System Reliability Prediction Research Articles

Enhancing water quality prediction for fluctuating missing data scenarios: A dynamic Bayesian network-based processing system to monitor cyanobacteria proliferation

Tackling the impact of missing data in water management is crucial to ensure the reliability of scientific research that informs decision-making processes in public health. The goal of this study is to ascertain the root causes associated with cyanobacteria proliferation under major missing data scenarios. For this purpose, a dynamic missing data management methodology is proposed using Bayesian Machine Learning for accurate surface water quality prediction of a river from Limia basin (Spain). The methodology used entails a sequence of analytical steps, starting with data pre-processing, followed by the selection of a reliable dynamic Bayesian missing value prediction system, leading finally to a supervised analysis of the behavioral patterns exhibited by cyanobacteria. For that, a total of 2,118,844 data points were used, with 205,316 (9.69 %) missing values identified. The machine learning testing showed the iterative structural expectation maximization (SEM) as the best performing algorithm, above the dynamic imputation (DI) and entropy-based dynamic imputation methods (EBDI), enhancing in some cases the accuracy of imputations by approximately 50 % in R2, RMSE, NRMSE, and logarithmic loss values. These findings can impact how data on water quality is being processed and studied, thus, opening the door for more reliable water management strategies that better inform public health decisions.

Flood damage cost prediction using random forest

Floods are one of nature's deadliest catastrophes, causing permanent and catastrophic damage on the socioeconomic system, agriculture and human life. The problems arise when floods could cause a lot of economic damage such as damage to buildings, agriculture and others. Flood damage estimation is a subject of study that has not received much attention. The objective of this research is to explore the Random Forest algorithm in the flood damage cost prediction. The damages specified by the Malaysia’s Department of Irrigation (JPS) are structures such as culverts, MTB bridges, riverbank ruins, concrete main channels, farm roads, hydrological stations, agricultural and water drainage, JPS pump houses and tyres in Terengganu. Terengganu is one of the states in Malaysia which has to endure floods during the monsoon season by the end of the year. The methods employed in this research include data collection, data pre-processing, backend engine coding and user interface design. This project was implemented using the Python programming language. The data were collected from the annual flood report provided by the JPS Negeri Terengganu. The research used the rainfall and streamflow data from the year 2012 to 2022 as attributes to forecast the cost of the JPS structures damages in Terengganu. The prediction results showed that the best model achieved the accuracy of 91.47% with a Mean Squared Logarithmic Error (MSLE) of 0.48 and Coefficient of Determination (R2) of 0.92. In the performance evaluation, the model with 80:20 training and testing data ratio produced the best result in predicting the flood damage cost. The potential enhancements to this research involve extending the scope to encompass all Malaysian states, incorporating diverse flooded structures and adding more input variables for a more improved and more reliable flood prediction system.

Enhanced Flood Detection Through Precise Water Segmentation Using Advanced Deep Learning Models

Floods are natural disasters that can result in significant social, economic, and environmental impacts. Timely and accurate flood detection is crucial for effective disaster management and mitigation. This paper addresses the importance of water segmentation in flood detection and water engineering applications, emphasizing the need for precise delineation of water areas in flood-hit regions. Accurate water segmentation not only aids in assessing the extent of flooding but also plays a vital role in predicting and preventing potential flood events. This study explores the application of advanced deep learning models, namely SegNet, UNet, and FCN32 for automated flood area segmentation. Leveraging a dataset comprising 290 images depicting flood-affected areas, the models are trained to accurately delineate water regions within the images. The experiment results demonstrate the efficacy of these models in effectively segmenting floodwaters. Among the tested models, SegNet emerges as the top performer, achieving an impressive precision rate of 88%. This superior performance underscores the potential of deep learning techniques in enhancing flood detection and response capabilities, paving the way for more efficient and reliable flood prediction systems.

