Coal mine underground explosion-proof electrical equipment state monitoring system

Abstract

The invention belongs to the technical field of coal mine safety monitoring, and provides a coal mine underground explosion-proof electrical equipment state monitoring system, which comprises a monitoring acquisition module, a period adjustment module, a confidence evaluation and virtual reconstruction module and a state evaluation module, and is provided with an integrated monitoring network comprising a main sensor group and a microenvironment sensor group, the method comprises the following steps: acquiring equipment state data and environment state data of explosion-proof electrical equipment, setting a credibility evaluation period, dynamically adjusting the period duration based on the environment state data, setting a credibility evaluation model, performing periodic credibility evaluation based on the acquired equipment state data, calculating a data credibility factor for each sensor in a main sensor group, and performing credibility evaluation on each sensor in the main sensor group. And judging whether a failure mark is added to the sensor or not, if so, performing virtual reconstruction on the equipment state data of the sensor, and performing quantitative calculation based on the collected or virtually reconstructed equipment state data and the corresponding credibility factor to obtain a high-confidence state parameter to perform state evaluation on the explosion-proof electrical equipment.

Description

Technical Field

[0001] This invention belongs to the field of coal mine safety monitoring technology, specifically a condition monitoring system for explosion-proof electrical equipment in underground coal mines. Background Technology

[0002] Coal mines are typically high-risk and complex environments, characterized by high concentrations of methane, dust, humidity, and severe vibrations. Explosion-proof electrical equipment, as a core component of underground power supply, ventilation, and transportation systems, directly determines the safety and efficiency of coal mine production. Faults such as insulation aging and component overheating can easily trigger major accidents like gas explosions and circuit failures. Therefore, precise condition monitoring of explosion-proof electrical equipment is a crucial aspect of ensuring safe coal mine production.

[0003] Current monitoring technology for explosion-proof electrical equipment in coal mines has significant shortcomings. Firstly, the monitoring dimensions are limited. Existing systems mostly collect parameters such as equipment temperature and current, ignoring the impact of micro-environmental factors like temperature, humidity, and methane concentration on equipment reliability and data acquisition accuracy. A comprehensive equipment-environment monitoring network has not been formed. Furthermore, some sensors lack explosion-proof characteristics and are prone to failure in harsh underground environments. Data transmission relies heavily on common communication protocols, resulting in weak anti-interference capabilities and low data reliability. Secondly, the evaluation mechanism is rigid. In harsh environments, the evaluation frequency cannot be increased, increasing the risk of missed detections. In stable environments, continuous high-frequency evaluation wastes computational resources. Thirdly, data processing capabilities are insufficient. There is a lack of multi-dimensional reliability evaluation models based on historical consistency, environmental coupling, and physical logic. There is no virtual data reconstruction mechanism after sensor failure, leading to data interruptions, delayed fault warnings, and high false alarm / missed alarm rates. This makes it difficult to meet the actual needs of real-time and reliable monitoring of explosion-proof electrical equipment in underground mines. Finally, there is a lack of intelligent scheduling systems that deeply integrate with condition monitoring. Current scheduling relies heavily on preset plans or manual intervention, failing to dynamically optimize and adjust based on real-time monitoring data.

[0004] To address the above problems, this invention proposes a condition monitoring system for explosion-proof electrical equipment in coal mines. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0006] The technical solution adopted by this invention to solve the technical problem is: a condition monitoring system for explosion-proof electrical equipment in coal mines, comprising:

[0007] Monitoring and data acquisition module: Set up an integrated monitoring network including a main sensor group and a micro-environment sensor group to collect equipment status data and environmental status data of explosion-proof electrical equipment;

[0008] Periodic adjustment module: Sets the credibility assessment period and adaptively adjusts the period duration based on environmental status data;

[0009] Confidence assessment and virtual reconstruction module: Set up a confidence assessment model, adopt a confidence assessment cycle, perform periodic confidence assessment based on the collected equipment status data, calculate the data confidence factor for each sensor in the main sensor group, determine whether the sensor has added a failure mark, and if so, perform virtual reconstruction of the sensor's equipment status data.

[0010] Condition assessment module: Based on the collected or virtually reconstructed equipment condition data and the corresponding confidence factor, it quantitatively calculates high-confidence condition parameters to assess the condition of explosion-proof electrical equipment.

