A multi-source perception-based full tailings filling proportioning optimization method

Abstract

The present application relates to the field of full tailings filling proportioning optimization, and particularly relates to a full tailings filling proportioning optimization method based on multi-source perception. First, based on the collected multi-source original industrial data, an original data set is established; based on the original data set, a dynamic synchronization mechanism based on flow state change trend is introduced, a pre-synchronization flow state change factor is constructed, the width of the synchronization time window is dynamically adjusted, the precise time synchronization of the original data set is carried out, the synchronized multi-source industrial data and the pre-synchronization flow state change factor after synchronization are obtained, and a particle blockage potential field analysis model is constructed to obtain the particle blockage potential field strength; based on the tailings moisture content data in the synchronized multi-source industrial data and the particle blockage potential field strength, the real-time slurry dynamic viscosity is obtained, and compared with the set target viscosity interval to generate corresponding adjustment instructions. The problem that the proportioning parameters are difficult to be adjusted in real time and accurately due to the tailings particle size fluctuation, moisture content change and slurry conveying state change is solved.

Description

Technical Field

[0001] This invention relates to the field of tailings backfill mix optimization, and more particularly to a method for optimizing tailings backfill mix based on multi-source sensing. Background Technology

[0002] Tailings backfilling technology is an important technique in the field of mine backfilling. It involves mixing tailings generated during mineral processing with cementing materials and water in a specific ratio to form a backfill slurry, which is then transported to the underground goaf to complete the backfilling operation. This achieves the resource utilization of tailings, stable support of the goaf, and control of surface subsidence. With the continuous expansion of deep mining scale, tailings backfilling systems are gradually developing towards long-distance, high-concentration, and continuous transportation, thus placing higher demands on the stability of slurry transportation and the precision of backfill ratio control. Existing tailings backfilling methods typically employ a fixed ratio control approach, controlling the tailings concentration, water-cement ratio, and cementing material addition based on preset laboratory parameters. However, because tailings generated in different mining areas, at different times, and under different mineral processing conditions exhibit significant fluctuations in particle size distribution, moisture content, and particle gradation, the particle contact state within the slurry continuously changes during actual transportation. When the proportion of ultrafine particles in tailings increases, free water inside the slurry is adsorbed by the particle surface, leading to increased slurry viscosity. Conversely, when the proportion of coarse particles increases, the settling velocity of particles inside the slurry increases significantly, easily leading to localized particle accumulation. Traditional fixed-ratio methods cannot adjust in real time according to changes in tailings conditions, thus easily causing a decrease in slurry transport stability. Furthermore, most existing methods only monitor the slurry transport status using pressure sensors or concentration meters and determine the risk of blockage based on pressure thresholds. However, a significant increase in transport pressure usually occurs after localized blockages have formed, making early identification of blockage risks difficult. Simultaneously, the sampling frequencies of different industrial sensing devices vary significantly. For example, pressure data is typically output in milliseconds, while industrial image data is updated in seconds. If a fixed time window is directly used for data synchronization, localized abnormal states are easily averaged over time, reducing the accuracy of abnormal state identification. Summary of the Invention

[0003] This invention provides a method for optimizing the proportion of tailings backfill based on multi-source sensing, in order to solve the problem that the proportion parameters are difficult to adjust in real time and accurately due to the fluctuation of tailings particle size, changes in moisture content and changes in slurry transport status in existing methods.

[0004] The present invention provides a method for optimizing the proportion of tailings backfill based on multi-source sensing, comprising the following steps:

[0005] S1. Based on the collected multi-source raw industrial data, establish a raw dataset; based on the raw dataset, introduce a dynamic synchronization mechanism based on the flow regime change trend, construct a pre-synchronization flow regime change factor, dynamically adjust the synchronization time window width, perform accurate time synchronization of the raw dataset, and obtain synchronized multi-source industrial data and synchronized pre-synchronization flow regime change factor.

[0006] S2. Based on the synchronized multi-source industrial data and the synchronized pre-synchronized flow regime change factor, a particle blockage potential field analysis model is constructed by introducing particle mass ratio, local particle aggregation density, and normalized local slurry flow capacity to obtain the particle blockage potential field intensity. Based on the tailings moisture content data and particle blockage potential field intensity in the synchronized multi-source industrial data, the basic slurry viscosity is introduced to obtain the real-time dynamic viscosity of the slurry. Based on the real-time dynamic viscosity of the slurry, it is compared with the set target viscosity range to generate corresponding adjustment commands.