A Design for Availability Approach for Use with PHM


 
 
 Prognostics and Health Management (PHM) methods are incorporated into systems for the purpose of avoiding unanticipated failures that can impact system safety, result in additional life cycle cost, and/or adversely affect the availability of the system. Availability is the probability that a system will be able to function when called upon to do so. A vailability depends on the system’ s reliability (how often it fails) and its maintainability (how efficiently and frequently it is pro-actively maintained, and how quickly it can be repaired and restored to operation when it does fail). Availability is directly impacted by the success of PHM. Increasingly, customers of critical systems are entering into “availability contracts” in which the customer either buys the availability of the system (rather than actually purchasing the system itself) or the amount that the system developer/manufacturer/supplier is paid for the system is a function of the availability achieved by the customer. Predicting availability based on known or predicted system reliability, operational parameters, logistics, etc., is relatively straightforward and can be accomplished using existing methods. However, while determining the availability that results from a set of events is straightforward, determining the events that result in a desired availability is not, and prediction of a system’s attributes to meet an availability requirement can only be accomplished using “brute force” search-based methods that are not general and become quickly impractical for real systems and when uncertainties are introduced. This paper presents a “design for availability” approach that starts with an availability requirement and uses it to predict the required logistics, design and operation parameters. The method is general and can be applied when the inputs to the problem are uncertain (even the availability requirement can be a probability distribution). The method is demonstrated on several examples with and without PHM.
 
 

Epileptic seizure prediction based on EEG using pseudo-three-dimensional CNN.

Epileptic seizures are characterized by their sudden and unpredictable nature, posing significant risks to a patient's daily life. Accurate and reliable seizure prediction systems can provide alerts before a seizure occurs, as well as give the patient and caregivers provider enough time to take appropriate measure. This study presents an effective seizure prediction method based on deep learning that combine with handcrafted features. The handcrafted features were selected by Max-Relevance and Min-Redundancy (mRMR) to obtain the optimal set of features. To extract the epileptic features from the fused multidimensional structure, we designed a P3D-BiConvLstm3D model, which is a combination of pseudo-3D convolutional neural network (P3DCNN) and bidirectional convolutional long short-term memory 3D (BiConvLstm3D). We also converted EEG signals into a multidimensional structure that fused spatial, manual features, and temporal information. The multidimensional structure is then fed into a P3DCNN to extract spatial and manual features and feature-to-feature dependencies, followed by a BiConvLstm3D input to explore temporal dependencies while preserving the spatial features, and finally, a channel attention mechanism is implemented to emphasize the more representative information in the multichannel output. The proposed has an average accuracy of 98.13%, an average sensitivity of 98.03%, an average precision of 98.30% and an average specificity of 98.23% for the CHB-MIT scalp EEG database. A comparison of the proposed model with other baseline methods was done to confirm the better performance of features through time-space nonlinear feature fusion. The results show that the proposed P3DCNN-BiConvLstm3D-Attention3D method for epilepsy prediction by time-space nonlinear feature fusion is effective.

Nomogram prediction model for the risk of intracranial hemorrhagic transformation after intravenous thrombolysis in patients with acute ischemic stroke.

Hemorrhagic transformation (HT) after intravenous thrombolysis (IVT) might worsen the clinical outcomes, and a reliable predictive system is needed to identify the risk of hemorrhagic transformation after IVT. Retrospective collection of patients with acute cerebral infarction treated with intravenous thrombolysis in our hospital from 2018 to 2022. 197 patients were included in the research study. Multivariate logistic regression analysis was used to screen the factors in the predictive nomogram. The performance of nomogram was assessed on the area under the receiver operating characteristic curve (AUC-ROC), calibration plots and decision curve analysis (DCA). A total of 197 patients were recruited, of whom 24 (12.1%) developed HT. In multivariate logistic regression model National Institute of Health Stroke Scale (NIHSS) (OR, 1.362; 95% CI, 1.161-1.652; p = 0.001), N-terminal pro-brain natriuretic peptide (NT-pro BNP) (OR, 1.012; 95% CI, 1.004-1.020; p = 0.003), neutrophil to lymphocyte ratio (NLR) (OR, 3.430; 95% CI, 2.082-6.262; p < 0.001), systolic blood pressure (SBP) (OR, 1.039; 95% CI, 1.009-1.075; p = 0.016) were the independent predictors of HT which were used to generate nomogram. The nomogram showed good discrimination due to AUC-ROC values. Calibration plot showed good calibration. DCA showed that nomogram is clinically useful. Nomogram consisting of NIHSS, NT-pro BNP, NLR, SBP scores predict the risk of HT in AIS patients treated with IVT.

