Three-dimensional electromagnetic tomography measuring device based on movable TMR sensor array

Abstract

The invention relates to a three-dimensional electromagnetic tomography measuring device based on a movable TMR sensor array. The three-dimensional electromagnetic tomography measuring device is used for carrying out three-dimensional imaging measurement on a solid phase which flows through a pipeline to be measured and has magnetic conductivity. The device takes an FPGA (Field Programmable Gate Array) as a core, and the program design comprises an excitation module, a channel control module, a digital demodulation module and upper computer software; digital signal demodulation is realized by a multiply-accumulate IP core in the FPGA; the system circuit comprises a corresponding excitation circuit, a front-end signal measurement circuit and a channel control circuit; the excitation circuit comprises a plurality of excitation coils and excitation signal generation units corresponding to the excitation coils, and all the excitation coils are fixedly distributed on the same pipeline section; the front-end signal measurement circuit comprises a sensor array, an excitation channel and a measurement channel. The invention also provides a three-dimensional image reconstruction method of the solid phase with magnetic conductivity, which is realized by using the device.

Description

Technical Field

[0001] The present invention belongs to the field of multiphase fluid imaging in fluidization technology, and relates to a three-dimensional electromagnetic tomography measurement device based on a movable TMR sensor array. Background Art

[0002] Solid fluidization technology is a technology that enables solid particles to be suspended and in a dynamic equilibrium state through the continuous flow of a fluid (gas or liquid). In this state, there is a large contact area and strong relative motion between the solid particles and the fluid, resulting in extremely high heat and mass transfer efficiency and good mixing effect between the two. Currently, solid fluidization technology has been widely applied in the chemical, energy, metallurgical, and food industries. During the solid fluidization process, the measurement of the solid phase is crucial. By identifying the distribution and motion state of solid particles, abnormal fluidization phenomena (such as large bubbles, slugging, and channeling) as well as wear and faults of fluidized bed equipment can be avoided, thereby improving production efficiency and product quality, achieving automation and intelligent control, and providing important basis for equipment design and industrial scale-up. Due to the existence of local particle accumulation and other situations, many traditional measurement methods cannot work properly, and the testing instruments are extremely easy to be damaged. In the practical application of fluidization technology, the solid phase usually has a high magnetic permeability, while the gas phase and liquid phase generally do not have this characteristic. This phase difference provides a basis for detection using the magnetic permeability-sensitive electromagnetic tomography (EMT) technology. At the same time, the EMT technology has the characteristics of non-contact, non-invasive, and high time resolution, making it an ideal detection method that can effectively monitor and optimize the fluidization process.

[0003] Due to the capacitive coupling of the coil and size limitations, the traditional EMT technology using the coil as a detection element has obvious deficiencies. In 2018, Wang et al. proposed a TMR-EMT system based on a magnetic-sensitive element (Tunneling MagnetoResistance, TMR) sensor [1]. Due to the characteristic that the output signal of the TMR sensor is proportional to the magnetic field strength near the element, the TMR sensor can be used as a detection element in the EMT system to replace the detection coil based on the electromagnetic induction principle. Compared with the coil, the TMR sensor has better detection independence, sensitivity, resolution, noise level, and smaller size. This makes the TMR-EMT system have significant advantages in terms of the number and position flexibility of the measurement elements compared with the traditional EMT technology. This enables the TMR-EMT system to have a higher number of independent measurements, which helps to improve the underdetermination of the inverse problem solution and improve the image reconstruction quality.

[0004] Traditional EMT technology mainly focuses on two-dimensional imaging. It obtains the object field distribution information of a specific cross-section through a two-dimensionally distributed sensor array, but the three-dimensional distribution of the target area cannot be directly obtained. Currently, the commonly used three-dimensional EMT technology obtains the axial information of the target area by arranging multiple layers of coils. However, limited by the excitation frequency and coil size, the accuracy of the reconstructed image is limited. By using TMR sensors instead of coils for measurement, three-dimensional image reconstruction with higher accuracy can be achieved. At the same time, due to the small size of TMR sensors, more sensors can be arranged near a target area of a certain volume, further increasing the number of independent measurements of the system. And by designing the movable structure of the sensor array, the object field information in different directions can be obtained, and the quality of image reconstruction can be improved through the method of compensating for the rearrangement of the positions of the TMR sensor array.

