Research paperDevelopment and application of maintenance decision-making support system for aircraft fleet
Introduction
Structure fatigue is a critical problem for aircraft arising from their nature as multiple-component structures, subjected to random dynamic loads during flight [1], [2], [3], [4]. The structure is replaced or repaired when the reliability of structure reaches the threshold value given by designer. The reliability analysis methods of aircraft structure are known as stress-life, damage-tolerance, etc. [5]. Currently, the stress-life method is used by the U.S. Navy to account the fatigue damage in aircraft [1], which directly relates service loads to a safe operating life according to the linear cumulative damage [2], [3]. The time of structural failure is determined from full-scale fatigue tests, in which the expected service loads are simulated and applied to a specimen in a laboratory environment. To ensure that the structure do not exceed the fatigue life limits during their service, based on the laboratory defined S-N curve and loading cycles in the service spectrum, the fatigue cumulative damage (FCD) is calculated and linearly summed to determine when the structure fails. When FCD reaches 1, either the structure is replaced, or its life is extended by repair. Stress-life method is widely used to determine aircraft and other mechanical structure fatigue life because of its simplicity and practicality [6], [7], [8]. However, the exact reliability value of the structure and its components cannot be determined by stress-life method [6], [9]. To overcome this weakness, as well as considering that the factors which affect the structure fatigue life exhibit considerable scatter [10], [11], [12], [13], [14], the probability-damage-tolerance (PDT) methods are widely studied in recent years, due to its capability in both of handling high reliability problems and taking the stochastic factors into full consideration. Despite their advantages over deterministic analysis methods (e.g., stress-life method), design organizations have been reluctant to adopt even the standard PDT methods or to include them as part of their risk analysis capability. Reason cited include: the complexity of failure modes; lack of available damage data; and safety issues [15].
Lots of studies relating to aircraft reliability have been discussed, the importance of structural and non-structural reliability of aircraft has been extensively studied as well [16], [17], [18], [19], [20]. It is difficult to establish a comprehensive reliability model which can take the whole aircraft reliability into consideration, since aircraft is a complex system involving mechanical, electrical and hydraulic sub-systems. A practical way is to establish different reliability models for different sub-systems. Each of these sub-systems contains components of the same type (e.g., structural components and non-structural components). Due to the research concerning maintenance of aircraft structure subjected to fatigue loads is insufficient. In addition, more and more advanced fatigue monitoring sensors are applied to collect the loading information of the key structures and the collected information are further used to determine the failure times of structures. Thus, this work discussed the reliability of structures subjected to fatigue loads. With improvement to technology and the increase of the Army's information infrastructure, many aircraft have been outfitted with Health and Usage Monitoring Systems (HUMS). Fatigue monitoring of airframes has developed over decades to the stage where it is now incumbent for aircraft to be fitted with an airborne fatigue monitoring systems [21], [22], [23]. With the arrival of these on-board monitoring systems, maintainers have access to a near real-time assessment of component health [24], [25], [26], [27], [28]. Gao et al. [29] proposed a new deep quantum inspired neural network to fault diagnosis for aircraft fuel system, which aimed not only to improve flight security, but also to reduce the huge cost due to regular maintenance. Medina-Oliva et al. [30] presented a knowledge structuring scheme of fleets in the marine domain based on ontologies for diagnostic purposes. Al-Dahidi et al. [31] proposed a remaining useful life estimation method based on condition monitoring data for large fleets of similar equipment. Xia et al. [32] proposed a condition-based maintenance policy for intelligent monitored multi-unit series systems with independent machine failure modes. Li et al. [33] used a general expression of stochastic process models depending on both the system age and the system state to describe the degradation processes of systems. In this work, we discussed a way to establish a new reliability evaluation method, in which the linear cumulative damage of stress-life method and crack propagation life model of PDT method are combined to make full use of the real-time load information collected by HUMS.
