A multi-model data-fusion based deep transfer learning for improved remaining useful life estimation for IIOT based systems

Early and accurate identification of brain tumors is essential for determining effective treatment strategies and improving patient outcomes. Artificial intelligence (AI) and deep learning (DL) techniques have shown promise in automating diagnostic tasks based on magnetic resonance imaging (MRI). This study evaluates the performance of four pre-trained deep convolutional neural network (CNN) architectures for the automatic multi-class classification of brain tumors into four categories: Glioma, Meningioma, Pituitary, and No Tumor. The proposed approach utilizes the publicly accessible Brain Tumor MRI Msoud dataset, consisting of 7023 images, with 5712 provided for training and 1311 for testing. To assess the impact of data availability, subsets containing 25%, 50%, 75%, and 100% of the training data were used. A stratified five-fold cross-validation technique was applied. The CNN architectures evaluated include DeiT3_base_patch16_224, Xception41, Inception_v4, and Swin_Tiny_Patch4_Window7_224, all fine-tuned using transfer learning. The training pipeline incorporated advanced preprocessing and image data augmentation techniques to enhance robustness and mitigate overfitting. Among the models tested, Swin_Tiny_Patch4_Window7_224 achieved the highest classification Accuracy of 99.24% on the test set using 75% of the training data. This model demonstrated superior generalization across all tumor classes and effectively addressed class imbalance issues. Furthermore, we deployed and benchmarked the best-performing DL model on embedded AI platforms (Jetson AGX Xavier and Orin Nano), demonstrating their capability for real-time inference and highlighting their feasibility for edge-based clinical deployment. The results highlight the strong potential of pre-trained deep CNN and transformer-based architectures in medical image analysis. The proposed approach provides a scalable and energy-efficient solution for automated brain tumor diagnosis, facilitating the integration of AI into clinical workflows.

As one of the most crucial issues in the condition monitoring of mechanical systems, the remaining useful life (RUL) estimation of bearings suffers from the limitation of run-to-failure data, which makes the cross-operating condition RUL estimation of bearings a challenging subject. To address this issue, a cross-operating condition degradation knowledge learning method is proposed by constructing a shared latent feature space across different operating conditions, in which the bearing degradation features are effectively extracted and the cross-domain feature differences are weakened. Focusing on learning the cross-operating condition degradation knowledge, the time-frequency representation (TFR) samples generated by the continuous wavelet transform (CWT) method in both the source and target domains were projected onto a shared latent feature space constructed by a novel joint dictionary matrix factorization (JDMF) method. The joint-domain dictionaries solved by the JDMF method have played a core role in extracting the domain-invariant features that do not change along with the operating conditions of bearings. As a result, the convolutional neural network (CNN) model trained with the source samples in the shared latent feature space can be directly generalized to the target bearings for the degradation indicator (DI) modeling and RUL estimation. The cross-operating condition bearing RUL estimation experiment executed in the PRONOSTIA platform has demonstrated the effectiveness and superiority of the proposed JDMF-CNN framework.

Remaining Useful Life (RUL) estimation of rotating machinery based on their degradation data is vital for machine supervisors. Deep learning models are effective and popular methods for forecasting when rotating machinery such as bearings may malfunction and ultimately break down. During healthy functioning of the machinery, however, RUL is ill-defined. To address this issue, this study recommends using anomaly monitoring during both RUL estimator training and operation. Essential time-domain data is extracted from the raw bearing vibration data, and deep learning models are used to detect the onset of the anomaly. This further acts as a trigger for data-driven RUL estimation. The study employs an unsupervised clustering approach for anomaly trend analysis and a semi-supervised method for anomaly detection and RUL estimation. The novel combined deep learning-based anomaly-onset aware RUL estimation framework showed enhanced results on the benchmarked PRONOSTIA bearings dataset under non-varying operating conditions. The framework consisting of Autoencoder and Long Short Term Memory variants achieved an accuracy of over 90% in anomaly detection and RUL prediction. In the future, the framework can be deployed under varying operational situations using the transfer learning approach.

