Optimal Maintenance Decisions Research Articles

Dynamic scheduling of hybrid flow shop problem with uncertain process time and flexible maintenance using NeuroEvolution of Augmenting Topologies

AbstractA hybrid flow shop is pivotal in modern manufacturing systems, where various emergencies and disturbances occur within the smart manufacturing context. Efficiently solving the dynamic hybrid flow shop scheduling problem (HFSP), characterised by dynamic release times, uncertain job processing times, and flexible machine maintenance has become a significant research focus. A NeuroEvolution of Augmenting Topologies (NEAT) algorithm is proposed to minimise the maximum completion time. To improve the NEAT algorithm's efficiency and effectiveness, several features were integrated: a multi‐agent system with autonomous interaction and centralised training to develop the parallel machine scheduling policy, a maintenance‐related scheduling action for optimal maintenance decision learning, and a proactive scheduling action to avoid waiting for jobs at decision moments, thereby exploring a broader solution space. The performance of the trained NEAT model was experimentally compared with the Deep Q‐Network (DQN) and five classical priority dispatching rules (PDRs) across various problem scales. The results show that the NEAT algorithm achieves better solutions and responds more quickly to dynamic changes than DQN and PDRs. Furthermore, generalisation test results demonstrate NEAT's rapid problem‐solving ability on test instances different from the training set.

Optimal Maintenance Decision Method for a Sensor Network Based on Belief Rule Base considering Attribute Correlation

Optimal maintenance decision for a sensor network aims to intelligently determine the optimal repair time. The accuracy of the optimal maintenance decision method directly affects the reliability and safety of the sensor network. This paper develops a new optimal maintenance decision method based on belief rule base considering attribute correlation (BRB-c), which is designed to address three challenges: the lack of observation data, complex system mechanisms, and characteristic correlation. This method consists of two sections: the health state assessment model and the health state prediction model. Firstly, the former is accomplished through a BRB-c-based health assessment model that considers characteristic correlation. Subsequently, based on the current health state, a Wiener process is used to predict the health state of the sensor network. After predicting the health state, experts are then required to establish the minimum threshold, which in turn determines the optimal maintenance time. To demonstrate the proposed method is effective, a case study for the wireless sensor network (WSN) of oil storage tank was conducted. The experimental data were collected from an actual storage tank sensor network in Hainan Province, China. The experimental results validate the accuracy of the developed optimal maintenance decision model, confirming its capability to efficiently predict the optimal maintenance time.

Multiple-failure mode division and condition-based maintenance decision making for systems with multi-indicator performance degradation

In prognostics and health management research, it is typically assumed that the health status of a system can be reflected by a single performance indicator. However, modern industrial systems are complex and often require multiple indicators to represent different aspects of system health. Furthermore, these indicators may be associated with multiple fault types, leading to multiple failure modes. This study focuses on systems with multi-indicator performance degradation. We construct a division model of multiple failure modes, and investigate the optimal condition-based maintenance decision under multiple failure modes. By constructing division models for systems with two-indicator performance degradation and systems with three-indicator performance degradation, a common multiple-failure mode division model for systems with multi-indicator performance degradation is to distinguish different fault types that may occur in such systems. Taking the systems with two-indicator performance degradation as an example, a condition-based maintenance strategy is developed that considers the differences in maintenance types and effectiveness corresponding to different faults types. A joint probability recursive model under the influence of strategies is derived, and a cost-rate model in a finite time horizon is established to determine the optimal inspection cycle and maintenance threshold for each indicator. Taking the steel rolling system as an application case, the correctness and effectiveness of the proposed strategy and model are verified. The results indicate that the division model can represent the complex relationships among multiple indicators and fault types and that the strategy considering multiple maintenance types and effects can reduce the cost rate of the system.

