Intelligent backfill process design method based on multi-dimensional data analysis and optimization control

Abstract

The invention provides a backfill process intelligent design method based on multi-dimensional data analysis and optimization control, which comprises the following steps: judging whether a material difference exceeds a preset threshold according to a classified material group, and if so, predicting non-uniform distribution of a soil body through a support vector machine model, and determining coordinates of a potential non-uniform region; side slope instability risk points are extracted from the coordinates of the potential non-uniform area, stress distribution under the backfilling technology is analyzed and simulated through finite elements, and a stress distribution map is obtained; for the stress distribution map, obtaining a material selection scheme, adjusting a material combination proportion through an optimization algorithm, and determining an optimized material matching sequence; generating an inner waste dump three-dimensional model through the backfill process parameters, and evaluating a long-term stability index by adopting a numerical simulation method to obtain a stability evaluation score; and optimized feedback data are extracted from the stability evaluation score, whether the long-term stability requirement is met or not is judged, and if yes, a final backfill process scheme is output to control settlement and slope instability.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a backfill process intelligent design method based on multi-dimensional data analysis and optimization control. BACKGROUND

[0002] In the design of the backfill process of the dump in the open-pit mine area, there is a core technical problem, that is, how to achieve the balance between settlement control and long-term stability of the slope under complex geological conditions through optimization of material selection and collocation.

[0003] The physical properties and mechanical parameters of the discarded materials are quite different, and the data collected by the sensors have high-dimensional and nonlinear characteristics. Although the k-means clustering algorithm can initially classify the material groups, the classification results may be fuzzy due to the discreteness of the material parameters, making it difficult to accurately determine whether the difference threshold is exceeded, which further affects the accuracy of the support vector machine model in predicting the uneven distribution of the soil body, and the extraction of the coordinates of the potential uneven area may be biased.

[0004] When the finite element analysis based on these coordinates simulates the stress distribution under the backfill process, the stress distribution map may not accurately reflect the actual stress state due to the material differences and the complexity of the boundary conditions, resulting in inaccurate identification of the slope instability risk points.

[0005] When optimizing the collocation ratio of the materials, the optimization algorithm may fall into local optimization due to the high-dimensional parameter space and nonlinear constraints, and the generated material collocation sequence may not be able to stably achieve the preset threshold when simulating the backfill process.

[0006] In addition, the construction of the three-dimensional model of the internal dump is limited by the evaluation accuracy of the long-term stability index by numerical simulation methods, and the stability evaluation score may be distorted due to the uncertainty of the geological parameters and the simplification of the model, making it difficult for the optimization feedback data to accurately reflect the actual engineering requirements.

[0007] This core problem runs through the whole process of material classification, stress analysis, process optimization, and stability evaluation, involving high-dimensional data processing, model prediction accuracy, and engineering applicability under complex geological conditions, and needs to be solved to ensure the scientificity and reliability of the backfill process. SUMMARY

[0008] The present application provides a backfill process intelligent design method based on multi-dimensional data analysis and optimization control, mainly including: The sample data of the discarded material is acquired, physical performance indexes and mechanical performance parameters are collected through sensors, a k-means clustering algorithm is used to classify the material differences, and the classified material groups are obtained.

[0009] Further, if the material difference exceeds the preset threshold, a support vector machine model is used to predict the uneven distribution of the soil body, and the coordinates of the potential uneven area are determined.

[0010] Further, the stress distribution atlas is obtained.

[0011] Further, the material combination ratio is adjusted by an optimization algorithm, and the optimized material matching sequence is determined.

[0012] Further, if the simulated settlement value is lower than the preset threshold, the backfill process parameters are output, otherwise the material matching sequence is iteratively adjusted.

[0013] Further, the stability evaluation score is obtained.

[0014] Further, if the long-term stability requirement is met, the final backfill process scheme is output to control settlement and slope instability.

[0015] The technical scheme provided by the embodiment of the application can include the following beneficial effects: This invention discloses an intelligent design method for backfilling processes based on multidimensional data analysis and optimization control. Addressing the settlement and slope instability problems of spoil heaps in open-pit mines, it collects physical and mechanical property data of materials using sensors, employs k-means clustering to classify material differences, accurately identifies areas of difference exceeding thresholds, and utilizes support vector machines to predict soil unevenness and determine the coordinates of potential risk points. Combined with finite element analysis to simulate the stress distribution of the backfilling process, a stress spectrum is generated. Then, an optimization algorithm adjusts the material mix ratio, outputting an optimized material sequence. Based on this sequence, the backfilling process is simulated to evaluate the settlement control effect. If the target is not met, the material mix is ​​iteratively optimized to ultimately generate backfilling process parameters that satisfy settlement control. A three-dimensional model of the spoil heap is constructed, and long-term stability is evaluated through numerical simulation, outputting the final process scheme. This invention achieves precise design and stability control of the backfilling process through seamless integration of data-driven and intelligent optimization, effectively reducing the risks of settlement and slope instability and improving the efficiency of mine environmental governance. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to specific embodiments.

