Health management method and system of coal mine dry separation system based on multi-source data fusion

Abstract

The application discloses a coal mine dry separation system health management method and system based on multi-source data fusion, and relates to the technical field of equipment health management. The method comprises the following steps: obtaining equipment operation parameters of multiple devices of a coal mine dry separation system, and obtaining working condition parameters; based on the equipment operation parameters and the working condition parameters, an equipment state recognition intelligent agent is constructed to recognize the health state of the equipment and obtain multiple equipment health parameters; based on the sequence of the coal mine dry separation system process, process influence parameters of the multiple devices are obtained, and in combination with the physical connection conditions and historical health parameters of the equipment, the equipment health parameters are evaluated for credibility to obtain equipment health recognition results; in combination with the equipment health recognition results and the working condition parameters, a coal mine dry separation system maintenance scheme is obtained to perform coal mine dry separation system health management. The application solves the technical problem that the health management effect is poor due to the neglect of the collaborative influence between the working conditions and the equipment in the coal mine dry separation process in the prior art.

Description

Technical Field

[0001] This invention relates to the field of equipment condition monitoring technology, specifically to a health management method and system for coal mine dry separation systems based on multi-source data fusion. Background Technology

[0002] As a core component for improving the quality and efficiency of raw coal and intelligently separating gangue, the continuous and stable operation of coal mine dry separation systems is crucial for ensuring mine production efficiency and economic benefits. To ensure system reliability, traditional equipment health management mainly relies on threshold monitoring of key operating parameters of individual devices or simple model analysis based on historical fault data. However, these existing methods have significant limitations in practical applications. Their health status assessments are often conducted in isolation, focusing only on the equipment's own sensor data and ignoring the highly dynamic operating conditions and inter-equipment synergies during coal mine dry separation, resulting in poor management of coal mine dry separation systems. Summary of the Invention

[0003] This application provides a health management method and system for coal mine dry separation systems based on multi-source data fusion, which is used to address the technical problem of poor health management results caused by neglecting the influence of working conditions and equipment coordination during the coal mine dry separation process in the prior art.

[0004] In view of the above problems, this application provides a method and system for health management of coal mine dry separation system based on multi-source data fusion.

[0005] In a first aspect, this application provides a health management method for a coal mine dry separation system based on multi-source data fusion, the method comprising: Obtain the equipment operating parameters of multiple devices in the coal mine dry separation system, and obtain the working condition parameters; Based on the equipment operating parameters and the working condition parameters, an intelligent agent for equipment status recognition is constructed to identify the health status of the equipment and obtain multiple equipment health parameters. Based on the process sequence of the coal mine dry separation system, process influence parameters of multiple devices are obtained, and combined with the physical connection status and historical health parameters of the devices, the health parameters of the devices are reliably evaluated to obtain the device health identification results. Based on the equipment health identification results and the operating parameters, a maintenance plan for the coal mine dry separation system is obtained, and health management of the coal mine dry separation system is carried out.

[0006] Secondly, this application provides a health management system for coal mine dry separation systems based on multi-source data fusion, including: The equipment operation status acquisition module is used to acquire the equipment operation parameters of multiple devices in the coal mine dry separation system, and to acquire the working condition parameters; The health status recognition module is used to construct an intelligent agent for equipment status recognition based on the equipment operating parameters and the working condition parameters, to perform equipment health status recognition, and to obtain multiple equipment health parameters. The identification result evaluation module is used to obtain process influence parameters of multiple devices based on the process sequence of the coal mine dry separation system, and to conduct a reliable evaluation of the device health parameters by combining the physical connection status and historical health parameters of the devices, thereby obtaining the device health identification result. The maintenance plan acquisition module is used to acquire a maintenance plan for the coal mine dry separation system by combining the equipment health identification results and the operating parameters, and to carry out health management of the coal mine dry separation system.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application proposes a health management method and system for coal mine dry separation systems based on multi-source data fusion. By systematically acquiring and fusing equipment operating parameters with working condition parameters reflecting material characteristics, and constructing an intelligent agent for equipment status recognition capable of learning the complex mapping relationship between the two, a more fundamental and accurate characterization of equipment health status is achieved, significantly improving the pertinence and reliability of health status recognition. Compared with traditional methods that rely solely on equipment threshold monitoring, the technical solution provided in this application significantly overcomes the problem of inaccurate early warnings caused by ignoring changes in actual production loads and the coupling effects of equipment, achieving the technical effects of improving the overall operational reliability of coal mine dry separation systems and optimizing the efficiency of maintenance resource allocation. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart illustrating the health management method for a coal mine dry separation system based on multi-source data fusion, as provided in an embodiment of this application.

[0010] Figure 2 A schematic diagram of the structure of a coal mine dry separation system health management system based on multi-source data fusion provided in this application embodiment.

[0011] The components represented by each number in the attached diagram are explained below: Equipment operation status acquisition module 100, health status recognition module 200, recognition result evaluation module 300, and maintenance plan acquisition module 400. Detailed Implementation

[0012] This application provides a health management method and system for coal mine dry separation systems based on multi-source data fusion, which addresses the technical problem of poor health management results caused by neglecting the influence of working conditions and inter-equipment coordination during the coal mine dry separation process in existing technologies.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, this application provides a health management method for a coal mine dry separation system based on multi-source data fusion, wherein the method includes: S10: Obtain the equipment operating parameters of multiple devices in the coal mine dry separation system, and obtain the working condition parameters.

