Power equipment state monitoring and early warning system based on distributed sensors

Abstract

This invention belongs to the field of power equipment monitoring technology, specifically a power equipment status monitoring and early warning system based on distributed sensors. It includes a multimodal sensor collaborative acquisition module, an edge node adaptive relay transmission module, a multi-dimensional status assessment module, a dynamic threshold early warning decision module, and an intelligent early warning information push module. The invention dynamically adjusts the acquisition cycle of each sensor through the multimodal sensor collaborative acquisition module, ensuring the temporal consistency and high quality of the output data. The edge node adaptive relay transmission module dynamically excludes unqualified nodes and selects the optimal transmission path based on an evaluation function, reducing node energy consumption and ensuring real-time reliable data transmission. The multi-dimensional status assessment module constructs an assessment system from multiple key dimensions to avoid misjudgments and accurately identifies and classifies abnormal power equipment statuses, formulating differentiated push strategies based on the early warning level to ensure that early warning information reaches key stakeholders in a timely manner.

Description

Technical Field

[0001] This invention relates to the field of power equipment monitoring technology, specifically a power equipment condition monitoring and early warning system based on distributed sensors. Background Technology

[0002] Power equipment refers to all kinds of power equipment and auxiliary devices that support the power system in completing the energy conversion and transmission of the entire process of power generation, transmission, transformation, distribution and consumption, and ensure the safe and stable supply of power. These devices are the core carriers of power grid operation, and their state stability directly determines the reliability of the power system. Therefore, it is necessary to monitor the operation of power equipment.

[0003] For example, Chinese invention patent with publication number CN114661735A discloses a power equipment operation status monitoring system, which monitors and displays the operation status of power equipment through a power distribution monitoring module, monitors and displays the environment in which the power equipment is located through an environmental monitoring module, and monitors and displays the energy consumption of the power equipment through an energy consumption monitoring module. It has the advantages of stable performance and low cost.

[0004] However, in practical applications, the above-mentioned inventions cannot dynamically adjust the acquisition strategy according to the real-time load rate of power equipment, nor can they automatically and reasonably select the relay transmission path based on the node status. Furthermore, they cannot accurately identify and classify the abnormal status of power equipment, nor can they formulate differentiated push strategies based on the warning level. This is not conducive to improving the real-time performance, accuracy, and reliability of power equipment status monitoring. Therefore, a solution is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a power equipment condition monitoring and early warning system based on distributed sensors to address the technical deficiencies mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a power equipment status monitoring and early warning system based on distributed sensors, including a multimodal sensing collaborative acquisition module, an edge node adaptive relay transmission module, a multi-dimensional status assessment module, a dynamic threshold early warning decision module, and an early warning information intelligent push module;

[0007] The multimodal sensing collaborative acquisition module monitors the temperature, partial discharge, vibration, and dissolved gas in oil of power equipment, and coordinates the various types of sensors involved in the power equipment to complete synchronous data acquisition, raw data preprocessing, and labeled data output.

[0008] The edge node adaptive relay transmission module dynamically monitors the operating status of distributed edge nodes, optimizes the data transmission path, and sends the complete data frame to the multi-dimensional status evaluation module.

[0009] The multi-dimensional condition assessment module quantifies the health status of power equipment from four dimensions: temperature, partial discharge, vibration, and dissolved gases in oil, generates a health index of the power equipment, and sends it to the dynamic threshold early warning decision module.

[0010] The dynamic threshold early warning decision module combines historical health index data of power equipment with real-time environmental parameters to dynamically adjust the early warning threshold and determine the early warning level of the power equipment. The intelligent early warning information push module formulates differentiated push strategies based on the early warning level and intelligently pushes early warning information.

