ArticlePublisher preview available

Manufacturing system maintenance based on dynamic programming model with prognostics information

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

The traditional maintenance strategies may result in maintenance shortage or overage, while deterioration and aging information of manufacturing system combined by single important equipment from prognostics models are often ignored. With the higher demand for operational efficiency and safety in industrial systems, predictive maintenance with prognostics information is developed. Predictive maintenance aims to balance corrective maintenance and preventive maintenance by observing and predicting the health status of the system. It becomes possible to integrate the deterioration and aging information into the predictive maintenance to improve the overall decisions. This paper presents an integrated decision model which considers both predictive maintenance and the resource constraint. First, based on hidden semi-Markov model, the system multi-failure states can be classified, and the transition probabilities among the multi-failure states can be generated. The upper triangular transition probability matrix is used to describe the system deterioration, and the changing of transition probability is used to denote the system aging process. Then, a dynamic programming maintenance model is proposed to obtain the optimal maintenance strategy, and the risks of maintenance actions are analyzed. Finally, a case study is used to demonstrate the implementation and potential applications of the proposed methods.
This content is subject to copyright. Terms and conditions apply.
J Intell Manuf (2019) 30:1155–1173
https://doi.org/10.1007/s10845-017-1314-6
Manufacturing system maintenance based on dynamic
programming model with prognostics information
Qinming Liu2·Ming Dong1·Wenyuan Lv2·Chunming Ye2
Received: 3 June 2016 / Accepted: 17 February 2017 / Published online: 27 February 2017
© Springer Science+Business Media New York 2017
Abstract The traditional maintenance strategies may result
in maintenance shortage or overage, while deterioration and
aging information of manufacturing system combined by sin-
gle important equipment from prognostics models are often
ignored. With the higher demand for operational efficiency
and safety in industrial systems, predictive maintenance with
prognostics information is developed. Predictive mainte-
nance aims to balance corrective maintenance and preventive
maintenance by observing and predicting the health status of
the system. It becomes possible to integrate the deteriora-
tion and aging information into the predictive maintenance
to improve the overall decisions. This paper presents an
integrated decision model which considers both predictive
maintenance and the resource constraint. First, based on
hidden semi-Markov model, the system multi-failure states
can be classified, and the transition probabilities among the
multi-failure states can be generated. The upper triangular
transition probability matrix is used to describe the system
deterioration, and the changing of transition probability is
used to denote the system aging process. Then, a dynamic
programming maintenance model is proposed to obtain the
optimal maintenance strategy, and the risks of maintenance
actions are analyzed. Finally, a case study is used to demon-
strate the implementation and potential applications of the
proposed methods.
BQinming Liu
lqm0531@163.com
1Department of Operations Management, Antai College of
Economics and Management, Shanghai Jiao Tong University,
Shanghai, China
2Department of Industrial Engineering, Business School,
University of Shanghai for Science and Technology,
Shanghai, China
Keywords Maintenance ·Dynamic programming ·
Prognosis ·Deterioration ·Aging
Introduction
Manufacturing system maintenance plays a critical role
in industrial equipment’s efficient usage in terms of cost,
availability and safety. Many effective system maintenance
strategies have been developed (Huynh et al. 2012;Wang
et al. 2016;Lu et al. 2015). Generally, system mainte-
nance can be classified into corrective maintenance (CM) and
preventive maintenance (PM). The corrective maintenance
involves the repair or replacement of failed components
(Kenne and Nkeungoue 2008). The preventive maintenance
is a schedule of maintenance actions aiming at the preven-
tion of system breakdowns and failures (Wang et al. 2015;
Zhong and Jin 2014). Recently, condition-based maintenance
(CBM) becomes more desirable in many application domains
where safety, reliability and availability of the system are con-
sidered critically. It has attracted researchers in recent years
by aiming to balance the maintenance cost, which is high in
PM, with failure cost, which is high in CM. In addition, CBM
can also increase productivity, efficiency and availability of
systems.
