Method for digitally acquiring and optimizing mineral deposit information data

Abstract

The invention relates to the technical field of mineral deposit data, in particular to a digitalized acquisition optimization method for mineral deposit information data. Based on geological evolution historical data of the mining area, a three-dimensional geological model is established, and formation and change of an ore deposit in different evolution stages are simulated; dynamically simulating key parameters of the ore deposit; according to prediction of the geologic model, adjusting acquisition parameters; setting a layout scheme of a sensor array in combination with a geological evolution simulation model; the sensor array adjusts the acquisition angle, direction and sensing area in real time through a built-in adaptive algorithm; carrying out real-time synchronous acquisition on multi-modal data from different sensors, carrying out feature extraction through a deep learning model, and identifying an association mode of each type of data; in the data fusion process, noise and redundant data are eliminated; the acquired data are analyzed in real time through an anomaly detection module; according to geological simulation data and a deployment scheme of acquisition equipment, an acquisition path and a strategy are designed.

Description

Technical Field

[0001] The present invention relates to the technical field of mineral deposit data, and in particular to a method for optimizing the collection of digital mineral deposit information data. Background Art

[0002] The "digital information analysis and display system and method for mineral deposits" proposed by China's invention patent 2018114678634 is based on the mineralization zone division map of the National Geological Information Center and Baidu Maps. Through the digital information display system of typical mineral deposits, combined with geological and mineral maps, rock and ore specimen exploration and scientific research data, the mineral deposit information is integrated and displayed. However, the current mineral deposit information data acquisition optimization method has obvious deficiencies in dynamic data acquisition, real-time monitoring, system intelligence and linkage of information display. First, the static characteristics of the existing methods are more obvious. Although the traditional mineral deposit data acquisition method can cover some geological information, it mostly relies on historical data and static data display, and does not fully consider the dynamic changes of the geological environment. Compared with the methods in the comparative documents, the current digital acquisition is often limited to the acquisition of fixed sampling points in the mining area, and lacks the ability to capture dynamic geological phenomena in real time. For example, when geological activities, mine earthquakes or fault slips occur in the mining area, it is difficult for the existing acquisition method to adjust the arrangement and acquisition frequency of sensors in real time according to the changes. In the comparison document, the combination of Baidu Map and the data of the National Geological Archive can better display the mineralization belt information of the mining area, but this is still a relatively static information display, and the dynamic update and feedback of the collected data cannot be realized. Therefore, the traditional collection method, like the technology shown in the comparison document, lacks a dynamic response mechanism to the complex geological activities in the mining area and cannot adapt to the rapidly changing conditions within the mining area. Especially in the process of mining area development, the lack of real-time collection and rapid adjustment will affect the timeliness and reliability of the data.

[0003] Secondly, the lack of data processing and fusion is also a major problem in traditional methods. The comparative document mentioned that a digital data database was established by digitally integrating the existing geological maps, rock and mineral specimens and scientific research data in the ore deposit laboratory. However, this data integration still mainly relies on historical data and geological maps, and does not combine real-time sensor data for dynamic analysis. Existing acquisition methods usually face the problem of multimodal data fusion, especially for heterogeneous data collected by different types of sensors (such as seismic sensors, gravity sensors, electromagnetic sensors, etc.), lack of effective fusion and processing methods. In practical applications, the complexity and diversity of mining areas require that sensor data can be fused and processed at multiple levels, but existing acquisition optimization methods often only process a single data type and lack the ability to comprehensively analyze cross-modal data. This leads to a large gap in the collected data in space and time, and cannot accurately reflect the three-dimensional structure and dynamic changes of the ore deposit. In contrast, although the technology in the comparative document can display typical ore deposit information in different mining areas, it mainly focuses on the data display level, rather than real-time data fusion and analysis, and cannot provide effective support for real-time mining area monitoring.

[0004] Furthermore, the existing methods are also insufficient in terms of intelligence and automation. The method in the comparative document uses Baidu Maps and positioning modules to achieve automatic positioning and information display of the mining area, which to some extent enhances the user's interactive experience, but the current acquisition method still relies on manual intervention and control. Most of the existing acquisition optimization systems require manual setting of acquisition strategies, sensor layouts, and data processing processes, and lack adaptive and intelligent adjustment mechanisms. For example, when geological activities or mine earthquakes occur in the mining area, the existing system cannot automatically identify abnormal situations and quickly adjust the acquisition strategy. The acquisition frequency, density, and location of the sensor are still fixed and cannot be flexibly adjusted according to real-time monitoring data. This fixed acquisition strategy is very likely to cause data redundancy or miss key data, which has an adverse impact on the development of the mining area. Summary of the invention

[0005] The purpose of the present invention is to provide a method for optimizing the digital collection of mineral deposit information data, thereby solving some of the drawbacks and shortcomings pointed out in the background technology.

