- Email: [email protected]
Implementation of Constraints Satisfaction Problems Methods for Solving the Periodic Maintenance Processes Scheduling
7th IFAC Conference on Manufacturing Modelling, Management, and Control International Federation of Automatic Control June 19-21, 2013. Saint Petersburg, Russia
Implementation of Constraints Satisfaction Problems Methods for Solving the Periodic Maintenance Processes Scheduling A.Fedotova*, V.Taratoukhine**, M.Ovsyannikov***, J.Becker**** * Department of computer automation, The Bauman Moscow State Technical University and European Research Centre in Information Systems (ERCIS) Lab Russia at Higher School of Economics, Russia (Tel: +7(926)5965359; email:[email protected]). ** University of Muenster, European Research Centre for Information Systems (ERCIS) HQ and ERCIS Lab Russia at Higher School of Economics, Germany (e-mail: [email protected]). ***Department of computer automation, The Bauman Moscow State Technical, Russia. **** University of Muenster, European Research Centre for Information Systems (ERCIS), Germany. Abstract: This paper reviews existing methods and techniques addressing the problem of maintenance support throughout the life cycle for high value manufacturing products. As part of this doctorate research the analysis of current methods of maintenance planning was conducted. In order to contribute to a more comprehensive solution, an advanced approach (algorithm) of periodic maintenance is presented. The authors believe that this approach will reduce the cost of maintenance of high value manufacturing products. The algorithm based on constraint programming methods is briefly presented and the future research directions are discussed. Keywords: Information systems, maintenance and repair of equipment, maintenance cost, periodic maintenance scheduling methods, methods of constraints satisfaction problems, domain ontology. are typically technology intensive, expensive and reliability critical requiring engineering services, such as maintenance support throughout the life cycle (Becker et al. 2007).
1. INTRODUCTION The competitiveness of the aircraft is influenced by a number of factors, including the operating system maintenance and repair of equipment. Maintenance and repair of equipment is an operation or a complex of operations intended to keep an item in the proper condition when used for intended purpose, on standby, for storage and transportation (GOST 18322-78 1978). British Standard Glossary of terms (3811:1993) defined maintenance as the combination of all technical and administrative actions, including supervision actions, intended to retain an item in, or restore it to, a state in which it can perform a required function. European Federation of National Maintenance Societies may be defined maintenance, repair, and operations as, "all actions which have the objective of retaining or restoring an item in or to a state in which it can perform its required function. The actions include the combination of all technical and corresponding administrative, managerial, and supervision actions."
From a purely industrial point of view, this means higher reliabilit\ RI D FRPSDQ\¶V WHFKQRORJLFDO FDSDFLWLHV DQG WKHLU constant availability at the peak of their features. Thus the importance of maintenance effectiveness considerably LQFUHDVHV DV ZHOO DV WKH ZKROH VHW RI D FRPSDQ\¶V DVVHW management processes (Becker et al. 2006, Scheer 1994). High value manufacturers are looking for additional cost savings through performance-based service contracts and by taking responsibilities for their product performance throughlife. The ability of owners and service providers to deploy best-in-class maintenance capability will be increasingly critical to meeting shareholder and public demands in WRPRUURZ¶V ZRUOG 7KLV UHTXLUHV D QHZ DQG LQWHJUDWHG IRFXV on research in maintenance within the high value manufacturing companies and their supply chains.
The increasing requirements for products quality and fluctuations in customer demand, as well as dynamic competitors impose significant requirements on the ability of manufacturers to provide customers with products in the required quantity, quality and for competitive price (Scheer 1994, Chapman 2006).
The most significant results to improve the efficiency of processes in the region can be achieved by providing integration platform to bring together ERP data and Enterprise Asset Management (EAM systems), based on current industry information platform (Becker et al. 2006). 2. ANALYSIS OF CURRENT ENTERPRISE ASSET MANAGEMENT (EAM)
Also, the cost and environmental impact of keeping long-life equipment and capital systems operating efficiently will have a key role (Taratoukhine et al. 2002), especially in high ± value products such as aerospace industry, automotive, and other industries. The high value products and infrastructure
978-3-902823-35-9/2013 © IFAC
In recent years, the importance and relevance of the problems of construction of system configurations for multicomponent systems has increased dramatically in many application areas: 341
10.3182/20130619-3-RU-3018.00634
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
- the low prevalence of specialists and consultants to implement; - the product is not available in many areas; - the lack of support for integration with other vendors' products; - support the operation only on the platform and limited use of advanced mathematical methods for periodic and nonperiodic; - preventive maintenance.
unit planning, resource allocation system, logistics system. Case studies EAM systems in the world testifies to their extremely high returns of investments. According number of studies the vast majority of the projects paid off in less than a year or two. Model is a reduction of 20% or more of the cost of maintenance, which is achieved by: - increase the availability, reliability and efficiency of plant and equipment / machinery (increase in production of 0.5% 1%); - improving the accuracy of planning maintenance of equipment (planned repairs share in the total number of repairs increased from 30% to 80%); - reducing the inefficiencies of human resources and materials and equipment (down the costs by 2% -5%); - to improve transparency and fairness including the cost of work on the management and maintenance of fixed assets; - increasing the validity of estimates of the cost of planned management and maintenance of fixed assets; - the construction of a single (correct and consistent) directory nomenclature capital assets.
Future literature review, industry interviews and case studies defined that it is a significant gap between current practice, integrated in EAM systems and new approaches for maintenance of high value manufacturing products. 3. MAINTENANCE SCHEDULING METHODS ANALYSIS Generally speaking, there are three types of maintenance in use: - Preventive maintenance, where equipment is maintained before break down occurs. This type of maintenance has many different variations and is subject of various researches to determine best and most efficient way to maintain equipment. Recent studies have shown that Preventive maintenance is effective in preventing age related failures of the equipment. For random failure patterns which amount to 80% of the failure patterns, condition monitoring proves to be effective. It can be divided into two subgroups: Scheduled maintenance (planned maintenance) and Condition-based maintenance. The effectiveness of the aircraft maintenance are determined mainly by scheduled maintenance. So we work with this type.