[0011] Furthermore, during operation, the multimodal sensing collaborative acquisition module acquires the real-time operating load of the power equipment and dynamically adjusts the acquisition cycle of each sensor according to the load level. If the power equipment is under high load, the acquisition cycle of the temperature sensor is set to 20s, the acquisition cycle of the partial discharge sensor is set to 30s, the acquisition cycle of the vibration sensor is set to 40s, and the acquisition cycle of the dissolved gas in oil sensor is set to 60s. When the power equipment is under low load, the acquisition cycle of each sensor is extended to twice the original cycle.

[0012] Furthermore, the multimodal sensing collaborative acquisition module performs time calibration on all distributed sensors through the BeiDou positioning system, controlling the timestamp error to less than or equal to 1ms, ensuring the temporal consistency of sensor data from different locations, and using the 3σ criterion to remove outliers, and then filtering high-frequency noise through an exponentially weighted moving average algorithm. After obtaining the smoothed acquisition data, a three-dimensional identifier of "sensor ID-device number-acquisition timestamp" is added to the preprocessed data to generate a standard data frame, and the standard data frame is transmitted to the edge node adaptive relay transmission module.

[0013] Furthermore, the edge node adaptive relay transmission module collects the remaining power E, signal reception strength RSSI and data buffer size of each edge node in real time. If the remaining power E of the edge node is less than 20%, or RSSI is less than -80dBm, or C is greater than 80%, the corresponding edge node is automatically excluded as a relay node.

[0014] Furthermore, edge nodes that were not excluded were marked as potential relay nodes. A transmission path evaluation function was constructed based on the remaining power E, the received signal strength RSSI, and the data buffer size C. All potential relay nodes were scored, and the path with the highest score was selected as the optimal transmission path.

[0015] Furthermore, the edge node adaptive relay transmission module adopts a block-based verification transmission mechanism, which divides the standard data frame into blocks and adds a CRC32 check code to each block. The receiving end sends back an acknowledgment frame after the verification is successful, and triggers retransmission if it fails. In addition, the edge node adaptive relay transmission module dynamically adjusts the transmission power according to the remaining power E of the node. When E≥50%, the power is 15dBm, and when 20%≤E<50%, the power is 10dBm.

[0016] Furthermore, the evaluation and analysis process of the multi-dimensional state assessment module is as follows:

[0017] Key indicators for each dimension were determined, including hot spot temperature deviation and temperature rise rate in the temperature dimension, discharge quantity and discharge frequency in the partial discharge dimension, vibration amplitude and frequency deviation in the vibration dimension, and methane concentration and acetylene concentration in the dissolved gas in oil dimension.

[0018] The linear normalization formula is used to map each indicator to the interval [0,1], where 0 indicates that the indicator is completely normal and 1 indicates that the indicator has reached the fault threshold. The analytic hierarchy process is used to construct a judgment matrix, calculate the weight of each dimension and indicator, and calculate the equipment health index HI by weighted summation formula.

[0019] Furthermore, the specific operation process of the dynamic threshold early warning decision module is as follows:

[0020] Health index data of power equipment over the past three months were collected. The historical health index mean HIavg and the historical health index standard deviation σHI were calculated. Based on HIavg, σHI, and the environmental correction factor k, the three-level early warning thresholds TH1, TH2, and TH3 were calculated, with TH1 > TH2 > TH3. The specific calculation formula is as follows:

[0021] Level 1 warning threshold: TH1 = HIavg + 2σHI × k;

[0022] Level II warning threshold: TH2 = HIavg + σHI × k;

[0023] Level 3 warning threshold: TH3 = HIavg + 0.5σHI × k;

[0024] The current health index HI is compared with the dynamic threshold. If HI≥TH1, a level 1 warning is triggered; if TH2≤HI<TH1, a level 2 warning is triggered; if TH3≤HI<TH2, a level 3 warning is triggered; if HI<TH3, no warning is triggered. The key dimensions that trigger the warning are analyzed and a warning reason report is generated. The warning level, reason report, current HI, and dynamic threshold are transmitted to the warning information intelligent push module.