[0016] In the health management of dry coal separation systems, traditional monitoring methods mainly rely on collecting the operating parameters of individual equipment, such as vibration, temperature, and current, and making isolated health judgments based on these. However, the actual load and wear state of the dry separation system are highly dependent on the real-time characteristics of the materials it processes, especially the particle size distribution and instantaneous throughput of coal gangue. These operating parameters directly determine the mechanical stress, thermal load, and wear rate of the equipment.

[0017] Step S10 in the method provided in this application embodiment includes: Acquire multiple devices from a coal mine dry separation system, and obtain the operating parameters of these devices based on their characteristics; Based on the current dry coal separation material, operating parameters are obtained, including material particle size distribution and instantaneous material flow rate.

[0018] In this embodiment of the application, the equipment operating parameters of multiple devices in the coal mine dry separation system are obtained, as well as the working condition parameters.

[0019] Specifically, first, the key equipment in the coal mine dry separation system is identified, including components such as the feeder, crusher, separator, and conveyor. Further, operating parameters are collected based on the characteristics of each piece of equipment. For example, the vibration velocity and operating current of the feeder are collected; the vibration acceleration and bearing temperature of the crusher are collected; the frequency of the separator's actions and operating current are collected; and the bearing temperature and motor current of the conveyor are collected. For instance, these parameters are read and cached at a sampling frequency of once per second, and finally integrated into a structured dataset of equipment operating parameters.

[0020] Furthermore, based on the current dry coal separation material, operating parameters are obtained, including material particle size distribution and instantaneous material flow rate. For example, images of falling material are captured above the transfer point of the conveyor belt at a rate of 5 frames per second. First, the images are preprocessed by grayscale conversion and median filtering to reduce noise and edge detection is performed. This further identifies the contours of each material in the image, and the equivalent particle size represented by each contour is calculated. Finally, the equivalent particle size of all materials in the current frame is statistically analyzed and divided into multiple particle size intervals, such as less than 50 mm, 50-100 mm, 100-150 mm, and greater than 150 mm. The proportion of the total pixel area of ​​the material in each interval to the total pixel area of ​​the material in the entire frame is calculated, and this is used as the material particle size distribution at the current moment. The instantaneous material flow rate is obtained by differential calculation of the amount of material passing through in the previous second, in tons per hour. Ultimately, the material particle size distribution and the instantaneous material flow rate are used together as operating parameters.

[0021] By synchronously acquiring equipment operating parameters and working condition parameters including material particle size distribution and total quantity, this application lays a comprehensive and accurate data foundation for the entire health management method.

[0022] S20: Based on the equipment operating parameters and the working condition parameters, construct an intelligent agent for equipment status recognition, perform equipment health status recognition, and obtain multiple equipment health parameters.

[0023] After obtaining multi-source data on equipment operating parameters and condition parameters, accurately identifying the health status of each piece of equipment becomes a core challenge. Traditional methods often employ alarm rules based on fixed thresholds or simple statistical analysis models. These methods struggle to characterize the nonlinear dynamic relationship between equipment operating parameters, complex and ever-changing condition parameters, and equipment health status.

[0024] Step S20 in the method provided in this application embodiment includes: Construct a device status recognition intelligent agent, wherein the device status recognition intelligent agent includes multiple recognition branches; The construction of an intelligent agent for device status recognition includes: Based on multiple devices in a coal mine dry separation system, construct multiple intelligent identification branches for device status; Based on multiple devices, obtain multiple historical device operating parameters and historical operating condition parameters, and obtain the corresponding historical device health parameters and device tags to form a historical sample set of devices; The device status intelligent recognition branches are trained separately using the device historical sample set until convergence, and then integrated to obtain the device status recognition intelligent agent. The device's operating parameters and working condition parameters are input into the device status recognition intelligent agent to identify and obtain multiple device health parameters.

[0025] In this embodiment of the application, an intelligent agent for equipment status recognition is constructed based on the equipment operating parameters and the operating condition parameters to identify the health status of the equipment and obtain multiple equipment health parameters.

[0026] Specifically, firstly, a device status recognition intelligent agent is constructed, wherein the device status recognition intelligent agent includes multiple recognition branches.

[0027] Based on multiple devices in a coal mine dry separation system, multiple intelligent identification branches for device status are constructed. In this embodiment, the device status identification agent includes multiple identification branches, each corresponding to a specific device in the system, such as a feeder branch, crusher branch, separator branch, and conveyor branch. Each intelligent identification branch for device status is an independent artificial neural network model. For example, it is constructed with a three-layer structure. Taking the crusher branch as an example, the number of nodes in the input layer is set according to the total number of all input features of the crusher, including its own equipment operating parameters and related working condition parameters. The hidden layer has 32 nodes and uses the ReLU function as the activation function. The output layer has one node, using the Sigmoid activation function to output a continuous value between 0 and 1, representing the health score of the crusher. A value closer to 1 indicates better health, while a value closer to 0 indicates a higher risk of failure. The identification branch structure for other devices is similar.

