Weak current equipment abnormality judgment method and system of internet of things

Abstract

The application relates to a weak-current equipment abnormality judgment method and system of an Internet of Things, which comprises the following steps: collecting multi-dimensional parameters such as temperature, current, voltage and vibration frequency by an Internet of Things node and generating a dynamic tracking identifier; performing dynamic parameter correlation analysis at an edge node, and establishing a real-time coupling degree relationship between parameters; grading and evaluating abnormality levels according to analysis results, and distinguishing between primary abnormality of single parameter deviation and advanced abnormality of multiple parameter deviation; starting a differentiated verification mechanism for different levels of abnormality by a cloud platform, performing horizontal comparison for primary abnormality, and performing full life cycle backtracking verification for advanced abnormality; and the system comprises a data acquisition module, an edge computing module, an analysis and evaluation module and a cloud analysis platform, and the system can realize the above method. Through dynamic coupling degree analysis and grading verification mechanism, the application significantly improves the accuracy of abnormality judgment and effectively reduces false positives caused by environmental interference.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of weak current equipment detection based on Internet of Things, in particular to a weak current equipment abnormality judgment method and system based on Internet of Things. BACKGROUND

[0002] In the field of intelligent buildings, industrial automation, etc., as the key infrastructure, the stable operation of weak current equipment is directly related to the reliability of the whole system. At present, the Internet of Things technology is generally used in the industry to monitor the state of weak current equipment, and various types of sensors are deployed to collect equipment operation parameters in real time. However, this monitoring method faces a prominent problem: the traditional abnormality judgment method mainly relies on the threshold alarm mechanism of a single parameter. When a parameter is detected to be out of the preset threshold range, it is determined to be abnormal. Although this method is simple and direct, it has obvious misjudgment problems. In the actual running environment, the parameters of weak current equipment often fluctuate due to environmental temperature fluctuations, transient power interference and other factors. These fluctuations are often within the range of equipment bearing and do not cause actual faults, but they frequently trigger false alarms.

[0003] The prior art attempts to improve this problem by introducing a simple multi-parameter correlation analysis, for example, monitoring temperature and current parameters at the same time, and determining abnormality only when both parameters exceed the threshold. But this improvement still stays at the static correlation level and cannot capture the complex dynamic coupling relationship between parameters. Especially when the equipment has early potential faults, each parameter often shows non-synchronous and nonlinear change characteristics, resulting in the existing method either missing real faults or continuing to produce a large number of false alarms. SUMMARY

[0004] Therefore, the technical problem to be solved by the present application is to overcome the problem of weak current equipment detection abnormality in the prior art, and to provide a weak current equipment abnormality judgment method and system based on Internet of Things, which improves the accuracy of abnormality judgment and effectively filters false alarms caused by other interference through dynamic coupling degree analysis.

[0005] To solve the above technical problems, the present application provides a weak current equipment abnormality judgment method based on Internet of Things, comprising the following steps:

[0006] Real-time collection of multi-dimensional operation parameters of the target weak current equipment through the Internet of Things node network, the multi-dimensional operation parameters at least including temperature parameters, current parameters, voltage parameters and vibration frequency parameters, and each Internet of Things node generates a unique dynamic tracking identifier for the collected data;

[0007] The multi-dimensional running parameters with dynamic tracking marks are transmitted to an edge computing node, and dynamic parameter correlation analysis is performed at the edge computing node, specifically including: establishing a dynamic mapping relationship between the current parameters and the historical normal parameter interval, calculating the real-time coupling degree between parameters, when a parameter deviates from its historical normal interval, analyzing the coupling degree change mode of other associated parameters;

[0008] Based on the dynamic parameter correlation analysis result, perform abnormality possibility grading evaluation: mark the phenomenon of single parameter deviation but normal coupling degree as primary abnormality, mark the phenomenon of multiple parameter coordinated deviation and abnormal coupling degree as advanced abnormality, and generate an evaluation report containing abnormality level and device state image;

[0009] Upload the evaluation report to the cloud analysis platform, and the cloud analysis platform starts different levels of verification mechanism according to the abnormality level: start horizontal comparison verification of the same type of device for primary abnormality, and start device full life cycle backtracking verification for advanced abnormality, and generate the final abnormality judgment result after excluding the possibility of false judgment through the verification mechanism.

[0010] In an embodiment of the present application, the dynamic tracking mark includes: a time stamp and device location information, wherein:

[0011] When performing dynamic parameter correlation analysis at the edge computing node, first, establish a same-area device group parameter comparison library according to the device location information in the dynamic tracking mark, and then construct a parameter time sequence change track according to the time stamp information;

[0012] When grading the abnormality possibility, identify the sudden deviation or gradual deviation mode of the parameters in combination with the time stamp information, and compare the running state differences of the same-area device group based on the device location information;

[0013] When verifying at the cloud analysis platform, select the same type of device in the same physical environment for horizontal comparison according to the device location information, and trace the historical evolution trend of the device parameters according to the time stamp information.

[0014] In an embodiment of the present application, the process of establishing a dynamic mapping relationship between the current parameters and the historical normal parameter interval, and calculating the real-time coupling degree between parameters includes:

[0015] A dynamic parameter window is established at the edge computing node, and current and historical parameter data are cached in a sliding time block manner, and features of various parameters are extracted in each time block, the extracted features including but not limited to: parameter fluctuation amplitude feature, change rate feature, periodic repetition feature and abnormal pulse feature;

[0016] Adaptive correlation matrix is adopted to construct dynamic mapping relationship between parameters, each element in the matrix represents the cooperative variation strength of two parameter feature vectors within a sliding time window, and the implicit correlation between parameters is identified by tracking the evolution trend of the matrix elements in real time;

[0017] The adaptive correlation matrix is used to perform coupling degree calculation, and the calculation process includes: first, identifying the leading parameters and following parameters in the parameter group, then analyzing the response delay time and amplitude matching degree of the following parameters when the leading parameters change, and finally comprehensively evaluating the real-time coupling strength between parameters according to the response characteristics.