Currently on the market there are over 400 suppliers EAM systems, most of which focused on the development of systems solutions for narrow local problems. The leading and presented in most regions are products of companies are IBM, IFS, Infor, Mincom, Oracle and SAP. The strengths of these systems are: - the integration with ERP-systems from different vendors; - most support systems functioning on a variety of platforms; - support for extensive functionality; - high scalability; - the company's presence in all global markets.
- Corrective maintenance, where equipment is maintained after break down. This maintenance is often most expensive because worn equipment can damage other parts and cause multiple damage.
Specific analysis of EAM systems was conducted for SAP AG solutions. For instance, the SAP Enterprise Asset Management (SAP EAM) solution helps organizations efficiently and sustainably manage the whole lifecycle of physical assets by means of interoperability IT systems. Tools for managing production equipment, roads, machinery, vehicles, facilities, and power grids help reduce operating costs, minimize risk, and manage capital expenditures more effectively. Powerful reporting functionality and analytics improve asset usage and performance. Core Business Processes interoperability supported by SAP EAM: Planning, Building, and Commissioning; Asset Visibility and Performance; Asset Operations and Maintenance; Operational Risk Management; Real Estate Lifecycle Management (SAP Enterprise Asset Management Maximize Return on Assets, Manage Risks, and Improve Equipment Effectiveness).
- Reliability centred maintenance often known as RCM, is a process to ensure that assets continue to do what their users require in their present operating context. The main parameters that characterize the perfection of the maintenance system are the scope and frequency of maintenance work, which have a direct impact on operating costs and the intensity of the aircraft operation. It is obvious that the smaller the amount of work and the greater the frequency of their performance, the lower the operating costs and the stronger the aircraft operation. The share of the maintenance cost can vary from 10 to 50 percents of total operating costs (Airline Maintenance Program Development Seminar, Fleet Maintenance Seminars, Boeing, Seattle, Washington, USA, 2008, Chapman 2006). The aircraft design features, the technical operation methods, regulatory and procedural requirements, the quality of work, operating parameters, environmental conditions and other factors influence on the scope and frequency of maintenance.
Despite that most of popular EAM systems can be considered as industry standards (in variety of the sectors of the economy), as part of the analysis we defined some of the key challenges: - the high cost of implementation; - limited investment in vendor development of innovative technologies;
The scope and frequency of work are determined by the maintenance program. It should be noted that the 342
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
Union and is widely used by the aircraft developers and operators.
requirements for the content of the maintenance program in Russia differ from those operating abroad. In the Russian Federation the maintenance proJUDP LV D GHYHORSHU¶V document, common to the aircraft type, whereas the maintenance program abroad is a document developed by the operator for each individual aircraft taking into account the conditions and parameters of its operation and maintenance system adopted by the operator based on the minimum maintenance requirements specified in the process of type certification.
Practice shows that MSG-3 method has a number of drawbacks. MSG-3 principles have been refined and developed to Operator/Manufacturer Scheduled Maintenance Development (ATA MSG-3 2007), which engineers and service technicians can use to the create and maintain aircraft operation. MSG-3 method of maintenance scheduling describes the method for developing the scheduled maintenance program. Non-scheduled maintenance results from scheduled tasks, normal operation or data analysis.
In this article methods of aircraft maintenance scheduling are considered. The objectives of an efficient airline maintenance program are:
Maintenance programs will be developed via use of a guided logic approach and will result in a task-oriented program. The logic's flow of analysis is failure-effect oriented. Items that, after analysis, have no scheduled task(s) specified, may be monitored by an operator's reliability program. The method for determining the maintenance program for systems/powerplant, including components and auxiliary power unit, uses a progressive logic diagram. This logic is the basis of an evaluation technique applied to each maintenance significant item (system, sub-system, module, component, accessory, unit, part, etc.), using the technical data available. Principally, the evaluations are based on the item's functional failures and failure causes. Before the actual MSG-3 logic can be applied to an item, the aircraft's significant systems and components must be identified. This process of identifying Maintenance Significant Items (MSI's) is a conservative process (using engineering judgment) based on the anticipated consequences of failure (ATA MSG-3 2007).
- to ensure realization of the inherent safety and reliability levels of the equipment; - to restore safety and reliability to their inherent levels when deterioration has occurred; - to obtain the information necessary for design improvement of those items whose inherent reliability proves inadequate; - to accomplish these goals at a minimum total cost, including maintenance costs and the costs of resulting failures. These objectives recognize that maintenance programs as such cannot correct deficiencies in the inherent safety and reliability levels of the equipment. The maintenance program can only prevent deterioration of such inherent levels. If the inherent levels are found to be unsatisfactory, design modification is necessary to obtain improvement (ATA MSG-3 2007).
In the Russian Federation this task is carried out by specialists of "Applied Logistics" Scientific Research Centre of CALS-technologies. They proposed a "Method of maintenance frequency choosing on the basis of quantitative analysis methods".
In the general statement in the presence of all the necessary initial information quantitative confirmation of maintenance program effectiveness is to provide a set of probabilistic aircraft safety performance, while minimizing maintenance cost (e.g., maintenance cost per flying hour). In this approach, the main difficulty is the quantitative confirmation of compliance with the aircraft requirements, which requires an adequate mathematical model allowing to assess the impact of maintenance on the aircraft safety and effectiveness.
Scheduled maintenance frequency is set so as to minimize the natural deterioration of inherent aircraft safety and reliability, its systems and components, preventing the level of aircraft airworthiness from exceeding the established standards. Scope of the aircraft scheduled maintenance is determined by a working group comprising representatives of airlines and a developer and is submitted for approval.