[0011] The adaptive adjustment method for the cycle duration is as follows:

[0012] Set the credibility assessment cycle and initialize the cycle length to a preset standard duration. Set the cumulative adverse environmental duration and initialize it to zero. Start and stop the cumulative adverse environmental duration based on environmental status data. When the cumulative adverse environmental duration reaches the upper limit, mark the time period from the initialization time of the cumulative adverse environmental duration to the end of the current credibility assessment cycle as the adjustment reference period. Use a linear adjustment model based on the ratio of the adjustment reference period to the cumulative adverse environmental duration. Combine the preset standard duration of the cycle length and the preset adjustment coefficient for data processing. Adaptively adjust and calculate the new cycle length to adjust the credibility assessment cycle, and reset the cumulative adverse environmental duration to zero.

[0013] The method for starting and stopping the cumulative timekeeping of severe environmental conditions is as follows:

[0014] Real-time acquisition of environmental status data is compared with the corresponding safe operation thresholds set based on coal mine safety regulations and standards. If any environmental status data exceeds the corresponding safe operation threshold, the cumulative duration of severe environmental conditions is started to accumulate; otherwise, the accumulation timer is stopped.

[0015] The data credibility factor is calculated as follows:

[0016] A credibility assessment model integrating multiple evaluation dimensions is adopted, including historical consistency dimension, environmental coupling dimension and physical logic dimension. At the end of each credibility assessment cycle, multi-dimensional factor scores are calculated for the sensors in the main sensor group to obtain the historical consistency factor score, environmental coupling factor score and physical logic factor score of the sensor. The corresponding weight coefficients are assigned for weighted fusion calculation and normalization to obtain the data credibility factor of the sensor.

[0017] The historical consistency factor score is calculated as follows:

[0018] The device status data collected by the sensors within the credibility assessment period ending at the current time is obtained, integrated according to the time series to obtain the historical data sequence, the standard deviation ratio and mean deviation of the device status data at the current time and the historical data sequence are calculated, and the standard deviation ratio and mean deviation are mapped to sub-factor scores between 0 and 1 through the exponential decay function and then weighted and fused to obtain the historical consistency factor score.

[0019] The environmental coupling factor score is calculated as follows:

[0020] An environmental impact quantification model is constructed and adopted. The environmental impact quantification model takes environmental state data as input features and equipment state data as output labels. The impact coefficients of each environmental state data are obtained by fitting with the least squares method to form an environmental impact transmission matrix. The environmental state data at the current moment is input into the environmental impact quantification model. The theoretical expected value of the equipment state data at the current moment is calculated by matrix operation and linear superposition. The matching degree between the equipment state data at the current moment and the theoretical expected value is calculated by the relative deviation algorithm and quantified and normalized to obtain the environmental coupling factor score.

[0021] The physical logic factor score is calculated as follows:

[0022] A physical relationship model is constructed and adopted. The physical relationship model is based on the principles of energy conservation, electrical characteristics and mechanical dynamics. It identifies the associated sensors of the sensor and establishes the mathematical relationship between the device status data collected by the sensor and the associated sensors. For the sensor being evaluated, the associated sensors of the sensor are identified through the physical relationship model. The device status data collected by the associated sensors is input into the physical relationship model. The theoretical prediction value of the device status data of the sensor being evaluated is output. The relative error between the device status data collected by the sensor at the current moment and the theoretical prediction value is calculated. The physical logic factor score is obtained by quantization and normalization through the error inverse ratio function.

[0023] The method for determining whether a sensor has a failure flag is as follows:

[0024] Each sensor in the main sensor group is assigned a reconstruction flag value and its initial value is set to 0. At the end of each confidence assessment cycle, the data confidence factor of each sensor is obtained. If the sensor data confidence factor does not meet the confidence standard, the reconstruction flag value of the corresponding sensor is incremented once. When the reconstruction flag value of any sensor reaches the preset reconstruction threshold, a failure mark is added to the sensor.

[0025] The virtual reconstruction method is as follows:

[0026] Sensors with failure markers are defined as failed sensors. For any failed sensor, a pre-built physical relationship model is used to identify the associated sensors of the failed sensor.

[0027] If there is no failed sensor among the associated sensors, the equipment status data of the associated sensors are input into the physical relationship model, the theoretical predicted value of the equipment status data of the failed sensor is output, the equipment status data collected by the failed sensor is updated and replaced with the calculated theoretical predicted value, and the data confidence factor of the failed sensor is updated to the average data confidence factor of the associated sensors, thus completing the virtual reconstruction. If there is a failed sensor among the associated sensors, the failed sensor in the associated sensors is given priority for virtual reconstruction, and the equipment status data after virtual reconstruction is used as the input of the physical relationship model for calculation.

[0028] The high-confidence state parameter is calculated as follows:

[0029] A confidence-weighted data fusion algorithm is adopted to perform weighted fusion calculation on the device status data collected by each similar sensor in the main sensor group and the corresponding data confidence factor. The weights are proportional to the data confidence factor. After normalization, the weights are summed to one, and the high confidence status parameters are calculated.

