Particle swarm optimisation clustering for cement kilning system fault recognition
| Autor: | Yun-Xing Shu, Ke-Hong Yuan, Wei Wei, Deyun Wang |
|---|---|
| Rok vydání: | 2014 |
| Předmět: |
Cement
Engineering business.industry Feature extraction Particle swarm optimization Pattern recognition Machine learning computer.software_genre Industrial and Manufacturing Engineering Kernel principal component analysis Fault recognition Kernel (image processing) Control and Systems Engineering Principal component analysis Artificial intelligence business Cluster analysis computer |
| Zdroj: | International Journal of Industrial and Systems Engineering. 17:477 |
| ISSN: | 1748-5045 1748-5037 |
| DOI: | 10.1504/ijise.2014.063965 |
| Popis: | Cement kilning system is the most important issue in the cement production process. Its fault recognition is the key issue which influences the quality of cement and the safety of the production process. In traditional ways, the efficiency in fault recognition is low, and the lag time of recognition is long. To improve the efficiency of the fault recognition, we propose a hybrid method based on both kernel principle component analysis (KPCA) and modified particle swarm optimisation (MPSO). First, KPCA extracts the non-linear features from the high-dimensional samples, which can eliminate the redundant information. Second, through the method of taking the features extracted by KPCA as the input of MPSO clustering, the MPSO clustering can obtain the optimal results of the recognition for cement kilning system. The simulation results show the efficiency and precision of the proposed algorithm for the practical cement kilning system. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|