Optimal Maintenance Plan Research Articles

Deep reinforcement learning for optimal planning of assembly line maintenance

Discovering the optimal maintenance planning strategy can have a substantial impact on production efficiency, yet this aspect is often overlooked in favor of production planning. This is a missed opportunity as maintenance and production activities are deeply intertwined. Our study sheds light on the significance of maintenance planning, particularly in the dynamic setting of an assembly line. By maximizing the average production rate and incorporating flexible planning windows, buffer content, and machine production states, a unique problem is addressed in which a policy for planning maintenance on the final machine of a serial assembly line is developed. To achieve this, novel average-reward deep reinforcement learning techniques are employed and pitted against generic dispatching methods. Using a digital twin with real-world data, experiments demonstrate the immense potential of this new deep reinforcement learning technique, producing policies that outperform generic dispatching strategies and practitioner policies.

Joint Optimization of Condition-Based Maintenance and Performance Control for Linear Multi-State Consecutively Connected Systems

Industrial systems such as signal relay stations and oil pipeline systems can be modeled as linear multi-state consecutively connected systems, which comprise sequentially ordered elements and fail if the first and the final elements are not connected. The performance level of each element is controllable, which determines how many elements an element can connect and affects its degradation rate. Accumulated degradation can cause element failure, which may lead to costly system failure. This paper aims to minimize long-term maintenance-related costs, including system failure costs. We provide optimal maintenance planning and performance control for every system degradation state through Markov decision process modeling and a dynamic programming algorithm. Load-sharing, restricted maintenance capacity, maintenance setup costs, and the structural characteristics of the system are considered in the model, all of which influence the optimal maintenance and performance control policy. Regarding degradation management, reducing the difference in degradation levels between elements, e.g., replacing more-degraded elements first, can be cost-effective. However, increasing the difference in degradation by maintenance or performance control can also lower maintenance-related costs in specific situations, which is discussed in numerical experiments. We also illustrate structural insights regarding the proposed model, including sensitivity analyses of maintenance capacity, setup costs, and the difference between preventive and corrective replacement costs.

Best practices to reduce methane emissions at gas transmission networks. A literature review and case studies

This paper studies methane emissions reduction at natural gas transmission systems. A literature survey on the topic is presented along with some characteristic case studies. Such case studies reveal that a significant reduction of methane emissions at gas transmission systems may be achieved with application of best practices, resulting in a large impact on decarbonization and environmental protection efforts. At a characteristic case study, the maintenance plan optimization of the scraper traps along a gas transmission network reduced methane emissions by 90%. Quantification of methane emissions reduction at the presented case studies may serve as quick reference for similar applications. The analysis aims to contribute to better understanding of methane emissions sources and the adoption of emissions reduction measures at gas transmission systems.

Integrated optimization of non-permutation flow shop scheduling and maintenance planning with variable processing speed

Integrated optimization of production scheduling and maintenance planning has received tremendous attention from academia and industry, but the impacts of variable processing speeds on the integrated optimization are either unconsidered or supposed to be constant in the most existing publications. To remedy this shortcoming, this paper addresses an integrated optimization problem of non-permutation flow-shop scheduling and maintenance planning (including preventive maintenance and replacement) with adjustable processing speeds for minimizing the total production cost. To better analyze the problem, there are five problem properties are derived to describe the relationship among processing speed, processing sequence and maintenance planning. Based on this, an improved genetic algorithm with variable processing speed selection (IGAVPSS) is proposed to deal with the concerned problem. In the numerical simulation, the effectiveness and superiority of extended advanced-postpone balancing strategy, proposed internal operators, IGAVPSS algorithm and variable processing speeds are firstly demonstrated through experimental comparisons. After that, a real application is presented to verify the benefits of the proposed IGAVPSS, and the results show that the newly generated scheduling plan based on the proposed IGAVPSS can reduce up to 42.29% of the total production cost compared with the existing company’s scheduling plan.

An approximate dynamic programming approach for the maintenance optimisation of networked critical infrastructures

The continuous performance of networked infrastructures, such as transportation, telecommunications, and power transmission, is critical to the economic development and social well-being of society. To ensure their reliability in continuously meeting prescribed demand, cost-effective maintenance is essential. To this end, the influence of maintenance actions measures on the reliability of an infrastructure need to be considered both during their implementation and after their completion. To address this problem and to consider the uncertainty of future states, we model the multi-period optimisation of an infrastructure’s maintenance planning as a stochastic dynamic programming problem. An approximate dynamic programming approach is developed to deal with the computational complexity and solve the described problem. The proposed approach can help deal with the uncertainty of future infrastructure states by determining the optimal maintenance plan for each possible state in advance.