PREDICTING FIRE ALARMS USING MULTI SENSOR DATA: A BINARY CLASSIFICATION APPROACH

Fires pose significant threats to human life, property, and the environment. Early detection of fire incidents is crucial to prevent extensive damage and to ensure the safety of occupants. Traditional fire alarm systems typically rely on a single type of sensor, such as smoke detectors or heat sensors, to detect specific fire indicators. These systems operate based on predefined thresholds and triggers. However, they can be prone to false alarms triggered by non-fire-related events (e.g., cooking fumes or dust) and may not provide early warning signs in certain scenarios. To address these limitations, researchers and engineers have turned to advanced technologies, such as multi-sensor data analysis and machine learning algorithms, to develop more reliable and efficient fire alarm prediction systems. On the other hand, the need for a more robust and accurate fire alarm prediction system stems from the shortcomings of traditional methods. False alarms not only lead to wasted resources but also desensitize occupants, potentially leading them to ignore genuine alarms. Additionally, a delayed response to a fire incident can result in severe consequences, making it essential to develop an intelligent system that can effectively and timely predict fire events. Therefore, this work presents the utilization of multi-sensor data and binary classification to develop a more reliable fire alarm prediction system. The experiments are conducted using a dataset collected from various sensor inputs, including air temperature, humidity, CO2 concentration, molecular hydrogen, ethanol gas, and air pressure etc. Then applied binary classification algorithm to learn patterns from the data and classify fire-related events accurately. The results showed promising improvements in prediction accuracy, reduced false alarm rates, and early detection of fire incidents.

OPTIMASI KINERJA LINEAR REGRESSION, RANDOM FOREST REGRESSION DAN MULTILAYER PERCEPTRON PADA PREDIKSI HASIL PANEN

Rice yield prediction is a significant challenge in the context of climate uncertainty and farmland variation. Erratic weather factors, along with land differences, make this prediction more complex. This research aims to address these issues using a machine learning approach. The method used involves three machine learning models namely Linear regression, Random Forest Regression, and ANN with MultiLayer Perceptron algorithm as well as the evaluation matrix RMSE (Root Mean Squared Error), MAE (Mean Absolute Error) and MAPE (Mean Absolute Percentage Error). This research focuses on testing the accuracy of the three models in the face of uncertain seasonal conditions and variations in agricultural land. The results showed that the MultiLayer Perceptron prediction model gave the best results with an error value of 0.094. The random forest regression method ranks second with an error value of 0.510, followed by Linear regression with an error value of 0.281. The importance of outlier testing in the model development process can be seen from the significant improvement in the performance of the MultiLayer Perceptron model. This research contributes to the development of a more reliable and dependable rice yield prediction system, especially in the midst of uncertain climatic conditions. Machine learning models, particularly MultiLayer Perceptron, can be an effective solution to increase agricultural productivity and reduce risks associated with weather changes and land variations.

Evaluate WRF-based Lightning Potential Index (LPI) lightning parameterization over Sri Lanka during second inter-monsoon in 2018

It is important to develop accurate and reliable lightning prediction system that can be contributed towards the safety of life, both concerning forecasting for the public safety and safety of aviation and electrical power. This study aims to evaluate WRF-based Lightning Potential Index(LPI) lightning parameterization and its applicability for predicting lightning over Sri Lanka during the second inter-monsoon. The WRF-ARW model 3.9.1 was used to produce predictions for three lightning events with various physical parameterization schemes with two nested domains with a resolution of 12km and 4km respectively. The model simulated LPI values were evaluated using the Earth Networks Global Lightning (ENGLN) dataset. Results show corresponding lightning simulations were produced with spatial distribution aligned with ground-based lightning data. Results consistently show a high correlation of the LPI index with an hourly CG flash rate over the three cases. Moreover, the WRF model was able to capture the lightning using LPI in Sri Lanka, suggesting that it can be used operationally to predict lightning potential region.