[0017] This embodiment of an intelligent design method for backfilling processes based on multidimensional data analysis and optimization control may specifically include: In one embodiment, the present invention provides a method for optimizing the classification and backfilling process of waste disposal materials. The aim is to effectively control settlement and slope instability of internal spoil heaps through precise analysis of material characteristics and scientific adjustment of process parameters. The following detailed description of the implementation of this method, with specific steps, aims to provide sufficient support for the technical solution through a logically clear process and diverse embodiments.

[0018] Step S1 involves acquiring waste material sample data. Physical and mechanical performance parameters are collected using sensors, and a specific clustering algorithm is employed to classify the material differences, resulting in categorized material groups. This process is fundamental to the entire method, aiming to extract key characteristic data from complex waste materials, laying the foundation for subsequent analysis. Specific implementation methods include collecting waste material samples from actual engineering sites, acquiring basic property data of the materials using various sensor devices, and scientifically grouping the materials through data processing and classification algorithms.

[0019] Step S11 involves acquiring physical and mechanical performance parameters from the waste material sample using sensors and storing them as a raw dataset. In practice, sensors may include pressure sensors, density meters, and humidity detection devices to measure physical performance parameters such as density, hardness, and moisture content of the material. Simultaneously, tensile or compression testing equipment is used to acquire mechanical performance indicators such as compressive strength and shear strength. After collection, this data is stored in a structured format as a raw dataset for subsequent processing. It should be noted that the selection and placement of sensors need to be adjusted according to the material type and engineering environment. For example, in an open-pit mine environment, the adaptability of the equipment to dust and temperature must be considered to ensure the accuracy of data acquisition.

[0020] Step S12 involves preprocessing the original dataset by using standardization to eliminate dimensional differences, resulting in a standardized dataset. Since different performance parameters have significantly different dimensions and numerical ranges—for example, density and compressive strength have completely different units and numerical ranges—directly using the original data could lead to biases in subsequent analyses. Therefore, the preprocessing stage uses standardization to unify the numerical ranges of different parameters to a similar scale. Typically, this process involves subtracting the mean from the value of each parameter and then dividing by the standard deviation, resulting in a standardized dataset with a mean of 0 and a standard deviation of 1. This processing method effectively eliminates the influence of dimensions, providing a fair data foundation for subsequent classification.

[0021] Step S13: If the standardized dataset contains missing values, the mean imputation method is used to fill in the missing values ​​to obtain the complete dataset. In actual data acquisition, some data may be missing due to equipment malfunctions or environmental interference; for example, the moisture content data of some material samples may not have been successfully recorded. To solve this problem, the mean imputation method is used to fill in missing values, that is, to estimate the missing values ​​based on the average value of the same parameter in other samples.

[0022] For example, if the density data for a certain sample is missing, the mean density of all other samples is calculated, and this value is used to fill the missing position. This method ensures the integrity of the dataset and avoids bias in subsequent classification results due to missing data.

[0023] Step S14: Based on the complete dataset, a specific clustering algorithm is used to classify the material characteristics, obtaining preliminary clustering results. The core of the clustering algorithm is to group similar materials into the same group based on their physical and mechanical properties. In this process, the algorithm groups data points according to the distance or similarity between them, for example, grouping materials with similar density and compressive strength into one category. The preliminary clustering results typically include multiple material groups, each representing a class of materials with similar characteristics. It should be noted that the preliminary clustering results may contain some errors, therefore subsequent steps will further optimize the classification effect.

[0024] Step S15 involves iteratively optimizing the parameters of the clustering algorithm and adjusting the cluster centers to obtain optimized clustering results. Based on the initial clustering, the number and position of the cluster centers are adjusted through multiple iterations to increase the similarity of materials within groups and the differences between groups.

[0025] For example, the initial clustering might set the number of cluster centers to 3, but during iteration, it might be found that adjusting the number of centers to 5 better reflects the distribution of material characteristics. Therefore, the parameters are dynamically adjusted based on the data characteristics. After multiple iterations, the optimized clustering results can more accurately reflect the true differences in materials, providing a reliable basis for subsequent analysis.

[0026] Step S16: Based on the optimized clustering results, determine the material difference classification and generate classified material groups. In this step, based on the optimized clustering results, the materials are divided into several groups, each group representing a specific combination of material characteristics.

[0027] For example, one group might contain materials with higher density and compressive strength, while another group might contain materials with lower density and higher moisture content. These material classifications provide the basic data for subsequent difference analysis and process optimization.

[0028] Step S17: Extract the performance parameter distribution of each material group from the classified material groups and generate a material characteristic report. For each material group, statistically analyze the distribution of its physical and mechanical property parameters. For example, the mean density, standard deviation, and maximum and minimum compressive strength of each material group can be calculated. These statistical data are used to generate detailed material characteristic reports describing the properties and differences of each material group. Such reports not only provide engineers with intuitive material classification information but also provide data support for subsequent process design and risk assessment.

[0029] In one embodiment, the specific implementation of steps S11 to S17 can be illustrated using the classification of waste materials in an open-pit mine as an example. Assume that in a certain mining environment, waste materials mainly include rock fragments, soil mixtures, and tailings slag. When collecting samples, the density and compressive strength data of the materials are obtained using a portable density meter and pressure sensor. During the preprocessing stage, it was found that the moisture content data of some samples was missing; therefore, a mean imputation method was used to fill the missing data based on the mean moisture content of other samples. During the clustering process, the materials were initially divided into 3 groups, but through iterative optimization, it was found that 5 groups better reflect the differences in material characteristics. Finally, a classification result and characteristic report containing 5 material groups were generated. Through this process, the applicable scenarios for different materials can be clearly distinguished. For example, high-density materials can be used for basic backfilling, while high-moisture-content materials need to be preprocessed before use.

