Tunnel construction ventilation frequency-control based on radial basis function neural network

Abstract

Reasonable and effective ventilation plays an important role in the safety of tunnel construction. Fans are usually designed to be are capable of satisfying the maximum demand of a tunnel. During the tunnel construction, the actual usage of the tunnel could be considerably less than the designed fan capacity. This leads to high energy consumption and low efficiency. Therefore, a system that can analyze in real-time the tunnel environment and calculate the actual demand is required for tunnel construction. In this study, a tunnel ventilation intelligent frequency control (TVIC) system is designed based on the radial basis function neural network (RBF NN). As a type of feedforward neural network, RBF NN is used to obtain the relationship between the fan operating frequency and various pollutant concentrations, the tunnel length, and the temperature. TVIC is composed of a safety-monitoring system, control system, communication system, and variable-frequency drive (VFD) fan. It can self-adjust the frequency of the fan according to the construction environment inside the tunnel, and has been used in the Huayingshan tunnel in southwest China for a year and a half. In addition, it displays good reliability and a satisfactory capacity for tunnel environmental improvement and energy conservation. Compared with the current manual control method, ventilation system was observed to reduce electricity consumption by 42% after using TVIC.