Research On MDEA Decarbonization Process Recommendation System Based On Combination Prediction Algorithm

With the rapid development of natural gas decarburization process,a lot of expert knowledge related to decarburization process has been accumulated,but at present,it mainly relies on the experience of professional technicians to realize the operation of decarbonization process,which has high technical requirements for operators,and in the existing case data,it is more difficult to optimize the decarburization process.In this paper,the simulation calculation of MDEA decarburization process parameters is realized by introducing a combined prediction algorithm,which simplifies the complex process of traditional decarburization process parameter optimization,based on the study of the selection of MDEA decarburization process parameters and the setting range of process parameters.Through the way provides certain guidance for the process optimization of decarbonization.First,based on the analysis of the MDEA decarburization process recommendation theory,by analyzing the MDEA decarburization process recommendation parameters,the structural characteristics of the existing process recommendation system model are studied;second,based on the characteristics of the combined prediction model of multiple algorithms,a genetic algorithm-based optimization is constructed BP neural network prediction model,support vector regression prediction model optimized based on particle swarm optimization algorithm,the initial single prediction model of the generated multi-algorithm combination prediction model;then,according to the mechanism of the multi-model combination prediction algorithm,the prediction model reconstructed through comparative analysis,Select the three prediction models with the best prediction effect as sub-models,use the game theory-based Chapter value for weight distribution,and construct a combined prediction model based on the Chapter value.After experimental analysis,the prediction accuracy is high and the proposed combined effect is better Advantages,and thus has a certain reference role for solving complex data.Finally,a software system for MDEA decarburization process recommendation based on combined prediction algorithm was constructed.The system can quickly and conveniently for the professional technicians,and recommend the best decarbonization process plan in the combined prediction results after training to users through process simulation.