[0005] References

[0006] [1]Chao W,He H,Cui Z,et al.Anovel EMT system based on TMR sensors forreconstruction of permeability distribution[J].Measurement Science andTechnology,2018,29.

[0007] [2]Shi Y,Li Q,Wang M,et al.A non-convex regularization methodcombined with Landweber method for image reconstruction in electricalresistance tomography[J].Flow Measurement and Instrumentation,2021,79:101917. Summary of the Invention

[0008] Aiming at the defects and deficiencies of the existing electromagnetic tomography technology in multiphase flow measurement mentioned above, the present invention provides a three-dimensional electromagnetic tomography measurement device based on a movable TMR sensor array. The technical solution of the present invention is as follows:

[0009] A three-dimensional electromagnetic tomography measurement device based on a movable TMR sensor array is used for three-dimensional imaging measurement of a high magnetic permeability solid phase flowing through a pipeline to be measured; the device is based on FPGA, and the program design includes an excitation module, a channel control module, a digital demodulation module and a host computer software; digital signal demodulation is implemented by the multiplication and accumulation IP core inside the FPGA; the system circuit includes a corresponding excitation circuit, a front-end signal measurement circuit and a channel control circuit;

[0010] The excitation circuit includes a plurality of excitation coils and an excitation signal generating unit corresponding to each excitation coil, and all the excitation coils are fixedly (uniformly) distributed on the same pipe cross section; the front-end signal measurement circuit includes a sensor array, an excitation channel and a measurement channel;

[0011] Assuming the number of excitation coils is N, the sensor array includes 3*N TMR sensor chips distributed on three sections of the outer circumference of the pipeline to be tested, the three sections are parallel to each other and are divided into three layers: upper, middle and lower; each excitation coil corresponds to three TMR sensor chips, the sensitive axes of the three TMR sensors are all arranged along the radial direction of the pipeline, and are located on the same plane as the central axis of the corresponding excitation coil, wherein the section where the TMR sensor array located in the middle layer is located coincides with the section where the center point of the excitation coil is located, and each TMR sensor chip can rotate around the pipeline axis on the section where it is located;

[0012] During measurement, the excitation module generates an excitation signal, which controls the corresponding N excitation coils through N excitation channels to generate a specific excitation magnetic field. If there is a solid phase with magnetic permeability in the target area, the excitation magnetic field will be distorted, and the boundary magnetic field strength of the distorted magnetic field is received by the three-layer TMR sensor array, and each TMR sensor chip corresponds to a measurement channel. The channel control module controls the excitation channel and the measurement channel. When one excitation channel is selected, all measurement channels are selected in turn, and the measurement value obtained by the TMR sensor is sent to the FPGA for digital demodulation, and then switched to the next excitation channel and the above process is repeated. The demodulated measurement value is transmitted to the host computer, and the measured boundary measurement value data is reconstructed through the host computer program.

[0013] Furthermore, all the excitation coils are evenly distributed on the same pipe section; the TMR chips distributed on each section of the outer circumference of the pipe to be tested are evenly spaced. The upper, middle and lower three layers of the sensor array are evenly spaced.

[0014] Furthermore, each TMR sensor chip can rotate around the pipe axis at a maximum angle of ±15° on the cross section where it is located.

[0015] Furthermore, the excitation signal generated by the excitation module is a sinusoidal voltage signal.

[0016] Furthermore, the excitation coils are all circular structures. The outer diameter of the circular coil is 23.5 mm, the inner diameter is 6.5 mm, and the thickness is 5 mm; the spacing between different layers of the sensor array is 10 mm.

