Research on Multi-sampling Rate Data Modeling Scheme for Hydrocracking Process
In order to improve the quality of petroleum refining products by controlling process parameters,it is necessary to construct a multi sampling rate soft sensing model for hydrocracking process.The research process constructed a multi sampling rate stack autoencoder algorithm model,and introduced the network structure,training method,and soft sensing modeling process of the model.In the performance testing phase,43 hydrocracking process variables were collected,and two quality variables,cetane number and 50%recovery temperature,were set for diesel products.The network structure was determined through experiments,and the performance was compared with the control group algorithm.It was found that the algorithm model established this time achieved the lowest root mean square error and the best quality prediction effect.
NETL
NSTL
万方数据