[0025] Furthermore, the environmental correction factor k is obtained through analysis as follows:

[0026] The temperature (Tenv) and humidity (Henv) of the operating environment of the power equipment are collected in real time. The environmental correction factor (k) is calculated using the formula: k = 1 + 0.012 × (Tenv - 25) + 0.006 × (Henv - 60).

[0027] Furthermore, the dynamic threshold early warning decision module communicates with the periodic hazard assessment module. The periodic hazard assessment module is used to set the monitoring period and analyze the degree of safety hazards of power equipment during the monitoring period. Through analysis, it determines whether to generate a high-hazard signal for the equipment. When a high-hazard signal is generated, it reminds the management personnel to strengthen the supervision of power equipment.

[0028] Furthermore, the specific analysis process of the periodic hazard assessment module is as follows:

[0029] The system acquires the number of times that warnings for power equipment were not processed within the specified time during the monitoring period and marks them as warning processing delay values. It also marks the total time that power equipment was out of service due to faults during the monitoring period as fault temporary operation time detection values. The system compares the warning processing delay values ​​and fault temporary operation time detection values ​​with the preset warning processing delay thresholds and preset fault temporary operation time detection thresholds respectively. If the warning processing delay value or the fault temporary operation time detection value exceeds the corresponding preset threshold, a high-risk equipment signal is generated.

[0030] If the delayed detection value and the fault temporary operation detection value do not exceed the corresponding preset threshold, the number of times a Level 1 warning is generated, the number of times a Level 2 warning is generated, and the number of times a Level 3 warning is generated during the monitoring period are obtained and marked as high-risk frequency value, medium-risk frequency value, and low-risk frequency value, respectively. The high-risk frequency value, medium-risk frequency value, and low-risk frequency value are weighted and summed to obtain the initial feature value. The initial feature value is compared with the preset initial feature threshold. If the initial feature value exceeds the preset initial feature threshold, a high-risk equipment signal is generated.

[0031] Compared with the prior art, the beneficial effects of the present invention are:

[0032] 1. This invention improves the real-time performance, accuracy, and reliability of power equipment status monitoring by dynamically adjusting the acquisition cycle of each sensor, ensuring the time consistency and high quality of the output data, dynamically eliminating unqualified nodes and selecting the optimal transmission path based on the evaluation function, constructing an evaluation system from multiple key dimensions to avoid misjudgment, accurately identifying and classifying the abnormal state of power equipment, and formulating differentiated push strategies based on the warning level.

[0033] 2. This invention uses a periodic hazard assessment module to analyze the number of delays in early warning processing, the duration of fault shutdowns, and the frequency of early warnings at all levels during the set monitoring period. This generates high-hazard signals for equipment to remind managers to strengthen supervision, enabling proactive perception and early intervention of long-term operational hazards in equipment. This further ensures the safe and stable operation of power equipment and makes up for the shortcomings of traditional real-time monitoring in long-term hazard prevention and control. Attached Figure Description

[0034] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0035] Figure 1 This is a system block diagram of Embodiment 1 of the present invention;

[0036] Figure 2 This is a system block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] Example 1: As Figure 1 As shown, the power equipment status monitoring and early warning system based on distributed sensors proposed in this invention includes a multimodal sensing collaborative acquisition module, an edge node adaptive relay transmission module, a multi-dimensional status assessment module, a dynamic threshold early warning decision module, and an early warning information intelligent push module.

[0039] The multimodal sensing collaborative acquisition module monitors the temperature, partial discharge, vibration, and dissolved gas in oil of power equipment. It also coordinates and schedules multiple types of sensors involved in the power equipment to complete synchronous data acquisition, raw data preprocessing, and labeled data output. By dynamically adjusting the acquisition cycle, it reduces data redundancy and ensures time sequence consistency through BeiDou time synchronization. After preprocessing, it provides a high-quality data source for subsequent transmission and analysis, reducing subsequent analysis errors.