[0007] Preferably, S1 specifically includes:

[0008] In the dynamic synchronization mechanism based on the flow regime change trend, the slurry transport flow rate data and transport pressure data in the original dataset are normalized to obtain normalized slurry volumetric flow rate and normalized transport pressure. Then, the normalized flow rate change rate and normalized pressure change rate are calculated. The square of the normalized flow rate change rate and the square of the normalized pressure change rate are added together and the square root is taken to calculate the degree of slurry transport fluctuation.

[0009] Preferably, S1 specifically includes:

[0010] Based on the tailings particle size distribution data in the original dataset, the particle size corresponding to a cumulative particle size distribution of 10%, the median particle size, and the particle size corresponding to a cumulative particle size distribution of 90% are calculated respectively. The difference between the particle size corresponding to a cumulative particle size distribution of 90% and the particle size corresponding to a cumulative particle size distribution of 10% is divided by the median particle size to calculate the particle size dispersion correction term. Based on the particle size dispersion correction term, the particle structure dispersion is calculated using a logarithmic function. The pre-synchronous flow state change factor is obtained by adding the slurry transport fluctuation degree to the particle structure dispersion degree.

[0011] Preferably, S1 specifically includes:

[0012] Based on the pre-synchronization flow regime change factor, the set maximum and minimum synchronization windows, a window sensitivity adjustment coefficient is introduced to calculate the dynamic compression of the synchronization window; the dynamic compression of the synchronization window is added to the minimum synchronization window to calculate the synchronization time window width; after synchronization based on the synchronization time window width, the synchronized multi-source industrial data and the synchronized pre-synchronization flow regime change factor are obtained.

[0013] Preferably, S2 specifically includes:

[0014] In the particle blockage potential field analysis model, the synchronized multi-source industrial data includes synchronized tailings particle size distribution data, conveying pressure data, pipeline vibration data, slurry surface texture image data, slurry conveying flow rate data, and tailings moisture content data. Based on the synchronized tailings particle size distribution data, the particles are divided according to the set particle size range, and the particle mass of each particle size range is counted. Based on the particle mass of each particle size range, the mass ratio of each type of particle is calculated.

[0015] Preferably, S2 specifically includes:

[0016] The difference between the synchronized transport pressure data and the average pressure within the synchronization time window is used as the pressure state characteristic data; a fast Fourier transform is performed on the synchronized pipeline vibration data to obtain the low-frequency vibration energy; based on the synchronized slurry surface texture image data, the slurry texture entropy value is calculated and subtracted from the slurry texture entropy value at the previous moment to obtain the texture entropy change; based on the obtained standard slurry density, combined with the pressure state characteristic data, low-frequency vibration energy, and texture entropy change, the local particle aggregation density is calculated.

[0017] Preferably, S2 specifically includes:

[0018] Based on the synchronized slurry delivery flow rate data and the obtained pipe cross-sectional area, the ratio of the synchronized slurry delivery flow rate data to the pipe cross-sectional area is used as the average flow velocity. This average velocity is then compared with the obtained standard stable delivery flow velocity to obtain the normalized local flow capacity of the slurry.

[0019] Preferably, S2 specifically includes:

[0020] Based on the particle mass ratio, a particle clogging contribution coefficient is introduced. The product of the particle mass ratio and the particle clogging contribution coefficient is summed, and the calculation result is multiplied by the pre-synchronized flow regime change factor to calculate the particle structure clogging contribution intensity under the current flow regime fluctuation conditions. The nonlinear enhancement capability of local particle aggregation state to clogging risk is calculated by the ratio of local particle aggregation density to the obtained critical packing density. Based on the normalized local flow capacity of the slurry, the attenuation effect of local flow capacity on clogging risk is calculated by an exponential function.

[0021] Preferably, S2 specifically includes:

[0022] The particle blockage potential field strength is obtained by multiplying the particle structure blockage contribution intensity under the current flow fluctuation conditions, the nonlinear enhancement of blockage risk by the local particle aggregation state, and the attenuation effect of the local flow capacity on blockage risk.