For system maintenance, condition monitoring is becom-
ing popular in industries because of its efficient role in
detecting potential failures. The use of condition monitoring
techniques can improve system availability and reduce down-
time. If a hidden defect is already presented, with the help of
condition monitoring, the failure may be identified, and main-
tenance actions may be taken. For an effective maintenance,
advance prediction of such a failure and its development are
very important for ordering spare parts and preparing main-
tenance personnel. Meanwhile, it requires careful plan well
before the failure actually occurs.
123
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
... For a parallel-series production line, the problem of single machine availability captures a growing interest in a variety of areas [11][12][13][14]. Pan et al. [15] proposed a scheduling model for single machine system incorporating production scheduling and machine maintenance, so as to maximize the machine's availability. ...
... e periodic maintenance includes periodic inspections, periodic repairs, and preventive maintenance. In this paper, periodic maintenance aims to obtain the optimal maintenance strategies in one life cycle of a parallel-series production line [11]. e integrated model can be compared with the periodic maintenance model integrated production by considering the influence of maximum available processing time U max and the number of production cycles N. e total cost of the two models can be computed based on the different U max and N. Table 8 shows that the integrated model has a better performance for saving cost based on different maximum available processing time U max , and the diff is proportional to U max . ...
Article
Full-text available
The maintenance of a production line is becoming more important with the development of demanding higher operational efficiency and safety in industrial system. However, a production line often operates under dynamically operational and environmental conditions and the production scheduling is also a very important factor for the maintenance of a production line. First, this paper proposes an integrated data-driven model that coordinates maintenance planning decisions with production scheduling decisions to solve the problem of scheduling and maintenance planning for a parallel-series production line. The degradation information is considered, and the total cost is to be minimized in the proposed model. Also, the total cost is related with production process and maintenance considering reliability of equipment. Then, in order to better describe the relationship between production and maintenance, the accumulative processing time of equipment is used as the input of its failure function. Also, an ability factor is developed to control its reduced level by adopting preventive maintenance. Finally, a case study is used to demonstrate the implementation and potential applications of the proposed model. The long-term wear test experiments are conducted at a research laboratory facility of Shanghai Pangyuan Machinery Co., Ltd. The result proves that the proposed method is feasible and efficient to solve the joint decision-making problem for a parallel-series production line with multivariety and small batch production. The proposed model in this paper is suitable for semiconductor manufacturing.
... Most scholars have turned their attention to the research on equipment maintenance strategies and the joint optimization research that combines various production conditions of equipment with maintenance [1][2][3][4][5]. Maintenance strategies are generally divided into three categories, including predictive maintenance (PM), corrective maintenance (CM), and condition-based maintenance (CBM) [6,7]. ...
Article
Full-text available
Aiming at the maintenance problem of a multi-component batch production system under dynamic demand, we propose a three-phase condition-based maintenance decision-making framework to reduce the number of downtime and maintenance costs. In phase 1, the minimum maintenance cost rate is used to determine each component’s initial preventive maintenance times and intervals. The maintenance cost-time ratio function is defined to evaluate the economics of preventive maintenance. In phase 2, according to the current production plan with batches as the periodic rolling update, the initial preventive maintenance time points are dynamically adjusted based on the advanced-postponed maintenance strategy. In phase 3, an opportunity maintenance model is established to reduce the number of batch production interruptions and the total cost. The numerical results show that the proposed CBM strategy performs effectively, and compared with the strategy that does not consider opportunistic maintenance, the number of downtimes is reduced by 11 times, and the total maintenance cost is reduced by 7%.
... With the rapid development of Internet of Things, information technology and Artificial Intelligence, Predictive maintenance (PdM) has gradually attracted great attention of scholars [1][2][3][4]. As a core of operation management in the industrial system, PdM plays a vital role in reducing the frequency of failures, improving operating efficiency, and ensuring product quality [5,6], and has become a research the prediction accuracy of TBF is of great significance and can bring tangible benefits to the industry. ...
