Machine Learning Device for Injection Molding Shrinkage Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mold design for injection molding faces challenges in accurately predicting molding shrinkage ratios, particularly due to variations in resin type, additive composition, and molding conditions, leading to increased costs and inefficiencies from repeated trial and error processes.
Innovation Solution
A machine learning device and design support system that acquire input data on molding conditions, including resin type, additive ratio, and mold temperature, and use supervised learning to generate a learned model predicting molding shrinkage ratios in both flow and vertical directions, allowing for precise prediction and optimization of mold design.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional trial and error methods are used for mold design, then molding shrinkage ratios can be obtained through actual measurement, but the number of trials increases leading to higher costs and longer development time
Solution Approach 1:
The patent applies preliminary action by performing flow analysis and residual stress analysis during the design phase to predict volume shrinkage and birefringence distribution before actual molding trials. This allows the molding shrinkage ratio to be calculated in advance based on the relationship between degree of birefringence and shrinkage anisotropy, eliminating the need for multiple trial mold manufacturing cycles.
2Measurement precision
If traditional trial and error methods are used for mold design, then accurate molding shrinkage ratios can be obtained, but the number of mold manufacturing trials increases leading to higher costs
Solution Approach 1:
The patent replaces the mechanical trial-and-error system with a computational system. Flow analysis and residual stress analysis are used to calculate volume shrinkage and predict birefringence distribution, which then allows calculation of molding shrinkage ratios through the established relationship with shrinkage anisotropy. This substitutes physical mold manufacturing trials with computational predictions.
3Adaptability or versatility
If molding conditions are set without advance prediction, then flexibility in design is maintained, but experience is required and multiple trials are needed to determine optimal conditions
Solution Approach 1:
The patent implements feedback by using the calculated volume shrinkage and predicted birefringence distribution to determine the molding shrinkage ratio through the relationship with shrinkage anisotropy. This feedback loop allows accurate prediction of molding shrinkage based on flow analysis and residual stress analysis results, enabling precise mold design without requiring operator experience or multiple trials.
Data Source
AI summary
A molding shrinkage ratio is predicted according to molding conditions set in advance in designing a mold. A machine learning device includes an input data acquiring unit that acquires input data including any molding condition including a type of resin, a type of additive, a blending ratio of the additive, a surface temperature of a mold, and a product of a holding pressure and a holding pressure time for any article molded by any injection molding machine, a label acquiring unit that acquires label data indicating a molding shrinkage ratio in a flow direction and a molding shrinkage ratio in a vertical direction perpendicular to the flow direction of a resin measured of the article molded at the molding condition, and a learning unit that executes supervised learning using the input data and the label data, and generates a learned model.