| Project/Area Number |
07555436
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (A)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 試験 |
|---|
| Research Field |
土木材料・力学一般
|
|---|
| Research Institution | University of Tokyo |
|---|
Principal Investigator |
UOMOTO Taketo University of Tokyo Institute of Industrial Science Prof., 生産技術研究所, 教授 (80114396)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
WATANABE Tadashi University of Tokyo Maeda Corporation, 主任研究員
KATO Yoshitaka University of Tokyo Institute of Industrial Science Assistant Lecture, 生産技術研究所・第5部, 助手 (80272516)
URA Tamaki University of Tokyo Institute of Industrial Science Prof., 生産技術研究所・第2部, 教授 (60111564)
|
|---|
| Project Period (FY) |
1995 – 1996
|
| Project Status |
Completed (Fiscal Year 1996)
|
| Budget Amount *help |
¥6,000,000 (Direct Cost: ¥6,000,000)
Fiscal Year 1996: ¥6,000,000 (Direct Cost: ¥6,000,000)
|
|---|
| Keywords | OPTIMIZING METHOD / NEURAL NETWAORK / CONCRETE PLANT / QUALITY CONTROL / 感度解析 / 季節変動 |
|---|
| Research Abstract |
This study is purposed to develop a management system. This system is easily applied to actual concrete plant and can do quality control much more accurately than conventional system. Neural network is introduced as a method for developing the management system and the possibility of optimized mix proportion in actual concrete plant using this method is verified. The results obtained in this study are as follows.
(1) Prediction method of concrete quality using neural network is found to be valid by the results of laboratory scale experiment. In particular, surface moisture can be predicted with high accuracy by measured weight of materials in mixer and measurement of change in mixing energy. The surface moisture was difficult to measure before this method.
(2) This management system was applied to actual concrete plant producing ordinary concrete and the prediction of concrete quality was found to be difficult. The reason was found to the change of concrete quality. Concrete quality is changed by the fluctuation of the material quality and temperature from season to season in actual plant. From these findings, the system is revised to a new system applicable in whole year by adopting parameters expressing seasonal fluctuation. Optimizing program has the limit in amount and range of data so that the program is found to diverse beyond this limit. This is due to the nature of neural network which is the core of this method and it is found that acquisition of wider data is needed.
(3) The revised system is applied to actual concrete plant for dam. The possibility of production management of concrete is confirmed on basic level with some problems remaining.
(4) This system can be refined to practical use level by applying and investigating repeatedly the system in actual plant in future.
|