[0030] The beneficial effects of this invention are as follows:

[0031] 1. This invention comprehensively improves the reliability of monitoring data and the accuracy of status assessment by introducing a dynamic reliability assessment and data virtual reconstruction mechanism. The system innovatively constructs a multi-dimensional reliability assessment model that integrates historical consistency, environmental coupling, and physical logic. It can accurately quantify the data reliability factor of each sensor and effectively identify sensor data distortion caused by environmental interference or self-degradation. Crucially, when the system determines that a sensor may fail, it can immediately trigger data virtual reconstruction based on a physical relationship model. It intelligently calculates replacement values ​​using reliable data from associated sensors, ensuring the continuous and uninterrupted flow of key monitoring data. This makes the comprehensive parameters used for status assessment highly confident, providing a solid and reliable data foundation for fault early warning and diagnosis, and fundamentally avoiding system misjudgment or monitoring function paralysis caused by a single sensor failure.

[0032] 2. This invention significantly enhances the system's adaptability, robustness, and maintenance efficiency by dynamically adjusting strategies based on environmental perception and achieving intelligent early warning. The system can adaptively shorten or extend the reliability assessment cycle according to the actual underground environmental quality, optimizing the allocation of computing resources while ensuring timely assessment. Furthermore, combined with the precise location alarm and data reconstruction capabilities of failure sensors, the system can maintain complete functionality even when some hardware fails, demonstrating excellent fault tolerance and robustness. Finally, based on a multi-level alarm mechanism with high-confidence state parameters, it can provide managers with clearly defined and precisely located early warning information, realizing a shift from passively responding to faults to proactively predicting risks, greatly improving the level of intelligence and maintenance efficiency of underground equipment safety management in coal mines. Attached Figure Description

[0033] The invention will now be further described with reference to the accompanying drawings.

[0034] Figure 1 This is a system module architecture diagram of a coal mine underground explosion-proof electrical equipment condition monitoring system according to an embodiment of the present invention;

[0035] Figure 2 This is a flowchart illustrating the specific steps of a coal mine underground explosion-proof electrical equipment condition monitoring system according to an embodiment of the present invention. Detailed Implementation

[0036] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0037] Example 1

[0038] Please see Figure 1 and Figure 2 As shown in the embodiment of the present invention, a condition monitoring system for explosion-proof electrical equipment in coal mines includes the following modules:

[0039] Monitoring and data acquisition module: Set up an integrated monitoring network including a main sensor group and a micro-environment sensor group to collect equipment status data and environmental status data of explosion-proof electrical equipment;

[0040] An integrated monitoring network is set up for monitoring the condition of explosion-proof electrical equipment in underground coal mines. The integrated monitoring network includes a main sensor group and a micro-environment sensor group. The main sensor group and the micro-environment sensor group collect data respectively to obtain the equipment status data and environmental status data of the explosion-proof electrical equipment.

[0041] Specifically, at the actual installation location of the explosion-proof electrical equipment, monitoring points are planned based on the equipment structure and operating characteristics of the explosion-proof electrical equipment, and the physical deployment of the main sensor group and the micro-environment sensor group is set according to the monitoring points;

[0042] The main sensor group integrates multiple types of high-precision industrial sensors, including vibration acceleration sensors, temperature sensors, current transformers and voltage sensors. These sensors are directly fixed and installed at the corresponding monitoring points on the outer shell and internal core components of the explosion-proof electrical equipment. They are used to collect equipment status data that reflects the operating status of the explosion-proof electrical equipment in real time. The equipment status data includes vibration amplitude, temperature value, current value and voltage value.

[0043] Among them, the micro-environment sensor group integrates special sensors for the working environment around the equipment, including explosion-proof temperature and humidity sensors, methane concentration sensors, dust concentration sensors and air pressure sensors. These sensors are deployed at the corresponding monitoring points planned inside the equipment cabin and in the key ventilation paths to collect environmental status data in real time. The environmental status data includes ambient temperature, ambient humidity, methane concentration, dust concentration and ambient air pressure.

[0044] It should be noted that environmental condition data affects both the reliability of explosion-proof electrical equipment and the reliability of data acquisition from the main sensor array.