[0030] Step S2 involves determining whether the material differences exceed a preset threshold based on the classified material groups. If so, a specific model is used to predict the uneven distribution of soil and determine the coordinates of potential uneven areas. This step aims to identify the impact of material characteristic differences on soil distribution, providing a basis for subsequent risk assessment. By quantitatively analyzing the differences between material groups and combining them with a spatial distribution prediction model, areas where uneven distribution may exist can be accurately located.

[0031] Step S21: Obtain material characteristic data from the categorized material groups and calculate the material difference values ​​between each group using statistical methods. During this process, for the performance parameters of each material group... For example, density and compressive strength, calculate the difference in mean or standard deviation between groups.

[0032] For example, if the average density of one group is 2.5, while that of another group is 1.8, then the density difference between the two groups is 0.7. By comprehensively calculating the differences in multiple parameters, the overall material difference between each group is obtained, which is used to determine whether further analysis is needed.

[0033] Step S22: If the material difference value exceeds a preset threshold, spatial distribution features are extracted from the material characteristic data to generate a feature dataset. The preset threshold is typically determined based on engineering experience or design requirements; for example, a significant material difference is considered to occur when the density difference exceeds 0.5 or the compressive strength difference exceeds a certain value. In this case, spatially relevant features are extracted from the material characteristic data. For example, the location coordinates and distribution density of materials at sampling points are used to generate a feature dataset containing spatial distribution information. This feature data will be used for subsequent distribution prediction.

[0034] Step S23: Based on the feature dataset, a specific model is used to predict the non-uniform distribution of soil, obtaining the probability of non-uniform distribution. The specific model predicts the possible non-uniform distribution of soil in different regions by analyzing the spatial distribution characteristics of materials.

[0035] For example, the model can determine whether there are significant density differences in certain areas based on the spatial variation trend of material density, thereby deriving a probability value for non-uniform distribution. The higher the probability value, the greater the likelihood of non-uniform distribution in that area.

[0036] Step S24: Extract high-probability regions from the uneven distribution probability and determine the initial coordinates of potential uneven regions. In the prediction results, filter out regions where the uneven distribution probability exceeds a certain threshold. For example, regions with a probability value greater than 0.8 are considered potential non-uniform regions, and their initial coordinates are recorded. These coordinates are typically expressed as the latitude and longitude of the sampling points or their relative positions, providing preliminary localization for subsequent precise analysis.

[0037] Step S25: Based on the initial coordinates, a spatial interpolation method is used to delineate the boundaries of the non-uniform region, obtaining accurate region coordinates. The spatial interpolation method estimates the data of the region surrounding the initial coordinates to draw the continuous boundary of the non-uniform region.

[0038] For example, data interpolation between initial coordinate points generates smooth boundary lines, thereby expanding the extent of the non-uniform region from discrete points to a continuous area, resulting in more accurate regional coordinates. This process improves positioning accuracy and provides more reliable spatial information for subsequent analysis.

[0039] Step S26: Obtain local features of soil distribution from precise regional coordinates, and use cluster analysis to verify the consistency of the uneven distribution in the region, thus determining the final coordinates. In this step, local features, such as local density variations or intensity distribution characteristics, are extracted from the soil distribution data within the precise regional coordinates, and cluster analysis is used to verify the consistency of these features. If the local features show significant differences, it indicates that uneven distribution does indeed exist in the region, and the coordinates of that region are ultimately determined as the analysis result.

[0040] Step S27: Based on the final coordinates, generate a spatial distribution map of the uneven area to determine the location of the uneven soil distribution. Based on the final coordinates, draw a spatial distribution map of the uneven area, clearly showing the distribution range and characteristic differences of each area. This spatial distribution map provides an intuitive basis for subsequent risk assessment and process optimization, helping engineers quickly identify potential problem areas.

[0041] In one embodiment, the specific implementation of steps S21 to S27 can be illustrated using the soil distribution analysis of a spoil heap within a mine as an example. Assume that in a certain spoil heap, the density difference between material groups reaches 0.6, exceeding a preset threshold of 0.5, thus requiring uneven distribution prediction. After extracting spatial distribution features from the material characteristic data, model prediction reveals that the uneven distribution probability in a certain area is as high as 0.85. Subsequently, spatial interpolation methods are used to accurately delineate the boundary of this area, ultimately generating a spatial distribution map. Through this process, the uneven soil distribution area can be accurately located, providing an important reference for subsequent slope instability risk analysis.

[0042] Step S3 involves extracting slope instability risk points from the coordinates of potential uneven regions and simulating the stress distribution under backfilling conditions using specific analytical methods to obtain a stress distribution map. This step aims to assess the impact of unevenly distributed areas on slope stability, identify potential high-risk areas by simulating stress changes during the backfilling process, and provide a basis for subsequent process optimization.

