In view of the research situation of backpressure extraction steam turbine technology for ultrasupercritical power generation units, the off-design condition calculation model and exergy analysis model of the unit was established, and the influences of steam mass flow, temperature, pressure, steam turbine back pressure and the BEST heat extraction stage on the thermal performance of the unit were analyzed. The exergy efficiency of unit increased first and then decreased with main steam flow rate, reaching the maximum value (52.42%) when main steam flow rate was 750 kg/s. Exergy efficiency of unit increased with the increase of main steam temperature and pressure but decreased with the increase of turbine back pressure. The thermal performance of unit was more affected by back pressure at low load. In addition, the BEST series has a significant effect on the thermal performance of the unit, and the BEST match degree with the system design scheme is achieved when the series is 4, and the thermal performance is optimal. The calculation method and BEST series selection scheme can provide reference for the scheme design and operation optimization of ultra-supercritical units.

Aiming at the problem of inaccurate prediction of NO_x emission concentration when the current coal-gas boiler gas mixture is uncertain and changing, a combined online prediction method based on attention mechanism is proposed. First, the characteristic variables of the model are determined by combining the maximum information coefficient method with the Pearson correlation coefficient method; Secondly, vector autoregressive(VAR) model is constructed online with sliding time window for linearly correlated characteristic variables to realize the prediction of NO_x emission concentration under the input of multi-dimensional time series linear correlation variables For non-linear-related feature variables, the relationship between NO_x emission concentration is predicted by constructing an online Recurrent extreme learning machine(OR-ELM) model online learning. Finally, Attention Mechanism(AM) is used to dynamically weight the two forecasting models to achieve trend forecasting. Through field data verification, it shows that the VAR-OR-ELM combined online prediction model constructed in this paper can accurately predict the variation trend of NO_x emission concentration after 10 minutes. Combining prediction accuracy and prediction time, the combined prediction model is better than other single prediction models.

The denitration efficiency is closely related to the uniformity of flue gas and reductant agent within the selective catalytic reduction (SCR) reactor for the coal-fired unit. Based on the established mathematical model for SCR denitration reaction, a user-defined subprogram is used to couple the multi-component flue gas flow with reaction process. The reliability and effectiveness of the CFD model are verified by comparing the measured and simulated data of the SCR performance of 330 MW level coal-fired units at different loads. According to the hydrodynamics of flue gas and reductant agent together with the chemical reaction process in SCR reactor, a new intensification scheme is proposed by optimizing the structure of deflectors in front of the ammonia injection grids. Furthermore, the effects of operational conditions on emission mass concentration of NO and NH₃ are investigated. The results indicates that, the maldistribution of the incoming flue gas to the ammonia injection grids leads to the poor mixing behavior of flue gas and reducing reagent. However, the denitration efficiency of the SCR reactor can be improved by about 3.37% through adjusting the upstream guiding plate structure and installing the baffle around the flue duct wall. Taking the SCR denitration device in this work as an example, the appropriate molar ratio of NH₃ to NO is 0.94 when the initial NO mass concentration is 650 mg/m³, which could meet with the emission limit for air pollutions of 50 mg/m³ for NO_x and 2.5 mg/m³ for NH₃, respectively.

Through the analysis of the principle of the existing methods for determination the water content in oil and the components of gear oil, as well as monitoring needs of water content in gear oil, it is found that the existing methods can not accurately and quickly detect the water content in gear oil. the new gear oil and running gear oil for wind power was selected as experiment object, through study of selection of test conditions and the accuracy of the results, it is found a method that can accurately and quickly detect the water content in the gear oil. This method is a combination of the Karl Fischer coulometry water analyzer and the heating furnace. With compressed nitrogen or air as the carrier gas, after the carrier gas is dried through the molecular sieve, it cross the sealed test bottle containing 1g oil sample and 1mL n-heptane by the sleeve air needle, Heat the bottle at 150 ℃, and transfer the water in the oil sample to enter the Karl Fischer water analyzer for detection.

The penetration rate of distributed photovoltaic power stations in the power system is increasing year by year, to ensure the safe and stable operation of the power grid, a distributed photovoltaic ultra-short-term power prediction method based on combined neural networks is proposed. Firstly, a 1DCNN&1DCNN-LSTM combined neural network model is constructed by using 1D convolutional neural network (1DCNN) and long short-term memory (LSTM) neural networks, to obtain multi location numerical weather prediction (NWP) information and historical power information, using combined neural network model for spatially correlated photovoltaic power prediction and time series prediction; and a fully connected neural network (FCNN) is added to the combined neural network model, which is used to learn and assign weights to the two prediction results, achieving ultra-short-term prediction of distributed photovoltaic power generation. The validation was conducted using measured data from a photovoltaic power station in Hebei, and the results showed that this method can effectively improve the accuracy of distributed photovoltaic prediction and has certain practical value.