[0018] In an embodiment of the present application, the process of establishing a dynamic mapping relationship further includes:

[0019] An adaptive reference model of parameters is established at the edge computing node, the temperature and humidity data of the environment where the device is located are collected in real time through the environment perception module, and the environment data are coded into an environment feature vector;

[0020] According to the environment feature vector, the corresponding historical parameter sub-interval library is activated, the sub-interval library is classified and stored according to different environmental conditions, and each sub-interval contains the parameter fluctuation range of at least three complete cycles of normal operation of the device under the environmental condition;

[0021] Dynamic calibration of parameter intervals is performed, the current environment feature vector is matched with the environment features of each sub-interval, when the matching degree exceeds a preset threshold, the sub-interval is taken as a reference interval, otherwise, the three sub-intervals with the highest matching degree are fused to generate a temporary reference interval;

[0022] When calculating the real-time coupling degree, a parameter deviation direction consistency factor is introduced, when multiple parameters deviate from their reference intervals in the same direction, the coupling degree calculation weight is enhanced, and when the deviation directions of the parameters are inconsistent, the coupling degree calculation weight is reduced.

[0023] In an embodiment of the present application, the device state image includes:

[0024] A parameter state matrix, taking a device physical structure diagram as a base plate, marking the real-time state of its associated parameters at the corresponding component positions, including temperature parameters, current parameters, voltage parameters and vibration frequency parameters;

[0025] An abnormal influence range map, based on a device electrical schematic diagram and a mechanical connection diagram, marked with different color areas, including a directly affected core area, a possibly affected associated area and a currently normal isolated area.

[0026] In an embodiment of the present application, the process of starting a horizontal comparison and verification of the same type of device for a primary abnormality includes:

[0027] Filtering the same power distribution circuit, the same physical installation location or the same type of device that has been recently maintained from the device account library as the comparison object based on the installation location characteristics and functional attributes of the abnormal device;

[0028] Collecting the current operating parameters of the comparison object, selecting the comparison object that matches the abnormal device working condition environment as the effective comparison group;

[0029] Comparing and analyzing the abnormal device parameters with the group fluctuation interval of the effective comparison group, marking the specific parameters that exceed the group interval and analyzing the relevance of the specific parameters with the device individual characteristics;

[0030] Generating a verification conclusion of group consistency anomaly, individual specificity anomaly or gradual group anomaly according to the comparison result, which corresponds to different judgments of environmental factor influence, device intrinsic fault or early stage of systematic deterioration, respectively.

[0031] In an embodiment of the present application, the process of full life cycle backtracking verification of high-level abnormal starting device includes:

[0032] Integrating the factory baseline parameters of the device, the historical maintenance records, the historical abnormal data and the environmental change characteristics to construct a complete life cycle digital archive;

[0033] Using a hierarchical backtracking mechanism to sequentially perform short-term parameter gradual change analysis, medium-term performance drift evaluation and full life cycle baseline comparison, and form a comprehensive diagnostic view covering different time scales;

[0034] Based on time series analysis and physical connection relationship tracking, the propagation path of the first abnormal point is tracked to generate a fault evolution chain containing time and space evolution characteristics and deterioration acceleration;

[0035] The comprehensive backtracking analysis result outputs an engineering decision conclusion including root cause judgment, recoverability evaluation, disposal priority suggestion and prevention measure scheme.

[0036] In an embodiment of the present application, the following steps are further included: according to the type and level of the final abnormality judgment result, adaptively triggering the corresponding control strategy, including: primary abnormality starting parameter fine-tuning instruction, high-level abnormality starting device isolation instruction and generating a maintenance work order.

[0037] To solve the above technical problems, the present application also provides a weak current device abnormality judgment system of Internet of Things, comprising:

[0038] A data acquisition module is configured to acquire multi-dimensional operating parameters of a target weak current device in real time through an Internet of Things node network, wherein the multi-dimensional operating parameters at least include temperature parameters, current parameters, voltage parameters and vibration frequency parameters, and each Internet of Things node is configured with a dynamic identification generation unit for generating a unique dynamic tracking identification for the collected data;

[0039] The edge computing module is in communication connection with the data acquisition module, and includes a function of receiving multi-dimensional operation parameters with dynamic tracking identification and performing correlation analysis on the dynamic parameters, including establishing a dynamic mapping relationship between the current parameters and the historical normal parameter interval, calculating the real-time coupling degree between the parameters, and analyzing the coupling degree change mode of other associated parameters when a parameter deviates from its historical normal interval;

[0040] The hierarchical evaluation module is in communication connection with the edge computing module, and is configured to perform abnormal possibility hierarchical evaluation based on the correlation analysis result of the dynamic parameters, mark a phenomenon that a single parameter deviates but the coupling degree remains normal as a primary abnormality, mark a phenomenon that multiple parameters deviate cooperatively and the coupling degree is abnormal as a high-level abnormality, and generate an evaluation report containing the abnormality level and the device state image;

[0041] The cloud analysis platform is in communication connection with the edge computing module, and is configured to start different levels of verification mechanisms according to the abnormality level in the evaluation report, start horizontal comparison verification of the same type of devices for the primary abnormality, start device full life cycle backtracking verification for the high-level abnormality, and generate a final abnormality determination result after excluding the possibility of misjudgment through the verification mechanism.

[0042] In an embodiment of the present application, the control strategy execution module is in communication connection with the cloud analysis platform, and adaptively triggers corresponding control strategies according to the type and level of the final abnormality determination result, including starting parameter fine-tuning instructions for the primary abnormality and starting device isolation instructions for the high-level abnormality and generating a maintenance work order.

[0043] The above technical solutions of the present application have the following advantages compared with the prior art:

[0044] The weak current device abnormality judgment method of the Internet of Things disclosed by the present application performs dynamic parameter correlation analysis at the edge computing node, no longer considers the over-limit condition of each parameter in isolation, but establishes an intelligent evaluation system based on the real-time coupling degree between the parameters.

[0045] Specifically, when it is detected that a parameter deviates from the historical normal interval, the coupling degree change mode of other associated parameters is analyzed synchronously, and this dynamic correlation analysis can effectively distinguish between normal fluctuations and real abnormalities, for example, when the temperature parameter alone increases while the associated parameters such as current and vibration remain in normal coupling relationship, it can be accurately judged that this may be a normal fluctuation caused by environmental temperature change; on the contrary, when multiple parameters deviate cooperatively and the coupling degree is abnormal, it can be determined that the device is a real fault.

[0046] And, through the hierarchical evaluation mechanism, the intelligent classification of "primary exception" and "high-level exception" is introduced, the accurate identification of the abnormal situation is realized, the multi-level verification mechanism adopted by the cloud analysis platform provides double protection for the judgment result, and the differentiated verification process is started for different levels of exceptions, which ensures the accuracy of the judgment and optimizes the allocation of system resources.