[0028] Furthermore, based on multiple devices, multiple historical device operating parameters and historical working condition parameters are acquired, along with corresponding historical device health parameters and device tags, forming a historical sample set of devices. For example, historical device operating parameters are exported from the data acquisition system based on device tags, and historical working condition parameters are acquired simultaneously. Corresponding historical device health parameters are generated for each device at each historical time point. Here, device tags refer to the type of equipment, such as a sorting machine or crusher. Historical device health parameters serve as the true values ​​for health scores, used as supervised training targets, and are obtained through manual analysis and annotation of historical maintenance records, inspection records, and operating logs. For instance, if historical records show that a crusher underwent preventative maintenance at a certain time, the historical device health parameters for the period prior to maintenance can be set to a lower value, such as 0.3; if records show that the equipment was in a fault-free and stable operating period, the corresponding historical device health parameters can be set to a higher value, such as 0.95; if records show that the equipment malfunctioned and was shut down for maintenance, the historical device health parameters corresponding to the historical device operating parameters at the time of the malfunction are set to 0. The historical operating parameters, historical working condition parameters, and corresponding historical equipment health parameters of the equipment at the same timestamp are integrated to form a training sample, and multiple training samples are integrated to obtain the equipment historical sample set.

[0029] Furthermore, the multiple intelligent identification branches for equipment status are trained separately using the historical sample set of the equipment until convergence, and an integrated intelligent agent for equipment status identification is obtained. For example, training uses a standard supervised learning process. Taking the crusher branch as an example, the historical sample set of the crusher is randomly divided into a training set and a validation set. During training, mean squared error is used as the loss function to measure the difference between the health score predicted by the model and the historical equipment health parameters. The Adam optimizer is selected to iteratively update the weight parameters within the model. The training process continues for multiple rounds, each round traversing the entire training set and calculating the loss. Simultaneously, after each round, the model performance is evaluated using the validation set, and the change in the validation set loss is observed. When the validation set loss no longer decreases significantly over multiple consecutive rounds, training stops, and the model is considered to have converged. After training, the final model weights for the crusher branch are saved. This training process is repeated for all equipment identification branches until all branch models are trained and converged.

[0030] Furthermore, multiple device operating parameters and operating condition parameters are input into the device status recognition intelligent agent to identify and obtain multiple device health parameters. The current operating parameters and operating condition parameters of multiple devices are input into the device recognition intelligent agent to identify and obtain the individual device health parameters of each device.

[0031] By constructing an intelligent agent for equipment status recognition, it is possible to learn and establish a high-order, nonlinear mapping model between equipment operating parameters, multi-dimensional operating condition parameters and equipment health status, and to output more refined and accurate equipment health parameters to characterize the equipment health status.

[0032] S30: Based on the process sequence of the coal mine dry separation system, obtain the process influence parameters of multiple devices, and combine the physical connection status and historical health parameters of the devices to perform a reliable assessment of the device health parameters and obtain the device health identification results.

[0033] In a coal mine dry separation system, which consists of multiple devices tightly connected according to specific processes, the health assessment of any single point may contain uncertainties or sporadic errors. Existing health management methods often directly use the assessment results of these isolated devices as the final conclusion, ignoring the inherent interconnectedness of the system.

[0034] Step S30 in the method provided in this application embodiment includes: A random combination of devices is selected, wherein the combination of devices includes two devices; Based on the process sequence of the coal mine dry separation system, the basic process influence parameters of the equipment combination are obtained. Obtain historical health parameters of the device combination and evaluate the obtained health consistency parameters; Obtain the physical connection status and physical distance of the device combination, and obtain physical impact parameters based on the physical connection status and physical distance; Based on the basic process impact parameters, the health consistency parameters, and the physical impact parameters, a reliable correlation result is obtained. Then, a new set of equipment combinations is obtained and evaluated to obtain a reliable correlation result, until the evaluation of all equipment combinations is completed. Based on multiple trusted association results of multiple device combinations of a single device, the device health parameters are corrected to obtain device health identification results; Specifically, based on multiple trusted association results of multiple device combinations for a single device, a trusted assessment of the device health parameters is performed to obtain device health identification results, including: Randomly select a device as the first device; Using each of the trusted association results of the first device as a weight, the current device health parameters of another device in the device combination including the first device are weighted and calculated to generate an association correction factor for the first device. Specifically, each trusted association result of the first device is used as a weight to perform a weighted calculation on the current device health parameters of another device in the device combination including the first device, generating an association correction factor for the first device, including: Randomly select a first device combination that includes the first device, obtain the trusted association result of the first device combination, and normalize to obtain the association weight; Obtain the device health parameters of the other device in the first device combination; Multiply the association weight by the device health parameter of the other device to obtain an association correction factor for the first device based on the combination of the first devices; A reliable assessment is performed on all the associated correction factors of the first device, and the device health identification result is obtained by combining the device health parameters of the first device.

[0035] In this embodiment of the application, based on the process sequence of the coal mine dry separation system, process influence parameters of multiple devices are obtained, and combined with the physical connection status and historical health parameters of the devices, the health parameters of the devices are reliably evaluated to obtain the device health identification results.

[0036] Specifically, a set of equipment combinations is randomly selected, wherein the equipment combination includes two pieces of equipment. For example, a set of equipment combinations is selected according to the process sequence, and each set contains two different pieces of equipment, such as a feeder and a crusher.