Article
Full-text available
Predictive maintenance (PdM) cannot only avoid economic losses caused by improper maintenance but also maximize the operation reliability of product. It has become the core of operation management. As an important issue in PdM, the time between failures (TBF) prediction can realize early detection and maintenance of products. The reliability information is the main basis for TBF prediction. Therefore, the main purpose of this paper is to establish an intelligent TBF prediction model for complex mechanical products. The reliability information conversion method is used to solve the problems of reliability information collection difficulty, high collection cost and small data samples in the process of TBF prediction based on reliability information for complex mechanical products. The product reliability information is fully mined and enriched to obtain more reliable and accurate TBF prediction results. Firstly, the Fisher algorithm is employed to convert the reliability information to expand the sample, and the compatibility test is also discussed. Secondly, BP neural network is used to realize the final prediction of TBF, and PSO algorithm is used to optimize the initial weight and threshold of BP neural network to avoid falling into local extreme value and improve the convergence speed. Thirdly, the mean-absolute-percentage-error and the Coefficient of determination are selected to evaluate the performance of the proposed model and method. Finally, a case study of TBF prediction for a remanufactured CNC milling machine tool (XK6032-01) is studied in this paper, and the results show that the feasibility and superiority of the proposed TBF prediction method.
... Production downtime is one of the most significant contributors to production inefficiency (Liu et al., 2012), resulting in lost profit. While planned production downtime occurs, for example, for scheduled maintenance based on regular schedules or predictive models (Boudjelida, 2019;Khatab, 2018;Liu et al., 2019;Sipos et al., 2014), unforeseen production stops are a result of failures in the production process, e.g., misconfiguration of a machine, intervention of a worker, or defective raw material. In the case of an unforeseen production stop, direct action from production workers is required to resolve the issue promptly and limit the financial loss (Mobley, 2002). ...
Article
Full-text available
In discrete manufacturing, the knowledge about causal relationships makes it possible to avoid unforeseen production downtimes by identifying their root causes. Learning causal structures from real-world settings remains challenging due to high-dimensional data, a mix of discrete and continuous variables, and requirements for preprocessing log data under the causal perspective. In our work, we address these challenges proposing a process for causal reasoning based on raw machine log data from production monitoring. Within this process, we define a set of transformation rules to extract independent and identically distributed observations. Further, we incorporate a variable selection step to handle high-dimensionality and a discretization step to include continuous variables. We enrich a commonly used causal structure learning algorithm with domain-related orientation rules, which provides a basis for causal reasoning. We demonstrate the process on a real-world dataset from a globally operating precision mechanical engineering company. The dataset contains over 40 million log data entries from production monitoring of a single machine. In this context, we determine the causal structures embedded in operational processes. Further, we examine causal effects to support machine operators in avoiding unforeseen production stops, i.e., by detaining machine operators from drawing false conclusions on impacting factors of unforeseen production stops based on experience.
Conference Paper
Full-text available
This paper aims to identify solution-oriented use cases of artificial intelligence (AI) across manufacturing, wholesale, transportation, warehousing, and retail industries in supply chain risk management (SCRM) that may be considered for prototyping or implementation. A systematic literature review of peer-reviewed papers published in the last six years across ProQuest databases was conducted to identify the AI use cases in SCRM. The articles have been categorized based on the type of risk in SCRM, the type of research work conducted by past researchers, and finally shortlisted based on the practical implications of the research. Papers categorized as solution-oriented use cases of AI in SCRM have been selected, covering the areas of demand risk, logistic risk, manufacturing risk, supply risk, and financial risk. This paper will enable supply chain executives and risk professionals to dive deep into the selected use cases and initiate prototyping and implementation of AI in their organisations.
Article
The process industry implements important maintenance strategies on the basis of the health status of system operation to reduce engineering maintenance cost and concurrently ensure reliability and safety. This study presents an architecture for knowledge-based prognostics and health management (K-PHM) by applying the trapezoidal interval type-2 fuzzy linguistic term sets. The proposed K-PHM methodology successfully provides operational information for diagnostics, prognostics, and subsequent actionable knowledge for health management in subjective decision making. It emphasizes knowledge-driven maintenance strategies to increase system reliability and safety within linguistic decision making. The main objective of this study is to focus on the uncertainty within the knowledge formation derived from data and subsequent information expressed with linguistic terms generated from domain experts in the K-PHM methodology. Fuzzy sets and fuzzy logic reasoning techniques are used in conjunction to enhance the capability of handling uncertainty throughout the PHM process. The effectiveness of the proposed K-PHM is verified with the authentic case study for the predictive maintenance of a chemical pumping system against abnormal vibration.