HUMS utilizes condition-based maintenance (CBM) concepts to minimize unscheduled failures and maintenance costs. CBM is a maintenance scheme that recommends maintenance decisions based on the information collected by condition monitors [34]. It mainly includes three steps: data acquisition, data processing and maintenance decision-making [35], [36], [37]. CBM attempts to avoid unnecessary maintenance tasks by taking maintenance actions only when there is evidence of abnormal behaviors of a physical asset or a system (e.g., structures) reaches its failure threshold level. A CBM program, if properly established and effectively implemented, can significantly reduce maintenance cost by reducing the number of unnecessary scheduled preventive maintenance operations or making full use of the system's design life (i.e., maintenance operations are applied only when the system reaches its failure threshold level) [38]. In a CBM framework, maintenance policies will be optimized to minimize the operational costs or maximize the availability of systems. As most existing researches on CBM for aircraft solely focused on reducing maintenance cost or maximizing the availability of single aircraft [39], [40], [41], [42], and there are few researches aimed both to minimize cost and to maximize the availability from the perspective of a fleet [43], [44], [45]. Feng et al. [46] developed a two-stage dynamic decision-making model aimed to minimize the maintenance cost for an aircraft fleet. Wijk [47] established a cost-effective optimization model for a phase-out scenario of an aircraft fleet. Al-Thani et al. [48] proposed an exact mixed integer programming model focused on maximizing availability for a specific homogeneous fleet type. In this work, a multi-objective decision-making model based on CBM (MODM-CBM) which concentrates on both reducing the maintenance cost and maximizing the availability of a fleet is established for an aircraft fleet. Sun et al. [49] proposed a ordering decision-making methods on spare parts for a new aircraft fleet based on a two-sample prediction.
The engineering application of maintenance decision for aircraft is being widely researched. The United States has the world's most advanced Integrated Maintenance Information System (IMIS) which mainly includes three parts: a Portable Maintenance Aids (PMA), a base-level Maintenance Information Workstations (MIWs) and a theater-level Maintenance Information Processing Center (MIPC) [50]. The employment of IMIS helps both to ensure the reliability of an aircraft fleet and to reduce the maintenance costs. In the US Air Force, a Core Automation Maintenance System (CAMS) [51] and a Reliability and Maintainability Information System (REMIS) [52] are developed to improve the efficiency of aviation equipment management, moreover, the two systems can help collect and analyze the information obtained during the maintenance process of aviation equipment. Additionally, the Pakistan Air Force developed a Logistics Management Information System (LMIS) to manage the real-time information of aviation equipment, LMIS can track the status of each aircraft in fleet and enhance the ability of maintenance support.
In light of the above studies and software development experience obtained from existing researches [53], [54], [55], this work developed a maintenance decision-making support system (MDMSS) for an aircraft fleet. MDMSS integrates the process of data acquisition (real-time status information of aircraft), data processing (reliability assessment of aircraft structures) and maintenance decision-making, moreover, it simplifies the complex process of equipment management.
The innovation points and improvements of this paper are listed as follows:
- •
A new reliability evaluation method based on the real-time load information is proposed for aircraft structural reliability, to make full use of real-time status information collected by the monitoring systems, as well as to handling the high reliability problems of aircraft structures. The new reliability evaluation method not only combines the practicality of the stress-life methods and the ability of handling high reliability problems of PDT methods, but also avoids the disadvantages of traditional stress-life methods and PDT methods (the real-time status information cannot be applied to help determine the reliability of aircraft structures).
- •
A multi-objective decision-making model based on CBM (MODM-CBM) is established for an aircraft fleet. MODM-CBM concentrates on both reducing the maintenance cost and maximizing the availability of a fleet. Thus, it overcomes the shortcomings of the traditional CBM research which focuses on single aircraft state prognostic and maintenance time.
- •
A maintenance decision-making support system (MDMSS) is developed. MDMSS integrates the process of data acquisition (real-time status information of aircraft), data processing (reliability assessment of aircraft structures) and maintenance decision-making, moreover, it simplifies the complex process of equipment management as well as takes a step further in engineering application.
The remainder of this paper is organized as follows. Section 2 presents the real-time status information processing. Section 3 introduces the proposed reliability evaluation method based on the real-time load information. Section 4 introduces the proposed multi-objective decision-making model based on CBM. Section 5 illustrates the development of maintenance decision-making support system, MDMSS. Section 6 introduces the system application of MDMSS. Section 7 concludes this paper.