[0047] From the technical effect, the weak current equipment abnormality judgment method of the Internet of Things can significantly improve the accuracy of abnormality judgment, effectively filter false positives caused by environmental interference through dynamic coupling degree analysis, and at the same time, the hierarchical evaluation mechanism can discover potential fault signs as soon as possible to avoid small faults from evolving into big problems; through the collaborative processing architecture of edge computing and cloud platform, the real-time requirement is considered while ensuring the analysis accuracy, which provides a complete solution for the intelligent operation and maintenance of weak current equipment. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to make the content of the present application more easily understood, the present application will be further described in detail below according to specific embodiments of the present application and in conjunction with the drawings, in which:

[0049] Figure 1 is a step flow chart of the weak current equipment abnormality judgment method of the Internet of Things of the present application;

[0050] Figure 2 is a step flow chart of the present application for constructing a dynamic mapping relationship and calculating a real-time coupling degree;

[0051] Figure 3 is a step flow chart for solving the influence of the environment on the weak current equipment abnormality judgment when constructing a dynamic mapping relationship of the present application;

[0052] Figure 4 is a step flow chart of the present application for starting the same type of equipment horizontal comparison verification for primary exceptions;

[0053] Figure 5 is a step flow chart of the present application for starting the equipment full life cycle backtracking verification for high-level exceptions;

[0054] Figure 6 is a structure framework diagram of the weak current equipment abnormality judgment system of the Internet of Things of the present application. DETAILED DESCRIPTION

[0055] The present application will be further described below in conjunction with the drawings and specific embodiments, so that those skilled in the art can better understand the present application and implement it, but the embodiments are not limiting to the present application.

[0056] Referring to Figure 1 , the present application discloses a weak current equipment abnormality judgment method of the Internet of Things, comprising the following steps:

[0057] S10, collecting multi-dimensional running parameters of the target weak current equipment in real time through the Internet of Things node network, the multi-dimensional running parameters at least including temperature parameters, current parameters, voltage parameters and vibration frequency parameters, each Internet of Things node generating a unique dynamic tracking identifier for the collected data;

[0058] In the embodiment, the collection of multi-dimensional running parameters breaks through the limitation of traditional single parameter monitoring, and provides a comprehensive data basis for subsequent analysis through the synchronous collection of multi-dimensional parameters. Compared with the prior art which only focuses on a few key parameters, this multi-dimensional data collection method can capture the complete portrait of the equipment running state, and especially the introduction of non-traditional monitoring parameters such as vibration frequency can discover potential faults of mechanical parts earlier. At the same time, the addition of the dynamic tracking identifier ensures the traceability of the data in the transmission and processing process, and creates conditions for cross-node and cross-period data correlation analysis.

[0059] S20, transmitting the multi-dimensional running parameters with the dynamic tracking identifier to an edge computing node, and performing dynamic parameter correlation analysis at the edge computing node, specifically including: establishing a dynamic mapping relationship between the current parameters and the historical normal parameter interval, calculating the real-time coupling degree between the parameters, and analyzing the coupling degree change mode of other associated parameters when a parameter deviates from its historical normal interval;

[0060] The traditional method usually performs simple threshold comparison in the cloud, while the present scheme establishes a dynamic mapping relationship between the current parameters and the historical normal interval at the edge side, and calculates the real-time coupling degree between the parameters. The working principle is that a dynamic model established through machine learning analyzes the interaction relationship between the parameters in real time, and when a parameter deviates from the normal interval, the system does not immediately alarm, but further analyzes the coupling degree change mode of other associated parameters.

[0061] This processing method has the following advantages: on the one hand, edge computing reduces data transmission delay, making the analysis more timely; on the other hand, dynamic coupling degree analysis can identify the nonlinear relationship between parameters, which is completely impossible for fixed threshold method. For example, some initial faults may be manifested as a slight change in multiple parameters at the same time, and each parameter is within the normal range, but the coupling degree is abnormal. This potential fault can only be effectively identified through the present scheme.

[0062] S30, based on the dynamic parameter correlation analysis result, performing abnormality possibility grading evaluation: marking a phenomenon that a single parameter deviates but the coupling degree remains normal as a primary abnormality, marking a phenomenon that multiple parameters deviate cooperatively and the coupling degree is abnormal as a high-level abnormality, and generating an evaluation report containing the abnormality level and the equipment state portrait;

[0063] According to the dynamic analysis results, the anomalies are divided into two levels of primary and advanced. The design principle of this grading mechanism is based on the general law of fault development: single parameter anomaly often corresponds to temporary disturbance or early potential problem, and multi-parameter coordinated anomaly usually indicates substantive failure. Compared with the alarm mode of the prior art, this grading evaluation has multiple advantages: first, it greatly reduces the false alarm rate and avoids overreaction to temporary fluctuations; second, by distinguishing the anomaly level, it provides more accurate decision-making basis for operation and maintenance personnel; finally, the generated device state portrait can intuitively show the fault characteristics, significantly improving the fault diagnosis efficiency, especially the identification ability of primary anomalies, so that the system can issue a warning before the fault causes substantial damage, realizing true preventive maintenance.

[0064] S40, upload the evaluation report to the cloud analysis platform, and the cloud analysis platform starts different levels of verification mechanisms according to the anomaly level: for primary anomalies, start the same type of device horizontal comparison verification, and for advanced anomalies, start the device full life cycle backtracking verification, and generate the final anomaly determination result after excluding the possibility of false judgment through the verification mechanism;

[0065] The cloud analysis platform designs differentiated verification paths according to anomalies, because different types of anomalies have different sources of false judgments, and targeted verification strategies must be used to effectively filter false alarms, including:

[0066] For primary anomalies, start the same type of device horizontal comparison verification. If the same parameters of other similar devices in the same environment also appear similar fluctuations, it is likely that the environmental disturbance (such as power grid transient, temperature change) leads to systemic phenomena, rather than device failure itself. This verification method can efficiently identify and filter out most of the false alarms caused by external environment;

[0067] For advanced anomalies, trigger full life cycle backtracking verification. Because real device failure often has progressive characteristics in the time dimension, by tracing the evolution trend of the same type of parameters in the history of the device (such as slow rise of vibration frequency, gradual distortion of current waveform), it can distinguish between sudden disturbance (such as lightning strike, operation impact) and real device degradation process. This verification mechanism is particularly good at excluding false judgments caused by short-term abnormal disturbances (such as lightning strikes, operation impacts), while improving the identification accuracy of potential faults.