[0023] Preferably, S2 specifically includes:

[0024] When the real-time dynamic viscosity of the slurry is higher than the upper limit of the target viscosity range, it is determined that the slurry flow resistance has increased. The tailings mass concentration is reduced, the water flow rate is increased, and the operating frequency of the delivery pump is increased simultaneously according to the set adjustment ratio. When the real-time dynamic viscosity of the slurry is lower than the lower limit of the target viscosity range, it is determined that the slurry is too fluid. The water flow rate is reduced, the tailings concentration is increased, and the delivery flow rate is reduced. When the real-time dynamic viscosity of the slurry is within the target viscosity range, the current tailings concentration, water-cement ratio, and delivery flow rate are maintained unchanged.

[0025] The beneficial effects of the technical solution of the present invention are:

[0026] 1. Through a dynamic synchronization mechanism based on the trend of flow regime changes, the synchronization time window width can be adjusted in real time according to the degree of change in the slurry flow state, avoiding the problem of averaging of abnormal states caused by fixed time windows, and improving the synchronization accuracy and data utilization of multi-source raw industrial data at the same time scale.

[0027] 2. By constructing a particle blockage potential field analysis model, the particle mass ratio, local particle aggregation density, and normalized local slurry flow capacity are coupled and analyzed. This allows for the early identification of particle contact network collapse trends before the conveying pressure increases significantly, thereby improving the real-time performance and accuracy of pipe blockage risk prediction.

[0028] 3. The dynamic viscosity of the real-time slurry is corrected by using the particle blocking potential field strength, so that the changes in the particle structure, the free water attenuation effect and the changes in local flow resistance can be uniformly mapped to the dynamic viscosity of the real-time slurry. Then, the tailings concentration, water-cement ratio and conveying velocity are dynamically adjusted according to the dynamic viscosity of the real-time slurry and the target viscosity range, thereby achieving dynamic and stable control in the entire tailings backfilling process. Attached Figure Description

[0029] Figure 1 This is a flowchart of a method for optimizing the proportion of tailings backfill based on multi-source sensing, as described in this invention. Detailed Implementation

[0030] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0032] The following description, in conjunction with the accompanying drawings, details a specific scheme for the tailings backfill ratio optimization method provided by this invention.

[0033] See attached document Figure 1 The diagram illustrates a flowchart of a method for optimizing the proportion of tailings backfill based on multi-source sensing, according to an embodiment of the present invention. The method includes the following steps:

[0034] S1. Based on the collected multi-source raw industrial data, establish the raw dataset; based on the raw dataset, introduce a dynamic synchronization mechanism based on the flow regime change trend, construct a pre-synchronization flow regime change factor, dynamically adjust the synchronization time window width, perform accurate time synchronization of the raw dataset, and obtain synchronized multi-source industrial data and synchronized pre-synchronization flow regime change factor.

[0035] By deploying multi-source industrial sensing devices, such as online laser particle size analyzers, microwave moisture meters, electromagnetic flow meters, pressure sensors, vibration sensors, and industrial cameras, in the entire tailings backfilling system, real-time data collection of the operating status during the backfilling process is achieved, obtaining multi-source raw industrial data for subsequent mix optimization analysis. This multi-source raw industrial data includes tailings particle size distribution data, tailings moisture content data, slurry transport flow rate data, transport pressure data, pipeline vibration data, and slurry surface texture image data. Specifically, tailings particle size distribution data is continuously collected using an online laser particle size analyzer installed between the tailings buffer silo outlet and the mixing tank inlet. This analyzer utilizes the principle of laser scattering to detect tailings particles in real time, outputting a particle size distribution curve and generating data on median particle size, proportion of coarse particles, and proportion of ultrafine particles. Tailings moisture content data is collected using a microwave moisture meter installed at the mixing tank inlet. This analyzer detects the attenuation of microwave signals by the tailings slurry to determine the free water content. Slurry flow rate data is collected in real time using an electromagnetic flow meter installed in the main slurry transport pipeline. The electromagnetic flow meter obtains the slurry volumetric flow rate data based on the induced potential change generated by the slurry cutting magnetic field lines. Transport pressure data is collected using a pressure sensor at the pump outlet. Pipeline vibration data is collected using a vibration sensor installed on the outer wall of the main transport pipeline. Slurry surface texture image data is continuously collected using industrial cameras positioned at the mixing tank observation window and the tailings return trough.

