| The rough set theory is an effective treatment of the uncertainty of mathematical theory. It has been widely used in many areas and provided a new theoretical tool for complex process modeling. Attribute reduction is an important topic of the rough set theory and it’s a NP-hard problem to find a minimal reduction. Due to the advantage of Ant colony algorithm in solving combinatorial optimization problems, this dissertation treats attribute reduction as a combinatorial optimization problem, and combines the rough set theory with ant colony algorithm to search the minimal reduction. In accordance with the problems in obtaining many minimal reduction sets of the same length, the dissertation takes the dependence degree of decision attributes on the non-core attributes to measure the solutions and selects the reduction set which has the maximum dependence degree, so as to find the optimal solution.The ammonia net value is an important indicator to reflect the efficiency of synthetic ammonia production. Accurate estimate of the ammonia net value of the ammonia production process is of great significance for the guidance of the ammonia production process operation. This dissertation takes an ammonia synthesis apparatus of a chemical plant as an object, combines rough set attribute reduction method with fuzzy neural network to set up a model for the ammonia net value of ammonia synthesis process. This paper has modeled the ammonia net value and verified the accuracy of the model by using the production data. Experimental results show the accuracy of the model meet the requirements and indicate the effectiveness of the algorithm.The ammonia net value modeling module of the software "operational optimization system of ammonia converter" is designed and developed. This module can reflect the changes of ammonia net value real-time so as to conduct the operator to adjust the value of the input variable to improve ammonia net value. |
PDF Full Text Request