UdZPraxis 2-2018

56 UdZ Praxis | Gastbeitrag during initial workshops. With a good cross-functio- nal team and an experienced moderator, about 60 to 70 % of failures can be identified and preventive measures defined. The next step is to upload the re- sults to a computerized system for the implementa- tion innormal operations and collectionof feedback for continues improvement. In this stage the user centricity of the software so- lution is key, because especially in the implementa- tion of processes and routines every complication, every failure or time loss will be remembered by the team. In the next weeks or months, the data will be tested and used. The team needs to be able to navigate easily through the equipment structu- res and failure catalogues, and to filter or add in- formation to the existing data. It is also necessary to identify any limitations of the current catalogue framework or equipment structure for further im- provement in the following steps. Priorities for improvement – data analysis Add the information gained in the operational test to table 1. In the following workshops, the goal is to conduct an RCA analysis in regards to the worst as- sets in the equipment structure, enabling the team to cover 80 to 90% of failures with the predefined catalogues. In this step, data visualization and data management of the software solutionareof utmost importance. Diagrams, individual KPIs, CAD draw- ings and other information linked to the object help to identify and visualize the problems with the as- sets anddirect you to the right strategy for improve- ment. The easier the information is provided, the easier it is for a variety of users to exchange their knowledge and helpwith the decision analysis. Putting the data to use – decision analysis The aimof the following workshops is to complete the middle and left parts of table 1, reviewing and assessing the current maintenance program. The teamneeds to view all necessary information, star- ting with the bottom ten assets in terms of availa- bility or failure rate and question the maintenance processes currently in place. While the data enab- les the team to get a more complete picture of the situation and to address problems, it is only a tool to adjust business processes. The use of sophistica- ted maintenance solutions with machine learning and other fancy technologies will not make up for any mistake made in earlier steps or shortcuts ta- ken, which prevent fine tuning of the systems. Since the adjustment and im- plementation will take up significant time, we offer an approach which inclu- des a combination of the tools used in the first two steps and postpones the use the high priced advanced decision analysis software to the second phase (see figure 3, page 17). The process needs to be done regularly to establish a continuous impro- vement process. At this stage, every team member needs to have access to all relevant information. The team needs to decide together on adjustments of the maintenance strategy and measures for specific assets, and track all decisions based on the accepted asset risks (RPN). The improvement process does not only consist of the improvement of assets, but also of setting the bar for failure occurrence higher on a regular basis. For instance, the company agreed to set the highest acceptable occurrence rate of failures (MTBF –Mean Time Between Failures) for the RPN criterion to 3 months and the lowest occurrence to 18months. Thus, every asset that has a likelihood of failure of more than once in three months requires a thorough RCM analysis, while for an asset that is likely to fail less than once in 1.5 years a less intensive maintenance strategy can be implemented. To improve asset availability, the company could alter the limits to 4 months / 2 years and so on. This assessment does not solve all the challenges in a modern production landscape; HSQE guidelines, for example, need to be regularly updated and the assets need to be improved accordingly. Intelligent and thorough data management comes first Before starting with smart maintenance and machine learning, get things done right. Big data and analytics are a great way to get the most out of your assets, but they are not always the biggest lever and require a solid data foun- dation. As shown it is possible to get more out of the resources you have with relatively simple tools by applying the right method and bringing together the right people. To turn a computer system into a working tool and take full ad- vantage of the capabilities of modern software solutions, specific steps must be taken, and both management and personnel need to be involved in sha- ping the future business processes. Only the right processes are able to gene- rate a solid data foundation and enable the RCMmethod towork and improve asset lifecycle management and overall costs. df · Kryukov For more information please contact: Dr.-Ing. Philipp Jussen FIR e. V. an der RWTHAachen, Bereichsleiter Dienstleistungsmanagement E-Mail: Philipp.Jussen@fir.rwth-aachen.de Dipl.-Ing. Florian Defèr FIR e. V. an der RWTH Aachen, Bereich Dienstleistungsmanagement E-Mail: Florian.Defer@fir.rwth-aachen.de Boris Kryukov, M.Sc. Research and Consulting Group GmbH, Senior consultant E-Mail: B.Kryukov@rcg-ag.com