the results indicate that the GPR-based model with a combined covariance function significantly improves the prediction accuracy. The proposed model can effectively overcome the over-learning and poor robustness issues of approaches such as RBFN and SVM. In addition, the GPR-based forecasting model has the advantages of simplicity in the training process and the capacity to provide a probabilistic output.

Recurrent Neural Networks Based Photovoltaic Power Forecasting Approach

The intermittency of solar energy resources has brought a big challenge for the optimization and planning of a future smart grid. To reduce the intermittency, an accurate prediction of photovoltaic (PV) power generation is very important. Therefore, this paper proposes a new forecasting method based on the recurrent neural network (RNN). At first, the entire solar power time series data is divided into inter-day data and intra-day data. Then, we apply RNN to discover the nonlinear features and invariant structures exhibited in the adjacent days and intra-day data. After that, a new point prediction model is proposed, only by taking the previous PV power data as input without weather information. The forecasting horizons are set from 15 to 90 min. The proposed forecasting method is tested by using real solar power in Flanders, Belgium. The classical persistence method (Persistence), back propagation neural network (BPNN), radial basis function (RBF) neural network and support vector machine (SVM), and long short-term memory (LSTM) networks are adopted as benchmarks. Extensive results show that the proposed forecasting method exhibits a good forecasting quality on very short-term forecasting, which demonstrates the feasibility and effectiveness of the proposed forecasting model.

A CAD System for the Early Detection of Lung Nodules Using Computed Tomography Scan Images

In this paper, a computer-aided detection system is developed to detect lung nodules at an early stage using Computed Tomography (CT) scan images where lung nodules are one of the most important indicators to predict lung cancer. The developed system consists of four stages. First, the raw Computed Tomography lung images were preprocessed to enhance the image contrast and eliminate noise. Second, an automatic segmentation procedure for human's lung and pulmonary nodule canddates (nodules, blood vessels) using a two-level thresholding technique and morphological operations. Third, a feature fusion technique that fuses four feature extraction techniques: the statistical features of first and second order, value histogram features, histogram of oriented gradients features, and texture features of gray level co-occurrence matrix based on wavelet coefficients was utilised to extract the main features. The fourth stage is the classifier. Three classifiers were used and their performance was compared in order to obtain the highest classification accuracy. These are; multi-layer feed-forward neural network, radial basis function neural network and support vector machine. The performance of the proposed system was assessed using three quantitative parameters. These are: the classification accuracy rate, the sensitivity and the specificity. Forty standard computed tomography images containing 320 regions of interest obtained from an early lung cancer action project association were used to test and evaluate the developed system. The images consists of 40 computed tomography scan images. The results have shown that the fused features vector resulting from genetic algorithm as a feature selection technique and the support vector machine classifier give the highest classification accuracy rate, sensitivity and specificity values of 99.6%, 100% and 99.2%, respectively.

Application of a new hybrid particle swarm optimization-mixed kernels function-based support vector machine model for reservoir porosity prediction: A case study in Jacksonburg-Stringtown oil field, West Virginia, USA

Porosity is a fundamental property that characterizes the storage capability of fluid and gas-bearing formations in a reservoir. An accurate porosity value can be measured from core samples in the laboratory; however, core analysis is expensive and time consuming. Well-log data can be used to calculate porosity, but the availability of log suites is often limited in mature fields. Therefore, robust porosity prediction requires integration of core-measured porosity with available well-log suites to control for changes in lithology and fluid content. A support vector machine (SVM) model with mixed kernel function (MKF) is used to construct the relationship between limited conventional well-log suites and sparse core data. Porosity is the desired output, and two conventional well-log responses (gamma ray [GR] and bulk density) and three well-log-derived parameters (the slope of GR, the slope of density, and [Formula: see text]) are input parameters. A global stochastic searching algorithm, particle swarm optimization (PSO), is applied to improve the efficiency of locating the appropriate values of five control parameters in MKF-SVM model. The results of SVM with different traditional kernel functions were compared, and the MKF-SVM model provided an improvement over the traditional SVM model. To confirm the advantage of the hybrid PSO-MKF-SVM model, the results from three models: (1) radial basis function (RBF)-based least-squares SVM, (2) multilayer perceptron artificial neural network (ANN), and (3) RBF ANN, are compared with the result of the hybrid PSO-MKF-SVM model. The results indicate that the hybrid PSO-MKF-SVM model improves porosity prediction with the highest correlation coefficient ([Formula: see text] of 0.9560), the highest coefficient of determination ([Formula: see text] of 0.9140), the lowest root-mean-square error (1.6505), average absolute error value (1.4050), and maximum absolute error (2.717).

Performance evaluation of RBF- and SVM-based machine learning algorithms for predictive mineral prospectivity modeling: integration of S-A multifractal model and mineralization controls

Definition of the efficient ore-forming processes, which are considered as mineralization controls is a fundamental stage in mineral prospectivity modeling. In this contribution, four efficient targeting criteria of geochemical, geological and structural data related to porphyry-type Cu deposits in Varzaghan district, NW Iran, were integrated. For creation of multi-element geochemical layer, a two-stage factor analysis was firstly conducted on ilr-transformed data of 18 selected elements and it was found that factor 1 (F1) is the representative of Cu-Au-Mo-Bi elemental association in the study area. Then, the combined model of multifractal inverse distance weighting (IDW) interpolation technique and spectrum-area (S-A) fractal method of F1 as the significant mineralization-related multi-element geochemical layer was integrated with geological-structural evidence layers. For this purpose, two supervised machine learning algorithms, namely radial basis function (RBF) neural network and support vector machine (SVM) with RBF kernel were used for generating data-driven predictive models of porphyry-Cu mineral prospectivity. Comparison of the generated models demonstrates that the former is more successful in delineating exploration targets than the latter one.