Real alternative to the quantitative approach is the rational combination of quality engineering analysis of the choice of technical operation methods and the aircraft maintenance with a quantitative optimization of maintenance frequency on based on the mathematical model scheduled maintenance influence on the on-board systems reliability and safety. This allows to analyse the impact of possible system failures and their components on the aircraft flight safety, regularity and efficiency ("Applied Logistics" 2010).
In determining the aircraft maintenance work frequency standard modes of operation are considered: The average flight hours for typical flight, the average annual plaque ("Applied Logistics" 2010, Sudov et al. 2006). Nowadays maintenance program scheduling is based on product passport data, organization standards, industry requirements, etc. In this case, lead time is rigidly fixed on the time axis by assigning scheduled maintenance intervals. Maintenance intervals are set cyclically. The work scope at every maintenance is formed depending on its type. Each maintenance type is provided with a list of works given by the documentation to be executed as part of the maintenance. The main disadvantages of this method are:
A qualitative approach was proposed in the 70's by United Airlines and later became widely known as a guidance document on the development of a scheduled maintenance program - ATA MSG-3 (LII. MM Gromov, State Research Institute of Civil Aviation 2010). This document is recognized by the U.S. FAA, JAA and EASA European
343
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
- the inability to make changes to the maintenance program, - lack of feedback from the servicing system; - lack of data accumulation on the servicing system; - low adaptive capacity.
Table 1. Works Cost of material Ri, ¼
Work time Ti, labour time hour
1 2
Maximum time period between works ûWmax, day 100 200
0 0
10 2
Total cost of works, Si=Ri+Ti*c (c=¼2, cost of 1hour of labour time), ¼ 20 4
3 4 5 6 7 8 9 10
200 300 300 200 200 100 100 200
10 20 20 10 20 10 10 10
3 3 2 1 10 1 1 20
16 16 24 12 40 12 12 50
‹ Work
Based on this analysis, the main goal of this research was defined as: to create a method and algorithm for improving scheduling of periodic maintenance actions (works) as universal maintenance procedures. This will greatly simplify the support of the whole system, expand the area of its use, and hopefully will reduce the maintenance cost of complex products (aircraft). As a complex system we used an aircraft Yakovlev Yak-42. The Yakovlev Yak-42 (NATO reporting QDPH ³&OREEHU´ is a 100/120-seat three-engine mid-range passenger jet. It was designed as a replacement for several obsolete Aeroflot jets as a mid-range passenger jet. It was also the first airliner produced in the Soviet Union to be powered by modern highbypass turbofan engines.
Table 2. The compatibility matrix
4. PROBLEM STATEMENT AND METHODS PROPOSED
1
W ^wk `, where wk Maintenance procedure is work scope n is k-work, Wn is work scope, which is included in the maintenance Pn. Illustration of this fact you can see in the Figure 1.
1 2 3 4 5 6 7 8 9 10
2
3
4 +
5
6
7
8
9
10
+ + + + + + + + +
To achieve this goal we reduce the problem to the general problem of configuration design, which refers to the distribution of specific works among procedures. There are two types of tasks on system configuration: selection and/or placement and hierarchy system. Our problem is the first type. Input: a set of elements/components, which is divided into subsets (Ovsyannikov et al. 2012).
Fig. 1.The task of maintenance scheduling. The problem statement is the following: to define maintenance procedures work scope and time under which the selected criteria will take the extreme (minimum). The selected criteria can be the following: 1. 2.
The essence of the constraint programming method is to describe a set of domain rules and constraints. Constraints may be presented in the form of equations, inequalities, logical expressions, etc.
The maintenance total cost should seek to minimize: S o min , where S is the maintenance total cost; Maximum procedure time of execution should seek to minimize: Tmax max( Ti ) o min , where T is
(DFK FRPSRQHQW ³:RUN´ KDV LWV RZQ VHW RI SURSHUWLHV WKDW DUH often dependent (calculated) on the properties of its components. Such dependence can be described as functional dependencies.
maximum procedure time of execution.
Based on industry case studies and number of practical researches conducted with number of the aerospace companies we defined an advanced approach for improving the regular maintenance of complex, high value manufacturing products, described generally in number of stages:
The problem has to be solved taking into account special constraints. The special constraints can be the following: 1.
2.
Time period between the works (ûW) must be less than or equal to the specified maximum time period between the works (ûWmax); If two works, determined by the compatibility matrix, are done in one maintenance procedure, their costs will be reduce by 20%. 344
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
- The functional decomposition of the work into a hierarchy of components and the composition of the work scope (maintenance/repair work domain ontology). - Description of regularities determining the inclusion of this work in the periodic maintenance as constraints. - Starting a mechanism to monitor the resulting complex maintenance work based on constraint satisfaction programming (CSP) (Rossi et al. 2006). The last stage consists of 3 substages: First substage: satisfaction of the constraints that are associated with the known parameters of the components. Method NC-1 (Benhamou et al. 1997, Cleary 1987).
Fig. 2. Backtracking. The method allows to perform a full search of all the components combinations, and to find all possible solutions.
This algorithm allows to remove all the values that do not satisfy the unary constraints from all the domains of the problem variables.
5.SOLUTIONS AND RESULTS
Consistency in the nodes is achieved by viewing each item in each field and checking whether it satisfies the value of the unary constraints on that variable. All values that violate the unary constraints are removed from these domains.
We distributed the works on maintenance procedures and calculated maintenance total cost and procedure times. You can see the result of first experiment in Table 3.
Having completed the algorithm, the initial problem is reduced to the task, which is consistent in the nodes.
Table 3. Distribution of works
Second substage: the pairwise comparison of directory components, which is aimed at removing the components not compatible with each other. Method of pre-screening Forward Checking (Gent Ian, Bartak). Using the method of Forward Checking the search space can be significantly reduced or completely resolved (Benhamou et al. 1997, Cleary 1987). The basic idea of the method can be expressed as follows: if the set of inequalities becomes inconsistent while assigning the domain value to a variable, inconsistent values should be removed from the domains of the other variables. After that the search can be continued. In case the current domain is empty there is roll back to the previous level (Gent Ian). Domain is a set of components (works) from the directory. The search should be started with the variables that have domains with the least number of values.