[0040] It should be noted that when the multimodal sensor collaborative acquisition module is running, it acquires the real-time operating load of the power equipment (such as transformer load rate and the number of switch opening and closing times), and dynamically adjusts the acquisition cycle of each sensor according to the load level. If the power equipment is under high load (load rate ≥ 80%), the acquisition cycle of the temperature sensor is set to 20s, the acquisition cycle of the partial discharge sensor is set to 30s, the acquisition cycle of the vibration sensor is set to 40s, and the acquisition cycle of the dissolved gas in oil sensor is set to 60s. When the power equipment is under low load (load rate < 40%), the acquisition cycle of each sensor is extended to twice the original cycle.

[0041] Furthermore, the multimodal sensing collaborative acquisition module uses the BeiDou positioning system to perform time calibration on all distributed sensors, controlling the timestamp error to less than or equal to 1ms, ensuring the temporal consistency of sensor data from different locations, and using the 3σ criterion to remove outliers (such as data with temperatures exceeding the physical range of -40℃ to 150℃). Then, it uses an exponentially weighted moving average algorithm to filter high-frequency noise, with the formula: yt=λ×et+(1-λ)×y(t−1).

[0042] Where yt: the data value after the t-th filtering;

[0043] et: The value of the original data collected in the t-th time;

[0044] y(t−1): The data value after the (t-1)th filtering;

[0045] λ: Smoothing coefficient, with a value of 0.3 to 0.6 (the higher the load, the larger λ is);

[0046] After obtaining the smoothed acquisition data, a three-dimensional identifier of "sensor ID-device number-acquisition timestamp" is added to the preprocessed data to generate a standard data frame, and the standard data frame is transmitted to the edge node adaptive relay transmission module.

[0047] The edge node adaptive relay transmission module dynamically monitors the operating status of distributed edge nodes, optimizes the data transmission path, and sends the complete data frame to the multi-dimensional status assessment module. It can dynamically optimize the path and transmission power to achieve low-power and high-reliability transmission, which is beneficial to solving the problems of signal attenuation and transmission interruption of distributed sensors.

[0048] Specifically, the edge node adaptive relay transmission module collects the remaining power E, signal reception strength RSSI and data buffer size of each edge node in real time. If the remaining power E of the edge node is less than 20%, or RSSI is less than -80dBm, or C is greater than 80%, the corresponding edge node is automatically excluded as a relay node.

[0049] Furthermore, edge nodes that were not excluded were marked as potential relay nodes. A transmission path evaluation function was constructed based on the remaining battery power E, the received signal strength RSSI, and the data buffer size C. All potential relay nodes were scored, and the path with the highest score was selected as the optimal transmission path. The transmission path evaluation function is as follows:

[0050]

[0051] Wherein, S: Transmission path score (value from 0 to 1, the higher the value, the better the path);

[0052] w1, w2, w3: Weighting coefficients (0.4, 0.3, 0.3 respectively, prioritizing node power and signal strength);

[0053] Emax: Maximum node capacity (unit: mAh);

[0054] RSSImax and RSSImin: Maximum and minimum signal strength (-40dBm and -100dBm, respectively).

[0055] Cmax: Maximum cache size for a node (unit: MB).

[0056] Furthermore, the edge node adaptive relay transmission module adopts a block-based verification transmission mechanism, dividing the standard data frame into blocks (each block is 1024 bytes), and attaching a CRC32 checksum to each block. The receiving end sends back an "acknowledgment frame" after the verification is successful, and triggers retransmission if it fails. At the same time, the edge node adaptive relay transmission module dynamically adjusts the transmission power according to the remaining power E of the node. The power is 15dBm when E≥50% and 10dBm when 20%≤E<50%, which significantly reduces energy consumption.

[0057] The multi-dimensional condition assessment module quantifies the health status of electrical equipment from four dimensions: temperature, partial discharge, vibration, and dissolved gases in oil. This avoids misjudgment based on a single parameter, generates a health index for the electrical equipment, and sends it to the dynamic threshold early warning decision module, ensuring the comprehensiveness and accuracy of the health status assessment. The specific assessment and analysis process of the multi-dimensional condition assessment module is as follows:

[0058] First, the key indicators for each dimension are determined. Among them, the temperature dimension includes hot spot temperature deviation and temperature rise rate; the partial discharge dimension includes discharge quantity and discharge frequency; the vibration dimension includes vibration amplitude and frequency deviation; and the dissolved gas in oil dimension includes methane concentration and acetylene concentration.

