Minimum Maintenance Cost Research Articles

A unified multi‐class anomaly detection model based on reverse distillation

AbstractThe multi‐class single‐model detection paradigm is a prevalent design for industrial anomaly detection, exhibiting suitability for varying industrial classes and dynamic, flexible production scenarios. This approach not only enhances model adaptability but also minimizes maintenance costs. However, the current popular methods are susceptible to the ‘copying shortcut’ phenomenon, which constrains their performance on benchmark datasets. To overcome this limitation, this article proposes a multi‐class anomaly detection model: UniRD, based on the reverse distillation method. This model creates an image corruptor that expands the dataset and generates ‘normal‐corrupt’ image pairs. During the training process, their correspondence is used to optimize the reverse distillation. This process greatly exploits and releases the potential of the student decoder. Furthermore, a teacher feature adaptation module is devised to enhance the compatibility between the pre‐trained model and the anomaly detection task. This has the effect of reducing the discrepancy between teacher and student features while ensuring the consistency of normal sample features. The comprehensive evaluation results in two mainstream datasets, MVTec and VisA, and demonstrates that the proposed method exhibits improvement in all indicators compared to benchmark methods. The proposed method attains the state‐of‐the‐art, substantiating its effectiveness and advancement.

Data augmentation in predictive maintenance applicable to hydrogen combustion engines: a review

Machine-learning-based predictive maintenance models, i.e. models that predict breakdowns of machines based on condition information, have a high potential to minimize maintenance costs in industrial applications by determining the best possible time to perform maintenance. Modern machines have sensors that can collect all relevant data of the operating condition and for legacy machines which are still widely used in the industry, retrofit sensors are readily, easily and inexpensively available. With the help of this data it is possible to train such a predictive maintenance model. The main problem is that most data is obtained from normal operating conditions, whereas only limited data are from failures. This leads to highly unbalanced data sets, which makes it very difficult, if not impossible, to train a predictive maintenance model that can detect faults reliably and timely. Another issue is the lack of available real data due to privacy concerns. To address these problems, a suitable data generation strategy is needed. In this work, a literature review is conducted to identify a solution approach for a suitable data augmentation strategy that can be applied to our specific use case of hydrogen combustion engines in the automotive field. This literature review shows that, among the different state-of-the-art proposals, the most promising for the generation of reliable synthetic data are the ones based on generative models. The analysis of the different metrics used in the state of the art allows to identify the most suitable ones to evaluate the quality of generated signals. Finally, an open problem in research in this area is identified and it is the need to validate the plausibility of the data generated. The generation of results in this area will contribute decisively to the development of predictive maintenance models.

An Optimization Model for Production Scheduling in Parallel Machine Systems

The efficiency and quality of the manufacturing industry are greatly influenced by production scheduling, which makes it a crucial aspect. A well-designed production scheduling scheme can significantly enhance manufacturing efficiency and reduce enterprise costs. This paper presents a tailored optimization model designed to address a more complex production scheduling problem that incorporates parallel machines and preventive maintenance. The proposed solutions aim to achieve a balance between job sequence and machine reliability, considering the minimum maintenance cost rate for determining maintenance cycles of deteriorating machines in real manufacturing scenarios. Furthermore, the objective of minimizing the maximum completion time guides machine assignment and job sequence based on maintenance constraints. The innovation lies in the introduction of a greedy algorithm that utilizes a water injection model to address this NP-hard integrated problem. A pre-distribution model is constructed using the water injection model, and its solution is utilized as input for constructing the production scheduling model, which aids in determining machine assignment and job sequence. This algorithm demonstrates remarkable effectiveness and efficiency, enabling the achievement of an optimal solution. A numerical example is presented to illustrate the computational process, accompanied by an extensive discussion of the results showcasing improved performance. Furthermore, the optimization model developed in this paper can be adapted to tackle the production scheduling problem with modifications tailored for parallel machines.

