Plastics machinery and technology supplier Milacron is releasing data-driven developments for the industrial Internet of Things that it says ensures optimum operation of its machines under wear and tear, which unlocks the full value of predictive analysis.
Branded as M-Powered, the company says the three new applications use predictive models to help operators, managers and service technicians go beyond point-failure prediction to use "next-generation, machine-learning algorithms that continuously quantify the impact of wear and tear."
The new applications, which are related to screw efficiency, screw tips and efficiency reports, will be available to manufacturers beginning in June with assistance from Milacron’s data science partners, ei3.
“Our data science team has crafted algorithms that give operators valuable insights into the inner workings of the Milacron machine; for the first time, it allows us to quantify the impact of wear and tear as it progresses. Operators can then take smart decisions by weighing that cost against the cost of maintenance. This was only possible by close collaboration between our data scientist and the engineering team from Milacron. This is a true milestone for the industry,” Stefan Hild, director of data science at ei3, said in a news release.
Predictive analytics provide pathways to leverage data to track machine condition and advise operators when it triggers maintenance actions.
The three new applications are being adapted from extensive research and testing on Milacron machinery in hopes to go beyond the typical break-fix methodology in injection molding and extrusion.
Edward Jump, M-Powered industrial IoT digital analytics leader at Milacron, said the developments help manufacturers reduce waste and improve efficiency in their operations.
“In real-world applications, true maintenance requirements are based on many variables. Through the adoption of machine learning and advanced analytics and AI, M-Powered can now monitor signals of impending failure,” Jump said in the release.
The new applications are the result of extensive research and testing on Milacron machinery, he added.
The application available in June involves the plasticizing screw, which is a component that conveys a resin while generating and using mechanical and conductive energy to provide a homogenous melt for molding an acceptable part.
As wear of this component occurs and the flight diameters of the screw begin to deteriorate, the screw's ability to efficiently convey material is reduced, consequently leading to increased recovery time, energy consumption and increased melt temperatures.
Milacron says the deterioration of the process is also slow and can be difficult to immediately detect. Since the plasticizing screw is a highly engineered component, it also can be expensive to replace and often comes with a long lead time, which can hamper productivity.
Milacron says its patent-pending system offers touch-free functions that can give insight into the screw’s health without adding sensors to the screw or barrel. This allows M-Powered users to order a new screw and replace it before wear and tear has a great impact on each cycle.
The application available in August involves the feed screw tip, which is the link between the machine and the mold that contains a check ring or non-return valve critical to the molding process. Screw tips are a high-wear item that often catch the brunt of the molding application and have a direct impact on processing quality and repeatability.
Milacron says M-Powered uses a proprietary shot-by-shot multivariate analysis to determine the state of the screw tip and indicate the effect of wear and tear on part quality, cycle time and operating costs. The application gives an alert when issues impact repeatability. Like the screw analytics, Milacron says this is an evolving machine learning process that takes distinct parts and operator adjustments into account.
The third application, which is expected to hit the market in September, provides efficiency reports from aggregated machine data to understand the impact on energy cost shot by shot.
Operators can see an efficiency summary for each machine or their entire fleet in real-time. Then, any costly offenders of cycle deviations can be corrected.