[0059] A linear normalization formula is used to map each indicator to the interval [0,1], where 0 represents that the indicator is completely normal and 1 represents that the indicator has reached the fault threshold. The linear normalization formula is as follows:

[0060]

[0061] Where x′: the normalized index value;

[0062] x: Original index value;

[0063] xmin: Lower limit of the normal threshold for the indicator;

[0064] xmax: Upper limit of the indicator failure threshold;

[0065] A judgment matrix was constructed using the Analytic Hierarchy Process (AHP), and the weights of each dimension and indicator were calculated (e.g., temperature dimension weight 0.3, partial discharge dimension weight 0.35, vibration dimension weight 0.2, dissolved gas in oil dimension weight 0.15). The equipment health index HI (value from 0 to 1) was calculated using a weighted summation formula. The closer HI is to 0, the better the condition of the electrical equipment; the closer HI is to 1, the higher the risk of electrical equipment failure. The formula is as follows:

[0066]

[0067] Among them, HI: Health Index of Power Equipment;

[0068] qi: The weight of the i-th dimension;

[0069] n: The number of dimensions, n=4 (corresponding to temperature, partial discharge, vibration, and dissolved gas in oil, respectively);

[0070] p: The number of key indicators in the corresponding dimension;

[0071] wij: The weight of the j-th indicator in the i-th dimension;

[0072] x′ij: The normalized value of the j-th indicator in the i-th dimension.

[0073] The dynamic threshold early warning decision module combines historical health index data of power equipment with real-time environmental parameters to dynamically adjust the early warning threshold, determine the early warning level of the power equipment, and achieve accurate early warning. This not only helps to solve the problems of false and missed early warnings caused by fixed thresholds, but also facilitates targeted countermeasures. The specific operation process of the dynamic threshold early warning decision module is as follows:

[0074] The temperature (Tenv) and humidity (Henv) of the operating environment of the power equipment are collected in real time. The environmental correction factor (k) is calculated using the formula: k = 1 + 0.012 × (Tenv - 25) + 0.006 × (Henv - 60). That is, when k > 1, the ambient temperature is higher than 25℃ or the humidity is higher than 60%RH, indicating that the operating environment of the power equipment is harsh and the warning threshold needs to be lowered.

[0075] Health index data of power equipment over the past three months were collected. The historical health index mean HIavg and the historical health index standard deviation σHI were calculated. Based on HIavg, σHI, and the environmental correction factor k, the three-level early warning thresholds TH1, TH2, and TH3 were calculated, with TH1 > TH2 > TH3. The specific calculation formula is as follows:

[0076] Level 1 warning threshold: TH1 = HIavg + 2σHI × k;

[0077] Level II warning threshold: TH2 = HIavg + σHI × k;

[0078] Level 3 warning threshold: TH3 = HIavg + 0.5σHI × k;

[0079] The current health index HI is compared with the dynamic threshold. If HI≥TH1, a level 1 warning is triggered; if TH2≤HI<TH1, a level 2 warning is triggered; if TH3≤HI<TH2, a level 3 warning is triggered; if HI<TH3, no warning is triggered. The key dimensions that trigger the warning (such as HI elevation caused by excessive partial discharge) are analyzed and a warning cause report is generated. The warning level, cause report, current HI, and dynamic threshold are transmitted to the warning information intelligent push module.

[0080] The intelligent early warning information push module formulates differentiated push strategies based on the early warning level, and intelligently pushes early warning information to designated objects to improve the efficiency of operation and maintenance response. For example, the target of the first-level early warning is the operation and maintenance manager, the maintenance team, and the director of the monitoring center. The push method is "SMS + telephone voice + system pop-up". The push content includes the device name, ID, early warning level, key abnormal parameters, emergency maintenance suggestions, and contact person.

