Industrial isomerization equipment internet of things data fusion analysis system empowered by sdc

Abstract

The application discloses an SDC-enabled industrial heterogeneous equipment Internet of Things data fusion analysis system, and relates to the technical field of Internet of Things.The system comprises a cloud production control platform in a coal mine, which determines belt transportation scene parameters according to a coal mining plan; performs skid fault prediction on the transportation equipment to obtain a predicted skid fault probability; sets a fault prediction time limit according to predicted underground environment fluctuation characteristics, the belt transportation scene parameters and the predicted skid fault probability; takes the fault prediction time limit as an optimization constraint, and takes the maximization of fault detection accuracy and efficiency as double optimization targets to obtain an optimal monitoring index fusion scheme; and downgrades the optimal monitoring index fusion scheme to an equipment abstraction layer to perform data acquisition, data uploading and data fusion, and performs belt skid fault prediction on the cloud production control platform. The technical problem that existing coal mine belt transmission equipment cannot effectively balance belt skid prediction efficiency and prediction accuracy, resulting in untimely fault detection or misjudgment, is solved.

Description

Technical Field

[0001] This application relates to the field of Internet of Things (IoT) technology, specifically to an SDC-enabled IoT data fusion and analysis system for heterogeneous industrial devices. Background Technology

[0002] With the continuous development of industry, belt conveyor systems in coal mine production face complex operating environments and higher safety requirements. However, traditional fault prediction methods typically employ fixed monitoring indicators and fusion schemes, which cannot adjust and optimize belt slippage phenomena in coal mine transportation in real time according to actual conditions. This makes it difficult to achieve the best balance between the accuracy and efficiency of fault detection, leading to untimely or misjudged faults and posing safety hazards to coal mine production. Summary of the Invention

[0003] This application provides an SDC-enabled IoT data fusion and analysis system for industrial heterogeneous equipment, which solves the technical problem that existing coal mine belt conveyor equipment cannot effectively balance the prediction efficiency and accuracy of belt slippage, resulting in untimely fault detection or misjudgment.

[0004] The technical solution to the above-mentioned technical problems in this application is as follows:

[0005] Firstly, this application provides an SDC-enabled industrial heterogeneous device IoT data fusion and analysis system, the system comprising:

[0006] The parameter determination module determines the parameters for the belt conveyor scenario based on the coal mining plan within the preset time zone on the cloud-based production control platform of the coal mine.

[0007] The fault prediction module, based on the belt conveyor scenario parameters, predicts slippage faults of the belt conveyor equipment within the preset time zone and obtains the predicted slippage fault probability.

[0008] The time limit setting module sets the fault prediction time limit based on the predicted downhole environment fluctuation characteristics, the belt conveyor scenario parameters, and the predicted slippage failure probability.

[0009] The solution acquisition module uses the fault prediction time limit as the optimization constraint and maximizes the accuracy and efficiency of fault detection as the dual optimization objectives to iteratively optimize and search for the monitoring index fusion scheme for slippage fault prediction, and obtains the best monitoring index fusion scheme.

[0010] The fault management module delegates the optimal monitoring indicator fusion scheme to the equipment abstraction layer, performs data collection, data uploading and data fusion, and performs belt slippage fault prediction on the cloud production management platform.

[0011] This application provides one or more technical solutions, which have at least the following technical effects or advantages:

[0012] This application provides an SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system. First, on a coal mine cloud-based production control platform, parameters for the belt conveyor scenario are determined. Then, slippage fault prediction is performed on the belt conveyor equipment to obtain the prediction probability. Next, a reasonable fault prediction timeframe is set considering factors such as underground environmental fluctuations. Using this timeframe as a constraint and aiming to maximize fault detection accuracy and efficiency, the optimal monitoring indicator fusion scheme is obtained. Finally, the scheme is deployed for data collection and other operations, effectively balancing prediction efficiency and accuracy to achieve belt slippage fault prediction. The fault prediction and detection scheme is dynamically adjusted according to the coal mining plan and actual environment, avoiding the drawbacks of traditional methods that use fixed monitoring indicators and fusion schemes. This system, through real-time optimization, reduces production stoppages and safety accidents caused by untimely or misjudged fault detection. Simultaneously, with the dual optimization goal of maximizing fault detection accuracy and efficiency, it can improve detection speed and reduce detection costs while ensuring detection quality.

[0013] Through the above technical solution, this application comprehensively considers factors such as the characteristics of underground environmental fluctuations, belt conveyor scenario parameters, and the predicted probability of slippage failure, and iteratively optimizes and searches the monitoring index fusion scheme for slippage failure prediction. It dynamically adjusts the monitoring indicators and fusion scheme according to the actual situation, thereby improving the accuracy and efficiency of fault detection and providing a guarantee for the safe and stable operation of coal mine belt conveyor systems. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application, 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the structure of the SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system provided in this application embodiment.

[0016] The components represented by each number in the attached diagram are explained below:

[0017] The module includes a parameter determination module S10, a fault prediction module S20, a time limit setting module S30, a solution acquisition module S40, and a fault control module S50. Detailed Implementation

[0018] This application provides an SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system to address the technical problem in existing coal mine belt conveyor equipment that cannot effectively balance belt slippage prediction efficiency and prediction accuracy, resulting in untimely fault detection or misjudgment.