Fatigue life prognosis of composite structures using a transferable deep reinforcement learning-based approach

Accurately predicting the remaining useful life (RUL) of Carbon Fiber Reinforced Polymer (CFRP) structures under fatigue loading is crucial for enhancing safety and minimizing maintenance costs, especially in industries like aerospace and automotive. However, the complex physical properties of CFRP, combined with the scarcity of real-world damage-condition data, make this task extremely challenging. To address these issues, we propose a novel deep reinforcement learning (DRL)-based prognostic method. Our approach integrates Denoising Autoencoder (DAE) and Transformer architectures to construct a powerful DRL Policy Network, capable of extracting high-quality features from X-ray records to capture the subtle progression of damage in CFRP structures. Additionally, we employ advanced data augmentation techniques to overcome the limitations of small datasets and introduce transfer learning to extend the model’s generalization capabilities across different CFRP structures. By pre-training on diverse CFRP datasets, our model achieves highly accurate RUL predictions for new designs, even with minimal labeled data from the target structure. Experimental results demonstrate that our method significantly outperforms current state-of-the-art (SOTA) techniques, offering a scalable, efficient, and practical solution for the real-world monitoring and prognostics of CFRP structures, with broad potential for industrial applications.

A Maintenance Policy for Central Tower Receiver Subjected to Creep-Fatigue Damage

This study focuses on the central receiver of solar tower plants, which is subjected to extreme heat fluxes, high temperatures, and thermal gradients leading to degradation mechanisms such as creep, fatigue, and corrosion. Although studies in literature have developed thermal models for receiver tubes to understand the damage process and estimate their lifetime, they have not addressed the uncertainty associated with receiver damage, which arises from random operating conditions and errors in creep damage predictive models. Considering the random and varying nature of damage, this paper suggests effective maintenance strategies that optimize the receiver's lifetime and minimize maintenance costs. The strategies utilize a simulation-optimization approach and incorporate uncertain operating conditions and creep-fatigue models. The illustration case study shows that the proposed maintenance strategy can reduce the maintenance cost up to 37% versus the interval replacement of all panels.

Intelligent Fault Diagnosis of Hydraulic System Based on Multiscale One-Dimensional Convolutional Neural Networks with Multiattention Mechanism.

Hydraulic systems are critical components of mechanical equipment, and effective fault diagnosis is essential for minimizing maintenance costs and enhancing system reliability. In practical applications, data from hydraulic systems are collected with varying sampling frequencies, coupled with complex interdependencies within the data, which poses challenges for existing fault diagnosis algorithms. To solve the above problems, this paper proposes an intelligent fault diagnosis of a hydraulic system based on a multiscale one-dimensional convolution neural network with a multiattention mechanism (MA-MS1DCNN). The proposed method first extracts features from multirate data samples using a parallel 1DCNN with different receptive fields. Next, a Hybrid Attention Module (HAM) is proposed, consisting of two submodules: the Correlation Attention Module (CAM) and the Importance Attention Module (IAM), which aim to meticulously and comprehensively model the complex relationships between channel features. Subsequently, to effectively utilize the feature information of different frequencies, the HAM is integrated into the 1DCNN to form the MA-MS1DCNN. Finally, the proposed method is evaluated and experimentally compared using the UCI hydraulic system dataset. The results demonstrate that, compared to existing methods such as Shapelet, MCIFM, and CNNs, the proposed method shows superior diagnostic performance.

Optimization of Maintenance Strategies in an Oil and Gas Production Facility to Minimize Maintenance Cost

Abstract: In this study, maintenance strategies in an oil and gas production facility were optimized to minimize maintenance costs. The centrifugal pump system in Port Harcourt Refinery was selected for this study because it significantly impacts on the productivity of the refining plant. The pump system’s failure mode effects and criticality analysis (FMECA) were examined, and an optimized maintenance task was developed for the system. The exponential reliability method was employed to analyse the pump’s failure data to determine the pump systems' failure rate and reliability while reliability centred maintenance (RCM) and linear programming (LP) model were employed to optimize the pump’s maintenance strategies and the pump’s maintenance labour force respectively. Broad-based results of the optimized maintenance strategies consisted of on-condition-directed (CD) maintenance, preventive maintenance (PreM), and Proactive maintenance (ProM) strategies to be carried out monthly. The optimized maintenance labour force result revealed that approximately three engineers and two technicians should be employed for the pump’s maintenance task with a minimum of N2, 585, 700 as cost of salaries for the labour force monthly as against the current maintenance labour force requiring fourteen engineers and thirteen technicians monthly. The results showed that the labour cost decreased from N180,000,000 per year to N31,028,400 per year (approximately 82.76% reduction) using the proposed optimized maintenance task compared to the current maintenance plan. Conclusion and recommendation were made that the maintenance optimization technique presented in this work should be adopted by the oil and gas processing and production companies in order to minimize maintenance cost.

