Maintenance Policy Research Articles

Adopting New Machine Learning Approaches on Cox’s Partial Likelihood Parameter Estimation for Predictive Maintenance Decisions

The Proportional Hazards Model (PHM) under a Condition-Based Maintenance (CBM) policy is used by asset-intensive industries to predict failure rate, reliability function, and maintenance decisions based on vital covariates data. Cox’s partial likelihood optimization is a method to assess the weight of time and conditions into the hazard rate; however, parameter estimation with diverse covariates problem could have multiple and feasible solutions. Therefore, the boundary assessment and the initial value strategy are critical matters to consider. This paper analyzes innovative non/semi-parametric approaches to address this problem. Specifically, we incorporate IPCRidge for defining boundaries and use Gradient Boosting and Random Forest for estimating seed values for covariates weighting. When applied to a real case study, the integration of data scaling streamlines the handling of condition data with diverse orders of magnitude and units. This enhancement simplifies the modeling process and ensures a more comprehensive and accurate underlying data analysis. Finally, the proposed method shows an innovative path for assessing condition weights and Weibull parameters with data-driven approaches and advanced algorithms, increasing the robustness of non-convex log-likelihood optimization, and strengthening the PHM model with multiple covariates by easing its interpretation for predictive maintenance purposes.

Optimization of a preventive maintenance strategy for a product with limited repair times

This paper proposes a base warranty and replacement warranty strategy for repairable products. It should be emphasized that the product failure rate is low in the early stage of the base warranty. Therefore, in order to reduce the manufacturer’s maintenance cost, the manufacturer does not carry out preventive maintenance in the early stage of the base warranty, and carries out regular preventive maintenance in the later stage of the base warranty. At the same time, if the number of product failures in the base warranty exceeds the specified limit, the manufacturer must replace products for customers free of charge according to the warranty contract. After the product is replaced, if the number of product failures does not exceed the limit, the manufacturer will continue to provide customers with the remaining base warranty period of the original product, and the warranty period will continue to be extended for a period of time. By minimizing the expected cost rate, it is shown that the optimal maintenance policy for warranty products is unique. Finally, we give a numerical example and study the effects of periodic preventive maintenance and other relevant parameters on the optimal solution.

Optimal maintenance policy for imperfect production systems using reliability function and defect rate

AbstractThis study examines how a manufacturer can improve the robustness of an imperfect production system by implementing maintenance activities, which can be measured in terms of system reliability and product failure rate. A two‐dimensional maintenance policy comprising system reliability and the product defect rate is proposed to assess maintenance activity costs. The optimal thresholds of the two dimensions are analyzed to investigate the trade‐off between cost and system quality. A numerical example is provided to verify the proposed model's effectiveness. The results showed that the less stable the system, the greater the total costs incurred; therefore, lower stringent thresholds may be set to prevent frequent maintenance. Moreover, a sensitivity analysis is performed to investigate the essential parameters that significantly affect maintenance decisions. The results showed that the thresholds of the reliability function and defect rate significantly impact total costs. Any inaccurate assessment of system usage could lead to incorrect estimations and a substantial increase in cost.

Contribution of Maintenance to Reconfigurable Manufacturing Systems: State of the Art and Challenges

In the modern industrial landscape, reconfigurable manufacturing systems (RMS) have emerged as a crucial element due to their rapid responsiveness and cost effectiveness to market fluctuations. This importance is demonstrated in the current industrial environment and shows significant potential in the coming era of manufacturing. Maintenance plays an important role in modern industry paradigms to ensure systems are reliable and operate efficiently. However, the integration of maintenance scheduling processes in RMS has not been significantly explored, opening new research directions. This literature review addresses this gap by systematically examining the existing works into two databases of peer-reviewed literature: Scopus and Web of Science. The purpose of this review is to provide a comprehensive understanding of the current state of the art of maintenance methods in RMS, highlighting the need for further investigation. The selected articles were analyzed and synthesized to provide insights into different maintenance strategies related to RMS under two criteria: Maintenance modeling, maintenance policy and optimization. This study also highlights the need to develop enhanced maintenance policies and methods specifically designed for RMS and describes the research challenges and opportunities that lie ahead. In doing so, it seeks to advance future research efforts and serve as a valuable resource for both researchers and practitioners in the field, paving the way for innovation and advances in RMS maintenance.