[0019] 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 some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0021] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessarily obscuring the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0022] Example 1, as Figure 1 As shown, this application embodiment provides an SDC-enabled industrial heterogeneous device IoT data fusion and analysis system, including:

[0023] S10: Parameter determination module, which determines the parameters of the belt conveyor scenario based on the coal mining plan within the preset time zone on the cloud-based production control platform of the coal mine;

[0024] In this embodiment, the cloud-based production control platform of the coal mine first determines the belt conveyor scenario parameters based on the coal mining plan within a preset time zone. The coal mining plan includes the coal mining volume and mining area within the preset time zone. The coal mining volume affects the belt conveyor load, and different mining areas will cause differences in transportation distance, transportation gradient, etc. For example, if the coal mining volume within the preset time zone increases significantly, the belt conveyor load will increase accordingly, and the belt conveyor equipment will need to output more power to ensure normal transportation, which will also increase the possibility of slippage failure. If the mining area changes, the transportation distance becomes longer, or the transportation gradient becomes steeper, which will also increase the difficulty of belt transportation and similarly increase the probability of slippage failure.

[0025] Then, through a detailed analysis of the coal mining plan, combined with the actual transportation environment of the coal mine, such as the type, length, and width of the belt, as well as the performance indicators of the drive equipment and transmission equipment, the parameters of the belt transportation scenario are determined on the cloud-based production control platform.

[0026] Specifically, step S10 in the system includes:

[0027] The belt conveyor scenario data is determined based on the coal mining plan within the preset time zone, wherein the belt conveyor scenario data includes belt load sequence and coal pile thickness sequence;

[0028] Based on the belt load sequence and coal pile thickness sequence, the average belt load, maximum belt load, average coal pile thickness, maximum coal pile thickness, and coal pile uniformity coefficient are calculated as parameters for the belt transport scenario.

[0029] In this embodiment of the application, firstly, belt conveyor scenario data is determined according to the coal mining plan within a preset time zone. The belt conveyor scenario data includes belt load sequence and coal pile thickness sequence. The belt load sequence refers to the sequence of coal weight data carried by the belt at different times within the preset time zone, and the coal pile thickness sequence is the sequence of coal thickness data accumulated on the belt at the corresponding time.

[0030] For example, if a large amount of coal is mined in a specific area during a certain period of the coal mining plan, the load on the belt conveyor will increase accordingly, and the thickness of the coal pile will also change.

[0031] Then, based on the belt load sequence and coal pile thickness sequence, the average belt load, maximum belt load, average coal pile thickness, maximum coal pile thickness, and coal pile uniformity coefficient are calculated as parameters for the belt transport scenario. The average belt load reflects the average load-bearing capacity of the belt within a preset time zone, and is calculated by summing all data in the belt load sequence and dividing by the number of data points. The maximum belt load is the maximum value in the belt load sequence, reflecting the maximum pressure the belt may face within the preset time zone, and evaluating the performance and reliability of the belt.

[0032] The average coal pile thickness is the average of all data in the coal pile thickness sequence, reflecting the average coal accumulation on the conveyor belt; the maximum coal pile thickness is the maximum value in the coal pile thickness sequence. If this value is too large, it may cause the conveyor belt to run unstably and increase the risk of slippage failure.

[0033] The coal pile uniformity coefficient measures the uniformity of coal accumulation on the conveyor belt. It is calculated by dividing the standard deviation of the coal pile thickness sequence by the mean coal pile thickness. A smaller coal pile uniformity coefficient indicates more uniform coal accumulation on the conveyor belt, resulting in more stable belt operation. Conversely, a larger coefficient indicates less uniform coal accumulation, leading to more uneven distribution of resistance and pressure on the conveyor belt, thus increasing the likelihood of slippage.

[0034] By accurately calculating and analyzing the parameters of belt conveyor scenarios, more reliable data support is provided for predicting belt slippage faults, thereby improving the accuracy and efficiency of fault prediction and ensuring the safe and stable operation of coal mine belt conveyor systems.

[0035] S20: Fault prediction module, based on the belt conveyor scenario parameters, performs slippage fault prediction on the belt conveyor equipment in the preset time zone and obtains the predicted slippage fault probability;

[0036] In this embodiment, a slippage fault prediction model is established using established belt conveyor scenario parameters, such as average belt load, maximum belt load, average coal pile thickness, maximum coal pile thickness, and coal pile uniformity coefficient, combined with historical operating data and fault records of belt conveyor equipment within a preset time zone. This model employs machine learning algorithms, such as generative adversarial networks (GANs), to identify the potential relationships and patterns between belt conveyor scenario parameters and slippage faults through learning from historical data and adversarial training.

[0037] When making predictions, the current belt conveyor scenario parameters are input into a trained prediction model. The model then calculates and analyzes based on learned patterns, outputting the predicted probability of slippage failure for the belt conveyor within a preset time zone. For example, if the average belt load is too high, the maximum thickness of the coal pile is too large, or the coal pile uniformity coefficient is too high, the model may predict a higher probability of slippage failure.

[0038] Specifically, step S20 in the system includes:

[0039] Based on the historical operation monitoring records of belt conveyor equipment, a sample belt conveyor scenario parameter set is collected, and the proportion of slippage failure events of different sample belt conveyor scenario parameters in the historical time zone is counted as the sample slippage failure probability, thus obtaining a sample slippage failure probability set. The time interval between the historical time zone and the preset time zone is the same.

[0040] Using the parameter set of the sample belt transport scenario as input and the probability set of the sample slippage failure as supervision, a generative adversarial network is trained until convergence to generate a slippage failure predictor.

[0041] Using the slippage fault predictor, slippage faults are predicted based on the belt conveyor scenario parameters, and the predicted slippage fault probability of the belt conveyor equipment within the preset time zone is output.

[0042] In this embodiment, firstly, a sample set of belt conveyor scenario parameters is collected based on the historical operation monitoring records of the belt conveyor equipment. The sample set of belt conveyor scenario parameters covers data such as the average belt load, maximum belt load, average coal pile thickness, maximum coal pile thickness, and coal pile uniformity coefficient over different time periods.

