Maintenance Scheduling Model Research Articles

Multi-objective predictive maintenance scheduling models integrating remaining useful life prediction and maintenance decisions

This paper presents two novel data-driven multi-objective predictive maintenance scheduling models that integrate remaining useful life (RUL) prediction into maintenance planning. First, we propose a deep learning ensemble model that includes a convolutional neural network and bidirectional long short-term memory network with a temporal self-attentional mechanism and Bayesian optimization method to predict system RUL effectively. Then, two novel multi-objective mixed-integer linear programming (MILP) models are developed based on the system-predicted RUL to deal with the system predictive maintenance problem, which aims to minimize the total maintenance completion time and maintenance-related costs simultaneously. To solve this problem, we design an iteration ϵ-constraint method to achieve Pareto solutions. Meanwhile, a fuzzy logic method is proposed to recommend a preferred Pareto solution for decision-makers. Finally, the aircraft engine C-MAPSS dataset from NASA validates that the effectiveness of the proposed data-driven multi-objective predictive maintenance scheduling models are effective. For the predictive maintenance of 20 aircraft engines, both the maintenance completion time and maintenance cost are reduced by more than 40%.

DivideMerge: A multi-vessel optimization approach for cooperative operation and maintenance scheduling in offshore wind farm

The optimization of maintenance scheduling and routing in offshore wind farms is crucial for the intelligent operation and maintenance of offshore wind energy. It involves determining the optimal timing, assigning vessels, planning routes, and efficiently completing maintenance tasks with maximum efficiency and minimal cost. Traditional branch-and-bound search methods, often used in multi-vessel cooperative maintenance scheduling models, face limitations in finding optimal solutions efficiently, while classic heuristic methods may compromise on global optimality. This paper introduces DivideMerge, a high-performance hybrid algorithm that synergizes heuristic and exact algorithms to address large-scale multi-vessel cooperative maintenance scheduling challenges in offshore wind farms. Initially, a maintenance task constraint decomposition heuristic method is utilized to break down the collaborative optimization scheduling problem into individual vessel scheduling sub-problems, ensuring adherence to the constraints of the original problem. Subsequently, the CP-SAT solver is employed to sequentially solve these sub-problems rapidly. The solutions for individual vessel scheduling are then merged to form a comprehensive solution for the multi-vessel cooperative maintenance scheduling problem. Computational results validate the effectiveness and robustness of DivideMerge, demonstrating a solution speed nearly 1000 times faster than the commercial Gurobi solver, thus offering a significant advancement in the field of offshore wind farm maintenance optimization.

An efficient generator maintenance scheduling model based on unit clustering and linear relaxation

The formulation of periodic maintenance schedules for generators is crucial for the efficient and reliable operation of power systems. Traditional generator maintenance scheduling models often have low temporal resolution and ignore significant information. To solve this problem, this paper proposes a mid/ long-term generator maintenance scheduling model with an hourly time resolution. Given the challenges of large integer variable quantities and computational complexity in directly solving mid/ long-term maintenance schedules for power system generation units, an efficient solution method is proposed based on unit clustering and linear relaxation. Initially, a basic maintenance model is established based on the Unit Commitment (UC) model. Subsequently, the basic UC maintenance model is linearly relaxed to obtain an efficient approximate model, termed the Relaxed Clustered Unit Commitment (RCUC) maintenance model, which involves the operation of annual regulating hydropower stations, energy storage devices, pumped storage stations, wind power plants, and photovoltaic (PV) plants. Simulation results from an adapted IEEE RTS-79 system and a provincial scale system case study are provided to demonstrate the effectiveness of the proposed model in coping with mid/long-term maintenance scheduling and annual hydropower allocation problems of large-scale power systems.

