A knowledge management system for series-parallel availability optimization and design

https://doi.org/10.1016/j.eswa.2006.08.023Get rights and content

Abstract

System availability is an important subject in the design field of industrial system as the system structure becomes more complicated. While improving the system’s reliability, the cost is also on the upswing. The availability is increased by a redundancy system. Redundancy Allocation Problem (RAP) of a series-parallel system is traditionally resolved by experienced system designers. We proposed a genetic algorithm based optimization model to improve the design efficiency. The objective is to determine the most economical policy of components’ mean-time-between-failure (MTBF) and mean time-to-repair (MTTR). We also developed a knowledge-based interactive decision support system to assist the designers set up and to store component parameters during the intact design process of repairable series-parallel system.

Introduction

In 1952, the Advisory Group on the Reliability of Electronic Equipment (AGREE) defined the reliability in a broader sense: reliability indicates the probability implementing specific performance or function of products and achieving successfully the objectives within a time schedule under a certain environment (Wang, 1992). In general, a higher priority is placed on quality control rather than reliability in the process of manufacturing. Nonetheless, high quality is not equivalent to high reliability. For example, a certain component, which has passed quality control procedure in conformity to the specifications, may lead to problems when operating with other components. This involves reliability design that is related to electrical or mechanical interface compatibility among spare parts.

With the rapid technological progress and increasing complexity of system structure, any failure of any component may lead to system malfunction or serious damage. For instance, a weapon system is a precise and sophisticated system that comprises several sub-systems, components and spare parts. Failure of even a single element will likely have adverse impact upon the operability of the weapon system, or even threat the national security.

System availability, a concept closely related to reliability, refers to the scale of measuring the reliability of a repairable system. Repairable system indicates a system that can be repaired to operate normally in the event of any failure, such as computer network, manufacturing system, power plant or fire prevention system. Availability comprises “reliability” and “recovery part of unreliability after repair”, indicating the probability that repairable systems, machines or components maintain the function at a specific moment” (Wang, 1992). It is generally expressed as the operable time over total time.

In recent years, reliability and availability have expanded their influence in various industries and fields, thus serve as an integral quality element in the organization system and manufacturing process. To maintain the reliability of sophisticated systems to a higher level, the system’s structural design or system components of higher reliability shall be required, or both of them are performed simultaneously (Henley & Kumampto, 1985).

The system structure is virtually designed under the limitations such as weight, volume or other technologies, so the reliability cannot be further improved. In this case, replacing highly reliable components can improve the system reliability. While improving the reliability of systems and components, the associated cost also increases. Thus, it is a very important topic for decision-makers to fully consider both the actual business and the quality requirements. Redundancy Allocation Problem (RAP) of a series-parallel system refers to difficult NP-hard problems (Chern, 1992). Redundancy allocation is designed depending upon the experience of system designers, with the advantages (Chisman, 1998): (1) time-saving and convenient policy-making depending upon years of experience and (2) decision making via experience in the absence of information. The disadvantages include: (1) decision-making is subjective, without scientific support or evidence and (2) individual experience-based decision cannot offer an accurate or optimal design, thus leading to excessive cost. Due to potential risks, the experience-based empirical law may not be universally applied (Li, 2001). Additionally, given the fact of difficult accumulation and inheritance of design expertise, it would be very helpful to transfer, accumulate and manage design knowledge by applying systematic methods and by employing information technologies. Jeang (1999) suggested that computer-aided simulation software could contribute to system design or parameterization. Many information systems were built and a wide variety of methods were used for the reliability design (Chen and Hsu, 2006, Liu and Yang, 1999, Moon et al., 1998, Varde et al., 1998). However, a well-defined knowledge system for reliability design and availability optimization was not found in the literature.

Under repairable series-parallel system framework, there are many methods to determine the optimal parameters of components, such as dynamic planning, integer programming, non-linear integer programming and heuristic or metaheuristic algorithms. As a member of metaheuristic algorithm, Genetic Algorithm (GA) has proved itself to be able to approaching optimal solution against any problem.

The purpose of this study is first to utilize Genetic Algorithms to determine MTBF (mean time between failure, MTBF) and MTTR (mean time to repair, MTTR) of various components, during the design phase of the repairable system, and to optimize availability parameters. We proposed an optimization model of repairable series-parallel system and utilized Genetic Algorithms to find solutions. We then constructed a knowledge-based information system so that the design knowledge can be stored and accumulated. The optimization model and GA procedures ensure that the cost-effective parameters of system availability can be obtained, which helps the system designers formulate optimal design policies and repair policies. The information system stored the system designs and parameters in the knowledge base and can be retrieved by significant features, which facilitates design complexity and increases design efficiency. Specifically, the objective of this study is threefold: (1) develop an optimization model of repairable series-parallel system availability and analyze the model behavior; (2) utilize genetic algorithms to obtain optimal parameter of system components in a cost-effective manner; and (3) construct a knowledge-based information system to accumulate the design knowledge.

