Transfer Component Analysis Research Articles

Feature Transfer Learning for Fatigue Life Prediction of Additive Manufactured Metals With Small Samples

ABSTRACTA feature transfer learning (FTL)‐based model is proposed to address small‐sample problems in fatigue life prediction of additively manufactured (AM) metals. Transfer component analysis (TCA) is studied for data alignment before model training. Correspondingly, two TCA improvement strategies are further considered to aggregate training data from distinct AM processing conditions. An experimental database consisting of 103 fatigue data is built for model evaluation. The results demonstrate that the proposed model outperforms conventional machine learning models and other transfer learning‐based models in terms of accuracy and data demand, showing good applicability for AM fatigue life assessment.

Similarity analysis to enhance Transfer Learning for Damage Detection

Abstract One significant challenge in machine learning for Structural Health Monitoring (SHM) is reusing previously trained classifiers. A classifier might be suitable for one situation but not for another. Transfer learning techniques try to overcome this difficulty. In SHM, it is common to use the modal parameters as features; however, they are highly influenced by boundary conditions, geometry, and the level of structural damage. This work proposes an approach that performs a similarity analysis to select features before applying transfer learning, aiming at improving classification and damage detection. The reasoning is that a higher similarity leads to a more efficient transfer of learning and, consequently, a better classification. Transfer learning is conducted via the domain adaptation technique known as Transfer Component Analysis (TCA), and cases with low similarity are compared to those with high similarity. Two datasets are analyzed. The first consists of a beam under different boundary conditions, and data is generated through numerical simulations. The second derives from an experimental setup of bolted joints with loosening damage. The proposed strategy, which uses a cosine-type similarity, is shown to improve the transfer learning classification.

Unsupervised domain adaptation for drive-by condition monitoring of multiple railway tracks

Monitoring railway tracks through drive-by vibration data collected by in-service trains offers a cost-effective and adaptable solution for inspecting multiple railway lines. However, numerous existing drive-by monitoring methods rely on supervised learning models, necessitating extensive labelled data for each line. In this paper, a novel framework is proposed based on Unsupervised Domain Adaptation (UDA) concept which facilitates the transfer of a geometric defects diagnosis model learned from one line to a new line without the need for any labelled data from the new line. The proposed framework learns the dynamic-based features that are sensitive to damage and also invariant to different railway tracks. It comprises three components: data pre-processing, UDA implementation, and damage diagnosis. The framework uses the data from the source domain, including corresponding labels, as well as the unlabelled data from the target domain as input. The outputs of the framework consist of the predicted labels for the target domain. The performance of the proposed framework is evaluated using a comprehensive dataset of field measurements of a high-speed train passing 4 different lines within the French high-speed rail network. The proposed UDA framework is implemented using four common UDA algorithms including Information-Theoretical Learning (ITL), Geodesic Flow Kernel (GFK), Transfer Component Analysis (TCA), and Subspace Alignment (SA). The results show that the proposed framework has a 14% increase in the anomaly detection accuracy compared to traditional unsupervised learning methods in which UDA is not used. Furthermore, this study investigates the impact of incorporating a percentage of target data labels during training (semi-supervised domain adaptation), along with various sensor layouts and different tuning parameters, on the accuracy of the proposed approach. The results show that the proposed framework can significantly facilitate the monitoring of railway track conditions using the data collected by in-service trains which could be great interest of railway owners.

Simultaneously predicting SPAD and water content in rice leaves using hyperspectral imaging with deep multi-task regression and transfer component analysis.