[0068] The technical solution constructs a complete closed-loop system of "dynamic perception-correlation analysis-hierarchical evaluation-verification error elimination", constructs a time-space continuous equipment state image through multi-dimensional parameter collection and dynamic tracking identification, provides rich and structured input for subsequent analysis; introduces real-time coupling degree calculation on the edge side, captures the implicit physical correlation between parameters, and thus distinguishes between real faults and random noise; through the cooperation of hierarchical evaluation and verification mechanism, a progressive diagnosis process is formed: first, preliminary screening is carried out through coupling degree analysis, and then false positives are eliminated through differential verification. The distributed computing capability of the Internet of Things is deeply integrated with the physical characteristics of equipment failure, and the contradiction between false positive rate and false negative rate in traditional methods is solved.

[0069] In combination with actual application, the abnormal performance of weak current equipment often has significant time correlation and space correlation, therefore, the dynamic tracking identification includes: time stamp and equipment location information, which are processed in combination with the time stamp and equipment location information in subsequent dynamic parameter correlation analysis, abnormal hierarchical evaluation and cloud platform verification, and the specific process includes:

[0070] When the edge computing node obtains the running parameters with the time stamp and equipment location information, first, the equipment location information is used to intelligently form a same-area equipment group comparison library, because the equipment under the same physical environment (such as the same power distribution cabinet, the same floor) will be disturbed by similar environment, after establishing this space correlation model, when the parameters of a certain equipment are abnormal, the running states of the neighboring equipment can be compared immediately, if similar fluctuations occur in multiple equipment in the same area at the same time, it can be determined that it is caused by environmental factors rather than equipment failure, and this group intelligence judgment based on location information significantly improves the anti-interference capability; at the same time, the time sequence change track of the parameters is constructed through the time stamp information, which not only includes simple historical data backtracking, but more importantly, identifies the pattern characteristics of parameter deviation-mutant deviation usually corresponds to operation error or instantaneous impact and other external interference, while gradual deviation may indicate equipment aging or potential failure, and this time domain pattern recognition enables the system to distinguish between sudden abnormalities and chronic faults.

[0071] When entering the abnormal hierarchical evaluation stage, the synergistic effect of time-space information is further exerted, the time stamp information is used to assist in judging the nature of the abnormality, for example, if a temperature parameter suddenly rises within a few minutes and its historical data presents periodic fluctuation characteristics, it can be inferred that it is temporary overload rather than equipment failure; and the equipment location information is used to perform state comparison in the space dimension, for example, if multiple sensors in the same cable trench simultaneously detect temperature rise, it is likely to be a systemic risk caused by poor heat dissipation rather than a single equipment problem; this time-space cross verification mechanism greatly improves the accuracy of primary abnormality judgment, effectively avoiding the overreaction of traditional methods to temporary fluctuations.

[0072] In the cloud platform verification link, the value of space-time information is most fully reflected. A natural reference system is formed by the same environment and same type device group selected based on the device location information, and through horizontal comparison, the individual device that is truly abnormal can be quickly identified. This analysis method based on spatial clustering is particularly suitable for identifying faults caused by local factors (such as poor contact and local overheating). The whole life cycle tracing supported by the time stamp information reveals deeper rules. For example, if a vibration parameter shows a slow upward trend in the past three months, even if the current value is still within the theoretical threshold range, the system can still give an early warning of bearing wear risk. The verification strategy combining space-time dimensions enables the system to quickly respond to sudden failures and capture progressive degradation, ensuring real-time while achieving predictive maintenance.

[0073] From the technical effect, the introduction of space-time identification enables the system to have three-dimensional judgment ability (parameter value, time, and space). Compared with the traditional single parameter threshold method, this multi-dimensional analysis method can identify "what kind of abnormality occurs at what time, where, and under what conditions", thereby upgrading the abnormality judgment from simple state monitoring to real fault root cause analysis. It is particularly worth noting that in this scheme, the space-time information is not used in isolation, but forms a closed loop through the cooperation mechanism of "edge real-time processing-cloud deep verification". The time stamp ensures the continuity of the analysis time sequence, and the location information guarantees the accuracy of the spatial correlation. The organic combination of the two ultimately realizes the accurate diagnosis of weak electric equipment abnormalities in complex industrial environments.

[0074] In actual engineering, the running state of weak electric equipment is affected by many factors such as load change and environmental interference. The parameter normal interval itself has dynamics, and there is a complex interaction relationship between different parameters. In this embodiment, by establishing a dynamic correlation model between parameters, the complex interaction relationship between different parameters is explored, which can distinguish between real faults and normal working condition fluctuations. Specifically, referring to FIG. 6, a dynamic mapping relationship between the current parameter and the historical normal parameter interval is established, and the process of calculating the real-time coupling degree between parameters includes: Figure 2

[0075] S21, a dynamic parameter window is established at the edge node to cache current and historical parameter data in a sliding time block manner. Feature extraction is performed on various parameters in each time block. The extracted features include but are not limited to: parameter fluctuation amplitude feature, change rate feature, periodic repetition feature, and abnormal pulse feature.

[0076] ​This sliding time block processing method not only retains the recent trend of parameter changes (such as data in the past 15 minutes), but also avoids the computational burden of processing the entire historical data; the feature extraction performed within each time block is not simply recording numerical values, but capturing the essential characteristics of parameter behavior - fluctuation amplitude reflects the degree of deviation, change rate identifies abnormal mutations, periodicity feature discovers regular disturbances, and pulse feature captures transient faults. This multi-dimensional feature description provides rich and accurate input for subsequent analysis.

[0077] S22, a dynamic mapping relationship between parameters is constructed using an adaptive correlation matrix, each element in the matrix representing the cooperative change strength of two parameter feature vectors within a sliding time window, and the evolution trend of the matrix elements is tracked in real time to identify the implicit correlation between parameters;

[0078] In this embodiment, each element in the matrix reflects the cooperative change strength of two parameter features in real time, thereby capturing the dynamic relationship between parameters; for example, it is found in engineering practice that the transformer temperature and load current have an approximately quadratic relationship under normal circumstances, but when the heat dissipation system is abnormal, this relationship will change. By tracking the evolution of the matrix elements in real time, this implicit correlation change can be found, which is more sensitive and reliable than monitoring temperature or current values alone; the matrix update mechanism uses a sliding window approach, which ensures real-time performance and maintains the stability of the analysis. This balanced design is specifically designed for the characteristics of industrial field data that are often disturbed by noise.

