縦流式道路トンネルにおけるオンライン学習形換気制御 Online-learning Type of Ventilation Control Systems in Longitudinal-flow Road Tunnels
The road-traffic flow causes tunnel vetilation load. It changes seasonally and year by year and so the tuning of tunnel ventilation systems takes a lot of time. With these points as background, this paper proposes a online-learning type of ventilation control system. This system uses two types of neural networks. One is NNVI (Neural Network for VI control) which learns the change of VI (Visibility Index) successively. The other is NNAF (Neural Network for Air-Flow control) which learns the change of air-flow successively. NNVI also outputs control instructions to keep VI higher than a permissible level in parallel with the above learning. NNAF also outputs the ventilation control instructions to keep air-flow higher than a permissible level so that air-flow does not turn over. This paper describes the construction of the proposed ventilation control system and the automatic learning method. And using the dynamic digital simulator it shows the estimation results of the control performance based on the road-traffic flow data.
- 電気学会論文誌. D, 産業応用部門誌 = The transactions of the Institute of Electrical Engineers of Japan. D, A publication of Industry Applications Society
電気学会論文誌. D, 産業応用部門誌 = The transactions of the Institute of Electrical Engineers of Japan. D, A publication of Industry Applications Society 117(8), 970-979, 1997-08
The Institute of Electrical Engineers of Japan