In this paper, we develop a speech enhancement network specialized for a loud and noisy environment, and improve the efficiency and accuracy of factory work through noise control. The network uses a Deep Neural Network (DNN) to separate the target voice and transmit it to the operator. Two types of microphones are used in the proposed method: one to acquire the target worker’s voice and the other to acquire only the noises in the factory. The proposed method utilizes the existing real-time single-channel speech enhancement network “DEMUCS”, and extend it to multi-channel speech enhancement networks. Experiments demonstrate that the proposed speech enhancement network performs better than the conventional method.