Multitask-Transfer-Learning Method for Random-Force Frequency Identification Considering Multisource Uncertainties

Accurate reconstruction of unknown external forces from measurable responses is critical for ensuring structural safety and minimizing maintenance costs of aircraft structures. This paper presents a novel multitask-transfer-learning method for random-force frequency identification that accounts for modeling and measurement uncertainties. A data-driven convolutional neural network (CNN) model is utilized to capture the relationship between the power spectral densities of external forces and measured responses, addressing the inherent ill-posedness of traditional model-driven force identification methods. To shorten the frequency-dependent training time in the full frequency domain, a transfer-learning strategy is implemented, fine-tuning hyperparameters from a CNN model trained at one source frequency to another target frequency. Furthermore, an iterative dimensionwise collocation method based on nonprobabilistic interval modeling is introduced to quantify the uncertain boundaries of external loads caused by multisource uncertainties. By incorporating a multitask-learning framework, the process of establishing CNN models for collocated samples is accelerated, reducing the computational effort for uncertainty quantification. The proposed method is validated through both numerical and experimental examples, demonstrating its accuracy, robustness, and computational efficiency for force identification in the full frequency domain, even under conditions of insufficient measurements, measurement noises, and material dispersions.

Spare Parts Management Strategy of High-Speed Railway Running Department Based on Performance Prediction

Spare parts management is a critical aspect of high-speed train health management, playing a vital role in maximizing in-service time and minimizing maintenance costs. However, traditional spare parts management methods, which rely solely on historical experience and suggest spare parts quantities or ratios in equipment manuals, often lack practicality and fail to meet real-world demands. To address these limitations, this paper proposes a performance prediction-based spare parts management strategy for high-speed trains. The strategy comprises three main components. First, a performance degradation model is developed using performance evaluation results to define a performance degradation envelope. Next, the required quantity or ratio of spare parts for multiple devices in different performance states is determined using the expected performance score method. Finally, the timing of spare parts orders is scientifically optimized by accounting for production and transportation lead times. To demonstrate the effectiveness of the proposed strategy, we conducted experiments using the spare parts management of a specific high-speed train running gear as a case study and compared it with existing spare parts management methods.

Corrosion measurement of thermally sprayed carbide coatings on stainless steel pipes

The petrochemical industries face a formidable challenge from corrosion, which leads to high maintenance costs and lost production. Traditional corrosion monitoring methods frequently fail to provide accurate and reliable results. As a result, these components must be replaced or repaired promptly to avoid increasing costs and limiting component life, which will reduce operation efficiency. Novel techniques with demands for high surface qualities in the future will be applied to develop alloys with better performance and more diverse applications. This study examines the corrosion resistance of cermet coatings, such as tungsten carbide (WC) and Chromium carbide (Cr3C2), in acidic solution by electrochemical impedance spectroscopy and electrochemical potentiodynamic polarisation. Additionally, utilizing optical microscopy techniques, the microstructural characteristics of coatings were assessed. According to the results of the Tafel polarisation test, Cr3C2 coating exhibited a lower rate of corrosion than other coatings. The Cr3C2 coated, WC coated, and uncoated samples reported corrosion rates of 0.924, 1.122, and 2.599 mmpy respectively. The corrosion resistance of Cr3C2 and WC coatings was increased by 64% and 53%, respectively, in comparison to uncoated specimens. This suggests that the coating maximizes economy, minimizes maintenance costs, extends the structure's lifespan, and reduces down the rate of corrosion.

