Remote intelligent diagnosis method and system for electrical equipment

Abstract

The invention provides a remote intelligent diagnosis method and system for electrical equipment, and is suitable for the technical field of equipment monitoring, and the method comprises the steps: carrying out the preprocessing and feature extraction of a collected operation state data set, and obtaining an operation state feature set, carrying out weighted fusion and equipment health quantitative evaluation on the running state feature set to obtain a real-time health index and residual service life probability distribution, and when the real-time health index is lower than an early warning threshold value, carrying out early warning; and performing fault identification and positioning processing on each functional structure in the equipment according to the operation state data set through a fault diagnosis rule to obtain fault feature data and fault positioning information, and performing data fusion processing on the residual service life probability distribution and the fault feature data according to the fault positioning information to generate a maintenance work order of the corresponding equipment. According to the invention, early warning of equipment abnormity is realized and maintenance and spare part scheduling are optimized, so that non-planned shutdown is significantly reduced and operation and maintenance efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of equipment monitoring technology, and in particular to a remote intelligent diagnostic method and system for electrical equipment. Background Technology

[0002] With the accelerated advancement of industrial digitalization and intelligent manufacturing, enterprises and institutions have an increasingly strong demand for improving equipment availability, reducing unplanned downtime, and lowering maintenance costs. At the same time, the wide distribution of equipment, the uneven skills of on-site maintenance personnel, and equipment aging make remote and efficient diagnostics a necessity. Furthermore, the energy and manufacturing industries have strict requirements for production line continuity and safety compliance, prompting enterprises to shift from passive maintenance to predictive maintenance, thereby driving the expansion of market demand for real-time multi-source sensing, intelligent diagnostics, and maintenance decision support capabilities.

[0003] Currently, mainstream remote diagnostic methods for electrical equipment mainly involve SCADA / IIoT-based data acquisition and threshold alarms, periodic manual inspections and infrared temperature or vibration detection, as well as several condition monitoring systems based on expert rules or single models. It is also common to upload collected data to the cloud for manual or offline analysis, or to use single-type analysis tools (such as spectrum analysis or trend thresholding) for fault identification and alarms. These methods rely on individual sensor parameters, threshold settings, and human experience, and the accompanying visualizations are mostly dashboard-style displays. On-site maintenance still heavily depends on manual judgment and spare parts preparation.

[0004] The existing methods mentioned above often rely on fixed thresholds or single-factor alarms, making it difficult to achieve dynamic weight fusion and accurate remaining life prediction across modalities and multiple operating conditions. Furthermore, fault location is mostly limited to the equipment level or parameter anomalies, and cannot directly generate quantifiable maintenance work orders and material support plans. As a result, it is difficult to guide rapid on-site repairs and coordinate with the supply chain in advance and accurately, and cannot fully reduce unplanned downtime and overall operation and maintenance costs. Summary of the Invention

[0005] In view of this, this application provides a remote intelligent diagnostic method and system for electrical equipment, which can solve the problems of insufficient predictability and operability in the prior art.

[0006] In a first aspect, embodiments of this application provide a method for remote intelligent diagnosis of electrical equipment, the method comprising: The operating status dataset of the target electrical equipment is collected in real time and synchronously, and the operating status dataset is preprocessed and features are extracted to obtain the operating status feature set of each equipment. Based on the detected environmental condition data, the operating status feature set is weighted and fused, and the equipment health is quantitatively assessed to obtain the real-time health index of each device and the corresponding probability distribution of its remaining service life. When the real-time health index is lower than the preset warning threshold, the fault identification and location processing of each functional structure in the device are performed according to the operating status dataset based on the preset fault diagnosis rules, so as to obtain the fault feature data and fault location information of the corresponding device. Based on the preset AR data structure and the fault location information, the remaining service life probability distribution and the fault feature data are fused together to generate a maintenance work order for the corresponding equipment.

[0007] In some embodiments, the preprocessing and feature extraction of the operating status dataset to obtain the operating status feature set of each device includes: The running state dataset is subjected to signal filtering processing to obtain a denoised state time series dataset; The denoised state time series dataset is subjected to analog-to-digital conversion to obtain a digital signal sequence of the running state; The digital signal sequence of the operating state is normalized to obtain a standardized dataset; Based on preset key feature types and preset device identifiers, the standardized dataset is subjected to time-domain and frequency-domain feature extraction and data classification processing to generate operating status feature sets for each device.

