Font Size: a A A

Research On Data Prediction Of Multi-effect Evaporation Salt Production Based On Feature Extraction

Posted on:2021-02-12Degree:MasterType:Thesis
Country:ChinaCandidate:Y HuaFull Text:PDF
GTID:2431330614962398Subject:Chemical engineering
Abstract/Summary:PDF Full Text Request
In the multi-effect evaporation salt production process,the smooth operation of the salt production process is very important.With the continuous process of salt production,many unstable factors will lead to the unstable process of salt production.These factors can be found in advance by predicting salt production data.Therefore,it is of great significance to study the prediction of multi-effect evaporation salt production data.In the process of multi-effect evaporation salt production,there are many salt-making devices,which affect the parameters closely.The influence of a single parameter on itself is sometimes delayed.Therefore,the data of multi-effect evaporation salt production has the feature of high dimensions,high complexity and time series.If the historical salt production data is directly used for data prediction,the prediction model will take a long time on training and the prediction effect is not good.Therefore,how to predict the data of multi-effect evaporation salt production is the main research problem in this paper.In view of the above problems,this article analyzes and improves the restricted Boltzmann machine,deep belief network,and autoencoder based on the characteristics of multi-effect evaporation salt production data,which is used for feature extraction of multi-effect evaporation salt production data,thereby solving the problem of high dimensions and complexity of salt production data.Based on this,combined with the time series information contained in salt production data,three multi-effect evaporation salt production data prediction models based on long short-term memory recurrent neural networks are proposed to solve the prediction problem of time-series salt production data.Experiments show that the three prediction models proposed in this paper can get good prediction results on multi-effect evaporation salt production data.In addition,this paper develops and designs a multi-effect evaporation salt production key data prediction system.This system displays the key data of multi-effect evaporation salt production in real time through a visual interface,and displays the predicted values of key data by drawing a curve in real time.In general,this article conducts prediction research on multi-effect evaporation salt production data,which can predict problems in advance and prevent them in advance.Therefore,this paper has certain theoretical research value and application value in terms of intelligent salt chemical process and production line optimization.
Keywords/Search Tags:multi-effect evaporation salt production, feature extraction, recurrent neural network, process optimization
PDF Full Text Request
Related items