Water content and chlorophyll content are important indicators for monitoring rice growth status. Simultaneous detection of water content and chlorophyll content is of significance. Different varieties of rice show differences in phenotype, resulting in the difficulties of establishing a universal model. In this study, hyperspectral imaging was used to detect the Soil and Plant Analyzer Development (SPAD) values and water content of fresh rice leaves of three rice varieties (Jiahua 1, Xiushui 121 and Xiushui 134). Both partial least squares regression and convolutional neural networks were used to establish single-task and multi-task models. Transfer component analysis (TCA) was used as transfer learning to learn the common features to achieve an approximate identical distribution between any two varieties. Single-task and multi-task models were also built using the features of the source domain, and these models were applied to the target domain. These results indicated that for models of each rice variety the prediction accuracy of most multi-task models was close to that of single-task models. As for TCA, the results showed that the single-task model achieved good performance for all transfer learning tasks. Compared with the original model, good and differentiated results were obtained for the models using features learned by TCA for both the source domain and target domain. The multi-task models could be constructed to predict SPAD values and water content simultaneously and then transferred to another rice variety, which could improve the efficiency of model construction and realize rapid detection of rice growth indicators. © 2024 Society of Chemical Industry.

Combining transfer learning and numerical modelling to deal with the lack of training data in data-based SHM

Structural health monitoring (SHM) involves continuously surveilling the performance of structures to identify progressive damage or deterioration that might evolve over time. Recently, machine learning (ML) algorithms have been successfully employed in various SHM applications, including damage detection. However, supervised ML algorithms often require labelled data for multiple possible damage states of the structure for successful damage identification. Although it may be feasible to gather such data for low-value structures, obtaining damage data for expensive structures such as aircraft could be highly challenging. Herein, this data insufficiency is addressed by combining Finite Element (FE) models with domain adaptation, specifically transfer component analysis (TCA) and joint domain adaptation (JDA). The proposed methodology is showcased in two case studies, a Brake–Reuß beam, where damage scenarios correspond to different torque settings on a lap joint and a wingbox laboratory structure where damage is introduced as saw-cuts. Supervised learning algorithms in the form of Artificial Neural Networks (ANNs) and K-Nearest Neighbours (KNNs) are trained based on FE data after domain adaptation is applied and are then tested with the experimental data. It is shown that even though the performance of classifiers in distinct scenarios of dual, three, four and five-class cases is sensitive to choices in the training stage, the use of TCA or JDA allows for the use of FE data for training and significantly reduces the need for expensive experimental damage data to be used for training. These results can pave the way for a broader use of ML algorithms in SHM of critical and/or expensive structures.

Enhancing quantitative 1H NMR model generalizability on honey from different years through partial least squares subspace and optimal transport based unsupervised domain adaptation

Honey is a nourishing and natural food product that is widely favored by a diverse group of consumers. Proton Nuclear Magnetic Resonance (1H NMR) is a powerful tool for quantitative analysis of honey and plays a crucial role in ensuring its quality. The 1H NMR technique necessitates the utilization of multivariate calibration models to facilitate the quantitative analysis of key compounds present in honey. However, maintaining consistent measurement conditions across different years is scarcely possible, which can significantly impact the distribution of training and test spectra, ultimately leading to reduced performance of predictive models. Unsupervised domain adaptation (UDA) methods have gained considerable attention for their ability to match distribution differences between the labeled source spectra and the unlabeled target spectra without costly annotation. To enhance the quantitative model generalizability on honey from different years, we propose a UDA method known as partial least squares subspace and optimal transport-based UDA (PLSS-OT-UDA). This approach eliminates distribution differences between the source subspace and target subspace via partial least squares (PLS) dimensionality reduction and OT. Firstly, the optimal latent variable weight matrix from the source domain (i.e., labeled 1H NMR data in 2017) is extracted with PLS. Next, the dimension of both source and target domains (i.e., unlabeled 1H NMR data in 2018) is reduced and their corresponding subspaces are obtained with weight matrix of the source domain. Finally, OT is then employed to align the distribution of the source and target domains within the subspace. Experimental results on the honey dataset demonstrate that the PLSS-OT-UDA outperforms traditional methods, including transfer component analysis (TCA), optimal transport for domain adaptation (OTDA), domain adaptation based on principal component analysis and optimal transport (PCA-OTDA), and subspace alignment (SA), with respect to generalization performance on three components: baume degree, sugar content, and water content.

