Manufacturing Quality Prediction Using Virtual Sensors and Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current quality assurance methods in industrial manufacturing are costly and time-consuming, particularly due to data insufficiency, poor data quality, non-measurable variables, and lengthy learning phases, which hinder the development of reliable predictive models for product quality assessment.
Innovation Solution
A method that combines simulative determination of manufacturing state variables through system simulation and sensory detection, using machine learning to associate these variables with product quality, allowing for the generation of predictive models even with limited data and improving data quality and correlation between input and output variables.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional quality assurance measures are implemented, then product quality can be ensured, but costs and time consumption increase significantly
Solution Approach 1:
The system performs preliminary quality assessment by simulating manufacturing processes and predicting quality outcomes before actual production. Virtual sensors calculate manufacturing state variables in advance, allowing quality issues to be identified and corrected beforehand, eliminating the need for time-consuming post-production inspection
Solution Approach 2:
Physical quality inspection mechanisms are replaced with a digital simulation and prediction system. The system uses virtual sensors, process simulations, and machine learning models to substitute traditional mechanical measurement and inspection methods, achieving faster and more efficient quality assessment
2Measurement precision
If sufficient data is collected for predictive modeling, then model accuracy improves, but data collection time increases
Solution Approach 1:
The system generates synthetic training data through process simulations before actual production begins. Virtual sensors calculate manufacturing state variables based on simulated process parameters, creating a comprehensive dataset in advance that enables immediate deployment of accurate predictive models without waiting for extensive real-world data accumulation
Solution Approach 2:
The system creates virtual copies of the manufacturing process and its data through simulation. Virtual sensors replicate the functionality of physical sensors by calculating manufacturing state variables from process data, generating synthetic datasets that mirror real production conditions and enable model training without requiring extensive physical data collection
3Measurement precision
If physical sensors are installed to measure all manufacturing state variables, then data quality improves, but system complexity and costs increase
Solution Approach 1:
Virtual sensors act as intermediaries between process parameters and quality outcomes. Instead of installing physical sensors for every manufacturing state variable, the system uses virtual sensors that calculate these variables from readily available process data through simulation models, bridging the gap with minimal additional hardware
Solution Approach 2:
Physical sensing systems are replaced with computational virtual sensors. The system substitutes mechanical and electronic sensor infrastructure with software-based calculations that derive manufacturing state variables from process data, dramatically reducing system complexity while maintaining data quality
4Measurement precision
If a long learning phase is allowed for model development, then predictive accuracy improves, but productivity during startup is reduced
Solution Approach 1:
The system completes the learning phase before production startup by generating synthetic training data through simulations. Virtual sensors provide comprehensive training datasets that enable model development in advance, allowing the system to skip the lengthy real-world learning phase and achieve full predictive accuracy immediately upon startup
Solution Approach 2:
The system uses simulated copies of the manufacturing process to accelerate learning. Virtual sensors generate synthetic training data that replicates real production conditions, enabling the model to learn from virtual examples without requiring extended periods of actual production data collection, thus maintaining high productivity from the start
Data Source
AI summary
Devices and methods for determining a product quality resulting from a manufacturing process are disclosed herein. In one example, the method includes simulatively determining one of a plurality of manufacturing state variables depending on a scattering of the manufacturing state variables, acquiring one of the manufacturing state variables by a sensor, and associatively determining the product quality depending on the manufacturing state variables.