[0006] The present invention solves the above-mentioned technical problems by adopting the following technical solutions: a digital collection optimization method for mineral deposit information data, comprising: S1, dynamic adjustment of collection parameters based on geological evolution simulation:

[0007] S1.1. Based on the geological evolution history data of the mining area, a three-dimensional geological model is established to simulate the formation and changes of the ore deposit at different evolution stages;

[0008] S1.2. Dynamically simulate the key parameters of the ore deposit, including ore body thickness, ore grade, and mineral distribution;

[0009] S1.3. Adjust the acquisition parameters including sensor sensitivity, acquisition spacing, and acquisition density according to the prediction of the geological model;

[0010] S2. Dynamic adaptive sensor array and spatial optimization:

[0011] S2.1. Combined with the geological evolution simulation model, set the layout plan of the sensor array and adjust the density of the sensor array according to the geological conditions of the mining area;

[0012] S2.2, the sensor array adjusts the acquisition angle, orientation and sensing area in real time through a built-in adaptive algorithm;

[0013] S2.3. During data collection, the sensor array is dynamically optimized according to geological changes, including increasing sensor density when mining activity increases, or reducing the frequency of data collection in stable areas;

[0014] S3. Real-time fusion and anomaly detection of multimodal data driven by deep learning:

[0015] S3.1. Real-time synchronous collection of multimodal data from different sensors, feature extraction through deep learning models, and identification of correlation patterns among various types of data;

[0016] S3.2, in the process of data fusion, eliminate noise and redundant data;

[0017] S3.3, through the anomaly detection module, the collected data is analyzed in real time to identify the geological anomalies in the mining area, including faults, mine earthquakes, and sudden changes in geological structures, and the collection strategy is dynamically adjusted according to the anomalies;

[0018] S4. Iteration of collection path and collection strategy:

[0019] S4.1. In the initial stage, design the acquisition path and strategy based on the geological simulation data and the deployment plan of the acquisition equipment;

[0020] S4.2. During the acquisition process, the acquisition data and model feedback are received in real time based on the Bayesian optimization algorithm, and the operation path of the acquisition equipment is adjusted to cover the valuable mineral deposit area in a short time;

[0021] S4.3. After each data collection iteration, the collection strategy is updated using feedback data to identify the path and collection frequency.

[0022] Furthermore, the dynamic optimization layout method includes: dynamically adjusting sensor layout, collection density, and path planning according to feedback data collected in real time, and describing the changes of the three-dimensional structure of the ore deposit, mineral distribution, and thickness over time through the following integral equation, and combining adaptive weights for collection optimization:

[0023]

[0024] Among them, G(t) represents the dynamic change of the three-dimensional structure of the ore deposit, which is updated with the change of time t and is one of the key variables of the collected data, used to reflect the internal spatial characteristics of the ore deposit; D(t) represents the temporal change of mineral distribution, reflecting the migration or distribution change of minerals during the collection time period; M(t) is the temporal change of ore body thickness, which is predicted to change with the change of geological conditions; α, β, γ are weight factors in the adaptive learning algorithm, corresponding to the influence weights of three-dimensional structure, mineral distribution, and ore body thickness, respectively; t0 and t1 represent the start and end time periods of the collection time, and the changes in geological characteristics during the entire collection period are comprehensively processed through integral operations.

[0025] Furthermore, the dynamic optimization layout method includes: adopting a progressive multi-scale acquisition strategy, initially using sensors to scan the entire mining area to identify potential key areas; then dynamically adjusting the acquisition strategy, using sensors to perform detailed acquisition in key areas; and using a hierarchical progressive acquisition strategy to allocate acquisition density and resources to different areas through a priority function:

[0026]

[0027] Among them, f s (A) is the acquisition priority of a certain area, indicating the dynamic allocation of acquisition resources in the area; P h Indicates the thickness of the ore body, which is a key parameter in the area and is used to indicate the geological importance of the area during the mining process; P d Represents mineral density, indicating the abundance of mineral distribution in the area; P r It is the benchmark reference value of the area, which is used to standardize the influence of thickness and density; λ1 and λ2 are weight coefficients, which are used to adjust the influence of ore body thickness and mineral density in different acquisition areas; ∈ is the adaptive adjustment coefficient, which is dynamically adjusted according to the actual acquisition situation; δ is the accuracy adjustment factor, which is used to control the accuracy requirements during data acquisition.

[0028] Furthermore, the dynamic optimization layout method includes: introducing changes in physical fields including stress field, gravity field and magnetic field, combining geological characteristics, and dynamically monitoring the structure of the ore deposit; through the linkage analysis of geology and physical fields, dynamically adjusting the sensor layout and acquisition frequency, and monitoring and adjusting the acquisition strategy through the physical field linkage equation:

[0029]

[0030] Among them, S represents the change of stress field, which is used to monitor the pressure or dynamic change of stress on the ore deposit; G represents the change of gravity field, which reflects the abnormal change of gravity inside and outside the ore deposit, indicating the collapse of underground structure or the change of mineral distribution; M represents the change of magnetic field, which reflects the change of mineral composition or structure in the mining area; S0, G0, M0 are the reference values ​​of stress field, gravity field and magnetic field respectively, which are used to standardize the change amplitude of each physical field; η is the linkage coefficient, which is used to control the strength of correlation between geological and physical fields.