Vibration-based Damage Diagnosis for a Population of Composite Structures Under Uncertainty via Hyper-Sphere and Transfer Component Analysis Based Methods

The problem of vibration-based damage diagnosis, including the detection and type characterization of damages in a population of nominally identical composite structures under operational uncertainty is investigated. This is an important problem for various types of fleets, such as aircraft, drones, fixed wing UAVs and others, which becomes highly challenging if the variability in the composite materials and the manufacturing of the population is significant, while the diagnosis of early-stage damages is pursued. Such a population of 27 composite coupons subjecting to operating variability due to alterations in the temperature, the excitation and the experimental setup is considered in this study. Furthermore, early-stage delamination and impact-induced damages are implemented in 10 of the coupons. Damage diagnosis is herein attempted via two data-driven methods that may account for various types of uncertainty, which are founded on a number of stochastic Multiple-Input Single-Output Transmittance Function AutoRegressive with eXogenous pseudo-eXcitation (MISO-TF-ARX) models. The first employs Hyper Spheres for the representation of the healthy subspace under uncertainty using the parameters of the MISO-TF-ARX models and thus abbreviated as HS-MISO-TF, while the other is based on the Transfer Component Analysis of these parameters, abbreviated as TCA-MISO-TF. The methods’ damage diagnosis performance is experimentally assessed using the above composite coupons and a high number of 7 500 inspection test cases for damage detection, whereas damage characterization is based on 450 test cases. The damage detection and characterization results are presented through Receiver Operating Characteristics curves and confusion matrices, respectively. The obtained results indicate adequate damage detection with the HS-MISO-TF method to achieve 82.1% correct detection for 5% of false alarms, with the TCA-MISO-TF following with 76.4% correct detection. Damage characterization is remarkable with the HS-MISO-TF to correctly characterizing the considered damage type for the 100% of the considered test cases and the TCA-MISO-TF following with 92.2%.

Multi-class Damage Detection on Experimental Frames through Transfer Component Analysis

Machine learning techniques have been revealed to be useful in damage identification applications, exploiting, for example, algorithms able to detect anomalies in the monitored data. However, these datasets are not always enough to train a damage classifier reliably. This happens because the lack, or the low number of labelled data, produces outcomes prone to overfitting and bias (due to the low attainable statistical significance of the dataset). To avoid this problem, Transfer Learning technique can be exploited to make up for the lack of data available on a structure (target) using data recorded on another structural system (source), more or less similar, which is rich in data referring to the same structural behavior to be classified (e.g., damage detection in columns due to crack propagation). In this work, a specific sub-application of Transfer Learning, named Transfer Component Analysis technique, is exploited on a benchmark system (numerically or experimentally) represented by a three-story aluminum scaled structure, subjected to increasing damage and mass variation over the three different floors. Special emphasis will be given to how accuracy is affected by data distribution, and in more detail, the authors will show how the increase in accuracy is related to the type of damage to be classified.

Optimisation of steel rolling schedule based on evolutionary multi-tasking transfer algorithm

Strip rolling is an important part of steel processing. The load distribution scheme in the rolling process directly affects production efficiency and product quality. A dynamic rolling process with changing speed is investigated to improve quality and reduce energy use. To balance the iterative calculation of the complex industrial evolution process and the requirements of high real-time, this study examine the rolling process by combining the multivariate, multi-constraint, and strong coupling characteristics of field measured data. On this basis, several conflicting rolling optimisation objectives in the process of rolling schedule optimisation were analysed, and a dimension reduction migration model based on transfer component analysis was established. In case of an inefficient transfer of feasible solutions, a multi-objective multifactorial evolutionary algorithm based on an explicit transfer solution strategy (MOMFEA-ETS) was proposed. The proposed algorithm obtained four average distance optimal values for eight practical rolling programming problems.

