Rubber Device Property Determination Using Vector-Compressed Neural Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining technical properties of rubber devices, such as vehicle tires, require high computing power and memory usage, leading to inefficient use of computer resources.
Innovation Solution
The method involves representing the condensed data set in a real vector space and assigning vector groups to different rubber device types using a processing device, allowing for faster calculations through trivial vector calculus in linear algebra, thereby optimizing computer resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods are used to determine technical properties of rubber devices, then accurate property determination is achieved, but high computing power and memory usage are required
Solution Approach 1:
The patent creates a virtual copy of the rubber device through neural network simulation, generating virtual measurement data that replicates physical testing without requiring actual computational heavy processing. The autoencoder-compressed data set serves as a simplified representation that preserves essential characteristics while enabling efficient analysis.
Solution Approach 2:
The patent transforms the problem by changing parameters from direct physical measurement to neural network-based virtual measurement. The autoencoder compresses the input data into a condensed representation, and the neural network generates virtual measurement data with different statistical characteristics than physical measurements, achieving accurate property determination with reduced computational demands.
2Reliability
If comprehensive technical data is processed to determine rubber device properties, then accurate predictions are achieved, but processing time and computational complexity increase
Solution Approach 1:
The patent performs preliminary compression of the technical data set using an autoencoder before the actual property determination process. This pre-processing step condenses the comprehensive technical data into a condensed data set that retains essential information while reducing the dimensionality and processing requirements for subsequent neural network analysis.
Solution Approach 2:
The patent divides the comprehensive technical data into distinct segments: the condensed data set generated by the autoencoder and the virtual measurement data generated by the neural network. This segmentation allows each component to be processed independently and efficiently, with the condensed data set serving as optimized input for the property prediction model.
3Measurement precision
If physical experimentation is conducted to determine rubber device properties, then empirical accuracy is achieved, but resource consumption and environmental impact increase
Solution Approach 1:
The patent replaces physical experimentation with virtual experimentation using neural networks. The system generates virtual measurement data that mimics the statistical characteristics of physical measurements without requiring actual physical tests, thereby eliminating the energy consumption and environmental impact associated with physical experimentation while maintaining measurement accuracy.
Solution Approach 2:
The patent substitutes the mechanical and physical testing system with an information-processing system based on neural networks and autoencoders. Instead of physically measuring rubber device properties through experimentation, the system processes technical data through computational models that replicate the measurement process virtually, achieving the same scientific objective with minimal environmental impact.
Data Source
Figure 1
AI summary
The invention relates to a method for determining a technical property of a rubber device (9), comprising the following steps: providing a device (2) with an artificial neural network (3), comprising an input layer (4) and an output layer (5).