[0043] Simultaneously, the percentage of slippage failure events in different sample belt conveyor scenarios within historical time zones is statistically analyzed to obtain the sample slippage failure probability set. The time interval between the historical time zone and the preset time zone is the same, ensuring data comparability and validity. For example, if the preset time zone is one month, then the historical time zone also selects a one-month period.

[0044] Secondly, using the parameter set of the sample belt conveyor scenario as input and the sample slippage failure probability set as supervision, a generative adversarial network is trained until convergence to generate a slippage failure predictor.

[0045] Generative Adversarial Networks (GANs) are machine learning models consisting of a generator and a discriminator. The generator attempts to generate samples similar to the real data distribution, while the discriminator distinguishes between generated and real samples. Through continuous adversarial training, the generator learns the inherent patterns in the data, thereby generating accurate predictions. During training, the parameters of the generator and discriminator are continuously adjusted until the GAN converges, at which point the generated slippage fault predictor exhibits high accuracy and reliability.

[0046] Finally, the generated slippage fault predictor is used to predict slippage faults based on the current belt conveyor scenario parameters, outputting the predicted slippage fault probability of the belt conveyor equipment within a preset time zone. Real-time collected belt conveyor scenario parameters are input into the predictor, which calculates and analyzes based on learned patterns, outputting the prediction results. For example, if the current average belt load continues to increase, the maximum coal pile thickness exceeds the normal range, or the coal pile uniformity coefficient increases, the predictor will correspondingly increase the predicted slippage fault probability, providing timely and effective early warning information so that appropriate measures can be taken to prevent and handle belt slippage faults, ensuring the safe and stable operation of the coal mine belt conveyor system.

[0047] For example, the specific steps for building and training a skid failure predictor based on a generative adversarial network are as follows:

[0048] First, data acquisition involves collecting historical operational monitoring records of the belt conveyor equipment, including average belt load, maximum belt load, average coal pile thickness, maximum coal pile thickness, coal pile uniformity coefficient, and corresponding slippage failure event records. Simultaneously, the historical time zone and the preset time zone are determined to ensure that the time intervals between them are the same. The collected data is then cleaned to remove outliers and missing values ​​to improve data quality. Finally, the data is normalized to unify data from different ranges onto the same scale, facilitating model learning.

[0049] Secondly, model building involves dividing the processed data into training and test sets according to a certain ratio, such as an 8:2 ratio. The training set is used to train the generative adversarial network (GAN), and the test set is used to evaluate the model's performance. The generator and discriminator are initialized, employing neural network structures such as multilayer perceptrons. Appropriate initial parameters are set for the generator and discriminator, including the number of neurons and the learning rate. For example, the generator can be set to 3 hidden layers, each containing 50 neurons, with a learning rate of 0.001; the discriminator can be set to 2 hidden layers, each containing 30 neurons, with a learning rate of 0.0005.

[0050] Next, the model is trained by inputting the training set data into the generative adversarial network (GAN). In each training iteration, the generator attempts to generate samples similar to the real data distribution, while the discriminator attempts to distinguish between the generated samples and the real samples. Using the parameter set of the sample belt conveyor scenario as input and the sample slippage failure probability set as supervision, the training framework is constructed using the Adam optimizer and the mean squared error (MSE) loss function. The batch size is set to 32, the total number of training epochs is 50, and an early stopping mechanism (patience=5) is introduced. When the validation set loss does not decrease for 5 consecutive epochs, the training process is automatically terminated, resulting in a trained slippage failure predictor. This process is repeated until the GAN reaches convergence.

[0051] Through the above steps, a slippage fault predictor built and trained based on a generative adversarial network can accurately predict the probability of slippage faults of belt conveyor equipment in a preset time zone according to the parameters of the belt conveyor scenario, thus providing a guarantee for the safe and stable operation of the coal mine belt conveyor system.

[0052] S30: Time limit setting module, which sets the fault prediction time limit based on the predicted downhole environment fluctuation characteristics, the belt conveyor scenario parameters and the predicted slippage failure probability;

[0053] In this embodiment, a fault prediction time limit is set by comprehensively predicting the downhole environment fluctuation characteristics, belt conveyor scenario parameters, and the predicted slippage failure probability. Downhole environment fluctuation characteristics affect the operation of belt conveyor equipment, and belt conveyor scenario parameters and the predicted slippage failure probability are also the basis for setting the fault prediction time limit.

[0054] Taking into account the predicted downhole environmental fluctuations, belt conveyor scenario parameters, and the predicted probability of slippage failures, the fault prediction time limit is dynamically set according to different situations. By reasonably setting the fault prediction time limit, the timeliness and accuracy of fault prediction are improved, effectively reducing the probability of slippage failures in belt conveyor equipment.

[0055] Specifically, step S30 in the system includes:

[0056] The predicted downhole environmental parameter sequence within the preset time zone is obtained for downhole environmental fluctuation analysis, and the downhole environmental fluctuation coefficient is output. The downhole environmental parameters include at least air humidity, downhole environmental temperature, dust concentration and gas concentration.

[0057] Based on the belt load sequence and coal pile thickness sequence, a transportation scenario fluctuation analysis is performed, and the transportation scenario fluctuation coefficient is output.

[0058] Based on the sample parameter-time limit correction coefficient comparison table, the prediction time limit correction coefficient is determined according to the downhole environment fluctuation coefficient, transportation scenario fluctuation coefficient and predicted slippage failure probability. The prediction time limit correction coefficient is negatively correlated with the downhole environment fluctuation coefficient, transportation scenario fluctuation coefficient and predicted slippage failure probability.

