A small-scale three-dimensional temperature field observation inversion method

Abstract

The application discloses a small-scale three-dimensional temperature field observation inversion method, which utilizes an acoustic receiving and transmitting system to collect original data of an observation water area, carries out grid segmentation on the environment through collected temperature-salinity-depth profile data, and carries out three-dimensional sound ray simulation; carries out signal correlation processing on the original data collected by each acoustic station, obtains sound ray propagation time, carries out sound ray multipath resolution and extraction, processes arrival peaks of different paths, eliminates abnormal data of each peak, and obtains propagation time and propagation length in different sound ray forms; the obtained multi-station data are divided into grids in the three-dimensional environment to calculate sound ray length and propagation time in each grid, a coefficient matrix is obtained, and a three-dimensional grid temperature field is constructed; inverse problem calculation is carried out through the Lagrange least square method; and the three-dimensional temperature field is subjected to visual processing. The application has high precision and small result error through signal mutual transmission of multiple acoustic stations.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of ocean, river and reservoir hydrological environment monitoring, and particularly relates to a small-scale three-dimensional temperature field observation inversion method. BACKGROUND

[0002] Temperature observation of small-scale water areas such as small-scale marine ranching, cold and warm current intersection, shallow hydrothermal vent and artificial upwelling is a hot spot in current marine scientific research. However, due to technical limitations, it is often difficult to obtain long-term, three-dimensional and synchronous observation data, which seriously limits the development of related scientific research.

[0003] In the prior art, in order to obtain monitoring data, the patent application with publication number CN101592531B provides a real-time monitoring method for vertical temperature field distribution of ocean, which first sets a special floating box floating on the water surface, the special floating box is provided with a temperature detection host, the temperature detection host is connected to one end of a transmission sensing optical cable vertical to the seabed, the other end of the transmission sensing optical cable is connected to a deep water anchoring mooring device, the deep water anchoring mooring device is anchored on the seabed, the temperature detection host emits a series of laser pulses into the transmission sensing optical cable vertical to the seabed deep into the ocean bottom, the temperature detection host demodulates the light pulse echo signal to calculate the temperature of the sea water around each point along the transmission sensing optical cable, and then obtains the vertical distribution of the ocean temperature field. The monitoring method is used for single-point data monitoring, and does not involve multi-point monitoring and data visualization processing.

[0004] In addition, the patent application with publication number CN102622514A provides a marine temperature field data processing method, converts the obtained signal data into grid data with a spatial coordinate system, and synthesizes the grid data at the same time into a multi-layer grid data set; through the function of query analysis, the point temperature change curve graph, temperature profile change graph and regional temperature change graph in the measured region can be obtained; here, two-dimensional modeling is adopted for digital observation, and three-dimensional temperature field observation cannot be realized.

[0005] At present, in the observation and prediction scale of estuary, lake, ocean and other environments, two-dimensional modeling is mainly adopted for digital observation, and related three-dimensional observation technology is still blank. The three-dimensional temperature field observation of underwater acoustic observation system can accumulate data for water environment prediction, and ultimately provide accurate three-dimensional scale results. SUMMARY

[0006] The present application provides a small-scale three-dimensional temperature field observation inversion method, and proposes a three-dimensional environmental stationing, observation and data processing method, which provides a train of thought for establishing a three-dimensional temperature field hydrological environment. The present application adopts the following specific technical solutions:

[0007] A small-scale three-dimensional temperature field observation inversion method, comprising the steps of:

[0008] (1) The terrain of the observation water area is scanned to determine the positions where the stations need to be arranged, and an acoustic transceiving system is arranged at each acoustic station to collect original data of the observation water area;

[0009] (2) The environment is grid-divided through the collected temperature-salinity-depth profile data, three-dimensional sound ray simulation is performed, and the ray simulation results containing reference sound ray length, reference propagation time and time window information are obtained;

[0010] (3) The original data collected by each acoustic station is subjected to signal correlation processing to obtain sound ray propagation time, and the ray simulation results in step (2) are combined to perform sound ray multipath resolution and extraction, the arrival peaks of different paths are processed, abnormal data of each peak is eliminated, and the propagation time and propagation length of different sound rays are obtained;

[0011] (4) The multi-station data obtained are divided into grids in the three-dimensional environment to calculate the sound ray length and propagation time in each grid, a coefficient matrix is obtained, and a three-dimensional grid temperature field is constructed;

[0012] (5) The inverse problem is calculated by the Lagrange least square method, a smoothness penalty term is added, an error threshold is set, if the inversion error exceeds the set value, the iterative calculation is returned, and the inversion error is calculated until the inversion error meets the requirements;

[0013] (6) The three-dimensional temperature field is visualized.