[0045] For the integrated monitoring network, an industrial communication protocol based on CAN bus is adopted. The main sensor group and the micro-environment sensor group are connected to the intrinsically safe data acquisition module. The intrinsically safe data acquisition module integrates signal conditioning circuit and analog-to-digital converter to perform anti-interference processing and digital sampling on the raw signals collected by the sensors in the main sensor group and the micro-environment sensor group. A unified acquisition frequency is set, and the acquisition period and acquisition time are determined based on the acquisition frequency. The time interval between adjacent acquisition times is one acquisition period. Based on the set acquisition frequency, the integrated monitoring network synchronously triggers the main sensor group and the micro-environment sensor group to perform real-time data acquisition at each acquisition time, and obtains equipment status data and environmental status data respectively, which are then uploaded to the well surface monitoring center in real time.

[0046] It should be noted that the purpose of this step is to build a three-dimensional monitoring network covering the explosion-proof electrical equipment itself and the surrounding environment, providing a comprehensive and real-time data source for the entire condition monitoring system. By selecting explosion-proof, high-precision sensors and adopting industrial-grade communication protocols, the stability, accuracy and reliability of data acquisition in the harsh environment of underground coal mines are ensured, providing a solid perception foundation for subsequent data credibility assessment and condition diagnosis.

[0047] Periodic adjustment module: Sets the credibility assessment period and adaptively adjusts the period duration based on environmental status data;

[0048] Set the credibility assessment period and initialize the period duration to the preset standard duration;

[0049] The cumulative duration of severe environmental conditions is set and initialized to zero. The environmental quality is evaluated in real time based on the environmental status data collected by the micro-environment sensor group. Specifically, at each collection moment, the environmental status data uploaded by the micro-environment sensor group is acquired in real time. Based on the coal mine safety regulations and standards, corresponding safe operation thresholds are set for each environmental status data. The real-time collected environmental status data is compared with the corresponding safe operation thresholds. If any environmental status data exceeds the corresponding safe operation threshold, the environmental quality at the current collection moment is judged to be severe; otherwise, the environmental quality at the current collection moment is judged to be normal.

[0050] When the environmental quality is poor, a cumulative timer is started to record the duration of the poor environmental quality. When the environmental quality is normal, the cumulative timer for the poor environmental quality stops.

[0051] The cumulative duration of environmental adverse events is compared with the preset upper limit of adverse event duration in real time. When the cumulative duration of environmental adverse events reaches the upper limit of adverse event duration, an adaptive adjustment mechanism for the duration of the current credibility assessment period is triggered at the end of the current period.

[0052] Specifically, the period from the initialization of the cumulative adverse environmental duration to the end of the current credibility assessment cycle is marked as the adjustment reference period. A linear adjustment model based on the ratio of the cumulative adverse environmental duration to the duration of the adjustment reference period is adopted. Combined with the preset standard duration of the cycle duration and the preset adjustment coefficient, data processing is performed, and the new cycle duration is adaptively adjusted and calculated. The new cycle duration = preset standard duration × [1 + preset adjustment coefficient × (0.5 - cumulative adverse environmental duration / duration of the adjustment reference period)];

[0053] The ratio of the cumulative duration of severe environmental conditions to the duration of the adjustment reference period represents the percentage of time during which the environment is in a severe state. 0.5 represents the equilibrium point of the percentage. When the percentage exceeds the equilibrium point, the cycle duration is shortened; when the percentage does not exceed the equilibrium point, the cycle duration is increased.

[0054] It should be noted that the longer the adjustment reference period, the longer the period of normal environmental quality, and the more stable the data collection reliability. By adaptively adjusting, the evaluation frequency can be reduced, thus reducing the waste of computing resources. Conversely, the shorter the adjustment reference period, the more unstable the data collection reliability. Increasing the evaluation frequency can prevent excessive deviation in the evaluation of data collection reliability within the reliability evaluation period.

[0055] For example, when the adaptive adjustment mechanism is triggered, the duration of the adjustment reference period is 60 minutes, the cumulative duration of severe environmental conditions is 15 minutes, the preset standard duration of the cycle is 60 minutes, the preset adjustment coefficient is 1.2, and the new cycle duration is calculated to be 78 minutes.

[0056] It should be noted that when the environmental severity is high, the time required for the natural decline back to the normal range will be longer. The total time required for the cumulative environmental severity duration to reach the upper limit is shorter. When the cumulative environmental severity duration is the same as the duration of the adjustment reference period, the cycle duration reaches the adjustable upper limit regardless of the environmental severity. In addition, the cumulative environmental severity duration assesses the cumulative deviation risk caused by continuous environmental interference to the reliability of data collection. The goal is not an immediate danger alarm, but an adjustment to the reliability assessment duration. For example, continuous high temperature and humidity may cause equipment drift and circuit aging, affecting long-term reliability. Long-term dust exceeding the standard may slowly contaminate the optical sensor window, causing the reading to gradually become inaccurate. Although a stable high concentration of methane has not reached the explosion limit, its continued existence means the accumulation of risks and the potential decline in the reliability of sensor data.