[0031] Furthermore, the method for constructing the anomaly detection module includes: collecting multimodal data of the mining area through various types of sensors, including seismic sensors, gravity sensors and electromagnetic induction sensors, which are responsible for collecting key data of seismic fluctuations, gravity changes, and electromagnetic field fluctuations respectively; using a data fusion algorithm to preprocess and denoise multi-source data to eliminate redundant data and noise; the data fusion uses a formula to integrate and preprocess multi-source data:

[0032] F fusion (x i ,y i ,z i ,t)=∫0 T (α·f1(x i ,t)+β·f2(y i ,t)+γ·f3(z i ,t))dt

[0033] Among them, x i represents the seismic fluctuation data collected by the sensor as a function of time t, reflecting the impact of seismic activity on the structure of the deposit; y i Represents gravity change data, records the gravity field anomaly in the mining area, and reflects the change of mineral density or structure through t; z i It is electromagnetic field change data, capturing the electromagnetic characteristics of different areas in the mining area, and is used to identify changes in mineral composition in the ore body; f1(x i ,t),f2(y i ,t),f3(z i , t) are acquisition functions, which are used to describe the changing trend of each sensor data over time t; α, β, γ are corresponding weight factors, which are used to adjust the importance of different types of data in the overall fusion and reflect the contribution of different sensor data to geological analysis; T represents the time range of data acquisition, and the integral calculation is used to summarize and fuse each data source within the time interval.

[0034] Furthermore, the method for constructing the anomaly detection module includes: establishing a model of normal geological activities by learning historical geological data; detecting abnormal phenomena occurring inside the mining area, including faults, mine tremors or sudden changes in geological structures, when the real-time collected data deviates significantly from the standard mode; identifying abnormal situations by comparing historical and real-time data; the anomaly detection module learns standard characteristic patterns under different geological conditions through an adaptive neural network, and uses a nonlinear analysis method to detect and compare the collected data in real time:

[0035]

[0036] Among them, D obs (t i ) represents the sensor at time t i The observed real-time data is the multimodal data collected at the current moment; D ref (t i ) represents the corresponding time t in the historical data i The reference data reflects the normal state of the geological structure; n is the number of samples collected in the time period, and the mean value is calculated through the data at multiple time points; What is calculated is the relative deviation between real-time observation data and historical reference data, which can quantify the degree of difference between current data and standard patterns; By taking the square root of the sum of squares, we can form the core metric for nonlinear anomaly detection, which is used to smooth the differences between different data.

[0037] Furthermore, the method for constructing the anomaly detection module includes: adjusting the data collection strategy according to the detected anomaly; the adjustment strategy includes dynamically increasing the collection frequency, adjusting the collection range, and dynamically deploying the sensor type;

[0038] When a mine earthquake signal is detected, the acquisition frequency of the sensors in the area is increased; when a regional gravity or magnetic field anomaly is detected, the sensors are redeployed to expand the acquisition range and capture the area affected by the anomaly prediction; the acquisition strategy is adjusted to dynamically manage the acquisition frequency and density through the formula:

[0039] f adjust (A,ω,t)=A·sin(ωt)+ΔA

[0040] Where A is the initial sensor acquisition density, which indicates the sensor distribution and acquisition frequency in a certain area under normal circumstances; ω is the acquisition frequency adjustment factor after an abnormal event is triggered, which is usually set according to the detected abnormal intensity; t represents time, and is used to capture the acquisition rhythm that changes with events in time when dynamically adjusting the acquisition frequency and density; sin(ωt) represents the fluctuation pattern of the acquisition frequency with time and abnormal detection, and ω controls the frequency of fluctuation; ΔA is the additional acquisition density that is dynamically increased according to the abnormal detection results, which increases the acquisition intensity in key areas or time periods.

[0041] Beneficial effects of the present invention:

[0042] 1. Improve collection accuracy and efficiency: By dynamically optimizing sensor layout and collection strategies, the system can adjust the collection density, path and frequency based on real-time feedback, thereby achieving accurate monitoring of key areas in the mining area, avoiding waste of resources and improving data collection efficiency.

[0043] 2. Comprehensive multi-dimensional data fusion: This method introduces multimodal sensors, including multiple data sources such as seismic, gravity and electromagnetic induction, and effectively integrates information through data fusion algorithms to form a more comprehensive geological feature of the deposit, ensure the effective use of different types of data, and improve the ability to capture the complexity and diversity of the deposit.

[0044] 3. Real-time anomaly detection and response: Through adaptive neural networks and nonlinear analysis methods, the system can identify anomalies in the mining area in real time, including faults, mine earthquakes and sudden changes in geological structures, and adjust the collection strategy in real time according to the detection results, so as to achieve efficient response and data supplementation and capture key data in a timely manner.

[0045] 4. Dynamic adaptive acquisition strategy: The system combines the linkage analysis of geology and physical fields, and can adaptively adjust the sensor layout and acquisition frequency according to the dynamic changes in the mining area, ensuring high-density data in key areas while reducing the acquisition frequency in stable areas, thereby achieving reasonable allocation and optimization of acquisition resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 The present invention is a flow chart of the method for optimizing the collection of digitalized mineral deposit information data.