[0008] In some embodiments, the step of performing weighted fusion and quantitative assessment of the operating status feature set based on detected environmental condition data to obtain the real-time health index and corresponding remaining service life probability distribution of each device includes: The current ambient temperature, current operating load rate, and current operating mode of each device are obtained to construct environmental operating condition data for each device. Based on the environmental condition data, the set of feature weight coefficients for each device is obtained by using preset working condition and weight mapping rules. The operating status feature set is weighted and fused according to the feature weight coefficient set to generate the real-time health index of each device. Based on the real-time health index, historical decay prediction data sequences for each device are extracted from a preset device lifecycle database. Based on the historical decay prediction data sequence and the preset health index failure threshold, the real-time health index is subjected to nonlinear regression decay fitting quantitative evaluation to generate the probability distribution of the remaining service life of each device.

[0009] In some embodiments, the step of identifying and locating faults in various functional structures of the device based on the operating status dataset using preset fault diagnosis rules, and obtaining corresponding fault feature data and fault location information of the device, includes: By using preset structural prior fault rules, the fault screening of each functional structure in the device is performed on the operating status dataset, and a preliminary fault type list of the corresponding device is generated. Based on the preset fault type and structural feature dataset and the preliminary fault type list, deep signal analysis and comparison are performed on the operating status dataset to identify fault feature data and fault functional structure. Based on the device identifier corresponding to the device, extract the corresponding structural connection diagram from the preset device engineering structure database; Based on the structural connection diagram and the fault function structure, the fault feature data is processed for fault source tracing and component location mapping to obtain fault location information.

[0010] In some embodiments, the step of performing data fusion processing on the remaining service life probability distribution and the fault feature data according to a preset AR data structure and the fault location information to generate a maintenance work order for the corresponding equipment includes: Based on the equipment identifier corresponding to the equipment, extract the corresponding 3D model of the equipment from the equipment engineering structure database; Based on the fault location information, virtual space anchoring and high-brightness annotation processing of the faulty components are performed in the three-dimensional model of the equipment to generate virtual space coordinate data of the fault point. Key prediction information is extracted from the remaining service life probability distribution, and fault description text is extracted from the fault feature data, so as to integrate the key prediction information and the fault description text into maintenance guidance text data; Based on the preset AR data structure, the virtual spatial coordinate data of the fault point and the maintenance guidance text data are fused and encapsulated to generate a maintenance work order for the corresponding equipment.

[0011] In some embodiments, the method further includes: Extract faulty component information from the maintenance work order, and query the inventory status in the preset spare parts management database based on the faulty component information to obtain the spare parts inventory status information of each material required for the maintenance of the corresponding equipment. When the spare parts inventory status information belongs to the preset inventory sufficient information type, a corresponding material collection instruction is generated according to the preset collection instruction format and the spare parts inventory status information, and the collection instruction is data-bound with the corresponding maintenance work order. When the spare parts inventory status information does not belong to the inventory sufficiency information type, a purchase requisition order for the corresponding material is generated according to the preset purchase order format and the faulty component information.

[0012] In some embodiments, the method further includes: The maintenance work orders are classified and processed according to the preset production line data to obtain the production line maintenance work order set for each production line. Based on the production line data and the probability distribution of remaining service life, a production line operation stability correlation assessment is performed on each work order in the production line maintenance work order set, and a maintenance priority identifier corresponding to each work order is generated. Based on the preset downtime and maintenance time window and the maintenance priority identifier, maintenance planning decisions are made for the set of production line maintenance work orders to generate the optimal maintenance strategy for each production line.

[0013] Secondly, this application also provides a remote intelligent diagnostic device for electrical equipment, the device comprising: The feature extraction module is used to collect the operating status dataset of the target electrical equipment in real time and perform preprocessing and feature extraction on the operating status dataset to obtain the operating status feature set of each equipment. The status assessment module is used to perform weighted fusion and quantitative assessment of the operating status feature set based on the detected environmental condition data, and to obtain the real-time health index of each device and the corresponding probability distribution of its remaining service life. The fault diagnosis module is used to identify and locate faults in various functional structures of the equipment according to the operating status dataset based on preset fault diagnosis rules when the real-time health index is lower than the preset warning threshold, and to obtain the corresponding fault feature data and fault location information of the equipment. The work order generation module is used to perform data fusion processing on the remaining service life probability distribution and the fault feature data according to the preset AR data structure and the fault location information, and generate a maintenance work order for the corresponding equipment.

[0014] Thirdly, this application also provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the remote intelligent diagnostic method for electrical equipment as described above.

[0015] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the remote intelligent diagnostic method for electrical equipment as described above.

[0016] In summary, this application includes at least the following beneficial technical effects: 1. Through real-time multimodal data acquisition and dynamic weighted health assessment, the health index can be quantified and the remaining service life can be predicted, thereby issuing maintenance windows and warnings in advance, significantly reducing sudden equipment downtime and production losses caused by downtime.

