Processing Result Estimation Model Using Inline Intermediary Data

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for generating models that estimate processing results face challenges when measurement data for certain parameters is not collected inline, leading to increased experimental requirements and inefficiencies.

Innovation Solution

A method that involves performing experiments to acquire type 1 and type 2 information for processing conditions, and type 3 and type 4 information for processing results, deriving relations between these information types, and generating a model that estimates type 4 information using type 2 and type 3 information as inputs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If measurement data for certain parameters is not collected inline, then experimental requirements increase, but model estimation accuracy deteriorates

Engineering Contradiction:

Improvemodel estimation accuracyVSAvoidexperimental requirements

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces type 3 information (intermediary variable) as a mediator between type 2 information (measurable input) and type 4 information (target output). By deriving relations through this intermediary, the system can estimate parameters without direct inline measurement, reducing experimental complexity while maintaining accuracy.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces direct physical measurement (mechanical/data collection system) with computational derivation. Instead of collecting type 4 information directly through measurement, the system substitutes this with mathematical derivation from type 2 and type 3 information, reducing the need for complex measurement infrastructure.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If direct measurement of all parameters is performed, then model accuracy improves, but measurement time and cost increase

Engineering Contradiction:

Improveparameter measurement accuracyVSAvoidmeasurement time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the necessary measurement data (type 2 and type 3 information) that can be obtained inline, and derives the remaining parameters (type 4 information) computationally. This selective extraction approach reduces measurement time while maintaining sufficient accuracy for the estimation model.

Inventive Principle:

Principle #2Taking out (Extraction)

3Ease of manufacture

If polynomial model fitting is used with accumulated data, then physical interpretability decreases, but ease of construction improves

Engineering Contradiction:

Improvemodel construction easeVSAvoidphysical interpretability

Core Design Contradiction:

Ease of manufactureVSLoss of information

Solution Approach 1:

The patent transforms the modeling approach by changing from direct polynomial fitting to a relational derivation approach using type 1, type 2, and type 3 information. This parameter transformation maintains mathematical simplicity for easy construction while restoring physical interpretability through the meaningful relationships between processing conditions and results.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS12223237B2Generation method, estimation method, generator, and estimator

Publication Date: 2025.02.11 PANASONIC INTELLECTUAL PROPERTY MANAGEMENT CO LTD

AI summary

An experiment of processing a device is performed to acquire type 1 information and type 2 information indicating processing conditions, and type 3 information and type 4 information indicating results of the processing, derive a first relation between the type 1 information, the type 2 information, and the type 3 information, and a second relation between the type 1 information, the type 2 information, and the type 4 information, and generate and output a model that estimates the type 4 information indicating a result of the processing by using the first relation and the second relation with the type 2 information and the type 3 information that are measured during the processing as inputs.