Injection Molding Cycle Monitoring for Predictable Part Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for monitoring injection molding processes struggle to reliably predict the quality of molded parts due to the complexity and non-linearity of these processes, which are further complicated by changes in environmental conditions and machine settings, often requiring large datasets not readily available at the start of production.
Innovation Solution
The method determines the nearest neighbors in cycle data to assess predictability criteria, checking if the quality spread of these neighbors meets certain thresholds, and if not, issues a message indicating unreliable predictions, triggering data collection or operator intervention to ensure quality assessment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If simulation models or statistical models are used to predict molded part quality, then prediction capability is improved, but the complexity of the system increases and large amounts of data are required which are often unavailable
Solution Approach 1:
The patent uses simple distance calculations and basic statistical comparisons instead of complex simulation or machine learning models. The approach treats quality prediction as a straightforward comparison between current cycle data and historical data, avoiding the need for sophisticated computational models while maintaining practical reliability
Solution Approach 2:
The system uses the forming machine's own historical data to predict quality without requiring external complex models. The machine's cycle data serves itself for prediction purposes through simple nearest-neighbor comparisons, eliminating the need for separate complex prediction systems
2Reliability
If complex testing systems are implemented to ensure quality monitoring, then monitoring effectiveness is improved, but the effort and system complexity required increases significantly
Solution Approach 1:
The patent extracts only the essential elements needed for quality prediction - cycle data parameters and quality data - from the complex forming process. By focusing solely on comparing these extracted data elements using simple distance calculations, the system achieves effective monitoring without implementing complex testing infrastructure
Solution Approach 2:
Instead of using complex models to predict quality and then verifying them, the patent inverts the approach by directly comparing current cycle data with historical data using simple distance metrics. This inverted approach achieves monitoring effectiveness through simplicity rather than complexity
3Measurement precision
If statistical models with large datasets are used for reliable predictions, then prediction accuracy is improved, but the system cannot adapt quickly at the start of production when data is limited
Solution Approach 1:
The patent uses a partial approach by relying on a limited number of nearest neighbors (k-value) from historical data rather than requiring large datasets. This partial action enables the system to provide predictions with acceptable accuracy even when data is limited, allowing quick adaptation at the start of production
Solution Approach 2:
The system prepares for future predictions by continuously storing and organizing historical cycle data and quality data. This preliminary accumulation of data in structured formats enables quick adaptation when production starts, as the system can immediately begin comparing current cycles against the pre-organized historical data
Data Source
Figure 1
Figure 2a~4
Figure 5
AI summary
A method for monitoring a forming process carried out in cycles, wherein a data collection is provided which includes at least the following data: - for past cycles, cycle data of the forming process carried out on a forming machine (1), and - for at least some of the past cycles, quality data of molded parts (2) produced with the forming machine (1), and wherein at least one further cycle of the forming process is carried out with the forming machine (1), and current cycle data of the at least one further cycle are collected, wherein the following further steps are carried out: - Determining at least two nearest neighbors in the form of the cycle data of at least two of the past cycles, such that the cycle data of the at least two nearest neighbors are closer to the current cycle data than those cycle data that do not belong to the at least two nearest neighbors.where, for the determination of the at least two nearest neighbors, only those past cycles are used for which quality data are included in the data collection; - checking a predictability criterion, whereby it is checked whether a quality variation of the quality data of the cycles of the at least two nearest neighbors is less than a maximum variation and/or greater than a minimum variation; and - if the predictability criterion is not met, issuing an initial message that a quality and/or a quality date of the molded part (2) produced with the at least one further cycle is not reliably predictable.