Bi-level corrected residual life-based maintenance for deteriorating systems under competing risks

The accurate prediction of residual life in deteriorating systems is a challenging task due to uncertainties and idealized assumptions during data acquisition and model construction. To address this challenge, the paper proposes a bi-level maintenance model that incorporates a corrected residual life (CRL) approach for a deteriorating system under competing risks. Specifically, the degradation failure is modeled using a Gamma process, while the competing sudden failure is described by a proportional hazards (PH) model. To reduce inconsistencies in residual life prediction, a CRL model is presented and continuously updated with a dynamic corrective factor at each monitoring epoch. Based on the updated CRL, a bi-level maintenance model is developed, which considers both system availability and maintenance cost objectives, enabling dynamic monitoring of the system's degradation state and residual life. The optimal decision variables in the maintenance model are determined through a multi-objective optimization algorithm formulated within a semi-Markov decision process (SMDP) framework. The unique aspect of this work lies in the consideration of prediction errors in residual life and the incorporation of multi-attribute optimization for maintenance design. The proposed method is validated through a case study on light-emitting diodes, confirming its effectiveness in improving residual life prediction and optimizing maintenance decisions.

Calendar-time-based and age-based maintenance policies with different repair assumptions

Traditional maintenance models usually assume that the time duration required for conducting corrective repairs is negligible for mathematical tractability. In this paper, we propose two new preventive maintenance policies that take non-negligible repair time into consideration: One is a calendar-time-based maintenance policy in which a system should be preventively replaced at either of two predetermined calendar time instants, depending on the system state; and the other is an age-based maintenance policy in which a system should be preventively replaced at a predetermined age. Under both policies, if any failure occurs before replacement, then the system will be repaired. In particular, four repair assumptions—perfect, imperfect, minimal, and worse-than-minimal—that correspond to different repair effectiveness, are considered. We derive the long-run average cost rates for the two maintenance policies under different repair assumptions, respectively, and investigate the existence and uniqueness of the solutions of optimal maintenance decisions. Numerical studies are then conducted to compare the two maintenance policies, providing guidelines on which policy should be adopted by system operators in various scenarios.

Multi-Objective Optimization for Sustainable Pavement Maintenance Decision Making by Integrating Pavement Image Segmentation and TOPSIS Methods

To provide a low-carbon economy maintenance strategy is the most challenging problem faced by pavement management authorities under the restricted budget and significant environmental repercussions. The development of a multi-objective optimization model for pavement maintenance decision making is essential to formulate pavements. Nevertheless, the existing automatic detection can only recognize and classify pavement distress. However, few studies are able to accurately determine the precise dimensions of specific distresses such as cracks and potholes, especially combined with the actual size of the image. This limitation hinders the ability to provide specific maintenance recommendations and make optimal maintenance decisions. Therefore, this paper develops a comprehensive and effective multi-objective decision-making framework for pavement maintenance. This framework consists of four distinct components: (1) recognizing the dimensions of pavement distresses based on the pavement image segmentation technique; (2) compiling a list of viable pavement maintenance strategies; (3) assessing the costs and carbon emissions of these strategies; and (4) optimizing decisions on pavement maintenance. We used the U-Net algorithm to accurately recognize the dimensions of pavement distresses, while an improved entropy-weighted TOPSIS model was proposed to determine the optimal pavement maintenance strategy with the lowest cost and carbon emissions. The results indicated that the pavement distress dimension recognition model achieved a high accuracy of 96.88%, and the TOPSIS model identified the optimal maintenance strategy with a score of 99.16. This maintenance strategy achieved a substantial reduction of 30.80% in carbon emissions and a cost reduction of 20.81% compared to the highest values among all maintenance strategies. This study not only provides a scientifically objective method for making pavement maintenance decisions but also offers specific, quantifiable maintenance programs, marking a stride towards more environmentally friendly and cost-effective road maintenance. It also contributes to the sustainability of pavement maintenance.

