1. Introduction
2. Experiments
2.1 Materials
2.2 Welding Equipment and Process
3. Genetic Algorithm
3.1 Genetic Algorithm Optimization Technique
3.2 Selecting Deep Learning Hyperparameters
Table 3
Hyper parameters from previous research^{11)} | |
---|---|
Number of hidden layer | 4 |
Number of node in each layer | 24, 24, 24, 24 |
Learning rate | 0.01 |
Batch size | 32 |
Dropout rate | 0.5 |
3.3 Genetic Algorithm-based Optimization
4. Results and Observations
4.1 Signal Analysis and Feature Extraction
Table 4
4.2 Results of Optimization Using a Genetic Algorithm
Table 5
Variables | Value |
---|---|
Initial number of chromosomes | 256 |
Minimum number of chromosomes | 64 |
Maximum generation | 15 |
mutation probability | 15 % |
Table 6
Table 7
Table 8
Total test data | True | Error | accuracy | |
---|---|---|---|---|
DNN structure from previous research | 361 | 323 | 38 | 89.4737 |
Optimized GA- DNN structure | 336 | 25 | 93.0748 |
4.3 System for Real-time Use After Optimization
5. Conclusions
1) Feature variables were derived by measuring the arc voltage waveforms that occur during the GMAW process in real-time and performing data preprocessing.
2) The new genetic algorithm that was proposed by this study reduced calculation time by around 30% compared to before and increased the number of first- generation chromosomes, thus improving the diversity of the initial chromosomes.
3) The results of verifying the optimization results showed that the DNN model that was optimized by the genetic algorithm had a prediction accuracy of 93.1%, which was an increase of 3.60% compared to a previous study’s DNN model prediction rate.