ûWmax, day
‹ Work
90 days
180 days
270 days
360 days
450 days
540 days
100
1
1
1
1
1
1
1
200
2
1
1
1
200
3
1
1
1
300
4
1
1
300
5
1
1
200
6
1
1
1
200
7
1
1
1
100
8
1
1
1
1
1
1
100
9
1
1
1
1
1
1
200 Ti, hours
10
Si, ¼
Third substage: the calculation and verification of compliance with the requirements of the design parameters. Backtracking method (Dechter et al. 2002).
1
1
1
12
48
17
48
12
53
44
166
94
166
48
216
Tmax10=53 hours S10=¼734 We carried out the series of experiments using the special constraints with the following results: Tmax 21=37 hours S21=¼734 Tmax 22=35 hours S22=¼734 Tmax 31=53 hours S31=¼671,6 Tmax 41=35 hours S41=¼699,2 Tmax 51=48 hours S51=¼693,6 Tmax 52=42 hours
Backtracking algorithm for solving constraint satisfaction problems (CSP) is compiled on the basis of a recursive depthfirst search (Dechter et al. 2002, Semenov 2003). This method involves the enumeration of all possible combinations by traversing the tree layout options with returns in case of the constraints are not satisfied, as shown in Fig. 2.
345
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
Activities within Cost Estimating. 9th ISPE International Conference on Concurrent Engineering: Research and Applications, Netherlands: A.A. Balkema Publisher. Becker, J., Vilkov, L., Taratoukhine, V., Kugeler, M., Rosemann (2007). M.Process management. Moscow, Russia.Eksmo. Becker, J., Taratoukhine, V., Vilkov, L., Rieke, (2006). Process Driven Value Assessment of ERP Solutions: An Overview on the Extended SAP Methodology. In: Proceedings of the 2nd International Conference on Information Management and Business (IMB2006). Sydney, Australia. S. 225-235. Scheer, A. (1994). Computer Integrated Manufacturing: Towards the factory of the future, 3rd ed. Berlin u.a. SAP Enterprise Asset Management Maximize Return on Assets, Manage Risks, and Improve Equipment Effectiveness: http://www.sap.com/community/ebook/2012_05_EAM_ Max/en/index.html#/page/1. Airline Maintenance Program Development Seminar, Fleet Maintenance Seminars, Boeing, Seattle, Washington, USA (2008). ATA MSG-3. Revision 2007.1. (2007). Operator/Manufacturer Scheduled Maintenance Development. :L:. Guidelines "Guidelines for the analysis of logistic support aircraft production," Research Center of CALStechnologies, "Applied Logistics" (2010). M.: - 204p. Guide for designers and operators to develop programs and certification of maintenance and repair of civil aircraft. LII. MM Gromov, State Research Institute of Civil Aviation. (1993). Sudov, E.V., Levin, A.I., Petrov, A.V., Chubarova, E.V. (2006). Technology integrated logistics support engineering products. - Moscow: Publishing house 'Information', - 232. Rossi, F., Van Beek, P., Walsh, T. (eds.) (2006). Handbook Of Constraint Programming. Elsevier. 978 p. Ovsyannikov, M., Bukhanov, S. (2012). Configuration management with using the method constraint programming. News of higher educational institutions. Mechanical Engineering. ‹ pp. 70-75. Benhamou, F., Older, W. (1997). Applying Interval Arithmetic to Real, Integer and Boolean Constraints. Journal of Logic Programming, 32, pp.1-24. Cleary, J. (1987). Logical Arithmetic. Future Comput. Syst. 2 (2), pp. 125±149. Gent, Ian, Artificial Intelligence: Constraint Programming [Electronic resource] / Gent Ian.- Access mode: http://www.dcs.st-and.ac.uk/~ipg/AI/. Bartak, R. On-line guide to Constraint Programming [Electronic resource] / Bartak R.-Access mode: http://ktilinux.ms.mff.cuni.cz/~bartak/constraints/. Dechter, R., Frost D. (2002). Backjump-based backtracking for constraint satisfaction problems. Artificial Intelligence, 136(2): pp. 147±188. Semenov, A.L. (2003). Constraint propagation methods: basic concepts. PSI'03. IMRO - Interval mathematics and methods of distribution constraint.
S52=¼686 Tmax 53=37 hours S53=¼704,4 And constructed the graph:
Fig. 3. Plot of maintenance total cost and maximum procedure time of execution. If we compare the value of maintenance cost before and after applying the algorithm, we can see dramatic fall. The figure seems to suggest that, on the whole, our algorithm makes possible to decrease the maintenance cost by about 8,5 %. Moreover the maximum procedure time went down 34%. This is due to applying constraint satisfaction methods and decision logic, which allow to use new constraints if it is necessary. 6. CONCLUSIONS AND PERSPECTIVES We would like to wrap up the paper by summarising the main points. In this research the analysis of current Enterprise Asset management (EAM) systems was conducted. The analysis of existing methods of maintenance scheduling was presented. The key challenges and opportunities were defined. Advanced approach (algorithm) for periodic maintenance were described. The algorithm is based on constraint programming methods. The practical case studies for specific industry (aerospace industry) which are enable us to evaluate the effectiveness of the proposed framework was created. Also, cost reduction analysis were conducted to evaluate economic effectiveness of the proposed method. Finally, integration of the proposed solution to current ERP/EAM systems as joint co-innovation program between Bauman Moscow State University/ERCIS Lab Russia and SAP will be considered. REFERENCES GOST 18322-78 (1978). System maintenance and repair of equipment. Terms and definitions. British Standard Glossary of terms (3811:1993). Scheer, A. (1994). Business Process Engineering - Reference Models for Industrial Enterprises. 2nd ed. Berlin u.a. Chapman, S. N. (2006). The Fundamentals of Production Planning and Control. Pearson Prentice Hall. Taratoukhine, V., Roy, R., Mishra, K., Souchouorkov R. (2002). Knowledge in the Commercial and Engineering 346
Implementation of Constraints Satisfaction Problems Methods for Solving the Periodic Maintenance Processes Scheduling A.Fedotova*, V.Taratoukhine**, M.Ovsyannikov***, J.Becker**** * Department of computer automation, The Bauman Moscow State Technical University and European Research Centre in Information Systems (ERCIS) Lab Russia at Higher School of Economics, Russia (Tel: +7(926)5965359; email:[email protected]). ** University of Muenster, European Research Centre for Information Systems (ERCIS) HQ and ERCIS Lab Russia at Higher School of Economics, Germany (e-mail: [email protected]). ***Department of computer automation, The Bauman Moscow State Technical, Russia. **** University of Muenster, European Research Centre for Information Systems (ERCIS), Germany. Abstract: This paper reviews existing methods and techniques addressing the problem of maintenance support throughout the life cycle for high value manufacturing products. As part of this doctorate research the analysis of current methods of maintenance planning was conducted. In order to contribute to a more comprehensive solution, an advanced approach (algorithm) of periodic maintenance is presented. The authors believe that this approach will reduce the cost of maintenance of high value manufacturing products. The algorithm based on constraint programming methods is briefly presented and the future research directions are discussed. Keywords: Information systems, maintenance and repair of equipment, maintenance cost, periodic maintenance scheduling methods, methods of constraints satisfaction problems, domain ontology. are typically technology intensive, expensive and reliability critical requiring engineering services, such as maintenance support throughout the life cycle (Becker et al. 2007).