[0081] Level 2 alerts are sent to operations and maintenance managers and the maintenance team via SMS and system messages. The alert content includes equipment status, anomaly details, and a recommendation to perform maintenance within 24 hours. Level 3 alerts are sent to on-site operations and maintenance personnel via system messages. The alert content includes abnormal equipment parameters and a recommendation to observe the equipment within 72 hours. Based on the alert level matching strategy, SMS interfaces (such as Alibaba Cloud SMS API), voice interfaces (such as iFlytek speech synthesis API), and system message interfaces are used to push information to the corresponding recipients, and the push time, recipient, and status (read / unread) are recorded.

[0082] Furthermore, if no "read" confirmation is received within 30 minutes of a Level 1 alert being sent, it will be automatically resent and the push method will be upgraded (e.g., push to the supervisor in charge will be added); if a Level 2 alert is not confirmed, it will be resent after 1 hour. Finally, the push records will be stored.

[0083] Example 2: Figure 2 As shown, the difference between this embodiment and Embodiment 1 is that the dynamic threshold early warning decision module is communicatively connected to the periodic hazard assessment module. The periodic hazard assessment module is used to set the monitoring period, preferably twenty days. The module analyzes the degree of safety hazards of the power equipment during the monitoring period to determine whether a high-hazard signal is generated. When a high-hazard signal is generated, it alerts management personnel to strengthen the supervision of the power equipment, facilitating the reasonable formulation of management plans for the power equipment and further ensuring the safe and stable operation of the power equipment. The specific analysis process is as follows:

[0084] The system acquires the number of times that warnings for power equipment were not processed within the specified time during the monitoring period and marks them as warning processing delay values. It also marks the total time that power equipment was out of service due to faults during the monitoring period as fault temporary operation time detection values. The system compares the warning processing delay values ​​and fault temporary operation time detection values ​​with the preset warning processing delay thresholds and preset fault temporary operation time detection thresholds, respectively. If the warning processing delay value or fault temporary operation time detection value exceeds the corresponding preset threshold, it indicates that the safety hazard level of the power equipment is high during the monitoring period, and a high equipment hazard signal is generated.

[0085] If the delayed detection value and the fault temporary operation detection value do not exceed the corresponding preset threshold, the number of times the first-level warning, the number of times the second-level warning, and the number of times the third-level warning are generated during the monitoring period are obtained and marked as high-risk frequency value, medium-risk frequency value, and low-risk frequency value, respectively. The initial feature value is obtained by weighted summation of the high-risk frequency value, medium-risk frequency value, and low-risk frequency value.

[0086] The system assigns corresponding preset proportional coefficients a1, a2, and a3 to high-risk, medium-risk, and low-risk frequency values, respectively, where a1 > a2 > a3 > 0. Each of these values ​​is then multiplied by its respective preset proportional coefficient, and the sum of these three products is marked as the initial characteristic value. It should be noted that a larger initial characteristic value indicates a higher degree of safety hazard to the power equipment during the monitoring period. The initial characteristic value is then compared with a preset initial characteristic threshold. If the initial characteristic value exceeds the preset initial characteristic threshold, it indicates a high degree of safety hazard to the power equipment during the monitoring period, and a high-hazard signal for the equipment is generated.

[0087] This invention dynamically adjusts the acquisition cycle of each sensor through a multimodal sensing collaborative acquisition module, ensuring the temporal consistency and high quality of the output data, thus avoiding analysis errors caused by the quality of the original data. The edge node adaptive relay transmission module dynamically excludes unqualified nodes and selects the optimal transmission path based on the evaluation function. Combined with a block-based verification transmission mechanism and dynamic power adjustment based on power consumption, it significantly reduces node energy consumption and ensures real-time and reliable data transmission, thereby solving the problems of signal attenuation and unstable transmission in distributed scenarios.

