In mass production environments that rely on machining processes, a common challenge is the continuation of operations with a broken tool without realizing it, leading to the production of defective items. By monitoring the condition of tools and detecting breakages in real time, it's possible to minimize the production of defective products.
MAZIN, Inc. is committed to developing AI solutions to address production challenges on a process-by-process basis. In the field of machining, one focus is on developing algorithms for tool condition monitoring. This is achieved by attaching current sensors to machine tools and analyzing the electrical current values, which correlate with cutting torque, as part of various experiments and analyses.
To avoid false detections that can arise from simply basing anomaly detection on the magnitude of current values, a more precise method is employed for mass production processes that involve repetitive cycles of machining the same product. This method involves comparing the current values from the current machining cycle with those from previous cycles. Such a comparison allows for more accurate anomaly detection, enhancing the ability to identify tool breakage or wear without interrupting the production flow unnecessarily.
This approach not only aims to reduce the rate of defective products but also contributes to maintaining high production efficiency and quality control by ensuring that tools are in optimal condition throughout the machining process.
This time, we introduce a crucial technology for comparing current values between the current and previous machining cycles: cycle detection technology.
A machining cycle can be identified in coordination with the NC (Numerical Control) device. However, retrofitting machines to establish this connection can sometimes be impractical or expensive.
MAZIN, Inc. has developed a cycle detection algorithm that can identify a machining cycle from time-series data without the need to interface directly with the NC control device. This algorithm allows for the automatic identification of cycles with just a few simple steps to register the machining process information, eliminating the need for physical modifications to the machine tool.
The advancement of cycle detection technology expands the variety of machine tools that can adopt tool anomaly detection systems.
We will continue to improve the cycle detection technology to make it compatible with an even broader range of machine tool types.
For more detailed information about the technology or inquiries related to research and development, please contact us.