[0043] Step S31: Obtain geological parameters from the coordinates of the non-uniform region, extract slope instability risk points using point cloud processing technology, and determine the coordinate set of risk points. In this process, based on the coordinates of the non-uniform region, obtain geological parameters within that region, such as soil density and cohesion, and analyze the spatial distribution characteristics of these parameters using point cloud processing technology to identify key points that may lead to slope instability.

[0044] For example, points with significantly lower density than the surrounding area may be at risk of instability. These points are recorded as a risk point coordinate set, providing a basis for subsequent analysis.

[0045] Step S32: Construct a computational grid based on the risk point coordinate set, and initialize the grid properties using a specific analysis method to obtain the initial stress field. For the risk point coordinate set, a computational grid covering the entire analysis area is constructed, with each grid cell corresponding to a specific spatial range. Subsequently, the properties of each grid cell are initialized according to geological parameters, such as assigning it initial stress values ​​and deformation characteristics, thereby generating the initial stress field. This initial stress field reflects the soil stress distribution state before the backfilling process is applied.

[0046] Step S33 involves adjusting the initial stress field using backfilling process parameters and updating the stress values ​​of the grid nodes through iterative calculations to obtain the dynamic stress distribution. In this step, backfilling process parameters, such as the type, thickness, and compaction method of the backfill material, are introduced to simulate the impact of backfilling on soil stress. Through multiple iterative calculations, the stress value of each grid node is updated, resulting in a dynamic stress distribution reflecting the influence of the backfilling process. This distribution visually demonstrates the stress change trend during backfilling, providing data support for risk assessment.

[0047] Step S34: If the stress value at a node in the dynamic stress distribution exceeds a preset threshold, the node is marked as a high-risk instability point, and a set of high-risk points is generated. The preset threshold is usually determined according to engineering design standards. For example, if the stress value at a certain node exceeds the ultimate shear strength of the soil, that node is considered to be at risk of instability. By screening all nodes in the dynamic stress distribution, nodes with stress values ​​exceeding the limit are marked, forming a high-risk point set. These high-risk point sets collectively reflect the instability areas that may be caused by backfilling processes.

[0048] Step S35 involves performing spatial interpolation analysis on the high-risk point set, using a specific interpolation algorithm to generate a continuous stress distribution map. Based on the location and stress values ​​of the high-risk point set, the stress distribution in the surrounding area is estimated using spatial interpolation methods, generating a continuous stress distribution map. This map visually displays the spatial trend of stress variation; for example, some areas may exhibit stress concentration, while other areas are more uniform. This map provides a more comprehensive understanding of the impact of backfilling techniques on slope stability.

[0049] Step S36: Extract contour data from the stress distribution map, and use vector field analysis to determine the instability probability distribution, thus obtaining a slope instability risk map. Based on the stress distribution map, extract contour data of stress values ​​and analyze the spatial gradient change and distribution characteristics of stress. Using vector field analysis, calculate the instability probability of each region; for example, the instability probability is usually higher in regions with larger stress gradients. Finally, generate a slope instability risk map, clearly showing the instability risk level of different regions.

[0050] Step S37: Based on the slope instability risk map, the region is segmented, and cluster analysis is used to divide the area into high-risk and low-risk zones, obtaining the zoning results. Based on the slope instability risk map, cluster analysis is used to divide the region into high-risk and low-risk areas.

[0051] For example, areas with an instability probability higher than a certain threshold can be classified as high-risk areas, while areas with a lower probability can be classified as low-risk areas. This zoning result can provide targeted guidance for subsequent process optimization, such as adopting stricter backfill control measures for high-risk areas.

[0052] In one embodiment, the specific implementation of steps S31 to S37 can be illustrated using the slope stability analysis of a spoil heap within an open-pit mine as an example. Assume that within a non-uniform area of ​​a spoil heap, multiple low-density points are identified as a risk point coordinate set using point cloud processing technology. After constructing a computational grid and initializing the stress field, simulation of the backfilling process reveals that the stress values ​​of some nodes significantly exceed the threshold; these nodes are marked as high-risk instability points. A continuous stress distribution map is generated through spatial interpolation, and contour data is further extracted to ultimately delineate high-risk and low-risk areas. This process helps engineers accurately identify slope instability risk areas, enabling targeted measures to be taken in subsequent backfilling processes to prevent instability accidents.

[0053] In another embodiment, the stress distribution simulation in step S3 can take into account the influence of different backfilling process parameters.

[0054] For example, in the first backfilling scheme, high-density materials are used for layered backfilling with smaller thicknesses each time, resulting in higher compaction. In the second backfilling scheme, a mixture of materials is used for one-time backfilling, resulting in lower compaction. Simulations of stress distribution under both schemes revealed that the stress values ​​in high-risk areas were significantly lower in the first scheme than in the second, indicating that layered backfilling and high compaction can effectively reduce the risk of instability. This comparative analysis provides an important reference for subsequent process optimization.

[0055] By implementing steps S1 to S3 above, we can progress from the classification of waste materials to the prediction of uneven soil distribution and the assessment of slope instability risks, laying a solid foundation for the optimization of subsequent backfilling processes. Each step is implemented in detail with diverse examples, ensuring the operability and applicability of the method.