Section snippets
Problem statement
Fig. 1 shows that two sections of the wing box which is a typical structure of aircraft are fitted with fatigue monitoring systems, as subjected to the random dynamic loads during flight. The specific locations of strain gauges (sensors) for the two sections are shown in Fig. 2.
As the real-time information collected by these strain gauges are micro strain data, a conversion process is required to convert the micro strain data into load information. It will cause calculation error when using the
Reliability evaluation method based on the real-time load information
With the development of technology, aircrafts are fitted with an airborne fatigue monitoring systems. These systems typically collect fatigue data is simply used for the calculation of the stress-life method or the inspection interval of the airframe, nevertheless, the exact reliability value of the airframe cannot be determined according to the collected information. The PDT methods cannot make full use of the collected real-time load information for reliability assessment, in spite of its
Multi-objective decision-making model based on CBM
More and more advanced fatigue monitoring sensors are applied to collect the loading information of the key structures and the collected information are used to determine the failure times of structures. Additionally, aircraft maintenance program develops from tradition corrective maintenance and preventive maintenance to CBM. Thus, a multi-objective decision-making model based on CBM for an aircraft fleet is established.
Development of maintenance decision-making support system, MDMSS
The maintenance decision-making support system, MDMSS, is developed for maintenance decision-making of an aircraft fleet. As shown in Fig. 6, MDMSS covers the process of data acquisition (real-time status information of aircraft), data processing (reliability assessment of aircraft structures) and maintenance decision-making, as well as databases. The framework of MDMSS is a typical 3-layer system based on Java Swing, and the general structure is shown in Fig. 13.
MDMSS is developed by
Outline of MDMSS
The MDMSS covers the whole process from collection of real-time status information to maintenance decision-making of a fleet. As shown in Fig. 14, the main window of MDMSS is shown when user login to the system, and on the top of the window, user can choose the system's main categories. The basic information of fleets stored in database are shown as a tree (located at the right side of the window). If user clicks the tree node, then the basic information of the fleet that is represented by the
Conclusion
Due to the traditional reliability assessment methods lack the ability to make the full use of real-time status information collected by the monitoring systems, a new reliability evaluation method based on the real-time load information is proposed for aircraft structural reliability. Furthermore, a multi-objective decision-making model based on CBM (MODM-CBM) is established. MODM-CBM aims not only minimizing the maintenance cost, but also maximizing the availability of fleet. Thus, it avoids
Acknowledgments
This work was supported by the Key National Natural Science Foundation of China (grant numbers U1533202); the Shandong Independent Innovation and Achievements Transformation Fund (grant numbers 2014CGZH1101), the Civil Aviation Administration of China (grant numbers MHRD20150104), and National Science-Technology Support Plan Project “the application paradigm of full lifecycle information closed-loop management for construction machinery” (grant numbers 2015BAF32B01-4). We confirm that there is
References (71)
- et al.
A probability method for prediction on high cycle fatigue of blades caused by aerodynamic loads
Adv Eng Software
(2011) - et al.
Fault detection and explanation through big data analysis on sensor streams
Expert Syst Appl
(2017) - et al.
Systems & control for the future of humanity, research agenda: current and future roles, impact and grand challenges
Ann Rev Control
(2017) - et al.
Sequential pattern mining applied to aeroengine condition monitoring with uncertain health data
Eng Appl Artif Intell
(2015) - et al.
Graph-based reasoning in collaborative knowledge management for industrial maintenance
Comput Ind
(2013) - et al.
Chapter 9 - structural health monitoring and influence on current maintenance
Reliability based aircraft maintenance optimization and applications
(2017) - et al.
DBN based failure prognosis method considering the response of protective layers for the complex industrial systems
Eng Fail Anal
(2017) - et al.
Assessment of diagnostic and prognostic condition indices for efficient and robust maintenance decision-making of systems subject to stress corrosion cracking
Reliab Eng Syst Saf
(2017) - et al.
Remaining useful life estimation for mechanical systems based on similarity of phase space trajectory
Expert Syst Appl
(2015) - et al.