Reverse Engineering of Algebraic Inequalities for System Reliability Predictions and Enhancing Processes in Engineering

Reverse Engineering of Algebraic Inequalities for System Reliability Predictions and Enhancing Processes in Engineering

Predicting the Planting Time of Bird's Eye Chili Based on Environmental Conditions Using Internet of Things (IoT) and Neural Network Method

In Indonesian cuisine, the red Tabasco pepper holds a significant place as a commonly used ingredient. However, the cultivation of this chili variety is not without its challenges, primarily due to the volatile nature of the chili prices. Farmers often struggle with the critical decision of when to plant Tabasco peppers to optimize their yields and income. Understanding the complexities of this decision-making process in the context of varying environmental conditions is crucial. Thanks to recent advances in Internet of Things (IoT) technology, innovative systems have emerged to address these challenges.This study focuses on the development of an IoT-based solution aimed at helping farmers in precisely determining the optimal planting time for Tabasco pepper. It uses five key criteria—average temperature (°C), average humidity (%), rainfall (mm), length of sunlight (hours) and groundwater usage data (m3) to make data-driven planting decisions. The urgent need for such a system becomes evident when considering the unpredictability of climate patterns and their direct impact on crop outcomes. Using historical data from 2019, obtained from the Jakarta Provincial Government Open Data DKI, and climate data from the Meteorological Agency, Climatology, and Geophysics (BMKG), the authors have successfully developed an IoT-based prototype. This prototype employs a neural network algorithm to analyze the aforementioned criteria. The result is a reliable prediction system that boasts an impressive accuracy rate of 91.26%. By offering this level of precision in determining the ideal planting time for Tabasco pepper, the system extends invaluable support to farmers, helping them optimize their cultivation practices and navigate the uncertainties of the chili market.

A Bayesian nonparametric system reliability model which integrates multiple sources of lifetime information

We present a Bayesian nonparametric system reliability model which scales well and provides a great deal of flexibility in modeling. The Bayesian approach naturally handles the disparate amounts of component and subsystem data that may exist. However, traditional Bayesian reliability models are quite computationally complex, relying on MCMC techniques. Our approach utilizes the conjugate properties of the beta-Stacy process, which is the fundamental building block of our model. These individual models are linked together using a method of moments estimation approach. This model is computationally fast, allows for right-censored data, and is used for estimating and predicting system reliability.

A machine learning-driven support vector regression model for enhanced generation system reliability prediction

PurposeThe purpose of this paper is to achieve high accuracy in forecasting generation reliability by accurately evaluating the reliability of power systems. This study uses the RTS-79 reliability test system to measure the method’s effectiveness, using mean absolute percentage error as the performance metrics. Accurate reliability predictions can inform critical decisions related to system design, expansion and maintenance, making this study relevant to power system planning and management.Design/methodology/approachThis paper proposes a novel approach that uses a radial basis kernel function-based support vector regression method to accurately evaluate the reliability of power systems. The approach selects relevant system features and computes loss of load expectation (LOLE) and expected energy not supplied (EENS) using the analytical unit additional algorithm. The proposed method is evaluated under two scenarios, with changes applied to the load demand side or both the generation system and load profile.FindingsThe proposed method predicts LOLE and EENS with high accuracy, especially in the first scenario. The results demonstrate the method’s effectiveness in forecasting generation reliability. Accurate reliability predictions can inform critical decisions related to system design, expansion and maintenance. Therefore, the findings of this study have significant implications for power system planning and management.Originality/valueWhat sets this approach apart is the extraction of several features from both the generation and load sides of the power system, representing a unique contribution to the field.

Traffic Prediction using Machine Learning

The traffic is one of the major issues faced by people in almost every city. Predicting the traffic has always been very challenging for transportation planning as well as the city manager. The growth of population and usage of vehicles has increased the need for a reliable traffic prediction system. The system should reduce congestion, avoid accidents, and optimize the traffic flow that enhance road safety. The use of machine learning algorithms would be more suitable in handling the traffic and maintaining the flow of vehicles. The paper focuses on the work related to handling the traffic using machine learning principles such as LSTM and Random Forest (RF) algorithm. The advanced online dataset for traffic forecasts is used. The system is tested to have improved accuracy with the help of feature engineering. The basic dataset had the following fields: Vehicle Id, Time, Date, Junction Id are the primary and significant parameters considered to anticipate the traffic at each zone. With this we have added the day and weekend component to enhance our result.

