Glass Container Cavity Identification via Physical Parameter Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for identifying the cavity of a glass container forming machine rely on reading a molded code, which can fail due to reader errors, contamination, or code deterioration, leading to unknown defects and inefficient corrective measures.
Innovation Solution
A method that determines characteristic parameters of glass containers, such as wall thickness, and uses a supervised learning algorithm to classify and identify the cavity of manufacture, allowing identification without relying on the molded code.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a molded code is used to identify the cavity of manufacture, then the identification method is simple and direct, but the identification may fail due to reader errors, contamination, or code deterioration
Solution Approach 1:
The patent introduces characteristic parameters (wall thickness, diameter, height) as intermediary features that indirectly identify the cavity of manufacture. Instead of directly reading the molded code, the system measures physical characteristics that vary by cavity and uses these measurements to infer the origin cavity, thereby avoiding the reliability issues of direct code reading.
Solution Approach 2:
The patent replaces the mechanical/optical code reading system with a measurement and classification system. Instead of using a code reader to detect visual/markers, the system uses dimensional measurement devices to capture physical characteristics and a classification algorithm to identify the cavity, substituting a more reliable physical measurement approach for the prone-to-failure code reading approach.
2Loss of information
If code reading is used for defect tracking, then the process is straightforward, but defects with unknown cavity origin cannot be properly addressed
Solution Approach 1:
The patent performs preliminary measurement and classification of containers during normal production. By continuously measuring characteristic parameters and classifying containers by their origin cavity before defects occur, the system ensures that cavity information is already captured and stored. When a defect is detected, the cavity identification is already available, eliminating the need for additional identification actions at the defect handling stage.
Solution Approach 2:
The patent implements a feedback loop where characteristic parameters are continuously measured and used to update the classification of containers by cavity. This ongoing feedback ensures that the system maintains accurate knowledge of which cavity produced which container, even as production conditions vary over time. The classification is continuously refined based on new measurements, ensuring reliable cavity identification for defect tracking.
3Measurement precision
If machine learning classification is implemented to identify cavities, then identification accuracy improves, but the system complexity and computational requirements increase
Solution Approach 1:
The patent changes the parameters used for identification from visual code features to physical dimensional parameters (wall thickness, diameter, height). These physical parameters naturally vary by cavity due to manufacturing variations, providing a reliable basis for classification. By using parameters that inherently encode cavity information through physical variation rather than artificial markers, the system achieves accurate identification without complex processing.
Solution Approach 2:
The patent uses simple, inexpensive measurement of physical characteristics as a disposable alternative to expensive, complex code reading systems. Instead of investing in high-precision code readers and maintenance, the system uses basic dimensional measurements that are cheap to obtain and process, achieving reliable identification through statistical classification rather than precise direct reading.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
The invention relates to a method for identifying the cavity of a glass container forming machine in which a glass container has been manufactured. The method comprises manufacturing containers (1), wherein each container (1) has a code (C) which identifies the cavity in which it has been manufactured. A characteristic parameter is determined for each container (1), and the code (C) which identifies the cavity in which it has been manufactured is read; a classification of the containers (1) is performed which comprises relating each container (1) with its characteristic parameter and with the cavity in which it has been manufactured; a new container (1) is manufactured; the characteristic parameter of the new container (1) is determined; and the characteristic parameter of the new container (1) is compared with the classification for identifying the cavity in which the new glass container (1) has been manufactured without reading the code (C).