[0057] The credibility assessment period is updated to the adaptively adjusted period, and the cumulative adverse environmental duration is reset to zero to restart the cumulative timing. This means that from the credibility assessment period starting from the current moment, the updated period will be used until the next adaptive adjustment and update of the period.

[0058] It should be noted that the purpose of this step is to set the credibility assessment cycle and dynamically adjust the cycle length based on environmental status data. By monitoring the severity of environmental conditions in real time, the credibility assessment frequency is adaptively adjusted to optimize system resource allocation and ensure the timeliness of data credibility assessment. When the environmental quality is poor, the assessment frequency is increased to prevent the accumulation of data collection credibility assessment deviations caused by environmental factors and improve the accuracy of system assessment. When the environmental quality is normal, the assessment frequency is reduced to reduce computing resource consumption and improve system operating efficiency. This achieves intelligent adaptive adjustment of the credibility assessment cycle. By linking real-time environmental quality assessment with cycle length, the shortcomings of a fixed assessment cycle under sudden environmental changes are overcome, and the system's responsiveness to harsh underground environments is improved.

[0059] Confidence assessment and virtual reconstruction module: Set up a confidence assessment model, adopt a confidence assessment cycle, perform periodic confidence assessment based on the collected equipment status data, calculate the data confidence factor for each sensor in the main sensor group, determine whether the sensor has added a failure mark, and if so, perform virtual reconstruction of the sensor's equipment status data.

[0060] Specifically, the credibility assessment model integrates multiple assessment dimensions, including historical consistency dimension, environmental coupling dimension, and physical logic dimension. At the end of each credibility assessment cycle, for any sensor in the main sensor group, multi-dimensional factor scores are calculated based on the device status data collected by the sensor and the environmental status data collected by the micro-environment sensor group.

[0061] For the historical consistency dimension, the consistency of the time series of device status data collected by sensors is quantitatively calculated. The device status data collected by sensors within the credibility assessment period ending at the current time is obtained and integrated according to the time series to obtain the assessment data sequence. All assessment data sequences within a preset duration ending at the current time are obtained and integrated to obtain the historical data sequence. The standard deviations of the current assessment data sequence and the historical data sequence are calculated and the ratio is processed to obtain the standard deviation ratio. The mean and absolute difference of the current assessment data sequence and the historical data sequence are calculated and the absolute difference is processed and the standard deviation of the historical data sequence is processed to obtain the mean deviation. The standard deviation ratio and the mean deviation are mapped to sub-factor scores between 0 and 1 through an exponential decay function and weighted fusion processing is performed to obtain the historical consistency factor score.

[0062] For the environmental coupling dimension, the impact of environmental state data on equipment state data is quantitatively calculated. A pre-constructed environmental impact quantification model is adopted, which is based on long-term historical monitoring data of the main sensor group and the micro-environment sensor group. The mathematical mapping relationship between environmental state data and equipment state data is established by multiple linear regression analysis. Environmental state data is used as input features and equipment state data is used as output labels. The impact coefficient of each environmental state data is obtained by fitting with the least squares method to form an environmental impact transmission matrix. The environmental state data collected by the micro-environment sensor group at the current moment is input into the pre-constructed environmental impact quantification model. Through matrix operations and linear superposition, the theoretical expected value of the equipment state data at the current moment is calculated. The matching degree between the equipment state data at the current moment and the theoretical expected value is calculated by the relative deviation algorithm and quantified as an environmental coupling factor score between 0 and 1.

[0063] For example, taking the data reliability assessment of the winding temperature sensor of an explosion-proof motor as an example, the environmental impact quantification model is constructed as follows: Based on the historical data of the long-term operation of the explosion-proof motor, through multiple linear regression analysis, the mathematical relationship between the winding temperature in the equipment status data and the ambient temperature, ambient humidity, and methane concentration in the environmental status data is established. The fitted environmental impact transfer matrix is: theoretical expected value of winding temperature = 25 + 0.8 × ambient temperature - 0.1 × ambient humidity + 0.05 × methane concentration, where 25 is the basic temperature constant term, and 0.8, -0.1, and 0.05 are the influence coefficients of ambient temperature, ambient humidity, and methane concentration, respectively. The theoretical expected value is calculated as follows: at the current acquisition time, the environmental status data collected by the micro-environment sensor group is: ambient temperature = 40°C, ambient humidity = 60%RH, and methane concentration = 0.8%VOL. Substituting this data into the above model: The theoretical expected value of winding temperature = 25 + 0.8 × 40 - 0.1 × 60 + 0.05 × 0.8 = 25 + 32 - 6 + 0.04 = 51.04°C, and the actual value collected by the winding temperature sensor in the main sensor group is 53.5°C. The matching degree is calculated using the relative deviation algorithm, and the relative deviation = |53.5 - 51.04| / 51.04 ≈ 0.048. Through the preset matching degree function: matching degree = exp(-k × relative deviation), where k is the sensitivity coefficient, and k = 10 is used for quantification: environmental coupling factor score = exp(-10 × 0.048) ≈ exp(-0.48) ≈ 0.62;