Combining spectrum, thermal, and texture features using machine learning algorithms for wheat nitrogen nutrient index estimation and model transferability analysis

The nitrogen nutrition index (NNI) has been extensively applied for the diagnosis of crop nitrogen status, providing insights into efficient nitrogen utilization and plant growth. In this study, we utilized a low-altitude unmanned aerial vehicle (UAV) platform, equipped with multispectral (MS), red–green–blue (RGB), and thermal infrared (TIR) cameras, to comprehensively capture wheat spectral information. The analysis of the relationship between NNI and relative yield revealed an initially linear relationship, which saturated for high NNI values. To enhance accuracy and minimize complexity, we employed a random forest (RF) – recursive feature elimination (RFE) method to select features as inputs for four machine learning (ML) models: back propagation neural network (BPNN), extreme learning machine (ELM), support vector regression (SVR), and Gaussian process regression (GPR). After feature selection, the prediction accuracies of single-sensor models were ranked as: MS > RGB > TIR. The R2 values for the four ML models were in the range of 0.54–0.75. Among multi-sensor combinations, the GPR with MS + RGB + TIR input features achieved the best results with R2 = 0.89 and RPD = 2.52. Further, the dataset was partitioned into six subsets based on location and cultivar variety to evaluate model transferability. The results showed that the transferability largely suffered during the bivariate conditions of different varieties at different locations; the transferability of the model was average improved by 11 % when GPR was combined with transfer component analysis (TCA). The accuracy and transferability of the NNI estimation models significantly improved, offering valuable guidance and methodological support for diagnosing the nitrogen nutrient status of wheat.

Modeling and Optimization of an Enhanced Soft Sensor for the Fermentation Process of Pichia pastoris.

This paper proposes a novel soft sensor modeling approach, MIC-TCA-INGO-LSSVM, to address the decline in performance of soft sensor models during the fermentation process of Pichia pastoris, caused by changes in working conditions. Initially, the transfer component analysis (TCA) method is utilized to minimize the differences in data distribution across various working conditions. Subsequently, a least squares support vector machine (LSSVM) model is constructed using the dataset adapted by TCA, and strategies for improving the northern goshawk optimization (INGO) algorithm are proposed to optimize the parameters of the LSSVM model. Finally, to further enhance the model's generalization ability and prediction accuracy, considering the transfer of knowledge from multiple-source working conditions, a sub-model weighted ensemble scheme is proposed based on the maximum information coefficient (MIC) algorithm. The proposed soft sensor model is employed to predict cell and product concentrations during the fermentation process of Pichia pastoris. Simulation results indicate that the RMSE of the INGO-LSSVM model in predicting cell and product concentrations is reduced by 47.3% and 42.1%, respectively, compared to the NGO-LSSVM model. Additionally, TCA significantly enhances the model's adaptability when working conditions change. Moreover, the soft sensor model based on TCA and the MIC-weighted ensemble method achieves a reduction of 41.6% and 31.3% in the RMSE for predicting cell and product concentrations, respectively, compared to the single-source condition transfer model TCA-INGO-LSSVM. These results demonstrate the high reliability and predictive performance of the proposed soft sensor method under varying working conditions.

Transfer learning in near infrared spectroscopy for stingless bee honey quality prediction across different months