Thesis
Full-text available
Prognosis and Health Management (PHM) is a very useful tool for improving the overall safety, reliability and cost of a system. Fault diagnosis and prognosis are two pillars of a successful PHM. Both have their own challenges and are mostly studied independent of each other. Therefore, performing both fault diagnosis and prognosis using a common metric can greatly ease the overall implementation of PHM process and hence make the system more safe and reliable. This work proposes an innovative integrated framework for performing fault diagnosis and prognosis using a common metric i.e. Energy Activity (EA). The underlying framework is in the first step validated using simulated spring-mass-damper system. The proposed framework is further improved to addresses a major challenge to PHM i.e. robustness of decision considering both system parameter uncertainties and measurement uncertainties. The improved framework is successfully applied to a real hydraulic two-tank system test bed system under uncertain conditions.
Article
In view of the maintenance of multiple power equipment operating in abnormal conditions in large-scale power plants, based on the prediction of fault state, a predictive maintenance scheduling method is proposed to schedule the maintenance activities. Firstly, based on the actual operating data, combined with the influencing factors of fault state deterioration by Pareto analysis, a time-variant function is improved to predict the deterioration state of potential fault in future maintenance interval. Then, maintenance priority is divided based on the fault state, considering the constraints of maintenance resources and equipment downtime, with the objective of minimizing the total maintenance cost, a scheduling model is built for the maintenance of multiple equipment. Finally, aiming at the continuity of maintenance time, a two-stage algorithm is proposed, which divides the maintenance time-window to transform the complex continuous time optimization problem into a combinatorial optimization problem of time periods, and then develops the optimal maintenance scheme. Taking the maintenance of multiple power transformers as an example, combined with the data resources provided by Yunnan power grid of China, the effectiveness of the improved prediction function of fault state is proved. In addition, by comparing with the traditional maintenance strategy based on the principle of first-fault-first-repair, the superiority of the proposed maintenance scheduling method is verified in reducing cost and improving system stability.
Article
Full-text available
Nowadays CPSs have drawn an upsurge of interests for their enormous potential towards the next generation smart systems where safety is a critical issue. In CPS, large number of IoT devices (sensors, actuators, etc.) are deployed to collect data to support safety critical smart-world CPS infrastructures such as smart city, smart health, smart manufacturing, and smart transportation. The degradation of physical components of safety critical smart-world CPSs would deteriorate the performance of the smart system and lead to loss of human life with significant damage to properties. Hence, planned preventive maintenance replacement of physical components of the system is vital to extend the lifetime of the system, to reduce maintenance cost and to avoid risks that may cause major harm to life and property. In this study, we focus on cost effective preventive replacement strategy that recovers failure of physical components of safety critical smart-world CPSs by mainly considering the deterioration state of the physical component on the availability of the system. In view of the degraded physical component of CPSs, we demonstrate the effects of involving maintenance actions during the deterioration state of physical components of CPSs on its availability. Since timely preventive replacement is crucial to support continuous and effective system operation, we compute the optimum time of the constant-interval of preventive replacement strategy for the physical components of CPSs. In this study, we use Weibull distribution as it is a generalized failure model that has been widely used for process equipment life data analyses. At last, the effects of the shape and the scale parameters on the optimal preventive replacement interval of the physical component are also demonstrated.
Article
Full-text available
A way to reduce the cost of regression testing consists of selecting or prioritizing subsets of test cases from a test suite according to some criteria. Besides greedy algorithms, cost cognizant additional greedy algorithms, multi-objective optimization algorithms, and Multi-Objective Genetic Algorithms (MOGAs), have also been proposed to tackle this problem. However, previous studies have shown that there is no clear winner between greedy and MOGAs, and that their combination does not necessarily produce better results. In this paper we show that the optimality of MOGAs can be significantly improved by diversifying the solutions (sub-sets of the test suite) generated during the search process. Specifically, we introduce a new MOGA, coined as DIV-GA (DIversity based Genetic Algorithm), based on the mechanisms of orthogonal design and orthogonal evolution that increase diversity by injecting new orthogonal individuals during the search process. Results of an empirical study conducted on eleven programs show that DIV-GA outperforms both greedy algorithms and the traditional MOGAs from the optimality point of view. Moreover, the solutions (sub-sets of the test suite) provided by DIV-GA are able to detect more faults than the other algorithms, while keeping the same test execution cost.