Capacity-based daily maintenance optimization of urban bus with multi-objective failure priority ranking

There is a specific correlation between bus operational failures and work intensity. The hysteresis effect identified the correlation between the number of repairs and the bus capacity of urban bus fleets. Therefore, administrators can determine whether to perform maintenance on a specific bus based on the severity of detected potential failures. At this stage, administrators should make a maintenance decision after considering maintenance costs, operational losses, and bus capacity. As a result, we propose a model for optimizing maintenance decisions with multi-objective failure priority ranking in this article. We employ the TOPSIS to prioritize the maintenance of failed buses based on indicators such as transport capacity loss, downtime, and maintenance cost. Finally, the study verifies the model's validity by utilizing real data from urban bus operation and maintenance data in China. The outcomes of the case study indicate that implementing the model can reduce operating expenses by 18.8143% when considering a corrective maintenance cost of 300. The model adoption provides a significant reduction in operating costs.

Component-wise Markov decision process for solving condition-based maintenance of large multi-component systems with economic dependence

Condition-Based Maintenance (CBM) of multi-component systems is a prevalent engineering problem due to its effectiveness in reducing the operational and maintenance costs of a system. However, developing the exact optimal maintenance decisions for a large multi-component system is computationally challenging, even not feasible, due to the exponential growth in system state and action space size with the number of components in the system. To address the scalability issue in CBM of large multi-component systems, we propose a Component-Wise Markov Decision Process(CW-MDP) and an Adjusted Component-Wise Markov Decision Process (ACW-MDP) to obtain an approximation of the optimal system-level CBM decision policy for large systems with heterogeneous components. We propose using an extended single-component action space to model the impact of system-level setup cost on a component-level solution. The theoretical gap between the proposed approach and system-level optima is also derived. Additionally, theoretical convergence and the relationship between ACW-MDP and CW-MDP are derived. The study further shows extensive numerical studies to demonstrate the effectiveness of component-wise solutions for solving large multi-component systems.

Optimal production and predictive maintenance decisions for deteriorated products under advance-cash-credit payments

In the 4th industrial revolution era, manufacturers tend to integrate predictive maintenance into production systems to maximize the lifespan of equipment and avoid costly disruptions. Using internet of things sensors and data analyses, predictive maintenance can significantly reduce the production downtime by detecting and predicting system problems before they become uncontrollable, resulting in a lower defective rate. Additionally, an advance-cash-credit (ACC) payment scheme is commonly applied in real-world business transactions to enhance cash flow flexibility among supply chain members. In this study, we developed a supplier–manufacturer chain, wherein a predictive-maintenance-adopting manufacturer received an ACC payment from a supplier, to demonstrate a supply chain management situation within the imperfect economic production quantity framework. The inventory system of the manufacturer was modeled as a cost minimization problem to determine the optimal replenishment cycle time and predictive maintenance effort for deteriorated products. The proposed model also adopted a discounted cash-flow analysis to consider the time value of the cost function, which was proven to be strictly convex in the replenishment time and predictive maintenance effort. Furthermore, a numerical and sensitivity analysis was conducted to illustrate the effectiveness of the proposed model and gain management insights.

Pipeline reliability assessment and predictive maintenance considering multi-crack dependent degradation

Cracks due to corrosion are one of the main reasons for natural-gas pipeline leaks. Making the reliability assessment, prediction, and maintenance decision of pipelines based on measurable crack data is a central issue at present. The failure of pipelines is usually a result of the cumulative impact of multiple cracks. The interaction between adjacent cracks accelerates crack propagation, and greatly affects the degradation mechanism of pipelines. In this study, the reliability prediction and maintenance decisions were studied by considering the dependent degradation between multiple cracks in pipelines. Firstly, the initiation and propagation of pipeline cracks were modeled using a non-homogeneous Poisson process and a Gamma process, respectively. The interaction between cracks was defined to be a function of the random crack distance, which could be reflected by the change of shape parameters in the Gamma process. Secondly, the pipeline’s failure was defined as the competitive failure of the number of cracks, the maximum crack depth, and the total crack depth. The reliability prediction model of a pipeline under this failure mode was determined. A non-periodic combined maintenance policy considering both the pipeline condition and its predictive reliability was then proposed, and an optimal predictive maintenance decision model was constructed to minimize the long-term average cost rate. Finally, the effectiveness of the proposed model and policy was verified by a numerical experiment and a crack dataset of a transnational pipeline.

