Search

Graphical abstract Abstract With the immense pressure to sustain competitiveness in manufacturing, the strategy of digitizing this industry sector is vital. With the onset of new ICT technology and big data capabilities, the physical asset and data computation is integrated in manufacturing through Cyber Physical Systems (CPS). This Industry 4.0 strategy will significantly improve maintenance of machines and processes. Current big-data approaches focus on data available in production systems for monitoring purposes. However, data processing to define critical characteristic values for condition monitoring and maintenance remains challenging. Large and special-purpose machine tools are constantly re-configured regarding process, workpiece and machine itself, thus increasing the complexity of determining limit values. Therefore, it is not possible to execute a robust condition monitoring without structured data-analyses considering different machine states. Fuzzy-clustering of machine states over time creates a stable pool representing different typical machine configuration clusters. Hence, new and discontinuous machine states can be gradually attributed to such clusters to interpret their key characteristic values and limits, even when the concrete configuration never occurred before. [ABSTRACT FROM AUTHOR]

Condition Monitoring and Maintenance prediction are permanent demands of state-of-the-art machine tools. Their IT-infrastructure including drive control systems and specific sensors deliver extensive amounts of data for monitoring. However to combine and link this data to define critical characteristic values remains challenging. The paper presents a novel approach to deal with these tasks. Firstly equal machine states are detected at different times by checking and comparing several parameters. If an equal state is detected an algorithm is executed which leads to a characteristic value. The value as well as its limits is self-adapting and time-depending by constant redefining based on the machine history. [ABSTRACT FROM AUTHOR]

Condition Monitoring and Maintenance prediction are permanent demands of state-of-the-art machine tools. Their IT-infrastructure including drive control systems and specific sensors deliver extensive amounts of data for monitoring. However to combine and link this data to define critical characteristic values remains challenging. The paper presents a novel approach to deal with these tasks. Firstly equal machine states are detected at different times by checking and comparing several parameters. If an equal state is detected an algorithm is executed which leads to a characteristic value. The value as well as its limits is self-adapting and time-depending by constant redefining based on the machine history. [ABSTRACT FROM AUTHOR]