[0037] Furthermore, based on the process sequence of the coal mine dry separation system, the basic process influence parameters of the equipment combination are obtained. For example, for an equipment combination (equipment A, equipment B), the rules for determining the basic process influence parameters are as follows: If equipment A is upstream of equipment B, for example, the crusher is upstream of the separator, then the process influence parameter of equipment A on equipment B is set to 0.8, indicating that a failure of A can directly affect B; if equipment B is upstream of equipment A, then this parameter is set to 0.4; if there is no direct process sequence relationship between the two, for example, two parallel conveyors, then this parameter is set to 0.2. Specific parameters can be set based on the actual scenario.

[0038] Further, historical health parameters of the device combination are obtained, and a health consistency parameter is evaluated. For example, the historical health score sequence of the device combination over the past 30 days is retrieved from the historical database. The Pearson correlation coefficient between these two historical health score sequences is calculated and normalized to map the correlation coefficient to a range of 0 to 1. The health consistency parameter is calculated as (Pearson correlation coefficient + 1) / 2. The higher this parameter, the more synchronized the historical health status trends of the two devices are.

[0039] Further, the physical connection status and physical distance of the equipment combination are obtained, and physical influence parameters are obtained based on the physical connection status and physical distance. Specifically, the physical connection status refers to whether there is a direct material transfer channel between the equipment, such as direct connection via belt or chute. The installation center position coordinates (X, Y) of each piece of equipment in the two-dimensional plane coordinate system are read. For any equipment combination (e.g., equipment A and equipment B), the system calculates its physical straight-line distance. For example, the crusher and the separator are closer on the drawing, and the calculated distance value is smaller; while the feeder and the end conveyor are farther apart, and the calculated distance value is larger. Further, the physical connection status between the equipment is read from the same layout drawing. It is found that "the downstream of the crusher is directly connected to the separator," while there is no direct connection record between the "feeder" and the "sorter," and they need to be transferred through the crusher. Further, physical influence parameters are calculated based on the physical connection status and physical distance. For example, define a basic connection coefficient: if equipment A and equipment B are directly connected in the relationship table, the basic connection coefficient is set to 1; if they are indirectly connected, i.e., need to be transferred through other equipment, then for each transferred equipment, the basic connection coefficient is multiplied by an attenuation coefficient less than 1, such as 0.6. For example, if the feeder to the sorter needs to be transferred through the crusher once, then its basic connection coefficient = 1 × 0.6 = 0.6. Further, set a baseline distance, such as based on the average distance between equipment. Attenuation factor = baseline distance / (baseline distance + actual distance). That is, the greater the distance, the smaller the attenuation factor. Finally, calculate the physical influence parameter = basic connection coefficient × attenuation factor. This result is a value between 0 and 1. For directly connected and close-range equipment combinations, the mutual influence may be greater, and the parameter value will be close to 1.0; for indirectly connected or far-range combinations, the parameter value will be lower.

[0040] Furthermore, by comprehensively considering the basic process impact parameters, the health consistency parameters, and the physical impact parameters, a credible correlation result is obtained. A new set of equipment combinations is then acquired and evaluated to obtain credible correlation results, until the evaluation of all equipment combinations is completed. For example, a weighted calculation is performed on the basic process impact parameters, health consistency parameters, and physical impact parameters to obtain a credible correlation result: credible correlation result = first weight × basic process impact parameter + second weight × health consistency parameter + third weight × physical impact parameter. Wherein, first weight + second weight + third weight = 1. The specific weights can be set based on the importance of the basic process impact parameters, health consistency parameters, and physical impact parameters. For example, in the dry coal separation process, if the failure of upstream equipment easily leads to the failure of downstream equipment, a larger third weight is set to increase the weight of the physical impact parameter. Further, a new set of equipment combinations is acquired, and the same method is used to evaluate and obtain credible correlation results, until the evaluation of all equipment combinations is completed.

[0041] Furthermore, based on multiple trusted association results of multiple device combinations of a single device, the device health parameters are corrected to obtain device health identification results.

[0042] Specifically, first, a device is randomly selected as the first device.

[0043] Furthermore, each of the trusted association results of the first device is used as a weight to perform a weighted calculation on the current device health parameters of another device in the device combination including the first device, thereby generating an association correction factor for the first device.

[0044] First, a first device combination containing the first device is randomly selected, and the trusted association result of the first device combination is obtained. The association weight is then normalized. Specifically, based on the sum of trusted association results for all device combinations containing the first device, the trusted association result of the first device combination is normalized to obtain the association weight. Association weight = Trusted association result of the first device combination / Sum of trusted association results for all device combinations containing the first device.

[0045] Furthermore, the device health parameters of the other device in the first device combination are obtained.

[0046] Furthermore, the association weight is multiplied by the device health parameter of the other device to obtain an association correction factor for the first device based on the combination of the first devices. Association correction factor = association weight × device health parameter of the other device. The obtained association correction factor reflects the reliability of the health status of the first device obtained based on the health status of the other device and the association weight between the other device and the first device.