1. INTRODUCTION The competitiveness of the aircraft is influenced by a number of factors, including the operating system maintenance and repair of equipment. Maintenance and repair of equipment is an operation or a complex of operations intended to keep an item in the proper condition when used for intended purpose, on standby, for storage and transportation (GOST 18322-78 1978). British Standard Glossary of terms (3811:1993) defined maintenance as the combination of all technical and administrative actions, including supervision actions, intended to retain an item in, or restore it to, a state in which it can perform a required function. European Federation of National Maintenance Societies may be defined maintenance, repair, and operations as, "all actions which have the objective of retaining or restoring an item in or to a state in which it can perform its required function. The actions include the combination of all technical and corresponding administrative, managerial, and supervision actions."
From a purely industrial point of view, this means higher reliabilit\ RI D FRPSDQ\¶V WHFKQRORJLFDO FDSDFLWLHV DQG WKHLU constant availability at the peak of their features. Thus the importance of maintenance effectiveness considerably LQFUHDVHV DV ZHOO DV WKH ZKROH VHW RI D FRPSDQ\¶V DVVHW management processes (Becker et al. 2006, Scheer 1994). High value manufacturers are looking for additional cost savings through performance-based service contracts and by taking responsibilities for their product performance throughlife. The ability of owners and service providers to deploy best-in-class maintenance capability will be increasingly critical to meeting shareholder and public demands in WRPRUURZ¶V ZRUOG 7KLV UHTXLUHV D QHZ DQG LQWHJUDWHG IRFXV on research in maintenance within the high value manufacturing companies and their supply chains.
The increasing requirements for products quality and fluctuations in customer demand, as well as dynamic competitors impose significant requirements on the ability of manufacturers to provide customers with products in the required quantity, quality and for competitive price (Scheer 1994, Chapman 2006).
The most significant results to improve the efficiency of processes in the region can be achieved by providing integration platform to bring together ERP data and Enterprise Asset Management (EAM systems), based on current industry information platform (Becker et al. 2006). 2. ANALYSIS OF CURRENT ENTERPRISE ASSET MANAGEMENT (EAM)
Also, the cost and environmental impact of keeping long-life equipment and capital systems operating efficiently will have a key role (Taratoukhine et al. 2002), especially in high ± value products such as aerospace industry, automotive, and other industries. The high value products and infrastructure
978-3-902823-35-9/2013 © IFAC
In recent years, the importance and relevance of the problems of construction of system configurations for multicomponent systems has increased dramatically in many application areas: 341
10.3182/20130619-3-RU-3018.00634
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
- the low prevalence of specialists and consultants to implement; - the product is not available in many areas; - the lack of support for integration with other vendors' products; - support the operation only on the platform and limited use of advanced mathematical methods for periodic and nonperiodic; - preventive maintenance.
unit planning, resource allocation system, logistics system. Case studies EAM systems in the world testifies to their extremely high returns of investments. According number of studies the vast majority of the projects paid off in less than a year or two. Model is a reduction of 20% or more of the cost of maintenance, which is achieved by: - increase the availability, reliability and efficiency of plant and equipment / machinery (increase in production of 0.5% 1%); - improving the accuracy of planning maintenance of equipment (planned repairs share in the total number of repairs increased from 30% to 80%); - reducing the inefficiencies of human resources and materials and equipment (down the costs by 2% -5%); - to improve transparency and fairness including the cost of work on the management and maintenance of fixed assets; - increasing the validity of estimates of the cost of planned management and maintenance of fixed assets; - the construction of a single (correct and consistent) directory nomenclature capital assets.
Future literature review, industry interviews and case studies defined that it is a significant gap between current practice, integrated in EAM systems and new approaches for maintenance of high value manufacturing products. 3. MAINTENANCE SCHEDULING METHODS ANALYSIS Generally speaking, there are three types of maintenance in use: - Preventive maintenance, where equipment is maintained before break down occurs. This type of maintenance has many different variations and is subject of various researches to determine best and most efficient way to maintain equipment. Recent studies have shown that Preventive maintenance is effective in preventing age related failures of the equipment. For random failure patterns which amount to 80% of the failure patterns, condition monitoring proves to be effective. It can be divided into two subgroups: Scheduled maintenance (planned maintenance) and Condition-based maintenance. The effectiveness of the aircraft maintenance are determined mainly by scheduled maintenance. So we work with this type.
