Z-BRE4K – Towards the demonstration of Predictive Maintenance solutions
Vignola, 25th March 2020
Z-BRE4K, the EC-funded project that aims to achieve zero downtimes caused by unexpected breakdowns in manufacturing plants, is now in the mid of Year 3 and after the consolidation and integration of the Predictive Maintenance (PdM) solutions within project’s end-users, is proceeding towards the validation and testing of Z-BRE4K solutions.
In the last months, technology providers and end-users have cooperated to integrate Z-BRE4K subcomponents at the shop floor level in the industrial environments of the 3 end-users: GESTAMP, SACMI-CDS and PHILIPS. This allowed to perform an initial demonstration at TRL6 and to produce deployment guidelines, including pilot area analysis, software topology designs and pilot installations specification. Now, at M30 of the project, the Consortium is ready for validating and testing Z-BRE4K solutions.
The automotive use case led by Gestamp considers a multistage zero-defect manufacturing cell for the frame components of light aluminum and steel components of automotive models. Here, Z-BRE4K solution consists of implementing cognitive maintenance strategies for full equipment and process production availability while addressing the quality control of instruments on the shop floor and measure it with the established Key Performance Indicators (KPIs). The solution exploits available information from the shop floor and connects with the MES and quality control system for data retrieval, then data are sent to the Z-BRE4K PdM Component which creates predictions, also with the support of a digital twin technology. This allows to optimise the working conditions and reduce breakdowns and down times of the machines, without affecting cycle time nor chassis product quality. At the moment, all the systems have been integrated in the shop floor and the Machine Learning (ML) techniques are giving outputs for the deployment of a Decision Support System (DSS) that will recommend a list of actions to avoid unexpected failures.
The second use case consists of introducing predictive maintenance functionalities within the CDS’ Compression Moulding machines used provided by Sacmi in the beverage sector. Now, after the deployment of a Condition Monitoring solution, the shopfloor of CDS, the plastic-products manufacturer, will be enhanced with PdM functionalities by the deployment of a DSS that provides information on the arising failure and the Remaining Useful Life (RUL) and suggests actions for avoiding failures from happening. The different kinds of information provided by the DSS are based on the output of different ML approaches used for the detection of anomalies and degradation of components of the Compression Moulding Machines.
The use case hosted by PHILIPS consists of a production line with cold forming tooling where the company aims to move from preventive to predictive maintenance. Ζ-BRE4K combines various data source and sensor readings for the implemented predictions algorithms which computes the expected useful life of tools and compares the results to manufacturer’s experience thresholds on the shop floor. Thanks to this, predictions are communicated to the DSS that creates suggestions on improving the manufacturing procedures and sends the notification to production managers or operators on the shop floor. Predictive maintenance implemented in this use case will improve the uptime of machinery, while reducing tooling time, stocking of parts and working hours and it will create a maintenance plan on the shop floor based on the predictions of the expected tool life including new maintenance orders for machinery and tools.
In the next months, project partners will test and evaluate the functioning of the Z-BRE4K solutions within the end-users facilities (WP6), and they will finally deliver the system’s results and lesson learnt which will answer to quality and maintenance measurements, predictive maintenance accuracy and performance, improvement of productivity and cost reduction goals set by the project.