The early faults of sliding bearings are highly concealed. To accurately predict their vibration amplitude, a deep learning model incorporating a YOLOv8-optimized CBAM attention mechanism is proposed. The CBAM module is embedded between the Backbone and Neck to enhance the model’s focus on critical vibration features. Additionally, an improved complete intersection over union loss function is employed to enhance object detection accuracy. Considering the nonlinear and non-stationary characteristics of vibration data, the empirical mode decomposition (EMD) method is integrated into the model to improve the accuracy of vibration state prediction. The experimental results show that, on the 600 MW steam turbine operation dataset, this method improves the detection accuracy by 2.85 percentage points and 8.50 percentage points compared with that of the conventional YOLOv8 and YOLOv7, respectively. Moreover, the root mean square error (RMSE) is reduces, and the mean absolute error (MAE) decreases. Furthermore, in high-noise environments, the model’s error fluctuation reduces by 30% compared with that of the conventional methods, demonstrating stronger generalization ability and stability.

With the increasing demand for flexible operation of power plant boilers, frequent variable-load operation leads to a wide range of fluctuations in pollutant concentrations and flue gas parameters. Modeling of key indicators such as single pollutant or flue gas parameter can no longer meet the actual demand, so it is necessary to consider the coupling of multiple key indicators for synergistic predictive modeling. Based on the historical operation data of coal-fired power plants, feature extraction is performed through kernel function mapping, and a long short-term memory neural network with a hard parameter sharing structure is constructed for multi task prediction modeling. The prediction model is optimized using uncertainty loss methods. The experimental results show that, the proposed prediction model exhibits high prediction accuracy under variable load conditions, and the prediction errors for the key metrics involved in this study are reduced by 25.5%, 41.8% and 4.7%, respectively. The proposed method is capable of predicting several key indicators of utility boilers under variable load conditions, which can assist power plants to achieve pollution control and optimize the thermal efficiency of combustion, and provide technical support for intelligent operation of power plants.

To solve the problem of severe ash accumulation and slagging on heating surface of boilers caused by a large proportion of blended economic coal, based on the close relationship between the ash fouling layer and the flue gas flow field parameters, the concept of cross-sectional “ash fouling characteristic field” is proposed, and a new intelligent soot blowing control system for boilers is developed, which includes functions such as characteristic field detection and generation, and benchmark field prediction. By comparing the difference in “drop value” and “concentration” between the benchmark feature field and the current feature field, the system can timely and accurately determine the appropriate blowing time, achieving “intelligent perception and on-demand blowing”. The new system solves the problem of lack of measurement points and low accuracy in existing model calculation methods, overcomes the disadvantage of high equipment cost in furnace observation methods, and uses on-site full section data collectors combined with intelligent prediction models for ash pollution characteristic fields to achieve low-cost and high-precision detection of ash and slag accumulation, effectively solving the problems of over blowing and under blowing. The actual application effect of the power plant shows that, after the new system was put into use for 3 months, the monthly blowing frequency decreased by 19.6%, and the monthly blowing steam consumption decreased by 229.0 tons, which is equivalent to a direct economic benefit of 284 000 yuan per year. In addition, the system also brings multiple indirect benefits, such as avoiding sudden coking that causes the unit to stop, extending the service life of the heating surface, and avoiding delayed soot blowing that leads to a decrease in boiler efficiency. The relevant control optimization experience can be used as a reference for similar units in the future.

The porous media methodology is used to describe the filter bags and their surface ash layer seepage flow. On this basis, the discrete particle model (DPM), as implemented in Fluent software, is employed to simulate the dynamics of dust particles movement and deposition in a baghouse filter. The effects of dust particle size on evolving morphology of the cake on the filter bags are investigated, and the influence of additional resistance of the filter cake on subsequent ash particle deposition is explored. The results show that, when the duration of dust removal is sufficiently long, the cake thickness and pressure loss in the baghouse filter increase in a linear fashion over time. For a constant dust mass flow rate, the pressure loss becomes higher with smaller dust particles. The smaller ash particles are transported more effectively by the upward airflow to the top of the vertical filter bag, whereas the larger particles, due to their weight, tend to settle in the middle area. This results in significant differences in the distribution of cake thickness. In contrast to conventional algorithms that neglect cake resistance, the proposed model, which incorporates this resistance through advanced development in commercial software, predicts a more uniform distribution of dust thickness, which is more consistent with the actual situations. This conclusion can provide references for improving the operational efficiency of baghouse filters.

To enhance the cybersecurity protection capabilities of power monitoring systems, a security reinforcement middleware for interal unidirectional safety isolating device for electric power has been designed. This middleware integrates compatibility adaptation, file format correction, encryption authentication, load balancing, and access control functions, addressing the security issues such as business system compatibility, hardware failures, and plaintext communication faced by isolation devices during the upgrading and reinforcement process. It enhances the security control of data transmission channels in power monitoring systems and achieves an “efficient and unobtrusive” and “standardized” security upgrade and reinforcement of the isolation devices. This middleware has been successfully applied to all thermal power, hydropower, and new energy power stations of Huaneng Group, strengthening the cybersecurity boundary protection capabilities of critical information infrastructure in power monitoring and ensuring the information security of power production.