[0047] Figure 2 This is a flowchart of the real-time fusion and anomaly detection of multimodal data driven by deep learning in the present invention.

[0048] Figure 3 This is a flow chart of the acquisition path and acquisition strategy iteration of the present invention.

[0049] Figure 4 This is a flow chart of the dynamic optimization layout method of the present invention.

[0050] Figure 5 The present invention is a flowchart of the method for constructing an anomaly detection module. DETAILED DESCRIPTION

[0051] The specific implementation modes of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0052] Combined with Figure 1 The digitalized acquisition optimization method of mineral deposit information data of the present invention aims to improve the efficiency and accuracy of mineral deposit data acquisition through scientific parameter adjustment and dynamic simulation. First, the acquisition parameters are dynamically adjusted based on geological evolution simulation. The primary task of this process is to establish a three-dimensional geological model with the help of the geological evolution history data of the mining area. This model not only reflects the spatial distribution of the mineral deposit, but also effectively simulates the formation and change of the mineral deposit at different evolution stages, helping to better understand the geological characteristics of the mining area. Next, the key parameters of the mineral deposit are dynamically simulated, including ore body thickness, ore grade and mineral distribution. The dynamic simulation of these key parameters is the basis, which can reveal the potential changes inside the mineral deposit, thereby providing an important reference for subsequent acquisition strategies. Finally, according to the prediction of the geological model, various acquisition parameters such as sensor sensitivity, acquisition spacing and acquisition density are adjusted to meet the needs of data acquisition at different evolution stages. For example, when the model predicts that the thickness of the ore body changes significantly, it is necessary to increase the sensitivity and acquisition frequency of the sensor to capture more detailed data; while in areas where the mineral deposit is stable, the acquisition spacing can be appropriately increased and the acquisition density can be reduced, thereby improving the acquisition efficiency and reducing unnecessary data redundancy.

[0053] The second step involves dynamic adaptive sensor array and spatial optimization, which aims to improve the effectiveness and accuracy of data collection through flexible sensor layout and real-time adjustment. First, the layout plan of the sensor array is set in combination with the geological evolution simulation model, and the density of the sensor array is adjusted according to the specific geological conditions of the mining area. This process ensures that the distribution of sensors can match the geological characteristics of the deposit, such as increasing the density of sensors in areas with thicker ore bodies or higher ore grades to obtain more detailed data. Next, the sensor array is equipped with a built-in adaptive algorithm that can adjust the acquisition angle, orientation and sensing area in real time. The sensors can work according to a fixed layout and flexibly adjust according to real-time data feedback, thereby maximizing the coverage and quality of data collection. Finally, during the data collection process, the sensor array dynamically optimizes the layout according to geological changes. For example, when the activity of the mining area increases, the system will actively increase the density of sensors to intensively collect more detailed geological information, while in stable areas, the acquisition frequency will be reduced in a timely manner, thereby reducing resource waste and optimizing data processing efficiency.

[0054] Combination Figure 2The third step involves real-time fusion and anomaly detection of multimodal data driven by deep learning, which aims to improve the integration and analysis capabilities of data through advanced algorithms in order to better understand the geological characteristics of the mining area. First, multimodal data from different sensors are collected in real time and synchronously. This process ensures that different types of data, such as seismic waves, gravity changes, and electromagnetic field data, can be integrated in the same time frame. Feature extraction through deep learning models can effectively identify the correlation patterns between different types of data, thereby revealing potential geological phenomena. For example, deep learning models can learn the complex relationship between seismic signals and geological structural changes. Then, during the data fusion process, noise and redundant data are removed. This link is the key to ensuring data quality. By removing irrelevant or repeated information, the accuracy and effectiveness of data analysis can be improved. Finally, the collected data is analyzed in real time through the anomaly detection module. The function of this module is to identify geological anomalies in the mining area, including faults, mine earthquakes, and sudden changes in geological structures. Anomaly detection can not only detect potential risks in a timely manner, but also dynamically adjust the acquisition strategy according to abnormal conditions to cope with complex changes in the mining area, such as increasing the data acquisition frequency or adjusting the sensor position when a mine earthquake is found.

[0055] Combination Figure 3 ,The fourth step involves the iteration of acquisition paths and acquisition strategies, which aims to improve the effectiveness and coverage of data by continuously optimizing the acquisition process. First, in the initial stage, the acquisition path and strategy are designed based on the geological simulation data and the deployment plan of the acquisition equipment. This process includes evaluating the geological characteristics of the mining area in order to formulate a reasonable acquisition plan to ensure that important areas in the ore deposit are covered first. Next, during the acquisition process, the acquisition data and model feedback are received in real time based on the Bayesian optimization algorithm. The key to this link is to use the ability of Bayesian optimization to dynamically adjust the operation path of the acquisition equipment to ensure that valuable ore deposit areas are covered to the greatest extent in a short time. The Bayesian optimization algorithm can intelligently predict which areas are most likely to contain rich mineral deposits based on the distribution of existing data, thereby guiding the acquisition equipment to adjust its direction and position. Finally, after each data acquisition iteration, the feedback data is used to update the acquisition strategy, identify the path and acquisition frequency, and further optimize the subsequent acquisition plan based on the actual acquisition results. For example, after a significant change in the thickness of the ore body is found in some areas, the future acquisition frequency needs to be adjusted to obtain more detailed data.

