Prediction method of deaerator water level in ultra-supercritical steam turbine fcb condition

Abstract

A method for predicting the water level of the deaerator under the FCB working condition of an ultra-supercritical steam turbine: Based on the RBF neural network model of the water level of the deaerator, starting from the actual data under different FCB working conditions, the data samples of the deaerator water level are sorted out, Use this sample to train the RBF neural network, and obtain the RBF neural network model of the deaerator water level after multiple iterations. The RBF neural model adopts a three-layer network structure, and the radial basis function adopts Gaussian functions. The RBF neural network model is used for FCB working conditions According to the prediction of the water level of the deaerator, the change of the water level of the deaerator under the actual working conditions is obtained. The calculation result of the present invention is very close to the measured value, and has ideal calculation accuracy. At the same time, the model has a simple structure and rapid calculation, and can achieve the purpose of predicting the water level of the deaerator under the FCB working condition, prevent the water level of the deaerator from being too low, and help the unit to run Personnel understand the operating status of the unit and maintain the safe and economical operation of the steam turbine.

Description

technical field [0001] The invention relates to a method for predicting the water level of a deaerator under the FCB (fast load shedding) working condition of an ultra-supercritical steam turbine based on a RBF (radial basis function) neural network. Background technique [0002] The water tank of the deaerator of the ultra-supercritical steam turbine is set up to ensure that the boiler has a certain feed water reserve. The ultra-supercritical unit adopts a once-through boiler. Since there is no steam drum, the deaerator is its only water outlet container, and its capacity is generally required to be Not less than the water supply required for 15-20 minutes of continuous operation under the rated load of the boiler. If the water level of the deaerator is too low, the insufficient water storage may endanger the safe operation of the boiler; The water level of the deaerator of the steam turbine will drop greatly under the FCB working condition, which seriously affects the safe...

AI Technical Summary

Problems solved by technology

[0002] The water tank of the deaerator of the ultra-supercritical steam turbine is set up to ensure that the boiler has a certain feed water reserve. The ultra-supercritical unit adopts a once-through boiler. Since there is no steam drum, the deaerator is its only water outlet container, and its capacity is generally required to be Not less than the water supply required for 15-20 minutes of continuous operation under the rated load of the boiler. If the water level of the deaerator is too low, the insufficient water storage may endanger the safe operation of the boiler; The water level of the deaerator of the steam turbine will drop greatly under the FCB working condition, which will seriously affect the safety of the unit operation

[0003] The existing method for predicting the water level of the deaerator under the FCB working condition of an ultra-supercritical steam turbine has low calculation accuracy, slow speed, and complex models, and it is difficult to accurately and real-time monitor the value of the water level of the deaerator

At present, there are no relevant research results on the prediction of the water level of the deaerator using the RBF neural network

Images