A novel dynamic predictive maintenance framework for gearboxes utilizing nonlinear Wiener process

Abstract In the context of advancing industrial automation, gearboxes, as pivotal components in power transmission systems, have a direct bearing on the operational efficiency and safety of the entire machinery. This study introduces a novel dynamic predictive maintenance framework for gearboxes using a nonlinear Wiener process. Comprehensive experiments validate the framework, demonstrating significant reductions in maintenance costs and improvements in reliability. First, a full-life degradation experiment was executed on the gearbox, leveraging the root mean square (RMS) value of the vibration signal as an indicator of system degradation. Subsequently, the signals from four vibration sensors were synthesized and normalized through Kernel Principal Component Analysis (KPCA), thereby enabling a more nuanced representation of the gearbox's degradation profile. The degradation trajectory was then modeled using a nonlinear Wiener process framework. The Wiener process's parameters and state variables were iteratively refined utilizing an online filtering algorithm grounded in Bayesian inference. This facilitated the derivation of the probability density function for the remaining useful life (RUL), thereby enabling a robust prediction of the gearbox's RUL. Finally, to minimize maintenance costs per unit of time, an optimization model for dynamic maintenance decision-making was formulated. The optimal maintenance timing was ascertained by solving this model. The empirical findings of this investigation demonstrate the efficacy of the proposed approach in executing dynamic predictive maintenance for gearboxes. This research endeavors to furnish novel theoretical underpinnings and pragmatic directives for the field of predictive maintenance in the context of gearboxes.

A Deep Learning-Based Approach for Identifying Defects in Solar Panels

In the solar energy sector, the task of monitoring and maintaining large photovoltaic (PV) system portfolios is essential for ensuring optimal performance and reliability. Prominent solar energy companies face challenges with their current fault detection methods, which are inefficient and resource- intensive. This paper addresses the critical need for improved fault detection in solar PV systems to maximize uptime and minimize maintenance costs. We employed advanced data preprocessing and augmentation techniques using Roboflow and developed a YOLOv8 segmentation model in Google Colab with GPU. This model was then deployed using Streamlit, providing a robust solution for identifying faulty solar modules. The proposed approach significantly enhances fault detection accuracy, achieving a minimum accuracy rate of 85%, thus ensuring reliable operation of the PV systems. Additionally, the implementation of this model contributes to a 15% reduction in system downtime and a 10% reduction in maintenance costs. By leveraging advanced machine learning techniques, our solution transforms the maintenance process, making it more efficient and cost-effective. Consequently, this work not only improves the reliability and performance of solar PV systems but also supports the broader goal of sustainable energy through more efficient resource usage.

Prediction of air compressor faults with feature fusion and machine learning

Air compressors are critical for many industries, but early detection of faults is crucial for keeping them running smoothly and minimizing maintenance costs. This contribution investigates the use of predictive machine learning models and feature fusion to diagnose faults in single-acting, single-stage reciprocating air compressors. Vibration signals acquired under healthy and different faulty conditions (inlet valve fluttering, outlet valve fluttering, inlet-outlet valve fluttering, and check valve fault) serve as the study’s input data. Diverse features including statistical attributes, histogram data, and auto-regressive moving average (ARMA) coefficients are extracted from the vibration signals. To identify the most relevant features, the J48 decision tree algorithm is employed. Five lazy classifiers viz. k-nearest neighbor (kNN), K-star, local kNN, locally weighted learning (LWL), and random subspace ensemble K-nearest neighbors (RseslibKnn) are then used for fault classification, each applied to the individual feature sets. The classifiers achieve commendable accuracy, ranging from 85.33% (K-star and local kNN) to 96.00% (RseslibKnn) for individual features. However, the true innovation lies in feature fusion. Combining the three feature types, statistical, histogram, and ARMA, significantly improves accuracy. When local kNN is used with fused features, the model achieves a remarkable 100% classification accuracy, demonstrating the effectiveness of this approach for reliable fault diagnosis in air compressors.

Anomaly Detection for Power Quality Analysis Using Smart Metering Systems.