Maintenance lifecycle cost analysis through Agent-Based Simulation

Proper maintenance management allows companies to reduce failures and breakdowns while keeping high asset productivity. Several variables and uncertainties affect the selection of the optimal maintenance policy, and the one that might seem more cost effective in the short-term, might reveal having a total higher cost over the complete asset lifecycle. In this sense, the use of simulation to quantify the maintenance costs over the entire asset lifecycle can be helpful, allowing companies to test different scenarios and compare their result in a dynamic way, looking at long-term effects instead of short-term ones. The paper proposes an agent-based simulation model that allows to quantify the total costs associated with different maintenance policies, as well as computing productivity and important maintenance indexes to help companies in evaluating the most suitable maintenance strategy, being it used internally, or as a selling point for maintenance services towards customers.

Integrating maintenance and energy problems through a Digital Twin-based decision support framework under the guidance of Asset Management

In a world where more than one third of the global energy consumption is directed to the industry sector, manufacturing companies must overhaul their business processes to move towards the improvement of sustainable performance. This work aims to contribute to this direction by integrating maintenance and energy problems through a Digital Twin-based decision support framework under the guiding role of Asset Management. The framework is meant as guideline for manufacturing companies to manage the integrated decision-making process and the related information flows and functional blocks. Maintenance or energy policies should consider both maintenance and energy objectives, handling the complexity and knowledge required for such an integration. A focus is given to the maintenance policy setting at the current release of the framework. This latter is tested and validated by an initial proof in a laboratory setting. In the future, it should be evaluated in its benefits within an industrial case.

The Conceptual of Pavement Management System Based on IoT, Big Data, and Data Mining in Indonesia

The country’s economic development and people’s lives require reliable transportation and logistics services. Therefore, road maintenance policies and techniques must be more comprehensive to ensure transportation infrastructure operations’ efficiency. The inefficiency of road maintenance, among others, is due to the failure to identify and predict the level of damage that can cause fundamental data anomalies. Therefore, the current scientific development of road maintenance must be able to take advantage of the convenience of technology. Widespread technology implementation on the Internet of Think (IoT), big data, and data mining (DM) can be a solution to developing a better pavement management system. However, implementing these technologies is challenging due to the weak conception and availability of currently available resources. This paper offers a pavement system management implementation concept that can be developed in Indonesia to improve the existing system. The study results show that IoT, big data, and DM are believed to support a more intelligent and comprehensive pavement management system. The concept offered consists of three main things: identification of damage and 3D pavement modeling, data analysis, decision support systems, and collaboration of intelligent solutions in implementing maintenance in the field. This paper is equipped with an illustration of the concept of road maintenance based on IoT, big data, and DM, which can be implemented intelligently and simply in Indonesia.

A Condition-Based Maintenance Policy Considering Batch Sizes for Warm Standby Systems With Priority to Repair

A Condition-Based Maintenance Policy Considering Batch Sizes for Warm Standby Systems With Priority to Repair

Optimal condition-based maintenance policy for multi-component repairable systems with economic dependence in a finite-horizon

Optimal condition-based maintenance policy for multi-component repairable systems with economic dependence in a finite-horizon

A condition-based maintenance policy with non-periodic inspection for k-out-of-n: G systems

A condition-based maintenance policy with non-periodic inspection for k-out-of-n: G systems

Preventive Maintenance Policy to Improve the Effectiveness of XYZ Machine

PT ABC which is a manufacturing company that produces steel pipes. The last period there was an XYZ machine that often downtime and disrupted the production process. The breakdowns indicate the ineffectiveness of machine. It was necessary to identify and analyze the problems that occurred and the factors that caused the problem. Measurement of machine effectiveness dan it caused is carried out using the OEE method. To avoid the recurrence of machine ineffectiveness, machine maintenance is carried out in a preventive manner according to the spare parts classification (there are 3 classifications of spare parts based on price, namely A under Rp.250,000, B between Rp.250,000 -Rp.500,000 and C above Rp.500,000). The results obtained for the average value of OEE for 1 year is 68%, which shows a value below the ideal OEE standard (85%) and needs to be evaluated. From the calculation of the largest losses value is in the reduce speed losses factor of 67%, it is necessary to minimize the difference in ideal cycle time with actual time (operation time). The best scheduling results is using the preventive maintenance policy method with the schedule of maintenance are: for classification A spare parts the maintenance every 3 months, class B every 5 months, and class C every 6 months.

