A synergistic treatment system for tin metal recovery from tantalum-niobium tailings

Abstract

The application discloses a kind of for tantalum niobium tailings in tin metal recovery's synergistic treatment system, specifically related to metallurgical solid waste resource utilization technical field, acoustic-electricity perception pre-agglomeration module exports pre-agglomeration probability index according to pre-agglomeration threshold value;Heat resistance chromatography module exports the temperature field and fluidized void fraction distribution of suspected agglomeration area;Inhibiting ingredient self-adapting module calculates inhibitor addition rate according to distribution;Wind material decoupling adjustment module exports decoupled wind speed and feed rate set value according to addition rate and bed pressure drop;Chain negative feedback blocking module detects pressure drop deviation to generate feedback correction, superimposed to wind speed set value and reversely adjust feed rate set value, while used to update pre-agglomeration threshold value;The application realizes the online identification of agglomeration precursor, accurate danger area precision inhibiting and the closed-loop cooperation of wind material decoupling, inhibits the nonlinear deterioration of coking loss flow, makes tin recovery rate keep stable when tailings fluctuation, and reduces inhibitor consumption and manual intervention.

Description

Technical Field

[0001] This invention relates to the field of metallurgical solid waste resource utilization technology, and more specifically, to a co-processing system for recovering tin metal from tantalum-niobium tailings. Background Technology

[0002] In the tantalum and niobium smelting process, a large amount of tailings containing cassiterite is generated. Industrially, tin metal is often recovered from this tailings using a high-temperature reduction volatilization method, converting cassiterite into tin vapor, which is then condensed and collected to obtain metallic tin or tin-rich products. This method can achieve a high tin recovery rate when processing materials with relatively stable compositions.

[0003] In actual production, the composition of tantalum-niobium tailings varies significantly depending on the source and batch. In particular, the content of low-melting-point alkali metal salts (such as potassium and sodium salts) and the moisture content in the tailings frequently change, sometimes accompanied by a certain amount of silicates and sulfates. When these fluctuations occur, the temperature in localized areas of the fluidized bed easily exceeds the critical agglomeration temperature, leading to melting of the particle surface and the formation of liquid bridges. The formation of liquid bridges reduces particle fluidity, deteriorates the local bed fluidization state, and subsequently causes coking and bed loss. Once loss occurs, the tin volatilization reaction will rapidly terminate, and the recovery rate may drop significantly. Simultaneously, dust carried in the high-temperature flue gas easily adheres to the inner wall of subsequent pipelines, causing blockages and affecting the continuous operation of the system. Therefore, this invention proposes a co-processing system for tin metal recovery from tantalum-niobium tailings to solve the above problems. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: A co-processing system for recovering tin metal from tantalum-niobium tailings includes: The acoustic and electrostatic sensing pre-agglomeration module collects acoustic emission signals and electrostatic signals of particles in the fluidized bed, and outputs a pre-agglomeration probability index based on the pre-agglomeration threshold. The thermal resistance synergistic chromatography module, based on the pre-agglomeration probability index, combines a temperature sensor array with electrical impedance tomography to output images of the temperature field distribution and fluidization porosity distribution in the suspected agglomeration region inside the bed. The melt-inhibiting batching adaptive module automatically calculates and outputs the inhibitor addition rate based on the degree of local agglomeration indicated in the temperature field distribution image and the fluidization porosity distribution image. The air-material decoupling adjustment module outputs the decoupled fluidizing velocity setpoint and feed rate setpoint based on the real-time detected bed pressure drop. The chain-type negative feedback blocking module detects the real-time deviation between the bed pressure drop and the target pressure drop, generates a feedback correction amount, and adds the feedback correction amount to the fluidization velocity setting value output by the air-material decoupling adjustment module. At the same time, it adjusts the feed rate setting value output by the air-material decoupling adjustment module in the opposite direction according to the feedback correction amount, and the feedback correction amount is also used to update the pre-agglomeration threshold.

[0005] In a preferred embodiment, the acoustic-electric sensing pre-agglomeration module calculates the pre-agglomeration probability index according to the following steps: The acquired acoustic emission signal is subjected to Hilbert transform within a preset time window to obtain the instantaneous amplitude envelope; The collected electrostatic signals are subjected to differential operations to obtain the electrostatic signal change rate sequence; Calculate the difference between the average value of the instantaneous amplitude envelope in the current time window and the average value of the instantaneous amplitude envelope in the previous time window; Calculate the first denominator value, which is the absolute value of the maximum instantaneous amplitude envelope value within the previous time window, plus one-thousandth of that maximum value, plus a preset absolute positive number; Divide the difference by the first denominator to obtain the first intermediate value; Furthermore, the absolute value of the first intermediate quantity is divided by the sum of the absolute value of the first intermediate quantity and one to obtain the first normalized intermediate quantity; Calculate the root mean square value of the electrostatic signal rate of change sequence within the current time window; Calculate the second denominator value, which is the root mean square value of the electrostatic signal change rate sequence in the previous time window plus one-thousandth of the root mean square value plus a preset absolute positive number; Divide the root mean square value within the current time window by the second denominator value to obtain the second intermediate quantity; Furthermore, the second intermediate quantity is divided by the sum of the second intermediate quantity and one to obtain the second normalized intermediate quantity; The product of the first normalized intermediate value and the second normalized intermediate value is multiplied by the current pre-clustering threshold to obtain the pre-clustering probability index, and the value is one when the pre-clustering probability index is greater than one. The length of the preset time window is preset based on the sampling frequency of the acoustic emission signal and the electrostatic signal and the size of the fluidized bed; the preset absolute positive number is 0.0001; when the pre-agglomeration threshold is not updated, its initial value is 0.5.