[0079] S23, based on the adaptive correlation matrix, the coupling degree is calculated, the calculation process including: first identifying the dominant parameter and the following parameter in the parameter group, then analyzing the response delay time and amplitude matching degree of the following parameter when the dominant parameter changes, and finally evaluating the real-time coupling strength between parameters based on the response characteristics;

[0080] The coupling degree calculation based on the correlation matrix adopts a dominant-following analysis framework, which is derived from the thinking mode of engineers when diagnosing faults: engineers usually first identify the parameter that first appears abnormal (dominant parameter), and then observe the response of other parameters (following parameters) to determine the nature of the fault. By identifying the dominant-following relationship in the parameter group, analyzing the response delay (such as temperature change usually lagging behind current change) and amplitude matching degree (such as whether the vibration increase amplitude is proportional to the current rise), the final coupling strength evaluation contains both numerical relationship and time sequence characteristics. This multi-dimensional coupling analysis can accurately distinguish between real faults (parameter coupling relationship damage) and measurement disturbances (only single parameter abnormality).

[0081] In engineering application environment, the normal range of operating parameters of the same device under different environmental conditions (such as high temperature in summer and low temperature in winter) has significant differences, and it is difficult for the traditional method to accurately reflect this complex change using fixed threshold or simple seasonal adjustment factor, therefore, when establishing the dynamic mapping relationship, the influence of environment on weak current equipment abnormality judgment should also be considered, and the method shown in Figure 3 further comprises the following steps:

[0082] S24, first, the environmental perception module is used to collect real-time key environmental data such as temperature and humidity, which are coded as feature vectors instead of simple numerical records, because the environmental influence in engineering field is often multi-factor coupled (such as the combined effect of high temperature and high humidity is much greater than the influence of single factor), and this coding method retains the interaction information between environmental factors, which lays a foundation for subsequent accurate matching.

[0083] S25, the environmental feature vector is used to activate the corresponding historical parameter sub-interval library, which uses the historical operation data of the device as the judgment benchmark, each sub-interval library stores the parameter fluctuation range of the device in a complete operation cycle (such as day and night, production batch, etc.) under specific environmental conditions, ensuring that the benchmark data has statistical representativeness, and through this environment-parameter association storage method, the closest historical working condition of the current environment can be automatically identified, solving the problem of poor environmental adaptability of traditional methods.

[0084] S26, when performing dynamic calibration of parameter interval, instead of simply selecting the most matched sub-interval, an intelligent fusion mechanism is designed: for the environment conditions with obvious matching, the corresponding sub-interval is directly used, and for the transition environment or abnormal environment (such as extreme weather), a temporary benchmark is generated through multi-sub-interval weighted fusion; this processing method not only ensures the accuracy of judgment under common environment, but also ensures the robustness of the system under abnormal environment. For example, when the temperature fluctuates greatly in spring, the sub-interval data of winter and summer will be automatically fused to generate a reasonable benchmark that conforms to the current transition season, avoiding misjudgment caused by environmental gradual change, the generation process of temporary benchmark interval considers the similarity weight of environmental characteristics, so that the final benchmark is more consistent with the current actual conditions, and this dynamic adjustment mechanism greatly improves the adaptability to climate change.

[0085] S27, the parameter deviation direction consistency factor introduced in the coupling degree calculation link, engineering experience shows that real faults usually cause the related parameters to change in the same direction (such as the current and temperature rising at the same time when overloaded), and measurement interference or local anomalies often show inconsistent parameter change directions. By analyzing the consistency degree of the parameter deviation direction, the systematic fault and the local anomaly can be effectively distinguished. This design makes the coupling degree evaluation more in line with the actual operation rules of the equipment. Specifically, when multiple parameters deviate upward or downward synchronously, the coupling weights between these parameters are enhanced, because such coordinated changes are more likely to reflect real faults; and when the parameter change directions are chaotic, the weights are reduced to avoid misjudging irrelevant fluctuations as related faults. This weight adjustment strategy based on physical mechanism significantly improves the accuracy of composite fault judgment and reduces false alarms caused by environmental fluctuations.

[0086] Specifically, when generating an evaluation report containing an abnormality level and a device state image, it is necessary to design a device state image from the perspective of engineering application, so that each visual element is targeted at the actual needs of field operation, and the device state image includes a parameter state matrix and an abnormality influence range diagram, wherein:

[0087] The parameter state matrix uses the physical structure of the equipment as a display board, and labels the real-time state of its associated parameters at the corresponding component position. For example, the temperature parameter is displayed in the form of a heat map at the heating part of the equipment, the current parameter is displayed in the form of a dynamic flow bar on the circuit trace, the vibration frequency parameter is displayed in the form of a frequency spectrum diagram at the mechanical moving part, and the voltage parameter is displayed in the form of a color step diagram at the power supply access point. This design enables engineers to intuitively locate abnormal parameters to specific components, eliminating the conversion process between parameter numbers and physical positions.

[0088] The abnormality influence range diagram is generated based on the electrical and mechanical connection relationship of the equipment, and is marked with different color areas, for example, a red area clearly identifies the part that needs to be handled first, and a yellow area prompts the associated components that may be involved. This display method helps the operation and maintenance team to predict the impact of the fault and prepare maintenance plans for the associated components in advance.

[0089] Referring to Figure 4 The process of starting the transverse comparison and verification of the same type of equipment based on the primary anomaly includes the following steps:

[0090] S411, based on the installation location characteristics and functional attributes of the abnormal equipment, the same power distribution circuit, the same physical installation location, or the same type of equipment that has been maintained recently are selected from the equipment account library as comparison objects;

[0091] Based on the physical location and electrical characteristics of the abnormal device, the same type of device with comparable performance is intelligently screened, the devices in the same power distribution circuit share the same power quality, the devices in the same installation location bear similar environmental stress, and the devices recently maintained provide a health reference, and this multi-dimensional screening strategy ensures the representativeness of the comparison sample.

[0092] S412, collect the current operating parameters of the comparison object, select the comparison object matched with the working condition environment of the abnormal device as the effective comparison group;

[0093] The working condition matching of the preliminary comparison object includes quantitative evaluation of three dimensions of load rate, running time and environmental similarity, and the devices with large differences in actual running state although they are of the same type are excluded to form a truly comparable effective comparison group, and this dynamic screening mechanism overcomes the limitations of the traditional method of simply comparing by device type.