[0056] Embodiment 1:

[0057] See also Figure 4In this embodiment, the dynamic optimization layout method includes dynamically adjusting the sensor layout, collection density, and path planning according to the feedback data collected in real time, and describing the changes of the three-dimensional structure, mineral distribution, and thickness of the ore deposit over time through the following integral equation, and combining adaptive weights for collection optimization:

[0058]

[0059] In this formula, G(t) represents the dynamic change of the three-dimensional structure of the ore deposit, which is updated with the change of time t and is used to reflect the internal spatial characteristics of the ore deposit; D(t) represents the temporal change of mineral distribution, reflecting the migration or distribution change of minerals during the acquisition time period; M(t) is the temporal change of ore body thickness, which is predicted to change with the change of geological conditions; α, β, γ are weight factors in the adaptive learning algorithm, corresponding to the influence weights of three-dimensional structure, mineral distribution, and ore body thickness, respectively; t0 and t1 represent the start and end time periods of the acquisition time, and the changes in geological characteristics during the entire acquisition period are comprehensively processed through integral operations.

[0060] The three-dimensional structure G(t) of a mineral deposit in a certain mining area undergoes multiple stages of changes within a year. The initial thickness of the deposit is 15 meters, and it is expected to increase to 20 meters in the sixth month due to geological activities, while the mineral distribution D(t) shows uneven migration in different areas. The collected data show that the ore grade (such as metal content) increases from 5% to 8% within six months. Through the real-time feedback of the sensor, the weight factors are set to α=0.5, β=0.3, γ=0.2, which reflects that the importance of the three-dimensional structure to the overall data is greater than the mineral distribution and ore body thickness. The set acquisition time period is t0=0, t1=12 months.

[0061] During the data collection process, by dynamically adjusting the sensor layout and collection density, the system receives feedback data in real time, which enables the density of sensors to be increased in key areas. For example, in areas where mine earthquakes occur frequently, the density is increased to 5 sensors per square kilometer, which is usually 2. At the same time, the optimization of path planning allows each data collection to cover more valuable areas in a short time.

[0062] Next, an integral calculation is performed to evaluate the changes in geological characteristics during the entire acquisition period. Assuming that within the 12 months of acquisition, the changes in the three-dimensional structure G(t) are recorded as: G0=15, G6=20, G12=18, and the average change can be simplified to G(t)=15+0.5t (linear change). The change in mineral distribution is recorded as D(t) increasing linearly from 5% to 8%, which can be expressed as D(t)=5+0.25t. The thickness of the ore body M(t) can also be expressed linearly as M(t)=15+0.5t.

[0063] Using these formulas, we can substitute them into the main integral formula:

[0064] F(G,D,M,t)=∫0 12 (0.5·(15+0.5t)+0.3·(5+0.25t)+0.2·(15+0.5t))dt

[0065] Further simplifying, we get:

[0066] F(G,D,M,t)=∫0 12 (0.5·(15)+0.3·(5)+0.2·(15)+(0.25·0.3+0.5·0.5)t)dt

[0067] =∫0 12 (7.5+1.5t)dt

[0068] After calculation, we get:

[0069] =[7.5t+0.75t 2 ] 1 0 2 =7.5×12+0.75×12 2 =90+108=198

[0070] The final calculation result is 198, indicating that within this year, the amount of geological feature information successfully obtained is 198 through dynamic optimization of sensor layout and acquisition strategy.

[0071] Then, a progressive multi-scale acquisition strategy was adopted to make data collection more efficient and accurate. Initially, the entire mining area was extensively scanned by sensors to identify potential key areas. A preliminary scan of a certain mining area showed several areas with ore bodies exceeding 15 meters in thickness, and the ore grade fluctuated between 5% and 10%. At this stage, the acquisition team set up 5 sensors, evenly distributed in the mining area, to monitor the ore deposit information in real time. The scan results showed that some areas showed higher mineral density and thickness, and these areas were marked as key acquisition areas.

[0072] Next, dynamically adjust the collection strategy and enter the detailed collection stage. Use the hierarchical progressive collection strategy to allocate the collection density and resources in different areas through the priority function. Here, the priority function is set as:

[0073]

[0074] Among them, P h Indicates the thickness of the ore body, set in the key area A, P h is 20 meters; d represents the mineral density, which is set to 8 tons / cubic meter; P ris the reference value of the area, which is set to a combination of 10 meters and 6 tons / cubic meter. The weight coefficients are set to λ1=0.6 and λ2=0.4, the adaptive adjustment coefficient ∈ ranges from 1 to 3, and the precision adjustment factor δ ranges from 1 to 2. With these settings, the collection priority of area A can be calculated:

[0075]

[0076] Substitute into the calculation, first calculate each part:

[0077]

[0078] When ∈=2, we get:

[0079] =2·(1.733) 0.6667 ≈2·1.464≈2.928

[0080] Therefore, the acquisition priority of area A is 2.928, indicating that this area should be prioritized for more intensive data collection. According to the priority, the sensor density of this area is further increased to 8 sensors per square kilometer, and data is collected intensively.

