| With the use of traditional Chinese medicine in China grow rapidly, medicine decocting machine demand is also increasing. However, the control system of medicine decocting machine have some defects.Such as: temperature control and detection are not accurate, system's anti-interference ability is poor. It is a difficult point to detect the temperature accurately. However the soft sensor technology can solve this problem well. Soft sensor technology is an emerging industrial technology has broad application background.This thesis introduces the soft sensor technology, The core of the soft sensor is building the digital model. The key to model is to find the primary variable and the right secondary variable. Modeling steps include data acquisition and processing; model building; model validation.This thesis introduces radial basis function neural network modeling and linear regression analysis. In the control system of medicine decocting machine, I select four secondary variables(pot body size, the grid voltage, external temperature, NTC detection value), Primary t variable is the concoction temperature. We use regression analysis method and RBF neural network method to build model. Through data analysis and simulation we prove the models are correct. and RBF neural network modeling method is more compact, higher accuracy.This thesis introduces how the medicine decocting machine works, analysis its structure and the features designed to achieve medicine decocting machine's hardware and software flow. Hardware parts include the temperature detection circuits, grid voltage fluctuation detection circuits and the digital temperature display circuits. To ensure the whole system's safety we also design some features, such as anti-dry heating, the limitation of longest heating time. Software components start from the medicine decocting machine's functions. According to these functions, we design software which can make the decocting program correctly . |
PDF Full Text Request