[0059] The ratio of the prediction time limit correction coefficient to the initial fault prediction time limit is used as the fault prediction time limit.

[0060] In this embodiment, firstly, a downhole environment fluctuation analysis is performed to obtain a predicted downhole environment parameter sequence within a preset time zone. These parameters include air humidity, downhole ambient temperature, dust concentration, and methane concentration. Increased humidity reduces the surface friction of the conveyor belt, increasing the risk of slippage. Excessively high or low downhole ambient temperatures affect the physical properties of the conveyor belt, causing it to soften or harden, thus impacting its normal operation. Excessively high dust concentrations accumulate on the surface of the conveyor belt and rollers, affecting the friction between them. Excessively high methane concentrations pose a safety hazard and may affect the normal operation of the conveyor belt system.

[0061] For example, a sudden increase in air humidity may make the belt surface slippery, increasing the risk of slippage; excessively high underground ambient temperature may affect the material properties of the belt, reducing its friction; excessive dust concentration may enter the belt drive system, affecting the normal operation of the equipment; and excessive gas concentration may pose safety hazards and even affect the normal operation of the equipment. The underground environment fluctuation coefficient is obtained by calculating the ratio of the standard deviation to the mean of the parameters in the parameter sequence and then weighting them. This coefficient reflects the degree of fluctuation of the underground environment within a preset time zone. The weights are positively correlated with the impact of fault prediction; that is, the larger the weight, the higher the impact of fault prediction.

[0062] Secondly, a transportation scenario fluctuation analysis is conducted based on the belt load sequence and the coal pile thickness sequence. The fluctuation of the belt load sequence reflects the change in the weight of coal carried by the belt, while the fluctuation of the coal pile thickness sequence reflects the change in the coal accumulation on the belt. By analyzing the transportation scenario fluctuations of the belt load sequence and the coal pile thickness sequence, a transportation scenario fluctuation coefficient is output, which reflects the degree of fluctuation in the belt transportation scenario within a preset time zone.

[0063] Secondly, a sample parameter-time limit correction coefficient comparison table was constructed based on sample data analysis. The prediction time limit correction coefficient was determined by matching the downhole environment fluctuation coefficient, transportation scenario fluctuation coefficient, and predicted slippage failure probability. Since the prediction time limit correction coefficient is negatively correlated with the downhole environment fluctuation coefficient, transportation scenario fluctuation coefficient, and predicted slippage failure probability, it indicates that when the downhole environment fluctuation coefficient, transportation scenario fluctuation coefficient, or predicted slippage failure probability increases, the prediction time limit correction coefficient will decrease accordingly.

[0064] For example, if the downhole environment fluctuates drastically, the transportation scenario is unstable, and the predicted probability of slippage failure is high, then the prediction time limit correction coefficient will be smaller, which means that the failure prediction time limit needs to be shortened.

[0065] Finally, the ratio of the prediction time limit correction coefficient to the initial fault prediction time limit is used as the fault prediction time limit. The initial fault prediction time limit is a pre-set base value, which can be customized according to the actual scenario, for example, 60 seconds. The final fault prediction time limit is obtained by calculating the ratio of the prediction time limit correction coefficient to the initial fault prediction time limit. Based on the actual situation such as the fluctuation characteristics of the underground environment, the parameters of the belt conveyor scenario, and the predicted probability of slippage failure, the fault prediction time limit is dynamically adjusted to improve the timeliness and accuracy of fault prediction, thereby effectively reducing the probability of slippage failure of belt conveyor equipment and ensuring the safe and stable operation of the coal mine belt conveyor system.

[0066] S40: The scheme acquisition module uses the fault prediction time limit as the optimization constraint and maximizes the accuracy and efficiency of fault detection as the dual optimization objectives to iteratively optimize and search the monitoring index fusion scheme for slippage fault prediction and obtain the best monitoring index fusion scheme.

[0067] In this embodiment, firstly, using the fault prediction time limit as an optimization constraint and maximizing the accuracy and efficiency of fault detection as dual optimization objectives, an iterative optimization search is performed on the monitoring index fusion scheme for slippage fault prediction. The monitoring index fusion scheme involves the combination and weight allocation of multiple monitoring indicators, and different combinations and weights will have different impacts on the accuracy and efficiency of fault detection.

[0068] In the iterative optimization search process, the initial monitoring indicator fusion scheme is first determined. This scheme is then applied to actual belt conveyor scenario data, and the accuracy and efficiency of fault detection under this scheme are calculated. Fault detection accuracy can be measured by calculating the degree of matching between the predicted results and the actual fault conditions, such as calculating the true positive rate and true negative rate. Fault detection efficiency can be measured by calculating the time required to complete one fault detection.

[0069] Then, based on the calculated fault detection accuracy and efficiency, the monitoring indicator fusion scheme is adjusted. This can involve changing the combination of monitoring indicators or adjusting the weights of individual indicators. For example, if a monitoring indicator is found to contribute little to fault detection under the current scheme, its weight can be reduced or it can be removed from the scheme; conversely, if a monitoring indicator can significantly improve the accuracy and efficiency of fault detection, its weight can be appropriately increased.

[0070] After each adjustment, the new monitoring indicator fusion scheme is applied to the actual data for evaluation again, and this process is repeated until a preset stopping condition is met. The stopping condition may be reaching the maximum number of iterations, or the accuracy and efficiency of fault detection no longer show significant improvement after multiple consecutive iterations.

[0071] By iteratively optimizing the search process, the optimal monitoring indicator fusion scheme can be obtained. Within a given fault prediction timeframe, the accuracy and efficiency of fault detection are maximized. Using the optimal monitoring indicator fusion scheme, slippage faults in belt conveyor equipment can be predicted more accurately.