[0014] In the present application, in step (1), the same frequency and different modal signals of the same family are used between each acoustic station in the acoustic transceiving system, and the same transmission mode is used between all stations to ensure that the acoustic signals of each other station can be received at the same interval time.

[0015] Preferably, the acoustic transceiving system comprises a float, an acoustic transducer, a temperature-salinity-depth profiler, a data acquisition system, a floating ball and a net position instrument; the float is located on the water surface, the data acquisition system is located inside the float and connected with the acoustic transducer; the acoustic transducer is arranged at different depths underwater, one end is connected with the floating ball, and the other end is fixed to a weight.

[0016] In the present application, each acoustic station of the float, the acoustic transducer, the temperature-salinity-depth profiler, the data acquisition system, the floating ball and the net position instrument in the acoustic transceiving system collects the original data of the water area; and the acoustic transceiving system uses high-frequency pseudo-random acoustic signals for two-way mutual transmission to finally obtain the observation experimental results.

[0017] Further, all acoustic transducers are fixed at a set water depth by the way of anchoring downward and floating upward, ensuring the position of the transducers unchanged, while the data acquisition system is placed in the water surface buoy connected with the transducers through cable.

[0018] Preferably, in step (2), the observed environment is surveyed by the instrument to obtain the topography of the observed environment; the obtained topography data are combined with GPS data and station position to input into the sound ray simulation program to obtain complete three-dimensional ray simulation results between stations, and reference propagation time of different sound rays is obtained.

[0019] Preferably, in step (3), the sound signal data obtained at each station in step (1) are subjected to signal correlation, and the sound ray peak values of different arrival times in the result are distinguished and compared with the reference propagation time of the sound ray simulation in step (2) to identify the sound rays of different arrival modes, remove noise data and obtain the peak values and observed arrival times.

[0020] Preferably, in step (4), when the grid division is performed, k×k×k cubic grids are divided in the three-dimensional region, and the sound rays are propagated on the grid, and the specific parameters and formulas are as follows:

[0021] For each sound ray, the following can be obtained: l ijk l 0k represents the length of the kth grid through which the jth sound ray between the ith station passes, C k and δC 0k represent the reference sound speed of the kth grid and the deviation of the actual sound speed in the kth grid from the reference sound speed C 0ij and δt ij represent the reference propagation time of the jth sound ray between the ith station and the deviation of the reference propagation time from the actual propagation time, respectively, l represents the total number of station surfaces, m represents the total number of sound rays between each station surface, and n represents the total number of cubic grids.

[0022] Further, when the grid division is performed, the sound ray length in each cubic grid is equal or the sound ray length difference is within a set range, and to improve the accuracy of the result, the cubic grid with a sound ray length close to 0 should be avoided.

[0023] In step (4), the specific process of constructing the vertical profile three-dimensional temperature field is as follows:

[0024] After Taylor expansion of the formula , the following is obtained: Let be the coefficient matrix, x=δC k be the vector to be inverted, and N be the solution error, y=δt ijThe measured reciprocal two-station sound propagation time difference is written as a matrix equation y=Ex+N;

[0025] The above matrix equation is solved by using a regularization inversion method, and an expected optimal solution x is obtained The value of lambda is determined by satisfying a set error threshold after iteration, and a three-dimensional H regularization matrix is introduced to smooth the result by moving average of multiple grids, and x=delta C is obtained k Then, the sound velocity C is further obtained k =delta C k +C 0j The sound velocity field in the entire observation area is obtained by interpolation, and the temperature field is inversely solved by using the sound velocity formula.

