摘要
针对传统电磁波传感器输出信号与持水率难以线性化的缺点,综合单片机和BP神经网络的优点,研制出基于MSP430F5529单片机和BP神经网络的原油持水率检测仪。通过MSP430F5529检测电磁波传感器输出相位差信号,经低通滤波和A/D转换后得到与持水率相对应的电压值,将实时温度和电压值送BP神经网络进行训练。该仪器首次将单片机与人工神经网络相结合用于原油持水率检测,具有响应时间短、精确度高、功耗低等优点,能方便实时地检测出原油持水率,具有很强的实用性。
Integrating the shortcoming that the output signal of the traditional electromagnetic sensor is difficult to linearize with the water-hold and the advantages of microcontroller and BP Neural Network,the crude water-hold detector of crude oil based on MSP430 F5529 MCU and BP Neural Network was developed.By detecting the phase shifted signals outputting from electromagnetic wave sensor from MSP430 F5529 and transforming the low-pass filter and AD into the voltage values which correspond to the water-hold,the real-time temperature and voltage values were sent to BP Neural Network to be trained.The instrument combines the microcontroller and artificial neural networks to detect the water-hold ofcrude oil for the first time,which has the advantages of short-response time,high accuracy and low-power consumption.Moreover,the water-hold of crude oil can be easily detected in real time by it,which is very practical.
出处
《仪表技术与传感器》
CSCD
北大核心
2017年第10期29-32,共4页
Instrument Technique and Sensor
基金
国家自然科学基金项目(61273179)