Relationship between the Maintenance Management Problems and the Maintenance Performance of High-Rise Residential Buildings

Building maintenance is the practice to retain the property habitable and functioning as it is designed for. However, the effectiveness of maintenance management appears to be the issue in housing context. Many maintenance management problems still exist up to date. They are likely to influence the service quality and resident satisfaction. Therefore, the objective of this research is to establish the relationship between the maintenance management problems and the maintenance performance. The research adopted quantitative approach to achieve the research aim. The approach comprised of literature review, questionnaire survey using google form that included both close-ended and open-ended questions. Based on the literature review, fourteen (14) maintenance management problems of high-rise residential buildings are identified. From the correlation analysis result, two (2) significance correlations are determined. The maintenance budget overrun is significantly correlated with the ratio of actual maintenance spent to planned maintenance cost (cost variance) of high-rise residential building, and a significant correlation occurs between unavailability of spare parts and ratio of actual to planned maintenance downtime (equipment effectiveness) of high-rise residential building. Besides, the opinions and strategies on solving the maintenance management problems were provided, including educating the residents about the Strata Management Act 2013, selecting qualified maintenance contractor, outsourcing the maintenance services, implementation of preventive maintenance (PM), hiring specialist to perform special maintenance work, providing training for maintenance staffs, formulating the building maintenance regulation, ensuring transparency of financial account, and setting up maintenance policy.

RELIABILITY AND AVAILABILITY IMPROVEMENT TECHNIQUES FOR NIGERIA ELECTRIC POWER STATIONS

The maintenance of plant in the electric power stations involves an annual expenditure of millions of dollars and it is therefore essential to establish the cost-effectiveness of the present maintenance policies. The major problems in this evaluation are the lack of information and lack of suitable techniques to analyze the information. Most items of electric power plants are complex mechanical units and as such it is extremely difficult to predict their failure although condition monitoring can sometimes be used. This study explores how to use reliability engineering to identify best practice for maintenance work in electric power stations. It identified how to identify failure before their occurrences and how the maintenance management functions can become more conscious of the sole of maintenance management in enhancing has become the prominent management issues for electric utilities.

Integration of production–maintenance planning and monitoring simple linear profiles via Hotelling's T2 control chart and particle swarm optimization

This paper introduces an integrated model that merges three key concepts: quality control through linear profile monitoring using Hotelling's T2 control chart, production cost evaluation, and maintenance policies. The objective function of the integrated model is to minimize total costs. This study is the first to jointly consider production and maintenance costs in the framework of profile monitoring using Hotelling’s T2 control chart for linear profiles without compromising on the versatility of simple linear profiles as a modeling tool, particularly in calibration applications. Appropriate assumptions are made for the proposed integrated model and the particle swarm optimization algorithm is implemented to determine decision variables. To evaluate the effectiveness of the integrated model, a numerical example and a comprehensive comparative analysis are conducted. Moreover, a sensitivity analysis to study the effects of various input parameters on the optimal solutions is performed. The findings suggest that adopting the integrated model with specific decision variables results in lower expected total costs when compared to the joint model and fixed parameter schemes.

Maintenance optimization of multi-unit balanced systems using deep reinforcement learning

Balanced systems are extensively utilized in various engineering fields. However, the study of maintenance optimization for balanced systems under dependent competing risks is very limited, and the existing method for orthodox multi-unit systems cannot solve the maintenance problem of balanced systems. Thus, we present a novel maintenance optimization method for systems with multiple balanced units. The units in the system are assumed to experience multiple failure processes due to continuous degradation and external shock damage. The system failure occurs in the following situations: (1) the deterioration levels of units on the symmetric positions exceed critical thresholds. (2) the deterioration level of any unit reaches the failure threshold. (3) the shock magnitude is greater than the hard failure threshold. By employing a customized deep reinforcement learning algorithm, we derive the optimal maintenance policy under a fixed planning horizon. Notably, our approach differs from previous research on maintenance optimization for balanced systems, the proposed method avoids the discretizing of the deterioration state. This significantly enhances computational efficiency, particularly when applied to multi-unit balanced systems. Numerous numerical studies illustrate the effectiveness of our proposed maintenance optimization method for multi-unit balanced systems.

