Intelligent equipment management method and system

Abstract

The invention relates to the field of control, in particular to an intelligent management method and system for equipment, and the method comprises the steps: collecting data in a performance test process of the equipment in real time, and carrying out the preprocessing of the obtained data; calculating a time sequence fault risk index of the real-time equipment based on the preprocessed data; normalizing the time sequence fault risk index and taking the normalized time sequence fault risk index as an anomaly weight, improving an LOF anomaly detection algorithm, and calculating a real-time anomaly score of the equipment according to the improved LOF anomaly detection algorithm; and obtaining an equipment performance test result based on the abnormal score, and making an equipment management strategy according to the equipment test result. According to the method, through multi-dimensional data fusion and time sequence analysis, the operation state of the equipment can be comprehensively and dynamically evaluated, potential fault risks can be found in time, and powerful support is provided for quality control and reliability improvement of the equipment.

Description

Technical Field

[0001] The present invention relates to the field of control, and more particularly to a method and system for intelligent equipment management. Background Art

[0002] Smart mines are the realization of intelligent, information-based, and automated processes in mine production, management, and safety through technologies such as the Internet of Things, big data, artificial intelligence, and automated control. Their goal is to improve mine production efficiency, reduce costs, enhance safety, minimize environmental pollution, and achieve sustainable development.

[0003] During the construction of smart mines, effective equipment management is a key component in ensuring efficient and safe mine operations. Real-time monitoring and analysis of mining equipment performance parameters not only enables timely detection of equipment anomalies and prevents equipment failures, but also optimizes equipment maintenance strategies, extending equipment lifespan and reducing operating costs, thereby achieving full lifecycle management of equipment.

[0004] The Local Outlier Factor (LOF) algorithm is an unsupervised anomaly detection method based on relative density. Its core idea is to identify anomalies by comparing the local density of the target object and the data points in its neighborhood. In equipment status monitoring scenarios, the performance parameters of the equipment have strong spatiotemporal correlations under normal operating conditions. That is, the data points not only cluster in space (multidimensional feature space) but also show a certain regularity in time. The LOF algorithm mainly detects anomalies based on local density differences in space (multidimensional feature space) and cannot capture the changing trends of data points in the time series. For gradual faults (such as insulation degradation), their characteristics may gradually appear in the time series, but there is no immediate obvious difference in spatial density. This makes it difficult for the LOF algorithm to detect such faults, thereby reducing detection accuracy. Summary of the Invention

[0005] To address the technical problem that the LOF algorithm has certain limitations, resulting in insufficient detection accuracy of device performance and even misjudging it as normal, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a method for intelligent device management includes: Collect data during the equipment performance test in real time and pre-process the acquired data; the data includes voltage and current data of the equipment; Calculating a real-time sequential failure risk index of the device based on the preprocessed data; the sequential failure risk index is positively correlated with the thermal power comprehensive failure index and the real-time voltage weight; Normalizing the timing fault risk index and using it as an anomaly weight, improving the LOF anomaly detection algorithm, and calculating the real-time anomaly score of the device based on the improved LOF anomaly detection algorithm; Obtain device performance test results based on anomaly scores, and formulate device management strategies based on the device test results.

[0007] This method collects real-time device performance data and calculates a sequential fault risk index based on preprocessed data. This index takes into account a comprehensive thermoelectric fault index (combining temperature and current) and real-time voltage weighting. The comprehensive thermoelectric fault index reflects the impact of current generation and the magnitude of the current on the device during operation. The voltage weighting considers the impact of voltage changes on device performance. By calculating this sequential fault risk index, the failure risk of the device at different time points can be dynamically assessed. By normalizing this index and using it as an anomaly weight, the LOF algorithm's sensitivity to gradual faults can be enhanced, improving the accuracy and reliability of fault detection.

[0008] Preferably, the data also includes temperature data of the device.