[0056] Step S4 involves obtaining a material selection scheme based on the stress distribution map, adjusting the material combination ratio through an optimization algorithm, and determining the optimal material combination sequence. This step aims to scientifically select backfill materials based on the stress distribution analysis obtained earlier, and adjust the material ratio through optimization methods to reduce stress concentration and improve slope stability. This fine-tuning of the material combination provides a reliable basis for subsequent backfilling processes.

[0057] Step S41: Obtain stress distribution map data and generate stress distribution data using a specific analysis method. In this process, detailed stress distribution information is extracted from the previously generated stress distribution map, including the magnitude of stress values ​​in each region, their distribution range, and stress gradient changes. This data is stored in a structured format for subsequent material selection and optimization analysis. It should be noted that the extraction of stress distribution data must cover the entire analysis area to ensure the comprehensiveness and representativeness of the data.

[0058] Step S42 involves extracting material property features from the stress distribution data using a specific analytical method to obtain a material property feature set. For the stress distribution data, the influencing factors of different regions on the stress value are analyzed. For example, high-stress areas may require backfilling with high-strength materials, while low-stress areas can use ordinary materials. By extracting features related to material properties, such as compressive strength, density, and their correlation with stress distribution, a material property feature set is generated. This feature set provides crucial input for subsequent optimization.

[0059] Step S43: Based on the material property feature set, construct an optimization objective function and determine the initial material combination ratio. In this step, an optimization objective function is constructed based on the material property feature set to measure the effect of the material combination ratio on reducing stress concentration and improving stability.

[0060] For example, the objective function can be set to minimize the stress value in high-stress areas, while also considering material costs and construction feasibility. The initial material mix ratio is usually determined based on experience or historical data. For example, the proportion of high-strength materials can be set to a certain percentage, with the remainder being ordinary materials.

[0061] Step S44 involves iteratively adjusting the material combination ratio using a specific optimization algorithm to obtain the optimized ratio. The optimization algorithm continuously adjusts the material combination ratio through multiple iterations to approximate the optimal solution of the objective function.

[0062] For example, in each iteration, the proportion of high-strength materials in high-stress regions is increased while the use of low-strength materials is reduced, and the change in the objective function value is observed. After multiple iterations, an optimized combination ratio is obtained, which can minimize material costs while meeting stress distribution requirements.

[0063] Step S45: If the optimized combination ratio meets the preset stress distribution threshold, a material combination sequence is generated. The preset stress distribution threshold is usually determined according to engineering design standards, such as requiring the stress value in high-stress areas to be below a certain safety value. If the optimized combination ratio can make the stress distribution meet this threshold, a material combination sequence is generated according to the ratio, specifying the order and quantity of each material used in the backfilling process. This sequence provides direct guidance for subsequent process simulation.

[0064] Step S46: Based on the material combination sequence, verify the stress analysis results and determine whether they meet the design requirements. In this step, based on the material combination sequence, re-simulate the stress distribution under the backfilling process to verify whether the optimized material combination can effectively reduce the stress value in high-stress areas. If the simulation results show that the stress distribution meets the design requirements, for example, the stress values ​​in all areas are within the safe range, then the material combination sequence is considered feasible; if not, further adjustments are needed.

[0065] Step S47: By adjusting and optimizing the parameters of the objective function, the material combination sequence is iteratively updated to determine the final material combination sequence. Regarding the issues discovered during verification, For example, if the stress value still exceeds the standard in some areas, the parameters of the objective function should be adjusted and optimized. For example, the weighting of high-stress areas can be increased, or new constraints, such as material availability limitations, can be introduced. Then, iterative optimization is performed again, updating the material combination sequence until the simulation results meet the design requirements, ultimately determining the final material combination sequence. This sequence will serve as a crucial input to the backfilling process.

[0066] In one embodiment, the specific implementation of steps S41 to S47 can be illustrated using the material optimization of a spoil heap in an open-pit mine as an example. Assuming that in the stress distribution map of a certain spoil heap, the high-stress area is concentrated at the bottom of the slope, after analyzing and extracting the material properties, an objective function is constructed, initially setting the proportion of high-strength materials at 30%, with the remainder being ordinary materials. After multiple rounds of iterative optimization, it is found that increasing the proportion of high-strength materials to 50% significantly improves the stress distribution and meets the safety threshold. Subsequently, a material combination sequence is generated, and its effect is verified in simulation. Finally, this sequence is determined as the basic scheme for the backfilling process. Through this process, stability and cost can be effectively balanced, and the utilization efficiency of backfill materials can be optimized.

[0067] Step S5 involves simulating the backfilling process from the optimized material combination sequence to assess the settlement control effect. If the simulated settlement value is lower than a preset threshold, the backfilling process parameters are output; otherwise, the material combination sequence is iteratively adjusted. This step aims to evaluate the impact of the material combination sequence on settlement control by simulating the backfilling process, ensuring that the settlement of the soil after backfilling is within an acceptable range, and providing reliable parameters for subsequent construction.

[0068] Step S51 involves obtaining the material combination sequence and backfilling process parameters, and generating simulated settlement values ​​through process simulation. In this process, based on the previously determined material combination sequence and combined with backfilling process parameters such as backfill thickness, layering method, and compaction degree, the settlement of the soil during backfilling is simulated. The simulated settlement values ​​reflect the amount of soil settlement in different areas after backfilling, and are typically calculated using numerical simulation methods to provide data support for subsequent evaluation.

