A method for classifying coarse and fine red mud

Abstract

This invention provides a method for classifying coarse and fine red mud, relating to the field of bauxite technology, comprising the following steps: inputting red mud slurry into a tank equipped with a stirring system to disperse red mud particles in the slurry, while simultaneously adding water to control rheological properties; inputting the red mud slurry into a gravity separation device for separation, causing coarse red mud to settle to the bottom to form underflow concentrate, while fine red mud and water are discharged as overflow tailings; diluting the overflow tailings obtained from the first separation with water again, and inputting them again into another gravity separation device to separate and form secondary underflow concentrate, while the remaining fine particles and water are discharged as final overflow tailings; this invention uses multi-stage gravity classification as the initial classification, ultrasonic classification to enhance the fineness of the overflow tailings classification, and, based on the principle of sound wave reflection, unifies the classification to a size distribution classification, resulting in a more complete classification benchmark, more reasonable grade matching, and ensuring the rationality of subsequent recycling and use.

Description

Technical Field

[0001] This invention relates to the field of bauxite technology, and more particularly to a method for classifying coarse and fine red mud. Background Technology

[0002] Alumina production generates a large amount of waste, red mud. Red mud is a red, powdery, highly water-containing, strongly alkaline solid waste remaining after aluminum is extracted from alumina in the aluminum industry. Red mud has a complex composition, containing alkali and small amounts of radioactive substances. The percentages of its main chemical components are shown in the table below. Due to its fine particle size, most domestic red mud dumps currently use either wet storage or dehydration followed by long-term storage. The former allows water to seep into the ground, polluting groundwater, while the latter, after long-term drying, easily causes dust to fly, seriously polluting the environment and endangering human health. While red mud is an industrial waste causing environmental pollution, it also has resource value. Laboratory experiments have shown that valuable metals such as Ca, Te, and Ti can be recovered from red mud. Based on the principles of reducing the amount and harm of solid waste, fully and rationally utilizing solid waste, and carrying out harmless disposal, the development and utilization of red mud can effectively promote environmental cleanliness, energy conservation and emission reduction, and the development of a circular economy. Therefore, the classification of red mud is essential in its development and utilization.

[0003] Traditional red mud grading mainly involves the following method: adding red mud slurry to a dilution tank, then sending the diluted slurry to a hydrocyclone assembly for coarse and fine red mud grading. The underflow from the hydrocyclone enters the coarse red mud slurry tank and is sent to the coarse red mud treatment system in the settling process. The overflow from the hydrocyclone enters the fine red mud slurry tank and is sent to the fine red mud treatment system in the settling process. However, the above method mainly relies on gravity settling grading. The different components within red mud are complex, and gravity may not necessarily match their shape and size. In addition to gravity grading, the quality of red mud also needs to consider factors such as size, shape, and distribution. Red mud particles graded by different gravity methods may not have the same size, shape, and distribution as the gravity grading level, affecting the rationality of subsequent recycling. Therefore, this invention proposes a method for grading coarse and fine red mud to solve the problems existing in the prior art. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a method for classifying coarse and fine red mud. This method uses multi-stage gravity classification as the initial classification, ultrasonic classification to enhance the fineness of overflow tailings classification, and the principle of sound wave reflection to unify the classification based on the above classification to a size distribution classification. This results in a more complete classification benchmark, more reasonable level matching, and ensures the rationality of subsequent recycling and use.

[0005] To achieve the objectives of this invention, the invention is implemented through the following technical solution: a method for classifying coarse and fine red mud, comprising the following steps:

[0006] S1: The red mud slurry is fed into a tank equipped with a stirring system, so that the red mud particles are dispersed in the slurry, and water is added simultaneously to control the rheological properties;

[0007] S2: The red mud slurry is fed into the gravity separation equipment for separation, so that the coarse red mud settles to the bottom to form the underflow concentrate, and the fine red mud and water are discharged as overflow tailings.

[0008] S3: The overflow tailings obtained from the first separation are diluted with water again and fed into another gravity separation device to separate and form secondary underflow concentrate. The remaining fine particles and water are discharged as the final overflow tailings.