[0081] After detailed collection, the system can analyze the data in real time and adjust the collection strategy through feedback. If subsequent data collection shows that the thickness of the ore body in the area has increased or the mineral distribution has changed, the system will adjust the collection density or path planning accordingly to ensure the effective use of resources and the accuracy of the data.

[0082] Finally, by introducing changes in physical fields including stress field, gravity field and magnetic field, combined with geological characteristics, the ore deposit structure is dynamically monitored to ensure the efficiency and accuracy of data collection. The stress field change (S) of a certain mining area is set to 300kPa, the gravity field change (G) is set to 0.04g, and the magnetic field change (M) is set to 20nT. The reference value So is set to 250kPa, Go is set to 0.02g, Mo is set to 15nT, and the linkage coefficient η ranges from 1 to 5, and η=2 is taken for calculation. In this case, the linkage equation can be expressed as:

[0083]

[0084] Substitute the specific values ​​into the formula:

[0085]

[0086] Calculate the parts:

[0087]

[0088] therefore,

[0089]

[0090] This indicates that the physical field of the current ore deposit is changing significantly, suggesting that intensive data collection is needed in this area. Combined with these results, the acquisition team immediately increased the density of sensor deployment to 10 per square kilometer in order to more effectively monitor the dynamic changes of the ore deposit.

[0091] In the process of dynamically adjusting the sensor layout and data collection frequency, the team will also monitor the data feedback in real time. If the frequency of changes in the stress field and gravity field is detected to increase, the system will further optimize the collection path and increase the frequency of data collection to ensure that valuable information is obtained at critical moments.

[0092] Embodiment 2:

[0093] Combination Figure 5 In this embodiment, the anomaly detection module construction method uses multiple types of sensors to collect multimodal data from the mining area by constructing an anomaly detection module to improve the accuracy of ore bed structure monitoring. It is assumed that ten seismic sensors, five gravity sensors and eight electromagnetic induction sensors are installed in a certain mining area. The seismic sensor is responsible for monitoring seismic fluctuations, the gravity sensor detects gravity changes in the mining area, and the electromagnetic sensor is used to capture electromagnetic field fluctuations. It is assumed that during the one-month monitoring process, the sensor recorded the following data: seismic fluctuation data (x i ) shows obvious abnormal values ​​in some periods, and the gravity change (y i ) data show that the gravity field in some areas suddenly fluctuates and the electromagnetic field changes (z i ) data indicate changes in electromagnetic characteristics within the mining area.

[0094] In order to process these multi-source data, the data fusion algorithm is first used to pre-process the collected data to remove redundant data and noise. According to the formula:

[0095] F fusion (x i ,y i ,z i ,t)=∫0 T (α·f1(x i ,t)+β·f2(y i ,t)+γ·f3(z i ,t))dt

[0096] Here, the time range T is set to 30 days, and α, β, and γ are 0.5, 0.3, and 0.2, respectively, indicating the weights of seismic fluctuations, gravity changes, and electromagnetic field changes in data fusion.

[0097] The change trend of earthquake wave data within 30 days is: f1(xi ,t)=2sin(0.1t)+3, the gravity change is: f2(y i ,t)=1.5t+1, the electromagnetic field changes to: f3(z i ,t)=0.5t 2 -2.

[0098] Substituting these functions into the formula, we can perform the integral calculation:

[0099] 1. Calculate the integral of f1:

[0100]

[0101] 2. Calculate the integral of f2:

[0102]

[0103] 3. Calculate the integral of f3:

[0104]

[0105] Substitute these results into the main formula for the final calculation:

[0106] F fusion =∫0 30 (0.5·110+0.3·705+0.2·4440)dt=0.5·110+0.3·705+0.2·4440

[0107] =55+211.5+888=1154.5

[0108] This fusion result (1154.5) provides a basis for subsequent anomaly detection, indicating that the geological activities in the mining area were relatively active during the monitoring period, and further analysis of the geological changes in the specific area is needed. When the fused data exceeds the preset threshold, the system will trigger the anomaly detection module, analyze the feedback from various sensors, and promptly identify potential geological risks, such as mine earthquakes or structural changes, and adjust the data collection strategy and frequency to ensure the timeliness and accuracy of the data.

[0109] The construction of the anomaly detection module relies on learning from historical geological data to build a model of normal geological activity. The research team in a certain mining area collected geological activity data for the past ten years and built a baseline model of normal geological activity. By comparing the real-time collected data with historical data, the team was able to identify abnormal phenomena within the mining area, such as faults, mine earthquakes, or sudden changes in geological structures.