[0017] 2. By adopting a diagnostic approach that combines structural prior rules with depth signal comparison, and then issuing the positioning results in the form of digital twins and AR work orders, on-site visual positioning and step-by-step guidance become possible, reducing reliance on highly skilled engineers and accelerating the speed of fault diagnosis and repair.

[0018] 3. Link diagnostic results with RUL predictions, spare parts inventory, and production line maintenance windows to automate material requisition / procurement and maintenance priority planning, thereby optimizing maintenance timing and material support, reducing inventory holdings and emergency procurement costs, and improving overall operation and maintenance efficiency. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic flowchart of an embodiment of a remote intelligent diagnostic method for electrical equipment provided in this application; Figure 2 This is a schematic diagram of an embodiment of an electronic device provided in this application; Figure 3 This is a structural block diagram of a remote intelligent diagnostic device for electrical equipment provided in this application. Detailed Implementation

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0022] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0023] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0024] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), unless otherwise expressly and specifically defined.

[0025] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0026] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0027] Firstly, please refer to Figure 1 , Figure 1 This is a schematic flowchart illustrating an embodiment of a remote intelligent diagnostic method for electrical equipment provided in this application. The remote intelligent diagnostic method for electrical equipment provided in this application includes the following steps.

[0028] Step S1: Real-time synchronous acquisition of the operating status dataset of the target electrical equipment, and preprocessing and feature extraction of the operating status dataset to obtain the operating status feature set of each equipment.

[0029] To achieve intelligent diagnostics of industrial electrical equipment, it is first necessary to construct a digital operational mirror of it. This process begins with a comprehensive perception and standardized reconstruction of the physical equipment's operating status. It should be understood that in the automated production line environment monitored in this application, electrical equipment distributed across various process stages of the production line, including drive motors, frequency converters, circuit breakers, contactors, and relays in the distribution cabinet, needs to be deployed in a heterogeneous distributed sensor network. This network consists of sensor nodes based on various physical principles. For example, Hall effect current sensors and resistive voltage divider sensors are embedded in the three-phase power supply circuit of the motor to capture instantaneous waveforms of current and voltage; insulated-gate bipolar transistor drive voltage sensors are installed on the DC bus side of the frequency converter to monitor the pulsation of the DC link voltage; PT100 platinum resistance temperature sensors are attached to the surface of the motor bearing housing and the frequency converter heat sink to track temperature rise changes; piezoelectric vibration acceleration sensors are installed axially and radially on the motor housing to sense the mechanical vibration spectrum; and ultra-high frequency partial discharge sensors are arranged inside the high-voltage switchgear to detect signs of insulation degradation. All sensor nodes, via Time-Sensitive Networking (TSN)-based Industrial Ethernet or 5G-U field wireless network, synchronize with a global clock and collect raw physical signals in parallel at a sampling frequency of no less than 1024 points per power frequency cycle, according to a unified timestamp sequence. This results in a multi-channel, time-stamped set of raw analog signals (i.e., the operating status dataset). This dataset is the foundation for all subsequent intelligent diagnostics and analysis; its completeness directly determines the timeliness of fault warnings and the reliability of diagnostic conclusions.

[0030] Because the raw signals in the operational status dataset inevitably contain various electromagnetic interferences and measurement noises, signal filtering is necessary to extract the true physical state. For example, transient electromagnetic interference caused by the start-up and shutdown of high-power equipment in the production line environment, high-frequency harmonic noise caused by the switching frequency of the frequency converter, and thermal noise from the sensor itself can severely contaminate the useful signal. This application employs a bandpass filter bank based on Chebyshev approximation theory to process the raw signal. The passband range is customized according to the physical characteristics of the target signal. For example, the passband for current and voltage signals is set to 1Hz to 2kHz to retain the power frequency and main harmonic components, the passband for vibration signals is set to 10Hz to 10kHz to cover the characteristic frequencies of bearings and gears, and the passband for partial discharge signals is set to 300MHz to 1.5GHz to capture nanosecond-level pulses. This filter bank achieves attenuation of stopband frequency components higher than 60dB while ensuring that the signal amplitude fluctuation within the passband is less than 0.1dB, thereby outputting a denoised state time-series dataset with a significantly improved signal-to-noise ratio. The above operations can eliminate the misleading influence of environmental interference on condition assessment. For example, if the motor current is not filtered, the inverter switching noise may be misjudged as the high-frequency characteristics of a rotor bar breakage fault, leading to a false alarm.