[0017] The present invention also provides a three-dimensional image reconstruction method for a solid phase with magnetic permeability implemented by using the above-mentioned device, including the following steps:

[0018] (1) Rotate the sensitive axes of the N TMR sensors in each layer to be parallel to the central axis of the N excitation coils; select a certain excitation channel through the channel control module, apply an excitation signal to the excitation coil of this excitation channel, and sequentially select all measurement channels. Send the measured values obtained by the TMR sensor chip to the FPGA for digital demodulation to obtain the boundary magnetic field intensity measurement values. Then switch to the next excitation channel and repeat the above process until the magnetic field intensity values between all excitation measurement pairs are obtained; the output values of the measurement results obtained under the N excitation coils and the N*3 TMR sensor chips form an M-dimensional boundary magnetic field intensity measurement value vector B, where M = N*N*3 is the number of independent measurements;

[0019] (2) Divide the imaging area into T pixel units, and obtain the sensitivity matrix S (M×T) by using the measured data;

[0020] (3) Use the boundary magnetic field intensity measurement value B in the empty field 0 to normalize the measurement data, and the normalized measurement result is B norm (i), and the calculation method of B norm (i) is B 0 (i) = (B(i) - B 0 (i)) / B

[0021] (4) Establish a linearized model B norm = Sg, where g is the gray value vector representing the relative change of the magnetic permeability of the substance; according to the obtained sensitivity matrix S and the normalized measurement result B norm , use the iterative method to calculate the unknown quantity g;

[0022] (5) Rotate each TMR sensor chip by a certain angle around the central axis of the corresponding excitation coil on the outer peripheral cross-section of the pipeline to be measured where it is located. The rotation angle does not exceed the set maximum rotation angle. At this time, the relative positions of each TMR sensor chip and the excitation coil change, so that the sensor array can receive the object field information at different positions; then repeat the above steps to obtain a new optimized gray vector value g j , where is the jth image reconstruction result;

[0023] (6) Take the arithmetic mean fusion of the K image reconstruction results Obtain the optimized gray-scale vector value after data fusion;

[0024] (7) Use G as the gray-scale value on each pixel for imaging.

[0025] Furthermore, in step (3), normalize the measurement result B norm (i) is calculated as B norm (i) = (B(i) - B 0 (i)) / B 0 (i), where i is the serial number of the excitation measurement pair.

[0026] Furthermore, in step (4), use the improved Landweber iteration method to calculate the unknown g k+1 = T(g k - λS T (S × g k - ΔB)), where the threshold function threshold λ is the iteration step size, and its value is p is the penalty term norm; stop the iteration after the iteration converges to the preset value. At this time, g k is the reconstructed optimized gray-scale value vector;

[0027] Furthermore, p takes the value of 0.8. When ||g k - g k-1 || < 0.001, stop the iteration. At this time, g k is the reconstructed optimized gray-scale value vector.

[0028] The device and method of the present invention can better measure and display the three-dimensional distribution of the high magnetic permeability solid phase in multiphase flow measurement, which is an extension and supplement to the previous electromagnetic tomography technology based on TMR. It gives full play to the advantage of the small volume of the TMR sensor, obtains more object field distribution information by changing the relative position of the TMR, and improves the quality of the reconstructed image by using the data fusion method. Description of the Drawings

[0029] Figure 1 : System schematic diagram of the three-dimensional electromagnetic tomography measurement device based on TMR.

[0030] Figure 2 : Schematic diagram of the excitation coil and sensor array structure

[0031] Figure 3 : Schematic diagram of the maximum rotation range of the TMR sensor relative to the initial position

[0032] Explanation of the Reference Numerals in the Drawings

[0033] 1. Insulated pipe; 2. Electromagnetic shielding layer; 3. Excitation coil; 4. TMR sensor; 5. Coil central axis; 4. TMR sensor sensitive axis. Specific embodiments

[0034] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0035] As Figure 1 shown, the device of the present invention is composed of a front-end circuit, an FPGA, a system circuit, and a host computer. Among them, the front-end circuit includes an excitation coil and a sensor array. The FPGA includes an excitation module, a channel control module, a digital demodulation module, and a USB communication module. The system circuit includes an excitation circuit, a measurement circuit, and a channel control circuit. This system selects the Cmod S6 development board with the Xilinx Spartan 6LX4 FPGA as the core as the core control device. The DDS IP core inside the FPGA can generate a 1 kHz sine excitation signal. The excitation circuit converts this sine digital signal into a voltage signal through digital-to-analog conversion, and after filtering and amplification, drives the excitation coil of the corresponding excitation channel through a power amplification circuit, thereby generating an excitation magnetic field. Among them, the digital-to-analog conversion is implemented using the high-speed DAC chip AD9764, and the filter amplification circuit can be implemented using a high-current power amplifier from Texas Instruments (such as OPA544).