[0069] Step S52: If the simulated settlement value is lower than the preset threshold, output the backfilling process parameters and determine the process parameters. The preset threshold is determined according to the engineering design requirements. For example, the maximum settlement value must not exceed a certain specific value to ensure soil stability. If the simulated settlement value meets this requirement, the current material combination sequence and backfilling process parameters are considered feasible, and these parameters are directly output as a guide for subsequent construction. These parameters include detailed information such as material type, proportion, and backfilling sequence.

[0070] Step S53: Based on the simulated settlement values ​​and settlement monitoring data, calculate the settlement control effect and obtain a control effect evaluation. In this step, the simulated settlement values ​​are compared with the actual settlement monitoring data to analyze the differences between the two and evaluate the actual effect of the current process parameters on settlement control.

[0071] For example, if the simulated settlement value is basically consistent with the monitoring data and both are within the safe range, the control effect is considered good; if there is a large deviation, further analysis of the reasons is needed to generate a control effect evaluation report.

[0072] Step S54: If the control effect assessment fails to meet the parameter optimization target, an iterative adjustment strategy is used to update the material combination sequence. Parameter optimization targets typically include requirements for settlement control accuracy and stability. If the assessment results show that the targets have not been met, For example, if the settlement value in some areas still exceeds the standard, the material combination sequence needs to be adjusted. Iterative adjustment strategies include increasing the proportion of high-strength materials and changing the backfill stratification method. Through multiple adjustments, new material combination sequences are generated to provide new inputs for subsequent simulations.

[0073] Step S55: A new material combination sequence is generated using sequence generation rules to obtain an optimized material sequence. Based on the iterative adjustment strategy and considering the issues identified in the settlement control effect evaluation, a new material combination sequence is generated according to specific rules.

[0074] For example, if a region is found to have a high settlement value, the proportion of high-strength materials used in that region is increased, and the backfill thickness is adjusted to generate an optimized material sequence. This sequence will be used in the next process simulation to verify its effectiveness.

[0075] Step S56: The process simulation is re-executed using an optimized material sequence to generate new simulated settlement values. In this step, based on the optimized material sequence, the backfilling process is re-simulated to calculate the soil settlement under the new material combination and process parameters, generating new simulated settlement values. These values ​​will be used to further evaluate the settlement control effect and determine whether the design requirements are met.

[0076] Step S57: Based on the new simulated settlement value and the control effect evaluation, determine whether the parameter optimization target has been achieved, and output the final backfilling process parameters. If the new simulated settlement value is lower than the preset threshold, and the control effect evaluation shows that the optimization target has been achieved, then the current material combination sequence and process parameters are considered to meet the requirements, and the final backfilling process parameters are output. These parameters will serve as the final guidance for construction, ensuring that the settlement of the soil after backfilling is controlled within a safe range.

[0077] In one embodiment, the specific implementation of steps S51 to S57 can be illustrated using the settlement control of a spoil heap in an open-pit mine as an example. Suppose that in a certain spoil heap, after initial material mix sequence simulation, it is found that the settlement value in the top area of ​​the slope exceeds a preset threshold. Comparison with monitoring data indicates that the control effect is unsatisfactory. Subsequently, the material mix sequence is adjusted, increasing the proportion of high-strength materials in this area and reducing the backfill thickness. After resimulating, the settlement value is found to be significantly reduced, meeting the design requirements, and the backfilling process parameters are finally output. Through this process, the settlement can be effectively controlled, avoiding structural problems caused by excessive settlement.

[0078] Step S6 involves generating a three-dimensional model of the internal spoil heap using backfilling process parameters, and then employing numerical simulation methods to evaluate long-term stability indicators and obtain a stability assessment score. This step aims to simulate the stability performance of the internal spoil heap during long-term use by constructing a three-dimensional model, providing a basis for determining the final process plan.

[0079] Step S61: Obtain backfilling process parameters and geological data, and construct a three-dimensional model of the internal spoil heap. In this process, based on the previously determined backfilling process parameters and combined with on-site geological data, For example, soil layer thickness and groundwater level are used to construct a three-dimensional model of the internal spoil disposal site. This model digitally represents the spatial structure and material distribution of the spoil disposal site, providing a foundation for subsequent simulations.

[0080] Step S62: Calculate the stress distribution in the three-dimensional model using numerical simulation. Based on the three-dimensional model, calculate the stress distribution of the soil site under the current backfilling process parameters using numerical simulation. For example, analyzing the magnitude and distribution characteristics of stress values ​​in different regions. This distribution reflects the stress changes that the soil site may face during long-term use, providing data support for stability assessment.

[0081] Step S63: If the stress distribution exceeds a preset threshold, adjust the material properties and recalculate the stress distribution. The preset threshold is usually determined according to engineering safety standards. If the simulation results show that the stress values ​​in some areas exceed the standard, the properties of the backfill material need to be adjusted. For example, increase the proportion of high-strength materials or change the compaction method. Then, recalculate the stress distribution, observe the effect of the adjustment, and ensure that the stress value is within a safe range.