[0110] To achieve this goal, the team used an adaptive neural network to learn standard characteristic patterns under different geological conditions. Suppose that during a certain period of time, the team collected the following data in real time: the observation data D at time t1 obs (t1) = 105, corresponding historical reference data D ref (t1) = 100; Observation data D at time t2 obs (t2) = 200, historical reference data is D ref (t2) = 150; Observation data D at time t3 obs (t3) = 180, historical reference data is D ref (t3)=190.

[0111] Anomaly detection using the formula:

[0112]

[0113] Here, b=3 (corresponding to three time points), and the calculation is as follows:

[0114] 1. For t1:

[0115]

[0116] Its square is 0.0025.

[0117] 2. For t2:

[0118]

[0119] Its square is 0.1111.

[0120] 3. For t3:

[0121]

[0122] Its square is 0.0028.

[0123] Substituting these results into the formula:

[0124]

[0125] The test result is 0.197. If the threshold is set to 0.15, this value exceeds the preset normal range, indicating that there are abnormal phenomena in the mining area during this time period.

[0126] After discovering anomalies, the team will further analyze the geological risks. For example, when anomalies of mine earthquakes are detected, the system will trigger an alarm and increase the frequency of acquisition in the area by rearranging sensors to obtain more detailed data and further identify the location of the fault and its activity.

[0127] When the anomaly detection module identifies a mine earthquake signal, the system will quickly adjust the data collection strategy to ensure that the dynamic information of geological changes is captured in a timely manner. For example, during a monitoring in a mining area, the sensor detected a mine earthquake signal with an amplitude of 1.5 times that of the historical data, indicating a high abnormal intensity. According to the preset response mechanism, the research team set the initial acquisition density of the sensor to A = 10 sensors, and the normal acquisition frequency of each sensor is once an hour.

[0128] In this case, the acquisition frequency adjustment factor ω after the abnormal event is triggered can be set to 2rad / h, that is, the acquisition frequency is increased by two times per hour. The team set ΔA to 5, indicating that the acquisition frequency of 5 sensors needs to be increased in the mine earthquake affected area. Through the formula:

[0129] f adjust (A,ω,t)=A·sin(ωt)+ΔA

[0130] At t = 0, calculate the acquisition frequency:

[0131] f adjust (10,2,0)=10·sin(2·0)+5=0+5=5

[0132] This means that at time t=0, after the mine tremor occurs, the area needs to increase the acquisition frequency by 5 times, making the total acquisition frequency 5 times / hour.

[0133] When time passes to t = 0.5 hours, the acquisition frequency becomes:

[0134] f adjust (10,2,0.5)=10·sin(2·0.5)+5=10·sin1+5≈10·0.8415+5≈8.415+5

[0135] =13.415

[0136] At this time, the system dynamically adjusts the collection frequency to about 13 times per hour, monitoring the area affected by mine earthquakes at a higher frequency.

[0137] If the monitoring is a regional gravity or magnetic field anomaly, such as a 2.5% fluctuation caused by a change in the gravity field, the team will redeploy the sensor to expand the collection range. Set the baseline gravity value to t0 = 9.81m / s 2 , calculate the gravity anomaly amplitude by the formula:

[0138] ΔG=0.025·G0=0.025·9.81≈0.24525m / s 2

[0139] This magnitude suggests that the team should expand the collection range in areas with significant gravity changes, increase the number of sensors from 10 to 15, and increase the monitoring frequency around the abnormal areas.

[0140] Through such dynamic adjustments, the research team can obtain key data in a timely manner when mine earthquakes and other abnormal events occur, so as to predict the risks in the mining area and ensure the safety and effectiveness of the mineral development process.

[0141] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A digital collection optimization method for mineral deposit information data, characterized in that The following steps are involved: S1. Dynamic adjustment of acquisition parameters based on geological evolution simulation: S1.

1. Based on the geological evolution history data of the mining area, a three-dimensional geological model is established to simulate the formation and changes of the ore deposit at different evolution stages; S1.

2. Dynamically simulate the key parameters of the ore deposit, including ore body thickness, ore grade, and mineral distribution; S1.

3. Adjust the acquisition parameters including sensor sensitivity, acquisition spacing, and acquisition density according to the prediction of the geological model; S2. Dynamic adaptive sensor array and spatial optimization: S2.

1. Combined with the geological evolution simulation model, set the layout plan of the sensor array and adjust the density of the sensor array according to the geological conditions of the mining area; S2.2, the sensor array adjusts the acquisition angle, orientation and sensing area in real time through a built-in adaptive algorithm; S2.

3. During data collection, the sensor array is dynamically optimized according to geological changes, including increasing sensor density when mining activity increases, or reducing the frequency of data collection in stable areas; S3. Real-time fusion and anomaly detection of multimodal data driven by deep learning: S3.

1. Real-time synchronous collection of multimodal data from different sensors, feature extraction through deep learning models, and identification of correlation patterns among various types of data; S3.2, in the process of data fusion, eliminate noise and redundant data; S3.3, through the anomaly detection module, the collected data is analyzed in real time to identify the geological anomalies in the mining area, including faults, mine earthquakes, and sudden changes in geological structures, and the collection strategy is dynamically adjusted according to the anomalies; S4. Iteration of collection path and collection strategy: S4.