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.

Multiagent Reinforcement Learning for Project-Level Intervention Planning under Multiple Uncertainties

Reinforcement learning (RL) has recently been adopted by infrastructure asset management (IAM) researchers for adding flexibility regarding uncertainties in preventive actions decision-making. However, this relatively recent line of research has not incorporated other sources of uncertainties, such as hazards apart from deterioration patterns, nor has such research considered managerial aspects of IAM, such as stakeholders’ utilities. This paper aims to provide a holistic framework that draws upon recent developments in IAM systems and microworlds, employs RL model training, and considers deterioration, hazards, and cost fluctuations as the main sources of uncertainties while also adopting managerial aspects into decision-making. Consistent with the existing practice of IAM, this framework brings flexibility in the face of uncertainties to the IAM decision-making process. Multi-agent RL models based on deep Q networks and actor-critic models are constructed and trained for taking intervention actions regarding elements of a real bridge in Indiana through its life cycle. Both models could lead to higher expected utilities and lower costs compared to the optimal maintenance, rehabilitation, and reconstruction (MRR) plans obtained by Monte Carlo simulation and heuristic optimization algorithms. The proposed framework can assist decision-making bodies and managers in the IAM domain with making updateable optimal and more realistic decisions based on the updated state of various complex uncertainties in a negligible amount of time.

Pavement maintenance planning using a risk-based approach and fault tree analysis

ABSTRACT Optimal planning of Maintenance and Rehabilitation (M&R) treatments due to limited resources and funding is vital. A pavement performance function formula is needed to model the M&R optimisation problem. By considering important features, a pavement deterioration function was developed with an R2 of 0.897. The essence of pavement deterioration is uncertain. Hence, disregarding the uncertainty may result in sub-optimal solutions. Therefore, two approaches were employed to optimise M&R treatments in a large-scale pavement network, including deterministic and risk-based. Both approaches aimed to minimise the required budget in the planning period constrained to acceptable pavement condition. A novel and robust metaheuristic algorithm was adjusted to solve the M&R planning problem deterministically. In the risk-based method, appropriate probability distribution functions were fitted to historical data of uncertain parameters. Monte Carlo simulation was performed to generate probability distributions of the required budget and condition of pavement sections. By using a risk-based strategy instead of the deterministic one, the required budget was reduced by 30%. Fault tree analysis was performed on the solutions obtained from the risk-based method to increase the pavement network’s reliability. The results of the mentioned analysis represented that only by spending 2.5% more budget network’s reliability could be significantly increased.

BLOCKCHAIN OF OPTIMAL MULTIPLE CONSTRUCTION PROJECTS PLANNING UNDER PROBABILISTIC ARRIVAL AND STOCHASTIC DURATIONS

With the rapid development of projects, firms are facing challenges in planning and controlling complex multiple construction projects. This research, therefore, aims at developing blockchain of optimal scheduling and sequencing of multiple construction projects under probabilistic arrival and stochastic durations. Each project task was considered as a block. Then, a framework for electronic project recording (EPR) system was developed. The EPRs are records for project tasks that make information available directly and securely to authorized users. In this framework, two optimization models were developed for scheduling and sequencing project blocks. The scheduling model aims to assign project tasks to available resources at minimal total cost and maximal the number of assigned project tasks. On the other hand, the sequencing model seeks to determine the start time of block execution while minimizing delay costs and minimizing the sum of task’s start times. The project arrival date and the task’s execution duration were assumed probabilistic and stochastic (normally distributed), respectively. The developed EPR system was implemented on a real case study of five projects with total of 121 tasks. Further, the system was developed when the task’s execution duration follows the Program Evaluation and Review Technique (PERT) model with four replications. The project costs (idle time and overtime costs) at optimal plan were then compared between the task’s execution duration normally distributed and PERT modelled. The results revealed negligible differences between project costs and slight changes in the sequence of project activities. Consequently, both distributions can be used interchangeably to model the task’s execution duration. Furthermore, the project costs were also compared between four solution replications and were found very close, which indicates the robustness of model solutions to random generation of task’s execution duration at both models. In conclusion, the developed EPR framework including the optimization models provided an effective planning and monitoring of construction projects that can be used to make decisions through project progress and efficient sharing of project resources at minimal idle and overtime costs. Future research considers developing a Blockchain of optimal maintenance planning.