[0036] The control circuit is used to control the application sequence of the coil excitation signal in the front-end circuit and the reading sequence of the measurement data of the TMR sensor. In the front-end circuit, each excitation coil corresponds to an excitation channel, and each TMR sensor corresponds to a measurement channel. The control circuit realizes the conduction of each excitation and measurement channel through the cooperation of a latch and an analog switch. The latch can convert the serial channel control word signal output by the channel control module of the FPGA into an 8-bit parallel signal, and at the same time control the conduction state of the analog switches of each channel. The control circuit can be implemented using the shift register chip 74HC595 and the analog switch chip DG412DY.

[0037] The measurement circuit first filters and amplifies the boundary measurement value signal obtained by the sensor array, and then sends it to the FPGA for signal processing through analog-to-digital conversion. The FPGA restores the amplitude and phase information of the measurement signal through the digital demodulation module, and converts the calculation result into a 32-bit floating-point number, which is sent to the host computer for image reconstruction through the USB communication module. The filter amplification circuit for the front-end measurement signal can use the operational amplifier chip LM2904.

[0038] Figure 2Schematic diagram of the position distribution of the excitation coil and the sensor array, which consists of the insulating pipeline 1 to be measured, the electromagnetic shielding layer 2, the excitation coil 3 and the TMR sensor 4. The 8 excitation coils are all circular structures. The outer diameter of the circular coil is 23.5 mm, the inner diameter is 6.5 mm, and the thickness is 5 mm. They are fixedly distributed on the same pipeline cross-section. All coils are tightly wound with copper wires with a wire diameter of 0.5 mm. The sensor array is located between the pipeline to be measured and the excitation coil, and includes 24 TMR sensor chips distributed on three pipeline cross-sections. As Figure 2 shown, the sensitive axis of the TMR sensor chip is set along the radial direction of the pipeline, pointing towards the inside of the pipeline, and is located in the same plane as the central axis of the coil. The TMR chips on each cross-section are equally spaced. The three cross-sections are parallel to each other and are divided into three layers: upper, middle and lower. Among them, the cross-section where the middle-layer sensor array is located coincides with the cross-section where the excitation coil is located. The cross-sections where the three-layer sensor arrays are located are equally spaced, and each layer can independently rotate around the central axis of the pipeline. The distance between the sensor arrays of different layers is 10 mm. Before starting the measurement, the sensitive axes of the 8 TMR sensor chips in each layer are placed in a position parallel to the central axes of the 8 coils. Under the action of the control circuit, each time a sine excitation signal is used to excite one coil through the corresponding excitation channel, and the measured values of each measurement channel are read in sequence. Then, the excitation channel is switched to repeat the above process. The collected data is sent to the upper computer for image reconstruction through filtering, amplification and demodulation via the USB communication module of the FPGA. The image reconstruction method is the improved Landweber iteration method, and the process is as follows:

[0039] 1) The imaging area is divided into 5379 pixel units, and the sensitivity matrix (192×5379) is calculated. 192 = 8×24 is the number of independent measurements, where 8 is the number of excitation coils and 24 is the number of TMR sensors. Define the combination of the m-th coil and the n-th TMR sensor as the m-n excitation measurement pair, and its sensitivity value to the magnetic field intensity value at the l-th pixel unit is S mn (l) = -jω(B AX (l)·B BX (l)+B AY (l)·B BY (l)+B AZ (l)·B BZ (l)), where B AX (l), B BX (l) and B AY (l), B BY (l) and B AZ (l), B BZ (l) respectively represent the magnetic induction intensity values in the X direction, Y direction and Z direction of the magnetic field generated by the excitation coil and the magnetic dipole in the sensitive field under different combinations. Thus, the sensitivity matrix S is obtained.

