Research On Temperature Distribution Reconstruction Algorithm Based On Acoustic Thermometry

Temperature is a physical quantity used to describe the degree of hotness or coldness of objects or environments, which has vital significance to daily life, agriculture and industry. On one hand, it can be used as important reference information to estimate environment condition or running status of equipments. On the other hand, it always is a directly or indirectly control variable needed to be effectively monitored, and then given timely coping strategies if necessary.Among the existing temperature measurement methods, most are contact-type while with limitation of single-point measuring. Moreover, they are not suitable in severe industrial environments for continuous working, where may accompanied by high temperature, high pressure and corrosion condition. Temperature, however, is not only a concept with spatial significance, has distribution features in one-, two- and threedimensional direction. But also, it is a physical quantity varying with time, generally changes rapidly and unpredictable. Based on temperature's spatial significance and timevarying characteristic, it is necessary to real-time monitor and analyze temperature distribution, thus to understand and master it more comprehensively and accurately, then adopt effective control.Acoustic thermometry is an emerging non-contact temperature measurement technology, which is mainly based on specific dependencies between certain acoustic characteristic parameters and temperature when acoustic waves propagate in a medium. Thus, temperature information can be deduced by measuring these characteristic parameters or their changes. Among methods of acoustic thermometry, acoustic velocity method is a common one, whose working principle is based on specific function relationship between acoustic velocity and temperature. Apart from non-contact feature, acoustic thermometry also has advantages of wide temperature measuring range, strong environment adaptability, real-time and continuous, and can be used for measuring temperature distribution in space. At the same time, the feature of unaffected by microwave electromagnetic makes it has a giant application potential in the microwave industry field.This paper takes acoustic velocity method as main research object, investigates theories and methods of temperature distribution reconstruction in two- and threedimensional directions, on the basis of acoustic thermometry's principle and single-path acoustic temperature measurement. Researches are carried out centering on critical influence factors of acoustic temperature distribution reconstruction, then targeted solutions to problems existing in current researches are proposed as combined with specific application requirements. Thus, this paper is hoping to provide further theoretical basis and technical references for practical application of acoustic thermometry in the future. Here, the specific contents of researches can be briefly summarized as the following aspects:(1) This paper first summarized the existing temperature measurement technologies as well as their advantages and disadvantages, and pointed out the superiorities of acoustic thermometry; reviewed the development history, domestic and foreign research situation, and existing problems of acoustic thermometry; introduced single-path acoustic temperature measurement method, analyzed the principle of temperature distribution reconstruction using acoustic thermometry and discussed its critical influence factors.(2) Classical reconstruction algorithm in the field of acoustic thermometry technology is studied, namely algorithm based on least square, whose advantages and disadvantages are discussed. Specially, an improved temperature distribution reconstruction algorithm is proposed, which aimed at the phenomenon that reconstruction results by classical algorithm will lack of temperature information near measured area border. It is called algorithm based on least squares method and Multiquadric interpolation, as it combines with advantages of least square method and radial basis function. Its basic idea is: firstly, uses least squares method to obtain average temperatures of measured area's divided blocks, and regards them as temperatures of these blocks' geometric centers; then, set up Multiquadric interpolation model to estimate temperatures of other places in the measured area. This improved algorithm not only can solve the problem of temperature lacking near measured area border, but also maintains advantages of simpleness and quickness, while keeps reconstruction accuracy almost the same. Therefore, the proposed improved algorithm can be used in situations which the classic reconstruction algorithm applicable to, thus to compensate or improve the temperature distribution reconstruction effect.(3) Researches related to algorithm based on radial basis approximation and singular value decomposition are carried out, which is directed against the shortcomings of existing temperature distribution reconstruction algorithms. On one hand, this algorithm benefits from radial basis' s extraordinary abilities in function expression, interpolation and approximation of scattered, sparse data. On the other hand, it combines with the enormous advantage of singular value decomposition in solving inversion problem. In brief, this algorithm firstly uses radial basis to construct inversion model of acoustic temperature distribution reconstruction, then solves this inversion model by singular value decomposition, thus to accomplish high-precision reconstruction of temperature distribution. Simulation results show the algorithm's ideal reconstruction performance both in qualitative and quantitative ways.In order to investigate the above-mentioned algorithm's reconstruction effect in the real environment, actual experiments using split type acoustic transducers are carried out under the laboratory conditions. In this part of work, a measurement method based on fitting rising edge of echo envelope for time-of-flight is proposed, thus to gained time-offlight of acoustic wave with high precision. Research via actual experiments indicates that, algorithm based on radial basis approximation and singular value decomposition has excellent performance to reflect temperature distribution in real environment, and also proves the current practicability of split type acoustic transducers.(4) There exists great realistic demand for temperature distribution reconstruction of circular measured area, however, related researches in current acoustic thermometry field are very few. Aimed at this contradiction, an acoustic method of temperature distribution reconstruction applicable to special circular measured area is proposed in this paper, which called RBF-C method for short. This part of work mainly contains temperature distribution reconstruction algorithm, layout of acoustic transducers, selection of acoustic paths, and division of measured area, et al., and reasonable designs are made according to theoretical and experimental analysis. Simulation testing results have proved that, under various noises, our proposed RBF-C method can obtain excellent reconstruction results over the whole circular measured area, for all temperature distributions with different complexities. Compared with the existing LS-C method, our proposed method is greatly improved in terms of reconstruction accuracy, and without temperature information lacking near measured area border.(5) Researches about three-dimensional temperature distribution reconstruction are carried out, which is on the basis of researches related to single-path and two-dimensional situations. In this part of work, three-dimensional temperature distribution reconstruction methods respectively specific to cube and cylinder measured areas are proposed. Simulation results show that, reconstruction method proposed for cubic area has ideal reconstruction effect for temperature distributions with various complex degrees; reconstruction method proposed for cylinder area has ideal reconstruction effect for simple temperature distributions, and reconstruction effect becomes slightly poor for complex situations, however its ability to qualitative reflect the features of temperature distributions is still outstanding.Actual temperature's three-dimensional spatial distribution feature brings about the necessity of three-dimensional temperature distribution reconstruction. However, realistic difficulties hinder the progress of related researches to varying degrees. Due to less of related references, almost all of researches in three-dimensional situation in this paper are theoretically carried out under hypothetical ideal conditions, without much considerations of reality constraints such as environmental interferences or limits of transducer performances. In view of the above, related contents are expected to be specially studied in subsequent in-depth researches.