[0072] Specifically, step S40 in the system includes:

[0073] Obtain a preset monitoring index set for slippage fault prediction, wherein the preset monitoring index includes drive equipment index, transmission equipment index, conveyor belt monitoring index and sensor type, and the sensor type includes at least coal pile sensor, gas sensor, dust sensor and temperature and humidity sensor.

[0074] Starting with the initial number of monitoring indicators, the monitoring indicator fusion schemes are enumerated using the preset monitoring indicator set. The power set of the preset monitoring indicator set is used as multiple initial monitoring indicator fusion scheme sets, wherein the initial number of monitoring indicators is 5.

[0075] A fault accuracy predictor and a fault prediction duration analyzer are constructed based on generative adversarial networks.

[0076] Using the fault prediction duration analyzer, multiple fault prediction durations are obtained based on the multiple initial monitoring indicator fusion scheme sets. Several qualified monitoring indicator fusion scheme sets are then selected with the constraint that the duration is less than the fault prediction time limit.

[0077] Using the fault accuracy predictor, several prediction accuracies are obtained by analyzing the fusion scheme set of several qualified monitoring indicators;

[0078] With the dual objectives of maximizing fault detection accuracy and efficiency, the monitoring index fusion scheme for slippage fault prediction is iteratively optimized and searched based on several fault prediction durations and several prediction accuracies to obtain the optimal monitoring index fusion scheme.

[0079] In this embodiment, firstly, a preset set of monitoring indicators is obtained. These indicators include drive equipment indicators, transmission equipment indicators, conveyor belt monitoring indicators, and various sensor types. Drive equipment indicators reflect the operating status of the drive motor, such as motor speed and power. Unstable motor speed may lead to uneven belt speed, increasing the risk of slippage. Transmission equipment indicators include the diameter and transmission efficiency of the transmission rollers. Wear on the transmission rollers affects the friction between the belt and the rollers. Conveyor belt monitoring indicators include the tension and thickness of the conveyor belt. Insufficient conveyor belt tension can easily cause slippage.

[0080] The sensor types include coal pile sensors, gas sensors, dust sensors, and temperature and humidity sensors. These sensors can monitor various parameters of the transportation scenario and the underground environment in real time.

[0081] Secondly, starting with an initial number of 5 monitoring indicators, i.e., the minimum number of parameters in the subset is 5, the fusion schemes of the preset monitoring indicator set are enumerated. The power set of the preset monitoring indicator set is used as a set of multiple initial monitoring indicator fusion schemes. The set of all subsets in the set, including from a single element to the entire set itself, generates numerous schemes containing different combinations of monitoring indicators.

[0082] Then, a fault accuracy predictor and a fault prediction time analyzer are constructed based on generative adversarial networks. The fault accuracy predictor can predict the accuracy of slippage fault detection based on different monitoring index fusion schemes; the fault prediction time analyzer can analyze the time required to complete a fault prediction under each scheme.

[0083] Furthermore, a fault accuracy predictor and a fault prediction duration analyzer are constructed based on generative adversarial networks. The set of monitoring indicator fusion schemes is used as input, and the predicted fault accuracy and predicted fault duration are output, respectively.

[0084] For example, taking the construction of a fault accuracy predictor as an example, the specific steps are as follows:

[0085] First, data collection involves using the power set of a preset monitoring indicator set as a fusion scheme set for multiple initial monitoring indicators.

[0086] Secondly, model building involves setting appropriate initial parameters for the generator and discriminator, including the number of neurons and the learning rate. For example, the generator can be set to have 3 hidden layers, each containing 50 neurons, with a learning rate of 0.001; the discriminator can be set to have 2 hidden layers, each containing 30 neurons, with a learning rate of 0.0005.

[0087] Next, the model is trained by inputting the monitoring index fusion scheme set data into the generative adversarial network (GAN). In each training iteration, the generator attempts to generate samples similar to the real data distribution, while the discriminator attempts to distinguish between the generated samples and the real samples. Using the parameter set of the sample belt conveyor scenario as input and the sample slippage failure probability set as supervision, the training framework is constructed using the Adam optimizer and the mean squared error (MSE) loss function. The batch size is set to 32, the total number of training rounds is 50, and an early stopping mechanism (patience=5) is introduced. When the validation set loss does not decrease for 5 consecutive rounds, the training process is automatically terminated, resulting in a trained fault accuracy predictor. This process is repeated until the GAN reaches convergence.

[0088] The same method was used to build and train the fault prediction duration analyzer.

[0089] Next, the fault prediction duration analyzer is used to analyze multiple initial monitoring indicator fusion scheme sets to obtain multiple fault prediction durations, which are then filtered based on the constraint that they are less than the fault prediction time limit. If the fault prediction duration of a certain scheme exceeds the fault prediction time limit, it will be excluded even if the fault detection accuracy of the scheme may be high. After filtering, several qualified monitoring indicator fusion scheme sets are obtained.

[0090] Then, the fault accuracy predictor is used to analyze the fusion scheme set of qualified monitoring indicators to obtain several prediction accuracies.

[0091] Finally, with the dual optimization objectives of maximizing fault detection accuracy and efficiency, an iterative optimization search was conducted on the monitoring indicator fusion scheme for slippage fault prediction based on several fault prediction durations and several prediction accuracies. During the process, the combination and weight of monitoring indicators were continuously adjusted, and the optimal monitoring indicator fusion scheme was found by comparing the fault prediction duration and prediction accuracy of different schemes. The optimal monitoring indicator fusion scheme can maximize the accuracy and efficiency of fault detection while meeting the fault prediction time limit, providing a more reliable guarantee for the safe and stable operation of coal mine belt conveyor systems.