[0006] In a preferred embodiment, the thermal resistance synergistic chromatography module outputs temperature field distribution images and fluidization porosity distribution images according to the following steps: All pre-clustering probability indices calculated within the most recent preset time period are arranged in ascending order of value. When the number of samples reaches the preset minimum number of samples, the first quartile of the sequence is taken as the first interval threshold, the median as the second interval threshold, and the third quartile as the third interval threshold. When the number of samples is less than the preset minimum number of samples, the first interval threshold, the second interval threshold, and the third interval threshold are taken as the preset first default value, the second default value, and the third default value, respectively. Compare the current pre-aggregation probability index with the first interval threshold, the second interval threshold, and the third interval threshold; When the pre-agglomeration probability index is less than the first interval threshold, a circular region with the geometric center of the bed as the center and a radius equal to the bed radius is designated as a suspected agglomeration region. When the pre-agglomeration probability index is between the first interval threshold and the second interval threshold, a circular region with a radius equal to the bed radius and concentric with the circular region is designated as a suspected agglomeration region. When the pre-agglomeration probability index is between the second interval threshold and the third interval threshold, the entire bed cross section is designated as a suspected agglomeration region. When the pre-agglomeration probability index is greater than the third interval threshold, it indicates that agglomeration has been highly concentrated in the easily coagulating area. To improve the spatial resolution of tomographic imaging, an annular sidewall region extending upward from the surface of the gas distribution plate to the bed height at a third preset ratio and with a sidewall thickness equal to the bed radius at a fourth preset ratio is designated as a suspected agglomeration region. The first preset ratio is less than the second preset ratio, and the third and fourth preset ratios are preset according to the flow characteristics of particles in the fluidized bed. Within the suspected clustering area, the discrete temperature values ​​collected by the temperature sensor array are reconstructed into a continuous temperature field using radial basis function interpolation, resulting in a temperature field distribution image. Within the same region, the internal conductivity distribution is inverted from the boundary voltage measurements of electrical impedance tomography using a conjugate gradient iterative algorithm. The conductivity distribution is converted into a fluidized void ratio distribution image using a pre-calibrated power function curve relating conductivity and void ratio.

[0007] In a preferred embodiment, the specific method for calculating the inhibitor addition rate is as follows: In the temperature field distribution image, areas with temperatures exceeding the preset softening temperature are marked as high-temperature areas, and in the fluidization porosity distribution image, areas with porosity below the preset critical porosity are marked as dense areas. The intersection of high-temperature areas and dense areas is taken as the dangerous agglomeration area. The pixel area of ​​all dangerous agglomeration areas on the temperature field distribution image is counted. The pixel area is multiplied by the unit height of the bed axis to obtain the estimated volume of the dangerous agglomeration area. The estimated volume is multiplied by the preset unit volume inhibitor consumption coefficient to obtain the inhibitor addition rate. The preset softening temperature, preset critical porosity, bed axial unit height, and unit volume inhibitor consumption coefficient are all pre-calibrated based on the material characteristics and process requirements of tantalum-niobium tailings.

[0008] In a preferred embodiment, the air-material decoupling control module outputs the decoupled fluidizing velocity setpoint and feed rate setpoint according to the following steps: The total adjustment is calculated based on the deviation between the current bed pressure drop and the target pressure drop; The total adjustment is decomposed into a fluidizing velocity adjustment component and a feed rate adjustment component according to a preset proportional coefficient, and the two adjustment components have opposite signs. The fluidizing velocity adjustment component is superimposed on the fluidizing velocity setting value of the previous moment, and the feed rate adjustment component is superimposed on the feed rate setting value of the previous moment to obtain the current decoupled fluidizing velocity setting value and feed rate setting value.

[0009] In a preferred embodiment, the real-time deviation between the bed pressure drop and the target pressure drop is detected, and a feedback correction amount is generated according to the following steps: The difference between the real-time detected bed pressure drop and the target pressure drop is used as the deviation signal. This deviation signal is then passed through the proportional amplification branch, the integral accumulation branch, and the differential prediction branch in sequence. Finally, the outputs of the three branches are summed to obtain the feedback correction amount. The feedback correction is added positively to the fluidization velocity setting value output by the air-material decoupling adjustment module and negatively to the feed rate setting value output by the air-material decoupling adjustment module. At the same time, the feedback correction is normalized in absolute value and used to update the pre-agglomeration threshold.

[0010] In a preferred embodiment, the specific steps for using the feedback correction amount to update the pre-agglomeration threshold are as follows: After taking the absolute value of the feedback correction, calculate the third denominator value, which is the sum of the maximum value of all feedback corrections since the system started and one-thousandth of the maximum value, plus a preset absolute positive number. The absolute value of the feedback correction is divided by the third denominator to obtain the normalization coefficient. The normalization coefficient is multiplied by the preset adjustable range width of the pre-agglomeration threshold to obtain the threshold adjustment amount. The current value of the pre-agglomeration threshold is added to the threshold adjustment amount to obtain the updated pre-agglomeration threshold. The updated pre-agglomeration threshold is limited to the preset minimum threshold and maximum threshold. The preset adjustable range width, preset minimum threshold, and preset maximum threshold are all preset according to system stability requirements, and the preset absolute positive value is 0.0001.

[0011] The technical effects and advantages of this invention are as follows: This invention utilizes an acoustic-electro-sensing pre-agglomeration module to collect acoustic emission and electrostatic signals and output a pre-agglomeration probability index based on a pre-agglomeration threshold, enabling online identification of early signs of particle agglomeration. This module can detect anomalies in acoustic emission signal energy fluctuations and electrostatic signal change rates in the early stages of liquid bridge formation, issuing warnings before visible coking or significant pressure drop fluctuations. Compared to existing methods that rely solely on macroscopic parameters such as temperature or pressure drop, this invention shortens the agglomeration detection window from minutes to seconds, providing ample response time for subsequent adjustments. When the pre-agglomeration probability index exceeds the threshold, the system can quickly trigger subsequent controls, effectively preventing the nonlinear abrupt change from micro-agglomeration to full-bed loss of fluidization, thus ensuring the continuous tin volatilization reaction and significantly improving the stability of the recovery process.

