| Blast furnace pulverized coal injection is a technology widely adopted on blast furnace iron making. Pulverized coal injection can reduce production costs as well as environmental pollution. Due to the flow of the gas-solid two phases flow in the pulverized coal injection tube is a complex process, how to detect gas-solid phase material is a solid-phase flow mass flow testing, one of unsolved problems.Aim at the complex flow characteristics for flow of pulverized coal. Based on capacitance measurement of pulverized coal flow signals, this paper achieved an accurate measurement of the amount of pulverized coal with soft-sensing technology. This paper built mathematical models of flow measurement with multiple nonlinear regression analysis method and achieved the automatic update of model parameters online. And the system is successfully applied for the mass flow measurement of blast furnace8on Guangdong Shaoguan Iron&Steel. Therefore, the main works of this thesis are as follows:(1) This paper optimized the structure and size of the capacitance sensor and made the sensor of concentration, velocity according to the parameters. At last, the author made a uniformity testing of dense detecting sensor.(2) Redesigned the circuit of main board and signal detecting on the basis of previous design, which have improved the stability of circuit.(3) Made analysis of common algorithms on calculation of speed, and analyzed features of field data. Besides, the author also checked the feasibility of the minimum mean square error algorithm and the waveform comparison algorithm.(4) Studied how to construct models for soft-sensing and made deep analysis on measurement model of Elman recurrent neural network and multiple nonlinear regression analysis. According to the field conditions, the author constructed model for mass flow measurement with nonlinear regression analysis. Besides, the author achieved the automatic update of model parameters online and got several models for different rate of pulverized coal injection using signals of concentration, velocity and temperature which obtained from industrial field. (5) Programmed PC software with C++Builder. The software includes the implementation of cross-correlation algorithm, waveform comparison algorithm and the minimum difference algorithm used for velocity calculation; implementation of multiple nonlinear regression analysis algorithm based on least squares technique, which is used for mass flow measurement; the implementation of algorithm which is used for automatic update of model parameters online as well as the implementation of correction algorithm used for flow correcting online. Besides, the author also accomplished data acquisition, data analysis and data saving as well as human machine interface and so on.(6) Made analysis of field data about velocity and flow and the result showed that the measurement system has good accuracy and can meet with filed desire. |
PDF Full Text Request