[0009] Preferably, the process of obtaining the thermoelectric comprehensive fault index includes: Obtain the device operating temperature data for the n moments before the current moment, and construct a neighboring temperature monitoring sequence for the current moment. Also obtain the device grounding current data for the n moments before the current moment, and construct a grounding current monitoring sequence for the current moment. Based on the neighboring temperature monitoring sequence and ground current monitoring sequence, the thermal fault index and cable abnormality factor of the device at the current moment are calculated respectively; The product of the thermal fault index and the cable abnormality factor is taken as the thermal and electrical comprehensive fault index of the equipment at the current moment.

[0010] By analyzing device operating temperature data, it's possible to detect overheating. Overheating is often an early sign of issues like insulation aging, poor contact, or overload. By analyzing temperature data from the previous n moments, it's possible to identify abnormal temperature fluctuations and identify potential thermal failure risks in advance. Ground current data also reflects the insulation condition of cables and the integrity of the grounding system. Abnormal changes in ground current may indicate cable insulation damage, poor grounding, or other issues. By analyzing the ground current monitoring sequence, potential cable anomalies can be detected promptly.

[0011] By integrating dual monitoring data from temperature sequences and ground current sequences, a joint analysis of thermal effects (equipment overheating) and electrical effects (current anomalies) can be achieved. Compared to single parameter evaluation, this can more comprehensively reflect the risk of complex equipment failures.

[0012] Preferably, the process of obtaining the voltage weight includes: The voltage data of the device at each moment before the current moment and the breakdown voltage of the device under normal conditions are obtained, and the ratio of the voltage data of the device at each moment to the breakdown voltage of the device under normal conditions is used as the voltage weight at the corresponding moment.

[0013] The ratio of the voltage data at each moment to the breakdown voltage is used as a voltage weight to quantify how close the device's voltage state is to its breakdown voltage at that moment. A higher voltage weight indicates that the device's voltage state is closer to its breakdown voltage at that moment, and the potential overvoltage risk is greater.

[0014] Preferably, the timing failure risk index satisfies the relationship: ; For the current moment The timing failure risk index of the device, Based on the current moment Before The comprehensive thermoelectric fault index at the moment, Based on the current moment Before The voltage weight at each moment, Based on the current moment Before Voltage data at a moment, is the breakdown voltage of the device under normal conditions, Indicates based on the current moment The number of previously selected historical moments.

[0015] Preferably, the process of obtaining the corresponding anomaly score in the improved LOF anomaly detection algorithm is as follows: The normalized timing fault risk index at the current moment is used as a weight to multiply the original LOF anomaly score to obtain the anomaly score after using the improved LOF anomaly detection algorithm at the current moment.

[0016] Preferably, obtaining the device performance test results based on the anomaly score and formulating a device management strategy according to the device test results includes: If the anomaly score calculated based on the improved LOF anomaly detection algorithm is greater than the preset anomaly detection threshold, the performance test of the device at the current moment is determined to be unqualified; otherwise, the performance test of the device at the current moment is determined to be qualified; Equipment that fails the performance test will be inspected and diagnosed to confirm the type and severity of the fault. Equipment with high severity will be shut down for inspection or temporary repair.

[0017] Preferably, the process of obtaining the thermal failure index includes: Perform trend detection on the neighboring temperature monitoring series and obtain a p-value; The device's standard temperature is preset, and for each element in the neighboring temperature monitoring sequence, its temperature deviation from the standard temperature is calculated. This temperature deviation is used as the input of the hyperbolic tangent function, and its output is positively valued. The sum of all elements in the neighboring temperature monitoring sequence is then calculated. The summation result is multiplied by the value p to obtain the thermal failure index.

[0018] First, the p-value obtained through trend detection can reflect the trend of the temperature monitoring series. If the temperature shows an upward trend, the p-value may increase, indicating that the equipment may be heading towards failure. Furthermore, by calculating the temperature deviation and processing it with the hyperbolic tangent function, the degree of temperature anomaly can be quantified into a specific value. The output range of the hyperbolic tangent function is 0 to 1, which can effectively suppress the impact of extreme temperature deviations and make the results more stable.