[0036] The raw industrial data acquired by multi-source industrial sensing devices, along with device numbers, acquisition timestamps, sampling spatial locations, data source identifiers, and sampling period information, forms the raw dataset. Due to significant differences in sampling frequencies among different industrial sensing devices, it is impossible to directly perform unified analysis on all data. For example, pressure sensors output data continuously at millisecond intervals, while slurry surface texture images acquired by industrial cameras are typically updated at intervals of several seconds. Directly synchronizing all data using a fixed time window would result in the averaging of rapidly changing data, making it impossible to promptly identify localized blockage trends. Therefore, a dynamic synchronization mechanism based on flow pattern changes is introduced. First, a dynamic time synchronization buffer is established, and the synchronization time window width is adjusted in real time using the dynamic changes in slurry flow. The specific implementation process is as follows:

[0037] First, read the continuous time period. The slurry transport flow rate and pressure data were normalized using the maximum-minimum normalization method to obtain the normalized slurry volumetric flow rate. With normalized delivery pressure Further calculate the normalized rate of change of flow. and normalized rate of change of pressure If the normalized flow rate continues to increase, it indicates that the flow state inside the slurry is beginning to fluctuate unstably. The continuous time period... The specific application scenario determines the time window. For example, during the normal and stable filling stage, a continuous time window of 30 to 60 seconds can be used. However, when a rapid increase in pressure, increased vibration of the bend, or a sudden change in tailings particle size is detected, the time window can be shortened to 5 to 10 seconds.

[0038] Because the normalized rate of change of flow rate and normalized rate of change of pressure of tailings slurry usually remain at a low level under stable transport conditions, but when the particle contact structure inside the slurry changes, the flow rate and pressure will fluctuate rapidly at the same time, a pre-synchronous flow state change factor is constructed. The pre-synchronous flow state change factor is derived step by step based on the slurry flow stability mechanism.

[0039] Since pressure fluctuations reflect changes in the internal resistance of the slurry, the normalized flow rate change rate and the normalized pressure rate change rate are used together as the basic quantities for evaluating flow dynamics. However, using only flow rate and pressure changes cannot accurately identify changes in tailings structure, because differences in the particle contact network formed by different particle sizes directly affect transport stability. Therefore, a particle size dispersion correction term is introduced to characterize the degree of particle size dispersion. The greater the difference between coarse and fine particles, the easier it is for unstable contact structures to form between particles; therefore, it is necessary to increase the weight of this type of data in the synchronization process.

[0040] Because the particle size dispersion may exhibit extreme fluctuations, a logarithmic function is further employed for smoothing to prevent local outliers from causing severe oscillations in the synchronization time window, ultimately forming a pre-synchronization flow regime change factor. The specific calculation formula is as follows:

[0041]

[0042] in, It is the pre-synchronization flow regime change factor, used to represent the intensity of flow regime change; Indicates the current time Normalized slurry volumetric flow rate; This represents the normalized rate of change of flow rate; Indicates the current time Normalized delivery pressure; This represents the normalized rate of change of pressure; This indicates the particle size at which the cumulative particle size distribution reaches 10%. Indicates the median diameter of the particle; This indicates the particle size at which the cumulative particle size distribution reaches 90%. , , All were calculated using tailings particle size distribution data; This represents the particle size dispersion correction term, used to characterize the degree of dispersion in particle gradation; Used to describe the degree of fluctuation in slurry transport; Used to describe the degree of dispersion of particle structure.

[0043] Furthermore, a timestamp-based sliding window synchronization technique is employed to dynamically calculate the synchronization time window width based on the pre-synchronized flow regime change factor, facilitating the synchronization processing of the original dataset. Specifically, a maximum synchronization window is pre-set based on the average sampling interval of various sensors and the slurry state change cycle obtained from the entire tailings database under historical stable transport conditions. (e.g. 5) with the minimum synchronization window (e.g., 0.5), then the real-time synchronization window is calculated according to the following relationship, the specific expression being:

[0044]

[0045] in, This indicates the width of the synchronization time window at the current moment; This indicates the maximum allowed synchronization window under stable slurry conditions; This represents the minimum allowed synchronization window under abnormal fluctuation conditions; This represents the current pre-synchronization flow regime change factor; The window sensitivity adjustment coefficient is determined using the five-fold cross-validation method, and the reference range is [value range missing]. ; This indicates the dynamic compression amount of the synchronization window.

