To address the limitations of existing methods in handling non-stationary data and extracting salient features for monitoring ash blockage in rotary air preheaters of power station boilers， a hybrid differential pressure prediction model that integrates variational mode decomposition （VMD）， convolutional neural network （CNN）， bidirectional long short-term memory （BiLSTM）， and frequency-enhanced channel attention mechanism （FECAM） was proposed. To further improve the decomposition quality of VMD， the goose optimization algorithm （GOOSE） was employed to adaptively optimize the number of modes and the penalty factor， thereby enhancing the model’s ability to suppress noise interference. Experimental validation using real-world operational data from a coal-fired power plant demonstrates that the proposed model reduces root mean square error （RMSE）， mean absolute error （MAE）， and mean absolute percentage error （MAPE） by 42.93%， 43.58%， and 44.72%， respectively， in single-step forecasting compared to the baseline BiLSTM model. Additionally， the model exhibits superior robustness and predictive stability in multi-step forecasting tasks， outperforming other time series prediction models.

With the rapid development of steam turbine technology and the continuous emergence of various new types of steam turbine generator sets， the traditional frame-type foundation has been continuously optimized into a complex foundation with irregular structure. It is difficult to establish a concise and reasonable mechanical model and carry out a correct dynamic analysis for this kind of irregular space structure with multiple mass points and degrees of freedom. Taking the 9F-class gas generator set as an example， based on the vibration response characteristics of components such as the foundation base plate， column pier， and frame under the action of horizontal radial， axial， and vertical disturbing forces， four basic vibration modes with single mass points and single degrees of freedom were established by applying the principle of structural dynamics. Simple dynamic analysis and vibration verification were carried out， and the theoretical calculation values were verified with the on-site measured data. The research results prove that the dynamic analysis and vibration calculation method for the complex foundation of this large-scale power machine is reasonable and feasible， and has certain reference value.

Under special operating conditions， such as in the event of a condensate system accident or condensate throttling for primary frequency regulation of the unit， the deaerator water level may drop. Blindly and rapidly replenishing condensate when the deaerator water level is low can lead to a rapid decrease in deaerator pressure， resulting in insufficient net positive suction head （NPSH） available for the feedwater pump. By simplifying the transient mass-energy balance calculation model of the deaerator， the calculation formula for the conservative maximum condensate flow rate allowed to prevent insufficient NPSH of the feedwater pump was obtained. The calculation results using design parameters are consistent with the parameter variation laws of deaerator simulation modeling and transient calculation. Based on this， a deaerator water level control method was proposed to prevent insufficient NPSH for the feedwater pump by limiting the condensate flow rate. The method has been verified through experiments and can ensure that the NPSH of the feedwater pump is always sufficient.

The full life cycle aging management of nuclear power plants is a key task for the "Hualong One" nuclear power units. In order to effectively manage the aging issues of the "Hualong One" nuclear power units， it is necessary to retain material aging samples for concrete and cables. By analyzing the failure mechanisms of concrete and cable materials and combining with the actual operating conditions of the nuclear power plant， the concrete in the reactor pressure vessel pit area and the top of the outer dome of the containment vessel， and the cables in the steam generator， pressure stabilizer， and main steam pipeline room were finally selected for sample retention， forming a list of aging samples for concrete and cables. The implementation nodes for the regular inspection of subsequent aging management were planned， and finally the problems encountered during the implementation of the retained samples were summarized.

Under deep peak regulation operations， the primary air temperature of a subcritical 300 MW lignite-fired boiler in the eastern part of Inner Mongolia was relatively low， resulting in insufficient drying capacity of the pulverizing system and adversely affecting the combustion process inside the furnace. Therefore， a retrofit plan was proposed to install a steam heater in the primary air duct of the air preheater to increase the hot primary air temperature. The results show that the scheme can increase the primary air temperature by at least 35 ℃， and the effect of temperature improvement is remarkable under deep peak regulation operations. The technology can also optimize the distribution of air volume in the furnace and improve the boiler thermal efficiency.

A 2×600 MW unit in a power plant employs a dual-tower series desulfurization technology. During operation， the oxidation air ducts inside the tower frequently broke， resulting in a deterioration of the slurry quality in the absorption tower and difficulties in gypsum dewatering. Field inspection and analysis revealed that insufficient mechanical strength and improper support arrangement were the primary causes of duct failure. Based on an in-depth analysis of the corrosive environment inside the tower， the differences in performance and price among three commonly used stainless steels for desulfurization were analyzed. It is recommended to replace the material of the oxidation air duct with 2205 duplex stainless steel， which has high strength and excellent corrosion resistance， to meet the performance requirements and achieve a higher cost-effectiveness. Increasing the wall thickness of the air duct and optimizing its fixation method can effectively reduce stress concentration and prevent the deformation of the air duct. The transformation has achieved remarkable results， providing valuable reference experience for similar upgrades.