A stochastic track maintenance scheduling model based on deep reinforcement learning approaches

A stochastic track maintenance scheduling model based on deep reinforcement learning approaches

A mathematical model for optimal turnaround maintenance planning and scheduling for a network of plants in process industry supply chain

Turnaround (TAM) is a periodic maintenance in which plants are shutdown to conduct inspections, repairs, preventive maintenance and overhauls. The shutdown periods range from three to four weeks and prolonging this period is expensive and therefore requires effective planning. The primary purpose of this paper is to develop a mathematical programming model to coordinate and plan TAM for several plants in an integrated supply chain network. A secondary objective is to develop a heuristic algorithm for solving the model when the network has large number of plants. The model and the algorithm are developed, tested and applied on a case study petrochemical supply chain. The exact algorithm takes 8.3 times the time of the heuristic. In addition the heuristic algorithm provides near optimal solution in much shorter time. The developed model provides a tool to plan TAM for several plants in an integrated supply in an optimal fashion.

Stochastic generation maintenance scheduling with inertia-dependent primary frequency regulation constraints

The problem of insufficient inertia and frequency security becomes a critical concern in generation maintenance scheduling due to the increasing penetration of renewable energy sources (RESs). To address the issue, this paper presents a stochastic generation maintenance scheduling with inertia-dependent primary frequency regulation constraints. First, inertia-dependent primary frequency regulation constraints consisting of the minimum inertia requirement, frequency nadir, and quasi-steady-state frequency deviation are formulated based on the dynamic frequency characteristic of power systems in response to large disturbances. Then, a two-stage stochastic optimization model for generation maintenance scheduling including frequency security constraints is proposed. Finally, a solution methodology based on Benders decomposition (BD) is used to promote the calculation speed of the proposed model. Numerical simulations demonstrate that the proposed generation maintenance scheduling could address the potential frequency security issues and promote integration of RESs associated with calculation speed.

Dynamic inspection and maintenance scheduling for multi-state systems under time-varying demand: Proximal policy optimization

Inspection and maintenance activities are effective ways to reveal and restore the health conditions of many industrial systems, respectively. Most extant works on inspection and maintenance optimization problems have assumed that systems operate under a time-invariant demand. Such a simplified assumption is oftentimes violated by a changeable market environment, seasonal factors, and even unexpected emergencies. In this article, with the aim of minimizing the expected total cost associated with inspections, maintenance, and unsupplied demand, a dynamic inspection and maintenance scheduling model is proposed for Multi-State Systems (MSSs) under a time-varying demand. Non-periodic inspections are performed on the components of an MSS and imperfect maintenance actions are dynamically scheduled based on the inspection results. By introducing the concept of decision epochs, the resulting inspection and maintenance scheduling problem is formulated as a Markov Decision Process (MDP). The Deep Reinforcement Learning (DRL) method with a Proximal Policy Optimization (PPO) algorithm is customized to cope with the “curse of dimensionality” of the resulting sequential decision problem. As an extra input feature for the agent, the category of decision epochs is formulated to improve the effectiveness of the customized DRL method. A six-component MSS, along with a multi-state coal transportation system, is given to demonstrate the effectiveness of the proposed method.

Bayer digestion maintenance optimisation with lazy constraints and Benders decomposition

This paper describes a maintenance scheduling model for digester banks. Digester banks are network-connected assets that lie on the critical path of the Bayer process, a chemical refinement process that converts bauxite ore into alumina. The banks require different maintenance activities at different due times. Furthermore, the maintenance schedule is subject to production-related constraints and resource limitations. Given the complexity of scheduling maintenance for large fleets of digester banks, a continuous-time, mixed-integer linear program is formulated to find the cost-minimising maintenance schedule that satisfies all required constraints. A solution approach that employs lazy constraints and Benders decomposition is proposed to solve the model. Unlike generic implementations of Benders decomposition, we show that the subproblems can be solved explicitly using a specialist algorithm. We solve the scheduling model for realistic scenarios involving two Bayer refineries based in Western Australia.