[0031] The filtered and denoised time-series dataset is still a continuous analog signal and must be converted to the digital domain before it can be processed by computing devices. In this embodiment, the analog-to-digital conversion (ADC) is performed by a high-precision ADC chip deployed on the sensor node or edge gateway. This chip must have at least 16-bit resolution and a synchronous sampling rate of 100 kS / s to ensure the ability to capture subtle signal features, such as distinguishing current transients as low as 0.001% of full scale during motor startup. The conversion process follows the Nyquist-Shannon sampling theorem to prevent frequency aliasing. After conversion, each analog voltage or current value is quantized into a discrete integer code value, forming a sequence of digital signals representing the operating status. This sequence serves as a bridge between the physical and digital worlds, allowing the detected data to be used in subsequent digitization algorithms.

[0032] Because the digital signal sequences of operating status originate from different types of sensors, their dimensions and numerical ranges vary greatly. For example, current signals may be in the ampere range (0-100A), temperature signals in degrees Celsius (0-150°C), and vibration signals in gravitational acceleration units (0-20g). Directly comparing or fusing these heterogeneous data would lead to features with large numerical magnitudes completely dominating the analysis results. Therefore, it is necessary to normalize the digital signal sequences of operating status. A maximum-minimum normalization algorithm is used to linearly map the signal of each channel to the interval [0, 1]. This process ensures that all features are on the same order of magnitude, laying a fair foundation for subsequent multi-parameter fusion evaluation. For example, without normalization, a 100A current change would completely mask a 0.1g vibration change, which could be a key indicator of early bearing wear.

[0033] Finally, key indicators characterizing equipment health status are extracted from the standardized dataset. Based on preset key feature types, the mean, root mean square (RMS), peak value, peak-to-peak value, skewness, and kurtosis of the signal are calculated in the time domain. For example, the RMS value of current reflects load magnitude, and the kurtosis of vibration is sensitive to impact faults. In the frequency domain, the signal is converted to the frequency domain using a Fast Fourier Transform (FFT), and the energy proportion, center-of-gravity frequency, and harmonic distortion of specific frequency bands are calculated. For example, analyzing the energy of specific sidebands in the current spectrum can diagnose rotor eccentricity faults in motors. All these extracted feature vectors are bound to specific physical devices and classified using preset device identifiers. The device identifier serves as a unique key, ensuring that features extracted from dozens of motors on a production line can be accurately categorized as "roller motor A" or "pump station motor B," thereby generating an independent and complete set of operating status features for each device. This feature set is an efficient condensation of high-dimensional raw data. It extracts gigabyte-level waveform data into kilobyte-level feature vectors, greatly reducing the burden of subsequent data transmission and calculation, while retaining the core information of the device's health status, providing direct input for the calculation of real-time health index and fault diagnosis.

[0034] Step S2: Based on the detected environmental condition data, perform weighted fusion and equipment health quantification assessment on the operating status feature set to obtain the real-time health index of each device and the corresponding probability distribution of remaining service life.

[0035] After constructing the operational status feature set, the equipment health assessment process enters the core quantitative analysis stage. This stage aims to integrate multi-dimensional, heterogeneous feature data into intuitive health indicators and predict future equipment failure risks. The entire process begins with the acquisition of environmental operating condition data. Temperature and humidity sensors, load current detection units, and programmable logic controller (PLC) communication interfaces distributed at the equipment site capture the current ambient temperature, current operating load rate, and current operating mode in real time. The current ambient temperature refers to the instantaneous temperature value of the cabinet or space where the equipment is located; the current operating load rate is calculated as the percentage of motor drive current to rated current; and the current operating mode is obtained from the control system, such as the constant torque, constant power, or sleep state of the frequency converter. These three types of parameters together constitute the environmental operating condition data corresponding to each piece of equipment. The significance of this data lies in providing a dynamic context for subsequent assessments, because the same set of feature values ​​has drastically different indicative meanings for health status under different operating conditions. For example, a motor operating at an ambient temperature of 45 degrees Celsius and a load rate of 90% might have a winding temperature of 110 degrees Celsius within the normal range; however, if it reaches the same temperature at an ambient temperature of 25 degrees Celsius and a load rate of 50%, it strongly suggests a malfunction in the cooling system or abnormal internal losses. Health assessments that ignore the impact of operating conditions will lead to serious misdiagnosis.

[0036] After acquiring environmental operating condition data, the system analyzes it using preset operating condition and weight mapping rules to obtain a set of feature weight coefficients for each device. This mapping rule is a decision logic library built upon domain expert knowledge and historical fault data statistics; essentially, it is a multi-dimensional lookup table or a set of fuzzy inference rules. The rules define the importance of various operating state characteristics in the overall health assessment under different operating condition combinations. For example, when the ambient temperature is consistently higher than a set threshold and the load rate is high, the rules significantly increase the weight coefficients of temperature-related characteristics (such as winding temperature rise rate and radiator temperature) and current-related characteristics (such as current harmonic distortion rate), while appropriately reducing the weight of vibration characteristics, because thermal stress and electrical stress are the dominant factors in equipment degradation under this condition. Conversely, when the equipment is in a frequent start-stop operation mode, the rules assign higher weights to impact characteristics (such as vibration peak and starting current peak). The significance of this step is to achieve adaptive adjustment of the assessment strategy, ensuring that the health assessment model can focus on the feature dimensions most likely to cause failure, thereby improving the sensitivity and accuracy of state perception.