A hybrid deep learning-based approach for rolling bearing fault prognostics

Predictive Maintenance (PdM) has the potential to revolutionize the industry by providing advanced techniques to assess the condition of an industrial system and yield key information that can help optimize maintenance planning and prevent unexpected faults and breakdowns. Nevertheless, PdM is far from being universally applied and it is still the subject of increasing research. Thus, developing new approaches has great relevance to help PdM become a practical reality for the industry. PdM can also bring benefits in terms of sustainability, by reducing human and material resources waste, which is one of the main objectives of Circular Manufacturing initiatives. In this context, rolling bearings are one of the most studied components, as most industrial systems with rotating mechanisms contain bearings, which are prone to a number of faults caused by natural and unnatural wear. In this work, an hybrid Deep Learning (DL) approach is proposed, combining a Convolutional Neural Network (CNN) with a Gated Recurrent Unit (GRU) network to predict Remaining Useful Life (RUL) using rolling bearing vibration data preprocessed with the Short-Time Fourier Transform (STFT). This model was trained and validated using the PRONOSTIA public dataset, which is a popular benchmark for rolling bearing prognostics. The obtained results are satisfactory, providing RUL estimates close to the true values in most test cases, proving the competitiveness of the approach and its potential.

ML2-enabled Condition-based Demand, Production, Inventory, and Maintenance Planning

Production planning and control is pivotal to meeting customer demand and maximizing profit. At the same time, machine breakdowns compromise these goals, which can be tackled with a good maintenance strategy. Here, advances in condition-based maintenance and prognostics and health management allow predicting the health state of production machines through sensor data and prescribing optimal demand, production, inventory, and maintenance plans. Here, machine learning (ML) is promising for accurate health predictions using sensor data and decision-making in complex, highly dynamic production environments. Thus, in this work, two ML algorithms are applied. First, a data-driven regression algorithm predicts the health of a machine. This forecast is forwarded to a reinforcement learning algorithm (i.e. proximal policy optimization, recently made famous by its application within ChatGPT) to optimize demand, production, inventory, and maintenance plans. A computational study shows excellent performances of the ML-based health prediction and planning algorithms, which surpass traditional maintenance strategies.

New approach to the optimal planning of maintenance and repairs of equipment for hydropower facilities

The effectiveness of maintenance and repairs of hydropower equipment largely determines the environmental safety and reliability of the operation of modern hydroelectric power plants. The article proposes a new approach to planning the maintenance and repairs of hydraulic equipment. The task is to determine the sequence of control actions of the system of maintenance and repairs of hydropower equipment, which, at a minimum cost, will ensure the maintenance of a given level of technical condition of the fleet of this equipment at a hydroelectric power plant for a specified period. It is taken into account that the equipment is at different stages of operation. The article considers the formulations of scheduling optimization problems for various industries. A mathematical formulation of the optimization problem for planning a multi-stage maintenance and repair process for hydropower facilities is formulated.

An impulsive and switched system based maintenance plan optimization in building energy retrofitting project

The maintenance time schedule optimization into the investment budget decision for building energy retrofitting and maintenance planning is incorporated. The aggregate performance and population dynamics of the totality of retrofitted items is described by an impulsive switched system model. The items are switched to different working states by maintenance actions with impulsive effects. Considering the retrofitting project with deterioration. The objective of optimal maintenance control is to maximize the overall energy savings and financial payback. The optimal maintenance control problem is showed to be feasible and bounded. The maintenance time schedule and intensities are simultaneously optimized by an improved differential evolution (DE) algorithm. Compared with the fixed maintenance time schedule, the proposed approach is validated by simulation results, specifically, 3.36% of energy saving and 23.09% of internal rate of return (IRR) can be improved.