[0009] S4: The final overflow tailings are fed into an ultrasonic classifier to further classify the liquid by utilizing the cavitation effect, micro-jet effect and acoustic flow effect of ultrasound in the liquid.

[0010] S5: Collect red mud slurry of different particle sizes after further classification at different parts of the ultrasonic classification equipment.

[0011] S6: Combining the above steps, collect the red mud slurry after N-level grading, emit sound waves of the same frequency into the N-level red mud slurry through a sound wave transmitter, and receive the reflected sound waves through a receiver.

[0012] S7: Process and analyze the received reflected acoustic wave signal, extract signal feature parameters, and infer the size and distribution of different grades of red mud slurry based on the signal feature parameters;

[0013] S8: Set a classification threshold and further divide and combine the N-grade red mud slurry according to its size and distribution to obtain a maximum of four-grade products.

[0014] A further improvement is that S1 includes the following steps:

[0015] Red mud slurry is added to a tank equipped with a stirring system, and stirring blades with variable frequency speed control are used to stir the red mud slurry to control the red mud particles to be dispersed to the greatest extent in the slurry.

[0016] During the mixing process, add 20-30% water compared to the red mud slurry, and control the mixing time to 30-70 minutes to allow the red mud slurry to reach optimal rheological properties.

[0017] A further improvement is that S2 includes the following steps:

[0018] The red mud slurry is fed into a gravity separation device for separation, and the pressure during separation is controlled to be ≥0.1 MPa;

[0019] Coarse red mud settles to the bottom to form bottom flow concentrate, while fine red mud and water are discharged as overflow tailings. In this process, a mixture of solid and liquid is retained.

[0020] A further improvement is that S3 includes the following steps:

[0021] Dilute the overflow tailings obtained from the initial separation with water, controlling the proportion of water added to 20-40% of the overflow tailings;

[0022] The diluted material is fed into another gravity separation device, and the pressure during separation is controlled to be ≥0.12 MPa;

[0023] The process yields a secondary underflow concentrate and a final overflow tailings, while retaining a solid-liquid mixture.

[0024] A further improvement is made in S4, where, during the ultrasonic grading process, the frequency of the ultrasonic wave is controlled to be between 50-100kHz, the power of the ultrasonic wave is between 60-150W, and the duration of action is between 100-350s.

[0025] A further improvement is made in S5, where the collection of different parts includes the following steps:

[0026] Collect the mixed red mud slurry of particle size from the final overflow tailings in the inlet area of ​​the ultrasonic grading equipment;

[0027] Collect the red mud slurry after ultrasonic particle size classification of the final overflow tailings in the ultrasonic action area;

[0028] In the bottom settling zone, collect the coarsest red mud particles from the final overflow tailings;

[0029] In the top overflow zone, the finest red mud particles and the suspension formed by water in the final overflow tailings are collected.

[0030] Further improvements are made in S7, where signal processing and analysis include filtering, amplification, and digitization. Specifically, based on the characteristics of the signal and the type of noise, a bandpass filter and a bandstop filter are used to receive the reflected sound wave signal to obtain the filtered signal. A programmable gain amplifier is used to make the amplitude of the reflected sound wave signal reach a level suitable for digitization. An analog-to-digital converter is used to convert the amplified analog signal into a digital signal. The digitized signal is stored in a computer and timestamped.

[0031] A further improvement lies in the following step S7: An algorithm is used to extract signal feature parameters, specifically including the following steps:

[0032] Calculate the mean, variance, and peak value of the signal in the time domain.

[0033] The Fast Fourier Transform (FFT) algorithm is used to transform the signal from the time domain to the frequency domain to analyze the spectral characteristics of the signal.

[0034] Wavelet transform is used to perform multi-scale analysis of signals and extract their features at different scales.

[0035] The envelope of the signal is extracted using an envelope detection algorithm to analyze the amplitude changes of the signal.

[0036] Based on the extracted feature parameters, the reflected acoustic characteristics of the red mud particles can be inferred.