1. In the initial stage, design the acquisition path and strategy based on the geological simulation data and the deployment plan of the acquisition equipment; S4.

2. During the acquisition process, the acquisition data and model feedback are received in real time based on the Bayesian optimization algorithm, and the operation path of the acquisition equipment is adjusted to cover the valuable mineral deposit area in a short time; S4.

3. After each data collection iteration, the collection strategy is updated using feedback data to identify the path and collection frequency.

2. The method for digitizing the mineral deposit information data according to claim 1 is characterized in that The dynamic optimization layout method includes: dynamically adjusting sensor layout, collection density, and path planning according to feedback data collected in real time, and describing the changes of the three-dimensional structure, mineral distribution, and thickness of the ore deposit over time through the following integral equation, and combining adaptive weights for collection optimization: Among them, G(t) represents the dynamic change of the three-dimensional structure of the ore deposit, which is updated with the change of time t and is one of the key variables of the collected data, used to reflect the internal spatial characteristics of the ore deposit; D(t) represents the temporal change of mineral distribution, reflecting the migration or distribution change of minerals during the collection time period; M(t) is the temporal change of ore body thickness, which is predicted to change with the change of geological conditions; α, β, γ are weight factors in the adaptive learning algorithm, corresponding to the influence weights of three-dimensional structure, mineral distribution, and ore body thickness, respectively; t0 and t1 represent the start and end time periods of the collection time, and the changes in geological characteristics during the entire collection period are comprehensively processed through integral operations.

3. The method for digitizing the collection and optimization of mineral deposit information data according to claim 2, characterized in that The dynamic optimization layout method includes: adopting a progressive multi-scale acquisition strategy, using sensors to scan the entire mining area in the early stage to identify potential key areas; then dynamically adjusting the acquisition strategy, using sensors to perform detailed acquisition in key areas; and using a hierarchical progressive acquisition strategy to allocate acquisition density and resources in different areas through a priority function: Among them, f s (A) is the acquisition priority of a certain area, indicating the dynamic allocation of acquisition resources in the area; P h Indicates the thickness of the ore body, which is a key parameter in the area and is used to indicate the geological importance of the area during the mining process; P d Represents mineral density, indicating the abundance of mineral distribution in the area; P r It is the benchmark reference value of the area, which is used to standardize the influence of thickness and density; λ1 and λ2 are weight coefficients, which are used to adjust the influence of ore body thickness and mineral density in different acquisition areas; ∈ is the adaptive adjustment coefficient, which is dynamically adjusted according to the actual acquisition situation; δ is the accuracy adjustment factor, which is used to control the accuracy requirements during data acquisition.

4. The method for digitizing the collection and optimization of mineral deposit information data according to claim 3 is characterized in that The dynamic optimization layout method includes: introducing changes in physical fields including stress fields, gravity fields and magnetic fields, combining geological characteristics, and dynamically monitoring the structure of the ore deposit; and dynamically adjusting the sensor layout and acquisition frequency by analyzing the geology and physical fields in conjunction.

5. The method for digitizing the collection and optimization of mineral deposit information data according to claim 1, characterized in that The method for constructing the anomaly detection module includes: collecting multimodal data of the mining area through various types of sensors, including seismic sensors, gravity sensors and electromagnetic induction sensors, which are responsible for collecting key data of seismic fluctuations, gravity changes, and electromagnetic field fluctuations respectively; using a data fusion algorithm to preprocess and denoise multi-source data to eliminate redundant data and noise.

6. The method for digitizing the collection and optimization of mineral deposit information data according to claim 5, characterized in that The method for constructing the anomaly detection module includes: establishing a model of normal geological activities by learning historical geological data; detecting abnormal phenomena occurring inside the mining area, including faults, mine tremors or sudden changes in geological structures, when the real-time collected data deviates significantly from the standard mode; identifying abnormal situations by comparing historical and real-time data; the anomaly detection module learns standard characteristic patterns under different geological conditions through an adaptive neural network, and uses a nonlinear analysis method to detect and compare the collected data in real time: Among them, D obs (t i ) represents the sensor at time t i The observed real-time data is the multimodal data collected at the current moment; D ref (t i ) represents the corresponding time t in the historical data i The reference data reflects the normal state of the geological structure; n is the number of samples collected in the time period, and the mean value is calculated through the data at multiple time points; What is calculated is the relative deviation between real-time observation data and historical reference data, which can quantify the degree of difference between current data and standard patterns; By taking the square root of the sum of squares, we can form the core metric for nonlinear anomaly detection, which is used to smooth the differences between different data.

7. The method for digitizing the collection and optimization of mineral deposit information data according to claim 6, characterized in that The method for constructing the anomaly detection module includes: adjusting the data collection strategy according to the detected anomaly; the adjustment strategy includes dynamically increasing the collection frequency, adjusting the collection range, and dynamically deploying the sensor type; When a mine earthquake signal is detected, the acquisition frequency of the sensors in the area is increased; when a regional gravity or magnetic field anomaly is detected, the sensors are redeployed to expand the acquisition range and capture the area affected by the anomaly prediction.

Images