Condition-based maintenance optimization for multi-component systems considering prognostic information and degraded working efficiency

Recent developments in prognostic models have led to increased use of prognostic information to improve the benefits of condition-based maintenance (CBM). Existing CBM decision-making methods are often limited by the assumption that prognostics information is available for all components and by the neglect of economic loss due to system degradation in assessing benefits. This paper aims to find optimal maintenance decisions with maximum net revenue for multi-component systems using prognostics information. In contrast to the reported works, three main contributions are made. First, considering that the remaining useful life (RUL) of only some components in the system can be predicted continuously, maintenance decisions, including planned inspection intervals and prognostic thresholds, are optimized jointly at the system-level. Second, the loss of efficiency due to system degradation is investigated when assessing the benefits of maintenance decisions. A degradation-related system efficiency model is proposed, and its parameters can be determined from historical data. Third, we extend our model with prognostic error modeling to derive insights on the effects of the prognostic model on the relative benefit of CBM. A case of maintenance optimization of wind turbine farms is provided to demonstrate and validate the proposed method.

Multiobjective Flexible Job-Shop Rescheduling With New Job Insertion and Machine Preventive Maintenance.

In the actual production, the insertion of new job and machine preventive maintenance (PM) are very common phenomena. Under these situations, a flexible job-shop rescheduling problem (FJRP) with both new job insertion and machine PM is investigated. First, an imperfect PM (IPM) model is established to determine the optimal maintenance plan for each machine, and the optimality is proven. Second, in order to jointly optimize the production scheduling and maintenance planning, a multiobjective optimization model is developed. Third, to deal with this model, an improved nondominated sorting genetic algorithm III with adaptive reference vector (NSGA-III/ARV) is proposed, in which a hybrid initialization method is designed to obtain a high-quality initial population and a critical-path-based local search (LS) mechanism is constructed to accelerate the convergence speed of the algorithm. In the numerical simulation, the effect of parameter setting on the NSGA-III/ARV is investigated by the Taguchi experimental design. After that, the superiority of the improved operators and the overall performance of the proposed algorithm are demonstrated. Next, the comparison of two IPM models is carried out, which verifies the effectiveness of the designed IPM model. Last but not least, we have analyzed the impact of different maintenance effects on both the optimal maintenance decisions and integrated maintenance-production scheduling schemes.

Data-driven system prognostics and diagnostics for SMT component placement machines

Prognostic-based maintenance analyses past events and predicts the future state of a machine based on the understanding of the degradation function of the machine’s components. Diagnostics-based maintenance tests equipment according to a fixed routine for a machine’s proper functioning and reliability. Current Surface-mount Technology (SMT) machines are not equipped with self-prognostic and diagnostic functions. In this paper, a system prognostic and diagnostic method is proposed, implemented in software, for estimating a machine’s health condition and faulty components of a SMT component placement machine outfitted with machine logs that consist of take-up count, miss count and time information. At each execution period the method processes features extracted from the machine logs to obtain a set of parity parameters, which are further used to analyse the machine. The prognostic algorithm computes the health status indicator of the component placement machine. The computed final status indicator is compared to a threshold value to check the system’s health condition. The diagnostic algorithm predicts and identifies the faulty pick-up nozzles and faulty input trays. The proposed algorithms minimise the effects of faulty components on production lines and assist to produce optimal maintenance decisions and reliability functions for equipment.