[0037] By comparing the correspondence between different characteristic parameters and known particle sizes, a mapping relationship between characteristic parameters and particle size is established.

[0038] A further improvement lies in the following steps in S7: Inferring the size and distribution of red mud particles based on the characteristic parameters of the acoustic signal specifically includes the following steps:

[0039] A machine learning method combining support vector machine (SVM) and neural network was used to establish a model between the characteristic parameters of acoustic signal and the size and distribution of red mud particles;

[0040] The model is trained using sample data with known particle size and distribution, enabling it to predict the particle size and distribution of unknown samples.

[0041] The extracted acoustic signal feature parameters are input into the trained model to obtain the predicted results of the size and distribution of red mud particles.

[0042] A further improvement is made in S8, based on the inferred overall range of size and distribution parameters of different grades of red mud slurry, using equal arithmetic values ​​as the standard, the numerical thresholds are divided into at most four sub-ranges. The size and distribution parameters of the N-grade red mud slurry are then incorporated into the numerical thresholds of the four sub-ranges and divided and combined again to obtain at most four grades of graded products.

[0043] The beneficial effects of this invention are as follows:

[0044] 1. This invention first uses gravity separation equipment to perform a first separation of red mud slurry, obtaining underflow concentrate and overflow tailings. Then, the overflow tailings are subjected to a second gravity separation to obtain a secondary underflow concentrate and a final overflow tailings. Gravity classification is more refined. Next, ultrasonic classification is used to further classify the final overflow tailings. The shape classification effect of ultrasound improves the classification precision of the final overflow tailings. Finally, the size and distribution of N-level red mud slurry are obtained by using the effect of sound wave emission, combined with signal feature parameter extraction and algorithm calculation. This is used as the final classification standard for merging and classification. In summary, multi-level gravity classification is used as the initial classification, ultrasonic classification is used to improve the precision of overflow tailings classification, and the principle of sound wave reflection is used to unify the classification to size distribution classification based on the above classification. The classification benchmark is more complete, the level adaptation is more reasonable, and the rationality of subsequent recycling and use is guaranteed.

[0045] 2. Based on the inferred overall range of size and distribution parameters of different grades of red mud slurry, this invention uses equal arithmetic values ​​as the standard to divide the data into at most four sub-ranges of numerical thresholds. The red mud slurry performance in each sub-range is similar. Based on the size and distribution parameters of the four sub-ranges, the N-grade red mud slurry in the aforementioned process is further divided and combined, resulting in at most four-grade divisions. This simplifies the division process. When referring to the data, one can either refer to the detailed N-grade division mentioned above or use the final four-grade division as the standard, providing multiple division options that are more scientific and reasonable. Attached Figure Description

[0046] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0047] To enhance understanding of the present invention, the present invention will be further described in detail below with reference to embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0048] Example 1

[0049] according to Figure 1 As shown in the figure, this embodiment proposes a method for classifying coarse and fine red mud, including the following steps:

[0050] S1: The red mud slurry is fed into a tank equipped with a stirring system, so that the red mud particles are dispersed in the slurry, and water is added simultaneously to control the rheological properties;

[0051] S2: The red mud slurry is fed into the gravity separation equipment for separation, so that the coarse red mud settles to the bottom to form the underflow concentrate, and the fine red mud and water are discharged as overflow tailings.

[0052] S3: The overflow tailings obtained from the first separation are diluted with water again and fed into another gravity separation device to separate and form secondary underflow concentrate. The remaining fine particles and water are discharged as the final overflow tailings.

[0053] S4: The final overflow tailings are fed into an ultrasonic classifier to further classify the liquid by utilizing the cavitation effect, micro-jet effect and acoustic flow effect of ultrasound in the liquid.

[0054] S5: Collect red mud slurry of different particle sizes after further classification at different parts of the ultrasonic classification equipment.

[0055] S6: Combining the above steps, collect the red mud slurry after N-level grading, emit sound waves of the same frequency into the N-level red mud slurry through a sound wave transmitter, and receive the reflected sound waves through a receiver.