Near infrared spectroscopy (NIRS) coupled with machine learning has demonstrated its capability as an efficient secondary measurement method in various applications. However, a NIRS predictive model might be unable to maintain its accuracy when it is applied to samples that were acquired from different periods. Particularly, the qualities of stingless bee honey are affected by various uncontrolled factors, e.g. ecosystem, origins, and weather. Since it is unrealistic to have a NIRS dataset that can represent unforeseen future changes, an algorithm that can adapt existing data for new samples is worth to be investigated. Thus, this study aims to evaluate the feasibility of homogenous transfer learning approaches to overcome data constraints in developing NIRS predictive models of stingless bee honey qualities across different months. First, the near infrared transmittance spectra were acquired across honey samples prior to their respective conventional quality measurements in four different analysis months. Then, data from different analysis months were used as source and target datasets. Next, joint distribution adaptation based partial least square (JDA-PLS) and transfer component analysis based PLS (TCA-PLS) were implemented to establish NIRS predictive models of moisture, hydroxymethylfurfural (HMF), and glucose quality. Results show that the proposed TCA-PLS outperformed JDA-PLS and PLS by achieving the lowest RMSEP of 3.401 % and 0.804 mg/kg in predicting the moisture and glucose of stingless bee honey, respectively, across different months. Lastly, finding shows that both JDA and TCA were inefficacy when the label distributions of the calibration and validation datasets were similar.

Combining transfer learning and hyperspectral imaging to identify bruises of pears across different bruise types.

Mechanical bruise is one of the most crucial factors affecting the quality of pears, which has a huge influence on postharvest transportation, storage, and sale of pears. To rapidly detect early bruises of pears across different bruise types, hyperspectral imaging technology coupled with transfer learning methods was performed in this study. Two transfer learning methods, that is, transfer component analysis (TCA) and manifold embedded distribution alignment (MEDA), were applied for two tasks (impact bruise→crush bruise, crush bruise→impact bruise). Supporting vector machine (SVM) was set as a baseline to conduct analysis and comparison of the transferability of the models. The result showed that, for task 1 (impact bruise→crush bruise), MEDA and TCA-SVM model achieved a classification accuracy of 93.33% and 91.11% in target domain, individually. For task 2 (crush bruise→impact bruise), MEDA and TCA-SVM model achieved an accuracy of 88.89% and 85.19% in target domain, respectively. Both the two models improved the accuracy compared with SVM models (84.44% for task 1; 77.04% for task 2). Overall, the results indicated that transfer learning approaches could perform pear bruise detection across different bruise types. Hyperspectral imaging in combination with transfer learning methods is a promising possibility for the efficient and cost-saving field detection of fruit bruises among different bruise types. PRACTICAL APPLICATION: The production and export of pears are faced with problems of mechanical damage due to vibration, collision, impact, and other factors, which cause chemical changes in color, odor, and taste. Sometimes the bruise was too slight to be ignored which would infect with other fruits in the future. In this study, we used hyperspectral imaging combined with transfer learning method could detect these slight bruises caused by different factors. Distinguishing different types of damage can provide a reference for quick judgment of the process causing damage and take prompt measures to reduce economic losses.

Transfer component analysis-driven domain adaptation approach for estimating the life of tunnel boring machine disc cutters

Estimating the life of disc cutters in tunnel boring machines is crucial for tunnel excavation efficiency. Accordingly, generating accurate predictions of the state of disc cutters can prevent unexpected downtimes while optimizing cutting performance. The literature presents investigations of the application of machine learning (ML) for estimating the disc cutter life. Nevertheless, diverse excavation conditions and data scarcity have stimulated research on transfer learning (TL) frameworks that tailor models for different domains, that is, domain adaptation (DA), by generating a common subspace in which the discrepancies among them are minimized. This study presents an end-to-end framework to estimate the disc cutter life through transfer component analysis, a DA-based method for projecting different domains by considering excavation project datasets in a common space; then, different ML models are trained, including linear, probabilistic, and neural methods. The results revealed that the Gaussian process regression model was superior to other methods, with R2 = 0.99. In conclusion, TL is a valuable tool for estimating disc cutter life across different domains, enabling accurate predictions and leveraging knowledge to mitigate data scarcity challenges. In addition, an explainable AI analysis is conducted to quantify and identify relevant features impacting negatively on the disc cutter life, evidencing great contribution from the cutter rotation speed, quartz content, screw rate and specific energy.

