Real-Time Quality Monitoring and Fault Detection for Injection Molding
The SenseLink™ QM provides an entire solution for process monitoring and part quality prediction utilizing multivariate analysis (MVA) technology from Umetrics. All necessary data acquisition, multivariate processing, and control functions come from one self contained, compact unit. Adding SenseLink™ QM to your injection molding processes will provide in-process fault detection along with contribution details which provide an understanding of your process not attainable from traditional SCADA and SPC approaches. The multivariate analysis provides a significant fault detection improvement over typical SPC type or univariate solutions currently used in molding which allows molders to eliminate manual and visual quality inspection and to achieve parametric release or "lights out" molding operations; additional, typical ROI benefits include:
- Lights-Out Manufacturing: Full Parametric release for 100% Quality Control.
- Reduced Operator "in-process" testing.
- Reduce routine visual inspection costs by 33%.
- Reduce number of exceptions by 50%.
- Defect Root Cause Analysis: Reduce MTTD (Mean-Time-To-Detect) by over 50%.
- Reduce Machine Downtime by 5%.
- Reduced Product Variability: Increase Cpk by 10%.
- Full ROI within 2 quarters
The SenseLink™ QM mounts to almost any molding machine and connects to existing machine sensors along with in-mold sensors, if available. Process data is acquired from analog, digital, serial and/or networked sources. All configuration and model creation functionality is accomplished via a web browser user interface, so no additional software is required.
FEATURES & BENEFITS
- Identify the quality of parts in-process
- Detected faults based on out of spec critical part dimensions as well as physical defects which include flash, short shots, sink marks, voids, burning, contamination, bubbles, surface appearance defects, and others
- Setup using automated modeling capabilities; does not require MVA training
- Models can adapt to acceptable process changes resulting from material viscosity changes, time, environment, etc.
- Process variables contributing to poor quality are easily identified
- Compact; easy to mount and integrate
- Reduced false alarms over typical univariate methods used in injection molding industry. For example, by increasing the number of process variables from 10 to 15, the number of false alarms increases nearly 50% (false alarm rate increases from 2.9% to 4.4%) when using univariate or SPC fault detection approaches, whereas, adding process variables to a MVA system does not increase the number of false alarms.
- Easy web browser configuration and data analysis; no additional software required
- Excellent multivariate prediction capabilities with existing machine sensors; does not require cavity sensors
- Automatic part containment
- Production, yield, OEE, and detailed multivariate reporting capabilities
- More efficient preventative maintenance scheduling
- Real-time process control with velocity to pressure (VP) switch over capabilities
Once SenseLink™ QM is installed on a system data can be collected, stored, and retrieved, as well as sent to a networked database for additional storage. Using multivariate techniques, the system builds a multivariate model around the optimized process with an acceptable processing window, established from a design of experiments (DOE). New data is then compared in real-time to the alarm limits developed by the model. The multivariate alarms are then triggered if the new data is off-spec, sending a reject signal to a part containment device for real-time part containment. Contribution plots are then shown, for each cycle, ranked by significance as to which variable or interaction of variables was responsible for the alarm. The results are available for web browser review and sent to local servers.
For slowly shifting or drifting processes, adaptive techniques are included to allow the model to adjust to the process change. As a result false alarms are minimized. Adaption can be reset to the model baseline after machine or mold maintenance is performed or as new process inputs are set. The multivariate models that are utilized for process defects detection consist of Principle Component Analysis (PCA) and Project to Latent Structures (PLS).
SenseLink™ QM is a data acquisition system with a multivariate data analysis engine that was developed for injection molding. The system, however, can be used with other plastics manufacturing processes or secondary operations to provide significant ROI benefits through improved process understanding, shortened process debug times, reduced scrap, and reduced product recalls. Plastics manufacturing processes and secondary operations that are or can be used with the SenseLink™ QM are:
- Blow Molding - Extrusion and Injection
- Insert Molding
- Thermoset Molding
- Ultrasonic and RF Welding
- Multi-shot Injection Molding
- Gas Assist Molding
- Painting Operations
- Form, Fill, and Seal Operations
Each SenseLink™ QM system is contained within a single, compact data acquisition and computing module. The module consists of the MKS RMU hardware platform with SenseLink data monitoring software, and Umetrics SIMCA-QM multivariate analysis software license.
The modules have expandable features, with a minimum size of 4" x 4" x 2.5". The number of process variables which can be collected is potentially unlimited. Data sources include analog, digital, serial, and Ethernet. The base unit can accommodate up to 64 analog input channels and up to 96 digital I/O channels, which can also be expanded by connection to additional Ethernet I/O modules.
Every effort has been taken to ensure the completeness and accuracy of MKS Instruments product information. For up to date information on this category from MKS please click here.