Flue gas emission is an inevitable procedure in the course of electricity generation, which would pose a severe threat to human health, and has an adverse effect on our environment. Due to the fact that the environment in practical flue gas desulfurization system fluctuates frequently, system parameters tend to vary constantly during the operating process, thus control performance with traditional strategies tends to be suboptimal in most cases. To address this problem, some insight into operating conditions must be gained prior to taking proper control strategy. Therefore, in this paper, based on actual measurements in 1000 MW Unit Wet Limestone FGD System for a coal-fired power plant, a kind of intelligent operating condition partition method is combined with the multi-model adaptive control strategy. Specifically, analysis and partition of operating condition is carried out in the first place, then adaptive multi-model control is implemented with the combination of parallel dynamic neural network and partition results. Additionally, the applicability of proposed control mode is investigated through different simulation examples. At the same time, to further enhance the flexibility of multi-model control structure, some possible improvements on it is also discussed.

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

Sulphur dioxide, as one of the most common air pollutant gases, brings considerable numbers of hazards on human health and environment. For the purpose of reducing the detrimental effect it brings, it is of urgent necessity to control emissions of flue gas in power plants, since a substantial proportion of sulphur dioxide in the atmosphere stems from flue gas generated in the whole process of electricity generation. However, the complexity and nondeterminism of the environment increase the occurrences of anomalies in practical flue gas desulphurization system. Anomalies in industrial desulphurization system would induce severe consequences and pose challenges for high-performance control with classical control strategies. In this article, based on process data sampled from 1000 MW unit flue gas desulphurization system in a coal-fired power plant, a multimodel control strategy with multilayer parallel dynamic neural network (MPDNN) is utilized to address the control problem in the context of different anomalies. In addition, simulation results indicate the applicability and effectiveness of the proposed control method by comparing with different cases.

SO₂ emissions are known to pose great harm to both human health and atmospheric air, and flue gas generated from coal-fueled power plant is the prime source of sulfur dioxide. For this reason, flue gas desulfurization (FGD) technology has found wide applications in most coal-fired power stations. Correctly describe the dynamic behavior of an FGD process is the precondition of controlling it effectively. However, FGD process modeling is by no means an easy task, as the underlying process dynamics are highly nonlinear in nature, meanwhile time-delay effect is significant therein. Long short-term memory (LSTM) network possesses remarkable long-term memory capability, hence it is anticipated to have a powerful identification capability. In this paper, the connection between deep learning and system identification is established, further a unidirectional/bidirectional LSTM deep network is designed and employed to identify a real FGD process. Simulation results clearly demonstrate the effectiveness of deep learning-based identification approach, and the superiority of deep LSTMs over other conventional identification models is also verified.

Under deep peak-shaving operation, coal-fired unit coordination systems exhibit strong multivariable coupling, large inertia, and significant uncertainties, posing challenges to decentralized control. Conventional decoupling methods rely on accurate models, while existing active disturbance rejection control (ADRC) schemes are mainly designed for single-input single-output systems and have limited capability in suppressing inter-loop coupling. To address this issue, this paper proposes a coupling-improved ADRC (CI-ADRC) strategy. By introducing cascaded coupling disturbance observers (CDOs), coupling disturbances are explicitly estimated and compensated in a feedforward manner, achieving effective decoupling without relying on explicit coupling-channel models, in contrast to conventional decoupling approaches. A unified quantitative tuning scheme is developed to simplify multivariable controller design, and closed-loop stability is analytically established. Simulation and experimental results on steam–steam two-input two-output and molten-salt thermal storage integrated systems demonstrate that CI-ADRC achieves superior dynamic performance and disturbance rejection compared with PID and IADRC. Taking the molten-salt thermal storage system as an example, in terms of the integral of absolute error (IAE), compared with PID and IADRC, reductions of 90.68%/83.90% (load), 83.79%/63.46% (main steam pressure), 52.20%/32.27% (intermediate temperature), and 98.10%/93.32% (molten salt heat absorption) are obtained. Moreover, Monte Carlo tests with ±10% parameter perturbations confirm enhanced robustness. Additional simulation and experimental results, as well as validation on a dual-loop temperature control laboratory and engineering implementation on a 1000 MW supercritical unit, further verify the feasibility and practical applicability of the proposed method.