Variable-search-space ribonucleic acid (RNA) genetic algorithm modeling method for continuous stirred tank reactor

Abstract

The invention discloses a variable-search-space ribonucleic acid (RNA) genetic algorithm modeling method for a continuous stirred tank reactor. The method comprises the following steps of: (1) respectively performing dimension removal normalization processing on each group of input data and corresponding output data which are acquired by the continuous stirred tank reactor, wherein the input data comprise cyclic coolant temperature and material inlet flow in an inner clamping sleeve of the continuous stirred tank reactor, and the output data comprise reactant concentration conversion rate to be monitored; (2) constructing a radial basis function (RBF) nerve network model of the continuous stirred tank reactor according to the input data and the output data which are subjected to dimension removal normalization processing, wherein a basis function in the RBF nerve network model is a thin plate spline function; and (3) based on the input data and the output data which are subjected to dimension removal normalization processing, optimizing values of hidden layer node numbers and basis function central points in the RBF nerve network model by a variable-search-space RNA genetic algorithm, so that the output error of the RBF nerve network model is minimum, wherein a weight value of an output layer of the RBF nerve network model is obtained by a recursive least-squares method during optimization.

Description

technical field [0001] The invention relates to a modeling method of a variable search space RNA genetic algorithm for a continuous stirring reactor. Background technique [0002] With the large-scale chemical production, the chemical process becomes more and more complex. In order to improve product quality and yield, it is usually necessary to establish a relatively high-precision model based on the analysis of the chemical process mechanism. However, due to the uncertainty in the chemical process and the existence of mutual coupling between variables, the modeling of the chemical process becomes a very difficult task. In chemical production, the main reaction is accompanied by unknown or unmeasurable side reactions, which makes the modeling problem more difficult. Usually, the traditional mechanism modeling method is adopted for chemical process modeling, that is, to analyze the mechanism of the chemical reaction process, consider the main factors, ignore the secondary ...

AI Technical Summary

Problems solved by technology

The control performance of CSTR is affected by many factors, mainly due to the lack of a clear enough understanding of its reaction mechanism, and the serious sensitivity and nonlinearity of the process itself, which make it quite difficult to model the CSTR process.

As an important unit of the chemical process, the accuracy and complexity of the continuous stirring reactor model seriously affect the subsequent control effect, and ultimately affect the quality of the product

The CSTR process is a kind of complex chemical process with strong nonlinearity, and the model obtained by traditional modeling methods is difficult to reflect the real characteristics

Images