A stochastic dynamic programming for maintenance planning of an emergency helicopter

Emergency/rescue helicopters are an essential part of the healthcare systems. Every helicopter should be kept in its best possible operational mode to save the life of the people in danger. For achieving this goal, using an optimal maintenance planning is crucial, which involves a periodic decision about repair or replacement of the components in the helicopter. But, in the considered case study, the spare part inventory is limited and cannot be easily replenished. Also, an observation in this case study is that a repaired component may have a shorter probabilistic lifetime (i.e. time to next failure) in contrast to a brand-new one. The best sequence of decisions can be made for each day of a given planning horizon, when different probabilistic trade-offs for every mission hour are taken into account. The problem is formulated in the framework of a stochastic dynamic program (SDP). For obtaining a fast near-optimal policy, an approximate dynamic program (ADP) is developed, which is based on the Monte-Carlo sampling of the possible paths in the corresponding transition graph of the SDP. A numerical experiment shows its gradual convergence toward an optimal policy as it spends more time in the search space of the SDP.

Research on coordinated optimization of unit maintenance plan and economic dispatch in the integrated energy system

The medium and long-term operation of the unit is an important part of the operation of the power energy system. With the development of IES, the coupling of various energy forms such as electricity, heat and natural gas has been continuously strengthened. When a CHP unit or a gas unit is shut down, its impact will involve a variety of energy systems. Therefore, It is of great significance to formulate a reasonable annual unit combination plan and a coordinated optimization plan for the maintenance plan. First of all, starting from the coupling mechanism of the power supply, heating and gas supply of the unit, consider the relationship between each system and equipment. On this basis, taking into account the system operation constraints, unit maintenance constraints and unit operation constraints, a combined dispatch model of IES unit combination and maintenance plan including power-to-gas (P2G) and electric boiler equipment are proposed. The accuracy and effectiveness of the proposed model are verified through simulation analysis of a numerical example. The calculation example results show that the overall arrangement of the unit maintenance plan and the unit combination plan in the integrated energy system can improve the operating economy of the system; the introduction of coupling equipment such as P2G and electric boilers can effectively improve the new energy consumption rate and reduce the system operating cost.

Development of an optimized maintenance scheduling for emergency rescue railway wagons using a genetic algorithm: a case study of Iran railways company

PurposeThis paper aims to optimize the maintenance scheduling of emergency rescue wagons in railway companies using a genetic algorithm (GA). It offers an integrated model for simultaneously solving maintenance planning, preventive maintenance, prognostic information and resource planning from which optimal levels of the system performance in terms of cost and repair time could be determined.Design/methodology/approachThis study initially evaluates the previous types of research in presenting the optimal model of the rescue train wagon for maintenance and repair planning and lists the identified criteria based on experts' opinions using fuzzy analytic hierarchy process (FAHP) and stepwise weight assessment ratio analysis (SWARA) techniques. Then, the final weight of the desired criteria is calculated. Later, the final decision matrix is evaluated by the experts. The final normal decision matrix is formed to select the optimal maintenance and repair planning plan based on the GA. Finally, two strategies including joint optimization strategy of preventive maintenance planning are compared with the independent preventive maintenance planning strategy.FindingsAccording to the primary results, three primary parameters, including technology, human damage caused by negligence and the average failure rate, should be considered for launching the GA model. Based on the results of the second part; by comparing the preventive maintenance planning strategy independently, the joint optimization strategy reduces the total cost of production up to 8.25%. Comparison results show that the total cost of joint optimization production is less than independent preventive maintenance and repair planning. Moreover, the value of total process time in joint optimization strategy was reduced by 1.2% compared to independent preventive maintenance (PM) planning (from 137.90 to 136.20 h).Originality/valueThe novelty of this paper lies on the application of the GAs to develop an optimized PM scheduling to localize the maintenance planning for maximizing productivity, avoid train accidents, reduce costs and increase efficiency and capacity.

Reliability-based optimization of imperfect preventive maintenance with Bayesian estimation

In this study, we present a sequential imperfect preventive maintenance model for a component subject to degradation. We use an age reduction reliability model for scheduling preventive maintenance, which is performed whenever the component’s reliability falls below the reliability threshold level, R. The problem is considered under lack of data, in which the Bayesian methodology is more appropriate than the frequentist view for estimating the unknown parameters of the model. We assume a fixed duration for preventive maintenance, whereas the duration of corrective maintenance is an exponential random variable. We model the total expected cost over all cycles and find the optimal preventive maintenance plan until a replacement based on the reliability threshold value. We also conducted a numerical study for sensitivity analysis of the developed model.