[0046] When the pre-synchronization flow regime change factor When the value is relatively small, it indicates stable slurry flow. The synchronization window is automatically increased, allowing data from different sampling frequencies to be fused over a longer period, thus improving overall data utilization. When the pre-synchronization flow regime change factor... As the denominator continues to increase, the denominator increases synchronously, resulting in a wider synchronization time window. The synchronization window is rapidly reduced, allowing only data with smaller time intervals to enter the same synchronization window. This ensures that pressure surges, vibration anomalies, and particle size fluctuations are precisely aligned within the same time scale, preventing the abnormal states from being averaged over time due to an excessively large synchronization window. After synchronization using a dynamic synchronization mechanism based on flow regime change trends, synchronized multi-source industrial data and synchronized pre-synchronized flow regime change factors are obtained. .

[0047] S2. Based on the synchronized multi-source industrial data and the synchronized pre-synchronized flow regime change factor, a particle blockage potential field analysis model is constructed by introducing particle mass ratio, local particle aggregation density, and normalized local slurry flow capacity to obtain the particle blockage potential field intensity. Based on the tailings moisture content data and particle blockage potential field intensity in the synchronized multi-source industrial data, the basic slurry viscosity is introduced to obtain the real-time dynamic viscosity of the slurry. Based on the real-time dynamic viscosity of the slurry, it is compared with the set target viscosity range to generate corresponding adjustment commands.

[0048] After data synchronization, a particle blockage potential field analysis model was constructed. Traditional pipe blockage prediction models typically rely on pressure thresholds, but pressure increases often occur in the later stages of blockage. Therefore, a new particle blockage potential field analysis model was established to identify the tendency of particle contact network collapse in advance. Specifically, this includes:

[0049] Based on synchronized multi-source industrial data, state features are extracted to form the input of the particle blockage potential field analysis model. The specific implementation process is as follows:

[0050] First, based on the synchronized tailings particle size distribution data, the particles were divided into preset particle size ranges, such as 0–20 micrometers, 20–50 micrometers, 50–100 micrometers, and above 100 micrometers. Then, the particle mass of each size range was statistically analyzed. , then calculate ,in, Indicates the first The percentage of particle size by mass is used to differentiate the contribution of particles of different sizes to blockage formation. It is the total number of particle categories. Indicates the first Particle quality, here .

[0051] Furthermore, based on synchronized transport pressure data, pipeline vibration data, slurry surface texture image data, and slurry transport flow rate data, local particle aggregation density is extracted. First, the difference between the synchronized transport pressure data and the average pressure within the synchronization time window is calculated as pressure state feature data. The synchronized pipeline vibration data were subjected to a fast Fourier transform to obtain the low-frequency vibration energy. High-frequency vibration energy. Since slurry mainly exhibits high-frequency micro-vibrations during normal transport, but significant low-frequency vibrations occur during localized accumulation, low-frequency vibration energy can serve as an important characteristic of impending blockage. The gray-level co-occurrence matrix is ​​used to calculate the slurry texture entropy value from the synchronized slurry surface texture image data, and this value is subtracted from the previous time step to obtain the change in texture entropy. Furthermore, the local particle aggregation density is calculated. The specific calculation formula is as follows:

[0052]

[0053] in, Indicates the local particle aggregation density; This is the standard slurry density, used to represent the reference density of the tailings slurry under normal and uniform conveying conditions. It is measured in real time using a densitometer. kg / m³; , , These represent the pressure response weighting coefficient, vibration response weighting coefficient, and texture entropy response weighting coefficient, respectively, with units of pressure state feature data. Low-frequency vibration energy Texture entropy change The reciprocal of is determined through an attention mechanism such as Transformer, with 8 attention heads and 64 dimensions. The reference value ranges are as follows: , , .

[0054] Based on the synchronized slurry transport flow rate data, the pipe cross-sectional area was obtained by consulting existing pipe design manuals. The ratio of the synchronized slurry transport flow rate data to the cross-sectional area was used as the average flow velocity. This average velocity was then compared with the standard steady-state transport velocity obtained from the whole tailings database to obtain the normalized local flow capacity of the slurry. .