[0092] Furthermore, to obtain the optimal monitoring indicator fusion scheme, including:

[0093] The reciprocal of the prediction time limit correction coefficient is set as the prediction time limit sensitivity, and the product of the prediction time limit sensitivity and the initial efficiency weight is set as the adaptation efficiency weight, wherein the initial efficiency weight is 0.4, and the adaptation efficiency weight is not less than 0.2 and not greater than 0.6.

[0094] The adaptation accuracy weight is obtained by subtracting the adaptation efficiency weight from 1.

[0095] With the dual optimization objectives of maximizing fault detection accuracy and efficiency, based on the adaptation efficiency weight and adaptation accuracy weight, after dimensionless processing of the several fault prediction durations and several prediction accuracies, several fusion scheme quality coefficients are obtained by weighted calculation. Among them, the fusion scheme quality coefficient is negatively correlated with the fault prediction duration and positively correlated with the prediction accuracy.

[0096] The optimal monitoring index fusion scheme is selected as the fusion scheme corresponding to the qualified monitoring index with the highest fusion scheme quality coefficient.

[0097] In this embodiment, firstly, the prediction time limit sensitivity is calculated. The prediction time limit sensitivity is the reciprocal of the prediction time limit correction coefficient, reflecting the sensitivity of the fault prediction time limit to system operation. The prediction time limit sensitivity is multiplied by the initial efficiency weight of 0.4 to obtain the adaptive efficiency weight. The adaptive efficiency weight is then limited to a range of no less than 0.2 and no greater than 0.6, ensuring that the weight of efficiency factors is within a reasonable range during fault detection—neither too high, leading to neglect of accuracy, nor too low, causing efficiency to be neglected.

[0098] Secondly, subtracting the adaptation efficiency weight from 1 yields the adaptation accuracy weight, demonstrating that the sum of the weights between fault detection accuracy and efficiency is always 1. By reasonably allocating these two weights, the accuracy and efficiency of fault detection can be balanced according to actual needs in different application scenarios.

[0099] For example, if the prediction time limit correction coefficient is 1, then the reciprocal of the prediction time limit correction coefficient is 1, which is set as the prediction time limit sensitivity. The prediction time limit sensitivity 1 is multiplied by the initial efficiency weight 0.4 to obtain the adaptation efficiency weight of 0.4. At this time, the adaptation accuracy weight is 0.6, which meets the requirement that the adaptation efficiency weight is not less than 0.2 and not greater than 0.6.

[0100] Then, with the dual optimization objectives of maximizing fault detection accuracy and efficiency, several fault prediction durations and several prediction accuracies are processed using dimensionless methods. Dimensionless processing eliminates the influence of dimensions between different indicators, obtained by dividing by the indicator's average value. The processed indicators are then weighted according to adaptation efficiency and adaptation accuracy weights to obtain several fusion scheme quality coefficients. Since the fusion scheme quality coefficient is negatively correlated with fault prediction duration and positively correlated with prediction accuracy, this means that the shorter the fault prediction duration and the higher the prediction accuracy, the larger the fusion scheme quality coefficient.

[0101] For example, if the fault prediction time for a certain target area is 10s and the average fault prediction time is 15s, then the fault prediction time obtained through dimensionless processing is 0.67; the fault prediction accuracy for a certain target area is 0.9. At this time, the quality coefficient of the fusion scheme = 0.4 × (1 - 0.67) + 0.6 × 0.9 = 0.672.

[0102] Finally, after obtaining the quality coefficients of several fusion schemes, the coefficients are compared. Since the dual optimization objective is to maximize fault detection accuracy and efficiency, the fusion scheme with the highest quality coefficient corresponding to the qualified monitoring index is selected as the optimal monitoring index fusion scheme. The optimal scheme achieves the best balance between fault detection accuracy and efficiency while meeting the fault prediction time limit.

[0103] S50: Fault management module, which decentralizes the best monitoring index fusion scheme to the equipment abstraction layer, performs data collection, data uploading and data fusion, and performs belt slippage fault prediction on the cloud production management platform.

[0104] In this embodiment, after obtaining the optimal monitoring indicator fusion scheme, the system enters the fault management phase. The optimal monitoring indicator fusion scheme is then delegated to the equipment abstraction layer. The equipment abstraction layer performs comprehensive data collection based on this scheme, utilizing various sensors and related indicators from a preset monitoring indicator set. The coal accumulation sensor monitors the coal accumulation on the conveyor belt in real time, the gas sensor continuously detects the underground gas concentration, the dust sensor records changes in dust concentration, and the temperature and humidity sensor acquires underground temperature and humidity data.

[0105] The collected data is transmitted to the cloud-based production management platform via an upload mechanism. During the upload process, data integrity and timeliness are ensured to prevent data loss or delays from impacting fault prediction. To guarantee the reliability of data transmission, encryption technology and redundant transmission methods can be employed.

[0106] Once the data reaches the cloud-based production control platform, it undergoes data fusion processing. Data fusion integrates and analyzes data from different sensors and monitoring indicators to extract more valuable information. Through data fusion, multiple factors such as the underground environment and transportation scenarios can be comprehensively considered to more accurately determine whether there are potential slippage hazards in belt conveyor equipment.

[0107] After data fusion is completed, the cloud-based production management platform will use the best monitoring indicator fusion scheme to predict belt slippage faults. If the prediction results indicate a high risk of faults, the system will promptly issue an early warning signal to remind relevant personnel to take corresponding preventive and handling measures, such as adjusting belt tension and cleaning dust from the transmission system.