Ball bearings are one of the most critical components of rotating machines. They ensure shaft support and friction reduction, thus their malfunctioning directly affects the machine’s performance. As a consequence, it is necessary to monitor the health conditions of such a component to avoid major degradations which could permanently damage the entire machine. In this context, HMS (Health Monitoring Systems) and PHM (Prognosis and Health Monitoring) methodologies propose a wide range of algorithms for bearing diagnosis and prognosis. The present article proposes an end-to-end PHM approach for ball bearing RUL (Remaining Useful Life) estimation. The proposed methodology is composed of three main steps: HI (Health Indicator) construction, bearing diagnosis and RUL estimation. The HI is obtained by processing non-stationary vibration data with the MODWPT (Maximum Overlap Discrete Wavelet Packet Transform). After that, a degradation profile is defined and coupled with crack initiation and crack propagation fatigue models. Lastly, a MB-HMM (Hidden Markov Model) is trained to capture the bearing degradation dynamics. This latter model is used to estimate the current degradation state as well as the RUL. The obtained results show good RUL prediction capabilities. In particular, the fatigue models allowed a reduction of the ML (Machine Learning) model size, improving the algorithms training phase.

Here we investigate the application of convolutional neural networks (CNNs) to predict the survival time of patients with Glioblastoma Multiforme (GBM) brain tumor. Our dataset consists of T1-weighted high-resolution MRI images of just 68 GBM patients. We compare two analytic methods for predicting survival time. The first consists of training a small convolutional neural network (CNN) and the second uses extracted deep features from a pre-trained CNN. Our method is completely automated, except for tumor region segmentation. In addition, we utilize a snapshot ensemble approach to boost test accuracy when dealing with limited availability of medical images for CNN training purposes. Our approach achieves an accuracy of 72.06% using a trained small network and 66.18% using a pre-trained deep CNN. Our results compare favorably with the accuracy of 54.41% using histogram of oriented gradients (HOG) features and a non-neural network classifier.

For health prognostic task, ever-increasing efforts have been focused on machine learning based methods, which are capable of yielding accurate remaining useful life (RUL) estimation for industrial equipment or components without exploring the degradation mechanism. A prerequisite ensuring the success of these methods depends on a wealth of run-to-failure data; however, run-to-failure data may be insufficient in practice. That is, conducting a substantial amount of destructive experiments not only is of high cost but also may cause catastrophic consequences. Out of this consideration, an enhanced RUL framework focusing on data self-generation is put forward for both noncyclic and cyclic degradation patterns for the first time. It is designed to enrich data from a data-driven way, generating realistic-like time-series to enhance current RUL methods. First, high-quality data generation is ensured through the proposed convolutional recurrent generative adversarial network, which adopts a two-channel fusion convolutional recurrent neural network. Next, a hierarchical framework is proposed to combine generated data into current RUL estimation methods. Finally, in this article the efficacy of the proposed method is verified through both noncyclic and cyclic degradation systems. With the enhanced RUL framework, an aero-engine system following noncyclic degradation has been tested using three typical RUL models. State-of-the-art RUL estimation results are achieved by enhancing capsule network with generated time-series. Specifically, estimation errors evaluated by the index score function have been reduced by 21.77$\%$ and 32.67$\%$ for the two employed operating conditions, respectively. Besides, the estimation error is reduced to zero for the lithium-ion battery system, which presents cyclic degradation.

The remaining useful life (RUL) estimation is an important metric that helps in condition-based maintenance. Damage data obtained from the diagnostics techniques are often noisy, and the RUL estimated from the data is less reliable. Estimating the probabilistic RUL by quantifying the uncertainty in the predictive model parameters using the noisy data increases confidence in the predicted values. In this work, Bayesian inference is used to inversely quantify the uncertainty in the model parameters of the predictive damage model in adhesively bonded joints, and quantified parameter uncertainty is represented by a probability distribution function (PDF). It is difficult to obtain a closed-form solution for the RUL from the PDF of model parameters due to the nonlinearity of the damage models. Instead, statistical simulations are used to generate the model parameter samples from the PDF, and the RUL is estimated from the samples. In the statistical simulations, the predictive damage model is evaluated for each parameter sample. Using a physics-based model as the predictive model increases computational time, which makes the statistical techniques to be computationally expensive. It is essential to reduce the computational time to estimate the RUL in a feasible time. In this work, real-time probabilistic RUL estimation is demonstrated in adhesively bonded joints using the Sequential Monte Carlo (SMC) sampling method and cloud-based computations. The SMC sampling method is an alternative to traditional Markov Chain Monte Carlo (MCMC) methods, which enables generating statistical parameter samples in parallel. The parallel computational capabilities of the SMC methods are exploited by running the SMC simulation on multiple cloud calls. This approach is demonstrated by estimating fatigue RUL in the adhesively bonded joint. The accuracy of probabilistic RUL estimated by SMC is validated by comparing it with RUL estimated by the MCMC and the experimental values. The SMC simulation is run on the cloud and the computational speedup of the SMC is demonstrated.