The experimental data and simulation studies on ammonia co-combustion in coal-fired power plants in recent years are investigated, with a focus on the effect of ammonia fuel on NO_x generation. Through experimental observation, numerical simulation, and theoretical analysis, the spatial distribution of soot and polycyclic aromatic hydrocarbons (PAHs) during combustion is directly observed using high-speed cameras, laser-induced ignition (LII) method, and laser-induced fluorescence (PAH-LIF) method. It finds out that, the generation of NO_x during ammonia combustion is significantly higher than that of the conventional hydrocarbon fuels. Reasonable design of ammonia nozzles and selection of appropriate injection positions can significantly reduce the NO_x generation. The ammonia blending combustion technology provides a promising approach for achieving low-carbon transformation of coal-fired power plants. Although there are still challenges in the industrial application of large-scale ammonia blending combustion technology, the application prospects of ammonia fuel in coal-fired power plants are broad through continuous experimentation and technological optimization. Future research should continue to focus on the generation and emission of other pollutants during the process of ammonia combustion, and explore in depth the transformation behavior of minerals in coal in the combustion environment, providing theoretical and practical support for the industrialization of ammonia combustion technology.

When thermal power units participate in deep peak loading, real-time acquisition of furnace temperature field is helpful to power plant boiler control and research of combustion process in the furnace. With the promotion of intelligent power generation, machine learning provides an important means for real-time acquisition of furnace temperature field. The principle and application of the three most commonly used online monitoring technologies of furnace temperature field, namely acoustic method, absorption spectral tomography and thermal radiation imaging, are summarized at first, and the advantages and disadvantages in the application of boiler furnace temperature measurement are reviewed. Then, the principle of the coupled machine learning and CFD prediction method is described in detail, indicating that the method is less affected in the harsh furnace environment, and the application research of the method in the combustion flame structure and parameters and the furnace temperature field is reviewed, demonstrating the feasibility of applying the method to the furnace temperature field, indicating it can accurately predict the furnace temperature field. Finally, the future development trend of furnace temperature field online monitoring technology and coupled machine learning and CFD prediction method is analyzed, so as to provide ideas for obtaining more accurate furnace temperature field in real time under the continuous advancement of intelligent construction of power station.

To address the issues of flashback and high NO_x emissions in hydrogen-enriched gas combustion, a swirl-stabilized burner was designed using mild premixed combustion technology, and a feasibility study was conducted for its application in industrial boilers. Using a combination of experimental and numerical simulation methods, the study explored the equivalence ratio adjustment range of the mild premixed burner and the influence of hydrogen blending ratio on flame shape and pollutant emission characteristics. The experimental results showed that the mild premixed burner can achieve a wide hydrogen blending heat ratio adjustment range of 0~100%. The addition of hydrogen promoted a more uniform flame distribution, and the flame height decreased with the increasing hydrogen blending ratio. Additionally, the NO_x mass concentration experienced a rapid increase (for hydrogen blending ratios less than 30%) followed by fluctuations around 60 mg/m³. As the hydrogen blending ratio increased, the critical equivalence ratio for local flashback decreased, narrowing the stable combustion range between the blowout limit and the partial flashback limit. The widest equivalence ratio stable combustion range was achieved at a hydrogen blending ratio of 40%, which was 0.54~1.06.

With the growth of installed capacity of renewable energy power generation, coal-fired units need to undertake more peaking tasks. In order to improve the operational flexibility of coal-fired units, a 1 000 MW unit is taken as the research object, and six heat storage configurations and four heat release configurations of molten salt coupling are proposed based on the Ebsilon software, and the thermo-economic indexes of different heat storage and heat release coupling configurations are analyzed comparatively. The results show that, the peak shifting capacity of the system in the heat storage stage is positively correlated with the pressure loss of the heat transfer steam, and the thermal economy of heating the deaerator outlet feedwater in the heat release stage is the best. The heat storage of the electrically heated molten salt has the highest thermal and exergy efficiency, and configuration D-a has the strongest peak shifting capacity, with a peak shifting depth of up to 23.61%, but it has the largest coal consumption rate and exergy loss. Configuration F-d has the best thermal economy, with peak shifting depth, thermal efficiency, fuel efficiency and coal consumption rate of 23.42%, 39.61%, 38.40% and 310.2 g/(kW·h), respectively.

At present, foreign brands almost occupy the entire fan lubrication market share, and domestic oil products lack opportunities to enter the market. To demonstrate the feasibility of domestic substitution of lubricating oil for wind turbine gearboxes, the testing and analysis database for domestic gearbox lubricating oil and foreign competitors is established by adopting laboratory analysis to test the physical and chemical performance indicators (including kinematic viscosity, viscosity index, flash point, pour point, moisture, acid value, demulsibility, foam characteristics, liquid phase corrosion, and copper corrosion), tribological performance indicators (sucn as maximum seizure free load, sintering load, friction coefficient, wear spot diameter, comprehensive wear), and other lubricating oil performance indicators (like antioxidant performance, ferrography, infrared spectroscopy, and PQ index). The comprehensive performance of domestic gearbox lubricating oil products is systematically evaluated, and the feasibility study of domestic substitution of wind turbine lubricating oil is completed. The results indicate that there is not much difference in the basic performance indicators between domestic brand fan gearbox lubricating oil products and imported brands, and the substitution is feasible.