With the promotion of China's "carbon peaking and carbon neutral" strategy, thermal power units are more involved in deep peak regulation. Under the conditions of deep peak regulation, the thermal power unit is insufficient in heat storage, and the primary frequency regulation capability decreases, resulting in a large deviation between the unit's primary frequency regulation capability calibrated under the rated operating condition and the actual frequency regulation capability, threatening the frequency security of the power grid. Aiming at this problem, an online estimation method of primary frequency regulation capability of deep peak regulation thermal power units based on LSTM neural network is proposed. The static model of steady-state unit design was improved to a dynamic model, considering the dynamic operation process of the unit by using the time sequence memory ability and nonlinear feature extraction ability of LSTM neural network, and the errors caused by the disturbance factors such as the load changing process and the historical action of primary frequency regulation were corrected. Based on the hierarchical modeling method, the sub-models with different neural network structures were designed for the different characteristics of the factors affecting the frequency regulation capacity, such as heat storage of the unit and steam turbine work performance, and the effects of furnace side were taken into account to improve the accuracy of frequency regulation estimation results. Compared with the traditional method used in the power system, the estimation result of this method has higher accuracy, and has better performance under different working conditions such as steady state and variable load.

A large amount of data is generated during steam turbine operation. In order to meet the requirements of high quality data driven by big data and simulation modeling, efficient data cleaning is very necessary. The semi-supervised data cleaning model of steam turbine is built by using the excellent nonlinear fitting ability of long and short memory layer for time series data. The model selects three boundary conditions of the unit as input to predict the cleaning data. Outliers are eliminated according to the residual difference between the predicted value and the actual value. Then, the predicted value of the model is used to fill the data to ensure the integrity of the data. The model is used to clean the data of a 650 MW unit in a power plant. To overcome the problems caused by sample imbalance in the selection of cleaning model indicators, the accuracy rate is improved and taken as the measurement index of cleaning effect. The results show that, the improved accuracy of the data cleaning model of the deep long and short memory network is higher than that of the other three common cleaning methods, which can effectively identify whether the data is abnormal, and can use the predicted value to fill the data to ensure the consistency of data before and after cleaning.

Distributed energy power stations are developing rapidly because of their cleanliness, environmental protection, economy and high efficiency. However, there are few data used for fault diagnosis of plant equipment, so a method to predict the health state and aging degree of equipment is urgently needed. Based on this, a prediction model which can analyze the running state of equipment and obtain the deterioration trend of equipment is proposed. Firstly, multi-dimensional data of the equipment is preprocessed, and an improved Mahalanobis distance based equipment health model of distributed energy power station is constructed quantitatively by combining the analytic hierarchy process (AHP) with Gaussian mixture distribution. Then, the combined prediction model based on the improved sparrow algorithm and short and long time memory neural network is established to predict the trend and correlation analysis of the deterioration of distributed energy power plant equipment. The experimental results show that the proposed fusion health model can predict equipment anomalies in the case of insufficient actual fault data of distributed energy power stations.

In order to use the market mechanism to reduce carbon dioxide emission and promote green low-carbon transition, countries around the world have successively built carbon emission trading markets. Carbon dioxide emission monitoring technology is the main technical method to achieve accurate carbon emission measurement. It is an important technical support to assist the carbon emission accounting system. This paper focuses on the analysis of the current situation of carbon dioxide emission monitoring and accounting in the power generation industry, and introduces the carbon dioxide emission monitoring methods in the power generation industry in detail, including emission factor based method, online monitoring method, carbon balance method, soft sensing method, and satellite monitoring method. In view of these monitoring methods, this paper systematically reviews the researches of carbon dioxide emission monitoring methods in the world, expounds the advantages and disadvantages of the monitoring methods, compares the methods from accuracy, timeliness, reliability and monitoring cost, and provides reliable technical solutions for carbon dioxide emission monitoring in the power generation industry. Finally, we make an outlook on future research directions and practical applications.

Chemical absorption method is an important way to apply and treat CO₂ from coal-fired power plants on a large scale, however, the traditional chemical absorption method with monoethanolamine as absorbent has been limited in its wide application because of the high energy consumption. In this paper, the research progress on the improvement of CO₂ capture process is reviewed with the new CO₂ capture solvents, the improvement of absorption process including intermediate cooling of absorber and solvent recirculation, Flash compression and regeneration process of steam/pentane direct purging are summarized, it was also pointed out that pilot-scale verification of new solvents based on the actual composition of flue gas, and the study of capture solvent degradation properties and volatile organic compound treatment processes were the research and development directions of carbon capture research, it points the way for future research on industrial carbon capture.