[0047] Further, a reliability assessment is performed on all the correlation correction factors of the first device, and the device health parameters of the first device are combined to obtain the device health identification result. For example, the average correlation correction value is obtained by averaging all the correlation correction factors of the first device. The average correlation correction value and the device health parameters of the first device are then weighted to obtain the device health identification result. For example, the initial scoring weight is set to 60%, and the average correction value weight is set to 40%, with the specific proportions adjustable based on the actual scenario. Then, the final device health identification result = device health parameters × 0.6 + average correlation correction value × 0.4. The obtained device health identification result is the final device health identification result after system correlation cross-validation and correction.

[0048] By conducting a reliable assessment of the initially identified equipment health parameters, the overall reliability and robustness of health status assessment are improved. The method provided in this application no longer treats equipment as isolated information silos, but fully utilizes the system's own process sequence, physical connections, and historical operational continuity to construct a multi-dimensional cross-validation network. The final output equipment health identification result is a more reliable comprehensive status judgment that has undergone systematic logical verification and correction, providing a solid and reliable core basis for generating accurate maintenance plans.

[0049] S40: Based on the equipment health identification results and the operating parameters, obtain a maintenance plan for the coal mine dry separation system and perform health management of the coal mine dry separation system.

[0050] Traditional maintenance decisions are often based on fixed time cycles, simple equipment health sequencing, or manual experience, lacking dynamism and foresight. This is especially true for dry coal separation systems, whose maintenance needs are closely related to production conditions.

[0051] Step S40 in the method provided in this application embodiment includes: Based on the device health identification results, historical health maintenance plans are obtained, and all devices are sorted by health status to generate a health status list. Based on the historical health maintenance plans and current operating conditions, a recommended maintenance plan is obtained; Based on the assessment of the wear acceleration impact of the aforementioned operating parameters, the health status list is adjusted to obtain a maintenance priority list; Specifically, based on the assessment of the wear acceleration impact of the aforementioned operating parameters, the health status list is adjusted to obtain a maintenance priority list, including: Calculate the percentage of large-diameter materials based on the particle size distribution. The wear coefficient under working conditions is calculated by comparing the large particle size material ratio parameter with the benchmark large particle size material ratio parameter. The wear coefficient under the working condition is used as an adjustment factor, and combined with the health status score of the corresponding equipment in the health status list for fusion calculation to obtain the updated health identification result. The updated health identification result is then reordered to generate the maintenance priority list. Based on the proposed maintenance plan and the maintenance priority list, a maintenance plan for the coal mine dry separation system is generated to carry out health management of the coal mine dry separation system.

[0052] In this embodiment of the application, a maintenance plan for the coal mine dry separation system is obtained by combining the equipment health identification results and the operating parameters, and the health management of the coal mine dry separation system is carried out.

[0053] Specifically, firstly, based on the equipment health identification results, historical health maintenance plans are obtained, and all equipment is sorted by health status to generate a health status list. Then, the equipment health identification results for each device at the current moment are obtained, and all devices are arranged in ascending order of health score to generate an ordered equipment list, i.e., the health status list. For example, the device at the top of the sorted list might be the crusher with the lowest health score, followed by the feeder with the second lowest score. Multiple historical health maintenance plans are obtained for each device, such as matching equipment type, main fault characteristics like vibration type, etc., to obtain the specific maintenance measures taken in historical cases, the required man-hours, spare parts lists, and records of the effects after treatment. This information together constitutes a basic historical health maintenance plan template.

[0054] Furthermore, based on the historical health maintenance plan and the current operating conditions, a recommended maintenance plan is obtained. For example, the currently collected operating condition parameters are compared with the estimated time required for the maintenance operation. If the current material flow rate remains high, it is recommended to postpone the maintenance operation to a preset maintenance period with lower system load, and this recommended time window will be noted in the recommended maintenance plan. Simultaneously, this time recommendation will be combined with the maintenance measures and spare parts list extracted from historical plans to form a preliminary draft maintenance plan with suggested execution times.

[0055] Furthermore, based on the assessment of the wear acceleration effect of the operating parameters, the health status list is adjusted to obtain a maintenance priority list.

[0056] Specifically, firstly, based on the particle size distribution of the material, the percentage of large-diameter materials is calculated. For example, if the total percentage of materials larger than 100 mm is 25%, then this parameter is 0.25.

[0057] Furthermore, the wear coefficient under operating conditions is calculated by comparing the large-particle-size material proportion parameter with the benchmark large-particle-size material proportion parameter. For example, the benchmark large-particle-size material proportion parameter is set based on the historical average value of the large-particle-size material proportion parameter during long-term stable system operation, and the wear coefficient under operating conditions is calculated as: large-particle-size material proportion parameter / benchmark large-particle-size material proportion parameter. For example, the wear coefficient under operating conditions is 0.25 / 0.15 = 1.67. This coefficient is greater than 1, indicating that the wear pressure on the equipment under current operating conditions is greater than the historical average level.

[0058] Furthermore, the wear coefficient is used as an adjustment factor, and combined with the health status score of the corresponding equipment in the health status list for fusion calculation to obtain an updated health identification result. This updated health identification result is then used to re-rank the equipment and generate the maintenance priority list. For example, equipment susceptible to impact wear from large materials, such as crushers and sorting machine screens, is identified in the health status list. This identification is based on historical health maintenance plans, such as equipment with high maintenance frequency due to impact wear from large materials. The equipment health identification results of these devices are then fused with the calculated wear coefficient, resulting in an updated health identification result of: original health score × wear coefficient. In other words, under severe operating conditions, the unhealthiness of these devices in the ranking is amplified. For example, a crusher with an original score of 0.6, multiplied by a wear coefficient of 1.67, will have an updated health identification result of 1 used for ranking, significantly increasing its urgency in the list.