[0055] Furthermore, based on synchronized multi-source industrial data, a pre-synchronized flow regime change factor is adopted. As a dynamic correction input term for the particle blocking potential field analysis model, because the more intense the flow fluctuations, the more unstable the particle contact structure, thus increasing the sensitivity of the blocking potential field to flow changes, ultimately forming the particle blocking potential field function, the specific expression of which is:

[0056]

[0057] in, This represents the strength of the particle blocking potential field; This represents the pre-synchronization flow state change factor after synchronization, i.e., the dynamic correction input of the particle blocking potential field analysis model; Indicates the first Percentage of particle-like mass; This represents the total number of particle categories; The critical bulk density is obtained directly from the whole tailings database; It is the first The clogging contribution coefficient of particles is determined according to the specific category, such as 1.7 for ultrafine particles, 1.3 for fine particles, 0.8 for medium-sized particles, and 1.5 for coarse particles. This is the particle aggregation sensitivity coefficient, used to describe the nonlinear amplification effect of particle aggregation degree on clogging risk. It is determined using the five-fold cross-validation method, and the reference value range is [value missing]. ; This is the velocity impact factor, used to describe the ability of flow velocity to suppress clogging risk. It is determined using the five-fold cross-validation method, and the reference value range is [value missing]. ; This indicates the intensity of particle blockage contribution under the current flow fluctuation conditions; This indicates the nonlinear enhancement of clogging risk by local particle aggregation. This indicates the attenuation effect of local flow capacity on the risk of blockage.

[0058] Through particle blocking potential field strength Real-time assessment of the stability of the particle contact network and the strength of the particle blockage potential field within the slurry. The larger the value, the higher the risk of localized particle accumulation.

[0059] Subsequently, the particle blocking potential field strength Continuing with the input of the dynamic rheological correction model, the dynamic rheological correction function is constructed as follows:

[0060]

[0061] in, Indicates the real-time dynamic viscosity of the slurry; This represents the basic slurry viscosity, a reference rheological parameter measured by a rotational rheometer under standard mix conditions; This is the obstruction potential field influence coefficient, used to describe the degree to which the particle contact network amplifies the viscosity of the base slurry. It is determined using the five-fold cross-validation method, and the reference value range is [value missing]. ; It refers to the mass percentage of ultrafine particles, such as the mass percentage of particles with a diameter of less than 20 micrometers. This is the synchronized tailings moisture content data; It is the water absorption enhancement coefficient of fine particles, determined by the five-fold cross-validation method, with a reference range of values. ; Used to describe the enhancing effect of particle blockage potential field on the internal shear resistance of slurry; This describes the effect of reduced effective flowing water caused by the adsorption of free water by fine particles.

[0062] Based on the final obtained real-time dynamic viscosity of the slurry The viscosity is compared with a preset target viscosity range based on user needs, and corresponding adjustment instructions are generated. When the real-time dynamic viscosity of the slurry... When the viscosity exceeds the upper limit of the target viscosity range, it is determined that the slurry flow resistance has increased. At this time, the tailings concentration is reduced according to a preset adjustment ratio, the water flow rate is increased, and the operating frequency of the delivery pump is increased simultaneously. The tailings concentration adjustment is proportionally reduced based on the viscosity deviation percentage obtained by subtracting the upper limit of the target viscosity range from the real-time dynamic viscosity of the slurry. The water flow rate is incrementally compensated based on the viscosity growth rate per unit time, and the delivery pump frequency is synchronously increased based on the pipeline pressure change. When the real-time dynamic viscosity of the slurry... If the viscosity is below the lower limit of the target viscosity range, the slurry is deemed to have excessive fluidity. In this case, the water addition should be reduced, the tailings concentration increased, and the conveying velocity appropriately decreased to improve the suspension stability of the slurry particles. When the real-time dynamic viscosity of the slurry... When the slurry is within the target viscosity range, the current tailings concentration, water-cement ratio, and conveying flow rate are kept constant, thereby achieving dynamic and stable control of the slurry rheological state.

[0063] In summary, a method for optimizing the proportion of tailings backfill based on multi-source sensing has been developed.