Currently on the market there are over 400 suppliers EAM systems, most of which focused on the development of systems solutions for narrow local problems. The leading and presented in most regions are products of companies are IBM, IFS, Infor, Mincom, Oracle and SAP. The strengths of these systems are: - the integration with ERP-systems from different vendors; - most support systems functioning on a variety of platforms; - support for extensive functionality; - high scalability; - the company's presence in all global markets.
- Corrective maintenance, where equipment is maintained after break down. This maintenance is often most expensive because worn equipment can damage other parts and cause multiple damage.
Specific analysis of EAM systems was conducted for SAP AG solutions. For instance, the SAP Enterprise Asset Management (SAP EAM) solution helps organizations efficiently and sustainably manage the whole lifecycle of physical assets by means of interoperability IT systems. Tools for managing production equipment, roads, machinery, vehicles, facilities, and power grids help reduce operating costs, minimize risk, and manage capital expenditures more effectively. Powerful reporting functionality and analytics improve asset usage and performance. Core Business Processes interoperability supported by SAP EAM: Planning, Building, and Commissioning; Asset Visibility and Performance; Asset Operations and Maintenance; Operational Risk Management; Real Estate Lifecycle Management (SAP Enterprise Asset Management Maximize Return on Assets, Manage Risks, and Improve Equipment Effectiveness).
- Reliability centred maintenance often known as RCM, is a process to ensure that assets continue to do what their users require in their present operating context. The main parameters that characterize the perfection of the maintenance system are the scope and frequency of maintenance work, which have a direct impact on operating costs and the intensity of the aircraft operation. It is obvious that the smaller the amount of work and the greater the frequency of their performance, the lower the operating costs and the stronger the aircraft operation. The share of the maintenance cost can vary from 10 to 50 percents of total operating costs (Airline Maintenance Program Development Seminar, Fleet Maintenance Seminars, Boeing, Seattle, Washington, USA, 2008, Chapman 2006). The aircraft design features, the technical operation methods, regulatory and procedural requirements, the quality of work, operating parameters, environmental conditions and other factors influence on the scope and frequency of maintenance.
Despite that most of popular EAM systems can be considered as industry standards (in variety of the sectors of the economy), as part of the analysis we defined some of the key challenges: - the high cost of implementation; - limited investment in vendor development of innovative technologies;
The scope and frequency of work are determined by the maintenance program. It should be noted that the 342
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
Union and is widely used by the aircraft developers and operators.
requirements for the content of the maintenance program in Russia differ from those operating abroad. In the Russian Federation the maintenance proJUDP LV D GHYHORSHU¶V document, common to the aircraft type, whereas the maintenance program abroad is a document developed by the operator for each individual aircraft taking into account the conditions and parameters of its operation and maintenance system adopted by the operator based on the minimum maintenance requirements specified in the process of type certification.
Practice shows that MSG-3 method has a number of drawbacks. MSG-3 principles have been refined and developed to Operator/Manufacturer Scheduled Maintenance Development (ATA MSG-3 2007), which engineers and service technicians can use to the create and maintain aircraft operation. MSG-3 method of maintenance scheduling describes the method for developing the scheduled maintenance program. Non-scheduled maintenance results from scheduled tasks, normal operation or data analysis.
In this article methods of aircraft maintenance scheduling are considered. The objectives of an efficient airline maintenance program are:
Maintenance programs will be developed via use of a guided logic approach and will result in a task-oriented program. The logic's flow of analysis is failure-effect oriented. Items that, after analysis, have no scheduled task(s) specified, may be monitored by an operator's reliability program. The method for determining the maintenance program for systems/powerplant, including components and auxiliary power unit, uses a progressive logic diagram. This logic is the basis of an evaluation technique applied to each maintenance significant item (system, sub-system, module, component, accessory, unit, part, etc.), using the technical data available. Principally, the evaluations are based on the item's functional failures and failure causes. Before the actual MSG-3 logic can be applied to an item, the aircraft's significant systems and components must be identified. This process of identifying Maintenance Significant Items (MSI's) is a conservative process (using engineering judgment) based on the anticipated consequences of failure (ATA MSG-3 2007).
- to ensure realization of the inherent safety and reliability levels of the equipment; - to restore safety and reliability to their inherent levels when deterioration has occurred; - to obtain the information necessary for design improvement of those items whose inherent reliability proves inadequate; - to accomplish these goals at a minimum total cost, including maintenance costs and the costs of resulting failures. These objectives recognize that maintenance programs as such cannot correct deficiencies in the inherent safety and reliability levels of the equipment. The maintenance program can only prevent deterioration of such inherent levels. If the inherent levels are found to be unsatisfactory, design modification is necessary to obtain improvement (ATA MSG-3 2007).
In the Russian Federation this task is carried out by specialists of "Applied Logistics" Scientific Research Centre of CALS-technologies. They proposed a "Method of maintenance frequency choosing on the basis of quantitative analysis methods".
In the general statement in the presence of all the necessary initial information quantitative confirmation of maintenance program effectiveness is to provide a set of probabilistic aircraft safety performance, while minimizing maintenance cost (e.g., maintenance cost per flying hour). In this approach, the main difficulty is the quantitative confirmation of compliance with the aircraft requirements, which requires an adequate mathematical model allowing to assess the impact of maintenance on the aircraft safety and effectiveness.
Scheduled maintenance frequency is set so as to minimize the natural deterioration of inherent aircraft safety and reliability, its systems and components, preventing the level of aircraft airworthiness from exceeding the established standards. Scope of the aircraft scheduled maintenance is determined by a working group comprising representatives of airlines and a developer and is submitted for approval.
Real alternative to the quantitative approach is the rational combination of quality engineering analysis of the choice of technical operation methods and the aircraft maintenance with a quantitative optimization of maintenance frequency on based on the mathematical model scheduled maintenance influence on the on-board systems reliability and safety. This allows to analyse the impact of possible system failures and their components on the aircraft flight safety, regularity and efficiency ("Applied Logistics" 2010).
