Improvement Of Anti-ambient Stray Sound Ability Of Acoustic CT Temperature Field Measurement System

Acoustic CT temperature field reconstruction is a non-contact temperature measurement method,which is a computer tomography technology that can detect the temperature distribution of the measured area without disturbing.This technology can be used in the fields of ocean temperature field monitoring,atmospheric temperature field monitoring,industrial furnace temperature field monitoring,and storage grain temperature field monitoring,has great research value.In acoustic CT temperature reconstruction system,the measurement accuracy of Time of flight(TOF)is a key to the accuracy of temperature reconstruction.In actual measurement,there will be various ambient stray sounds,which will interfere with useful acoustic signals.Although the commonly used cross-correlation time delay estimation method has a good suppression effect on non-correlated noise interference,it has weak suppression ability on correlated noise interference.Therefore,when the volume of ambient noise is large,the estimation error of the cross-correlation time delay estimation method will be large or even the estimation will fail.This paper focused on improving the ability of acoustic CT temperature field measurement system to resist environmental stray sound,and following work was mainly completed.Aiming at the continuous interference of ambient stray sounds,an acoustic TOF measurement method based on the improved FastICA algorithm and the cross-correlation method is proposed.Aiming at the problem of under-determined blind source separation when the number of observation signals is less than the number of source signals,the phase space reconstruction method is used to expand the dimensions of the observed signals.Aiming at the problem that the FastICA algorithm is sensitive to the initial value,the steepest descent method is used to obtain a better initial iterative value.Aiming at the failure of TOF estimation caused by the uncertainty of FastICA algorithm,a strategy to re-estimate the failed data is proposed.Using the useful and spurious sound data collected in the simulated granary,the algorithm verification experiment was carried out.The results show that the method proposed in this paper can still estimate the TOF well when the ambient stray sound is strong and the cross-correlation time delay estimation method fails.In view of the strong,short-term burst noise that may be encountered during the measurement process,this paper proposes a method for distinguishing abnormal data and two methods for processing abnormal data: the TOF abnormal data processing method similar to the moving average and the method to eliminate the sound line containing abnormal data.Simulation experiments show that the temperature field reconstruction after abnormal data processing is better,and it is closer to the temperature field reconstruction without interference.This paper also studies the tolerance of the system to eliminate sound lines,and further verifies the feasibility of the method of eliminating abnormal sound lines.