[0019] In summary, using the hyperbolic tangent function to process temperature deviations and summing the positive values ​​can prevent negative deviations from interfering with the results. Furthermore, multiplying by the p-value further enhances sensitivity to trend changes, making the thermal fault index calculation more stable and reducing the likelihood of false positives and negatives.

[0020] Preferably, the process of obtaining the cable abnormality factor includes: The mode of the ground current monitoring sequence is obtained, the square of the difference between each element in the ground current monitoring sequence and the mode is calculated and summed, and the summed result is divided by the mode to obtain the cable abnormality factor.

[0021] Selecting the mode (rather than the mean) can avoid interference from extreme outliers and is more suitable for describing actual operating conditions. By calculating the sum of the squares of the differences between each monitored value and the mode, the contribution of outliers can be amplified (the squaring operation is more sensitive to larger deviations), thereby highlighting atypical current fluctuations.

[0022] In a second aspect, an intelligent equipment management system includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, any one of the intelligent equipment management methods described above is implemented.

[0023] The beneficial effects of the present invention are: This method collects real-time device performance data, combines multi-dimensional features with dynamic time series analysis, calculates a time series fault risk index, and introduces anomaly weights to improve the LOF anomaly detection algorithm. This effectively addresses the traditional LOF algorithm's inaccurate detection of gradual faults. This approach not only improves the ability to detect gradual faults but also enables real-time assessment of device performance, providing strong support for device maintenance and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a method flow chart of steps S1 to S4 in a device intelligent management method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0026] Reference Figure 1 A device intelligent management method includes steps S1 to S4, specifically as follows: S1: Acquire data during the equipment performance test and preprocess the acquired data.

[0027] When testing device performance, data is collected using three sensors: a temperature sensor collects operating temperature data, a current sensor collects ground current data, and a voltage sensor collects voltage data. This data is collected every second, meaning it's collected once per second.

[0028] During data collection, environmental interference (such as electromagnetic interference and temperature fluctuations) or sensor errors (such as sensor accuracy and aging) may affect the collected data, resulting in anomalies such as noise. To eliminate the impact of this noise on subsequent analysis results, a moving average filter algorithm is used to denoise the collected data.

[0029] In addition, the different data collected (temperature, current, voltage) have different dimensions (units). The denoised data is further normalized (such as maximum value normalization). The value range of all data is normalized to between 0 and 1, eliminating the impact of different data dimensions.

[0030] After the above-mentioned preprocessing steps of denoising and normalization, relevant monitoring data during the preprocessed equipment performance test process are obtained.

[0031] S2: Calculate the real-time sequential failure risk index of the equipment based on the preprocessed data; the sequential failure risk index is positively correlated with the thermal power comprehensive failure index and the real-time voltage weight.

[0032] When the equipment is working, it will generate heat due to the thermal effect of electric current. The normal temperature rise has limited impact. However, faults such as excessive load and poor contact can cause abnormal temperature rise, accelerate component aging, and cause other faults. Therefore, by analyzing the changing trend of the equipment's operating temperature, the risk of equipment overheating can be assessed.

[0033] In one embodiment, the device operating temperature data of the n moments before the current moment (n is 20 in the embodiment of the present invention) is obtained, and a neighbor temperature monitoring sequence of the current moment is constructed. Then, a Cox-Stuart trend test is used to test whether the above neighbor temperature monitoring sequence has an upward trend, and a p-value is output. The smaller the p-value, the more significant the upward trend of the sequence.

[0034] Another thing to consider is that if the operating temperature of the equipment is within the normal range, even if the temperature fluctuates, it will not have a significant impact on the normal operation of the equipment, so this temperature rise is a normal phenomenon. Therefore, it is also necessary to pay attention to abnormal temperature rises that exceed the normal range to avoid misjudging normal operating conditions as faults.

