Noise Suppression For Measuring Acoustic Time Of Flight

Acoustic temperature field reconstruction technology based on acoustic time of flight has the advantages of non-contact,convenient maintenance,and being able to measure large space objects.It has prominent advantages in high-temperature fields such as industrial furnaces.This technology can also be used in fields such as submarine hydrothermal vents,grain storage,integrated circuit wafer temperature monitoring,and has high research value.In acoustic temperature field reconstruction techniques,the measurement accuracy of acoustic time of flight is the key to determining the accuracy of temperature field reconstruction.Acoustic wave time of flight measurement belongs to time delay estimation,and cross correlation method is often used.This method has good performance in suppressing low correlation noise.However,the observation of the same noise source,such as the operating sound of nearby equipment or ambient noise,at each acoustic receiver is correlated with each other.When the noise is large,the performance of the cross correlation method will decrease sharply.In order to improve the system’s resistance to environmental noise,this thesis has conducted research on noise suppression in acoustic time-of-flight measurement using simulated granaries as the experimental background.The main tasks are as follows:In order to suppress non-stationary noise such as noisy human voices and car horns that may be encountered during the measurement process,this thesis proposes a time of flight measurement method called LMDCC,which first uses Local Mean Decomposition(LMD)to denoise and then calculates the relative delay of the denoised signal using the cross correlation method.Firstly,based on the level of environmental noise,it is determined whether to reduce the probability of modal aliasing by truncating the signal;Then,perform LMD decomposition on the signal and reconstruct the denoised signal based on the useful component screening method proposed in this thesis;Finally,the measured value of acoustic wave flight time is obtained through cross correlation operation on the denoised signal.Algorithm validation experiments were conducted using measured data collected from simulated granaries.The experiment shows that the method proposed in this thesis can still maintain good accuracy and stability in acoustic time-of-flight measurement in low signal-to-noise ratio environments and other comparison algorithms that fail.A method for measuring the time of flight of sound waves is proposed,which first uses spectral subtraction with residual noise suppression and then performs cross correlation operations,in response to the interference of persistent stationary noise such as wind and rain.Firstly,the residual noise problem caused by spectral subtraction overestimation is improved by using over subtraction,spectral smoothing mechanism,and endpoint detection,forming a spectral subtraction(RSSS)method with residual noise suppression.By combining the RSSS method with basic cross correlation time delay estimation,a new acoustic time of flight noise reduction measurement method,the RSSSCC method,was obtained.Algorithm validation experiments were conducted using measured data collected from simulated granaries.The results show that the method proposed in this thesis is fast and effective,with good accuracy and stability.Its ability to suppress noise in low signal-to-noise ratio environments is significantly superior to the comparison algorithm.