A multi-objective differential evolutionary algorithm for optimal sustainable pavement maintenance plan at the network level

Sustainable highway pavement maintenance is important for achieving sustainability in the transportation sector. Because the three aspects included in sustainability metrics (environment, economy, and society) often contradict each other, maximising the sustainability performance of highway pavements is difficult, especially at the network level. This study developed a novel multi-objective heuristic algorithm to formulate sustainable highway pavement network maintenance plans considering carbon emissions (CE), life cycle agency cost (LCAC), and pavement long-term performance (LTP). The proposed algorithm is a new variant of multi-objective differential evolution (MODE) that incorporates self-adaptive parameter control and hybrid mutation strategies embedded in its framework (MOSHDE). Three state-of-the-art multi-objective heuristics, namely, the non-dominated sorting genetic algorithm II(NSGA-II), classic MODE, and multi-objective particle swarm optimisation (MOPSO), as well as the proposed MOSHDE, were applied to an existing highway pavement network in China for performance evaluation. Compared with other heuristic algorithms, the proposed self-adaptive parameter control strategy enables the automatic adjustment of the control parameters, avoiding the time-consuming process of selecting them and enhancing the robustness and applicability of differential evolution. The hybrid mutation strategy uses both exploration and exploitation operators for the mutation operations, thus leveraging both global and local searches. The results of the numerical experiment demonstrate that MOSHDE outperforms the other tested heuristics in terms of efficiency and quality and diversity of the obtained approximate Pareto set. The optimal solutions obtained by the proposed method correspond to a proactive maintenance policy, as opposed to the reactive maintenance policy commonly adopted in current practice. In addition, these solutions are more cost-effective and environmentally friendly and can provide better pavement performance to highway users over the project life cycle. Therefore, the proposed MOSHDE may help practitioners in the transportation sector make their highway infrastructure more sustainable.

Multi-level optimization of maintenance plan for corroded pipelines considering several modes of failure

Multi-level optimization of maintenance plan for corroded pipelines considering several modes of failure

RBM-MCDM framework for optimization of maintenance and inspection intervals of small unmanned aircrafts

PurposeUnmanned aircraft applications are quickly expanded in different fields. These systems are complex that include several subsystems with different types of technologies. Maintenance and inspection planning is necessary to obtain optimal performance and effectiveness. The failure rate in these systems is more than commercial and manned aircraft since they are usually cheaper. But maintenance and operation planning are difficult because we deal with a system that has multi-components, multi-failure models, and different dependencies between subsystems without any advanced health monitoring system. In this paper, this matter is considered and a framework to determine optimal maintenance and inspection plan for this type of system is proposed to improve system reliability and availability. The new criteria according to this field are proposed.Design/methodology/approachMaintenance of unmanned systems influences their readiness; also, according to the complexity of the system and different types of components, maintenance programming is a vital requirement. The plan should consider several criteria and disciplines; thus, multicriteria decision approaches may be useful. On another side, the reliability and safety of unmanned aircraft are the most important requirements in the design and operation phases. The authors consider these parameters and develop a framework based on risk-based maintenance to overcome the problems for unmanned systems. This framework consists of two stages: at the first stage, the critical components and failure modes are determined by FMEA, and in the second stage, the priority of maintenance tasks is determined by a fuzzy multicriteria weighted decision system. In this study, fourteen criteria with different levels of importance are developed and proposed to find the best plan for maintenance and inspection intervals. These criteria have been extracted from the literature review, the author's experience, and expert opinions.FindingsA novel framework for risk-based maintenance has been proposed. Risk determination and risk criteria are the most important factors in this framework. Risks are determined by FMEA, and new criteria are proposed that are used for decision-making. These criteria are proposed based on practical experience and experts' opinions for the maintenance process in the aeronautic industry. These are evaluated by industrial cases, and this framework capability has been demonstrated.Research limitations/implicationsThe proposed framework and criteria for small unmanned aircraft have been developed based on a practical point of view and expert opinion. Thus for implementation in other aeronautic industries, the framework may need a minor modification.Practical implicationsTwo important subsystems of an unmanned aircraft have been studied, and the capabilities of this method have been presented.Originality/valueThis research is original work to determine a maintenance program for unmanned aircraft that their application has rapidly grown up. Practical and design parameters have been considered in this work.