[0037] Next, the system performs a weighted fusion calculation on the set of operational status features based on the feature weight coefficient set. This process aggregates dozens or even hundreds of feature indicators into a single, scalarized real-time health index. The calculation is not a simple linear weighted average, but rather employs a comprehensive function that considers the nonlinear interactions between features. Specifically, each feature value is first multiplied by its dynamic weight, and then all weighted feature values ​​are input into a fusion engine based on the Martin system or a multivariate health benchmark model. This engine calculates the Mahalanobis distance or similarity between the current weighted feature vector and the benchmark vector established by the device in a "fully healthy" state, ultimately mapping and generating a real-time health index between 0 and 100. A value of 100 represents an ideal health state, and 0 represents complete failure. This aggregation process greatly simplifies the complexity of status monitoring, providing maintenance personnel with a clear and concise "dashboard" of device health. For example, a pump motor may have both a slight current imbalance and increased bearing vibration. Through weighted fusion, these two independent features are combined into a clear health index of 85, which intuitively indicates that the equipment has entered the "attention" zone and needs to be monitored more closely, without the need for maintenance personnel to interpret the complex changes of multiple feature parameters separately.

[0038] After generating a real-time health index, the system extracts historical degradation prediction data sequences for each device based on this index from a pre-defined device lifecycle database. This database maintains an independent time-series file for each registered device, continuously recording its daily health index, key operating conditions, and maintenance history. When conducting a new assessment, the system uses the device's unique identifier as an index to retrieve all health index records for that device over a past period (e.g., one year) from the database, forming a historical degradation prediction data sequence. This sequence provides direct evidence of the slow deterioration of device performance over time and is the sole data basis for lifespan prediction. Its core value lies in achieving completely personalized prediction, because even two motors of the same model will have completely different health degradation trajectories due to differences in installation location, load characteristics, and maintenance history. Predictions relying on a general failure rate model for the entire device family are far less accurate than personalized predictions based on the device's own operating history.

[0039] Finally, the system performs a nonlinear regression-based quantitative assessment of the real-time health index based on historical decay prediction data sequences and a preset health index failure threshold. The health index failure threshold is a threshold value defined according to the critical failure state of the equipment; for example, it can be set to 20. When the health index falls below this value, the equipment is considered to have lost its function. The assessment process first uses a nonlinear regression algorithm, such as curve fitting based on the Levenberg-Marquardt algorithm, to model the historical decay data sequence and find the mathematical function that best describes its downward trend, such as an exponential decay function or a power function. This fitted function captures the rate of degradation of the equipment's health status. Subsequently, the system uses the current real-time health index as a starting point and extrapolates along the trajectory of this fitted function to calculate the time required for the health index curve to reach the preset failure threshold, which is the predicted remaining service life. Due to the fluctuations and uncertainties in historical data, this extrapolation result is not a definite value, but rather a probability distribution of the remaining service life expressed as a probability density function, for example, "the probability of a remaining service life of 90 days is 80%, and the probability of 120 days is 15%." This probability distribution quantifies the uncertainty of the forecast, providing a scientific basis for making maintenance decisions with manageable risks. For example, if the probability distribution of the remaining service life of a critical robot servo motor on a high-cycle automated production line shows that the probability of failure within the next three weeks is less than 1%, then its maintenance can be safely scheduled within the planned downtime window of the following month. Conversely, if the predicted distribution of another conveyor chain motor shows that its probability of failure within a week is as high as 30%, then an early warning must be triggered immediately and an emergency inspection must be arranged to avoid huge production losses caused by unplanned downtime.

[0040] The above operations, through a coherent process from context awareness and dynamic weighting to data-driven prediction, transform complex multi-source feature information into decision indicators with clear physical meaning and forward-looking perspectives. This process not only answers the question of "is the equipment healthy at this moment?", but more importantly, accurately predicts "how long it can continue to operate reliably," laying the core technological foundation for the transformation of the operation and maintenance model from passive response to proactive prediction. The entire evaluation system closely aligns with the actual logic of equipment operation and management in industrial fields, ensuring the direct usability and high value of the output results in operation and maintenance decision-making.

[0041] Step S3: When the real-time health index is lower than the preset warning threshold, the fault identification and location processing of each functional structure in the device is performed according to the operating status dataset based on the preset fault diagnosis rules, so as to obtain the fault feature data and fault location information of the corresponding device.