Section snippets

Reliability of series-parallel system

Series-parallel system indicates sub-systems in which several components are connected in parallel, and then in series, or sub-systems that several components are connected in series, and then in parallel. A series-parallel system can be improved by four methods (Wang, 1992): (1) use more reliable components; (2) increase redundant components in parallel; (3) utilize both #1 and #2; and (4) enable repeatedly the allocation of entire system framework. For the framework of series-parallel system,

Manufacturing cost

The manufacturing cost varies with different product specifications. For electronic components, a longer MTFB of manufactured components represents a lower failure (λ) and higher strength, indicating that the components feature high reliability. The product quality is thus ensured once the failure rate of components declines to a desired level. In principle, while the failure rate (λ) is lower, the components are more difficult to fabricate, leading to a sharp increase in the manufacturing cost

System analysis

The system is configured for installation at the R&D department’s site with the database applications installed. The system users are primarily series-parallel system designers of the R&D department. The system database is updated while the system receives commands from users or generates computational results. While executing the optimization process, the system responds to the users with real-time result, which provides feedback for users to control the optimization process. Specifically, the

Conclusion

In the intellectual economy era, the design of repairable series-parallel system is inefficient if relying merely on empirical method. It tends to cause increasing design cost due to the difficulty of inheriting design experience. Applying soft computing techniques such as Genetic Algorithms to analyze and optimize the design problems of repairable series-parallel system appears to be very helpful in facilitating the decision-making of system parameter design.

We proposed an optimization model

References (57)

  • Y.-C. Hsieh et al.

    Genetic algorithms for reliability design problems

    Microelectronics Reliability

    (1998)
  • H.-Z. Huang

    Fuzzy multi-objective optimization decision-making of reliability of series system

    Microelectron Reliability

    (1997)
  • Y.-J. Hwang et al.

    A knowledge-based approach to the evaluation of fault trees

    Reliability Engineering and System Safety

    (1996)
  • M. Kumral

    Reliability-based optimisation of a mine production system using genetic algorithms

    Journal of Loss Prevention in the Process Industries

    (2005)
  • C.M.F. Lapa et al.

    A model for preventive maintenance planning by genetic algorithms based in cost and reliability

    Reliability Engineering and System Safety

    (2006)
  • S.-S. Lin et al.

    Statistical tolerance analysis based on beta distributions

    Journal of Manufacturing Systems

    (1997)
  • T.I. Liu et al.

    Design for quality and reliability using expert system and computer spreadsheet

    Journal of the Franklin Institute

    (1999)
  • S. Martorell et al.

    Alternatives and challenges in optimizing industrial safety using genetic algorithms

    Reliability Engineering and System Safety

    (2004)
  • Y.B. Moon et al.

    AEWS: an integrated knowledge-based system with neural networks for reliability prediction

    Computers in Industry

    (1998)
  • M. Samrout et al.

    New methods to minimize the preventive maintenance cost of series-parallel systems using ant colony optimization

    Reliability Engineering and System Safety

    (2005)
  • P.V. Varde et al.

    An operator support system for research reactor operations and fault diagnosis through a connectionist framework and PSA based knowledge based systems

    Reliability Engineering and System Safety

    (1998)
  • T. Yokota et al.

    Genetic algorithm for non-linear mixed integer programming problems and its applications

    Computers and Industrial Engineering

    (1996)
  • P.-S. You et al.

    An efficient heuristic for series-parallel redundant reliability problems

    Computers and Operations Research

    (2005)
  • W.Y. Yun et al.

    Multi-level redundancy optimization in series systems

    Computers and Industrial Engineering

    (2004)
  • A. Birolini

    Reliability engineering theory and practice

    (1999)
  • R.L. Bulfin et al.

    Optimal allocation of redundant components for large systems

    IEEE Transactions on Reliability

    (1985)
  • J.A. Chisman

    Using discrete simulation modeling to study large-scale system reliability/availability

    Computers and Operations Research

    (1998)
  • D.W. Coit et al.

    Reliability optimization of series-parallel systems using genetic algorithm

    IEEE Transactions on Reliability

    (1996)
  • Cited by (88)

    • Availability analysis of Tube Light Manufacturing industry using numerical technique

      2022, Materials Today: Proceedings
      Citation Excerpt :

      The Several authors have addressed the problems of reliability and availability of the system such as Gupta et al. [12] proposed a decision support model based upon the Markovian approach applied to the feed water subsystem of coal based power plant. Juang et al. [15] applied an evolutionary technique i.e. GA to optimize availability of a hybrid system. Azaron et al. [1] have been developed the reliability evaluation and optimization method for non-repairable different subsystems which were cold standby.

    • A new fuzzy methodology-based structured framework for RAM and risk analysis

      2019, Applied Soft Computing Journal
      Citation Excerpt :

      Since for an industrial system failure is an unavoidable event which takes place due to various failure causes (associated with different subsystem/equipment) therefore prevention or elimination of such failure causes is on top priority for the system analyst [1,2]. For the elimination or prevention of failure causes, it is essential for the analyst to study or analyze the performance/behavior issues of a system [3,4]. However, under real life situation, due to various complexities related to technological advancement, vague data/information or human error etc., performance analysis of an industrial system up to desired level of accuracy has become extremely difficult [5].

    • Optimization of the redundancy allocation problem: Genetic algorithm and Monte Carlo simulation with discrete events

      2023, Journal of the Faculty of Engineering and Architecture of Gazi University
    • Cold Standby System Availability

      2023, 2023 International Telecommunications Conference, ITC-Egypt 2023
    View all citing articles on Scopus
    View full text