The problem of Power Quality analysis is becoming crucial to ensuring the proper functioning of complex systems and big plants. In this regard, it is essential to rapidly detect anomalies in voltage and current signals to ensure a prompt response and maximize the system's availability with the minimum maintenance cost. In this paper, anomaly detection algorithms based on machine learning, such as One Class Support Vector Machine (OCSVM), Isolation Forest (IF), and Angle-Based Outlier Detection (ABOD), are used as a first tool for rapid and effective clustering of the measured voltage and current signals directly on-line on the sensing unit. If the proposed anomaly detection algorithm detects an anomaly, further investigations using suitable classification algorithms are required. The main advantage of the proposed solution is the ability to rapidly and efficiently detect different types of anomalies with low computational complexity, allowing the implementation of the algorithm directly on the sensor node used for signal acquisition. A suitable experimental platform has been established to evaluate the advantages of the proposed method. All the different models were tested using a consistent set of hyperparameters and an output dataset generated from the principal component analysis technique. The best results achieved included models reaching 100% recall and a 92% F1 score.

Resilience-oriented adaptive predictive maintenance optimization for continuous process manufacturing systems considering mission profile variation

Continuous process manufacturing systems (CPMSs) are typical phased mission systems that require high standards of operational stability, reliability, and safety. With the variation of production mission profiles, CPMSs are required to run in diverse modes or conditions in different phases; therefore, to meet these standards, maintenance decisions applied to CPMSs should be adapted to such variations. Considering that the concept of “resilience” provides a systematic solution to evaluate system adaptability via “disruption absorption” and “recoverability,” this paper proposes a CPMS resilience evaluation model and utilizes it as guidance for the optimization of CPMS predictive maintenance (PdM). The proposed method consists of the following steps: (1) applying a customized Seasonal Trend Decomposition model to predict the future trend of production mission profile variations, (2) assessing the production mission accomplishment capability of CPMS based on a Gamma process model of equipment performance degradation, (3) using disruption response ratio to evaluate CPMS resilience based on mission accomplishment capability, and (4) proposing a Simulated Annealing Q-Learning algorithm for adaptive PdM optimization, which keeps resilience above a threshold level while minimizing maintenance costs. The applicability and effectiveness of the proposed method are validated by an industrial case study of a nuclear fuel rod shielding component CPMS.

Improving Software Defect Prediction With a Combination of Feature Selection Based On Ant Colony Optimization and Ensemble Technique

Software defect prediction plays a vital role in enhancing software quality and minimizing maintenance costs. This study aims to improve software defect prediction by employing a combination of Ant Colony Optimization (ACO) for feature selection and ensemble techniques, particularly Gradient Boosting. The research utilizes three NASA MDP datasets: MC1, KC1, and PC2, to evaluate the performance of four machine learning algorithms: Random Forest, Support Vector Machine (SVM), Decision Tree, and Naïve Bayes. Data preprocessing involved handling class imbalances using the SMOTE technique and transforming categorical data into numerical representations. The results indicate that the integration of ACO and Gradient Boosting significantly enhances the accuracy of all four algorithms. Notably, the Random Forest algorithm achieved the highest accuracy of 99% on the MC1 dataset. The findings suggest that combining ACO-based feature selection with ensemble techniques can effectively boost the performance of software defect prediction models, offering a robust approach for early detection of potential softwaredefects and contributing to improved software reliability and efficiency.

Multi-objective preventive maintenance strategy and optimization considering unavailability and cost: A case study on VOBC

The Vehicle On-Board Controller (VOBC) is a core component of urban rail transit trains, with high maintenance costs and serious consequences in case of failure. Ensuring long-term VOBC operation while minimizing maintenance costs is of paramount importance. This paper proposes a multi-objective preventive maintenance (PM) strategy and optimization considering system unavailability and maintenance costs. Firstly, we provide a statement of the problem and analyze the impact of different maintenance actions on VOBC. Next, the decreasing process parameter is introduced to adjust the PM interval dynamically. The mathematical models of system unavailability and maintenance costs are derived to construct a multi-objective PM strategy. Then, we develop a hybrid heuristic algorithm including multi-objective particle swarm optimization and genetic algorithm (MOPSO-GA), which is used to obtain the optimal Pareto frontier for maintenance strategies. In the study, we select maintenance strategies by introducing an aversion factor to fulfill the preferences of different decision-makers. A comprehensive sensitivity analysis is performed to illustrate the impact of several critical parameters on the multi-objective PM strategy. The case study on Beijing subway’s VOBC is also executed. The outcomes indicate that the proposed method can dynamically adjust to match the preferences of diverse decision-makers and the MOPSO-GA algorithm has excellent robustness on most parameters. This approach allows for creating field maintenance strategies catering to the varied maintenance requirements of urban rail transit VOBC.