An iterative improvement approach for high-speed train maintenance scheduling

The high-level maintenance plan is the train-set overhaul arrangement that is scheduled annually and manually in China. Previous studies have investigated this issue based on deterministic daily mileages. However, future daily mileages are difficult to predict, which may cause existing plans to be inaccurate and unfeasible in practice. Therefore, this study only considers the historical daily mileages as raw inputs, and the future daily mileages are considered as ranges to generate wider maintenance time windows for train-sets. Subsequently, a 0–1 integer linear programming model for high-level maintenance scheduling is formulated. After obtaining a high-level maintenance plan, the planned daily mileages of all train-sets are calculated and verified. To make all planned daily mileages feasible, we design an iterative algorithm to adjust the time windows and update the plan. A real-world case study is conducted using the data of 124 CRH2 EMU train-sets belonging to China Railway Shanghai Group to prove the effectiveness of the model and the algorithm. The commercial solver Gurobi is used to solve this case. A program supporting high-level maintenance scheduling has been developed, and it has been used by planners for testing.

Investments and maintenance spending for public facilities: what do we know and what do we want to know

PurposeThere is widespread and long-lasting worry related to the condition of public purpose buildings and public investments. Public buildings make up a huge capital stock and proper maintenance and investments are important for public policy. Notwithstanding, the relevant research literature is fragmented and spread across several fields. The authors take stock of earlier and more recent research and suggest some ideas for future research.Design/methodology/approachThe authors summarize the relevant literature and discuss implications of various theoretical assumptions and empirical findings for maintenance and investment strategies.FindingsA better understanding of the role of public facilities in public service provision is important. Relevant topics for further research are the impact of technological changes, both in buildings and service provision, economic issues including macroeconomic shocks and trends that influence public funding and demand for public services, and advancing maintenance scheduling models to consider a portfolio of facilities. Further, the empirical literature suffers from a lack of relevant data to gauge both the condition of public facilities and their impact on public services.Originality/valueThere is widespread worry that poor facilities adversely impact public services, but the size and significance of this impact are an open question. This paper contributes by taking stock of the existing research on public facilities, maintenance, and investments, and suggest ideas for further work.

Integrated maintenance, inventory and quality engineering decisions for multi-product systems

Abstract It is essential for manufacturers to consider the interrelation among quality, inventory, and maintenance decisions to detect imperfect quality products, keep the production system in good operating condition, and manage quality and inventory costs. Hence, this paper aims to develop an integrated model of inventory planning, quality engineering, and maintenance scheduling in which the expected total cost per time unit is minimised by determining the sample size, sampling interval, control limit coefficient, along with production cycle time. In this regard, an imperfect multi-product manufacturing system is considered, in which the inventory shortage in satisfying the demand for each product type and the idle time during the production cycle are not allowed. It is assumed that the process starts in an in-control condition where most produced units are conforming. However, due to the occurrence of an assignable cause (AC), the process mean moves to an out-of-control condition in which a significant fraction of non-conforming units is produced. The efficiency of the proposed mathematical model is evaluated by a numerical example, and then the sensitivity of the proposed model to important inputs is analysed. Finally, a comparative study based on the Taguchi design approach is given to confirm the capability of the proposed model to achieve remarkable cost savings.

Tri-objective generator maintenance scheduling model based on sequential strategy.

A multi-objective modeling approach is required in the context of generator maintenance scheduling (GMS) for power generation systems. Most multi-objective modeling approaches in practice are modeled using a periodic system approach that caters for a fixed maintenance window. This approach is not suitable for different types of generating units and cannot extend the generator lifespan. To address this issue, this study proposes a tri-objective GMS model with three conflicting objectives based on the sequential system approach that accounts for operating hours and start-up times. The GMS model’s objectives are to minimize the total operation cost, maximize system reliability and minimize violation. The main difference between the proposed tri-objective GMS model and other multi-objective GMS models, is that the proposed model uses a sequential strategy based on operating hours and start-up times. In addition, the proposed model has considered the most important criteria in scheduling the generator maintenance, and this reflects the real-life requirements in electrical power systems. A multi-objective graph model is also developed to generate the maintenance units scheduling and used in developing the proposed Pareto ant colony system (PACS) algorithm. A PACS algorithm is proposed to implement the model and obtain solution for GMS. The performance of the proposed model was evaluated using the IEEE RTS 26, 32, and 36-unit systems dataset. The performance metrics used comprise the GMS model objectives. The experimental results showed that the obtained solution from the proposed tri-objective GMS model was a robust solution by considering the different initial operational hours of the units.