[0094] S413, compare the abnormal device parameters with the group fluctuation interval of the effective comparison group, mark the specific parameters that exceed the group interval and analyze their association with the individual characteristics of the device;

[0095] After obtaining the effective comparison group, instead of using a fixed theoretical threshold, a dynamic normal interval is established based on the actual running data of the group device, and the parameter deviation of the abnormal device is compared with this living benchmark, and this group intelligence analysis method can adapt to environmental changes, for example, when the environmental temperature generally rises, causing the temperature parameters of all similar devices to rise synchronously, it can identify that this is a normal phenomenon rather than an individual device failure.

[0096] S414, generate verification conclusions of group consistency anomaly, individual specificity anomaly or gradual group anomaly according to the comparison results, corresponding to different judgments of environmental factor influence, device intrinsic failure or early system degradation;

[0097] Specifically, three typical conclusions are formed through verification decision, each of which corresponds to different engineering disposal schemes: group anomaly prompts to check the public system (such as power supply and heat dissipation), individual anomaly points to device intrinsic maintenance, and gradual anomaly warns that batch replacement may be needed; This structured output directly interfaces with the field maintenance process, so that the Internet of Things data analysis result can be seamlessly integrated into the existing operation and maintenance system.

[0098] Referring to Figure 5 The process of starting the full life cycle backtracking verification of the advanced abnormal device includes the following steps:

[0099] S421, integrate the factory reference parameters of the device, the history maintenance records, the historical abnormal data and the environmental change characteristics to construct a complete life cycle digital file;

[0100] By integrating the full-dimensional historical data of the equipment, a life cycle digital archive is constructed, which is not simply a data accumulation, but a spatio-temporal alignment and structured processing of the factory baseline parameters, historical maintenance records, abnormal data and environmental change characteristics, forming a traceable "genetic map" of the equipment; for example, when intermittent signal interruption occurs in a certain communication equipment, the signal stability test curve of the equipment at the time of leaving the factory, the batch information of the past replaced modules, the handling records of similar faults in history and the temperature and humidity change data of the installation environment can be quickly retrieved from the archive, and the organic integration of these information provides a stereoscopic data basis for subsequent analysis.

[0101] S422, using a hierarchical backtracking mechanism to sequentially perform short-term parameter gradual change analysis, medium-term performance drift evaluation and full life cycle benchmark comparison, forming a comprehensive diagnostic view covering different time scales;

[0102] The hierarchical backtracking mechanism is used for multi-time scale analysis, and the technical mechanism is that different types of fault clues are distributed in different time dimensions, among which: short-term parameter gradual change analysis focuses on data changes in the last 72 hours, through minute-level parameter reproduction, it can capture subtle signs before failure, such as periodic current fluctuations of a power module that occur 48 hours before complete failure; medium-term performance drift evaluation focuses on parameter change trends within 3 months, this cross-cycle analysis can identify progressive degradation of the equipment, such as the slow increase of connector contact resistance; full life cycle benchmark comparison compares the current state with the factory standard to determine whether the equipment has undergone essential changes; specifically, the analysis of these three time dimensions is not isolated, but mutually verified - short-term abnormalities may find explanations in medium-term trends, and long-term deviations can provide reference benchmarks for medium-term changes, which greatly improves the reliability of diagnostic conclusions.

[0103] S423, based on time series analysis and physical connection relationship tracking, the propagation path of the first abnormal point is generated, and a fault evolution chain containing spatio-temporal evolution characteristics and degradation acceleration is generated;

[0104] The generation process of the fault evolution chain integrates physical connection analysis and time series pattern recognition, first determines the spatio-temporal coordinates of the abnormal origin, then along the electrical connection and mechanical transmission path of the equipment, combined with the component replacement information in the maintenance record, constructs the propagation model of the fault, and especially pays attention to the time delay effect and amplitude amplification effect of abnormal propagation in this process, for example, when an abnormal motor current is found, the vibration change of the mechanical load driven by the motor is traced back, and the time correlation and amplitude relationship between the two are analyzed, to determine whether it is a problem of the motor itself or caused by the load abnormality, this physical mechanism-based tracing method can effectively avoid misjudgment that may be caused by pure data-driven analysis.

[0105] S424, the comprehensive backtracking analysis result outputs an engineering decision conclusion including root cause judgment, recoverability assessment, treatment priority suggestion, and preventive measure scheme;

[0106] The engineering decision conclusion generation of the embodiment is not a simple abnormality classification, but forms a treatment scheme including four key dimensions: the root cause judgment is based on full life cycle data tracing, and distinguishes different reasons such as design defects, material aging, improper operation, or environmental factors; the recoverability assessment comprehensively considers the current state of the equipment, maintenance history, and service life, and judges the economic efficiency and technical feasibility of maintenance; the treatment priority suggestion determines the response urgency according to the fault propagation speed and influence range; the preventive measure scheme specifically proposes suggestions such as design improvement, operation specification, or monitoring enhancement, and this structured output mode enables the complex analysis result to be directly converted into an engineering action guide, greatly improving the practical value of the Internet of Things monitoring system.

[0107] Specifically, in actual application, after the weak current equipment abnormality is judged, the following steps are further included:

[0108] S50, according to the type and level of the final abnormality judgment result, the corresponding control strategy is adaptively triggered:

[0109] The parameter fine-tuning instruction is started for the primary abnormality, the equipment isolation instruction is started for the senior abnormality, and a maintenance work order is generated, and the historical normal parameter interval and the coupling degree benchmark of the edge computing node are updated.

[0110] After the abnormality judgment is completed, a differentiated response mechanism is designed according to the different types and levels of the abnormality, wherein: for the primary abnormality, an alarm or shutdown is not immediately triggered, but parameter fine-tuning is first attempted, and this flexible processing mode fully considers the parameter fluctuation characteristics commonly seen in industrial environments; for example, when it is detected that the temperature parameter of a power distribution cabinet is slightly higher, but the related parameters such as current and vibration are normal, the ventilation equipment speed is automatically adjusted or the load is slightly reduced, this mild intervention avoids production interruption and effectively controls the risk.

[0111] The generation of the parameter fine-tuning instruction is not a simple numerical adjustment, but a calculation of a safe and economical adjustment range in combination with the equipment operation history and environmental conditions, so as to ensure that a new problem is not caused by excessive adjustment.