Article
Full-text available
This paper investigates the maintenance problem for a flow line system consisting of two series machines with an intermediate finite buffer in between. Both machines independently deteriorate as they operate, resulting in multiple yield levels. Resource constrained imperfect preventive maintenance actions may bring the machine back to a better state. The problem is modeled as a semi-Markov decision process. A distributed multi-agent reinforcement learning algorithm is proposed to solve the problem and to obtain the control-limit maintenance policy for each machine associated with the observed state represented by yield level and buffer level. An asynchronous updating rule is used in the learning process since the state transitions of both machines are not synchronous. Experimental study is conducted to evaluate the efficiency of the proposed algorithm.
Article
In this paper, we present a joint optimization of economic manufacturing quantity (EMQ) and maintenance policy for a production facility subject to deterioration and condition monitoring (CM) at the times the production runs are completed. The production facility deterioration is described by a hidden continuous-time Markov process. CM provides partial information about the hidden state of the production facility. The objective is to develop a jointly optimal lot sizing and maintenance policy using multivariate Bayesian control approach. The posterior probability statistic is updated at each sampling epoch using Bayes’ rule. When the posterior probability crosses a control limit, the production system is stopped and full inspection is initiated, followed possibly by preventive maintenance (PM). Production will resume when all available inventory is depleted or when PM action is completed, whichever occurs later. We also assume that the production and demand rates are constant over time. The problem is formulated and solved in the semi-Markov decision process (SMDP) framework. The objective is to minimize the long-run expected average cost per unit time. The shortage and set-up costs are considered in the model along with the maintenance, inventory holding, and lost production costs. A numerical example is provided and a sensitivity analysis is performed. A comparison with the age-based maintenance policy shows an outstanding performance of the new model and the control policy proposed in this paper.
Article
The maintenance strategies optimization can play a key role in the industrial systems, in particular to reduce the related risks and the maintenance costs, improve the availability, and the reliability. Spare part demands are usually generated by the need of maintenance. It is often dependent on the maintenance strategies, and a better practice is to deal with these problems simultaneously. This article presents a stochastic dynamic programming maintenance model considering multi-failure states and spare part inventory. First, a probabilistic maintenance model called hidden semi-Markov model with aging factor is used to classify the multi-failure states and obtain transition probabilities among multi-failure states. Then, spare parts inventory cost is integrated into the maintenance model for different failure states. Finally, a double-layer dynamic programming maintenance model is proposed to obtain the optimal spare parts inventory and the optimal maintenance strategy through which the minimum total cost can be achieved. A case study is used to demonstrate the implementation and potential applications of the proposed methods.
Article
Considering the objectives of "condition based preventative maintenance strategy", condition monitoring and fault diagnosis of an MR M-type load tap changer (LTC) was assessed by applying several oil analysis techniques, followed by interpretation of the derived data. The interpretation revealed the oil low cleanliness level as well as in progress thermal fault with light coking or increased resistance of the contacts or severe arcing. As a result, overhauling and visual inspection were recommended in spite of the suggested overhaul criteria by the manufacturer, which advised no overall repair for the pilot LTC at that time. Finally by overhauling the pilot LTC, the observations from physical inspection were compared to the extracted results from oil analyses data interpretation, which were in close agreement with each other.
