Research For Optimization Method Of Six-stage Time Ratio In Intermittent Coal Gasification System

At present, our country uses a fixed bed intermittent gas furnace, gas furnace in the gasification process with 150 step sequence to a cycle, during this period a variety of raw materials into the furnace, gasification reaction and half water gas collection intermittently. Each cycle is divided into six stages, the length of time for each stage has a direct impact on the amount of use gasification agent within a cycle, temperature changes, yield and quality of semi-water gas also determine the length of the reaction process, efficiency and output case, therefore six stages time ratio in a cycle is an important process control parameters in the production process. For time proportioning of six stages, based on the analysis of several linear and nonlinear parameter optimization, this paper uses dynamic parameters optimization method based on BP neural network to optimize time ratio of six-stage in intermittent coal gasification process. Main work is as follows:Firstly, for the missing data in coal gasification process, based on the study and research in theoretical knowledge of analysis similarity, we propose a compensation algorithm based on analysis similarity of data, validate the validity of data preprocessing methods through simulation in improving data quality.Secondly, for the depth study in characteristics, scope and the advantages and disadvantages of some commonly used linear and nonlinear parameter optimization method and compare their differences. The actual industry, time ratio data on six-stage have complex characteristics of small difference, irregular, noise, strong coupling relationship of each variable in coal gasification process. This paper presents a dynamic time ratio optimization methods combined with BP neural network. The basic steps of the method is to use the correlation analysis methods to analyze process data, use the large correlation variables to establish the dynamic linear model, iterative Optimizing the difference between real data and dynamic linear model by BP neural network.Thirdly, simulate time ratio on six-stage in intermittent coal gasification process with the data in production of Shandong Runyin chemical, verify the validity proposed method. The results show that this method has high accuracy and good control effect.