JaLCDOI 10.18926/14148
フルテキストURL Mem_Fac_Eng_39_1_28.pdf
著者 石丸 和史| 小西 正躬| 今井 純| 西 竜志|
抄録 In reactor furnace, due to high temperature and high pressure, data can be measured only near the furnace wall. In this paper, the way to estimate temperature distribution in a reactor furnace using measured data near the furnace walls and to control temperature distribution to the desired temperature distribution was studied. In the estimation, SQP method is employed using measured data near the furnace walls. As the result, the whole temperature distribution in a furnace could be obtained from such limited data. Furthermore, to control the temperature distribution in a reactor furnace, gas flow from multiple tuyeres and supplying material for controlling temperature distribution in a reactor furnace were determined by the SQP method. It was shown that temperature distribution in a furnace was regulated to achieve various desired distribution. Thus, it was verified that complicated temperature distribution in a reactor furnace could be controlled by combining furnace simulation and SQP method.
出版物タイトル Memoirs of the Faculty of Engineering, Okayama University
発行日 2005-01
39巻
1号
開始ページ 28
終了ページ 39
ISSN 0475-0071
言語 English
論文のバージョン publisher
NAID 120002308270
JaLCDOI 10.18926/14146
フルテキストURL Mem_Fac_Eng_39_1_21.pdf
著者 鳥越 崇史| 小西 正躬| 今井 純| 西 竜志|
抄録 In this paper, neuro based intelligent diagnosis methods for electro-mechanical control system are proposed. A self organizing map neural network (SOM) is used to classify measured data of the target system as a qualitative diagnostic method. Besides of the above procedure, it is expected to attain more efficient maintenance by a quantitative estimation of failure. For the purpose, new method is proposed using a hierarchical neural network (HNN). In the method, classified results by SOM are processed for the quantitative diagnosis. Hierarchical neural network can identify inner structure of the relations between failure causes and its results that enables a quantitative diagnosis.
出版物タイトル Memoirs of the Faculty of Engineering, Okayama University
発行日 2005-01
39巻
1号
開始ページ 21
終了ページ 27
ISSN 0475-0071
言語 English
論文のバージョン publisher
NAID 120002308342
JaLCDOI 10.18926/14099
フルテキストURL Mem_Fac_Eng_OU_40_1_23.pdf
著者 菊池 純二| 小西 正躬| 西 竜志| 今井 純|
抄録 In steel works, iron ores are stored in ore yard and sent to various plants of down stream accoding to transfer requests. To attain stable operation, it is neccesary to keep a certain allowable stock level in ore tanks of all plants. To this purpose, ore transfer routing method has been developed based on decentralized agent method. In case of disaster such as a big earthquake or a big fire, damages in the facilities of industrial complex may be unavoidable. In this paper, decentralized optimization method is tested to cope with such emergencies. Decentralized agents corresponding to kind of ore make their own transfer route plan exchanging information with others. As the application of the proposed method, transfer scheduling in ore yard in steel works are made in case of destruction of transfer facilities.
出版物タイトル Memoirs of the Faculty of Engineering, Okayama University
発行日 2006-01
40巻
1号
開始ページ 23
終了ページ 30
ISSN 0475-0071
言語 English
論文のバージョン publisher
NAID 120002308160
JaLCDOI 10.18926/14098
フルテキストURL Mem_Fac_Eng_OU_40_1_16.pdf
著者 阿部 快洋| 小西 正躬| 西 竜志| 今井 純|
抄録 In this study, auto tuning of PID control gains in hot strip looper controller is made based on RNN model. Neuro emulator is employed to model the characteristics of looper dynamics. Combining neuro emulator and RNN model, auto tuning system of PID control gains is constructed. As the inputs to RNN, plural evaluation functions which reflect individual preference of human experts. Further, Self learning mechamism is embeded to RNN model which enables adaptation to the change in rolling chracteristics. Through numerical experiments, the effect of the proposed method is ascertained.
出版物タイトル Memoirs of the Faculty of Engineering, Okayama University
発行日 2006-01
40巻
1号
開始ページ 16
終了ページ 22
ISSN 0475-0071
言語 English
論文のバージョン publisher
NAID 120002308615
JaLCDOI 10.18926/14097
フルテキストURL Mem_Fac_Eng_OU_40_1_9.pdf
著者 貫名 高史| 小西 正躬| 西 竜志| 今井 純|
抄録 Controls of the temperature, pressure and flowing quantity are important for the stable operation of the product quality in the distillation tower. The usual measuring way of product quality estimation is made by the off-line analysis. In this paper, online estimation method of product quality is studied for improving the product quality. The estimation method based on stochastic analysis was developed for online estimation. In this paper, the data of temperature, pressure and flow volum in the distillation tower are treated. As the estimation models, RNN (Recurrent Neural Net Work) and PLS (Partial Least Square Regression Method) were adopted. The actual plant data were used in the analysis. Both PLS and RNN models could compensate each other to improve the accuracy in estimation.
出版物タイトル Memoirs of the Faculty of Engineering, Okayama University
発行日 2006-01
40巻
1号
開始ページ 9
終了ページ 15
ISSN 0475-0071
言語 English
論文のバージョン publisher
NAID 120002308095