Research On Non-dispersive Infrared SF₆ Gas Sensor And Air Pressure Compensation Method

The rapid development of electric power system has resulted in the large-scale application of SF₆ gas with good insulation and arc extinguishing functionality.However,the SF₆ gas is easy to react with the oxygen and water vapor in the air under the action of discharging.The various toxic and acidic substances produced can damage the normal operation of power equipment and seriously threaten the life and health of workers.In view of the current non-dispersive infrared SF₆ gas sensor in the actual detection in the process of nonlinear error caused by environment air pressure fluctuation problem,this paper proposes a hybrid PSO-BP neural network air pressure compensation model.Using its good nonlinearmapping and generalization ability to compensate the sensor detection errors caused by environmental air pressure,the accurate detection of SF₆ gas concentration in power system can be implemented.The specific research contents are as follows:1?The measuring principle of several typical SF₆ gas sensors is briefly introduced,with the technical advantages of the non-dispersive infrared SF₆ gas sensor highlighted.Then the research status and the method of air pressure compensation at home and abroad are described.2?The absorption principle of infrared spectrum is analyzed and the infrared spectrogram of SF₆ gas is drawn with the characteristic absorption peak selected.Based on the non-dispersive infrared detection technology,a single beam dual wavelength optical path structure is designed,which effectively eliminates the interference of optical path interference,light source power instability and zero drift of photoelectric devices.The loss of infrared radiation in the propagation process is reduced by designing a reasonable optical gas chamber.3?The test platform of the non-dispersive infrared SF₆ gas sensor is set up,then thecalibration experiment is carried out on the sensor,based on which the relation between SF₆ gas concentration and fractional absorbance is obtained.In addition,the verification experiment on the influence of test environment air pressure on the sensor is carried out.The air pressure compensation effect of PSO-BP neural network method and RBF neural network method is then analyzed.The experimental results show that the maximum measurement error of the sensor in the range of 0?2000ppm gas concentration and 100?120Kpa air pressure is reduced from 626ppm to 20ppm.The relative standard deviation is 1.32%,the full range accuracy is ± 1%FS,and the relative error is 2%.The SF₆ gas sensor has the advantages of high detection accuracy,good miniaturization and strong interference-proof ability and has wide application prospect in power system.