[0035] In one embodiment, the normal temperature range of the device is set with reference to the technical manual of the device and industry standards and specifications. In the embodiment of the present invention, the maximum value of its normal temperature range is set to 40°C. For the convenience of analysis, the maximum value of the normal temperature range is recorded as the standard temperature, and the thermal failure index of the device at each moment is calculated based on the above p value and the standard temperature.

[0036] For example, a calculation formula for the thermal failure index is given:

[0037] Where, For the current moment Thermal failure index of the equipment, For the current moment The p-value obtained after trend detection of the corresponding neighbor temperature monitoring series, is the standard temperature, For the current moment The corresponding neighbor temperature monitoring sequence The value of the element, represents the hyperbolic tangent function, is the total number of elements contained in the neighbor temperature monitoring sequence. Indicates the temperature deviation between the current monitored temperature and the standard temperature.

[0038] Among them, when the device is within the normal operating temperature range, that is, the temperature does not exceed the above setting 40℃, The value is 0, which has no effect on the occurrence of equipment failure; on the contrary, when the equipment operating temperature exceeds the normal operating temperature range, that is, greater than 40℃, the value is , indicating a significant impact on equipment failure. The more pronounced the equipment's operating temperature rises and the greater the temperature exceeds the normal operating range, the higher the thermal failure index. The equipment may be experiencing overheating or abnormal temperature rise, potentially accelerating component aging and causing other failures.

[0039] Furthermore, under normal circumstances, the sum of the currents flowing through the three cores of a three-core cable in equipment should be zero, and no induced current will flow through the metal shield. However, in practice, distortion in the cores can cause the sum of the electromotive force vectors to be non-zero, generating ground current. This ground current can reflect the severity of the cable fault and the aging of the cable insulation.

[0040] In one embodiment, the ground current data of the device at n moments before the current moment (n is 20 in this embodiment of the present invention) is obtained, a ground current monitoring sequence at the current moment is constructed, and then the cable abnormality factor of the device is calculated.

[0041] For example, a calculation formula for the cable abnormality factor is given:

[0042] Where, For the current moment The cable abnormality factor of the equipment, For the current moment The corresponding ground current monitoring sequence The value of the elements, For the current moment The mode in the corresponding ground current monitoring sequence, The total number of elements included in the ground current monitoring sequence.

[0043] In the formula, when the monitoring value with mode As the deviation of increases, the numerator Increase, leading to Increase.

[0044] It's important to note that during cable operation, the resistance loss caused by current flowing through the conductor, as well as the dielectric loss of the insulation material due to electric and magnetic fields, are converted into heat energy. If this heat cannot be dissipated promptly, the cable temperature will continue to rise. Abnormal device temperatures can accelerate cable aging and damage, increasing the risk of cable failure. By considering the thermal failure index, the impact of temperature on cable failure can be factored into the assessment.

[0045] In one embodiment, the current time calculated above is Thermal failure index of the equipment and the current moment Cable abnormality factor of the device The product of .

[0046] The thermal fault index and cable anomaly factor reflect the risk of equipment failure from two different perspectives: temperature and ground current. Combining the two can improve the accuracy of fault assessment.

[0047] However, considering only thermoelectric factors may not accurately reflect the true risk of equipment under high-voltage conditions. For example, even if the temperature and ground current are within normal ranges, if the test voltage approaches the breakdown voltage, the equipment may still fail due to insufficient dielectric strength of the insulation material. When the voltage approaches the breakdown voltage, the impact of thermoelectric factors on the equipment is magnified. Including voltage in the calculation can more accurately reflect this amplification effect, thereby improving the accuracy of risk assessment.

[0048] In one embodiment, the voltage data of the device at each moment before the current moment and the breakdown voltage of the device under normal conditions are obtained, and the ratio of the voltage data of the device at each moment to the breakdown voltage of the device under normal conditions is used as the voltage weight at the corresponding moment. The timing failure risk index of the device at the current moment is calculated based on the weight and the thermoelectric comprehensive failure index obtained by the above calculation.

