Causal Inference Model for Manufacturing Software Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for probability calculation using Bayesian networks in manufacturing settings fail to effectively generate actionable application software for failure diagnosis and improvement.
Innovation Solution
A software generation method that constructs an expanded causal inference model from manufacturing log data and environment configuration information, contracts it to a targeted causal relation, and generates application software based on this model for specific diagnostic purposes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Bayesian network probability calculation is performed for failure diagnosis, then diagnostic capability is improved, but no actionable improvement measures are provided
Solution Approach 1:
The system implements feedback by automatically generating improvement measures based on diagnosis results. The software generation unit creates actionable software that feeds back to the manufacturing process, enabling continuous improvement cycles rather than just providing static diagnostic information.
Solution Approach 2:
The system performs self-service by automatically generating improvement software without requiring manual intervention. The software generation unit autonomously creates application software based on the diagnosis results, eliminating the need for manual analysis and software development.
2Adaptability or versatility
If manual software development is performed for each diagnostic requirement, then software functionality is improved, but development time and complexity increase
Solution Approach 1:
The system uses copying by generating software through template-based automatic creation. Instead of manually developing software from scratch, the software generation unit copies and adapts predefined software templates based on diagnosis results, significantly reducing development time while maintaining functionality.
Solution Approach 2:
The system implements universality by creating a multi-functional software generation platform that can produce various types of application software (improvement measures, countermeasures, optimization software) from a single diagnostic system, eliminating the need for separate development processes for each software type.
3Measurement precision
If comprehensive manufacturing data is collected for accurate diagnosis, then diagnostic accuracy is improved, but data processing complexity increases
Solution Approach 1:
The system applies extraction by isolating and focusing on only the necessary data elements required for specific diagnostic purposes. The causal inference model extracts relevant relationships from comprehensive manufacturing data, separating essential diagnostic information from unnecessary data complexity.
Solution Approach 2:
The system uses segmentation by dividing the comprehensive data processing task into modular components through the causal inference model. The model segments data processing into distinct causal relationships and dependencies, making complex data processing manageable and systematic.
Data Source
AI summary
The software generation method uses a computer, wherein the computer includes a control unit and a storage unit; the storage unit stores manufacturing log data that includes sensor data acquired in one or both of a manufacturing process and an inspection process, and environmental configuration information that relates to a manufacturing device or an inspection device from which the sensor data are acquired for each component or product; and the control unit reads the manufacturing log data from the storage unit, reads the environment configuration information from the storage unit, constructs a causal inference model based on the manufacturing log data, constructs an expanded causal inference model by expanding the causal inference model using the environment configuration information, generates a contracted model by contracting the expanded causal inference model to a causal relation of prescribed target data of interest, and generates prescribed application software by reading the contracted model.