[0042] In an equipment health monitoring system, when the real-time health index falls below a preset warning threshold, it indicates that the equipment has progressed from a sub-healthy state to a stage of manifest fault development. At this point, the system automatically triggers the refined fault diagnosis process in step S3. In this application, this process aims to transform macroscopic health degradation warnings into specific, actionable maintenance guidelines. Its core technology lies in integrating equipment mechanism knowledge with data-driven analysis to achieve accurate fault identification and physical location.

[0043] The cornerstone of the fault diagnosis process is the pre-defined fault diagnosis rules, which consist of two parts: structural prior fault rules and a fault type and structural feature dataset. The structural prior fault rules are an expert knowledge base built upon the physical structure and working principles of the equipment. They encode domain experts' deep understanding of specific functional structural failure modes using production rules in the form of "IF-THEN". For example, for the DC bus capacitor of a frequency converter, a typical rule might be: "IF The ripple coefficient of the DC bus voltage exceeds the threshold α AND The fundamental effective value of the output voltage is normal THEN There is an aging fault in the DC bus capacitor (confidence level 85%)". The fault type and structural feature dataset is a vast data warehouse, storing the characteristic fingerprints of various known fault modes in high-frequency sampled data. For example, a bearing inner race fault will show peaks at specific passing frequencies and their harmonics in the vibration acceleration signal envelope spectrum. Together, these two constitute the "brain" of the diagnostic system, closely linking abstract logical judgments with specific numerical characteristics.

[0044] Diagnosis begins with screening the operational status dataset for faults in various functional structures of the equipment using pre-defined structural prior fault rules. The system loads a corresponding subset of rules based on the equipment type (e.g., asynchronous motor, frequency converter) and efficiently traverses it using an optimized, decision tree-like path. This traversal order references the MPT algorithm, comprehensively considering the fault probability and detection cost of each functional block, prioritizing the inspection of components with high failure rates and ease of diagnosis. The system matches the real-time collected operational status dataset, including three-phase current, voltage waveforms, and vibration signals, with the conditions in the rules. For example, it calculates the negative-sequence component of the current to determine if there is a power imbalance and analyzes the peak value of the vibration signal to detect mechanical impact. All triggered rule conclusions are collected to generate a preliminary list of fault types for the corresponding equipment. The significance of this step lies in performing a rapid, low-computational-cost preliminary screening, converging an infinite number of fault possibilities into a finite, high-probability set of suspected faults. This is similar to an experienced engineer conducting a preliminary examination on-site to quickly pinpoint several possible directions of the fault, thus providing a target for subsequent precision instrument testing and avoiding the enormous computational overhead of blindly performing a comprehensive spectrum analysis.

[0045] After obtaining a preliminary list of fault types, the diagnosis enters the precise authentication phase. Based on a pre-defined dataset of fault types and structural features, along with the preliminary list, the system performs in-depth signal analysis and comparison on the operating status dataset. This phase abandons rule-based logic and instead activates signal processing algorithms to directly extract microscopic evidence related to suspected faults from the raw data. The system invokes corresponding specialized analysis algorithms based on suspected faults in the list. For example, if the preliminary list includes "motor bearing outer race fault," the system will initiate an envelope spectrum analysis program to demodulate the vibration acceleration signal using a Hilbert transform, then calculate its spectrum to find characteristic spectral lines at the bearing outer race's passing frequency. If the list includes "inverter IGBT open circuit fault," a current waveform analysis program will be initiated to meticulously compare the symmetry of the three-phase current waveforms and the rate of change of current at switching moments. By performing pattern matching and similarity calculations between the analysis results and typical fault features stored in the fault type and structural feature dataset, the system can identify precise fault feature data (such as "the amplitude at the bearing outer ring passing frequency reaches 0.5g, exceeding the normal benchmark by 10 times") and confirm the specific faulty functional structure (such as "motor drive end bearing"). The significance of this in-depth analysis step lies in providing solid data evidence for preliminary screening conclusions, elevating experience-based speculation to measurement-based diagnosis, and greatly improving the objectivity and accuracy of the results.

[0046] To translate the confirmed faulty functional structure into operable location commands for on-site maintenance personnel, the system needs to map the abstract "functional structure" to specific physical components. This mapping process is achieved by querying a pre-set equipment engineering structure database. The system initiates a query to this database based on the equipment identifier corresponding to the equipment (such as asset code "MCC-001A"). The equipment engineering structure database is a central repository containing detailed engineering information for all controlled equipment, storing electrical schematics, mechanical assembly drawings, bills of materials, and most importantly—structural connection diagrams. These diagrams clearly illustrate the electrical connections and physical adjacencies between the various functional modules within the equipment in a graphical manner. For example, for a frequency converter, the diagram would show the interconnections of components such as the rectifier module, DC bus, inverter IGBT module, control board, and cooling fan.