[0040] 2) Use the measured value B of the boundary magnetic field strength in the empty field 0 Normalize the measurement data, and the normalized measurement result is B norm The calculation method of (i) is B norm (i) = (B(i) - B 0 (i)) / B 0 (i), where i is the excitation measurement pair number

[0041] 3) Establish a linearization model B norm = Sg, where g is the gray value vector representing the relative change of the magnetic permeability of the substance. Through the above steps, the sensitivity matrix S and the normalized measurement result B can be obtained norm , from which the unknown quantity g can be calculated using the improved Landweber iteration method [2] k+1 = T(g k - λS T (S × g k - ΔB)), where the threshold function Threshold λ is the iteration step size, and its value is p is the penalty term norm, and its value is 0.8; when ||g k - g k-1 || < 0.001, stop the iteration. At this time, g k is the reconstructed optimal gray value vector; in this embodiment, the improved Landweber iteration method proposed in Reference 2 is used to calculate the gray value vector g of the relative change of the magnetic permeability of the substance. In practical applications, other optimization methods can also be used, such as the original Landweber iteration method, and the technical effects will be different

[0042] 4) Rotate each layer of the TMR sensor array around the pipeline. At this time, the relative position between the sensor chip and the excitation coil changes, so that the sensor array can receive the object field information at different positions. Fuse the object field information data at multiple positions, and use the fused data for image reconstruction, which can further improve the image reconstruction quality. As Figure 3 shown, the value range of the rotation angle of each layer of TMR sensor is -15° to 15°. Repeat the above steps to obtain a new optimal gray vector value g j (j = 1, 2... K), where j is the number of image reconstructions. Arithmetically average and fuse the optimal gray value vectors at different positions to obtain the optimal gray G after data fusion. Perform three-dimensional imaging based on the gray values of each pixel in G

Claims

1. A three-dimensional electromagnetic tomography measurement device based on a movable TMR sensor array, used for three-dimensional imaging measurement of high magnetic permeability solid phase flowing through the pipeline to be measured; the device is based on FPGA, and the program design includes an excitation module, a channel control module, a digital demodulation module and a host computer software; digital signal demodulation is implemented by the multiplication and accumulation IP core inside the FPGA; the system circuit includes a corresponding excitation circuit, a front-end signal measurement circuit and a channel control circuit; The excitation circuit includes a plurality of excitation coils and an excitation signal generating unit corresponding to each excitation coil, and all the excitation coils are fixedly distributed on the same pipe cross section; the front-end signal measurement circuit includes a sensor array, an excitation channel and a measurement channel; Assuming the number of excitation coils is N, the sensor array includes 3*N TMR sensor chips distributed on three sections of the outer circumference of the pipeline to be tested, the three sections are parallel to each other and are divided into three layers: upper, middle and lower; each excitation coil corresponds to three TMR sensor chips, the sensitive axes of the three TMR sensors are all arranged along the radial direction of the pipeline, and are located on the same plane as the central axis of the corresponding excitation coil, wherein the section where the TMR sensor array located in the middle layer is located coincides with the section where the center point of the excitation coil is located, and each TMR sensor chip can rotate around the pipeline axis on the section where it is located; During measurement, the excitation module generates an excitation signal, which controls the corresponding N excitation coils through N excitation channels to generate a specific excitation magnetic field. If there is a solid phase with magnetic permeability in the target area, the excitation magnetic field will be distorted, and the boundary magnetic field strength of the distorted magnetic field is received by the three-layer TMR sensor array, and each TMR sensor chip corresponds to a measurement channel. The channel control module controls the excitation channel and the measurement channel. When one excitation channel is selected, all measurement channels are selected in turn, and the measurement value obtained by the TMR sensor is sent to the FPGA for digital demodulation, and then switched to the next excitation channel and the above process is repeated. The demodulated measurement value is transmitted to the host computer, and the measured boundary measurement value data is reconstructed through the host computer program.