Cross-Project Defect Prediction Using Transfer Learning with Long Short-Term Memory Networks

With the increasing number of software projects, within-project defect prediction (WPDP) has already been unable to meet the demand, and cross-project defect prediction (CPDP) is playing an increasingly significant role in the area of software engineering. The classic CPDP methods mainly concentrated on applying metric features to predict defects. However, these approaches failed to consider the rich semantic information, which usually contains the relationship between software defects and context. Since traditional methods are unable to exploit this characteristic, their performance is often unsatisfactory. In this paper, a transfer long short-term memory (TLSTM) network model is first proposed. Transfer semantic features are extracted by adding a transfer learning algorithm to the long short-term memory (LSTM) network. Then, the traditional metric features and semantic features are combined for CPDP. First, the abstract syntax trees (AST) are generated based on the source codes. Second, the AST node contents are converted into integer vectors as inputs to the TLSTM model. Then, the semantic features of the program can be extracted by TLSTM. On the other hand, transferable metric features are extracted by transfer component analysis (TCA). Finally, the semantic features and metric features are combined and input into the logical regression (LR) classifier for training. The presented TLSTM model performs better on the f-measure indicator than other machine and deep learning models, according to the outcomes of several open-source projects of the PROMISE repository. The TLSTM model built with a single feature achieves 0.7% and 2.1% improvement on Log4j-1.2 and Xalan-2.7, respectively. When using combined features to train the prediction model, we call this model a transfer long short-term memory for defect prediction (DPTLSTM). DPTLSTM achieves a 2.9% and 5% improvement on Synapse-1.2 and Xerces-1.4.4, respectively. Both prove the superiority of the proposed model on the CPDP task. This is because LSTM capture long-term dependencies in sequence data and extract features that contain source code structure and context information. It can be concluded that: (1) the TLSTM model has the advantage of preserving information, which can better retain the semantic features related to software defects; (2) compared with the CPDP model trained with traditional metric features, the performance of the model can validly enhance by combining semantic features and metric features.

Grinding Surface Roughness Measurement Combined with Simulation Data and Transfer Learning

AbstractThe surface‐roughness machine vision measurement method requires training numerous samples with uniform roughness values to construct a prediction model. Insufficient actual sample size will make the model lack generalization ability. So, a grinding surface‐roughness measurement method combined with simulation data and transfer learning is proposed. First, through surface simulation, inverse fitting, and three‐dimensional (3D) modeling, more realistic simulation samples are obtained, real samples are prepared, and a simulation imaging system and an actual imaging experiment to acquire surface images are designed. Second, the data distribution difference between the simulation domain and the actual domain through transfer component analysis and the use of the adjusted simulation‐domain data and its label to train a support vector regression model is adjusted. Finally, the model predicts the surface roughness of the uncalibrated actual‐domain sample. For shared features of transferable simulation images and actual images, using the high sensitivity of color information, image features are enhanced and a proportion of pure color region index is designed, strongly correlated within the domain but less differentiated between domains. The results show that the transfer learning prediction accuracy based on this index exceeds 90%, verifying the feasibility of the proposed method and providing a new improvement strategy for sample expansion in visual roughness measurement.

A Bearing Fault Diagnosis Method Based on Improved Transfer Component Analysis and Deep Belief Network

Domain adaptation can handle data distribution in different domains and has been successfully applied to bearing fault diagnosis under variable working conditions. However, most of these methods ignore the influences of noise and data distribution discrepancy on marking pseudo labels. Additionally, most domain adaptive methods require a large amount of data and training time. To overcome the aforementioned challenges, firstly, sample rejection and pseudo label correction using K-means (SRPLC-K-means) were developed and explored to filter the noisy samples and correct the pseudo labels to obtain pseudo labels with higher confidence. Furthermore, a bearing fault diagnosis method based on the improved transfer component analysis and deep belief network is proposed, which can achieve subdomain adaptation and improve the compactness of the samples, leading to a complete bearing fault diagnosis under variable working conditions that is faster and more accurate. Finally, the results of the comparative tests confirmed that the proposed method could boost the average accuracy of 0.73%, 0.99%, and 5.55% in the three tests than the state-of-the-art methods, respectively. Moreover, the comparison of the time required for a fault diagnosis using different methods shows that compared to the end-to-end models, the proposed method reduces the time required by 594.9 s and 1431.6 s, respectively.