[0012] This invention utilizes a thermal resistance-assisted tomography module, based on a pre-agglomeration probability index combined with a temperature sensor array and electrical impedance tomography, to output temperature field distribution images and fluidization porosity distribution images of suspected agglomeration areas. Combined with an adaptive melt-inhibition batching module, this automatically calculates and outputs the inhibitor addition rate based on these distributions. This design combines agglomeration localization with precise batching, ensuring that inhibitors are added only to areas where dangerous agglomeration occurs, avoiding waste caused by uniform dosing across the entire bed and potential interference with normal fluidization zones. Dangerous agglomeration areas are identified by the intersection of the temperature field and porosity images. The volume is then estimated using pixel area and unit height along the bed axis, multiplied by the inhibitor consumption coefficient per unit volume to obtain the addition rate, achieving on-demand adjustment of inhibitor dosage. This reduces inhibitor consumption while effectively suppressing the melting and agglomeration of low-melting-point components, extending the continuous operation cycle of the fluidized bed.

[0013] This invention uses a chain-like negative feedback blocking module to detect the real-time deviation between the bed pressure drop and the target pressure drop, generating a feedback correction amount. This correction amount is then superimposed on the fluidization velocity setpoint output by the air-material decoupling adjustment module, while simultaneously adjusting the feed rate setpoint in the reverse direction. This feedback correction amount is also used to update the pre-agglomeration threshold. This structure constitutes a fully closed-loop collaborative control system from agglomeration identification to air-material decoupling and threshold adaptation. When a positive deviation in pressure drop occurs, the feedback correction amount increases the air velocity with positive polarity and decreases the feed rate with negative polarity, quickly suppressing bed expansion or clogging tendencies. Simultaneously, after absolute value normalization, this correction amount dynamically adjusts the pre-agglomeration threshold, enabling the system to adapt to the fluctuating characteristics of different batches of tailings. This multi-stage linkage and threshold self-tuning mechanism effectively breaks the positive feedback chain of local agglomeration → uneven fluidization → increased temperature → intensified coking, allowing the tin recovery rate to remain at its original level even when the tailings composition fluctuates drastically, and the system possesses the ability to automatically recover to a steady state. Attached Figure Description

[0014] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a schematic diagram of a co-processing system for recovering tin metal from tantalum-niobium tailings according to the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0016] Reference Figure 1 The following examples were obtained: Example 1: A co-processing system for tin metal recovery from tantalum-niobium tailings, comprising: The acoustic and electrostatic sensing pre-agglomeration module collects acoustic emission signals and electrostatic signals of particles in the fluidized bed, and outputs a pre-agglomeration probability index based on the pre-agglomeration threshold. The thermal resistance synergistic chromatography module, based on the pre-agglomeration probability index, combines a temperature sensor array with electrical impedance tomography to output images of the temperature field distribution and fluidization porosity distribution in the suspected agglomeration region inside the bed. The melt-inhibiting batching adaptive module automatically calculates and outputs the inhibitor addition rate based on the degree of local agglomeration indicated in the temperature field distribution image and the fluidization porosity distribution image. The air-material decoupling adjustment module outputs the decoupled fluidizing velocity setpoint and feed rate setpoint based on the real-time detected bed pressure drop. The chain-type negative feedback blocking module detects the real-time deviation between the bed pressure drop and the target pressure drop, generates a feedback correction amount, and adds the feedback correction amount to the fluidization velocity setting value output by the air-material decoupling adjustment module. At the same time, it adjusts the feed rate setting value output by the air-material decoupling adjustment module in the opposite direction according to the feedback correction amount, and the feedback correction amount is also used to update the pre-agglomeration threshold.

[0017] The acoustic-electric sensing pre-agglomeration module calculates the pre-agglomeration probability index according to the following steps: The acquired acoustic emission signal is subjected to Hilbert transform within a preset time window to obtain the instantaneous amplitude envelope. Specifically, the DC component is first removed from the acoustic emission signal sequence within the current time window, i.e., the arithmetic mean of all sampling points within the window is subtracted, resulting in a zero-mean signal. Then, a Fast Fourier Transform (FFT) is performed on this zero-mean signal to obtain its frequency domain representation. The Fourier coefficients corresponding to the negative frequency components in the frequency domain representation are set to zero, while the coefficients of the positive frequency components are multiplied by two. An Inverse Fourier Transform is then performed to obtain the analytic signal. Finally, the modulus of the analytic signal is taken, which is the instantaneous amplitude envelope. This process is equivalent to convolving the original signal with a filter whose impulse response is 1 / πt and then taking the modulus, where t represents time and π is the constant pi. The length of the preset time window is set according to the sampling frequency of the acoustic emission signal and the diameter of the fluidized bed. The sampling frequency is usually between 1 kHz and 10 kHz. The larger the bed diameter, the longer the window length. The typical window length is between 1 second and 10 seconds. For example, when the sampling frequency is 2 kHz and the bed diameter is 1 meter, the window length is 5 seconds.

[0018] The acquired electrostatic signals are subjected to differential operations to obtain a sequence of electrostatic signal change rates. The differential operation involves calculating the difference between two adjacent sampling points and dividing by the sampling interval, reflecting the drastic instantaneous changes in the electrostatic signal. The difference between the average value of the instantaneous amplitude envelope within the current time window and the average value of the instantaneous amplitude envelope within the previous time window is calculated. This difference represents the trend of acoustic emission signal energy change between two adjacent time windows; a positive value indicates energy enhancement, and a negative value indicates energy decay.

