Dynamic Modeling of Flue Gas Desulfurization Process via Bivariate EMD-Based Temporal Convolutional Network

Sulfur dioxide (SO2) can cause detrimental impacts on the ecosystem. It is well known that coal-fired power plants play a dominant role in SO2 emissions, and consequently industrial flue gas desulfurization (IFGD) systems are widely used plants. To remove effectively such ultra-low emission standard be satisfied, IFGD modeling has become urgently necessary. chemical process with long-term dependencies between time steps, it typically exhibits strong non-linear behavior. Furthermore, rendered non-stationary due to frequent changes boiler loads. The above-mentioned properties make truly formidable problem, since chosen model should have capability of learning dependencies, dynamics processes simultaneously. Previous research this area fails take all above points into account at time, calls for novel approach so satisfactory performance achieved. In work, bivariate empirical mode decomposition (BEMD)-based temporal convolutional network (TCN) proposed. our approach, BEMD employed generate relatively stationary processes, while TCN, which possesses memory ability uses dilated causal convolutions, serves each subprocess. Our method was validated using operating data from system station China. Simulation results show yields desirable performance, demonstrates its effectiveness dynamic problem.