[0049] Then the above timing failure risk index satisfies the relationship:

[0050] For the current moment The timing failure risk index of the device, Based on the current moment Before The comprehensive thermoelectric fault index at the moment, Based on the current moment Before The voltage weight at each moment, Based on the current moment Before Voltage data at a moment, is the breakdown voltage of the device under normal conditions, Indicates that it is based on the current moment The number of previously selected historical moments.

[0051] in, The historical voltage Converted to a ratio relative to the breakdown voltage. If the value is close to or exceeds 1, it indicates that the voltage is close to the device's tolerance limit and the risk is increasing.

[0052] Representing historical moments The product of the thermoelectric comprehensive fault index and the corresponding voltage ratio. If the thermoelectric comprehensive fault index is high and the voltage is close to the breakdown value at a certain moment, the contribution is significant.

[0053] Integrated equipment in the past The fault trend and voltage status at each moment are normalized and averaged to quantify the timing fault risk at the current moment. A high value indicates that the equipment is at high risk of failure in the near future and early warning or maintenance measures are required.

[0054] The timing failure risk index of the real-time device can be obtained similarly according to the above operation.

[0055] By calculating the time series fault risk index at each moment, the abnormality of the equipment performance test data at each moment in the time series can be reflected.

[0056] S3: normalizing the timing fault risk index and using it as an anomaly weight, improving the LOF anomaly detection algorithm, and calculating the real-time anomaly score of the device according to the improved LOF anomaly detection algorithm.

[0057] Although traditional LOF (Local Outlier Factor) anomaly detection algorithms can detect anomalies in data, they may not fully consider the actual operating status and failure risks of the equipment.

[0058] By multiplying the normalized timing failure risk index by the traditional LOF anomaly score, the targeted nature of anomaly detection can be enhanced. If a device inherently has a high failure risk (i.e., a large normalized timing failure risk index), then even if the traditional LOF score is low, the improved score may increase, making it easier to detect potential problems.

[0059] The improved anomaly score is expressed as follows:

[0060] Where, For the current moment Anomaly score after device improvement, is the normalized current moment The timing failure risk index of the device, is the current moment calculated using the traditional LOF anomaly detection algorithm The anomaly score of the device.

[0061] S4: Obtain device performance test results based on the anomaly score and formulate device management strategies based on the device test results.

[0062] In one embodiment, the anomaly detection threshold is set to 1.2. If the current moment If the anomaly score after equipment improvement is greater than the anomaly detection threshold, it indicates that there are certain problems with the equipment's performance test results and that there may be risks in actual application. In this case, quality inspectors are required to promptly handle the issue and conduct further inspections, repairs, or adjustments to the problematic equipment to ensure its quality and safety. For example, in a smart mining power system, based on real-time collected equipment performance data and calculated anomaly scores, the equipment's operating parameters can be adjusted in real time to ensure that the equipment is always in optimal operating condition. Based on the transformer's real-time load, temperature, and other performance data, combined with its anomaly score, the transformer's cooling system operating mode can be dynamically adjusted (such as increasing the cooling fan speed or starting auxiliary cooling equipment) to ensure that the transformer operates within a safe and efficient temperature range, improve power transmission efficiency, and reduce energy loss.

[0063] In addition, equipment with anomaly scores greater than the quality inspection threshold should be inspected and diagnosed to confirm the type and severity of the fault. For faults that may affect production safety or cause major downtime, emergency measures should be taken immediately, such as shutdown inspections and temporary repairs.

[0064] If the improved anomaly score is less than the quality detection threshold, it indicates that the device performance test result has passed.

[0065] The system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the device intelligent management method according to the first aspect of the present invention is implemented.

[0066] The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and therefore will not be described in detail here.

[0067] It should be noted that those skilled in the art may make various modifications and improvements without departing from the scope of the present invention, and these modifications and improvements fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent for this invention shall be based on the appended claims.

