Multi-parameter Field Tomography Reconstruction Of Inhomogeneous Flow Field Based On Ultrasonic Refrection/Reflection Effect

China is committed to achieving carbon peak before 2030 and carbon neutrality before 2060.To achieve this goal,rational and efficient energy conversion and utilization,as well as improved operating efficiency of power machinery,are essential.The development of reliable combustion and flow diagnostic techniques,including temperature and velocity field measurements,plays a crucial role in revealing the essence and laws of combustion phenomena,promoting the rational organization of combustion flow fields within power machinery,enhancing fuel utilization efficiency,and ensuring the safe and efficient operation of equipment.Due to the non-invasive nature,ability to penetrate opaque media,and compact detection equipment of acoustic tomography-based diagnostic method,it is highly suitable for combustion diagnostics in confined spaces such as combustion chambers or turbines where large-scale wall openings are not feasible.However,the accurate simultaneous reconstruction of temperature and velocity fields based on acoustic tomography is still a challenging problem limited by the difficulty of ultrasound transmission processes tracking and corresponding reconstruction algorithms realization,as well as the high underdetermination and severe ill-conditioning issues caused by the mutual coupling of multiple physical fields during simultaneous reconstruction.This thesis focuses on the construction of linear and nonlinear acoustic tomography,based on ultrasound transmission,refraction,and reflection/refraction fusion models.A systematic study and analysis of the simultaneous reconstruction and inversion methods for the temperature and velocity fields of combustion and flow are conducted.The specific research contents of this thesis are as follows:Based on the wave equation of sound wave in fluid media,a functional relationship between the temperature and sound velocity of medium is derived.Under the geometric acoustics assumption of high frequency sound wave,a combustion flow field information acquisition model based on ultrasound transmission model is constructed.An improved Tikhonov regularization method based on radial basis functions is proposed for sparse measurement data,which reduces the dimensionality of the measured flow field and adds prior information of continuity to the flow field.Parametric investigations are conducted on the type,shape parameters,and center point configuration of radial basis functions.The velocity field under typical conditions and the experimental data based on tomographic particle image velocity measurement are reconstructed.These verifications prove the applicability of the velocity field reconstruction model and algorithm based on acoustic tomography.Based on the linear acoustic tomography,taking into account the refraction effect of sound waves caused by changes in sound propagation refractive index due to large temperature gradients in the temperature field,a fast two-point ray tracing method is proposed to solve the problem,which is different from traditional boundary value problem solving methods.And as a result,the ultrasound transmission refraction model is constructed.As for the problem of sparse measurement signals caused by the limited number of transducers in confined spaces such as combustion chambers and gas turbines,ultrasound refraction and reflection fusion model based on the image source model is established.This model takes into account the first and second reflected echoes of the sound waves from the walls,achieving accurate tracking of ultrasound transmission process and effective increasement of measurement data.The introduction of convex optimization theory into the inverse problem solving process has led to the proposal of a novel method for the simultaneous reconstruction of velocity and temperature fields using radial basis functions and alternating direction method of multipliers.This approach has effectively reduced the influence of shape parameters on the instability of the solutions,thereby enhancing the anti-noise capability of the reconstruction process.The proposed method has been applied to simultaneously reconstruct the temperature and velocity fields based on experimental data from a Karman vortex street.The impact of different orders of reflected echoes,the number and layout arrangements of transducers on the reconstruction results has been investigated.When the velocity field needed to be reconstructed cannot be neglected compared to the speed of sound,the flow speed cannot be ignored compared with the sonic speed,and the time of flight with respect to the temperature and velocity fields is nonlinear.And the assumption of linear relationship no longer holds.The simultaneous reconstruction of the temperature and velocity fields in the flow field based on the ultrasound transmission refraction model in nonlinear acoustic tomography is investigated by using the covariance matrix adaptive evolutionary strategy algorithm.Box Constraints and regularization terms based on smooth prior information are added to the objective function to improve the multi-solution of this ill-posedness problem and increase the algorithm’s noise resistance.By studying the single-parameter regularization and dual-parameter regularization,a regularization parameter adaptive selection strategy is proposed.In order to improve the quasi-real-time reconstruction performance of nonlinear acoustic tomography,single-objective optimization is transformed into a many-objective optimization problem with four objectives.The simultaneous reconstruction of temperature and velocity fields based on many-objective optimization is studied.A novel knee point-driven evolutionary algorithm with improved environmental selection strategy is proposed.The effects of different measurement noise,population size,function evaluation times,and regularization norms on the reconstruction accuracy are investigated.Currently,the placement of transducers often relies on rough heuristic methods.Therefore,based on the prior information that the linear independence of vectors incorporates maximum measurement information,a linear independence degree metric is proposed to optimize the transducer array.This method achieves the co-optimization and reconstruction of transducer placement and parameter field based on both offline and online modes,maximizing the accuracy of parameter field reconstruction.A multiparameter field reconstruction experimental platform based on acoustic tomography is established.The improved Tikhonov regularization method based on radial basis functions is used to reconstruct typical velocity fields.The simultaneous reconstruction of typical temperature and velocity fields is achieved using linear and nonlinear acoustic tomography-based inverse problem methods.Physical experiments are conducted to validate the research results presented in the previous sections.The reconstructed results are compared with the point measurement values obtained from thermocouples and hot wire anemometers,which prove the effectiveness of the multi-parameter field reconstruction measurement system based on acoustic tomography and the applicability of the inverse problem optimization algorithm for different acoustic tomography imaging frameworks,providing guidance for the improvement of algorithms and models.