Container Treatment Plant Control Using Minor Defect Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing container treatment plants fail to consider valuable information from containers with minor defects, leading to inadequate treatment and potential quality issues.
Innovation Solution
Implement a method using machine learning-based container inspection models to evaluate spatially resolved sensor data, including a set of container part features, for precise control and regulation of treatment processes, and a similarity analysis with reference data to detect subtle defects and plant anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a threshold-based defect evaluation method is used to classify containers as defect-free or defective, then the inspection process is simple and fast, but valuable information from containers with minor defects is lost and cannot be utilized for process optimization
Solution Approach 1:
The patent segments the defect evaluation into multiple levels: a first evaluation based on threshold comparison for quick classification, and a second evaluation that segments containers with minor defects for further analysis. This hierarchical segmentation allows fast processing of clearly defective containers while preserving detailed information from borderline cases for process optimization.
Solution Approach 2:
The patent changes the evaluation parameter from a binary defect-free/defective classification to a multi-level assessment that includes a defect severity value. This parameter transformation enables the system to distinguish between critical defects requiring rejection and minor defects providing valuable process information, resolving the contradiction between speed and information retention.
2Ease of operation
If only defect-free containers are considered for process control, then the evaluation is straightforward, but the treatment plant cannot optimize based on subtle defects that indicate process deviations
Solution Approach 1:
The patent implements feedback by using sensor data from containers with minor defects to provide information about process deviations. This feedback loop allows the treatment plant to adjust operating parameters based on subtle defects, improving process control accuracy while maintaining evaluation simplicity through automated multi-level assessment.
Solution Approach 2:
The patent performs preliminary evaluation of all containers using threshold-based methods to identify defect-free containers quickly. Containers with minor defects are then preliminarily flagged for secondary evaluation, allowing the system to maintain simplicity for the majority of containers while ensuring reliability through additional analysis of borderline cases.
3Loss of information
If all containers undergo detailed analysis, then no information is lost, but the processing time and computational resources increase significantly
Solution Approach 1:
The patent applies partial action by performing detailed analysis only on containers with minor defects that fall within a specific severity range, rather than analyzing all containers equally. This selective approach preserves complete information for relevant cases while avoiding unnecessary processing time for clearly defective or obviously perfect containers.
Solution Approach 2:
The patent segments the container stream into different evaluation groups: containers with clear defects (fast rejection), containers with minor defects (detailed analysis for process optimization), and defect-free containers (standard processing). This segmentation enables the system to allocate processing resources efficiently, providing complete information where needed while minimizing overall processing time.
Data Source
AI summary
To carry out a container inspection task, a sensor device optically detects sensor data and camera images relating to the container parts, and a real-time evaluation device evaluates spatially resolved sensor data in real time using a machine learning container inspection model which includes a set of parameters which are set to values which were learned as a result of a machine learning method. A set of container part features based on a machine learning method is predetermined, and the detected, spatially resolved sensor data are evaluated in relation to a plant inspection task different from the container inspection task, based on the predetermined set of container part characteristics, a plant inspection variable being determined depending on the inspection result of the carried out plant inspection task, which is provided for the control and/or regulation of the container treatment plant.