[0019] Calculate the first denominator value, which is the absolute value of the maximum instantaneous amplitude envelope within the previous time window, plus one-thousandth of that maximum value, plus a preset absolute positive number. The preset absolute positive number is a value between 0.0001 and 0.001, for example, 0.0005. The rationale is that after the signal is normalized to the 0-1 interval, this value is much smaller than the minimum resolvable change of the signal, while avoiding a denominator of zero. Divide the difference by the first denominator value to obtain the first intermediate quantity. Divide the absolute value of this first intermediate quantity by the sum of its absolute value and one to obtain the first normalized intermediate quantity. This normalization operation maps any non-negative real number to an open interval between 0 and 1. For example, if the absolute value of the first intermediate quantity is 2, then the first normalized intermediate quantity is 2 divided by 3, which equals 0.6667; if it is 0, the result is 0.

[0020] Calculate the root mean square (RMS) value of the electrostatic signal rate of change sequence within the current time window. The RMS value is the square root of the sum of the squares of each sample point, divided by the number of sample points, and used to characterize the average intensity of the signal rate of change.

[0021] Calculate the second denominator value, which is the root mean square value of the electrostatic signal rate of change sequence in the previous time window, plus one-thousandth of that root mean square value, plus a preset absolute positive number. This preset absolute positive number is the same as the previously mentioned value, ranging from 0.0001 to 0.001. Divide the root mean square value in the current time window by the second denominator value to obtain the second intermediate quantity. Divide this second intermediate quantity by the sum of the second intermediate quantity and one to obtain the second normalized intermediate quantity. This normalization logic is the same as the first normalized intermediate quantity, ensuring that the output is between 0 and 1.

[0022] The pre-agglomeration probability index is obtained by multiplying the product of the first and second normalized intermediate values ​​by the current pre-agglomeration threshold. The current pre-agglomeration threshold is a dynamically updated parameter, with an initial value set between 0.4 and 0.6, for example, 0.5. This initial value is determined based on the statistical results of the pre-agglomeration probability index under background noise signals during fluidized bed operation without load. When the pre-agglomeration probability index is greater than one, it is set to one to conform to the probability definition.

[0023] The length of the preset time window is pre-set based on the sampling frequencies of the acoustic emission and electrostatic signals and the size of the fluidized bed. Specifically, the window length is shortened by 0.2 seconds for every 1 kHz increase in sampling frequency and increased by 1 second for every 0.5 m increase in bed diameter, with a typical range of 1 to 10 seconds. The preset absolute positive value is between 0.0001 and 0.001. This value ensures that the denominator is non-zero and does not significantly affect the calculation accuracy when the signal is normalized to the 0 to 1 range. The initial value of the pre-agglomeration threshold before it is updated is between 0.4 and 0.6. This range was obtained through statistical analysis of the acoustic and electrostatic signals of different tantalum-niobium tailings samples using a laboratory-scale test setup.

[0024] The thermal resistance synergistic chromatography module outputs temperature field distribution images and fluidization porosity distribution images according to the following steps: All pre-agglomeration probability indices calculated within a recent preset time period are arranged in ascending order of value. This preset time period is set based on the typical timescale of aggregation phenomena in a fluidized bed from their inception to their development to a detectable level. Experimental data shows that a time period between 30 and 120 seconds can balance response speed and statistical stability; for example, 60 seconds. When the sample size reaches the preset minimum sample size, the first quartile of the sequence is taken as the first interval threshold, the median as the second interval threshold, and the third quartile as the third interval threshold. The preset minimum sample size is determined to be 4 based on the minimum sample size required for quartile calculation. For example, if the sampling frequency is calculated every 5 seconds within 60 seconds, 12 samples are obtained, which meets the condition. When the sample size is less than the preset minimum sample size, the first interval threshold, second interval threshold, and third interval threshold are set to the preset first default value, second default value, and third default value, respectively. The first, second, and third default values ​​are determined based on historical statistical data of tantalum and niobium tailings under normal fluidized bed operating conditions, with typical value ranges of 0.2 to 0.3, 0.4 to 0.6, and 0.7 to 0.8, respectively, for example, 0.25, 0.50, and 0.75.

[0025] The current pre-agglomeration probability index is compared with the first interval threshold, the second interval threshold, and the third interval threshold. The comparison result is used to determine the magnitude relationship. For example, if the current pre-agglomeration probability index is 0.62, the first interval threshold is 0.25, the second interval threshold is 0.50, and the third interval threshold is 0.75, then 0.62 is between the second and third interval thresholds.

[0026] When the pre-agglomeration probability index is less than the first interval threshold, a circular region with the geometric center of the bed as its center and a radius equal to the bed radius is designated as a suspected agglomeration region. The first preset ratio is set based on the principle that agglomeration rarely occurs under low probability conditions, and only the central region needs to be monitored; an empirical value of 0.25 to 0.35 is used, for example, 0.3. When the pre-agglomeration probability index is between the first and second interval thresholds, a circular region with a radius equal to the bed radius and concentric with the aforementioned circular region is designated as a suspected agglomeration region. The second preset ratio is greater than the first preset ratio, and is taken as 0.5 to 0.7, for example, 0.6. When the pre-agglomeration probability index is between the second and third interval thresholds, the entire bed cross-section is designated as a suspected agglomeration region. When the pre-agglomeration probability index is greater than the third interval threshold, it indicates that agglomeration has become highly concentrated in the easily coagulating region. To improve the spatial resolution of the tomographic imaging, an annular sidewall region extending upward from the gas distribution plate surface to the bed height at a third preset ratio, with a sidewall thickness equal to the bed radius, is designated as a suspected agglomeration region. The third preset ratio is based on the fact that coking mainly occurs at the bottom, and is set to 0.2 to 0.4, for example, 0.3. The fourth preset ratio is based on fluidization experiment data showing that agglomeration is prone to occur in the sidewall region, and is set to 0.1 to 0.3, for example, 0.2.

