Research On 3D Temperature Field Reconstruction Techniques Of Acoustic CT

At present,temperature field detection technology has been widely used in industrial and agricultural and environmental monitoring and other fields.Acoustic computed tomography temperature field measurement technology needs to arrange a plurality of acoustic wave transceiver devices around the area to be measured to form a plurality of acoustic wave paths through the measured area,measure the acoustic wave propagation time of each path,and use an appropriate reconstruction algorithm to inversely measure the temperature of the measured area.Distribution,with non-contact,does not interfere with the measured temperature field,a wide range of temperature measurement and measurement of a wide range of advantages.This paper describes the principle of acoustic temperature measurement,reconstruction of the temperature field by the acoustic method,evaluation criteria for temperature field reconstruction,and factors affecting reconstruction.Acoustic computed tomography temperature field reconstruction algorithm can be divided into two categories: one is that the number of pixels required to be measured in the region is not more than the number of effective acoustic wave paths,and the other type of allowed pixel number is greater than the number of acoustic wave paths.For the first kind of algorithm,the least squares method,singular-value decomposition method,algebra reconstruction technique and simultaneous iterative reconstruction technique is deduced.These four algorithms are utilized to reconstruct the typical temperature field.The reconstruction results show that for no-noise flight time,the four reconstruction algorithms have similar reconstruction results.Least squares method and singular-value decomposition method as non-iteration algorithm reconstruction faster and the reconstruction results are basically the same,when the problem is not serious,their reconstruction accuracy is high and have a certain anti-noise ability,when the problem is seriously ill,the algorithm Affected by the ill-posedness,the temperature field reconstruction error is larger.Both the algebra reconstruction technique and the simultaneous iterative reconstruction technique are iterative algorithms,and the reconstruction result is less affected by the problem state.The algebra reconstruction technique is greatly affected by the noise,while the simultaneous iterative reconstruction technique is robust against noise.Of the four algorithms,the simultaneous iterative reconstruction technique is the best.However,simultaneous iterative reconstruction technique has the disadvantage of low convergence efficiency.For this reason,a step-length-optimized simultaneous iterative reconstruction technique is proposed.The iterative step size is adaptively adjusted by the minimum residual vector.Compared with the simultaneous iterative reconstruction technique,the proposed algorithm can better balance the temperature field reconstruction accuracy and reconstruction efficiency,the convergence speed is faster,the temperature field reconstruction effect is better,and the algorithm adaptability are stronger.This paper also studies the reconstruction algorithm based on radial basis function and regularization.The algorithm does not require the number of acoustic paths more than the number of spatial pixels to divide,and is more suitable for complex temperature field reconstruction.The selection of radial basis function shape parameters is investigated.The comprehensive evaluation index of reconstruction error was introduced,and the selection of regularization parameters under different noise levels was discussed.For the typical temperature field,the temperature field reconstruction was performed using the noiseless and noisy acoustic time-of-flight data,and compared with the reconstruction results of the four classical reconstruction algorithms.The comparison results are shown that the algorithm has the best reconstruction results,especially for noise-free data.