1993 Fiscal Year Final Research Report Summary
APPLICATION OF FUZZY THEORY FOR HYDRAULICS AND HYDROLOGY
| Project/Area Number |
04660257
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
農業土木
|
|---|
| Research Institution | KYUSHU UNIVERSITY |
|---|
Principal Investigator |
SHIKASHO Shiomo FACULTY OF AGRICULTURE, ASSOCIATE PROFESSOR, 農学部, 助教授 (80038265)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
HIRAMATSU Kazuaki FACULTY OF AGRICULTURE, ASSISTANT PROFESSOR, 農学部, 助手 (10199094)
MORI Ken FACULTY OF AGRICULTURE, ASSISTANT PROFESSOR, 農学部, 助手 (50117272)
TOHARA Yoshio FACULTY OF AGRICULTURE, PROFESSOR, 農学部, 教授 (80038182)
|
|---|
| Project Period (FY) |
1992 – 1993
|
| Keywords | Fuzzy Inference / Facilities Control / Fuzzy Linear Regression Analysis / Neural Networks / Numerical Experiments / Parameter Identification / Runoff Analysis / Water Quality Analysis |
|---|
| Research Abstract |
In runoff analysis of low-lying drainage basin, there are many problems, such as the prediction of hydrograph and the effective control of the drainage facilities. Therefore, we tried to establish these hydraulic and hydrological problems introducing Fuzzy Theory.
The results are summarized as follows :
(1).The Fuzzy Linear Regression Analysis (fuzzy LRA) was applied to identification of runoff model parameters. From the results, it can be concluded that the fuzzy LRA is an efficient and practical method as well as the other methods in the past, and that the merit of this formulation is to be able to obtain easily parameters by soiving the linear programming problem and to add another constraint conditions which might be obtained from expert knowledge of parameters.
(2).For the case of drainage facilities operation, we applied the method of 2 step fuzzy inference to the model drainage basin. It was concluded, from the simulations, that the operation of drainage facilities using fuzzy control theory is a practical method of the water management in low-lying basin.
(3).It has been important to develop the effective numerical methods to analyze flow through nonprismatic channel sections. At present from hydraulic and numerical experiments, we are studying for the computer codes served as convinient tools for these flow analyzes.
(4).Analyzes of river discharge and water quality are difficult to solve exactly by conventional methods of the mathematical programming. Then, we tried on application of a multilayred neural network to these problems. We confirmed effectiveness of the neural networks through several computational experiments.
|