[0056] S7: Process and analyze the received reflected acoustic wave signal, extract signal feature parameters, and infer the size and distribution of different grades of red mud slurry based on the signal feature parameters;

[0057] S8: Set a classification threshold and further divide and combine the N-grade red mud slurry according to its size and distribution to obtain a maximum of four-grade products.

[0058] Red mud with a composition of Fe2O3-55%, Al2O3-18%, and Na2O-2.5% is mixed with water in a tank and stirred to maximize particle dispersion. First, it passes through a separation device with four groups activated at a pressure of 0.1 MPa, recovering 30% of the high-grade iron concentrate. The remaining 70% of the tailings is then scavenged through a separation device with two groups activated at a pressure of 0.15 MPa, recovering 20% ​​of the low-grade iron concentrate. The remaining 50% of the tailings achieves a red mud yield of 50% and an iron recovery rate of over 55%. Multi-stage gravity classification is used as the initial classification, ultrasonic classification further refines the classification of the overflow tailings, and the principle of sound wave reflection is used to unify the classification based on size distribution. This results in a more comprehensive classification standard, more reasonable grade matching, and ensures the rationality of subsequent recycling and use.

[0059] Example 2

[0060] according to Figure 1 As shown in the figure, this embodiment proposes a method for classifying coarse and fine red mud, including the following steps:

[0061] The red mud slurry is fed into a tank equipped with a stirring system to disperse the red mud particles in the slurry, while water is added simultaneously to control rheological properties. The specific steps include: adding the red mud slurry to the tank with a stirring system; using variable frequency speed-regulating stirring blades to stir the red mud slurry and control the maximum dispersion of red mud particles in the slurry; during the stirring process, adding 20-30% clean water compared to the red mud slurry, and controlling the stirring time for 30-70 minutes to achieve optimal rheological properties in the red mud slurry; through a unique stirring blade design and variable frequency speed regulation technology, efficient dispersion of red mud particles in the slurry is achieved, laying a solid foundation for subsequent classification.

[0062] The red mud slurry is fed into a gravity separation device for separation, causing the coarse red mud to settle to the bottom to form the underflow concentrate, while the fine red mud and water are discharged as overflow tailings. The specific steps include: feeding the red mud slurry into the gravity separation device for separation, controlling the pressure during separation to be ≥0.1 MPa; the coarse red mud settles to the bottom to form the underflow concentrate, while the fine red mud and water are discharged as overflow tailings. During this process, the solid-liquid mixture is retained.

[0063] The overflow tailings obtained from the initial separation are diluted with water again and then fed into another gravity separation device to form a secondary underflow concentrate. The remaining fine particles and water are discharged as the final overflow tailings. Specifically, the process includes the following steps: diluting the overflow tailings obtained from the initial separation with water at a ratio controlled to 20-40% of the overflow tailings; feeding the diluted material into another gravity separation device, controlling the pressure during separation to ≥0.12 MPa; obtaining the secondary underflow concentrate and the final overflow tailings, while retaining the solid-liquid mixture during this process; through the further dilution and selection of the tailings, valuable components in the fine particles are effectively screened out, reducing resource waste.

[0064] By combining gravity separation and ultrasonic classification technologies, efficient and precise classification of red mud particles is achieved. This improves classification efficiency and accuracy while reducing energy consumption. The entire classification process requires no chemical reagents, making it environmentally friendly and in line with the concept of green and sustainable development.

[0065] The final overflow tailings are fed into an ultrasonic classifier, which utilizes the cavitation effect, micro-jet effect and acoustic flow effect of ultrasound in liquids for further classification. During the ultrasonic classification process, the frequency of the ultrasound is controlled between 50-100kHz, the power of the ultrasound is between 60-150W, and the action time is between 100-350s.

