Inkjet Printing Parameter Recommendation via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inkjet printing systems face challenges in adjusting print parameter values efficiently, particularly for operators without experience, leading to time-consuming adjustments and resource wastage, as existing methods rely heavily on operator experience and are limited to specific printing conditions and media types.
Innovation Solution
A printing system equipped with a machine learning-based recommended candidate value search model that evaluates and outputs optimal print parameter values, such as conveyance speed and drying temperature, based on input printing conditions, allowing for easy and experienced-independent adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If parameter values are adjusted on the basis of operator experience, then printing quality can be optimized for experienced operators, but adjustment time increases significantly and resources are wasted for inexperienced operators or new printing conditions
Solution Approach 1:
The patent replaces the mechanical system of manual parameter adjustment based on operator experience with an automated information processing system. The management server automatically determines optimal parameter values by retrieving printing condition data, comparing it with stored reference data, and outputting recommended parameters, thereby eliminating the need for operators to manually adjust parameters based on experience.
Solution Approach 2:
The system enables self-service by allowing the management server to autonomously determine and recommend optimal parameter values without requiring operator expertise. The server automatically processes printing condition information, retrieves relevant reference data, and generates parameter recommendations, making the system self-sufficient in providing optimized printing parameters.
2Manufacturing precision
If parameter adjustment is performed through repeated test printing, then optimal printing quality can be achieved, but ink and print media are wasted
Solution Approach 1:
The patent applies preliminary action by determining optimal parameter values before actual printing occurs. The management server retrieves printing condition data, compares it with stored reference data from similar printing conditions, and outputs recommended parameter values in advance, eliminating the need for repeated test printing and preventing waste of ink and print media.
3Device complexity
If a single model outputs recommended parameter values directly, then the system is simpler, but it cannot handle diverse printing conditions and media types effectively
Solution Approach 1:
The patent segments the parameter recommendation system into two distinct functional modules: a reference data generation unit that creates reference data from historical printing information, and a recommended value output unit that uses this reference data to determine optimal parameters for new printing conditions. This segmentation allows the system to handle diverse printing conditions effectively while maintaining manageable complexity.
4Ease of manufacture
If fabric data and ink data are used as input for parameter recommendation, then recommendations can be generated for fabric printing, but the system cannot be applied to other print media types
Solution Approach 1:
The patent achieves universality by designing a reference data generation unit that can process various types of printing condition information beyond just fabric data and ink data. The system is configured to handle diverse input data types including print medium information, enabling it to provide parameter recommendations for multiple media types such as paper, fabric, and other materials, making the system multi-functional and broadly applicable.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A recommended value output unit (116) is provided in a print control device (110) included in an inkjet printing apparatus, the recommended value output unit (116) including: a recommended candidate value search model (61) that is constructed on the basis of print result logs (6) transmitted from a large number of inkjet printing apparatuses (10) by a management server (20), and obtains a plurality of recommended candidate values (63) for a print parameter on the basis of an input printing condition (62); and a recommended candidate value evaluation unit (1161) that selects output targets as a plurality of recommended values (64) from the plurality of recommended candidate values (63) by evaluating each of the plurality of recommended candidate values (63).