t-SNE and variational auto-encoder with a bi-LSTM neural network-based model for prediction of gas concentration in a sealed-off area of underground coal mines

A deep learning network is introduced to predict concentrations of gases in the underground coal mine enclosed region using various IoT-enabled gas sensors installed a metallic chamber. The air sucked automatically at specific intervals from sealed-off site utilizing solenoid valve, suction pump, and programmed microprocessor. monitor content communicate concentration surface server room through wireless cloud storage media. t-SNE_VAE_bi-LSTM model proposed this study as prediction that combines t-SNE, VAE, bi-LSTM networks. model's t-SNE method aims minimize dimensionality recorded concentration; VAE layer intends retrieve inner characteristics low-dimensional concentration. Finally, given Bi-LSTM tries forecast CH4, CO2, CO, O2, H2 gases. accuracy compared with existing two models, namely auto-regressive integrated average moving (ARIMA) chaos time series (CHAOS). experiment findings demonstrate forecasted mean square error (MSE) more accurate, it has lesser MSE value 0.029 0.069 for CH4; 0.037 0.019 CO2; 0.092 0.92 CO; 1.881 1.892 O2; 1.235 1.200 than ARIMA CHAOS respectively.