[0059] Furthermore, based on the suggested maintenance plan and the maintenance priority list, a maintenance plan for the coal mine dry separation system is generated to conduct health management of the coal mine dry separation system. For example, the maintenance priority list is used as the primary basis for decision-making, and the specific measures in the suggested maintenance plan are bound and integrated with the equipment order in the priority list. For instance, a final plan document is generated, clearly listing the equipment sequence in descending order of urgency at the top of the document, and providing detailed maintenance steps extracted from historical plans, a list of required spare parts, and the system's recommended time window for the top-priority equipment. This plan is pushed to maintenance management personnel through a human-machine interface such as an industrial computer monitor or a mobile terminal app, guiding them to perform targeted and timely maintenance activities, thereby achieving proactive health management of the coal mine dry separation system.

[0060] By fusing reliable equipment health identification results with real-time operating parameters, this application achieves a leap from static response to dynamic optimization in maintenance plan generation. The resulting maintenance plan for the coal mine dry separation system is highly contextualized and forward-looking. The method provided in this application not only prioritizes equipment based on its absolute health status, but more importantly, it dynamically adjusts the urgency and priority of maintenance by analyzing current material characteristics in real time and assessing the additional wear pressure or performance impact coefficient of these operating conditions on specific equipment. Ultimately, this method outputs an optimized maintenance plan that integrates maintenance objects, priorities, and recommended timing, significantly improving the scientific, economic, and systematic nature of maintenance activities, and effectively ensuring the continuous and stable operation of the dry separation system under high load and variable operating conditions.

[0061] Example 2, as Figure 2As shown, based on the same inventive concept as the coal mine dry separation system health management method with multi-source data fusion provided in Embodiment 1, this embodiment of the invention also provides a coal mine dry separation system health management system with multi-source data fusion, including: The equipment operation status acquisition module 100 is used to acquire the equipment operation parameters of multiple devices in the coal mine dry separation system and to acquire the working condition parameters. The health status recognition module 200 is used to construct an intelligent agent for equipment status recognition based on the equipment operating parameters and the working condition parameters, to perform equipment health status recognition, and to obtain multiple equipment health parameters. The identification result evaluation module 300 is used to obtain process influence parameters of multiple devices based on the process sequence of the coal mine dry separation system, and to conduct a reliable evaluation of the device health parameters in combination with the physical connection status and historical health parameters of the devices, so as to obtain the device health identification result. The maintenance plan acquisition module 400 is used to acquire the maintenance plan of the coal mine dry separation system by combining the equipment health identification results and the operating parameters, and to carry out health management of the coal mine dry separation system.

[0062] In one embodiment, the device operating status acquisition module 100 is further configured to: Acquire multiple devices from a coal mine dry separation system, and obtain the operating parameters of these devices based on their characteristics; Based on the current dry coal separation material, operating parameters are obtained, including material particle size distribution and instantaneous material flow rate.

[0063] In one embodiment, the health status recognition module 200 is further configured to: Construct a device status recognition intelligent agent, wherein the device status recognition intelligent agent includes multiple recognition branches; The construction of an intelligent agent for device status recognition includes: Based on multiple devices in a coal mine dry separation system, construct multiple intelligent identification branches for device status; Based on multiple devices, obtain multiple historical device operating parameters and historical operating condition parameters, and obtain the corresponding historical device health parameters and device tags to form a historical sample set of devices; The device status intelligent recognition branches are trained separately using the device historical sample set until convergence, and then integrated to obtain the device status recognition intelligent agent. The device's operating parameters and working condition parameters are input into the device status recognition intelligent agent to identify and obtain multiple device health parameters.

[0064] In one embodiment, the real-time analysis and computing module 300 is further configured to: A random combination of devices is selected, wherein the combination of devices includes two devices; Based on the process sequence of the coal mine dry separation system, the basic process influence parameters of the equipment combination are obtained. Obtain historical health parameters of the device combination and evaluate the obtained health consistency parameters; Obtain the physical connection status and physical distance of the device combination, and obtain physical impact parameters based on the physical connection status and physical distance; Based on the basic process impact parameters, the health consistency parameters, and the physical impact parameters, a reliable correlation result is obtained. Then, a new set of equipment combinations is obtained and evaluated to obtain a reliable correlation result, until the evaluation of all equipment combinations is completed. Based on multiple trusted association results of multiple device combinations of a single device, the device health parameters are corrected to obtain device health identification results; Specifically, based on multiple trusted association results of multiple device combinations for a single device, a trusted assessment of the device health parameters is performed to obtain device health identification results, including: Randomly select a device as the first device; Using each of the trusted association results of the first device as a weight, the current device health parameters of another device in the device combination including the first device are weighted and calculated to generate an association correction factor for the first device. Specifically, each trusted association result of the first device is used as a weight to perform a weighted calculation on the current device health parameters of another device in the device combination including the first device, generating an association correction factor for the first device, including: Randomly select a first device combination that includes the first device, obtain the trusted association result of the first device combination, and normalize to obtain the association weight; Obtain the device health parameters of the other device in the first device combination; Multiply the association weight by the device health parameter of the other device to obtain an association correction factor for the first device based on the combination of the first devices; A reliable assessment is performed on all the associated correction factors of the first device, and the device health identification result is obtained by combining the device health parameters of the first device.