[0064] For the physical logic dimension, quantitative calculations are performed based on the working principle of the equipment and physical laws. A pre-constructed physical relationship model is adopted. The physical relationship model is based on the principles of energy conservation, electrical characteristics and mechanical dynamics. It identifies the associated sensors of the sensor and establishes the mathematical relationship between the equipment status data collected by the sensor and the associated sensors.

[0065] Among them, associated sensors refer to other sensors that have a direct or indirect physical relationship with the sensor being evaluated in terms of data. For example, for a current sensor, associated sensors include voltage sensors; for a vibration acceleration sensor, associated sensors include current sensors and temperature sensors. The mathematical association is a mathematical equation between the device state data. For example, for current sensors and voltage sensors, according to the law of power conservation, current value = power / voltage value; for vibration sensors and temperature sensors, according to mechanical dynamics, vibration amplitude is positively correlated with temperature, and the theoretical predicted value of vibration amplitude = temperature influence coefficient × current temperature measurement value + vibration reference value. The temperature influence coefficient and vibration reference value are calculated by fitting historical data through a linear model.

[0066] For the sensor currently being evaluated, the associated sensors are identified through the physical relationship model. The device status data collected by the associated sensors is input into the physical relationship model, and the theoretical prediction value of the device status data of the sensor currently being evaluated is output. The relative error between the device status data collected by the sensor at the current moment and the theoretical prediction value is calculated and quantified into a physical logic factor score between 0 and 1 through the error inverse ratio function.

[0067] Weighting coefficients are set for the historical consistency factor score, environmental coupling factor score, and physical logic factor score. The weighting coefficients are configured based on the degree of influence of each evaluation dimension on data credibility. The credibility factor score of the sensor is obtained by weighted fusion calculation. The credibility factor score is normalized to obtain the data credibility factor of the sensor.

[0068] Each sensor in the main sensor group is assigned a reconstruction flag value and its initial value is set to 0. At the end of each confidence assessment cycle, the data confidence factor of each sensor is obtained. The data confidence factor is compared with the preset confidence standard in real time. If the sensor data confidence factor does not meet the confidence standard, the reconstruction flag value of the corresponding sensor is incremented once.

[0069] The system continuously monitors the reconstruction flag values ​​of each sensor. When the reconstruction flag value of any sensor reaches the preset reconstruction threshold, the sensor is determined to be in a potential failure state. A failure flag is added to the sensor, and the sensor with the failure flag is defined as a failed sensor. A location alarm is triggered for the failed sensor, the failed sensor is located, and a failure alarm is generated and sent to the administrator terminal to notify the repair. Until the repair is completed, the failure flag of the sensor is cleared and the corresponding reconstruction flag value is initialized to 0.

[0070] It should be noted that triggering the location alarm is done in parallel with subsequent steps;

[0071] For virtual reconstruction of data triggered by failed sensors, specifically, for any failed sensor, a pre-built physical relationship model is used to identify the associated sensors of the failed sensor;

[0072] If there is no failed sensor among the associated sensors, the equipment status data of the associated sensors are input into the physical relationship model, the theoretical predicted value of the equipment status data of the failed sensor is output, the equipment status data collected by the failed sensor is updated and replaced with the theoretical predicted value, and the data confidence factor of the failed sensor is updated to the average data confidence factor of the associated sensors, thus completing the virtual reconstruction.

[0073] If there is a failed sensor among the associated sensors, the failed sensor among the associated sensors will be virtually reconstructed first. The virtually reconstructed device status data will be used as the input of the physical relationship model to virtually reconstruct the current failed sensor. If no failed sensor can be virtually reconstructed first, that is, no associated sensor without a failed sensor can be found, the main sensor group is judged to have failed and the subsequent steps will be stopped until the repair is completed and the main sensor group's acquisition effect is restored.