The modeling of risk perception in the use of structural health monitoring information for optimal maintenance decisions

This paper proposes an approach to select a maintenance strategy from a predefined set of choices considering the decision maker’s behavioral risk profile. It is assumed that the damage state is characterized by a continuous state parameter probabilistically inferred from observable sensor data. This work applies an engineering application of consequence-based decision-making incorporating the acceptable risk intensity of the decision-maker, i.e., the decision-maker’s (individual or an organization) valuation of the outcome of a decision, using a risk profile model. The utility of a decision-maker is subjective, and this paper considers the fact that different decision-makers mentally assign a different importance factor (the utility) to the seriousness or urgency to take necessary actions with the increasing intensity of structural damage. The approach herein incorporates a layer of human psychology on selecting appropriate maintenance strategies that not only depend on the posterior distribution of unmeasurable damage state but also consider the behavioral risk profile of the decision-maker. The collective decision-making of an organization consisting of many individuals is also investigated. The approach is exemplified in a case study involving life cycle monitoring of a miter gate, part of a lock system enabling navigation of inland waterways.

Reliability-Centered Maintenance Strategy for Redundant Power Networks Using the Cut Set Method

In this study, we present an efficient maintenance strategy for securing stable power supplies in power facilities. In general, industrial customers that may face large outage costs from power supply interruptions prepare highly reliable systems with redundant network topologies. When establishing a maintenance schedule for an industrial power system, the condition of the equipment should be inspected first. However, frequent inspections may result in excessive inspection costs, and infrequent inspections may reduce the reliability of the system. Therefore, it is necessary to determine an optimal maintenance and inspection interval that balances the demand for reliable electricity and the need to limit maintenance and inspection costs. Accordingly, considering the topology of the equipment in terms of the system as a whole, a reliability-centered maintenance algorithm was derived to determine the specific combinations of equipment conditions that cause power outages. Generally, the more redundant the elements in the system were, the less the load experienced power outages. We also examined the unavailability, which is an indicator that predicts the probability of the load becoming unpowered. Finally, we obtained an optimal inspection interval and maintenance decision table for events that occur in practice.

Semi-supervised clustering-based method for fault diagnosis and prognosis: A case study

Recent increased enthusiasm towards Industrial Artificial Intelligence (IAI) coupled with advancements in smart sensor technologies have resulted in simultaneous incorporation of several Condition Monitoring (CM) technologies within manufacturing/industrial sectors. Smart utilization of CM data leads to enhanced safety, reliability and availability of manufacturing systems. Conventional system monitoring techniques, however, cannot efficiently cope with such rich CM information content. In this regard, the paper proposes a novel hybrid Maintenance Decision Support System (MDSS) for fault diagnostic and prognostics considering event-triggered CM data. The proposed MDSS is a hybrid framework designed by coupling Machine Learning (ML)-based models and statistical techniques. More specifically, the MDSS is a time-dependent Proportional Hazard Model (PHM) augmented with semi-supervised ML approaches and Reinforcement Learning (RL) to find an optimal maintenance policy for systems subject to stochastic degradations with focus on cost minimization. The developed hybrid model is capable of inferring and fusing high-volume CM data sources in an adaptive and autonomous fashion to recommend optimal maintenance decisions without human intervention, which is a step-forward contribution in the maintenance context. To evaluate the structure and performance of the proposed model, comprehensive ML-based solutions are developed based on a dataset provided by NASA containing run-to-failure and CM data associated with aircraft engines.

Automated asphalt pavement damage rate detection based on optimized GA-CNN

Pavement needs to be maintained from the moment its service life begins. The maintenance strategy is mainly based on pavement quality indexes, such as the road damage rate (DR). Accurate pavement distress detection is generally costly and complicated. An inappropriate pavement distress detection strategy could yield a low efficiency of budget usage and untreated road diseases. This study describes an innovative vision-based pavement crack detection strategy that provides a direct pavement surface condition index (PCI) for a specific pavement location. This strategy was achieved with the application of the convolutional neural network (CNN) algorithm to mine a database that contains over five thousand pavement distress images to classify the pavement crack type. To improve the accuracy of the network, a genetic algorithm (GA) was employed to optimize the model. To simultaneously consider the efficiency and accuracy, an enhanced image processing method was used to measure the DR of the pavement. The model input image size is 100 × 100 pixels and the test results showed that the proposed method has satisfactory performance in robustness, accuracy, and processing speed. The accuracy of classifying different crack types reached 98%, and the processing time for an image is 0.047 s. This research, for the first time, proposed an ellipse fitting method to calculate the block of mesh cracks. The overall accuracy in detecting the pavement damage rate reached 90%. This study demonstrates the potential of an innovative deep learning method in automated pavement quality index calculations to provide direct guidance for long-term asphalt pavement optimal rehabilitation and maintenance (R&M) decisions.

