UdZPraxis 2-2018

55 UdZ Praxis | Gastbeitrag Setting the foundation – data gathering and handling To get the right data in the right quality is the first and usually completely un- derestimated step in the process of data driven maintenance execution. The good news is that there is a variety of software available with little or no need for alteration, since processes are fairly similar in the production industries. The bad news is that the challenge usually does not lie in the selection of a pro- gram, but in the implementation and process deployment. While the software investment commonly is around €100,000 – 150,000, the implementation, trai- ning and consulting cost exceed this usually by far. Typically the lack of sufficient master data, various non-formalized data sour- ces and the lack of user acceptance and benefits stand in the way of an easy implementation of the maintenance data system. Essential insights generated inmultiplemaintenance data implementations are: • Generatingdatacostsmoney:Alwaysthinkfirstbeforeyoucreateandpos- sibly over-engineer equipment structures, workflows and maintenance routines. Focus on the important tasks, equipment and orders. It is al- ways possible to add detail; it is very hard to take it away from a running system. • Maintenance data is company data: When implementing maintenance datamanagement, always keep inmind that there is more than themain- tenance engineering team is involved in, such as asset management and themaintenance process. There need to be sufficient links to accounting, purchasing, controlling, production, and so on. • Focus on the data, not the software solution: The software must have a flexible interface allowing to upload massive data sets from Excel as well as to make amendments for collections of records simultaneously in the system itself. • Easy to work with: The less time the input and management of the data needs, the more time the maintenance engineering team has for produc- tive work and for actually using the generated data. • Centralizedknowledge:Convertallexperientialknowledge,HSQE (health, safety, quality and environment) and OEM (original equipment manufacturer) manuals into digital job descriptions, instructions and task lists and identify the appropriate level of competence, which allows the system in the future to allocate resources accordingly du- ring work planning. • Cluster failures togainabetter asset understan- ding: Define a library of all actual and possible failures and their causes. It is vital to keep in mind during specification and standardizati- on of failures that only a combination of both the equipment name and failure name give a unique character to the event. Therefore, attempts to put too much specification in the designation of failures leads to over-growth of the catalog and towrong conclusions. • Priorities for continuous improvement: De- velop a measuring system with reasonable rankingcriteriaof failures and remedial actions (e.g., Risk Priority Number – RPN), which help to track progress while reviewing the main- tenance strategy, and measure the results. There are well known and proven methodologies like FMECA (Failure Mode, Effect, and Criticality Analysis) and RCA (Root Cause Analysis) that pro- vide – even with basic analytical tools like Excel – a robust approach for the implementation of risk ba- sed maintenance. Acasestudybelow(seefigure2, page16) represents all insights listed above in one Excel sheet. In the im- plementation phase, a cross-functional team starts to fill in the left part of the table (FMECA and RCA) Figure 3: Recommended approach for datamanagement and analysis