[0027] Within a suspected clustering area, discrete temperature values ​​collected by the temperature sensor array are reconstructed into a continuous temperature field using radial basis function interpolation, resulting in a temperature field distribution image. The specific implementation of radial basis function interpolation is as follows: Each temperature sensor's spatial coordinates are used as nodes, and the temperature measured at that node is used as the function value. Multiple quadratic surface basis functions are selected. The form of this basis function is the square root of a bracket plus a bracket shape parameter multiplied by the square of the distance from the bracket, where the distance is the Euclidean distance between the interpolation point and the sensor node, and the shape parameter ranges from 0.1 to 0.5. The weighting coefficients of each basis function are obtained by solving a system of linear equations. Then, a weighted sum is calculated for any location within the region to obtain the continuous temperature value. For example, if 16 thermocouples are arranged on a bed cross-section, interpolation can be used to obtain a distribution image of one temperature data point per square centimeter across the entire cross-section.

[0028] Within the same region, the internal conductivity distribution is inverted using the conjugate gradient iterative algorithm on the boundary voltage measurements from electrical impedance tomography. The specific implementation of the conjugate gradient iterative algorithm is as follows: A finite element model is established, and the region is divided into a grid. The initial conductivity distribution is set to a uniform value. The boundary voltage is calculated using the forward problem, and the residual is obtained by comparing it with the measured boundary voltage. The objective function is constructed as the sum of squared residuals. The conductivity distribution is updated by searching along the negative gradient direction using the conjugate gradient method. The number of iterations is set to 20 to 50 steps until the residual decreases to less than one percent of the initial residual. For example, in a 32-electrode system, the conductivity correction is calculated for 2000 grid cells in each iteration, ultimately obtaining the conductivity value for each cell.

[0029] The conductivity distribution is converted into a fluidized bed porosity distribution image using a pre-calibrated power function relationship curve between conductivity and porosity. The calibration method for the power function relationship is as follows: In a laboratory fluidized bed, standard materials with different porosities are used, with samples taken at intervals of 0.1 porosities from 0.4 to 0.9. Their conductivity is measured, and the data points are fitted to the form where conductivity equals a coefficient multiplied by a power of the porosity, where the coefficient and the exponent are the fitting results. For example, for tantalum-niobium tailings particles, the calibration result is a coefficient of 0.12 Siemens per meter and an exponent of 1.8. After obtaining the power function relationship, the conductivity of each grid cell is substituted into the inverse function, i.e., the porosity equals the conductivity divided by the coefficient and the exponent raised to the power of one. This gives the porosity of that cell. All cells are combined to form the porosity distribution image.

[0030] The specific method for calculating the inhibitor addition rate is as follows: Regions in the temperature field distribution image whose temperature exceeds the preset softening temperature are marked as high-temperature regions. The preset softening temperature is determined based on the softening point of low-melting-point alkali metal salts in tantalum-niobium tailings. Experimental data shows that the softening temperature range is between 600 degrees Celsius and 800 degrees Celsius. For example, if we take 700 degrees Celsius, and the temperature value of a certain pixel is 720 degrees Celsius, then it is marked as a high-temperature region.

[0031] In the fluidized bed porosity distribution image, regions with porosity below a preset critical porosity are marked as dense regions. The preset critical porosity is set according to the critical value at which particles in the fluidized bed begin to show significant contact and friction. The empirical value is 0.4 to 0.5, for example, 0.45. If the porosity of a certain pixel is 0.35, it is marked as a dense region.

[0032] The intersection of the high-temperature region and the dense region is taken as the dangerous clustering region. The intersection operation is a pixel-level logical AND, that is, pixels that simultaneously meet the conditions of high temperature and denseness constitute dangerous clustering regions. For example, if a region has a temperature of 750 degrees Celsius and a porosity of 0.4, it belongs to the dangerous clustering region.

[0033] The pixel area of ​​all dangerous clusters on the temperature field distribution image is calculated. The pixel area is the total number of pixels in the dangerous cluster area multiplied by the actual physical area corresponding to a single pixel. The area of ​​a single pixel is determined by the spatial resolution of the temperature sensor array. For example, if each pixel corresponds to 1 square centimeter, then there are 500 pixels in the dangerous cluster area, and the area is 500 square centimeters.

[0034] The estimated volume of the dangerous agglomeration region is obtained by multiplying the pixel area by the unit height of the bed axis. The unit height of the bed axis is determined based on the slice thickness of electrical impedance tomography, with a typical value of 1 cm to 5 cm. For example, if we take 2 cm, the estimated volume is 500 square centimeters multiplied by 2 cm, which equals 1000 cubic centimeters.

[0035] Multiply the estimated volume by a pre-set inhibitor consumption coefficient per unit volume to obtain the inhibitor addition rate. The inhibitor consumption coefficient per unit volume is determined based on the chemical composition of tantalum-niobium tailings and the experimental calibration results of inhibitor neutralization efficiency. The value ranges from 0.01 g to 0.05 g per cubic centimeter. For example, if we take 0.02 g per cubic centimeter, then the addition rate is 1000 cubic centimeters multiplied by 0.02 g per cubic centimeter, which equals 20 g per second.

