On the occasion of the BI-MU fair in Milan, MCE presented a new and interesting application, concerning Predictive Maintenance on Cloud MindSphere, developed on the cloud platform in collaboration with the 40Factory company.
For some years now MCM and its Engineering MCE section have embarked on a research and development path in the field of predictive maintenance, which is bringing interesting and important results. After an initial phase of study, a series of modules and applications have been created which constitute a now complete solution.
In principle, the modules for interfacing with the machines were developed, responsible for the acquisition of high frequency data. In this case, the approach was to group the data from multiple sensors to obtain a compact and realistic image of the condition of the component or controlled group. Once the data had been acquired, data fusion and sensory data analysis algorithms were developed, also in collaboration with research institutes, to try to predict the wear trends of mechanical components, before failures and therefore episodes of stopping of machines. Finally, applications have been created that implement the management of the service and the display of results, allowing users to always have under control the updated status of the machine and their monitored components.
The developed application is located in this last area. The module shows the current operating status of the machine and the health status of all controlled components, updated in real time. The application also provides tools for analyzing historical data that allow you to easily identify possible anomalies and go into detail in the deepening of specific situations detected.
The software was developed in the cloud and in particular the MindSphere platform proposed by Siemens was chosen. After some experience of direct development of applications on the cloud, the need was felt for the presence of an infrastructure that would guarantee reliability and good performance even in the presence of large volumes of data, as well as security on access management for users, in based on role and organization. Given the lack of knowledge of the MindSphere infrastructure, MCE decided to collaborate for development with the 40factory company which, despite being a young company, has gained a good knowledge of this infrastructure. The collaboration proved to be very positive as it was possible to share and compare the different experiences developed on cutting-edge technologies for the analysis and processing of data and for their presentation, through the creation of web based interfaces and views.
The machines that MCM produces generally remain in service for more than twenty years and therefore wear is very slow, as is their replacement. A predictive maintenance system is therefore particularly convenient because it should allow users to exploit their means of production throughout their life cycle, keeping the episodes of failure and relative downtime contained, even in old age. Even though the predictive maintenance system has been in operation for a relatively short time, it is proving to be so efficient that MCM is convinced to equip all new machines produced with the data collection module (Flight Recorder). At the same time, we are also proceeding with the sensorization of the machines already in service, with perhaps a decade of work behind them and on which it is more interesting to detect the degradation and criticality of the components, when they actually approach the phase in which the probability that problems occur increases.
The work carried out for the realization of this latest application will allow to support new proactive maintenance services and will bring advantages both to the final users of the plants and to the MCM manufacturer. Users will have the opportunity to see and manage their systems even remotely, with the security of data and access to the cloud guaranteed by the MindSphere platform. The manufacturer will have the possibility to access in real time all of its machines installed and distributed all over the planet. This will allow him to be much more effective in servicing interventions but also to collect significant data on the durability and reliability of the components that will allow him to build increasingly robust and reliable machines.