[0047] Finally, based on the retrieved structural connection diagram and the identified fault functional structures, the system performs fault source tracing and component location mapping on the fault characteristic data. This process analyzes the propagation path of the fault signal and its dependency on the functional structure. For example, deep signal analysis confirms the fault functional structure of "inverter bridge arm output abnormality." Combined with the structural connection diagram, the system can trace this abnormality back to a specific IGBT power module and its drive circuit. Ultimately, the system outputs precise fault location information, which is no longer a functional description like "inverter fault," but a specific component-level location such as "inverter A, inverter unit, third phase upper bridge arm IGBT module (part number: IGBT-XX-003)." This step is the final step in the entire diagnostic process. Its significance lies in seamlessly connecting the diagnostic conclusions in the digital space to the maintenance operations in the physical space, directly informing maintenance personnel which equipment to go to, which cabinet door to open, and which specific part to replace, thereby minimizing fault finding time, improving maintenance efficiency, and ensuring the rapid recovery and continuous stable operation of the automated production line.

[0048] Step S4: Based on the preset AR data structure and the fault location information, perform data fusion processing on the remaining service life probability distribution and the fault feature data to generate a maintenance work order for the corresponding equipment.

[0049] After accurately diagnosing and locating the fault, the system enters the maintenance information visualization and structured output stage (i.e., step S4). The core task of this step is to transform the abstract diagnostic conclusions generated in the previous steps and existing in the digital space into augmented reality maintenance work orders that on-site maintenance personnel can intuitively perceive and efficiently execute. This process constructs a maintenance guidance interface that connects the digital world and physical reality by integrating the equipment's 3D model, spatial positioning technology, and structured text data.

[0050] The maintenance work order generation process begins with accessing the digital twin model of the physical entity of the equipment. Based on the equipment identifier corresponding to the equipment, the system initiates a query request to the equipment engineering structure database. This database not only stores the two-dimensional structural connection diagrams used in the previous diagnostic phase but also includes high-precision three-dimensional models of the equipment obtained through laser scanning or computer-aided design. This model is a precise digital reproduction of the geometry, internal component layout, and spatial relationships of the physical equipment, such as a frequency converter or a motor control center. For example, for a specific model of frequency converter (equipment identifier: VFD-2024-05), its three-dimensional model will accurately show the three-dimensional shape, size, and relative position of all maintainable components, such as the casing, cooling fan, terminals, capacitor banks, and IGBT power modules. The significance of obtaining this model is to provide an accurate digital base for subsequent augmented reality overlay, ensuring that the virtual information is perfectly aligned with the real equipment in space.

[0051] After acquiring the 3D model of the equipment, the system performs virtual spatial anchoring and high-brightness annotation processing on the faulty component within the model based on the fault location information. The fault location information comes from step S3 and is typically in the format "Equipment Identifier: VFD-2024-05; Faulty Component: Inverter Unit Third Phase Upper Arm IGBT Module". Based on this information, the system accurately locates the corresponding virtual component among thousands of parts in the 3D model by matching the part number or function name. Subsequently, spatial anchoring is performed on the virtual component, calculating and recording its precise position, orientation, and bounding box in the 3D model coordinate system. Simultaneously, high-brightness annotation processing is applied to the component, typically using dynamic, semi-transparent colored specular rendering, such as a striking red pulse flashing effect, to make it stand out in the 3D scene. This processing generates virtual spatial coordinate data for the fault point, which includes not only the 3D coordinates of the fault point but also rendering style information. The significance of this step lies in transforming the textual description of the fault location into a spatial guide that cannot be ignored by the eye, solving the difficulty for on-site personnel to find tiny faulty components in complex equipment. For example, it can directly guide maintenance personnel to accurately locate a specific IGBT module inside the inverter from the outside, without having to repeatedly consult drawings or disassemble unrelated parts.

[0052] Simultaneously, the system extracts key textual information from the remaining service life probability distribution and fault characteristic data in parallel. From the remaining service life probability distribution, the system extracts key predictive information, including the expected value of the remaining service life, key confidence intervals (e.g., the range of remaining days at 90% confidence level), and risk level. This information is formatted into concise statements such as "Expected remaining service life: 15 days (90% confidence interval: 10-22 days), Risk level: High". From the fault characteristic data, the system extracts fault description text, including confirmed fault modes, severity, and possible cause analysis, such as "Fault mode: IGBT module overheating aging; Characteristic evidence: Junction temperature estimate consistently exceeds 150°C, on-state voltage drop increases by 15%". Subsequently, the system integrates the key predictive information with the fault description text and supplements it with standard safety precautions and preliminary handling suggestions to form a complete maintenance guidance text. The significance of this text integration lies in providing context and decision-making basis for maintenance actions, enabling maintenance personnel not only to know "where it is broken", but also to understand "how serious it is" and "why it is broken", thereby making the correct maintenance strategy, such as whether to carry out preventive replacement or emergency shutdown.