[0065] In one embodiment, the monitoring result analysis module 400 is further configured to: Based on the device health identification results, historical health maintenance plans are obtained, and all devices are sorted by health status to generate a health status list. Based on the historical health maintenance plans and current operating conditions, a recommended maintenance plan is obtained; Based on the assessment of the wear acceleration impact of the aforementioned operating parameters, the health status list is adjusted to obtain a maintenance priority list; Specifically, based on the assessment of the wear acceleration impact of the aforementioned operating parameters, the health status list is adjusted to obtain a maintenance priority list, including: Calculate the percentage of large-diameter materials based on the particle size distribution. The wear coefficient under working conditions is calculated by comparing the large particle size material ratio parameter with the benchmark large particle size material ratio parameter. The wear coefficient under the working condition is used as an adjustment factor, and combined with the health status score of the corresponding equipment in the health status list for fusion calculation to obtain the updated health identification result. The updated health identification result is then reordered to generate the maintenance priority list. Based on the proposed maintenance plan and the maintenance priority list, a maintenance plan for the coal mine dry separation system is generated to carry out health management of the coal mine dry separation system.

[0066] In this embodiment of the application, a maintenance plan for the coal mine dry separation system is obtained by combining the equipment health identification results and the operating parameters, and the health management of the coal mine dry separation system is carried out.

[0067] In summary, the embodiments of this application have at least the following technical effects: This application proposes a health management method and system for coal mine dry separation systems based on multi-source data fusion. By systematically acquiring and fusing equipment operating parameters and working condition parameters reflecting material characteristics, and constructing an intelligent agent capable of learning the complex mapping relationship between the two, it achieves a more essential and accurate characterization of equipment health status, significantly improving the pertinence and reliability of health status identification. Specifically, this application enables health assessment to closely align with the actual production of coal mine dry separation. For example, it can distinguish between reasonable load conditions exceeding normal vibration range caused by an increased proportion of large gangue being processed and genuine early mechanical failures, thereby reducing false alarms and missed alarms. Furthermore, by introducing process influence parameters based on process sequence, physical connection relationships between equipment, and historical health data, a reliable assessment mechanism is constructed. This mechanism can cross-validate and logically correct the initial health parameters output by the intelligent agent, utilizing the spatiotemporal correlation of equipment status in the system to filter abnormal fluctuations or identify hidden faults. Therefore, the final equipment health identification results not only reflect the individual state but also demonstrate their true impact and reliability within the overall system, enhancing the holistic view of health management and the robustness of decision-making. Based on this, by combining equipment health identification results with real-time operating parameters, a maintenance plan that matches the current production conditions can be dynamically generated. This method not only prioritizes maintenance based on health status but also proactively assesses the accelerated wear impact on specific equipment based on real-time material characteristics. This allows for predictive adjustments to maintenance priorities and recommendations for targeted inspections and maintenance items, transforming maintenance activities from a passive response or fixed-cycle model to a proactive, precise, and production-synchronized predictive maintenance model. Compared to traditional methods that rely solely on equipment threshold monitoring, the technical solution provided in this application significantly overcomes the problem of inaccurate early warnings caused by ignoring actual production load changes and equipment coupling effects, achieving the technical effects of improving the overall operational reliability of coal mine dry separation systems and optimizing the efficiency of maintenance resource allocation.

[0068] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0069] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0070] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A health management method for coal mine dry separation systems based on multi-source data fusion, characterized in that, include: Obtain the equipment operating parameters of multiple devices in the coal mine dry separation system, and obtain the working condition parameters; Based on the equipment operating parameters and the working condition parameters, an intelligent agent for equipment status recognition is constructed to identify the health status of the equipment and obtain multiple equipment health parameters. Based on the process sequence of the coal mine dry separation system, process influence parameters of multiple devices are obtained, and combined with the physical connection status and historical health parameters of the devices, the health parameters of the devices are reliably evaluated to obtain the device health identification results. Based on the equipment health identification results and the operating parameters, a maintenance plan for the coal mine dry separation system is obtained, and health management of the coal mine dry separation system is carried out.

2. The health management method for a coal mine dry separation system based on multi-source data fusion according to claim 1, characterized in that, Obtain the equipment operating parameters of multiple devices in the coal mine dry separation system, and obtain the operating condition parameters, including: Acquire multiple devices from a coal mine dry separation system, and obtain the operating parameters of these devices based on their characteristics; Based on the current dry coal separation material, operating parameters are obtained, including material particle size distribution and instantaneous material flow rate.

3. The health management method for a coal mine dry separation system based on multi-source data fusion according to claim 1, characterized in that, Based on the equipment operating parameters and the working condition parameters, an intelligent agent for equipment status recognition is constructed to identify the equipment health status and obtain multiple equipment health parameters, including: Construct a device status recognition intelligent agent, wherein the device status recognition intelligent agent includes multiple recognition branches; The device's operating parameters and working condition parameters are input into the device status recognition intelligent agent to identify and obtain multiple device health parameters.

