Development Of Micro Sugar In Sugar Factory Steam Condensed Water Detecting Device Based On RBF Neural Network

In the process of sugar production, Sugar companies need to use a lot of fuel, such as coal and bagasse, which is not only huge energy consumption, but also cause serious environmental pollution by the wastewater. In order to save energy and protect the environment, many sugar companies began recycling the steam condensed water, which from the process of boiling sugar and with the high temperature. The steam condensed water can be used as feed water into the boiler again, not only to achieve the purpose of saving water, but also can use the heat to saving energy. But these recover steam condensed water often contains trace amounts of sugar. If the sugar content exceeds the limit (50ppm) in steam condensed water, it will cause great security risk to the production equipment. Therefore, it is necessary to test the sugar concentration in steam condensed water before them into the boiler. Recently most of sugar companies are still using the artificial chemical colorimetric technology to detect micro sugar. Therefore, this topic designed a device which is based on TMS320F2812to detect the micro sugar in steam condensed water, and providing a precise and rapid method to detect micro sugar for sugar factory. The main research work is as follows:This topic propose a method which by detecting the dielectric loss, conductivity and temperature data of steam condensed water to measure micro sugar content. In device hardware system we select DSP chip TMS320F2812as the core of hardware system, and includes a sinusoidal signal generating circuit which DDS technology as signal generator for device, a signal detection circuit which used of high-speed AD chip, a temperature detection circuit based on high-precision digital temperature sensor, a liquid crystal display module circuit and a serial communications interface circuit. In software System we use the Fourier transform spectrum analysis algorithm to measure dielectric loss accurately, and introducing of the RBF neural network algorithm. Through network training and network arithmetic programming, the relationship between measured data and sugar content of the steam condensed water was obtained, which achieved the purpose of data processing precisely and rapidly.After completion of the device, we conduct a comparative experiment. By comparing the detect results and process of experiments, it shows that the device we designed have high precise and speed in detecting micro sugar, and easy to operate. The device can be used to detect the micro sugar concentration less than100ppm in steam condensed water.