An exact MILP model for joint switch placement and preventive maintenance scheduling considering incentive regulation

Power delivery with high reliability is one of the main objectives of any distribution system. In this regard, system operators are adapting the traditional distribution systems with sectionalized switches and optimal preventive maintenance (PM) scheduling techniques as two effective methods for reliability improvement. However, they suffer due to the lack of a practical model in which switch placement (SP) and PM scheduling problems are considered simultaneously. Furthermore, consideration of incentive regulation schemes, such as reward-penalty scheme (RPS) is highly questionable in the separated reliability improvement problems. In this paper, a mix-integer linear programming model is presented wherein both SP and PM scheduling problems are considered. Rather than using conventional methods for linearization, an exact formulation is proposed to avoid the impact of approximations on the final decision. Moreover, RPS is also regarded as an incentive regulation scheme to balance the reliability level and financial performance. The results of test and real network implementation confirm that a higher performance is reached in terms of reliability and financial issues when the proposed model is utilized compared to the other ones. Besides, analysis demonstrates that the proposed model behaves effectively in both mid and long terms.

Exploiting dynamic programming in optimizing reliability-centered maintenance

Maintenance Scheduling (MS) is one of the most persistent issues that might arise in a manufacturing facility. It is crucial to do routine maintenance on machinery in order to avoid unplanned breakdowns. This type of failure could cause costly manufacturing process interruptions. Numerous strategies have been developed in an effort to deal with MS. But because each system has specific requirements and limitations, this is a particularly challenging problem. The scheduling of maintenance for machine units at a manufacturing facility that produces aluminum in moderate amounts is explored in this study using a dynamic programming approach. The Maintenance Scheduling model is put into practice using a Reliability-Centered Maintenance (RCM) strategy after an investigation of the architecture and infrastructure of the plant. By maintaining machines at acceptable dependability values and minimizing maintenance costs, this method is created to optimize the maintenance schedule. Here, factors like reliability and failure rate that affect the MS problem are discussed and investigated. Applying the model to a situation that represents the aluminum manufacturing allows it to be tested in a variety of different circumstances. The results of applying the model to the test cases are given, followed by a discussion of the results. The results obtained are reasonable and show that the dynamic programming strategy is a successful way to fix the MS issue that the manufacturing plant's machines are experiencing.

A Stochastic Markov Model for Maintenance Scheduling in the Presence of Online Monitoring System

Active electricity distribution systems (AEDS) play an important role on the social welfare that is highly relies on the reliability of these systems. In this regard, maintenance management tries to decrease component failures via repairing component that are more possible to be failed in operation time. In this manuscript, a stochastic MINLP maintenance scheduling model is proposed in which the reliability of AEDS's equipment is considered through employing a three-state Markov model. This model helps us to embed the uncertainty of AEDS's equipment service restoration time by considering online monitoring system for it and FLISR functions in AEDS. In this paper, minimizing the total cost of AEDS is determined as objective function considering total customer interruption cost, total cost of energy not supplied, total cost of prevented maintenance actions, total corrective maintenance actions cost implemented in the AEDS and total cost of implementing online monitoring for AEDS's component in maintenance scheduling. The proposed framework is applied in two different cases, Case I is maintenance scheduling without online monitoring system and Case II is maintenance scheduling with online monitoring system. The aforementioned cases are implemented in an IEEE Test System as well as a real scale distribution network. Moreover, several sensitivity analyses are implemented for indicating the efficiency of the proposed model.

Learning to branch in the generation maintenance scheduling problem

To maximize the reliability index of a power system, this study modeled a generation maintenance scheduling problem that considers the network security constraints and rationality constraints of the generation maintenance practice in a power system. In view of the computational complexity of the generation maintenance scheduling model, a variable selection method based on a support vector machine (SVM) is proposed to solve the 0–1 mixed integer programming problem (MIP). The algorithm observes and collects data from the decisions made by strong branching (SB) and then learns a surrogate function that mimics the SB strategy using a support vector machine. The learned ranking function is then used for variable branching during the solution process of the model. The test case showed that the proposed variable selection algorithm — based on the features of the proposed generation maintenance scheduling problem during branch-and-bound — can increase the solution efficiency of the generation-scheduling model on the premise of guaranteed accuracy.