[0108] Specifically, step S50 in the system includes:

[0109] The optimal monitoring indicator fusion scheme is deployed to the equipment abstraction layer. Data is collected and uploaded regularly according to the monitoring indicators in the optimal monitoring indicator fusion scheme. After data fusion on the cloud production management platform, a monitoring fusion dataset is obtained.

[0110] On the cloud-based production management platform, an adaptive fault risk predictor is obtained by matching the optimal monitoring indicator fusion scheme, and belt slippage fault risk is predicted based on the monitoring fusion dataset.

[0111] In this embodiment, firstly, after the optimal monitoring indicator fusion scheme is implemented at the equipment abstraction layer, the equipment abstraction layer strictly follows the monitoring indicators in the scheme and collects data at specified time intervals to ensure comprehensive capture of the operating status data of the belt conveyor equipment at different time periods. Each collected data is promptly transmitted to the cloud-based production management platform through a stable data upload channel to ensure data timeliness.

[0112] Secondly, after receiving the uploaded data, the cloud-based production control platform processes the data using data fusion algorithms. The data fusion process requires comprehensive consideration of the correlations and influences between different monitoring indicators, deeply integrating data from various sensors and monitoring points to obtain a more comprehensive and representative monitoring fusion dataset.

[0113] Next, on the cloud-based production management platform, a suitable fault risk predictor is matched based on the optimal monitoring indicator fusion scheme. The adaptation process selects the prediction model and algorithm based on the characteristics and requirements of the monitoring indicators in the scheme. The risk predictor is built based on deep learning. By predicting risks, it matches the corresponding model according to the number of indicators. The adapted fault risk predictor uses the monitoring fusion dataset, combined with previously calculated information such as the downhole environment fluctuation coefficient, transportation scenario fluctuation coefficient, and prediction time limit correction coefficient, to perform a comprehensive prediction of belt slippage fault risks for belt conveyor equipment.

[0114] During the prediction process, the fault risk predictor conducts in-depth data analysis and mining to identify potential factors and abnormal patterns that may lead to slippage failures. If the prediction results indicate a risk of slippage failure in the belt conveyor equipment, the system will quickly activate the early warning mechanism. This enables accurate prediction and timely control of slippage failures in belt conveyor equipment, providing strong support for the safe production and efficient operation of coal mines.

[0115] In summary, compared to existing technologies, this application improves the accuracy and efficiency of fault detection by optimizing the monitoring indicator fusion scheme for predicting slippage faults in coal mine belt conveyor systems. During the selection and optimization of the monitoring indicator fusion scheme, both fault prediction duration and prediction accuracy were comprehensively considered. Through iterative optimization search and dual optimization objectives, the optimal monitoring indicator fusion scheme was obtained. In the fault management phase, the optimal monitoring indicator fusion scheme was applied to the actual data acquisition, uploading, and fusion process. Combined with a suitable fault risk predictor, the risk of belt slippage faults was accurately predicted. This provides a more reliable guarantee for the safe and stable operation of coal mine belt conveyor systems, reduces equipment damage and production interruptions caused by slippage faults, and lowers the operating costs of coal mining enterprises.

[0116] In summary, the embodiments of this application have at least the following technical effects:

[0117] This application provides an SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system. First, it determines the parameters of the belt conveyor scenario on a coal mine cloud-based production control platform. Then, it predicts belt conveyor slippage faults and obtains the prediction probability. Next, it sets a reasonable fault prediction time limit based on factors such as underground environmental fluctuations. Using this time limit as a constraint and aiming to maximize fault detection accuracy and efficiency, it obtains the optimal monitoring indicator fusion scheme. Finally, it implements the scheme for data collection and other operations, effectively balancing prediction efficiency and accuracy to achieve belt slippage fault prediction. The fault prediction and detection scheme is dynamically adjusted according to the coal mining plan and actual environment, avoiding the drawbacks of traditional methods that use fixed monitoring indicators and fusion schemes. Through real-time optimization, this system reduces production stoppages and safety accidents caused by untimely or misjudged fault detection. Simultaneously, with the dual optimization goal of maximizing fault detection accuracy and efficiency, it can improve detection speed and reduce detection costs while ensuring detection quality. Through the above technical solution, this application comprehensively considers factors such as the characteristics of underground environmental fluctuations, belt conveyor scenario parameters, and the predicted probability of slippage failure, and iteratively optimizes and searches the monitoring index fusion scheme for slippage failure prediction. It dynamically adjusts the monitoring indicators and fusion scheme according to the actual situation, thereby improving the accuracy and efficiency of fault detection and providing a guarantee for the safe and stable operation of coal mine belt conveyor systems.

[0118] 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.

[0119] 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.