[0036] The preset softening temperature, preset critical porosity, axial height per unit volume of the bed, and inhibitor consumption coefficient per unit volume are all pre-calibrated based on the material characteristics and process requirements of tantalum-niobium tailings. The calibration method involves taking representative tailings samples and conducting inhibition tests at different temperatures and porosities in a laboratory fluidized bed, recording the parameters at which the inhibition effect is optimal as calibration values. For example, for tailings from a certain mine, the calibrated preset softening temperature is 700 degrees Celsius, the preset critical porosity is 0.45, the axial height per unit volume of the bed is 2 cm, and the inhibitor consumption coefficient per unit volume is 0.02 g / cm³. These parameters can be fine-tuned according to the fluctuations in the composition of different batches of tailings.

[0037] The air-material decoupling control module outputs the decoupled fluidizing velocity setpoint and feed rate setpoint according to the following steps: The total adjustment is calculated based on the deviation between the current bed pressure drop and the target pressure drop. This deviation is the measured current bed pressure drop minus the target pressure drop, expressed in Pascals. The total adjustment is calculated using proportional control, meaning the total adjustment equals the deviation multiplied by the proportional gain coefficient. The proportional gain coefficient is obtained through on-site tuning based on the pressure drop response characteristics of the fluidized bed, with a value ranging from 0.2 to 2.0, for example, 0.8. If the current bed pressure drop is 5000 Pa, the target pressure drop is 4500 Pa, and the deviation is 500 Pa, then the total adjustment is 500 multiplied by 0.8, which equals 400.

[0038] The total adjustment is decomposed into a fluidization velocity adjustment component and a feed rate adjustment component according to a preset proportional coefficient, with the two components having opposite signs. The preset proportional coefficient is pre-calibrated based on the air-material decoupling matrix, with a value range of 0.3 to 0.7, for example, 0.5. The fluidization velocity adjustment component is equal to the total adjustment multiplied by this proportional coefficient, and the feed rate adjustment component is equal to the total adjustment multiplied by one minus this proportional coefficient, with one component taking a negative sign. Specifically, if the total adjustment is positive, the fluidization velocity adjustment component is positive, and the feed rate adjustment component is negative, indicating that the air velocity is increased while the feed rate is decreased; conversely, the opposite is true. For example, if the total adjustment is 400 and the proportional coefficient is 0.5, then the fluidization velocity adjustment component is 200, and the feed rate adjustment component is -200.

[0039] The fluidizing velocity adjustment component is superimposed on the previous fluidizing velocity setpoint, and the feed rate adjustment component is superimposed on the previous feed rate setpoint to obtain the current decoupled fluidizing velocity and feed rate setpoints. The previous setpoints are stored in system memory, and the initial values ​​are set according to process requirements. The initial range of the fluidizing velocity setpoint is 0.5 m / s to 2.0 m / s, and the initial range of the feed rate setpoint is 10 kg / h to 50 kg / h. The superposition operation uses addition. For example, if the previous fluidizing velocity setpoint was 1.2 m / s and the adjustment component was 0.1 m / s, the current setpoint is 1.3 m / s. If the previous feed rate was 30 kg / h and the adjustment component was -3 kg / h, the current setpoint is 27 kg / h. These two values ​​are output for use by subsequent control loops.

[0040] The real-time deviation between the bed pressure drop and the target pressure drop is detected, and a feedback correction is generated according to the following steps: The real-time bed pressure drop is measured by a differential pressure transmitter installed in the dense phase zone of the fluidized bed, with a sampling frequency of 1 Hz to 10 Hz, for example, reading once every 0.2 seconds. The target pressure drop is preset according to process requirements, typically ranging from 3000 Pa to 6000 Pa, for example, 4500 Pa. The difference between the real-time measured bed pressure drop and the target pressure drop is used as the deviation signal. If the current measured pressure drop is 5000 Pa and the target pressure drop is 4500 Pa, then the deviation signal is 500 Pa.

[0041] The deviation signal is passed sequentially through a proportional amplification branch, an integral accumulation branch, and a differential prediction branch. The outputs of the three branches are then summed to obtain the feedback correction amount, specifically: The output of the proportional amplifier branch is the deviation signal multiplied by a proportional coefficient. This proportional coefficient is set according to the proportional relationship between the pressure drop deviation and the wind speed regulation amount, and the value range is from 0.1 to 1.0. For example, if it is 0.3, then the proportional output is 500 multiplied by 0.3 equals 150. The output of the integral accumulation branch is the integral of the deviation signal over time multiplied by the integration coefficient. The integration coefficient ranges from 0.01 to 0.1. For example, if it is 0.05, the integration time constant is 1 second. If the average deviation over the past second is 500 Pa, then the integral output is 500 multiplied by 1 multiplied by 0.05, which equals 25. The output of the differential prediction branch is the rate of change of the deviation signal multiplied by the differential coefficient. The rate of change is the difference between two adjacent deviations divided by the sampling interval. The differential coefficient ranges from 0.05 to 0.5. For example, if the deviation changes from 500 Pa to 520 Pa, the rate of change is 20 Pa per second. Then the differential output is 20 multiplied by 0.1, which equals 2. The sum of the outputs from the three branches is 150 plus 25 plus 2, which equals 177. This value is the feedback correction amount.

[0042] The feedback correction is positively added to the fluidization velocity setting output by the air-material decoupling adjustment module and negatively added to the feed rate setting output by the same module. Positive addition means the feedback correction is added to the original velocity setting, while negative addition means the feedback correction is subtracted from the original feed rate setting. For example, if the fluidization velocity setting output by the air-material decoupling module is 1.3 m / s and the feedback correction is equivalent to 0.1 m / s, the resulting velocity setting is 1.4 m / s. If the feed rate setting is 27 kg / h, subtracting the feed rate conversion value corresponding to 0.1 (the conversion factor is determined based on the process, for example, 1 kg / h per unit of feedback correction) results in a feed rate of 26.9 kg / h. Meanwhile, the feedback correction amount is used to update the pre-agglomeration threshold after absolute value normalization. The specific method of absolute value normalization is to divide the absolute value of the feedback correction amount by a reference value so that the result falls in the range of 0 to 1. For example, if the reference value is the maximum allowable correction amount of 200, then the absolute value of 177 is 0.885 after normalization.