Opportunistic maintenance scheduling with deep reinforcement learning

The great complexity of advanced manufacturing processes combined with the high investment costs for manufacturing equipment makes the integration of maintenance scheduling a challenging, but similarly crucial task. Opportunistic maintenance scheduling holds the potential to increase the operational performance by considering the opportunity cost of maintenance measures. At the same time, reinforcement learning (RL) has proved to be able to handle complex scheduling tasks. Therefore, RL constitutes a promising approach to develop an integrated maintenance scheduling model to consider order dispatching and maintenance scheduling in a single decision support system. This paper models a real-world use case of a semiconductor front-end wafer fabrication by using discrete-event simulation. In the simulated scenarios, the performance of the integrated dispatching and maintenance scheduling is regulated by both complex novel heuristics adapted to opportunistic maintenance and reinforcement learning. The results show that the RL policy is able to learn a competitive joint scheduling strategy by including internal and external opportunistic opportunities. This indicates that opportunistic maintenance scheduling, with and without RL, holds the potential to improve the performance not only of semiconductor manufacturing but capital-intense machinery industry alike.

Remaining-Useful-Life prognostics for opportunistic grouping of maintenance of landing gear brakes for a fleet of aircraft

Several studies have proposed Remaining-Useful-Life (RUL) prognostics for aircraft components in the last years. However, few studies focus on integrating these RUL prognostics into maintenance planning frameworks. This paper proposes an optimization model for opportunistic maintenance scheduling of aircraft components that integrates RUL prognostics and that groups the maintenance of these components to reduce costs. We illustrate our approach for the maintenance of a fleet of aircraft, each equipped with multiple landing gear brakes. RUL prognostics for the landing gear brakes are obtained using a Bayesian regression model. Based on these RUL prognostics, we group the replacement of brakes using an integer linear program. As a result, we obtain a cost-optimal RUL-driven opportunistic-maintenance schedule for the brakes of a fleet of aircraft. Compared with traditional maintenance strategies, our approach leads to a reduction of up to 20% of the total maintenance costs.

Multi-Objective Model and Variable Neighborhood Search Algorithms for the Joint Maintenance Scheduling and Workforce Routing Problem

This paper addresses a problem faced by maintenance service providers: performing maintenance activities at the right time on geographically distributed machines subjected to random failures. This problem requires determining for each technician the sequence of maintenance operations to perform to minimize the total expected costs while ensuring a high level of machine availability. To date, research in this area has dealt with routing and maintenance schedules separately. This study aims to determine the optimal maintenance and routing plan simultaneously. A new bi-objective mathematical model that integrates both routing and maintenance considerations is proposed for time-based preventive maintenance. The first objective is to minimize the travel cost related to technicians’ routing. The second objective can either minimize the total preventive and corrective maintenance cost or the failure cost. New general variable neighborhood search (GVNS) and variable neighborhood descent (VND) algorithms based on the Pareto dominance concept are proposed and performed over newly generated instances. The efficiency of our approach is demonstrated through several experiments. Compared to the commercial solver and existing multi-objective VND and GVNS, these new algorithms obtain highly competitive results on both mono-objective and bi-objective variants.

Multivariate Analyses for Finding Significant Track Irregularities to Generate an Optimal Track Maintenance Schedule

We first discuss the relationship between the optimal track maintenance scheduling model and an efficient detection method for abnormal track irregularities given by the longitudinal level irregularity displacement (LLID). The results of applying the cluster analysis technique to the sampling data showed that maintenance operation is required for approximately 10% of the total lots, and these lots were further classified into three groups according to the degree of maintenance need. To analyze the background factors for detecting abnormal LLID lots, a principal component analysis was performed; the results showed that the first principal component represents LLIDs from the viewpoints of the rail structure, equipment, and operating conditions. Binomial and ordinal logit regression models (LRMs) were used to quantitatively investigate the determinants of abnormal LLIDs. Binomial LRM was used to characterize the abnormal LLIDs, whereas ordinal LRM was used to distinguish the degree of influence of factors that are considered to have a significant impact on LLIDs.