Blower Performance Prediction Using 2D Blade Parameter Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Neural network models struggle to accurately predict blower performance due to the inability to learn local correlations between blade geometry parameters when inputting them as a one-dimensional array.
Innovation Solution
Convert one-dimensional array data into two-dimensional array data by replicating and arranging parameter values to enhance the learning capability of convolutional neural networks (CNNs) for improved performance prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If blade geometry parameters are input as a one-dimensional array to the neural network, then the input structure is simple, but the model cannot learn local correlations between parameters resulting in lower prediction accuracy
Solution Approach 1:
The patent transforms the one-dimensional array of blade geometry parameters into a two-dimensional array structure. This dimensional transformation enables the neural network to capture local correlations between parameters by creating a spatial structure where adjacent elements represent related geometric features, thereby improving prediction accuracy without significantly increasing overall complexity
Solution Approach 2:
The patent segments the blade geometry parameters into multiple channels or feature maps within the two-dimensional array structure. This segmentation allows different aspects of the blade geometry to be processed independently and then combined, enabling the model to learn local patterns and correlations more effectively
Data Source
AI summary
According to one embodiment, a performance prediction method includes a conversion process, a training process, and a prediction process. The conversion process converts one-dimensional array data including a plurality of parameter values arranged in one dimension into two-dimensional array data. The training process trains a machine learning model based on training data, and generates a trained model that takes two-dimensional array data as input and provides predicted performance data as output. The prediction process applies two-dimensional array data of a processing target to the trained model to generate predicted performance data for the processing target.