This article focuses the denitrification processes control of selective catalytic reduction, which faces great challenges caused by high-order dynamics, strong nonlinearity, wide-range load variations, and multisource disturbances. A modified active disturbance rejection control based on gain scheduling (MADRC-gs) is proposed. A parameter-switching methodology with the selected scheduling parameter is provided and the visual analyses of performance guarantees are analyzed. MADRC-gs is applied to a practical denitrification process of an in-service 660MW power plant. Actual operational data illustrate that MADRC-gs can reduce hourly average integral absolute error by about 22.61% and 43.18% in high-load and low-load ranges, respectively. Even though the power plant experiences lifting and lowering powers frequently, MADRCgs can still obtain better control performance compared to the original controller and MADRC, and show a promising application potential in energy and chemical industries.

Since that time delay is dominant in most industrial systems, it can reduce the stability and control performance of the closed-loop system. This article proposes a delay-inertia-compensation active disturbance rejection control (DIC-ADRC) for time-delay dominant industrial systems, where these systems can be approximated as second-order plus time-delay (SOPTD) systems. Different from a regular pure-delay compensation part, a compensation part combined with a pure-delay and first-order inertia is added to the control signal before it is sent to the extended state observer, where the SOPTD system can be approximated as a combination of a low-frequency integrator and a pure-delay. Quantitative tuning rules are derived to facilitate the application of DIC-ADRC, allowing controller parameters to be computed from the model and adjusted as needed. A stability theorem is also proposed and rigorously proven, ensuring closed-loop stability when its conditions are met. Comparative simulations and dual-tank experiments demonstrate that the proposed control method exhibits strong robustness and satisfactory control performance, while also validating the effectiveness of the DIC-ADRC quantitative tuning rules. Finally, further validation of DIC-ADRC is conducted through field application on the main steam pressure process. Compared to the original controller, DIC-ADRC demonstrated superior control performance. The results from both simulations and experiments reveal that DIC-ADRC has excellent control performance on time-delay systems and can be widely used in real industrial systems.

Circulating fluidized bed boilers (CFBBs) are widely applied in energy, metallurgy, the chemical industry and other fields, mainly due to their high combustion efficiency and low pollution emissions. However, the CFBB combustion system, as a typical third-order plus time delay (TOPTD) system, has inherent characteristics: large inertia, significant time delays and strong coupling. Coupled with the difficulty in establishing an accurate mathematical model, traditional control methods struggle to achieve the desired control performance. Active disturbance rejection control (ADRC) has prominent advantages, such as low dependence on the controlled plant’s accurate model and strong disturbance rejection ability, but it has obvious limitations in dealing with systems with large inertia and large time delays. To address this problem, this paper proposes an input-compensated active disturbance rejection control (ICADRC) method. An input-compensated part composed of a second-order inertial link and a time delay link is introduced into the ESO input channel, which is specially optimized for the characteristics of TOPTD systems. A set of quantitative parameter tuning rules unique to ICADRC is established via the equivalent approximation method, and a dedicated MATLAB auto-tuning toolbox for ICADRC is developed for TOPTD systems. Simulation experiments are conducted on the CFBB combustion system, and the results show that the proposed ICADRC exhibits superior setpoint tracking performance, disturbance rejection performance and robustness compared with ADRC, DADRC, and SIMC-PI. Under nominal operating conditions, the ${\mathrm{IAE}}_{\mathrm{sum}}$ of ICADRC is reduced by 36.2% relative to DADRC and by 54.3% relative to SIMC-PI. Specifically, under fixed parameter perturbations, the variation amplitude of ICADRC’s performance index is only 2.1%, significantly lower than the 5.1% for DADRC, 6.1% for ADRC, and 7.3% for SIMC-PI.

Day-ahead dispatch of integrated energy systems (IESs) is commonly solved by mixed-integer linear programming when objectives, constraints, and forecasts are fully specified. In practice, however, operators often express temporary preferences in natural language, and day-ahead forecasts inevitably deviate from realized operation. These two conditions make a fixed optimization interface difficult to use without additional modeling effort. This paper proposes DARC (Decompose-Act-Repair-Critique), a constraint-diagnostic LLM agent framework for day-ahead dispatch of campus-level microgrid-type IES with under-specified operating requirements. DARC combines three language modules, namely a Decomposer for temporal structure, a Resolver for numerical schedule generation, and a Critic for root-cause diagnosis, with deterministic repair, constraint checking, and metric evaluation. In all main experiments, the Decomposer is instantiated by its deterministic rule-based variant for reproducibility, so the reported results reflect a loop with two LLM modules (Resolver and Critic). The Projector repairs candidate schedules where possible, while the Checker supplies mathematical facts that ground the Critic’s feedback. DARC does not model forecast uncertainty through sets or scenarios; robustness to forecast deviation is pursued operationally, through margin-aware repair and checker-grounded revision, and is assessed empirically. Experiments on a campus-level microgrid IES testbed show that DARC can incorporate natural-language preferences, improve realized feasibility under forecast noise relative to both a point-forecast MILP and an interval-robust MILP baseline in terms of strictly feasible configurations, at higher operating cost, and produce diagnostic feedback that is robustly more useful than an ungrounded LLM diagnoser under judge models from three families, and at least as useful as checker facts alone, with its advantage concentrated in actionable revision guidance. DARC is therefore positioned not as a replacement for MILP, robust formulations, or model predictive control on fully specified problems, but as a complementary interface for operational settings where preferences and forecasts are not completely formalized.