[0088] The multi-dimensional status assessment module constructs an assessment system from multiple key dimensions and quantifies and generates a health index, breaking away from the limitations of traditional single-parameter assessments and avoiding misjudgments caused by one-sided assessment dimensions. This makes the equipment health status assessment more comprehensive and accurate. The dynamic threshold early warning decision module breaks through the limitations of fixed thresholds, enabling accurate identification and graded early warning of abnormal equipment status, effectively solving the problem of frequent false and missed early warnings in existing technologies.

[0089] Furthermore, the intelligent early warning information push module formulates differentiated push strategies based on the early warning level, ensuring that early warning information can reach key roles in a timely manner, significantly shortening operation and maintenance response time, improving fault handling efficiency, and significantly enhancing the real-time performance, accuracy, and reliability of power equipment status monitoring and early warning.

[0090] In the technical solution of this invention, the threshold, preset value, preset range, etc., are set for result comparison and analysis to determine whether it is good or bad. The magnitude of these values ​​is set and stored based on a combination of large-scale model analysis of sample data and human experience. It can also be appropriately adjusted based on seasonal or common-sense influence conditions. Similarly, the preset weight coefficients, influence factors, etc., are set based on the magnitude of the influence of each parameter on the result, and the specific values ​​are allocated to ultimately reflect the influence on the result. This is also set and stored based on a combination of large-scale model analysis of sample data and human experience. It can also be appropriately adjusted based on seasonal or common-sense influence conditions.

[0091] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, enabling those skilled in the art to better understand and utilize it. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A distributed sensor based power equipment condition monitoring and early warning system, characterized in that, It includes a multimodal sensing collaborative acquisition module, an edge node adaptive relay transmission module, a multi-dimensional status assessment module, a dynamic threshold early warning decision module, and an early warning information intelligent push module; The multimodal sensing collaborative acquisition module coordinates and schedules multiple types of sensors involved in power equipment to complete synchronous data acquisition, raw data preprocessing and labeled data output. The edge node adaptive relay transmission module dynamically monitors the operating status of distributed edge nodes and optimizes data transmission paths. The multi-dimensional status assessment module quantifies the health status of power equipment from four dimensions: temperature, partial discharge, vibration, and dissolved gases in oil, generating a health index for the power equipment. The dynamic threshold early warning decision dynamically adjusts the early warning threshold and determines the early warning level of the power equipment. The early warning information intelligent push module formulates differentiated push strategies based on the early warning level and intelligently pushes early warning information. The specific operation process of the dynamic threshold early warning decision module is as follows: Health index data of power equipment over the past three months were collected. The historical health index mean HIavg and the historical health index standard deviation σHI were calculated. Based on HIavg, σHI, and the environmental correction factor k, the three-level early warning thresholds TH1, TH2, and TH3 were calculated, with TH1 > TH2 > TH3. The specific calculation formula is as follows: Level 1 warning threshold: TH1 = HIavg + 2σHI × k; Level II warning threshold: TH2 = HIavg + σHI × k; Level 3 warning threshold: TH3 = HIavg + 0.5σHI × k; The current health index HI is compared with the dynamic threshold. If HI≥TH1, a Level 1 warning is triggered; if TH2≤HI<TH1, a Level 2 warning is triggered; if TH3≤HI<TH2, a Level 3 warning is triggered. If HI < TH3, no warning is given; The system analyzes the key dimensions that trigger the warning and generates a warning reason report. It also transmits the warning level, reason report, current HI, and dynamic threshold to the intelligent warning information push module. The environmental correction factor k is obtained through analysis as follows: The temperature (Tenv) and humidity (Henv) of the operating environment of the power equipment are collected in real time. The environmental correction factor (k) is calculated using the environmental correction factor formula: k = 1 + 0.012 × (Tenv - 25) + 0.006 × (Henv - 60). The dynamic threshold early warning decision module communicates with the periodic hidden danger assessment module. The periodic hidden danger assessment module is used to set the monitoring period and analyze the degree of safety hazards of power equipment during the monitoring period. The analysis is used to determine whether a high hidden danger signal of equipment is generated. When a high hidden danger signal of equipment is generated, it reminds the management personnel to strengthen the supervision of power equipment. The specific analysis process of the periodic hazard assessment module is as follows: The system acquires the number of times that warnings for power equipment were not processed within the specified time during the monitoring period and marks them as warning processing delay values. It also marks the total time that power equipment was out of service due to faults during the monitoring period as fault temporary operation time detection values. The system compares the warning processing delay values ​​and fault temporary operation time detection values ​​with the preset warning processing delay thresholds and preset fault temporary operation time detection thresholds respectively. If the warning processing delay value or the fault temporary operation time detection value exceeds the corresponding preset threshold, a high-risk equipment signal is generated. If the delayed detection value and the fault temporary operation detection value do not exceed the corresponding preset threshold, the number of times a Level 1 warning is generated, the number of times a Level 2 warning is generated, and the number of times a Level 3 warning is generated during the monitoring period are obtained and marked as high-risk frequency value, medium-risk frequency value, and low-risk frequency value, respectively. The high-risk frequency value, medium-risk frequency value, and low-risk frequency value are weighted and summed to obtain the initial feature value. The initial feature value is compared with the preset initial feature threshold. If the initial feature value exceeds the preset initial feature threshold, a high-risk equipment signal is generated.