Claims

1. A device intelligent management method, characterized in that: include: Collect data during equipment performance testing in real time and pre-process the acquired data; The data includes voltage and current data of the device; Calculate the real-time equipment timing failure risk index based on pre-processed data; The timing fault risk index is positively correlated with the thermal power comprehensive fault index and the real-time voltage weight; Normalizing the timing fault risk index and using it as an anomaly weight, improving the LOF anomaly detection algorithm, and calculating the real-time anomaly score of the device based on the improved LOF anomaly detection algorithm; Obtain device performance test results based on anomaly scores, and formulate device management strategies based on the device test results.

2. A device intelligent management method according to claim 1, characterized in that: The data also includes temperature data of the device.

3. The device intelligent management method according to claim 2, characterized in that: The process of obtaining the thermoelectric comprehensive fault index includes: Obtain the device operating temperature data for the n moments before the current moment, and construct a neighboring temperature monitoring sequence for the current moment. Also obtain the device grounding current data for the n moments before the current moment, and construct a grounding current monitoring sequence for the current moment. Based on the neighboring temperature monitoring sequence and ground current monitoring sequence, the thermal fault index and cable abnormality factor of the device at the current moment are calculated respectively; The product of the thermal fault index and the cable abnormality factor is taken as the thermal and electrical comprehensive fault index of the equipment at the current moment.

4. The device intelligent management method according to claim 3, characterized in that: The process of obtaining the voltage weight includes: The voltage data of the device at each moment before the current moment and the breakdown voltage of the device under normal conditions are obtained, and the ratio of the voltage data of the device at each moment to the breakdown voltage of the device under normal conditions is used as the voltage weight at the corresponding moment.

5. The device intelligent management method according to claim 4, characterized in that: The timing failure risk index satisfies the relationship: ; For the current moment The timing failure risk index of the device, Based on the current moment Before The comprehensive thermoelectric fault index at the moment, Based on the current moment Before The voltage weight at each moment, Based on the current moment Before Voltage data at a moment, is the breakdown voltage of the device under normal conditions, Indicates based on the current moment The number of previously selected historical moments.

6. A device intelligent management method according to claim 5, characterized in that: The process of obtaining the corresponding anomaly score in the improved LOF anomaly detection algorithm is as follows: The normalized timing fault risk index at the current moment is used as a weight to multiply the original LOF anomaly score to obtain the anomaly score after using the improved LOF anomaly detection algorithm at the current moment.

7. The device intelligent management method according to claim 6, characterized in that: Obtaining the device performance test results based on the anomaly score and formulating a device management strategy based on the device test results includes: If the anomaly score calculated based on the improved LOF anomaly detection algorithm is greater than the preset anomaly detection threshold, the performance test of the device at the current moment is determined to be unqualified; otherwise, the performance test of the device at the current moment is determined to be qualified; Equipment that fails the performance test will be inspected and diagnosed to confirm the type and severity of the fault. Equipment with high severity will be shut down for inspection or temporary repair.

8. The device intelligent management method according to claim 3, characterized in that: The process of obtaining the thermal failure index includes: Perform trend detection on the neighboring temperature monitoring series and obtain a p-value; Set the device's standard temperature and calculate the temperature deviation from the standard temperature for each element in the neighboring temperature monitoring sequence. Use the temperature deviation as the input of the hyperbolic tangent function, perform a positive operation on its output, and then sum the values ​​for all elements in the neighboring temperature monitoring sequence. The summation result is multiplied by the value p to obtain the thermal failure index.

9. The device intelligent management method according to claim 3, characterized in that: The process of obtaining the cable abnormality factor includes: The mode of the ground current monitoring sequence is obtained, the square of the difference between each element in the ground current monitoring sequence and the mode is calculated and summed, and the summed result is divided by the mode to obtain the cable abnormality factor.

10. An intelligent equipment management system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the device intelligent management method according to any one of claims 1 to 9 is implemented.

Images