[0026] Compared with the prior art, the method has the following beneficial effects:

[0027] 1. The three-dimensional temperature field environment observation method is proposed by using the signal transmission between multiple acoustic stations, which has high precision and small result error.

[0028] 2. The signal mode of simultaneous transmission and reception of the acoustic station is adopted, and the observation results of multiple stations are obtained in a small scale range, which is compared with the data simulated by the sound line, and the result is obtained by solving the inverse problem.

[0029] 3. The inverse problem processing mode of the three-dimensional grid is established, and the result can be directly obtained by the observation data. DETAILED DESCRIPTION

[0030] Figure 1 The flowchart of the method of the present application is shown in the figure;

[0031] Figure 2 The station arrangement diagram of the method of the present application is shown in the figure;

[0032] Figure 3 The three-dimensional sound line simulation diagram in one embodiment of the present application is shown in the figure;

[0033] Figure 4 The three-dimensional grid temperature field result diagram between three stations in one embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0034] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, however, the present application can also be implemented in other ways different from those described herein, therefore, the present application is not limited to the specific embodiments disclosed below.

[0035] As shown in the figure, a small-scale three-dimensional temperature field observation inversion method comprises the following steps: Figure 1

[0036] ​Step 1: Topographic scanning is performed on the area to be observed to determine the positions where the stations need to be deployed, and a three-dimensional station deployment experiment is established. An acoustic transceiver system is arranged at each acoustic station to collect raw data of the observed water area.

[0037] In this embodiment, the acoustic transceiver system includes a buoy, an acoustic transducer, a temperature-salinity-depth profiler, a data acquisition system, a floating ball, and a net position instrument. The three-dimensional station deployment experiment requires that the acoustic transducer be deployed at different depths, and the position of the ultrasonic transceiver should be kept as constant as possible. All acoustic stations need to be time-synchronized using GPS and record the station positions. The acoustic signal uses the same transmission and reception mode, and different modes of the same signal family are used. The same transmission and reception mode allows each acoustic station to receive bidirectional acoustic signals at the same time, and the distance of the "direct path" obtained is used as the distance correction parameter between stations.

[0038] Each station is arranged as shown in Figure 2 The end of the acoustic transducer is anchored with a weight, and the other end is connected with a floating ball. The floating ball, line, ultrasonic transceiver, line, and weight are all underwater. The data acquisition system and battery are placed inside the buoy, which is fixed to the water surface by two anchors. A cable is connected between the buoy and the transducer, and the cable needs to be kept in a relaxed state. Figure 2 S1, S2, and S3 are the station numbers of the three stations, and L12, L23, and L31 are the station spacing numbers between two stations.

[0039] Step 2: High-precision three-dimensional acoustic ray simulation is performed using the collected temperature profile and topographic data and the station deployment data to obtain possible acoustic ray forms, reference propagation times, and acoustic ray lengths. The original observation data are also collected and correlated between two stations to obtain the propagation time of the observed acoustic ray.

[0040] In this step, the depth topography of the observed environment is obtained by using a depth sounder, an ADCP (acoustic Doppler current profiler), and other instruments. The obtained topographic data are combined with GPS data and station deployment positions to input into the acoustic ray simulation program to obtain complete three-dimensional ray simulation results between stations and reference propagation times of different acoustic rays.

[0041] Step 3: The simulation acoustic rays and reference propagation times in Step 2 are combined to compare the propagation time results after processing the observation data, to distinguish different paths in the simulation acoustic rays, to select all possible identified acoustic rays, and to distinguish and extract the relevant peaks and corresponding propagation times in all observation data.

[0042] Step 4: The three-dimensional acoustic ray area is divided into a three-dimensional grid, and the acoustic ray length and corresponding reference sound speed in each three-dimensional grid are solved. The observation data error values are removed as output quantities for solving the three-dimensional temperature field.