Maturity detection of single maize seeds based on hyperspectral imaging and transfer learning

Seed maturity is closely related to seed germination, seed storage and the yield of the next generation of crops and it is an important indicator of seed quality. Maize is an important food, feed, and industrial raw material crop in China. Hyperspectral imaging technology has the advantages of spectral technology and machine vision technology and it can obtain information inside and outside the seed without damaging it. On the basis of hyperspectral imaging technology, transfer learning algorithms were used in this study to explore the transfer feasibility of maize seed maturity models amongst different varieties. Three transfer tasks were designed between different varieties on the basis of the model established for the Xianyu 335 variety. Cross-variety detection models were constructed using the classical machine-learning algorithm partial least square-linear discriminant analysis (PLS-LDA), the transfer learning algorithm transfer component analysis (TCA), and manifold embedded distribution alignment (MEDA). Results showed that the PLS-DA direct transfer model had varying degrees of model failure in all three transfer tasks. However, the average recognition accuracies of the TCA migration model and the MEDA migration model were 89.3% and 91.2%, respectively. Both of the two Transfer learning models can effectively solve the problem of different data feature distributions amongst different varieties, making the model more universal. It solves the problem of transferring the single-grain maize seed maturity detection model amongst different varieties, providing effective, stable and universal technical support for maize seed detection, which is conducive to the promotion of single-grain sowing technology and the improvement of agricultural automation level.

Performance prediction of industrial robot harmonic reducer via feature transfer and Gaussian process regression

This paper addresses the problem of identifying faults in the harmonic reducers of industrial robots by analysing their vibration signals. In order to solve the problem of obtaining fault data and rotation error from harmonic reducers in service, an accuracy performance prediction method based on transfer learning and Gaussian process regression (GPR) is proposed. The Euclidean distance between the spectral sequence of each component is proposed as the fitness index to optimise the transition bandwidth of the filter banks. The optimised empirical wavelet transform (OEWT) is used for signal decomposition to obtain sensitive frequency bands. A feature transfer method based on semi-supervised transfer component analysis (SSTCA) is proposed to achieve target domain feature transfer under missing data conditions. A prediction model based on GPR is established using the mapped features to predict the performance and accuracy of the harmonic reducer. The effectiveness of the proposed method is verified through model evaluation indicators and degradation experiments.

Cross-Condition Fault Diagnosis of an Aircraft Environmental Control System (ECS) by Transfer Learning

Fault diagnosis models based on machine learning are often subjected to degradation in performance when dealing with data that are differently distributed than the training data. Such an occasion is common in reality because machines usually operate under various conditions. Transfer learning is a solution for the performance degradation of cross-condition fault diagnosis problems. This paper studies how transfer learning algorithms transfer component analysis (TCA) and joint distribution alignment (JDA) improve the cross-condition fault diagnosis accuracy of an aircraft environmental control system (ECS). Both methods work by transforming the source and target domain data into a feature space where their distributions are aligned to allow a uniform classifier to act accurately in both domains. This paper discovered that both TCA and JDA produce significantly more accurate results than traditional methods on target domains with unlabelled ECS data taken at different operating conditions than the source domain. Additionally, when dealing with unlabelled data from unknown conditions bearing a different composition of classes in the target domain, TCA is found to be more robust and accurate, generating an average predictive accuracy of 95.22%, which demonstrates the ability of transfer learning in solving similar problems in the real-world application of fault diagnosis.