[0112] For the senior abnormality, a more strict equipment isolation process is started, which is different from the simple power-off protection, and the isolation instruction of the present scheme is executed in stages: firstly, the current running state and data snapshot are recorded, then the orderly shutdown is executed according to the equipment operation procedure, and finally the maintenance work order containing detailed diagnostic information is generated; this standardized isolation procedure avoids secondary damage caused by sudden power-off, and also reserves complete fault scene data for subsequent maintenance.

[0113] The generation process of the maintenance work order integrates equipment profile information, abnormal feature analysis and historical maintenance records, not only points out the fault phenomenon, but also provides possible root cause analysis and maintenance suggestions, greatly improving the efficiency of on-site maintenance. For example, when a motor system has multiple parameter abnormalities such as current, temperature and vibration, the generated work order will mark the correlation pattern of abnormal parameters, prompt to check the bearing or winding, and attach the handling record of the last similar fault. This intelligent work order system significantly shortens the fault troubleshooting time.

[0114] In addition, after each abnormality treatment, whether it is parameter fine-tuning or equipment isolation, the reference benchmark of the edge computing node will be updated according to the treatment effect. For primary abnormalities, the historical normal parameter interval is mainly updated, and a gradual adjustment strategy is adopted to absorb new normal operation data while retaining the original benchmark characteristics, avoiding benchmark drift caused by single abnormality. For advanced abnormalities, the coupling degree benchmark is updated synchronously, and the correlation model between parameters is recalibrated. This dynamic learning capability enables the system to adapt to changes in operating characteristics caused by long-term factors such as equipment aging and environmental changes. For example, after seasonal maintenance, the performance of a weak current equipment changes. The system can quickly adapt to the new operating state by continuously updating the benchmark data, maintaining monitoring accuracy.

[0115] To implement the above method, referring to Figure 6 The application also discloses a weak current equipment abnormality judgment system for Internet of Things, which can implement the above method. The system comprises:

[0116] A data acquisition module is configured to acquire multi-dimensional operating parameters of the target weak current equipment in real time through the Internet of Things node network, wherein the multi-dimensional operating parameters at least include temperature parameters, current parameters, voltage parameters and vibration frequency parameters. Each Internet of Things node is provided with a dynamic identification generation unit for generating a unique dynamic tracking identification for the acquired data.

[0117] An edge computing module is in communication connection with the data acquisition module and comprises a module for receiving multi-dimensional operating parameters with dynamic tracking identification and performing correlation analysis on dynamic parameters, including establishing a dynamic mapping relationship between current parameters and historical normal parameter intervals and calculating the real-time coupling degree between parameters. When a parameter deviates from its historical normal interval, the coupling degree change pattern of other associated parameters is analyzed.

[0118] A hierarchical evaluation module is in communication connection with the edge computing module and is configured to perform abnormality possibility hierarchical evaluation based on the correlation analysis results of dynamic parameters. Single parameter deviation but normal coupling degree is marked as a primary abnormality, and multiple parameter deviation and abnormal coupling degree are marked as an advanced abnormality. An evaluation report containing abnormality level and equipment state image is generated.

[0119] The cloud analysis platform is in communication connection with the edge computing module, and is configured to start different levels of verification mechanisms according to the abnormality level in the evaluation report, to start the same type of device horizontal comparison verification for the primary abnormality, to start the device full life cycle backtracking verification for the high-level abnormality, and to generate a final abnormality determination result after excluding the possibility of misjudgment through the verification mechanism.

[0120] Specifically, the system further comprises a control strategy execution module in communication connection with the cloud analysis platform, which is configured to adaptively trigger corresponding control strategies according to the type and level of the final abnormality determination result, including starting the parameter fine-tuning instruction for the primary abnormality, and starting the device isolation instruction for the high-level abnormality and generating a maintenance work order.

[0121] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0122] Obviously, the above embodiments are only examples for clarity and are not limited to the embodiments. Based on the above description, other different forms of changes or variations can be made by those skilled in the art. Here, it is not necessary and impossible to enumerate all the embodiments. The obvious changes or variations derived therefrom are still within the protection scope of the present application.

Claims

1. A method for judging an abnormality of a weak current device of an Internet of Things, characterized by comprising: The method comprises the following steps: Real-time collection of multi-dimensional running parameters of the target weak current equipment through an Internet of Things node network, the multi-dimensional running parameters at least including temperature parameters, current parameters, voltage parameters and vibration frequency parameters, each Internet of Things node generating a unique dynamic tracking identifier for the collected data; Transmission of the multi-dimensional running parameters with the dynamic tracking identifier to an edge computing node, dynamic parameter correlation analysis being performed at the edge computing node, specifically including: establishment of a dynamic mapping relationship between the current parameters and historical normal parameter intervals, calculation of real-time coupling degrees among the parameters, analysis of a coupling degree change mode of other associated parameters when a certain parameter deviates from its historical normal interval; Based on the dynamic parameter correlation analysis result, performing abnormality possibility grading evaluation: marking a phenomenon of single parameter deviation but normal coupling degree as a primary abnormality, marking a phenomenon of multiple parameter collaborative deviation and abnormal coupling degree as a high-level abnormality, and generating an evaluation report containing the abnormality level and the equipment state image; Uploading the evaluation report to a cloud analysis platform, the cloud analysis platform starting different levels of verification mechanisms according to the abnormality level: starting horizontal comparison verification of the same type of equipment for the primary abnormality, and starting equipment full life cycle backtracking verification for the high-level abnormality, and generating a final abnormality determination result after excluding the possibility of misjudgment through the verification mechanism.

2. The weak current device anomaly determination method for the Internet of Things according to claim 1, characterized in that: The dynamic tracking identifier comprises: a time stamp and device location information, wherein: When performing dynamic parameter correlation analysis at the edge computing node, first, a same-region device group parameter comparison library is established according to the device location information in the dynamic tracking identifier, and then a parameter time sequence change track is constructed according to the time stamp information; When performing abnormality possibility grading evaluation, the time stamp information is used to identify the sudden deviation or gradual deviation mode of the parameters, and the device location information is used to compare the running state differences of the same-region device group; When verifying at the cloud analysis platform, the same type of equipment in the same physical environment is selected for horizontal comparison according to the device location information, and the historical evolution trend of the equipment parameters is traced according to the time stamp information. 3.The IoT weak current device anomaly determination method of claim 1, wherein: The process of establishing a dynamic mapping relationship between the current parameters and the historical normal parameter intervals and calculating the real-time coupling degrees among the parameters comprises: At the edge computing node, a dynamic parameter window is established to cache the current and historical parameter data in a sliding time block manner, and features of various parameters are extracted in each time block, the extracted features including but not limited to: parameter fluctuation amplitude features, change rate features, periodic repetition features and abnormal pulse features; An adaptive correlation matrix is used to construct the dynamic mapping relationship between the parameters, each element in the matrix representing the collaborative change strength of two parameter feature vectors within the sliding time window, and the evolution trend of the matrix elements is tracked in real time to identify the implicit correlation between the parameters; Based on the adaptive correlation matrix, the coupling degree is calculated, the calculation process including: first, identifying the dominant parameters and the following parameters in the parameter group, then analyzing the response delay time and amplitude matching degree of the following parameters when the dominant parameters change, and finally evaluating the real-time coupling strength between the parameters based on the response characteristics.