In determining the aircraft maintenance work frequency standard modes of operation are considered: The average flight hours for typical flight, the average annual plaque ("Applied Logistics" 2010, Sudov et al. 2006). Nowadays maintenance program scheduling is based on product passport data, organization standards, industry requirements, etc. In this case, lead time is rigidly fixed on the time axis by assigning scheduled maintenance intervals. Maintenance intervals are set cyclically. The work scope at every maintenance is formed depending on its type. Each maintenance type is provided with a list of works given by the documentation to be executed as part of the maintenance. The main disadvantages of this method are:
A qualitative approach was proposed in the 70's by United Airlines and later became widely known as a guidance document on the development of a scheduled maintenance program - ATA MSG-3 (LII. MM Gromov, State Research Institute of Civil Aviation 2010). This document is recognized by the U.S. FAA, JAA and EASA European
343
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
- the inability to make changes to the maintenance program, - lack of feedback from the servicing system; - lack of data accumulation on the servicing system; - low adaptive capacity.
Table 1. Works Cost of material Ri, ¼
Work time Ti, labour time hour
1 2
Maximum time period between works ûWmax, day 100 200
0 0
10 2
Total cost of works, Si=Ri+Ti*c (c=¼2, cost of 1hour of labour time), ¼ 20 4
3 4 5 6 7 8 9 10
200 300 300 200 200 100 100 200
10 20 20 10 20 10 10 10
3 3 2 1 10 1 1 20
16 16 24 12 40 12 12 50
‹ Work
Based on this analysis, the main goal of this research was defined as: to create a method and algorithm for improving scheduling of periodic maintenance actions (works) as universal maintenance procedures. This will greatly simplify the support of the whole system, expand the area of its use, and hopefully will reduce the maintenance cost of complex products (aircraft). As a complex system we used an aircraft Yakovlev Yak-42. The Yakovlev Yak-42 (NATO reporting QDPH ³&OREEHU´ is a 100/120-seat three-engine mid-range passenger jet. It was designed as a replacement for several obsolete Aeroflot jets as a mid-range passenger jet. It was also the first airliner produced in the Soviet Union to be powered by modern highbypass turbofan engines.
Table 2. The compatibility matrix
4. PROBLEM STATEMENT AND METHODS PROPOSED
1
W ^wk `, where wk Maintenance procedure is work scope n is k-work, Wn is work scope, which is included in the maintenance Pn. Illustration of this fact you can see in the Figure 1.
1 2 3 4 5 6 7 8 9 10
2
3
4 +
5
6
7
8
9
10
+ + + + + + + + +
To achieve this goal we reduce the problem to the general problem of configuration design, which refers to the distribution of specific works among procedures. There are two types of tasks on system configuration: selection and/or placement and hierarchy system. Our problem is the first type. Input: a set of elements/components, which is divided into subsets (Ovsyannikov et al. 2012).
Fig. 1.The task of maintenance scheduling. The problem statement is the following: to define maintenance procedures work scope and time under which the selected criteria will take the extreme (minimum). The selected criteria can be the following: 1. 2.
The essence of the constraint programming method is to describe a set of domain rules and constraints. Constraints may be presented in the form of equations, inequalities, logical expressions, etc.
The maintenance total cost should seek to minimize: S o min , where S is the maintenance total cost; Maximum procedure time of execution should seek to minimize: Tmax max( Ti ) o min , where T is
(DFK FRPSRQHQW ³:RUN´ KDV LWV RZQ VHW RI SURSHUWLHV WKDW DUH often dependent (calculated) on the properties of its components. Such dependence can be described as functional dependencies.
maximum procedure time of execution.
Based on industry case studies and number of practical researches conducted with number of the aerospace companies we defined an advanced approach for improving the regular maintenance of complex, high value manufacturing products, described generally in number of stages:
The problem has to be solved taking into account special constraints. The special constraints can be the following: 1.
2.
Time period between the works (ûW) must be less than or equal to the specified maximum time period between the works (ûWmax); If two works, determined by the compatibility matrix, are done in one maintenance procedure, their costs will be reduce by 20%. 344
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
- The functional decomposition of the work into a hierarchy of components and the composition of the work scope (maintenance/repair work domain ontology). - Description of regularities determining the inclusion of this work in the periodic maintenance as constraints. - Starting a mechanism to monitor the resulting complex maintenance work based on constraint satisfaction programming (CSP) (Rossi et al. 2006). The last stage consists of 3 substages: First substage: satisfaction of the constraints that are associated with the known parameters of the components. Method NC-1 (Benhamou et al. 1997, Cleary 1987).
Fig. 2. Backtracking. The method allows to perform a full search of all the components combinations, and to find all possible solutions.
This algorithm allows to remove all the values that do not satisfy the unary constraints from all the domains of the problem variables.
5.SOLUTIONS AND RESULTS
Consistency in the nodes is achieved by viewing each item in each field and checking whether it satisfies the value of the unary constraints on that variable. All values that violate the unary constraints are removed from these domains.
We distributed the works on maintenance procedures and calculated maintenance total cost and procedure times. You can see the result of first experiment in Table 3.
Having completed the algorithm, the initial problem is reduced to the task, which is consistent in the nodes.
Table 3. Distribution of works
Second substage: the pairwise comparison of directory components, which is aimed at removing the components not compatible with each other. Method of pre-screening Forward Checking (Gent Ian, Bartak). Using the method of Forward Checking the search space can be significantly reduced or completely resolved (Benhamou et al. 1997, Cleary 1987). The basic idea of the method can be expressed as follows: if the set of inequalities becomes inconsistent while assigning the domain value to a variable, inconsistent values should be removed from the domains of the other variables. After that the search can be continued. In case the current domain is empty there is roll back to the previous level (Gent Ian). Domain is a set of components (works) from the directory. The search should be started with the variables that have domains with the least number of values.