In order to ensure the safe operation of the generating set， a detailed analysis was conducted on the typical corrosion problems of two waste heat boilers in a gas-fired power plant. Specific types of corrosion included low-temperature corrosion， erosion corrosion， corrosion under insulation layer， uniform dissolved oxygen corrosion， atmospheric oxygen corrosion and stress corrosion. The results show that the main causes of corrosion include low-pressure recirculation system not being put into operation， the outer protection of the insulation layer not being tight， internal leakage of the valve， impurities in the pipeline， small curvature radius of the elbow， moisture in the equipment after the furnace is shut down， rainwater seeping into the insulation layer， and the pipe material being sensitive to stress corrosion. Therefore， effective corrosion prevention and control measures are proposed to ensure the safe operation of the equipment throughout its service life.

In the long-term operation of the pulverized coal pipelines in the No. 4 ultra-supercritical unit of a power plant， frequent issues such as support and hanger failure， crossarm deformation， and pulverized coal leakage posed serious safety risks to the unit. The causes of abnormal pipeline expansion were studied through on-site cold and hot state inspection of supports and hangers， disassembly analysis of bellows compensators， and stress simulation calculation using CAESAR Ⅱ software. The results show that improper installation and vibration of the V-type compensator at the coal mill outlet cause the hanger unloading. The bellows compensator at the burner inlet fails due to pulverized coal blockage and hardening of aluminosilicate fibers， unable to compensate for the thermal expansion of the boiler， causing the rigid hanger of the crossarm to be overloaded and deformed. By repairing the blockage in the inner cavity of the compensator， optimizing the pre-tension amount and adjusting the force on the hanger， the abnormal problems of the pulverized coal pipeline， as well as the overloading and complete unloading of the hanger， are successfully solved， providing an important reference for the design， installation and maintenance of the pulverized coal pipeline and its supports and hangers in coal-fired boilers.

To address the problem of poor drainage of high pressure heaters under deep peak regulation conditions， taking a 650 MW unit as the research object and combining the operating characteristics of high pressure heaters during deep peak regulation of the unit， an optimized drainage system for high pressure heaters based on a booster pump was proposed. This system adds a drainage bypass to the original normal drainage pipeline， uses a booster pump to overcome the water level pressure difference between the high pressure heater and the deaerator， and realizes the automatic switching of the drainage pipeline at low loads through the coordination of valve groups， meeting the requirements for normal operation of high pressure heaters drainage under deep peak regulation.

Theoretical modeling and simulation analyses of primary frequency control for the digital electric hydraulic control system （DEH） of 1 000 MWclass thermal power generating units were carried out. Firstly， an electro-hydraulic servo system model considering nonlinear factors was established. Secondly， a variable-parameter steam turbine model was proposed， in which the error does not change with the variation of the operating point of the unit load. A DEH optimization model considering nonlinear factors and the influence of load changes was established. This model is more consistent with the actual operating characteristics of deep peak regulation units and can maintain good accuracy and consistency within a wide range of deep peak regulation load intervals. A Simulink simulation model was established for a 1 000 MW class unit to analyze its primary frequency control performance under deep peak regulation conditions， and the results were compared with the unit’s actual test data， validating the accuracy of the established primary frequency control model.

Yaw deviation of wind turbines is commonly observed in operational wind farms， and the lack of appropriate measurement methods hinders accurate deviation assessment， preventing turbines from achieving optimal performance. To address the issue of static yaw deviation in wind turbine generator system， a quantitative analysis method based on operating condition segmentation and bidirectional binning was proposed. Utilizing big data technology and feature models， the method conducted both qualitative and quantitative analyses of historical operational data to achieve precise identification and quantification of static yaw deviation. Based on the operation data of a wind farm， support vector regression （SVR） was used for data preprocessing， and the yaw deviation was quantitatively evaluated by combining the operating condition segmentation and bidirectional binning method of wind speed and power. Results show that this method can effectively identify the yaw static deviation of wind turbine generator system， providing quantitative input for the yaw correction control and power curve optimization of wind farms， thereby providing effective support for improving the operation and maintenance efficiency of wind farms.

In order to consume green electricity and promote the construction of low carbon industrial parks， a system integration solution of "green electricity-energy storage-heating" based on molten salt energy storage technology was proposed. By constructing a molten salt energy storage system， the renewable energy consumption capacity can be effectively improved， while significantly reducing the carbon emissions of industrial parks. Based on the current situation of using thermal power plants to supply steam for an industrial park， a design was conducted using an electric heating molten salt energy storage system. The economy of using the molten salt energy storage system was evaluated by analyzing equipment selection and system operation. The results show that this scheme is expected to reduce the annual heating cost by 22% and the annual carbon emissions by approximately 3 241 t of the park. The molten salt energy storage system provides a replicable solution for building low carbon energy systems in industrial parks， and has important practical value for promoting energy structure adjustment in the industrial field.