4. The weak current device anomaly judgment method of the Internet of Things according to claim 3, characterized in that: The process of establishing the dynamic mapping relationship further comprises: A parameter adaptive benchmark model is established at the edge computing node, real-time temperature and humidity data of the environment where the equipment is located are collected through an environment perception module, and the environment data are coded into an environment feature vector; According to the environment feature vector, a corresponding historical parameter sub-interval library is activated, the sub-interval library is stored in different environment conditions, and each sub-interval contains a parameter fluctuation range of at least three complete cycles of normal operation of the equipment under the environment condition; A parameter interval dynamic calibration is performed, the current environment feature vector is matched with the environment features of each sub-interval, when the matching degree exceeds a preset threshold, the sub-interval is taken as a benchmark interval, otherwise, the three sub-intervals with the highest matching degree are weighted and fused to generate a temporary benchmark interval; When calculating the real-time coupling degree, a parameter deviation direction consistency factor is introduced, when multiple parameters deviate from their benchmark intervals in the same direction, the coupling degree calculation weight is enhanced, and when the parameter deviation directions are inconsistent, the coupling degree calculation weight is reduced.

5. The weak current device anomaly judgment method of the Internet of Things according to claim 1, characterized in that: The device state image includes: A parameter state matrix, taking a device physical structure diagram as a base plate, marking the real-time state of the associated parameters of the corresponding components, including temperature parameters, current parameters, voltage parameters and vibration frequency parameters; An abnormal influence range map, based on a device electrical schematic diagram and a mechanical connection diagram, marked with different color areas, including a directly affected core area, a possibly affected associated area and a currently normal isolated area.

6. The weak current device anomaly determination method for the Internet of Things according to claim 1, characterized in that: The process of starting horizontal comparison and verification of the same type of equipment for primary abnormalities includes: Based on the installation location characteristics and functional attributes of the abnormal equipment, the same power distribution circuit, the same physical installation location or the same type of equipment that has been recently maintained are selected from the equipment account library as comparison objects; The current operating parameters of the comparison objects are collected, and the comparison objects matching the working condition environment of the abnormal equipment are selected as effective comparison groups; The abnormal equipment parameters are compared and analyzed with the group fluctuation interval of the effective comparison groups, the specific parameters exceeding the group interval are marked and the correlation with the individual characteristics of the equipment is analyzed; According to the comparison results, the verification conclusions of group consistency abnormalities, individual specificity abnormalities or gradual group abnormalities are generated, which correspond to different judgments of environmental factor influence, equipment intrinsic failure or early stage of systematic deterioration.

7. The weak current device anomaly determination method for the Internet of Things according to claim 1, characterized in that: The process of starting device full life cycle backtracking verification for advanced abnormalities includes: Integrating the factory benchmark parameters of the equipment, the historical maintenance records, the historical abnormal data and the environmental change characteristics to construct a complete life cycle digital file; Using a hierarchical backtracking mechanism to sequentially perform short-term parameter gradual change analysis, medium-term performance drift evaluation and full life cycle benchmark comparison to form a comprehensive diagnostic view covering different time scales; Based on time series analysis and physical connection relationship, the propagation path of the first abnormal point is tracked to generate a fault evolution chain containing time and space evolution characteristics and deterioration acceleration; The backtracking analysis results are output to generate engineering decision conclusions including root cause judgment, recoverability evaluation, disposal priority suggestion and prevention measure scheme. 8.The method of claim 1, wherein: Further comprising the following steps: According to the type and level of the final abnormality judgment result, the corresponding control strategy is adaptively triggered, including parameter fine-tuning instructions for primary abnormalities, equipment isolation instructions for advanced abnormalities and generation of maintenance work orders. 9.A system for judging an abnormality of a weak current device of an Internet of Things, characterized by comprising: ​ The data acquisition module is configured to acquire multi-dimensional operation parameters of the target weak current equipment in real time through the Internet of Things node network, wherein the multi-dimensional operation parameters at least include temperature parameters, current parameters, voltage parameters and vibration frequency parameters, each Internet of Things node is configured with a dynamic identification generation unit for generating a unique dynamic tracking identification for the acquired data; The edge computing module is in communication connection with the data acquisition module, and includes a module for receiving the multi-dimensional operation parameters with the dynamic tracking identification and performing correlation analysis on the dynamic parameters, including establishing a dynamic mapping relationship between the current parameters and the historical normal parameter interval, calculating the real-time coupling degree between the parameters, and analyzing the coupling degree change mode of other associated parameters when a parameter deviates from its historical normal interval; The hierarchical evaluation module is in communication connection with the edge computing module, and is configured to perform abnormal possibility hierarchical evaluation based on the dynamic parameter correlation analysis result, mark a single parameter deviation but normal coupling degree phenomenon as a primary abnormality, mark a multi-parameter coordinated deviation and abnormal coupling degree phenomenon as a high-level abnormality, and generate an evaluation report containing the abnormality level and the equipment state image; The cloud analysis platform is in communication connection with the edge computing module, and is configured to start different levels of verification mechanisms according to the abnormality level in the evaluation report, start horizontal comparison verification of the same type of equipment for the primary abnormality, start equipment full life cycle backtracking verification for the high-level abnormality, and generate a final abnormality judgment result after excluding the possibility of misjudgment through the verification mechanism. 10.The IoT weak current device anomaly determination system of claim 9, wherein: Further comprising: The control strategy execution module is in communication connection with the cloud analysis platform, and is configured to adaptively trigger corresponding control strategies according to the type and level of the final abnormality judgment result, including starting parameter fine-tuning instructions for the primary abnormality, starting equipment isolation instructions for the high-level abnormality and generating a maintenance work order.

Images