[0043] The specific steps for using feedback correction to update the pre-agglomeration threshold are as follows: After taking the absolute value of the feedback correction, the third denominator is calculated. This third denominator is the sum of the maximum value of all feedback corrections since system startup and one-thousandth of that maximum value, plus a preset absolute positive number. The maximum value of all feedback corrections since system startup is obtained by comparing and updating the storage of each generated feedback correction in real time. One-thousandth of the maximum value is used to prevent the denominator from being too small, and the preset absolute positive number is used to further ensure that the denominator is always positive. This preset absolute positive number is set according to the minimum resolvable value after signal normalization and takes a value between 0.0001 and 0.001, for example, 0.0001. If the maximum value of the feedback correction since system startup is 200, then one-thousandth of the maximum value is 0.2, and the preset absolute positive number is 0.0001. Therefore, the third denominator is 200 plus 0.2 plus 0.0001, which equals 200.2001.

[0044] Divide the absolute value of the feedback correction by the third denominator to obtain the normalization coefficient. The normalization coefficient maps the absolute value of the feedback correction to an open interval between 0 and 1. For example, if the absolute value of the feedback correction is 177 and the third denominator is 200.2001, then the normalization coefficient is 177 divided by 200.2001, which is approximately equal to 0.884.

[0045] The threshold adjustment amount is obtained by multiplying the normalization coefficient by the preset adjustable range width of the pre-agglomeration threshold. The typical value range of the preset adjustable range width is 0.2 to 0.6. For example, if it is 0.4, then the threshold adjustment amount is 0.884 multiplied by 0.4 equals 0.3536.

[0046] The current value of the pre-agglomeration threshold is then added to the threshold adjustment amount to obtain the updated pre-agglomeration threshold. For example, if the current pre-agglomeration threshold is 0.5, the updated threshold is 0.5 plus 0.3536, which equals 0.8536. The updated pre-agglomeration threshold is limited to between a preset minimum threshold and a preset maximum threshold. The preset minimum threshold typically ranges from 0.1 to 0.2, for example, 0.1. The preset maximum threshold typically ranges from 0.8 to 0.9, for example, 0.9. Therefore, 0.8536 is limited to 0.9; if it is lower than 0.1, it is set to 0.1.

[0047] The preset adjustable range width, preset minimum threshold, and preset maximum threshold are all pre-set according to system stability requirements, and their typical value ranges are described above. System stability requirements are obtained through fluidized bed cold-state experiments and hot-state operation data: an adjustable range width of 0.2 to 0.6 ensures that threshold updates are neither sluggish nor overly sensitive; a minimum threshold of 0.1 to 0.2 avoids frequent false triggers due to excessively low thresholds; and a maximum threshold of 0.8 to 0.9 avoids threshold saturation and loss of adjustment capability. The preset absolute positive number ranges from 0.0001 to 0.001, for example, 0.0001. This value is much smaller than one-thousandth of the maximum feedback correction value (usually greater than 1), and is only used to avoid the risk of division by zero, without affecting normalization accuracy.

[0048] The above-mentioned models or function formulas are all dimensionless and numerical calculations. The models or function formulas are obtained by software simulation based on a large amount of collected data to obtain the most recent real situation. The preset parameters in the models or function formulas are set by those skilled in the art according to the actual situation.

[0049] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0050] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0051] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0052] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A synergistic processing system for tin metal recovery from tantalum-niobium tailings, characterized by, include: The acoustic and electrostatic sensing pre-agglomeration module collects acoustic emission signals and electrostatic signals of particles in the fluidized bed, and outputs a pre-agglomeration probability index based on the pre-agglomeration threshold. The thermal resistance synergistic chromatography module, based on the pre-agglomeration probability index, combines a temperature sensor array with electrical impedance tomography to output images of the temperature field distribution and fluidization porosity distribution in the suspected agglomeration region inside the bed. The melt-inhibiting batching adaptive module automatically calculates and outputs the inhibitor addition rate based on the degree of local agglomeration indicated in the temperature field distribution image and the fluidization porosity distribution image. The air-material decoupling adjustment module outputs the decoupled fluidizing velocity setpoint and feed rate setpoint based on the real-time detected bed pressure drop. The chain-type negative feedback blocking module detects the real-time deviation between the bed pressure drop and the target pressure drop, generates a feedback correction amount, and adds the feedback correction amount to the fluidization velocity setting value output by the air-material decoupling adjustment module. At the same time, it adjusts the feed rate setting value output by the air-material decoupling adjustment module in the opposite direction according to the feedback correction amount, and the feedback correction amount is also used to update the pre-agglomeration threshold.

2. The synergistic processing system for tin metal recovery from tantalum-niobium tailings according to claim 1, characterized in that, The acoustic-electric sensing pre-agglomeration module calculates the pre-agglomeration probability index according to the following steps: The acquired acoustic emission signal is subjected to Hilbert transform within a preset time window to obtain the instantaneous amplitude envelope; The collected electrostatic signals are subjected to differential operations to obtain the electrostatic signal change rate sequence; Calculate the difference between the average value of the instantaneous amplitude envelope in the current time window and the average value of the instantaneous amplitude envelope in the previous time window; Calculate the first denominator value, which is the absolute value of the maximum instantaneous amplitude envelope value within the previous time window, plus one-thousandth of that maximum value, plus a preset absolute positive number; Divide the difference by the first denominator to obtain the first intermediate value; Furthermore, the absolute value of the first intermediate quantity is divided by the sum of the absolute value of the first intermediate quantity and one to obtain the first normalized intermediate quantity; Calculate the root mean square value of the electrostatic signal rate of change sequence within the current time window; Calculate the second denominator value, which is the root mean square value of the electrostatic signal change rate sequence in the previous time window plus one-thousandth of the root mean square value plus a preset absolute positive number; Divide the root mean square value within the current time window by the second denominator value to obtain the second intermediate quantity; Furthermore, the second intermediate quantity is divided by the sum of the second intermediate quantity and one to obtain the second normalized intermediate quantity; The product of the first normalized intermediate value and the second normalized intermediate value is multiplied by the current pre-clustering threshold to obtain the pre-clustering probability index, and the value is one when the pre-clustering probability index is greater than one.