2. The distributed sensor based power equipment condition monitoring and warning system as claimed in claim 1, wherein, When the multimodal sensing collaborative acquisition module is running, it acquires the real-time operating load of the power equipment and dynamically adjusts the acquisition cycle of each sensor according to the load level. If the power equipment is under high load, the acquisition cycle of the temperature sensor is set to 20s, the acquisition cycle of the partial discharge sensor is set to 30s, the acquisition cycle of the vibration sensor is set to 40s, and the acquisition cycle of the dissolved gas in oil sensor is set to 60s. When the power equipment is under low load, the acquisition cycle of each sensor is extended to twice the original cycle.

3. The power equipment condition monitoring and early warning system based on distributed sensors according to claim 2, characterized in that, The multimodal sensing collaborative acquisition module performs time calibration on all distributed sensors through the BeiDou positioning system, removes outliers using the 3σ criterion, filters high-frequency noise using an exponentially weighted moving average algorithm, obtains smoothed acquisition data, adds three-dimensional labels to the preprocessed data, and generates standard data frames.

4. The power equipment condition monitoring and early warning system based on distributed sensors according to claim 1, characterized in that, The edge node adaptive relay transmission module collects the remaining power E, signal reception strength RSSI and data buffer size of each edge node in real time. If the remaining power E of the edge node is less than 20%, or RSSI is less than -80dBm, or C is greater than 80%, the corresponding edge node is automatically excluded as a relay node. Furthermore, edge nodes that were not excluded were marked as potential relay nodes. A transmission path evaluation function was constructed based on the remaining power E, the received signal strength RSSI, and the data buffer size C. All potential relay nodes were scored, and the path with the highest score was selected as the optimal transmission path.

5. The power equipment condition monitoring and early warning system based on distributed sensors according to claim 4, characterized in that, The edge node adaptive relay transmission module adopts a block-based verification transmission mechanism, which divides the standard data frame into blocks and adds a CRC32 check code to each block; and the edge node adaptive relay transmission module dynamically adjusts the transmission power according to the remaining power E of the node.

6. The power equipment condition monitoring and early warning system based on distributed sensors according to claim 1, characterized in that, The evaluation and analysis process of the multi-dimensional status assessment module is as follows: determine the key indicators of each dimension, map each indicator to the [0,1] interval using the linear normalization formula, construct the judgment matrix using the analytic hierarchy process, calculate the weights of each dimension and indicator, and calculate the equipment health index HI using the weighted summation formula.

Images