[0066] In different parts of the ultrasonic classifier, red mud slurry of different particle sizes after further classification is collected. The collection at different parts includes the following steps: collecting the mixed red mud slurry of particle size from the final overflow tailings in the inlet area of ​​the ultrasonic classifier; collecting the red mud slurry of the final overflow tailings after ultrasonic particle size classification in the ultrasonic action area; collecting the coarsest red mud particle slurry from the final overflow tailings in the bottom settling zone; and collecting the finest red mud particle slurry and the suspension formed by water in the top overflow zone. In ultrasonic classification, sound waves interact with red mud particles during propagation. This interaction causes the particles to be subjected to pressure and vibration from the sound waves, thus affecting their motion. By adjusting parameters such as the frequency, intensity, and duration of the sound waves, effective separation of red mud particles of different sizes, shapes, and densities can be achieved. An ultrasonic generator produces sound waves of a certain frequency and intensity, which are then transmitted to the red mud suspension. As the sound waves propagate in the suspension, they interact with the red mud particles through scattering, reflection, and transmission. These interactions cause particles to be subjected to pressure and vibration from the sound waves, resulting in changes in displacement and velocity. Since particles of different sizes, shapes, and densities respond differently to sound waves, effective particle separation can be achieved by adjusting the parameters of the sound waves. Generally, larger particles experience greater sound wave pressure and are therefore easier to separate. Ultrasonic classification technology offers higher efficiency and precision. It can achieve effective particle separation in a shorter time and allows for more precise control over particle size, shape, and density. Furthermore, ultrasonic classification technology is highly adaptable and can meet the classification needs of different types and properties of red mud particles.

[0067] In summary, after collecting N-graded red mud slurry, a sound wave of the same frequency is emitted into the N-grade red mud slurry via a sound wave transmitter, and the reflected sound wave is received by a receiver. During propagation, the sound wave undergoes reflection and transmission when it encounters interfaces between different media. The size, shape, and density of the red mud particles affect the propagation path and reflection characteristics of the sound wave. Therefore, by emitting sound waves into the red mud slurry and receiving the reflected sound wave signals, relevant information about the red mud particles can be analyzed, thereby achieving grading. To achieve sound wave reflection grading, appropriate equipment is required. This equipment includes a sound wave transmitter, a receiver, a signal processing unit, and a control system. The sound wave transmitter emits sound waves into the red mud slurry, the receiver receives the reflected sound wave signals, the signal processing unit processes and analyzes the received signals, and the control system controls the entire grading process. The sound wave reflection grading method does not require direct contact with the red mud particles, avoiding the wear and contamination problems that may exist in traditional grading methods. By accurately measuring parameters such as the propagation time and amplitude of the sound wave signal, high-precision grading of red mud particles can be achieved. The acoustic reflection grading method can monitor the particle distribution in red mud slurry in real time, providing an important basis for quality control in the production process.

[0068] The received reflected acoustic wave signals are processed and analyzed to extract signal feature parameters, and the size and distribution of different grades of red mud slurry are inferred based on the signal feature parameters;

[0069] Signal processing and analysis includes filtering, amplification, and digitization. Specifically, based on the characteristics of the signal and the type of noise, bandpass and bandstop filters are used to receive the reflected sound wave signal to obtain the filtered signal. A programmable gain amplifier is used to make the amplitude of the reflected sound wave signal reach a level suitable for digitization. An analog-to-digital converter is used to convert the amplified analog signal into a digital signal. The digitized signal is stored in a computer and timestamped.

[0070] The algorithm extracts signal feature parameters, specifically including the following steps: calculating the mean, variance, and peak value of the signal in the time domain; using the Fast Fourier Transform (FFT) algorithm to transform the signal from the time domain to the frequency domain and analyze the spectral characteristics of the signal; using wavelet transform to perform multi-scale analysis of the signal and extract the signal features at different scales; extracting the envelope of the signal using the envelope detection algorithm to analyze the amplitude changes of the signal; inferring the reflected acoustic characteristics of red mud particles based on the extracted feature parameters; and establishing a mapping relationship between feature parameters and particle size by comparing the correspondence between different feature parameters and known particle sizes.