[0074] It should be noted that the purpose of this step is to comprehensively evaluate the reliability of sensor data through multi-dimensional factor score calculation, improve the accuracy and timeliness of identifying failed sensors, trigger location alarms and virtual data reconstruction for failed sensors, maintain normal system operation when sensors are potentially failed, reduce data interruption and equipment downtime caused by sensor failures, and the virtual reconstruction mechanism is based on physical relationship models and associated sensor data to replace failed sensor data, ensure the continuity of status monitoring, improve system fault tolerance, introduce reconstruction flag values ​​and reconstruction threshold mechanisms to realize progressive judgment and early warning of sensor failure, and provide software redundancy when hardware fails, combined with virtual reconstruction technology, enhance the intelligent maintenance and self-repair characteristics of the system.

[0075] Condition assessment module: Based on the collected or virtually reconstructed equipment condition data and the corresponding confidence factor, it quantitatively calculates high-confidence condition parameters to assess the condition of explosion-proof electrical equipment.

[0076] Specifically, a confidence-weighted data fusion algorithm is adopted to perform weighted fusion calculation on the equipment status data collected by each similar sensor in the main sensor group and the corresponding data confidence factor. The weights are proportional to the data confidence factor, and normalization is performed to ensure that the sum of each weight is one. Based on the weighted fusion calculation, high confidence status parameters reflecting the true operating status of the equipment are obtained, including comprehensive temperature value, comprehensive vibration amplitude value, comprehensive current value and comprehensive voltage value.

[0077] Referring to the relevant standards of the National Coal Mine Safety Regulations, multi-level alarm thresholds are set for high-confidence status parameters, including normal threshold, attention threshold, warning threshold and danger threshold, forming a graded alarm judgment standard. Based on the high-confidence status parameters calculated in real time, the high-confidence status parameters are compared with the corresponding multi-level alarm thresholds in real time to determine whether the equipment status warning is triggered.

[0078] If a high-confidence status parameter exceeds the normal threshold, the warning level is automatically assessed based on the degree of exceedance. When the parameter is within the attention threshold range, it is marked as a Level 1 warning; when the parameter is within the warning threshold range, it is marked as a Level 2 warning; and when the parameter reaches or exceeds the danger threshold, it is marked as a Level 3 emergency warning. Simultaneously, the faulty equipment is accurately located, and detailed warning information including the equipment number, installation location, warning level, and occurrence time is generated and sent to the administrator terminal in real time through an intrinsically safe communication network.

[0079] The technical solution of this invention is as follows: An integrated monitoring network including a main sensor group and a micro-environment sensor group is set up to collect equipment status data and environmental status data of explosion-proof electrical equipment. A reliability assessment cycle is set and its duration is dynamically adjusted based on the environmental status data. A reliability assessment model is set up, and a reliability assessment cycle is used to perform periodic reliability assessments based on the collected equipment status data. A data reliability factor is calculated for each sensor in the main sensor group. It is determined whether a failure marker has been added to the sensor. If so, the equipment status data of the sensor is virtually reconstructed. Based on the collected or virtually reconstructed equipment status data and the corresponding reliability factor, high-confidence status parameters are quantitatively calculated to assess the status of the explosion-proof electrical equipment.

[0080] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A condition monitoring system for explosion-proof electrical equipment in coal mines, characterized in that: include: Monitoring and data acquisition module: Set up an integrated monitoring network including a main sensor group and a micro-environment sensor group to collect equipment status data and environmental status data of explosion-proof electrical equipment; Periodic adjustment module: Sets the credibility assessment period and adaptively adjusts the period duration based on environmental status data; Confidence assessment and virtual reconstruction module: Set up a confidence assessment model, adopt a confidence assessment cycle, perform periodic confidence assessment based on the collected equipment status data, calculate the data confidence factor for each sensor in the main sensor group, determine whether the sensor has added a failure mark, and if so, perform virtual reconstruction of the sensor's equipment status data. Condition assessment module: Based on the collected or virtually reconstructed equipment condition data and the corresponding confidence factor, high-confidence condition parameters are quantitatively calculated to assess the condition of explosion-proof electrical equipment.

2. The condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 1, characterized in that: The adaptive adjustment method for cycle duration is as follows: The credibility assessment period is set and its duration is initialized to a preset standard duration. The cumulative adverse environmental duration is set and initialized to zero. The cumulative adverse environmental duration is started and stopped based on environmental status data. When the cumulative adverse environmental duration reaches the upper limit, the time period from the initialization time of the cumulative adverse environmental duration to the end of the current credibility assessment period is marked as the adjustment reference period. A linear adjustment model based on the ratio of the adjustment reference period duration to the cumulative adverse environmental duration is adopted. Combined with the preset standard duration of the period and the preset adjustment coefficient, data processing is performed to adaptively adjust and calculate a new period duration to adjust the credibility assessment period. Finally, the cumulative adverse environmental duration is reset to zero.