4. The health management method for a coal mine dry separation system based on multi-source data fusion according to claim 3, characterized in that, Constructing an intelligent agent for device status recognition includes: Based on multiple devices in a coal mine dry separation system, construct multiple intelligent identification branches for device status; Based on multiple devices, obtain multiple historical device operating parameters and historical operating condition parameters, and obtain the corresponding historical device health parameters and device tags to form a historical sample set of devices; The device status intelligent recognition branches are trained separately using the device historical sample set until convergence, and then integrated to obtain the device status recognition intelligent agent.

5. The health management method for a coal mine dry separation system based on multi-source data fusion according to claim 1, characterized in that, Based on the process sequence of the coal mine dry separation system, process influence parameters of multiple devices are obtained. Combined with the physical connection status and historical health parameters of these devices, a reliable assessment of the device health parameters is performed to obtain device health identification results, including: A random combination of devices is selected, wherein the combination of devices includes two devices; Based on the process sequence of the coal mine dry separation system, the basic process influence parameters of the equipment combination are obtained. Obtain historical health parameters of the device combination and evaluate the obtained health consistency parameters; Obtain the physical connection status and physical distance of the device combination, and obtain physical impact parameters based on the physical connection status and physical distance; Based on the basic process impact parameters, the health consistency parameters, and the physical impact parameters, a reliable correlation result is obtained. Then, a new set of equipment combinations is obtained and evaluated to obtain a reliable correlation result, until the evaluation of all equipment combinations is completed. Based on multiple trusted association results of multiple device combinations for a single device, the device health parameters are corrected to obtain device health identification results.

6. The health management method for a coal mine dry separation system based on multi-source data fusion according to claim 5, characterized in that, Based on multiple trusted association results of multiple device combinations for a single device, a trusted assessment of the device health parameters is performed to obtain device health identification results, including: Randomly select a device as the first device; Using each of the trusted association results of the first device as a weight, the current device health parameters of another device in the device combination including the first device are weighted and calculated to generate an association correction factor for the first device. A reliable assessment is performed on all the associated correction factors of the first device, and the device health identification result is obtained by combining the device health parameters of the first device.

7. The health management method for a coal mine dry separation system based on multi-source data fusion according to claim 6, characterized in that, Using each of the trusted association results of the first device as a weight, a weighted calculation is performed on the current device health parameters of another device in the device combination including the first device to generate an association correction factor for the first device, including: Randomly select a first device combination that includes the first device, obtain the trusted association result of the first device combination, and normalize to obtain the association weight; Obtain the device health parameters of the other device in the first device combination; The association weight is multiplied by the device health parameter of the other device to obtain an association correction factor for the first device based on the combination of the first devices.

8. The health management method for a coal mine dry separation system based on multi-source data fusion according to claim 1, characterized in that, Based on the equipment health identification results and the operating parameters, a maintenance plan for the coal mine dry separation system is obtained, and health management of the coal mine dry separation system is carried out, including: Based on the device health identification results, historical health maintenance plans are obtained, and all devices are sorted by health status to generate a health status list. Based on the historical health maintenance plans and current operating conditions, a recommended maintenance plan is obtained; Based on the assessment of the wear acceleration impact of the aforementioned operating parameters, the health status list is adjusted to obtain a maintenance priority list; Based on the proposed maintenance plan and the maintenance priority list, a maintenance plan for the coal mine dry separation system is generated to carry out health management of the coal mine dry separation system.

9. The health management method for a coal mine dry separation system based on multi-source data fusion according to claim 8, characterized in that, Based on the assessment of the wear acceleration impact of the aforementioned operating parameters, the health status list is adjusted to obtain a maintenance priority list, including: Calculate the percentage of large-diameter materials based on the particle size distribution. The wear coefficient under working conditions is calculated by comparing the large particle size material ratio parameter with the benchmark large particle size material ratio parameter. The wear coefficient under the operating conditions is used as an adjustment factor, and combined with the health status score of the corresponding equipment in the health status list for fusion calculation to obtain an updated health identification result. The updated health identification result is then used to re-sort the equipment and generate the maintenance priority list.

10. A health management system for coal mine dry separation systems based on multi-source data fusion, characterized in that, A health management method for a coal mine dry separation system for implementing multi-source data fusion as described in any one of claims 1-9, the system comprising: The equipment operation status acquisition module is used to acquire the equipment operation parameters of multiple devices in the coal mine dry separation system, and to acquire the working condition parameters; The health status recognition module is used to construct an intelligent agent for equipment status recognition based on the equipment operating parameters and the working condition parameters, to perform equipment health status recognition, and to obtain multiple equipment health parameters. The identification result evaluation module is used to obtain process influence parameters of multiple devices based on the process sequence of the coal mine dry separation system, and to conduct a reliable evaluation of the device health parameters by combining the physical connection status and historical health parameters of the devices, thereby obtaining the device health identification result. The maintenance plan acquisition module is used to acquire a maintenance plan for the coal mine dry separation system by combining the equipment health identification results and the operating parameters, and to carry out health management of the coal mine dry separation system.

Images