Article
In this paper, we consider the joint optimization of economic manufacturing quantity (EMQ) and preventive maintenance (PM) policy for a production facility subject to deterioration and condition monitoring (CM). Unlike the previous joint models of EMQ and maintenance policy which used traditional maintenance approaches, we propose the proportional hazards model (PHM) to consider CM information as well as the age of the production facility. The deterioration process is determined by the age and covariate values and the covariate process is modeled as a continuous-time Markov process. The condition information is available at each inspection epoch, which is the end of each production run. The hazard rate is estimated after obtaining the new information through CM. The problem is formulated and solved in the semi-Markov decision process (SMDP) framework. The objective is to minimize the long-run expected average cost per unit time. Also, the mean residual life (MRL) of the production facility is calculated as an important statistic for practical applications. A numerical example is provided and a comparison with the age-based policy shows an outstanding performance of the new model and the control policy proposed in this paper.
Article
Health prognosis for equipment is considered as a key process of the condition-based maintenance strategy. This paper presents an integrated framework for multi-sensor equipment diagnosis and prognosis based on adaptive hidden semi-Markov model (AHSMM). Unlike hidden semi-Markov model (HSMM), the basic algorithms in an AHSMM are first modified in order for decreasing computation and space complexity. Then, the maximum likelihood linear regression transformations method is used to train the output and duration distributions to re-estimate all unknown parameters. The AHSMM is used to identify the hidden degradation state and obtain the transition probabilities among health states and durations. Finally, through the proposed hazard rate equations, one can predict the useful remaining life of equipment with multi-sensor information. Our main results are verified in real world applications: monitoring hydraulic pumps from Caterpillar Inc. The results show that the proposed methods are more effective for multi-sensor monitoring equipment health prognosis.
Article
A novel optimal preventive maintenance policy for a cold standby system consisting of two components and a repairman is described herein. The repairman is to be responsible for repairing either failed component and maintaining the working components under certain guidelines. To model the operational process of the system, some reasonable assumptions are made and all times involved in the assumptions are considered to be arbitrary and independent. Under these assumptions, all system states and transition probabilities between them are analyzed based on a semi-Markov theory and a regenerative point technique. Markov renewal equations are constructed with the convolution of the cumulative distribution function of system time in each state and corresponding transition probability. By using the Laplace transform to solve these equations, the mean time from the initial state to system failure is derived. The optimal preventive maintenance policy that will provide the optimal preventive maintenance cycle is identified by maximizing the mean time from the initial state to system failure, and is determined in the form of a theorem. Finally, a numerical example and simulation experiments are shown which validated the effectiveness of the policy. (c) 2013 Elsevier B.V. All rights reserved.
Article
This paper addresses the problem of finding a robust and stable schedule for a single machine with availability constraints. The machine suffers unexpected breakdowns and follows the Weibull failure function. A joint model for integrating run-based preventive maintenance (PM) into the production scheduling problem is proposed, in which the sequence of jobs, the PM times and the planned completion times of jobs are proactively determined simultaneously. Aiming at optimizing the bi-objective of system robustness and stability, a genetic algorithm based on the properties of the optimal schedule is proposed. The experimental results demonstrate that the proposed algorithm is efficient and effective under practical problem sizes. In addition, the impact of degree of uncertainty on the performance and the tradeoff between robustness and stability are explored in detail.
Article
Mechanical devices are playing a crucial role in modern industry. With the ever-growing demands of multiple function and high performance, the unpredicted failures of mechanical device might greatly increase maintenance cost during its lifetime. As a key state indicator of mechanical device, the degradation of some important performance provides substantial information for failure prognosis. More and more attention has been paid to the degradation-based failure time prediction. However, even mechanical devices of the same type might show greatly diverse degradation processes under different working environments. It is still a challenge to identify global degradation pattern and then predict the failure time of a specific mechanical device based on its degradation sequence. This paper proposes a novel approach for failure time prediction with the degradation sequence of mechanical device. The proposed approach combines the exponential regression and parametric empirical Bayesian (PEB) technology. Firstly, exponential regression is adopted to represent the local degradation pattern and then local failure time observations can be computed. Secondly, according to the rule that local failure time observations manifest, appropriate prior assumption is made and the posterior distribution is estimated by PEB technology. Herein, two prior assumptions are considered, including the exchangeable PEB and linear PEB case. The global failure time distribution can be predicted with the estimated prior and posterior distribution. Finally, three case studies are implemented to validate the proposed approach, including the simulation case, crack case and precision case of machine tool.