3. The condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 2, characterized in that: The method for starting and stopping the cumulative timekeeping of severe environmental conditions is as follows: Real-time acquisition of environmental status data is compared with the corresponding safe operation thresholds set based on coal mine safety regulations and standards. If any environmental status data exceeds the corresponding safe operation threshold, the cumulative duration of severe environmental conditions is started to accumulate; otherwise, the accumulation is stopped.

4. The condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 1, characterized in that: The data credibility factor is calculated as follows: A credibility assessment model integrating multiple evaluation dimensions is adopted, including historical consistency dimension, environmental coupling dimension, and physical logic dimension. At the end of each credibility assessment cycle, multi-dimensional factor scores are calculated for the sensors in the main sensor group to obtain the historical consistency factor score, environmental coupling factor score, and physical logic factor score of the sensor. Corresponding weight coefficients are assigned for weighted fusion calculation and normalization to obtain the data credibility factor of the sensor.

5. The condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 4, characterized in that: The historical consistency factor score is calculated as follows: The system acquires device status data collected by sensors within the credibility assessment period ending at the current time, integrates it according to the time series to obtain a historical data sequence, calculates the standard deviation ratio and mean offset between the device status data at the current time and the historical data sequence, maps the standard deviation ratio and mean offset to sub-factor scores between 0 and 1 respectively through an exponential decay function, and performs weighted fusion processing to obtain the historical consistency factor score.

6. The condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 4, characterized in that: The environmental coupling factor score is calculated as follows: An environmental impact quantification model is constructed and adopted. The environmental impact quantification model uses environmental state data as input features and equipment state data as output labels. The impact coefficients of each environmental state data are obtained by fitting using the least squares method, forming an environmental impact transmission matrix. The environmental state data at the current moment is input into the environmental impact quantification model. The theoretical expected value of the equipment state data at the current moment is calculated by matrix operations and linear superposition. The matching degree between the equipment state data at the current moment and the theoretical expected value is calculated by the relative deviation algorithm and quantified and normalized to obtain the environmental coupling factor score.

7. The condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 4, characterized in that: The physical logic factor score is calculated as follows: A physical relationship model is constructed and adopted. Based on the principles of energy conservation, electrical characteristics, and mechanical dynamics, the physical relationship model identifies the associated sensors of the sensor and establishes the mathematical relationship between the device status data collected by the sensor and the associated sensors. For the sensor currently being evaluated, the associated sensors of the sensor are identified through the physical relationship model. The device status data collected by the associated sensors is input into the physical relationship model, and the theoretical prediction value of the device status data of the sensor currently being evaluated is output. The relative error between the device status data collected by the sensor at the current moment and the theoretical prediction value is calculated, and the physical logic factor score is obtained by quantization and normalization through the inverse error ratio function.

8. The condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 1, characterized in that: The method for determining whether a sensor has a failure flag is as follows: Each sensor in the main sensor group is assigned a reconstruction flag value with an initial value of 0. At the end of each confidence assessment cycle, the confidence factor of each sensor is obtained. If the confidence factor of the sensor data does not meet the confidence standard, the reconstruction flag value of the corresponding sensor is incremented. When the reconstruction flag value of any sensor reaches the preset reconstruction threshold, a failure flag is added to the sensor.

9. A condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 7, characterized in that: The virtual reconstruction method is as follows: Sensors with failure markers are defined as failed sensors. For any failed sensor, a pre-built physical relationship model is used to identify the associated sensors of the failed sensor. If there are no failed sensors among the associated sensors, the equipment status data of the associated sensors are input into the physical relationship model, the theoretical predicted value of the equipment status data of the failed sensor is output, the equipment status data collected by the failed sensor is updated and replaced with the calculated theoretical predicted value, and the data confidence factor of the failed sensor is updated to the average data confidence factor of the associated sensors, thus completing the virtual reconstruction. If there are failed sensors among the associated sensors, the failed sensors among the associated sensors are given priority for virtual reconstruction, and the equipment status data after virtual reconstruction is used as the input of the physical relationship model for calculation.

10. A condition monitoring system for explosion-proof electrical equipment in coal mines according to claim 1, characterized in that: The high-confidence state parameter is calculated as follows: A confidence-weighted data fusion algorithm is adopted to perform weighted fusion calculation on the device status data collected by each similar sensor in the main sensor group and the corresponding data confidence factor. The weights are proportional to the data confidence factor. After normalization, the weights are summed to one, and the high confidence status parameters are calculated.

Images