With the development of boiler technology, requirement for the control of water level in drum has become higher and higher. Precision measurement and control of drum water level has become the key to stable operation of the boiler. However, the suppression and elimination of the drum water sloshing is still difficult to realize. The research of the mechanism of sloshing water as well as how to eliminate the adverse effects caused by the water sloshing has become a challenging task.In this paper, the mechanism of water sloshing in-depth discussions has been carried out based on the experiment, it is concluded that the time series of water level sloshing has chaotic characteristics. A neural network for prediction has been established in accordance with the corresponding chaotic parameters, and the precise prediction has been implemented.(1)A rock drum water level simulation bench has been established based onn similar principle by the analysis of the mechanism of the boiler drum water level. The time series has been sampled with the use of water level sensors, voltage amplifier, as well as devices such as data acquisition system.(2)The noise signal mixing in the water level sloshing signals has been removed by the use of empirical mode decomposition (EMD) and Hilbert transformation which called Hilbert-Huang transformation. This method makes the most useful information retained.(3)It is the first time to apply the new theory called chaotic properties of the same linear transformation into the analysis of the time sieries after noise removing, the conclusion that the water level sloshing time series has chaotic character has been reached, and by the detailed calculations, the chaotic parameters has been known: correlation dimension is 2.3457, Lyapunov exponent is 0.8269.(4)The theory of neural network has been used to predict the sloshing water level time series. Units of the input layer is 3, the hidden layer nodes is 5,and the output layer nodes is 1, BP algorithm, chaos optimization algorithm and genetic algorithm has been applied into optimization of the parameters of neural network, it is concluded that chaos optimization algorithm is more suitable for prediction of the sloshing water level.
|