Decision modelling for maintenance and safety

Kari Laakso, Miki Sirola, Jan Holmberg

    Research output: Contribution to journalArticleScientificpeer-review


    Several objectives which may be conflicting must be considered when making strategic maintenance decisions. Studies of decision cases have shown that the methods of decision analysis provide a systematic and clear basis for making complex decisions under multicriteria conditions. Such decision criteria might be system safety, availability, maintenance cost and work culture. A decision analysis approach provides a structure for identifying and ranking the significant decision options, decision criteria and information needs for making of the decisions. For instance, the results of probabilistic safety assessment and maintenance analysis, together with other engineering knowledge, can be used in a decision analysis to solve a decision problem.

    Experiences from practical cases of decision analysis have indicated that a better decision support procedure could be achieved by the introduction of tentative decision models and a tailored planning tool. Decision models were constructed based on knowledge obtained from earlier decision and reliability analyses and the collection of expert knowledge. A spreadsheet tool for qualitative and quantitative decision analysis was defined and developed. The tool helps to prepare, perform and document the decision analysis sessions with the decision models being stored in the tool. At present, decision criteria for strategic decision cases such as plant shutdown or continued operation in a fault situation, need of component replacements by repeating faults, ranking of optional modifications and the need for condition monitoring have been modeled.
    Original languageEnglish
    Pages (from-to)13-17
    Number of pages5
    JournalInternational Journal of COMADEM
    Issue number3
    Publication statusPublished - 1999
    MoE publication typeA1 Journal article-refereed


    Dive into the research topics of 'Decision modelling for maintenance and safety'. Together they form a unique fingerprint.

    Cite this