ûWmax, day
‹ Work
90 days
180 days
270 days
360 days
450 days
540 days
100
1
1
1
1
1
1
1
200
2
1
1
1
200
3
1
1
1
300
4
1
1
300
5
1
1
200
6
1
1
1
200
7
1
1
1
100
8
1
1
1
1
1
1
100
9
1
1
1
1
1
1
200 Ti, hours
10
Si, ¼
Third substage: the calculation and verification of compliance with the requirements of the design parameters. Backtracking method (Dechter et al. 2002).
1
1
1
12
48
17
48
12
53
44
166
94
166
48
216
Tmax10=53 hours S10=¼734 We carried out the series of experiments using the special constraints with the following results: Tmax 21=37 hours S21=¼734 Tmax 22=35 hours S22=¼734 Tmax 31=53 hours S31=¼671,6 Tmax 41=35 hours S41=¼699,2 Tmax 51=48 hours S51=¼693,6 Tmax 52=42 hours
Backtracking algorithm for solving constraint satisfaction problems (CSP) is compiled on the basis of a recursive depthfirst search (Dechter et al. 2002, Semenov 2003). This method involves the enumeration of all possible combinations by traversing the tree layout options with returns in case of the constraints are not satisfied, as shown in Fig. 2.
345
2013 IFAC MIM June 19-21, 2013. Saint Petersburg, Russia
Activities within Cost Estimating. 9th ISPE International Conference on Concurrent Engineering: Research and Applications, Netherlands: A.A. Balkema Publisher. Becker, J., Vilkov, L., Taratoukhine, V., Kugeler, M., Rosemann (2007). M.Process management. Moscow, Russia.Eksmo. Becker, J., Taratoukhine, V., Vilkov, L., Rieke, (2006). Process Driven Value Assessment of ERP Solutions: An Overview on the Extended SAP Methodology. In: Proceedings of the 2nd International Conference on Information Management and Business (IMB2006). Sydney, Australia. S. 225-235. Scheer, A. (1994). Computer Integrated Manufacturing: Towards the factory of the future, 3rd ed. Berlin u.a. SAP Enterprise Asset Management Maximize Return on Assets, Manage Risks, and Improve Equipment Effectiveness: http://www.sap.com/community/ebook/2012_05_EAM_ Max/en/index.html#/page/1. Airline Maintenance Program Development Seminar, Fleet Maintenance Seminars, Boeing, Seattle, Washington, USA (2008). ATA MSG-3. Revision 2007.1. (2007). Operator/Manufacturer Scheduled Maintenance Development. :L:. Guidelines "Guidelines for the analysis of logistic support aircraft production," Research Center of CALStechnologies, "Applied Logistics" (2010). M.: - 204p. Guide for designers and operators to develop programs and certification of maintenance and repair of civil aircraft. LII. MM Gromov, State Research Institute of Civil Aviation. (1993). Sudov, E.V., Levin, A.I., Petrov, A.V., Chubarova, E.V. (2006). Technology integrated logistics support engineering products. - Moscow: Publishing house 'Information', - 232. Rossi, F., Van Beek, P., Walsh, T. (eds.) (2006). Handbook Of Constraint Programming. Elsevier. 978 p. Ovsyannikov, M., Bukhanov, S. (2012). Configuration management with using the method constraint programming. News of higher educational institutions. Mechanical Engineering. ‹ pp. 70-75. Benhamou, F., Older, W. (1997). Applying Interval Arithmetic to Real, Integer and Boolean Constraints. Journal of Logic Programming, 32, pp.1-24. Cleary, J. (1987). Logical Arithmetic. Future Comput. Syst. 2 (2), pp. 125±149. Gent, Ian, Artificial Intelligence: Constraint Programming [Electronic resource] / Gent Ian.- Access mode: http://www.dcs.st-and.ac.uk/~ipg/AI/. Bartak, R. On-line guide to Constraint Programming [Electronic resource] / Bartak R.-Access mode: http://ktilinux.ms.mff.cuni.cz/~bartak/constraints/. Dechter, R., Frost D. (2002). Backjump-based backtracking for constraint satisfaction problems. Artificial Intelligence, 136(2): pp. 147±188. Semenov, A.L. (2003). Constraint propagation methods: basic concepts. PSI'03. IMRO - Interval mathematics and methods of distribution constraint.
S52=¼686 Tmax 53=37 hours S53=¼704,4 And constructed the graph:
Fig. 3. Plot of maintenance total cost and maximum procedure time of execution. If we compare the value of maintenance cost before and after applying the algorithm, we can see dramatic fall. The figure seems to suggest that, on the whole, our algorithm makes possible to decrease the maintenance cost by about 8,5 %. Moreover the maximum procedure time went down 34%. This is due to applying constraint satisfaction methods and decision logic, which allow to use new constraints if it is necessary. 6. CONCLUSIONS AND PERSPECTIVES We would like to wrap up the paper by summarising the main points. In this research the analysis of current Enterprise Asset management (EAM) systems was conducted. The analysis of existing methods of maintenance scheduling was presented. The key challenges and opportunities were defined. Advanced approach (algorithm) for periodic maintenance were described. The algorithm is based on constraint programming methods. The practical case studies for specific industry (aerospace industry) which are enable us to evaluate the effectiveness of the proposed framework was created. Also, cost reduction analysis were conducted to evaluate economic effectiveness of the proposed method. Finally, integration of the proposed solution to current ERP/EAM systems as joint co-innovation program between Bauman Moscow State University/ERCIS Lab Russia and SAP will be considered. REFERENCES GOST 18322-78 (1978). System maintenance and repair of equipment. Terms and definitions. British Standard Glossary of terms (3811:1993). Scheer, A. (1994). Business Process Engineering - Reference Models for Industrial Enterprises. 2nd ed. Berlin u.a. Chapman, S. N. (2006). The Fundamentals of Production Planning and Control. Pearson Prentice Hall. Taratoukhine, V., Roy, R., Mishra, K., Souchouorkov R. (2002). Knowledge in the Commercial and Engineering 346