3. The synergistic processing system for tin metal recovery from tantalum-niobium tailings according to claim 1, characterized in that, The thermal resistance synergistic chromatography module outputs temperature field distribution images and fluidization porosity distribution images according to the following steps: All pre-clustering probability indices calculated within the most recent preset time period are arranged in ascending order of value. When the number of samples reaches the preset minimum number of samples, the first quartile of the sequence is taken as the first interval threshold, the median as the second interval threshold, and the third quartile as the third interval threshold. When the number of samples is less than the preset minimum number of samples, the first interval threshold, the second interval threshold, and the third interval threshold are taken as the preset first default value, the second default value, and the third default value, respectively. Compare the current pre-aggregation probability index with the first interval threshold, the second interval threshold, and the third interval threshold; When the pre-agglomeration probability index is less than the first interval threshold, a circular region with the geometric center of the bed as the center and a radius equal to the bed radius is designated as a suspected agglomeration region. When the pre-agglomeration probability index is between the first and second interval thresholds, a circular region with a radius equal to the bed radius and concentric with the circular region is designated as a suspected agglomeration region. When the pre-agglomeration probability index is between the second and third interval thresholds, the entire bed cross-section is designated as a suspected agglomeration region. When the pre-agglomeration probability index is greater than the third interval threshold, it indicates that agglomeration has been highly concentrated in the easily coagulating region. To improve the spatial resolution of tomographic imaging, an annular sidewall region extending upward from the surface of the gas distribution plate to the bed height at a third preset ratio and with a sidewall thickness equal to the bed radius at a fourth preset ratio is designated as a suspected agglomeration region. The first preset ratio is less than the second preset ratio. Within the suspected clustering area, the discrete temperature values ​​collected by the temperature sensor array are reconstructed into a continuous temperature field using radial basis function interpolation, resulting in a temperature field distribution image. Within the same region, the internal conductivity distribution is inverted from the boundary voltage measurements of electrical impedance tomography using a conjugate gradient iterative algorithm. The conductivity distribution is converted into a fluidized void ratio distribution image using a pre-calibrated power function curve relating conductivity and void ratio.

4. The co-processing system for tin metal recovery from tantalum-niobium tailings according to claim 1, characterized in that, The specific method for calculating the inhibitor addition rate is as follows: Regions in the temperature field distribution image whose temperature exceeds the preset softening temperature are marked as high-temperature regions. Regions in the fluidization porosity distribution image whose porosity is lower than the preset critical porosity are marked as dense regions. The intersection of the high-temperature region and the dense region is taken as the dangerous agglomeration region. The pixel area of ​​all dangerous agglomeration regions on the temperature field distribution image is counted. The pixel area is multiplied by the unit height of the bed axis to obtain the estimated volume of the dangerous agglomeration region. The estimated volume is multiplied by the preset unit volume inhibitor consumption coefficient to obtain the inhibitor addition rate.

5. A co-processing system for recovering tin metal from tantalum-niobium tailings according to claim 1, characterized in that, The air-material decoupling control module outputs the decoupled fluidizing velocity setpoint and feed rate setpoint according to the following steps: The total adjustment is calculated based on the deviation between the current bed pressure drop and the target pressure drop; The total adjustment is decomposed into a fluidizing velocity adjustment component and a feed rate adjustment component according to a preset proportional coefficient, and the two adjustment components have opposite signs. The fluidizing velocity adjustment component is superimposed on the fluidizing velocity setting value of the previous moment, and the feed rate adjustment component is superimposed on the feed rate setting value of the previous moment to obtain the current decoupled fluidizing velocity setting value and feed rate setting value.

6. A co-processing system for tin metal recovery from tantalum-niobium tailings according to claim 1, characterized in that, The real-time deviation between the bed pressure drop and the target pressure drop is detected, and a feedback correction is generated according to the following steps: The difference between the real-time detected bed pressure drop and the target pressure drop is used as the deviation signal. This deviation signal is then passed through the proportional amplification branch, the integral accumulation branch, and the differential prediction branch in sequence. Finally, the outputs of the three branches are summed to obtain the feedback correction amount. The feedback correction is added positively to the fluidization velocity setting value output by the air-material decoupling adjustment module and negatively to the feed rate setting value output by the air-material decoupling adjustment module. At the same time, the feedback correction is normalized in absolute value and used to update the pre-agglomeration threshold.

7. A co-processing system for recovering tin metal from tantalum-niobium tailings according to claim 1, characterized in that, The specific steps for using feedback correction to update the pre-agglomeration threshold are as follows: After taking the absolute value of the feedback correction, calculate the third denominator value, which is the sum of the maximum value of all feedback corrections since the system started and one-thousandth of the maximum value, plus a preset absolute positive number. The absolute value of the feedback correction is divided by the third denominator to obtain the normalization coefficient. The normalization coefficient is multiplied by the preset adjustable range width of the pre-agglomeration threshold to obtain the threshold adjustment amount. The current value of the pre-agglomeration threshold is added to the threshold adjustment amount to obtain the updated pre-agglomeration threshold. The updated pre-agglomeration threshold is limited to the preset minimum threshold and maximum threshold.

Images