Novel wind-speed prediction system based on dimensionality reduction and nonlinear weighting strategy for point-interval prediction

In the context of today’s energy shortage, wind energy plays a crucial role as one of the most widely used renewable energy sources. However, in order to fully utilize the potential of wind energy, it is necessary to build a precise and reliable wind-speed prediction system. Conventional wind-speed prediction models frequently rely on a solitary model for forecasting, and sometimes these models cannot accurately fit the nonlinear characteristics of the data, resulting in poor prediction performance. Therefore, to address this deficiency, a new wind-speed point-interval prediction system is proposed in this study. The fuzzy information granulation technology is adopted in this system to reduce the dimension of data and solve the problem of redundancy in wind-speed data. In addition, this paper uses the multi-objective dragonfly algorithm and combines sub-models using the multi-nonlinear weight strategy to address the issue of inadequate precision exhibited by an individual prediction model. In order to verify the effectiveness of this wind-speed prediction system, the Penglai wind farm in Shandong Province is used as the dataset to construct the model, and several reference models are selected to conduct point prediction and interval prediction comparison experiments respectively. The experimental results show that the combined prediction system has the best prediction performance, with MAPE values of 14.16%, 15.02% and 15.74%, respectively. Compared with other benchmark models, the average improvement rates for the multi-step prediction were 14.97%, 16.48% and 19.11%, respectively.

Towards explainable motion prediction using heterogeneous graph representations

Motion prediction systems play a crucial role in enabling autonomous vehicles to navigate safely and efficiently in complex traffic scenarios. Graph Neural Network (GNN)-based approaches have emerged as a promising solution for capturing interactions among dynamic agents and static objects. However, they often lack transparency, interpretability and explainability — qualities that are essential for building trust in autonomous driving systems. In this work, we address this challenge by presenting a comprehensive approach to enhance the explainability of graph-based motion prediction systems. We introduce the Explainable Heterogeneous Graph-based Policy (XHGP) model based on an heterogeneous graph representation of the traffic scene and lane-graph traversals. Distinct from other graph-based models, XHGP leverages object-level and type-level attention mechanisms to learn interaction behaviors, providing information about the importance of agents and interactions in the scene. In addition, capitalizing on XHGP’s architecture, we investigate the explanations provided by the GNNExplainer and apply counterfactual reasoning to analyze the sensitivity of the model to modifications of the input data. This includes masking scene elements, altering trajectories, and adding or removing dynamic agents. Our proposal advances towards achieving reliable and explainable motion prediction systems, addressing the concerns of users, developers and regulatory agencies alike. The insights gained from our explainability analysis contribute to a better understanding of the relationships between dynamic and static elements in traffic scenarios, facilitating the interpretation of the results, as well as the correction of possible errors in motion prediction models, and thus contributing to the development of trustworthy motion prediction systems.The code to reproduce this work is publicly available at https://github.com/sancarlim/Explainable-MP/tree/v1.1.

A reliability prediction method considering degradation self-acceleration effect in DC-link electrolytic capacitor

The reliability of DC-link electrolytic capacitors is crucial to ensure the quality of power supply systems. The degradation of capacitor parameters may lead to a higher temperature and thus accelerate degradation as a self-accelerating effect. In this article, an improved reliability prediction method for DC-link electrolytic capacitors is proposed, as existing methods have not adequately accounted for the self-acceleration effect. The degradation under dynamic stress is obtained by cumulative computations and the stress is updated according to the degraded parameters. The degradation models are converted into degradation rate models to overcome the computational challenges associated with small-step iterations that may make traditional methods unaffordable. The proposed method for developing the degradation rate model is widely applicable and achieves satisfactory accuracy. To demonstrate the practicality of the proposed method, a case study of a boost motor drive system is presented. The appropriate iteration step can be determined by comparing the results of the lifetime distributions obtained using different iteration steps. Degradation paths considering self-acceleration effects can be obtained, enabling more precise system quality analysis and reliability prediction.

Machine learning approach to evaluate TdP risk of drugs using cardiac electrophysiological model including inter-individual variability.