[0043] When the grid is divided, k x k x k cubic grids are divided in the three-dimensional region, and the sound rays are distributed on the grid and propagate, and the specific parameters and formulas are as follows:

[0044] For each sound ray, the following can be obtained: l ijk is the length of the jth sound ray between the ith station and the kth grid, C 0k and δC k represent the reference sound speed of the kth grid and the deviation of the actual sound speed in the kth grid from the reference sound speed C 0k , respectively, t 0ij and δt ij represent the reference propagation time of the jth sound ray between the ith station and the reference propagation time and the deviation of the actual propagation time, respectively, and l represents the number of station surfaces (for example, 5 stations, a total of 10 station surfaces), m represents the total number of sound rays in each station surface, and n represents the total number of cubic grids.

[0045] Step 5: Establish a three-dimensional temperature field inversion method, taking a 3x3x3 grid division as an example, as follows: the jth sound ray on the ith profile:

[0046]

[0047] The above formula is obtained by Taylor expansion:

[0048]

[0049] The above formula is written in matrix form as follows:

[0050]

[0051] wherein, is defined as the coefficient matrix, x = δC k is defined as the vector to be inverted, N is defined as the solution error, y = δt ij is the measured two-station sound propagation time difference, which is written as a matrix equation y = Ex + N.

[0052] Step 6: Calculate the sound ray length and propagation time of each three-dimensional grid, construct the coefficient matrix, construct the three-dimensional temperature field, and obtain the inversion error in each grid. The regularization inversion method is used to solve the equation. x is the expected optimal solution The value of λ is determined by the expected error within the set threshold; at the same time, a three-dimensional H regularization matrix is introduced to smooth the results by moving average among multiple grids. The solution x = δC k is obtained, and the sound speed C k = δC k + C 0kIt is also found that the sound velocity field in the entire observation area can be obtained by interpolation, and the temperature field can be obtained by applying the sound velocity formula.

[0053] C(T,S,D)=1448.96+4.591T-0.05304T 2 +2.734×10 -4 T 3 +1.340(S-35)+1.630×10 -2 D+1.675×10 -7 D 2 -1.025×10 -2 T(S-35)-7.139×10 -13 TD 3

[0054] In the above formula, C, T, D, and S represent the speed of sound, temperature, depth, and salinity, respectively; the H matrix is ​​based on... Figure 3 The grid numbers for the instances in the example are as follows:

[0055]

[0056] Step 7: Finally, perform visualization processing of the three-dimensional temperature field.

[0057] To verify the effectiveness of this invention, an observation experiment was conducted on a specific water area of ​​the Huangcai Reservoir in Changsha to obtain three-dimensional temperature field information in a vertical profile. Specifically:

[0058] It includes the following devices: three sets of 50kHz frequency transceivers, three sets of acoustic transceiver systems, floats, weights, TD, CTD, ADCP, etc.

[0059] Step 3 identifies and extracts multiple transmission lines on three transmission profiles within the entire 3D environment: the direct path, the surface reflection path, and the bottom reflection path, and calculates the ray length and propagation time of each transmission path for each grid.

[0060] After obtaining the ray length and propagation time for each 3D grid, a corresponding program was written in MATLAB to construct a 3D temperature field using a grid partitioning method. This optimizes the current approach, which only observes the average temperature field of layered paths along vertical profiles, thus obtaining more refined temperature information. Figure 4 The image shows the vertical temperature field information at a certain moment in a three-dimensional environment. The changes in ambient temperature can be observed through each isothermal surface.

[0061] The three-dimensional temperature field more intuitively reflects the distribution and variation trend of temperature at different locations in the entire three-dimensional environment during the observation period than the two-dimensional temperature field, demonstrating the effectiveness of this method in observing the three-dimensional temperature environment in small-scale water bodies.