[0071] The method for inferring the size and distribution of red mud particles based on acoustic signal feature parameters includes the following steps: First, a machine learning method combining Support Vector Machine (SVM) and neural networks is used to establish a model relating acoustic signal feature parameters to the size and distribution of red mud particles. Second, the model is trained using sample data with known particle sizes and distributions to predict the particle size and distribution of unknown samples. Third, the extracted acoustic signal feature parameters are input into the trained model to obtain the predicted size and distribution of red mud particles. Fourth, the influence of factors such as the shape, density, and concentration of red mud particles on acoustic reflection is considered when building the model. Fifth, the prediction results are verified and calibrated when applying the model to ensure its accuracy and reliability.

[0072] A classification threshold is set, and the N-grade red mud slurry is further divided and combined according to its size and distribution to obtain a maximum of four-grade products. Specifically, based on the inferred overall range of size and distribution parameters of different grades of red mud slurry, the numerical thresholds for dividing it into at most four sub-ranges are used as the standard. The size and distribution parameters of the N-grade red mud slurry are then substituted into the numerical thresholds of the four sub-ranges and divided and combined again to obtain a maximum of four-grade products.

[0073] This method for classifying coarse and fine red mud first uses gravity separation equipment to separate the red mud slurry into underflow concentrate and overflow tailings. Then, the overflow tailings undergo a second gravity separation to obtain secondary underflow concentrate and final overflow tailings, further refining the gravity classification. Next, ultrasonic classification further classifies the final overflow tailings, using ultrasound to improve the classification precision based on shape. Finally, the size and distribution of N-level red mud slurry are obtained through sound wave emission, signal feature parameter extraction, and algorithm calculation, and used as the final classification standard for merging and grading. In summary, multi-stage gravity classification serves as the initial classification, ultrasonic classification enhances the precision of overflow tailings classification, and the principle of sound wave reflection unifies the classification based on size distribution. This results in a more complete classification benchmark, more reasonable level matching, and ensures the rationality of subsequent recycling and use. Furthermore, based on the inferred overall range of size and distribution parameters of different grades of red mud slurry, this invention uses equal arithmetic values ​​as a standard to divide the data into at most four sub-ranges of numerical thresholds. The red mud slurry performance in each sub-range is similar. Based on the size and distribution parameters of the four sub-ranges, the N-grade red mud slurry in the aforementioned process is further divided and combined, resulting in at most four-grade divisions. This simplifies the division process. When referring to the data, one can either refer to the aforementioned detailed N-grade division or use the final four-grade division as the standard, providing multiple division options that are more scientific and reasonable.

[0074] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method of classifying coarse and fine red mud, characterised by, The method comprises the following steps: S1: input the red mud slurry into a tank with a stirring system, so that the red mud particles are dispersed in the slurry, and water is added to control the rheological property; S2: input the red mud slurry into a gravity separation device for separation, so that the coarse red mud is settled at the bottom to form a bottom flow concentrate, and the fine red mud and water are discharged as an overflow tailing; S3: add water to the overflow tailing obtained in the first separation again to dilute the overflow tailing, input the overflow tailing into another gravity separation device again, separate the overflow tailing to form a secondary bottom flow concentrate, and discharge the remaining fine particles and water as a final overflow tailing; S4: input the final overflow tailing into an ultrasonic classification device, and further classify the final overflow tailing by using the cavitation effect, microjet effect and acoustic streaming effect of ultrasonic waves in the liquid; S5: collect the red mud particle slurries of different particle sizes after further classification at different positions in the ultrasonic classification device The method comprises the following steps: collect the red mud slurry of mixed particle sizes in the final overflow tailing at the inlet area of the ultrasonic classification device, collect the red mud slurry after ultrasonic particle size grading of the final overflow tailing at the ultrasonic action area, collect the red mud particle slurry of the coarsest particle size in the final overflow tailing at the bottom settling area, and collect the red mud particle slurry of the finest particle size in the final overflow tailing and the suspension formed by water at the top overflow area; S6: collect the red mud slurry after N-stage classification according to the above steps, emit the same frequency of acoustic waves to the N-stage red mud slurry through an acoustic wave emitter, and receive the reflected acoustic waves through a receiver; S7: process and analyze the received reflected acoustic wave signals, extract signal characteristic parameters, and infer the size and distribution of the red mud slurries of different stages according to the signal characteristic parameters; S8: set a division threshold, divide and combine the N-stage red mud slurries according to the size and distribution again, and obtain at most four-stage classification products, specifically: according to the overall range of the size and distribution parameters of the red mud slurries of different stages inferred, take the equal difference value as the standard, divide at most four numerical threshold values of the sub-ranges, input the size and distribution parameters of the N-stage red mud slurries into the numerical threshold values of the four sub-ranges, divide and combine again, and obtain at most four-stage classification products.