[0082] Step S64 involves using a specific analysis method to simulate long-term stability and obtain the deformation trend. In this step, based on the adjusted stress distribution, the deformation of the internal spoil heap during long-term use is simulated. For example, trends such as soil settlement and slope slippage. By analyzing deformation trends, we can determine whether the soil site can remain stable in the long term, providing a basis for subsequent assessments.

[0083] Step S65: Calculate the safety factor and determine the stability score based on the deformation trend. Calculate the safety factor for the internal spoil heap based on the deformation trend. For example, the safety factor is obtained by comparing the ratio of deformation to allowable deformation. If the safety factor is higher than a certain threshold, the soil site is considered to have high stability and is given a higher stability score; if it is lower than the threshold, further adjustments are needed.

[0084] In step S66, if the safety factor is lower than a preset threshold, environmental factors are introduced, and the long-term stability is re-simulated. Environmental factors include external conditions such as rainfall and temperature changes. If the safety factor does not meet the standard, these factors are included in the simulation to analyze their impact on the stability of the soil field.

[0085] For example, rainfall may increase soil moisture content, thereby reducing stability. After resimulating, observing the change in the safety factor provides a reference for subsequent optimization.

[0086] Step S67: By weighted calculation, the safety factor and environmental factors are integrated to obtain the final stability score. In this step, the influence of the safety factor and environmental factors are comprehensively considered, and the final stability score is generated through a weighted calculation method.

[0087] For example, the safety factor has the main weight, while environmental factors have the secondary weight. The final score reflects the overall stability of the soil field in long-term use.

[0088] In one embodiment, the specific implementation of steps S61 to S67 can be illustrated using the long-term stability assessment of a spoil heap within an open-pit mine as an example. Assume that in a certain spoil heap, after constructing a three-dimensional model based on backfilling process parameters, simulation reveals that the stress value at the bottom of the slope exceeds the standard. Subsequently, the material properties are adjusted, increasing the proportion of high-strength materials, and after recalculation, the stress distribution returns to normal. In the long-term stability simulation, the safety factor is found to decrease slightly after introducing rainfall, but through weighted calculation, the final stability score still meets the safety standard. This process provides a reliable basis for determining subsequent process schemes, ensuring the stability of the spoil heap during long-term use.

[0089] Step S7 involves extracting optimization feedback data from the stability assessment score to determine if the long-term stability requirements are met. If so, the final backfilling process plan is output to control settlement and slope instability. This step aims to synthesize the results of previous analyses to generate the final backfilling process plan, ensuring the long-term stability and safety of the internal spoil heap.

[0090] Step S71: Obtain raw geological data from the geological database, remove outliers using data cleaning methods, and obtain a standardized geological dataset. During this process, extract raw data related to the internal spoil heap from the geological database. For example, soil layer thickness, groundwater level, etc., and outliers are removed through data cleaning methods. For example, data points that significantly deviate from the normal range. Subsequently, the data is standardized to ensure consistency in the dimensions of different parameters, generating a standardized geological dataset.

[0091] Step S72: If the key parameters in the standardized geological dataset meet the preset stability threshold, then the relationship between the geological data and the settlement trend is analyzed using a specific analysis method to obtain the settlement change trend. Key parameters include soil density, cohesion, etc. If these parameters meet the stability requirements, their relationship with the settlement trend is analyzed. For example, by comparing historical data, the impact of soil density changes on settlement can be determined, and settlement trend can be generated to provide a basis for subsequent assessment.

[0092] Step S73: Based on the settlement trend, a specific algorithm is used to classify the slope stability and determine whether there is a risk of instability. Based on the settlement trend, the stability performance of the slope in different regions is analyzed. For example, areas with large settlement may be at risk of instability. Using a classification method, slope stability is divided into two categories: stable and unstable. If the classification result is stable, the process proceeds to the next optimization step.

[0093] Step S74: If the slope stability classification result is stable, historical feedback information is extracted from the optimization feedback database to obtain backfill process optimization parameters. The optimization feedback database stores backfill process data and effect feedback from historical projects. If the current slope stability classification is stable, relevant parameters are extracted from it. For example, successful material ratios and backfilling methods from the past can serve as a reference for current process optimization.

[0094] Step S75: The backfilling process optimization parameters are iteratively adjusted using a specific optimization algorithm to generate a preliminary backfilling process plan. Based on historical feedback information, the backfilling process parameters are adjusted using an iterative optimization method. For example, optimizing material ratios and backfill thickness can generate a preliminary backfilling process plan. This plan will be used for subsequent verification to ensure its applicability in the current engineering environment.

[0095] Step S76: Based on the preliminary backfilling process plan and settlement monitoring data, a specific analysis method is used to predict the long-term settlement trend and determine whether the long-term stability criteria are met. In this step, based on the preliminary backfilling process plan and actual settlement monitoring data, the settlement trend of the soil site during long-term use is predicted. For example, analyze whether the settlement will gradually stabilize over time. If the prediction results show that the settlement trend meets the long-term stability criteria, proceed to the final solution output step.