[0062] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A small-scale three-dimensional temperature field observation inversion method, characterized in that, The method comprises the steps of: (1) performing topographic scanning on the observed water area to determine the positions where the acoustic stations need to be arranged, and arranging acoustic transceiving systems at each acoustic station to collect original data of the observed water area; (2) performing grid division on the environment according to the collected temperature-salinity-depth profile data, performing three-dimensional ray simulation, and obtaining ray simulation results containing reference ray length, reference propagation time and time window information; (3) performing signal correlation processing on the original data collected by each acoustic station to obtain ray propagation time, combining the ray simulation results in step (2), performing ray multipath resolution and extraction, processing the arrival peaks of different paths, eliminating abnormal data of each peak, and obtaining propagation time and propagation length of different rays; In step (3), the signal correlation of the acoustic signal data obtained at each station in step (1) is performed, and the ray peaks of different arrival times in the results are distinguished, compared with the reference propagation time of the ray simulation in step (2), the rays of different arrival modes are identified, the noise data is removed, and the peak value and the observed arrival time are obtained; (4) dividing the obtained multi-station data into grids in the three-dimensional environment to calculate the ray length and propagation time in each grid, obtaining a coefficient matrix, and constructing a three-dimensional grid temperature field; (5) performing inversion calculation by the Lagrange least square method, adding a smoothing penalty term, setting an error threshold, returning to iterative calculation if the inversion error exceeds the set value, and stopping until the inversion error meets the requirements; (6) performing visualization processing on the three-dimensional temperature field.

2. The small-scale three-dimensional temperature field observation inversion method according to claim 1, characterized by, In step (1), the same frequency and different modal signals of the same family are used between each acoustic station in the acoustic transceiving system, and the same transmission mode is used between all stations to ensure that the acoustic signals of each other station can be received at the same interval time.

3. The small-scale three-dimensional temperature field observation inversion method according to claim 1, characterized by, The acoustic transceiving system comprises a float, an acoustic transducer, a temperature-salinity-depth profiler, a data acquisition system, a floating ball and a net position instrument; the float is located on the water surface, the data acquisition system is located inside the float and connected with the acoustic transducer; the acoustic transducer is arranged at different depths underwater, one end is connected with the floating ball, and the other end is fixed to a weight.

4. The small-scale three-dimensional temperature field observation inversion method according to claim 3, characterized by, The acoustic transceiving system uses high-frequency pseudo-random acoustic signals for mutual transmission.

5. The small-scale three-dimensional temperature field observation inversion method according to claim 1, characterized by, In step (2), the observed environment is surveyed by the instrument to obtain the topography of the observed environment; the obtained topographic data are combined with GPS data and station arrangement positions to input the acoustic ray simulation program to obtain complete three-dimensional ray simulation results between stations, and reference propagation times of different rays are obtained.

6. The small-scale three-dimensional temperature field observation inversion method according to claim 1, characterized by, In step (4), when the grid division is performed, k×k×k cubic grids are divided in the three-dimensional region, and the rays are propagated on the grid body, and the specific parameters and formulas are as follows: For each sound ray, we can get: l ijk C 0k is the length of the jth sound ray between the ith station and the kth grid, C k and δC 0k represent the reference sound speed of the kth grid and the deviation of the actual sound speed in the kth grid from the reference sound speed C 0ij and δt ij represent the reference propagation time of the jth sound ray between the ith station and the reference propagation time and the deviation of the actual propagation time, l represents the total number of station surfaces, m represents the total number of sound rays in each station surface, and n represents the total number of cubic grids.

7. The small-scale three-dimensional temperature field observation inversion method according to claim 6, characterized by, When the grid division is performed, the ray length in each three-dimensional grid is equal or the difference of the ray length is within a set range, and to improve the accuracy of the results, the three-dimensional grid with a ray length close to 0 should be avoided.

8. The small-scale three-dimensional temperature field observation inversion method according to claim 6, characterized by, The specific process of constructing a vertical profile three-dimensional temperature field is as follows: After Taylor expansion of the equation we get Let be the coefficient matrix, x = δC k be the vector to be inverted, and N be the solution error, y = δt ij be the measured travel time difference, which can be written as a matrix equation y = Ex + N. The matrix equation is solved by regularization inversion method, and the optimal solution of x is expected The value of λ is determined by satisfying the set error threshold after iteration; at the same time, the three-dimensional H regularization matrix is introduced, and the result is smoothed by moving average among multiple grids; the solution of x = δC is obtained k After that, the sound velocity C k = δC k + C 0k The sound velocity field in the entire observation area is obtained by interpolation, and the temperature field is inversely solved by using the sound velocity formula.

Images