2. The three-dimensional electromagnetic tomography measurement device according to claim 1, characterized in that: All excitation coils are evenly distributed on the same pipe section; TMR chips distributed on each section of the outer periphery of the pipe to be tested are evenly spaced. The upper, middle and lower sections of the sensor array are evenly spaced.

3. The three-dimensional electromagnetic tomography measurement device according to claim 1, characterized in that: Each TMR sensor chip can rotate around the pipe axis at a maximum angle of ±15° on its cross section.

4. The three-dimensional electromagnetic tomography measurement device according to claim 1, characterized in that: The excitation signal generated by the excitation module is a sinusoidal voltage signal.

5. The three-dimensional electromagnetic tomography measurement device according to claim 1, characterized in that: The excitation coils are all circular in structure, with an outer diameter of 23.5 mm, an inner diameter of 6.5 mm, and a thickness of 5 mm; the spacing between sensor arrays in different layers is 10 mm.

6. A method for reconstructing a three-dimensional image of a solid phase having magnetic permeability using the device according to any one of claims 1 to 5, comprising the following steps: (1) Rotate the sensitive axes of the N TMR sensors in each layer to be parallel to the central axes of the N excitation coils; A certain excitation channel is selected through the channel control module, an excitation signal is applied to the excitation coil of this excitation channel, and all measurement channels are selected in turn, and the measurement value obtained by the TMR sensor chip is sent to the FPGA for digital demodulation to obtain the boundary magnetic field strength measurement value, and then the next excitation channel is switched and the above process is repeated until the magnetic field strength values ​​between all excitation measurement pairs are obtained; the output values ​​of the measurement results obtained under N excitation coils and N*3 TMR sensor chips constitute an M-dimensional boundary magnetic field strength measurement value vector B, where M=N*N*3 is the number of independent measurements; (2) Divide the imaging area into T pixel units and use the measured data to obtain the sensitivity matrix S (M × T); (3) The measured data is normalized using the boundary magnetic field strength measurement value B0 under the empty field. The normalized measurement result is B norm The calculation method of (i) is B norm (i) = (B(i) - B0(i)) / B0(i), where i is the number of the excitation measurement pair; (4) Establish linear model B norm =Sg, where g is the gray value vector representing the relative change of the material's magnetic permeability; according to the obtained sensitivity matrix S and the normalized measurement result B norm , use the iterative method to calculate the unknown quantity g; (5) Each TMR sensor chip is rotated around the central axis of the corresponding excitation coil on the outer peripheral cross section of the pipeline to be tested by a certain angle, and the rotation angle does not exceed the set maximum rotation angle. At this time, the relative position of each TMR sensor chip and the excitation coil changes, so that the sensor array can receive object field information at different positions; then repeat the above steps to obtain a new optimized gray vector value g j , where is the j-th image reconstruction result; (6) Take the arithmetic mean fusion of K image reconstruction results Get the optimal grayscale vector value after data fusion; (7) Use G as the grayscale value of each pixel for imaging.

7. The three-dimensional electromagnetic tomography measurement device according to claim 1, characterized in that: In step (3), normalize the measurement result B norm The calculation method of (i) is B norm (i) = (B(i) - B0(i)) / B0(i), where i is the number of the excitation measurement pair.

8. The three-dimensional electromagnetic tomography measurement device according to claim 1, characterized in that: In step (4), the improved Landweber iterative method is used to calculate the unknown quantity g k+1 =T(g k -λS T (S×g k -ΔB)), where the threshold function Threshold λ is the iteration step length, which is p is the norm of the penalty term; the iteration stops when it converges to the preset value. k is the reconstructed optimal gray value vector.

9. The three-dimensional electromagnetic tomography measurement device according to claim 8, characterized in that: The p value is 0.

8.

10. The three-dimensional electromagnetic tomography measurement device according to claim 8, characterized in that: When ||g k -g k-1 The iteration stops when ||<0.001, then g k is the reconstructed optimal gray value vector.

Images