Accurate Identification of the Sex and Species of Silkworm Pupae Using Near Infrared Spectroscopy

The present study proposes a novel method to discriminate the sex and species of silkworm pupae using NIR spectroscopy (800–2778 nm). The spectra from 840 silkworm pupae were collected then divided into a calibration set (700) and a test set (140) using the Kennard–Stone (KS) algorithm. The recognition models were built using the radial basis function and neural network (RBF–NN) and support vector machine (SVM) approaches. The species and sex identification results using the RBF–NN and SVM models based on full spectral data achieved a low accuracy of 5% and 33.57%, respectively. To improve the accuracy and decrease the processing time, both principal component analysis (PCA) and linear discriminant analysis (LDA) were used to reduce the data dimensions. The performance of the optimized SVM model (92.14%) was much better than the RBF–NN model (19.29%) based on PCA. Overall, the best discrimination results were obtained using the RBF–NN and SVM models based on LDA, providing an accuracy of 100%. These promising results have shown that the LDA–SVM and LDA–RBF–NN models can accurately recognize the sex and species of silkworm pupae using NIR spectroscopy.

Spatial modelling with Euclidean distance fields and machine learning

SummaryThis study introduces a hybrid spatial modelling framework, which accounts for spatial non‐stationarity, spatial autocorrelation and environmental correlation. A set of geographic spatially autocorrelated Euclidean distance fields (EDF) was used to provide additional spatially relevant predictors to the environmental covariates commonly used for mapping. The approach was used in combination with machine‐learning methods, so we called the method Euclidean distance fields in machine‐learning (EDM). This method provides advantages over other prediction methods that integrate spatial dependence and state factor models, for example, regression kriging (RK) and geographically weighted regression (GWR). We used seven generic (EDFs) and several commonly used predictors with different regression algorithms in two digital soil mapping (DSM) case studies and compared the results to those achieved with ordinary kriging (OK), RK and GWR as well as the multiscale methods ConMap, ConStat and contextual spatial modelling (CSM). The algorithms tested in EDM were a linear model, bagged multivariate adaptive regression splines (MARS), radial basis function support vector machines (SVM), Cubist, random forest (RF) and a neural network (NN) ensemble. The study demonstrated that DSM with EDM provided results comparable to RK and to the contextual multiscale methods. Best results were obtained with Cubist, RF and bagged MARS. Because the tree‐based approaches produce discontinuous response surfaces, the resulting maps can show visible artefacts when only the EDFs are used as predictors (i.e. no additional environmental covariates). Artefacts were not obvious for SVM and NN and to a lesser extent bagged MARS. An advantage of EDM is that it accounts for spatial non‐stationarity and spatial autocorrelation when using a small set of additional predictors. The EDM is a new method that provides a practical alternative to more conventional spatial modelling and thus it enhances the DSM toolbox.HighlightsWe present a hybrid mapping approach that accounts for spatial dependence and environmental correlation. The approach is based on a set of generic Euclidean distance fields (EDF). Our Euclidean distance fields in machine learning (EDM) can model non‐stationarity and spatial autocorrelation. The EDM approach eliminates the need for kriging of residuals and produces accurate digital soil maps.

Comparison of spatial interpolation approaches for in-core power distribution reconstruction

Abstract This paper investigated the influence of various types of spatial interpolation algorithms in the reactor in-core power distribution reconstruction. These algorithms include different kinds of kernel function used in radial basis function (RBF) methods or support vector machine regression (SVR) methods, different orders of polynomial trend surface analysis (TSA), and various forms of distance weight average (DWA) methods, geo-statistics interpolation method. A typical pressurized water reactor core with 157 fuel assemblies and 33 measurement instruments located is analyzed. The validations of these methods under the measurement core status and predictive core status have been provided. The criterions of relative root mean square error (RRMSE) have been applied to guarantee the accuracy of these algorithms. The comparison of the different spatial interpolation algorithms shows that the DWA basis methods usually perform much better and stable than other methods. PEM and SVR methods have very poor performance in high signal deviation situation, but they can effectively eliminate measurement error in the opposite situation. The fitting methods TPS0, PEM4 could not be used in in-core power distribution reconstruction (IPDR). TPS1 is the best choice for no parameter RBF methods. While for other RBF basis methods, optimization algorithm should be used to search the optimized model parameters. The performance of RBF_TPS1, DWA_OK, SVR_Gauss, RBF_Gauss fall into a same group. Three-dimension surfaces of fitting results are compared. The factors are discussed that affect the reconstructed results of the methods, including detectors number, detectors design pattern and detector measurement properties, variability of the fitting surface. Suggestions to select an appropriate spatial interpolator method are provided.