Study On Condition Monitoring And Early Warning Of Hydropower Units

With the rapid development of new energy industry in China,the installed capacity and single unit capacity of hydro-electric generating sets are becoming larger and larger,accounting for a larger and larger proportion in the whole power generation system.Hydro-generator set is a core component of hydro-generator,and its safety level directly affects the safety of the whole power station and even the power grid.,and its safety level directly affects the safety of the entire power station and even the power grid.Hydropower unit is a system which is affected by many factors,constantly changes with time,and requires high safety factor.Any carelessness may lead to an accident.Therefore,it is extremely important to monitor the status of hydropower unit and make an accurate prediction and early warning of the monitoring results.With the rapid development of big data and advanced science and technology,it is of great significance for the power industry and even the national economy to conduct research on multi-source data monitoring and early warning for hydro-generator.This paper first reviews the research status of hydropower unit condition monitoring and early warning.It is found that the current industry has not considered the impact of pre-operation data such as design,manufacturing,installation and commissioning on the state assessment system,real-time monitoring of state information,and lack of health status.It is found that the pre-run data has a great influence on the model structure and threshold analysis of the state assessment system.Especially for the discrete unit and multi-species equipment,the data before the operation is essential.Therefore,this paper conducts some research on the status monitoring of multi-source data impulse hydropower units.Firstly,it constructs the system architecture based on the status monitoring of multi-source data hydropower units,which follows the international regulations IEC61850 and IEC61970,and summarizes the technical means to realize this architecture.On the basis of deep research on the structure of the impact-type hydro-generator set,according to the national standard and the industry standard,the important state quantity that the unit needs to be monitored and the corresponding sensor type selection and installation are summarized.On this basis,a state monitoring and early warning system for the unit temperature monitoring module is studied and designed.Secondly,this paper briefly introduces the gray prediction theory and the particle swarm optimization algorithm.Considering that the background value interpolation coefficient and the boundary value of the GM(1,1)model have a certain influence on the final prediction result,the particle swarm optimization algorithm(PSO)is introduced.Attempts to use the advantages of the particle swarm optimization algorithm to find the interpolation coefficients and edge values that minimize the average relative error to improve the prediction accuracy.Then the two algorithms are fused,and the grey particle swarm prediction model based on nonlinear sequence is proposed.Then the proposed nonlinear particle swarm prediction model is used to predict the operating temperature of the hydroelectric generating set,and the simulation analysis is carried out.It can be seen that the improved model proposed in this paper has its superiority and can be used to predict the temperature of hydropower units,helping staff to find problems in advance.Finally,based on the previous research,the framework of hydropower unit temperature monitoring and early warning system was constructed,hardware selection and software application were carried out,and the appeal algorithm was integrated into the unified information platform architecture to monitor the temperature of hydropower units.The design and implementation of the early warning system can meet the basic needs of forecasting and early warning.