[0096] Step S77: If the long-term settlement trend meets the stability criteria, the final backfilling process plan is output. This final backfilling process plan integrates all previous analysis results, including material matching sequences and backfilling process parameters, ensuring that the soil site can effectively control settlement and slope instability during long-term use. This plan will serve as the final guidance for construction, providing comprehensive support for the project.

[0097] In one embodiment, the specific implementation of steps S71 to S77 can be illustrated using the determination of the final process scheme for an internal spoil heap in an open-pit mine as an example. Assuming that the key parameters of the standardized geological dataset meet stability requirements in a certain internal spoil heap, and the slope stability is classified as stable by analyzing the settlement trend, then backfill parameters from successful cases are extracted from the historical feedback database. A preliminary scheme is generated through iterative optimization, and its effectiveness is verified in long-term settlement trend prediction. Finally, a backfill process scheme containing detailed information such as material proportions and layering methods is output. This scheme can effectively guide construction and ensure the long-term stable operation of the spoil heap.

[0098] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A smart design method for backfilling process based on multidimensional data analysis and optimization control, characterized in that, include: Acquire sample data of waste materials, collect physical performance indicators and mechanical performance parameters through sensors, and store them as raw datasets; Preprocessing operations are performed on the original dataset, and a standardization method is used to eliminate dimensional differences to generate a standardized dataset; For the missing values ​​in the standardized dataset, the mean imputation method is used to fill in the missing values ​​to form a complete dataset; Based on the complete dataset, a clustering algorithm is applied to classify the material properties and generate preliminary clustering results; By iteratively optimizing the parameters of the clustering algorithm and adjusting the cluster centers, an optimized clustering result is obtained. Based on the optimized clustering results, material difference classifications are determined, and material groups are generated. The method extracts the performance parameter distribution of each material group from the categorized material groups to form material characteristic data. Through multi-dimensional analysis of the material data, it constructs a classification model to distinguish differences in material characteristics, thus providing a data foundation for subsequent backfilling process design. The categorized material groups serve as input for subsequent process optimization, combining with the material characteristic data to drive precise backfilling process design. The method also includes in-depth mining of the distribution characteristics of material performance parameters to support dynamic adjustment of process parameters. The generation process of the material characteristic data covers the entire process from raw data collection to classification optimization, ensuring the accuracy and reliability of material classification and laying the foundation for intelligent backfilling process design.

2. The intelligent design method for backfilling process according to claim 1, characterized in that, The step of extracting the performance parameter distribution of each group of materials from the classified material groups to form material characteristic data includes: Obtain the physical and mechanical properties of each material group from the classified material groups; Statistical analysis was performed on the physical and mechanical performance parameters to calculate the distribution characteristics of the performance parameters of each group of materials. Based on the performance parameter distribution characteristics, a material characteristic dataset is constructed, covering the performance distribution range and key indicator values ​​of each group of materials. The material property dataset is stored as input data for subsequent process optimization.

3. The intelligent design method for backfilling process according to claim 1, characterized in that, The step of determining material difference classification and generating classified material groups based on the optimized clustering results includes: Based on the optimized clustering results, the difference values ​​between each material group are calculated; The differences were analyzed hierarchically to determine the material difference classification criteria. According to the material difference classification standard, the materials are divided into multiple groups to form the classified material groups; The material groups are verified, the classification boundaries are adjusted, and the final material classification results are generated.

4. The intelligent design method for backfilling process according to claim 2, characterized in that, The step of obtaining the physical and mechanical property parameters of each group of materials from the classified material groups includes: For each group of materials in the aforementioned material classification, extract the corresponding physical performance parameter data; The physical performance parameter data is cleaned to remove outliers; Simultaneously, mechanical performance index data from the classified material groups are extracted and integrated into a material performance dataset; The material performance dataset is standardized to form a unified set of performance parameters.

5. The intelligent design method for backfilling process according to claim 1, characterized in that, The step of performing preprocessing operations on the original dataset, using a standardization method to eliminate dimensional differences, and generating a standardized dataset includes: Perform dimensional analysis on the parameter values ​​in the original dataset to determine standardization rules; The data in the original dataset are transformed according to the standardization rules to generate the standardized dataset; The standardized dataset is subjected to consistency verification, the data format is adjusted, and the final standardized result is generated.

6. The intelligent design method for backfilling process according to claim 3, characterized in that, The step of performing stratified analysis on the difference values ​​to determine the material difference classification criteria includes: The difference values ​​are divided into intervals to generate the difference value distribution intervals; For each of the aforementioned difference value distribution intervals, calculate the material characteristic distribution features within the interval; Based on the material characteristic distribution features, the material difference classification criteria are determined; Generate a material classification hierarchy structure according to the aforementioned material difference classification criteria; The material classification hierarchy is optimized and adjusted to form the final classification standard.

7. The intelligent design method for backfilling process according to claim 6, characterized in that, The step of generating a material classification hierarchy structure according to the material difference classification standard includes: Based on the material difference classification standard, the material data is hierarchically divided to generate an initial classification hierarchy; Feature extraction is performed on the material data of each level in the initial classification hierarchy to determine the differences between levels; Based on the differences between the levels, the level boundaries are adjusted to generate the material classification hierarchy structure. The consistency of the material classification hierarchy is verified, the hierarchy division results are adjusted, and the final hierarchy structure is formed.