[0064] The order of the embodiments is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0065] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0066] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for optimizing the proportion of tailings backfill based on multi-source sensing, characterized in that, Includes the following steps: S1. Based on the collected multi-source raw industrial data, establish a raw dataset; based on the raw dataset, introduce a dynamic synchronization mechanism based on the flow regime change trend, and construct a pre-synchronization flow regime change factor, the calculation formula of which is: ； in, It is a pre-synchronous flow regime change factor; Indicates the current time Normalized slurry volumetric flow rate; Indicates the current time Normalized delivery pressure; This indicates the particle size at which the cumulative particle size distribution reaches 10%. Indicates the median diameter of the particle; This indicates the particle size at which the cumulative particle size distribution reaches 90%. The synchronization time window width is dynamically adjusted to achieve precise time synchronization of the original dataset, resulting in synchronized multi-source industrial data and synchronized pre-synchronized flow regime change factors. The specific expression for the synchronization time window width is: ； in, This indicates the width of the synchronization time window at the current moment; This indicates the maximum allowed synchronization window under stable slurry conditions; This represents the minimum allowed synchronization window under abnormal fluctuation conditions; This represents the current pre-synchronization flow regime change factor; This represents the window sensitivity adjustment coefficient; S2. Based on synchronized multi-source industrial data, including synchronized tailings particle size distribution data, conveying pressure data, pipeline vibration data, slurry surface texture image data, slurry conveying flow rate data, and tailings moisture content data, and a synchronized pre-synchronized flow regime change factor, particle mass ratio, local particle aggregation density, and normalized local slurry flow capacity are introduced. Based on the synchronized tailings particle size distribution data, particles are divided according to a set particle size range, and the particle mass of each particle size range is statistically analyzed. Based on the particle mass of each particle size range, the mass ratio of each type of particle is calculated. The difference between the synchronized conveying pressure data and the average pressure within the synchronization time window is used as the pressure state characteristic data. The synchronized pipeline vibration data were subjected to a fast Fourier transform to obtain the low-frequency vibration energy. Based on the synchronized slurry surface texture image data, the slurry texture entropy value is calculated and subtracted from the slurry texture entropy value at the previous moment to obtain the change in texture entropy. ; The formula for calculating local particle aggregation density is: ； in, Indicates the local particle aggregation density; It is the standard slurry density; , , These represent the pressure response weighting coefficient, vibration response weighting coefficient, and texture entropy response weighting coefficient, respectively. Based on the synchronized slurry delivery flow rate data and the obtained pipe cross-sectional area, the ratio of the synchronized slurry delivery flow rate data to the pipe cross-sectional area is used as the average flow velocity. This average velocity is then compared with the obtained standard steady-state delivery velocity to obtain the normalized local slurry flow capacity. A particle blockage potential field analysis model is constructed to obtain the particle blockage potential field strength, the specific expression of which is: ； in, This represents the strength of the particle blocking potential field; This represents the pre-synchronization flow state change factor after synchronization; Indicates the first Percentage of particle-like mass; This represents the total number of particle categories; Indicates the critical packing density; It is the first The clogging contribution coefficient of particulate matter; It is the particle aggregation sensitivity coefficient; It is a speed-related factor; Based on the tailings moisture content data and particle blockage potential field strength, the basic slurry viscosity is introduced to obtain the real-time dynamic viscosity of the slurry. The calculation formula is as follows: ； in, Indicates the real-time dynamic viscosity of the slurry; Indicates the viscosity of the base slurry; It is the influence coefficient of the blocking potential field; This represents the strength of the particle blocking potential field; It is the mass percentage of ultrafine particles; This is the synchronized tailings moisture content data; It is the water absorption enhancement coefficient of fine particles; Based on the real-time dynamic viscosity of the slurry, it is compared with the set target viscosity range to generate corresponding adjustment instructions.

2. The method for optimizing the proportion of tailings backfill based on multi-source sensing according to claim 1, characterized in that, S1 specifically includes: In the dynamic synchronization mechanism based on the flow regime change trend, the slurry transport flow rate data and transport pressure data in the original dataset are normalized to obtain normalized slurry volumetric flow rate and normalized transport pressure, and then the normalized flow rate change rate and normalized pressure change rate are calculated.

3. The method for optimizing the proportion of tailings backfill based on multi-source sensing according to claim 1, characterized in that, S2 specifically includes: When the real-time dynamic viscosity of the slurry is higher than the upper limit of the target viscosity range, it is determined that the slurry flow resistance has increased. The tailings mass concentration is reduced, the water flow rate is increased, and the operating frequency of the delivery pump is increased simultaneously according to the set adjustment ratio. When the real-time dynamic viscosity of the slurry is lower than the lower limit of the target viscosity range, it is determined that the slurry is too fluid. The water flow rate is reduced, the tailings concentration is increased, and the delivery flow rate is reduced. When the real-time dynamic viscosity of the slurry is within the target viscosity range, the current tailings concentration, water-cement ratio, and delivery flow rate are maintained unchanged.

Images