Selective Maintenance Policy of Complex Systems With Maintenance Priority Indexes

The maintenance priority indexes-based (MPIs-based) selective maintenance policy (SMP) for a complex system with degradation components is studied in this work. The SMP is performed in a scheduled break when the system just completed the current mission. During the break, the MPIs, including maintenance importance, structure importance, and reliability importance, are comprehensively considered to support the system for executing the next mission. The objective of the MPIs-based SMP is to find the optimal maintenance decision for the system subject to maintenance cost and time constraints. Besides, the maintenance quality related to component degradation, as well as the economic dependence, are also considered as the maintenance information to be integrated when performing a maintenance decision. The complicated optimization problem is solved by using a simulated annealing algorithm (SAA). The simulation example of an aero-engine control system is presented to depict the maintenance decision process and illustrate the advantage of MPIs-based SMP.

Automated decision making in highway pavement preventive maintenance based on deep learning

Decision-making in highway preventive maintenance (PM) is generally costly and complicated. An inappropriate maintenance strategy could yield a low efficiency of budget usage and untreated road distress. This study describes an innovative predictive maintenance strategy that provides direct maintenance guidance to specific highway mileposts. This was achieved with the application of the artificial neural network (ANN) algorithm to mine a maintenance database. Ten-year distress measurement data at 100-m intervals, traffic load data, climatic history, and maintenance records of a chosen highway were regarded as the input data of the ANN model. A data quality control method was proposed to ensure asphalt pavement performance improvement continuity over time based on the idea of the maintenance year as the starting point for prediction. The backpropagation neural network (BPNN) model and a hybrid neural network (HNN) were applied to predict five indexes of the highway asphalt pavement performance, and the genetic algorithm (GA) was employed to optimize the hyperparameters of these models. The results indicate that the GA enhanced HNN model could increase the accuracy by 35% on average compared with traditional ANN in predicting the highway asphalt distress performance. Furthermore, a notable agreement is attained when comparing the predicted indexes to the whole-year measurement data invalidation with average coefficient of determination (R2) reaches 0.74. This study demonstrates the potential of an innovative ANN method in highway distress prediction to provide direct guidance for long-term highway asphalt pavement optimal rehabilitation and maintenance (R&M) decisions.

Dynamic Predictive Maintenance Scheduling Using Deep Learning Ensemble for System Health Prognostics

Modern engineering systems are usually equipped with a variety of sensors to measure real-time operating conditions. Based on the condition monitoring data from multiple sensor sources, this paper deals with a dynamic predictive maintenance scheduling using deep learning ensemble for system health prognostics. The deep learning ensemble model is composed of deep autoencoder and bidirectional long short-term memory in series, and aims to accurately estimate the system health state and remaining useful life. The deep autoencoder is used to extract the deep representative features hidden in condition monitoring data, whereas the inclusion of the bidirectional long short-term memory allows learning the temporal correlation information of features in both forward and backward time directions. Thus, the combination of the two models forms an effective model. With obtained prognostic information, optimal maintenance decisions are determined by two designed rules. The first rule presets a reliable remaining useful life threshold to dynamically judge whether maintenance is carried out at current inspection time, while the second one selects a degradation state to dynamically determine whether to place a spare part order. The performance of the proposed dynamic predictive maintenance strategy is measured by the maintenance cost per unit operating time (cost rate) using the aero-engine dataset from NASA. Its effectiveness is demonstrated by comparing with recent publications.