Prognostics uncertainty reduction by fusing on-line monitoring data based on a state-space-based degradation model
Mech Syst Sig Process
(2014)
Deep quantum inspired neural network with application to aircraft fuel system fault diagnosis
Neurocomputing
Predictive diagnosis based on a fleet-wide ontology approach
Knowl-Based Syst
Remaining useful life estimation in heterogeneous fleets working under variable operating conditions
Reliab Eng Syst Saf
A proactive condition-based maintenance strategy with both perfect and imperfect maintenance actions
Reliab Eng Syst Saf
Production-driven opportunistic maintenance for batch production based on MAM-APB scheduling
Eur J Oper Res
An approach to operational aircraft maintenance planning
Decis Support Syst
A condition-based maintenance strategy for heterogeneous populations
Comput Ind Eng
Prognostic modelling options for remaining useful life estimation by industry
Mech Syst Sig Process
Optimal replacement policies for two-unit machines with running costs: II
Stochastic Processes Appl
Optimization of multipurpose process plant operations: a multi-time-scale maintenance and production scheduling approach
Comput Chem Eng
Heuristic hybrid game approach for fleet condition-based maintenance planning
Reliab Eng Syst Saf
Phase-out maintenance optimization for an aircraft fleet
Int J Prod Econ
A model and optimization-based heuristic for the operational aircraft maintenance routing problem
Transp Res Part C
Chapter 8–modernization of reliability and maintainability information system (REMIS) into the global combat support system-air force (GCSS-AF) framework
Information systems transformation
jMetal: a Java framework for multi-objective optimization
Adv Eng Software
A decision support system for concrete bridge deck maintenance
Adv Eng Software
Software architecture knowledge for intelligent light maintenance
Adv Eng Software
ANN modeling of pull-off adhesion of concrete layers
Adv Eng Software
Determination of the length of hydraulic jumps using artificial neural networks
Adv Eng Software
Predicting component reliability and level of degradation with complex-valued neural networks
Reliab Eng Syst Saf
A two-parameter model for crack growth simulation by combined FEM–DBEM approach
Adv Eng Software
A simple second order approximation for stochastic crack growth analysis
Probab Eng Mech
Statistical aspects of some fatigue crack growth data
Eng Fract Mech
Probabilistic models of fatigue crack propagation and their experimental verification
Probab Eng Mech
On the prediction of the residual fatigue life of cracked structures repaired by the stop-hole method
Int J Fatigue
Cited by (36)
Knowledge transfer for adaptive maintenance policy optimization in engineering fleets based on meta-reinforcement learning
2024, Reliability Engineering and System SafetyTackling uncertainties in aircraft maintenance routing: A review of emerging technologies
2022, Transportation Research Part E: Logistics and Transportation ReviewCitation Excerpt :El Afia and Sarhani (2017) established maintenance prediction ML models that estimate and schedule maintenance activities based on the remaining useful life of aircraft. Lin et al. (2017) used big data collected from real-time aircraft component health status to establish a multi-objective condition-based maintenance decision and evaluation model. Baptista et al. (2018) employed an auto-regressive moving average (ARMA) model and data-driven techniques to forecast the fault events for predictive maintenance planning.
A blockchain-based system to enhance aircraft parts traceability and trackability for inventory management
2021, Expert Systems with ApplicationsReinforcement learning-driven maintenance strategy: A novel solution for long-term aircraft maintenance decision optimization
2021, Computers and Industrial EngineeringCitation Excerpt :In addition, they also pointed out that the incorporation of dynamic (changing) criteria and insertions of the preferences of the decision-maker require further investigations. Specifically talking about the aircraft maintenance research, in (Lin et al., 2017), a comprehensive decision supporting approach is proposed for both minimizing the maintenance cost and maximizing the availability of an aircraft fleet. In (Sheng & Prescott, 2019), an overall aircraft maintenance scenario is modeled using a colored Petri net framework, and the general maintenance performance index like mission capable/abort rate can be assessed.
A decision support framework and prototype for aircraft dispatch assessment
2020, Decision Support SystemsVoice-driven fleet management system for agricultural operations
2019, Information Processing in AgricultureCitation Excerpt :It is evident that non-automatically generated information has a significant impact in time and user’s concentration on other parallel tasks in the field. In industry, Fleet Management Systems are widely used for a number of years [19–22]. Presently, such systems have advanced into providing enterprises with integrated management tools that connect all different parts of the business.