Prognostics or Remaining Useful Life (RUL) Estimation from multi-sensor time series data is useful to enable condition-based maintenance and ensure high operational availability of equipment. We propose a novel deep learning based approach for Prognostics with Uncertainty Quantification that is useful in scenarios where: (i) access to labeled failure data is scarce due to rarity of failures (ii) inherent noise ispresent in the sensor readings. The two scenarios mentioned are unavoidable sources of uncertainty in the RUL estimation process, often resulting in unreliable RUL estimates. To address (i), we formulate RUL estimation as an Ordinal Regression (OR) problem and propose LSTM-OR: deep Long Short Term Memory (LSTM) network-based approach to learn the OR function. We show that LSTM-OR naturally allows for the incorporation of censored operational instances in training along with the failed instances, leading to more robust learning. To address (ii), we propose a simple yet effectiveapproach to quantify predictive uncertainty in the RUL estimation models by training an ensemble of LSTM-OR models. Through empirical evaluation on the publicly available turbofan engine benchmark datasets, we demonstrate that LSTMOR is at par with commonly used deep metric regressionbased approaches for RUL estimation when sufficient failed instances are available for training. Importantly, LSTM-OR outperforms these metric regression-based approaches in the practical scenario where failed training instances are scarce, but sufficient operational (censored) instances are additionally available. Furthermore, our uncertainty quantification approach yields high-quality predictive uncertainty estimateswhile also leading to improved RUL estimates compared to single best LSTM-OR models.

Prognostics, in other words, remaining useful life (RUL) estimation is a core task of prognostics and health management (PHM). Reliable RUL predictions can reduce maintenance costs, improve production efficiency, and avoid unexpected downtime. Lots of models for RUL predictions have been proposed; however, noise and the nonlinear nature of degradation phenomena often leads to poor prognostics results, and the acquired engineered system data are usually subject to a high level of uncertainty. This makes the RUL estimation models less than satisfactory. Accurate RUL estimation and prediction not only rely on an accurate model but also depend on the adjustments of model parameters to track the variation. In this paper, an ensemble model combining the health index synthesis (HIS) approach and improved particle filtering (PF) is introduced. HIS approach was used to obtain the synthesized health index (SHI) for an engineered system with multiple sensors, which indicated the system's degradation model, while the improved PF approach was used to adjust the parameters of the degradation model obtained from the HIS approach and optimized the RUL estimation results. The performance of the prognostics approach introduced in this paper was demonstrated by using turbofan engine degradation data sets, which was supplied by NASA Ames, and results were compared with several usually used methods.

A cutting-edge solution enables accurate prediction of the remaining useful life (RUL) of drill string components before deployment in the field. This proactive approach allows informed decisions about such components to be made prior to actual drilling, significantly reducing the risk of field fatigue-related failures. The solution provides an efficient and cost-effective method for extending the lifespan of critical components, such as BHA connectors, ultimately improving overall operational efficiency. The solution seamlessly integrates drill plan data with results from offset well analysis to estimate the RUL of various drill string components, particularly different BHA connectors. By examining offset wells with similar equipment and geological conditions, it extracts essential anticipated operational parameters. Combining a physics-based model of the relevant drill string components with a data-driven approach and modeling capabilities, the solution ensures reliable decision-making regarding the suitability of specific components for particular drilling operations. Additionally, it facilitates the optimal selection of drill string components for the intended job. This innovative solution has undergone extensive testing, leveraging historical data to validate its reliability and effectiveness. A comprehensive comparison of predicted RUL for monitored components before field runs with actual post-job RUL estimates based on recorded drilling time consistently demonstrates favorable outcomes. This rigorous validation confirms the accuracy and dependability of the approach. By implementing this solution, drilling engineers can optimize drill string component selection, configurations, and maintenance operations, minimizing uncertainties in RUL estimation and reducing the risk of costly failures. Integrating predicted RUL estimates into routine post-job maintenance before starting new drilling runs enables well-informed decisions, preventing unnecessary maintenance and mitigating premature component retirements. The solution is now thoroughly validated and ready for seamless integration into a comprehensive drilling platform, encompassing maintenance systems, well construction planning, and execution. This unified strategy for drill string component management maximizes equipment utilization, enhances drilling performance, and ensures reliability throughout the drilling process. Operators adopting this integrated approach can optimize fleet management, reduce nonproductive time, minimize associated costs, and ultimately maximize revenue. This solution highlights the significant value digital technology brings to traditional mechanical systems in the oil and gas industry by enhancing functionality and creating new service opportunities. The described algorithm and its validation process are novel to the industry. Several instances analyzed demonstrate the methodology's application, providing valuable comparisons to post-job results and underscoring the approach's robustness.