2. A method of classifying thick and thin red mud according to claim 1, characterised in that: The S1 comprises the following steps: The red mud slurry is added into a tank with a stirring system, and the red mud slurry is stirred by using stirring blades with variable frequency speed regulation, so that the red mud particles are dispersed in the slurry to the maximum extent; During the stirring process, 20-30% of clean water is added compared with the red mud slurry, and the stirring time is controlled to be 30-70 min, so that the red mud slurry reaches the best rheological property.

3. A method of classifying thick and thin red mud according to claim 1, characterized by: The S2 comprises the following steps: The red mud slurry is input into a gravity separation device for separation, and the pressure during the separation is controlled to be greater than or equal to 0.1 Mpa; The coarse red mud is settled at the bottom to form a bottom flow concentrate, and the fine red mud and water are discharged as an overflow tailing, and the mixed slurry of solid and liquid is retained during the process.

4. A method of classifying thick and thin red mud according to claim 1, characterized by: The S3 comprises the following steps: Water is added to the overflow tailing obtained in the first separation to dilute the overflow tailing, and the proportion of the added water is controlled to be 20-40% of the overflow tailing; The diluted material is input into another gravity separation device, and the pressure during the separation is controlled to be greater than or equal to 0.12 Mpa; A secondary bottom flow concentrate and a final overflow tailing are obtained, and the mixed slurry of solid and liquid is retained during the process.

5. A method of classifying thick and thin red mud according to claim 1, characterized by: In the S4, during the ultrasonic classification process, the frequency of the ultrasonic waves is controlled to be between 50-100 kHz, the power of the ultrasonic waves is controlled to be between 60-150 W, and the action time is controlled to be between 100-350 s.

6. A method of classifying thick and thin red mud according to claim 1, characterized by: In the S7, the processing and analysis of the signal includes filtering, amplification, and digitization, specifically: according to the characteristics of the signal and the type of noise, a band-pass filter and a band-stop filter are used to receive the reflected acoustic wave signal to obtain the filtered signal, a programmable gain amplifier is used to make the amplitude of the reflected acoustic wave signal reach a level suitable for digitization, an analog-to-digital converter is used to convert the amplified analog signal into a digital signal, the digitized signal is stored in a computer and is time-stamped.

7. A method of classifying thick and thin red mud according to claim 6, characterised in that: In the S7, the signal characteristic parameters are extracted using an algorithm, specifically including the following steps: Calculate the mean, variance, and peak time-domain characteristic parameters of the signal; Use the fast Fourier transform (FFT) algorithm to convert the signal from the time domain to the frequency domain and analyze the spectral characteristics of the signal; Use wavelet transform to perform multi-scale analysis on the signal and extract the characteristics of the signal at different scales; Extract the envelope line of the signal using an envelope detection algorithm to analyze the amplitude changes of the signal; According to the extracted characteristic parameters, the reflected acoustic wave characteristics of the red mud particles are inferred; By comparing different characteristic parameters with the corresponding relationship of known particle sizes, a mapping relationship between the characteristic parameters and the particle sizes is established.

8. A method of classifying thick and thin red mud according to claim 7, characterised in that: In the S7, the size and distribution of the red mud particles are inferred according to the acoustic wave signal characteristic parameters, specifically including the following steps: Use a machine learning method combining support vector machines (SVM) and neural networks to establish a model between the acoustic wave signal characteristic parameters and the size and distribution of the red mud particles; Use sample data of known particle sizes and distributions to train the model to make the model predict the particle sizes and distributions of unknown samples; Input the extracted acoustic wave signal characteristic parameters into the trained model to obtain the predicted results of the size and distribution of the red mud particles.

Images