Introduction: Predicting ventricular arrhythmia Torsade de Pointes (TdP) caused by drug-induced cardiotoxicity is essential in drug development. Several studies used single biomarkers such as qNet and Repolarization Abnormality (RA) in a single cardiac cell model to evaluate TdP risk. However, a single biomarker may not encompass the full range of factors contributing to TdP risk, leading to divergent TdP risk prediction outcomes, mainly when evaluated using unseen data. We addressed this issue by utilizing multi-in silico features from a population of human ventricular cell models that could capture a representation of the underlying mechanisms contributing to TdP risk to provide a more reliable assessment of drug-induced cardiotoxicity. Method: We generated a virtual population of human ventricular cell models using a modified O'Hara-Rudy model, allowing inter-individual variation. and Hill coefficients from 67 drugs were used as input to simulate drug effects on cardiac cells. Fourteen features (, , , , , , , , , , , , qNet, qInward) could be generated from the simulation and used as input to several machine learning models, including k-nearest neighbor (KNN), Random Forest (RF), XGBoost, and Artificial Neural Networks (ANN). Optimization of the machine learning model was performed using a grid search to select the best parameter of the proposed model. We applied five-fold cross-validation while training the model with 42 drugs and evaluated the model's performance with test data from 25 drugs. Result: The proposed ANN model showed the highest performance in predicting the TdP risk of drugs by providing an accuracy of 0.923 (0.908-0.937), sensitivity of 0.926 (0.909-0.942), specificity of 0.921 (0.906-0.935), and AUC score of 0.964 (0.954-0.975). Discussion and conclusion: According to the performance results, combining the electrophysiological model including inter-individual variation and optimization of machine learning showed good generalization ability when evaluated using the unseen dataset and produced a reliable drug-induced TdP risk prediction system.

Can system reliability be predicted from average component reliabilities?

The paper reveals that a prediction of system reliability on demand based on average reliabilities on demand of components is a fundamentally flawed approach. A physical interpretation of algebraic inequalities demonstrated that assuming average component reliabilities on demand entails an overestimation of the system reliability on demand for systems with components logically arranged in series and series-parallel and underestimation of the reliability on demand for systems with components logically arranged in parallel. The key reason for these discrepancies is the variability of components from the same type. Techniques for countering variability by promoting asymmetric response through inversion have also been introduced. The paper demonstrates that variability during assembly operations can affect negatively the reliability of mechanical systems. Accordingly, techniques for reducing the variability of stresses during assembly operations have been discussed. Finally, the paper provides a discussion related to the reasons for the relatively slow adoption of domain-independent methods for improving reliability despite their numerous advantages.

Prediction of extreme response of an innovative HSR integral bridge subjected to crosswind and high-speed train

The randomness of responses in a train-bridge interaction (TBI) system induced by stochastic track irregularity and crosswind has been addressed. However, the extreme value distribution (EVD) and reliability prediction of TBI system are little known. Due to the complexity of TBI system, the prediction approach is critical to the issues of extreme value and reliability evaluation, involving the computational accuracy and efficiency. Thus, the Monto Carlo-based global maximum method, probability density evolution-based virtual time process method (VPM), and average conditional exceeding rate method (ACERM) are considered in this study. The optimal approach for the EVD prediction of bridge buffeting responses and the TBI system with or without the consideration of crosswinds is identified. An innovative 3 × 70 m integral bridge is chosen as the study case, which is an emerging type of bridge structures used in the Chinese high-speed railway lines. Its extreme responses under the single and dual stochastic excitations of track irregularity and wind turbulence are fully investigated in this study. The study results show that the VPM and ACERM are efficient and accurate to evaluate the extreme responses of a TBI system. The ACERM is recommended in assessing the train running safety due to its high computational efficiency and reasonable accuracy. When the crosswind is not a consideration, the high-speed train running safety can be ensured with high probability if the train travels on the integral bridge below the design speed of 350 km/h. However, the failure probability of wheel-rail lateral contact forces reaches 51% if predicted under the mean wind speed of 30 m/s and train speed of 120 km/h. Finally, the advantages of the integral bridge over conventional bridges (e.g., continuous girder bridge) are discussed.