| Sensor technology is the foundation of information science,and is also one of the important pillars of modern information technology on IoT.With the development of modern industry,we put forward higher requirements on measuring the precision of sensor.Meanwhile,with the development of optical fiber technology,the Reflective Fiber-Optical Pressure Sensor?ROFPS?encapsulated by polymer jacket,is widely used for their advantages including its light weight,small size,far transmission distance,good corrosion resistance,good performance of anti-electromagnetic interference,etc.Due to temperature has great influence on measuring precision of ROFPS,this study aims at researching on the temperature compensation for it.By employing data fusion technology,this paper solves the problem that the influence of environment temperature on measuring the precision of ROFPS.The main contents are as follows:?1?The working theory of ROFPS is analyzed and some frequently-used data fusion algorithms are introduced,like multiple regression,traditional BP neural network and particle swarm optimization BP neural network?PSO-BP?algorithm.?2?For the problem that temperature has a great influence on optical fiber pressure sensor,we make a compensation for the fiber optic pressure sensor with multi-sensor information fusion technology.ROFPS is the main sensor and the DS18B20 temperature sensor is provided as an auxiliary sensor,then the two dimensional temperature calibration experiment is carried out.According to the data of calibration experiment,we calculate the temperature sensitivity coefficient is 8.4808 × 10-3/? and the additional temperature error is 32.3120%.?3?Establishing the compensation model of multiple regression,traditional BP neural network and PSO-BP neural network algorithm to make a compensation for the fiber optic pressure sensor,and we compare the compensation effect of several fusion algorithms.?4?A real-time pressure measuring and compensating system with wireless transmission is designed by using MSP430F149 Single Chip Microcomputer?SCM?.This design mainly including three parts,the first part,according to the pressure gathered by ROFPS and temperature collected by the temperature sensor,we transmit this data in this two parts to Upper SCM;the second part is about nRF24L01 short distance wireless transmission module;the third part,after receiving the algorithm from wireless module;the Lower SCM will deal with the transmission data from wireless module,and some other peripheral hardware circuits like alarm,display,etc will be also disposed.On the IAR software platform,we use C language to design the program of each functional module,which makes the system work properly.In this paper,by applying binary regression analysis and the traditional BP neural network algorithm to compensate for optical fiber pressure sensor,the research results show that the sensitivity coefficient is improved from the original 8.4808 × 10-3/? to 1.9412 × 10-3/?and 1.0991 × 10-3/? respectively;the additional temperature error is improved from the original 32.3120% into 7.4615% and 4.1877%.Both these Two kinds of algorithms are helpful to improve the performance of optical fiber pressure sensor.However,by employing PSO-BP neural network fusion algorithm,the compensated temperature sensitivity coefficient and the additional temperature error are 2.2734 × 10-4/? and 0.8622%.From the above data analysis,PSO-BP neural network could increase temperature coefficient of sensitivity of optical fiber pressure sensor to 1 order of magnitude and 2 orders of magnitude for additional temperature error.Compared with the former two algorithms,PSO-BP neural network algorithm achieves better compensation effects on temperature. |
PDF Full Text Request