A Transformer Maintenance Interval Optimization Method Considering Imperfect Maintenance and Dynamic Maintenance Costs

As one of the most critical components of the power grid system, transformer maintenance strategy planning significantly influences the safe, economical, and sustainable operation of the power system. Periodic imperfect maintenance strategies have become a research focus in preventive maintenance strategies for large power equipment due to their ease of implementation and better alignment with engineering realities. However, power transformers are characterized by long lifespans, high reliability, and limited defect samples. Existing maintenance methods have not accounted for the dynamic changes in maintenance costs over a transformer’s operational lifetime. Therefore, we propose a maintenance interval optimization method that considers imperfect maintenance and dynamic maintenance costs. Utilizing defect and maintenance cost data from 400 220 KV oil-immersed transformers in northern China, we employed Bayesian estimation for the first time to address the distribution fitting of defect data under small sample conditions. Subsequently, we introduced imperfect maintenance improvement factors to influence the number of defects occurring in each maintenance cycle, resulting in more realistic maintenance cost estimations. Finally, we established an optimization model for transformer maintenance cycles, aiming to minimize maintenance costs throughout the transformer’s entire lifespan while maintaining reliability constraints. Taking a transformer’s strong oil circulation cooling system as an example, our method demonstrates that while meeting the reliability threshold recognized by the power grid company, the system’s maintenance cost can be reduced by 41.443% over the transformer’s entire life cycle. Through parameter analysis of the optimization model, we conclude that as the maintenance cycle increases, the factors dominating maintenance costs shift from corrective maintenance to preventive maintenance.

A hybrid reliability assessment method based on health index construction and reliability modeling for rolling bearing

AbstractThe assessment of rolling bearing reliability is vital for ensuring mechanical operational safety and minimizing maintenance costs. Due to the difficulty in obtaining data on the performance degradation and failure time of rolling bearings, traditional methods for reliability assessment are challenged. This paper introduces a novel hybrid method for the reliability assessment of rolling bearings, combining the convolutional neural network (CNN)‐convolutional block attention module (CBAM)‐ bidirectional long short‐term memory (BiLSTM) network with the Wiener process. The approach comprises three distinct stages: Initially, it involves acquiring two‐dimensional time‐frequency representations of bearings at various operational phases using Continuous Wavelet Transform. Subsequently, the CNN‐CBAM‐BiLSTM network is employed to establish health index (HI) for the bearings and to facilitate the extraction of deep features, serving as input for the Wiener process. The final stage applies the Wiener process to evaluate the bearings’ reliability, characterizing the HI and quantifying uncertainties. The experiment is performed on bearing degradation data and the results indicate the effectiveness and superiority of the proposed hybrid method.

Efficient opportunistic maintenance strategies via pruning in parallel–series systems with economic dependence

Opportunistic maintenance stands out as an effective strategy for minimizing maintenance costs and enhancing system efficiency. Despite its importance, current opportunistic maintenance models are often tailored to specific system types, particularly those with series structures, or they offer heuristic approaches for multi-component redundant systems. This paper shifts focus to the most common of such systems, the parallel–series system. We approach the optimization of opportunistic maintenance in a parallel–series system through a Markov Decision Process framework, utilizing phase-type approximations of individual component hazard functions. Addressing the challenge of the curse of dimensionality in the action space, this study conducts a thorough structural analysis of the optimal opportunistic maintenance policy, deriving a pruning procedure, which can reduce the extensive combinatorial action space into a more manageable linear complexity without sacrificing the strategy’s effectiveness in parallel–series systems. Building on this more manageable action space, we introduce a more efficient algorithm leveraging the proposed pruning procedure for determining the optimal opportunistic maintenance policy. The effectiveness and practical applicability of this approach are rigorously demonstrated through comprehensive numerical experiments.