Prognostics and health management (PHM) is one of the research hotspots in reliability, where remaining useful life (RUL) estimation is a typical application scenario. In this article, a feature learning method based on dynamic Bayesian networks (DBNs) is proposed to improve the RUL estimation accuracy of the regression models. The best feature set is obtained with the conditional dependencies represented by the DBN structure. A local modeling method is applied here to reduce the computation for high-order DBN construction. The strength of the connections between variables together with a contribution index of variables in the DBN structure are defined to represent the feature importance of the variables. Feature transfer is carried out with feature importance under different operating conditions for a further improvement. Nonlinear regression models such as support vector regression (SVR) and Gaussian mixture regression (GMR) are built based on the learned features and used to estimate the RUL. The turbofan engine dataset C-MAPSS is used to validate the effectiveness of the proposed method. Compared with other recent RUL estimation models, the proposed method has a faster modeling speed and higher prediction accuracy.

Remaining useful life estimation based on discriminating shapelet extraction

We propose a Bayesian evidence framework to facilitate transfer learning from pre-trained deep convolutional neural networks (CNNs). Our framework is formulated on top of a least squares SVM (LS-SVM) classifier, which is simple and fast in both training and testing, and achieves competitive performance in practice. The regularization parameters in LS-SVM is estimated automatically without grid search and cross-validation by maximizing evidence, which is a useful measure to select the best performing CNN out of multiple candidates for transfer learning, the evidence is optimized efficiently by employing Aitken's delta-squared process, which accelerates convergence of fixed point update. The proposed Bayesian evidence framework also provides a good solution to identify the best ensemble of heterogeneous CNNs through a greedy algorithm. Our Bayesian evidence framework for transfer learning is tested on 12 visual recognition datasets and illustrates the state-of-the-art performance consistently in terms of prediction accuracy and modeling efficiency.

Deep learning based methods have recently been successfully explored in hyperspectral image classification field. However, training a deep learning model still requires a large number of labeled samples, which is usually impractical in hyperspectral images. In this paper, a simple but effective feature extraction method is proposed for hyperspectral image classification. Specifically, a pretrained deep convolutional neural network based on the ImageNet dataset is used to extract spatial features of a hyperspectral image. Recently, it is easy to obtain a pretrained convolutional neural network on the Internet. Note that the pretrained models are trained by using the ImageNet dataset. This means that the proposed method does not need labeled hyperspectral samples to train the deep model. Therefore, the proposed method alleviates the problem of lacking labeled samples and avoids the artificial design of feature extraction rules. Finally, the extracted features are stacked with spectral features as the input of a support vector machine classifier. The proposed method is conducted on three widely used hyperspectral image datasets. The experimental results demonstrate that the proposed method could outperform the conventional feature extraction methods and deep learning based methods.

Deep learning has been used to train neural networks to estimate the Remaining Useful Life (RUL) of a machine given sensor signals from that machine. This has resulted in some accurate state-of-the-art RUL estimation models. Here we present a novel way to construct a generative model to estimate the RUL and quantify its uncertainty. It takes the form of a Linear Gaussian State Space Model (LGSSM) and is trained using the Kalman Filter; hence, uncertainty is quantified using this LGSSM instead of by other popular methods like Monte Carlo Dropout or Deep Ensembles. This means we can train by directly using the marginal log-likelihood loss and don’t require multiple samples to represent the uncertainty. However, this method is limited to using Gaussian distributions to quantify uncertainty. We avoid needing to use non-linear variants of the filter by processing the sensors using a neural network to represent noncausal sensor sequences as a “control variable” in the LGSSM. This noncausal representation is shown to be important for achieving state-of-the-art performance. The model is tested on a turbofan engine and dust filter dataset. The code can be found on GitHub. 1 1 https://github.com/StarMarco/Kalman_DVAE • Deep Neural Networks encode time window of sensor signals. • These are used as control/forcing variables of a Linear Gaussian State Space Model (LGSSM). • The “observables" of the LGSSM are the Remaining Useful Life (RUL) values of the machine. • The Kalman filtering algorithm is used to estimate the marginal log-likelihood of these RUL estimates. • After training the LGSSM can be evaluated over time to estimate the RUL trajectories.