Intelligent Modeling and Forecasting for Breeding Pool Water Quality Assessment and Machine Condition Prediction
| Autor: | 呂碩欽 |
|---|---|
| Rok vydání: | 2017 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 105 In this paper, we hope that we can obtain the best forecast result for the fish pond water quality of the aquaculture fishery and the failure of the high power pulsed magnetron sputtering machine. We predicted the data by gray theory, exponential smoothing method and recurrent neural network, and explored its relevance. In the gray theory, the data is accumulated into more simply incremental values, and then obtain its parameters for predictions by the least squares method. The exponential smoothing method is a kind of prediction method which does not discard the past data but gives a gradually reduced degree of influence. Finally, the Elman neural network, which is a model of simulating biological nerve conduction, is a kind of recursive neural network, which is divided into three parts: structure, incentive function and learning rule. In this study, we apply the above three methods to the fish pond water quality of the aquaculture fishery and the failure of the high power pulsed magnetron sputtering machine. Compared with the results of the experiment, in order to select the appropriate method for the above two cases, gray prediction method is more suitable for fish pond water quality assessment. In this case, its ability is more excellent than other method, although some backward situation. We are predicted by elman neural network for the malfunction prevention of the machine. This method has less error than the other two methods. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|