Injection Mold Model for Faster Machine Tuning and Predictive Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for tuning injection molding machines when switching molds require significant time, material, and energy, as they necessitate a new tuning process from scratch, leading to inefficiencies in production flexibility and scheduling.
Innovation Solution
A computer-implemented method generates an extended mold model by correlating mold cavity data and machine parameters using machine learning algorithms, enabling predictive control and optimization, allowing for faster adaptation to new machines and improved quality prediction across the injection cycle, including filling, packing, and cooling phases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a new tuning process is performed from scratch when switching molds, then the machine parameters can be optimized for the specific mold, but significant time, material, and energy are consumed
Solution Approach 1:
The system performs preliminary actions by pre-processing injection cycle data from multiple machines and storing it in a database. When a mold is switched to a new machine, the system quickly retrieves and applies pre-analyzed parameters instead of performing a complete tuning process from scratch, significantly reducing setup time while maintaining optimization quality
Solution Approach 2:
The system cushions against the time loss by pre-processing and storing optimal parameter sets for different machine-mold combinations before they are actually needed. This preparatory data storage acts as a buffer that enables rapid parameter transfer when molds are switched between machines
2Manufacturing precision
If a new tuning process is performed from scratch when switching molds, then the machine parameters can be optimized for the specific mold, but material and energy waste increases
Solution Approach 1:
The system pre-processes and stores optimal parameter combinations before mold switching occurs. This eliminates the need to generate parameters through trial-and-error testing on the new machine, thereby preventing material waste from test shots and energy waste from extended tuning cycles
Solution Approach 2:
The system creates copies of optimized parameter sets from reference machines and applies them to target machines. This copying approach avoids the need to re-discover optimal parameters through expensive material consumption and energy-intensive testing processes
3Measurement precision
If complete injection cycle data is collected and processed using machine learning algorithms, then predictive control accuracy is improved, but data processing complexity increases
Solution Approach 1:
The system segments the complex data processing task into distinct phases: data collection during injection cycles, preprocessing to remove irregularities, classification of injection cycles, and machine learning model generation. This segmentation makes the complex process more manageable and enables parallel processing of different data streams
Solution Approach 2:
The system introduces an intermediary preprocessing step that cleans and organizes raw sensor data before feeding it to machine learning algorithms. This intermediary layer removes irregular data values and standardizes formats, reducing the complexity burden on the downstream analysis components
4Productivity
If machine learning algorithms are used to generate extended mold models, then production efficiency and quality control are enhanced, but computational resources and processing time are consumed
Solution Approach 1:
The system performs machine learning model generation as a preliminary action during periods when data is naturally available (during normal production cycles). By preparing predictive models in advance rather than in real-time during critical production moments, the system balances computational energy consumption with productivity gains
Data Source
AI summary
A computer implemented method for generating a mold model for production predictive control and computer program products thereof. The method comprises receiving first parameters about molding machine sensors and second parameters about mold cavity; classifying each injection cycle of a plurality of injection cycles of a first injection molding machine considering the first and second parameters and quality or characteristics of injected given parts in the machine; processing the first and second parameters to remove undesired or irregular data values thereof; merging the first and second parameters providing a global group of processed parameters; executing a machine learning algorithm on the global group of processed parameters generating an extended mold model; and using said generated extended mold model for further monitoring and control of the mold in further injection processes in the first injection molding machine and/or for optimizing a production process of the mold in the first molding machine.