Smartphone-enabled predictive maintenance : development and implementation of a reference architecture and processes

Predictive maintenance (PdM) is a hot topic in the field of manufacturing. However, its industry-wide realization lacks accepted integration concepts. Small and medium-sized enterprises (SMEs), in particular, tend to postpone PdM initiatives, primarily due to the high costs and effort of creating interoperability with established as well as in-use machines. PdM requires machine data to be proactively maintained. Therefore, in-use machines without integrated sensors must be replaced or cost-intensively upgraded. Furthermore, it is not advisable to invest in upgrades of existing machines, as they are cost-intensive, and their remaining lifespan is unknown as well as difficult to predict. One promising approach to applying PdM to these kinds of machines is the use of retail smartphones. With up to 16 sensors onboard, they offer an opportunity to cost-effectively collect required data without being tied to a single machine. Following a design science research approach, we present a reference software architecture consisting of a mobile and server application and reference processes for smartphone-enabled PdM to provide a lightweight approach, especially for SMEs. Together with five manufacturers and a software developer, we demonstrated and evaluated our artifacts using the software prototypes in a real-world setting.