Decision making in operator-machine assignment problems: an optimization approach in U-shaped production lines

Decision-making is a complex problem in all fields of production, because decision makers have to find the most suitable solution alternative among the numerous, while the solution of the problem is limited by multiple, usually conflicting constraints. In the Industry 4.0 era, one of the most widely used production lines are the U-shaped production lines, which focus on the elimination of waste through a high utilization of workers. It means, that worker-machine assignment plays a significant role in the optimization of production processes in a U-shaped production line. The current investigation proposes a novel absorbing Markov chain (AMC) optimization approach to support the decision making regarding worker selection and assignment. The proposed approach integrates the AMC and the performance analysis of the production process. The numerical results validated the decision making model and showed that optimal worker-machine assignment can lead to about 20% production cost reduction, while the key performance indicators (KPI) of the U-shaped production line are significantly increased. The sensitivity analysis of influencing factors made it possible to identify the critical factors of this decision making problem, such as the qualification of operators, time of technological and logistics operations, and maintenance policy.

A condition-based maintenance policy for reconfigurable multi-device systems

A condition-based maintenance policy for reconfigurable multi-device systems

Incorporating cascading effects analysis in the maintenance policy assessment of torrent check dams against torrential floods

In mountainous regions, protection infrastructures designed to mitigate the impacts of torrential floods often consist of a complex system of several structural components (check dams). Over time, the efficacy of this system in protecting downstream assets diminishes as the structural components deteriorate. The extent of deterioration is influenced by the interdependencies between the failure modes of individual components, as well as those between multiple components of the system. Understanding and quantifying the chain of failure events, known as cascading effects, is a critical scientific challenge that remains largely unexplored. In this study, we propose a novel approach that employs physics-based models to examine the deterioration of a series of check dams over time, while considering failure dependencies and bidirectional interactions between consecutive dams. The results obtained from this approach reveals that the absence of a downstream dam accelerates the deterioration rate of upstream dams, while its presence serves to stabilize them. We further incorporate stochastic deterioration and maintenance processes using Stochastic Petri nets to support decision-making regarding maintenance actions for each dam, while also considering economic factors. Strategies involving minor operations achieved cost-effectiveness and prolonged satisfactory performance of the dams, with notable impacts from upstream and downstream dam presence on maintenance costs. We illustrate our approach using a case study of the Faucon torrent in France, where we model the deterioration of three consecutive check dams subjected to torrential floods over a period of 100 years.

Quantifying Reliability Indices of Garbage Data Collection IOT-based Sensor Systems using Markov Birth-death Process

The aim of this paper is to analyze the performance of a Garbage data collecting sensor network system (GDCSNS) through mathematical modelling and a reliability approach. The reliability assessment of such a system is essential to ensuring that it can collect data related to garbage at different locations consistently and accurately. After the determination of the reliability measures of the system, the next aim is to identify the weakest sensors of the system so that a timely maintenance strategy for the weakest sensor can be planned to avoid disruption in the collection of data from the sensor systems. In the considered system, three sensors have been installed at various location in the city that send the information to the center office (hub point) and then from the center office to the person who is responsible for collecting the garbage from the location and dumping it at some predefined places. These sensors collect data related to garbage level, weight, and other information and send it to computers at the city's central office. Markov modelling has been used to model the system. Based on the mathematical model, a state transition diagram and a set of Kolmogorov time-dependent differential equations have been obtained. The various state probabilities (explicit expressions) related to the performance of the system, namely, Reliability, Mean time to failure, have been obtained to understand the different maintenance policies that can be used. A sensitivity analysis has also been performed to determine the weakest sensor among the sensors.

A scalable multi-objective maintenance optimization model for systems with multiple heterogeneous components and a finite lifespan

Delivering after-sales maintenance services is a challenging task for original equipment manufacturers who must tailor their offers to customers’ needs. To accommodate this challenge, we propose a multi-objective optimization model to derive optimal maintenance policies for systems with a large number of heterogeneous components over a finite lifespan. Based on the practical assumption of a fixed time interval between two consecutive scheduled downs, we develop a decomposition approach that enables the fast computation of the optimal policy per component and the optimal time interval between two consecutive scheduled visits. We also analyze the structural properties of the optimal policies for both age-based and condition-based maintenance and prove that they have a non-static control-limit structure. Using a set of computational experiments, we first investigate the computational tractability of our model for systems with an increasing number of components. Then, we apply our model to a real case study to demonstrate how it can be used to derive optimal maintenance policies tailored to different customer needs.