[0120] 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. An SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system, characterized in that: The industrial heterogeneous equipment IoT data fusion and analysis system includes: The parameter determination module determines the parameters for the belt conveyor scenario based on the coal mining plan within the preset time zone on the cloud-based production control platform of the coal mine. The fault prediction module, based on the belt conveyor scenario parameters, predicts slippage faults of the belt conveyor equipment within the preset time zone and obtains the predicted slippage fault probability. The time limit setting module sets the fault prediction time limit based on the predicted downhole environment fluctuation characteristics, the belt conveyor scenario parameters, and the predicted slippage failure probability. The solution acquisition module uses the fault prediction time limit as the optimization constraint and maximizes the accuracy and efficiency of fault detection as the dual optimization objectives to iteratively optimize and search for the monitoring index fusion scheme for slippage fault prediction, and obtains the best monitoring index fusion scheme. The fault management module delegates the optimal monitoring indicator fusion scheme to the equipment abstraction layer, performs data collection, data uploading and data fusion, and performs belt slippage fault prediction on the cloud production management platform; Among them, the parameters for belt conveyor scenarios are determined based on the coal mining plan within the preset time zone, including: The belt conveyor scenario data is determined based on the coal mining plan within the preset time zone, wherein the belt conveyor scenario data includes belt load sequence and coal pile thickness sequence; Based on the belt load sequence and coal pile thickness sequence, the average belt load, maximum belt load, average coal pile thickness, maximum coal pile thickness, and coal pile uniformity coefficient are calculated as parameters for the belt transport scenario. The fault prediction time limit is set based on the predicted downhole environment fluctuation characteristics, the belt conveyor scenario parameters, and the predicted slippage failure probability, including: The predicted downhole environmental parameter sequence within the preset time zone is obtained for downhole environmental fluctuation analysis, and the downhole environmental fluctuation coefficient is output. The downhole environmental parameters include at least air humidity, downhole environmental temperature, dust concentration and gas concentration. Based on the belt load sequence and coal pile thickness sequence, a transportation scenario fluctuation analysis is performed, and the transportation scenario fluctuation coefficient is output. Based on the sample parameter-time limit correction coefficient comparison table, the prediction time limit correction coefficient is determined according to the downhole environment fluctuation coefficient, transportation scenario fluctuation coefficient and predicted slippage failure probability. The prediction time limit correction coefficient is negatively correlated with the downhole environment fluctuation coefficient, transportation scenario fluctuation coefficient and predicted slippage failure probability. The ratio of the prediction time limit correction coefficient to the initial fault prediction time limit is used as the fault prediction time limit.

2. The SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system according to claim 1, characterized in that, Based on the belt conveyor scenario parameters, slippage fault prediction is performed on the belt conveyor equipment within the preset time zone to obtain the predicted slippage fault probability, including: Based on the historical operation monitoring records of belt conveyor equipment, a sample belt conveyor scenario parameter set is collected, and the proportion of slippage failure events of different sample belt conveyor scenario parameters in the historical time zone is counted as the sample slippage failure probability, thus obtaining a sample slippage failure probability set. The time interval between the historical time zone and the preset time zone is the same. Using the parameter set of the sample belt transport scenario as input and the probability set of the sample slippage failure as supervision, a generative adversarial network is trained until convergence to generate a slippage failure predictor. Using the slippage fault predictor, slippage faults are predicted based on the belt conveyor scenario parameters, and the predicted slippage fault probability of the belt conveyor equipment within the preset time zone is output.

3. The SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system according to claim 1, characterized in that, Using the fault prediction time limit as the optimization constraint and maximizing fault detection accuracy and efficiency as the dual optimization objectives, the monitoring index fusion scheme for slippage fault prediction is iteratively optimized and searched to obtain the optimal monitoring index fusion scheme, including: Obtain a preset monitoring index set for slippage fault prediction, wherein the preset monitoring index includes drive equipment index, transmission equipment index, conveyor belt monitoring index and sensor type, and the sensor type includes at least coal pile sensor, gas sensor, dust sensor and temperature and humidity sensor. Starting with the initial number of monitoring indicators, the monitoring indicator fusion schemes are enumerated using the preset monitoring indicator set. The power set of the preset monitoring indicator set is used as multiple initial monitoring indicator fusion scheme sets, wherein the initial number of monitoring indicators is 5. A fault accuracy predictor and a fault prediction duration analyzer are constructed based on generative adversarial networks. Using the fault prediction duration analyzer, multiple fault prediction durations are obtained based on the multiple initial monitoring indicator fusion scheme sets. Several qualified monitoring indicator fusion scheme sets are then selected with the constraint that the duration is less than the fault prediction time limit. Using the fault accuracy predictor, several prediction accuracies are obtained by analyzing the fusion scheme set of several qualified monitoring indicators; With the dual objectives of maximizing fault detection accuracy and efficiency, the monitoring index fusion scheme for slippage fault prediction is iteratively optimized and searched based on several fault prediction durations and several prediction accuracies to obtain the optimal monitoring index fusion scheme.

4. The SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system according to claim 3, characterized in that, To obtain the best monitoring indicator fusion solution, including: The reciprocal of the prediction time limit correction coefficient is set as the prediction time limit sensitivity, and the product of the prediction time limit sensitivity and the initial efficiency weight is set as the adaptation efficiency weight, wherein the initial efficiency weight is 0.4, and the adaptation efficiency weight is not less than 0.2 and not greater than 0.

6. The adaptation accuracy weight is obtained by subtracting the adaptation efficiency weight from 1. With the dual optimization objectives of maximizing fault detection accuracy and efficiency, based on the adaptation efficiency weight and adaptation accuracy weight, after dimensionless processing of the several fault prediction durations and several prediction accuracies, several fusion scheme quality coefficients are obtained by weighted calculation. Among them, the fusion scheme quality coefficient is negatively correlated with the fault prediction duration and positively correlated with the prediction accuracy. The optimal monitoring index fusion scheme is selected as the fusion scheme corresponding to the qualified monitoring index with the highest fusion scheme quality coefficient.

5. The SDC-enabled industrial heterogeneous equipment IoT data fusion and analysis system according to claim 1, characterized in that, The optimal monitoring indicator fusion scheme is deployed to the equipment abstraction layer to perform data acquisition, data uploading, and data fusion, and belt slippage fault prediction is performed on the cloud-based production management platform, including: The optimal monitoring indicator fusion scheme is deployed to the equipment abstraction layer. Data is collected and uploaded regularly according to the monitoring indicators in the optimal monitoring indicator fusion scheme. After data fusion on the cloud production management platform, a monitoring fusion dataset is obtained. On the cloud-based production management platform, an adaptive fault risk predictor is obtained by matching the optimal monitoring indicator fusion scheme, and belt slippage fault risk is predicted based on the monitoring fusion dataset.

Images