[0053] Finally, based on the preset AR data structure, the virtual spatial coordinate data of the fault point and the maintenance guidance text data are fused and encapsulated. The AR data structure is a predefined data packet format specification optimized for mobile AR devices. Under this specification, the system packages the virtual spatial coordinate data of the fault point, the maintenance guidance text data, and the corresponding 3D model reference link of the device into a structured data object. This data object also contains device feature point descriptors for AR recognition to ensure stable tracking of virtual content in the camera image. After encapsulation, a maintenance work order for the corresponding device is generated. This work order is not a traditional paper document or a simple electronic list, but an interactive set of instructions that can be loaded and executed by a dedicated application on AR glasses or a tablet. When maintenance personnel arrive at the site and scan the actual device with the AR device, the application will accurately overlay flashing highlighted marks onto the actual faulty component based on the spatial coordinate data in the work order, and display the maintenance guidance text as a floating card on the side of the screen. The significance of this final step lies in achieving unambiguous transmission and immersive interaction of maintenance information. It presents the complex intelligent diagnostic results from the background in the most intuitive and error-free way to the first-person perspective of the maintenance operation, greatly reducing the technical threshold, shortening the training time, and ensuring the standardization and accuracy of maintenance operations, ultimately guaranteeing the efficiency and reliability of maintenance operations on automated production lines.

[0054] Beyond the core process of the remote intelligent diagnostic method for electrical equipment, the system integrates two key extended functional modules. These modules, as optional implementations, together form a complete closed loop from fault diagnosis to maintenance execution and even production scheduling. The first extended module focuses on the automated preparation of maintenance resources. When a maintenance work order is generated, the system automatically extracts the faulty component information contained in the work order. This information records the specific material number, name, and specifications that need to be replaced or repaired in a structured data format. Subsequently, based on this faulty component information, the system initiates a real-time query to a pre-set spare parts management database. This database is a dynamic inventory list linked to an enterprise resource planning system or warehouse management system, recording the latest inventory quantity, storage location information, and supplier details of all spare parts. The purpose of the query is to obtain the spare parts inventory status information of each material required for the maintenance of the corresponding equipment. This status information not only includes binary results of "available" or "absent," but may also include the specific available quantity at different warehouse locations, the quantity awaiting inspection, and the estimated arrival time of orders in transit.

[0055] After obtaining spare parts inventory status information, the system compares it with preset inventory sufficiency information types. Inventory sufficiency information types are typically defined by a rule set based on a safety stock model. For example, for frequently used spare parts, the rule might require "current available inventory greater than or equal to 2 pieces" to be considered sufficient; for infrequently used or expensive spare parts, the rule might be set to "current available inventory greater than or equal to 1 piece." When the system determines that the current spare parts inventory status information belongs to a preset inventory sufficiency information type, it automatically generates a corresponding material retrieval instruction based on a preset retrieval instruction format. This instruction format is standardized and includes information such as the material requisition number, material code, requested quantity, suggested storage location, and maintenance work order number. The generated retrieval instruction is automatically sent to the warehouse department through the enterprise's internal workflow system. Simultaneously, the system binds this retrieval instruction to the corresponding maintenance work order, allowing maintenance personnel to directly see the spare parts retrieval status as "pending retrieval" or "outbound" on the work order interface, thus achieving synchronization of logistics and information flow. The significance of this is that it eliminates the spare parts confirmation step before maintenance, avoiding the embarrassment and delays caused by maintenance personnel arriving on-site only to find that no materials are available. For example, when a fault is diagnosed in a servo drive encoder and a work order is generated, the system immediately finds that there are 3 encoders of that model in stock in the central warehouse. It then automatically generates a material requisition form and notifies the warehouse manager to prepare the materials. Maintenance personnel can pick up the spare parts on their way to the equipment site, which greatly reduces the overall maintenance time.

[0056] Conversely, when the spare parts inventory status information does not fall under the "sufficient inventory" category—meaning inventory is below the safety threshold or there is no inventory—the system automatically generates a purchase requisition order for the corresponding materials based on a preset purchase order format and information about the faulty component. The purchase order format is also standardized, including supplier information, material specifications, purchase quantity, required date, and associated cost center. The system automatically fills in these fields and can send the order to the purchasing department for approval or place an order directly with a certified supplier. The significance of this automated process lies in directly coupling fault diagnosis with supply chain response, triggering a replenishment mechanism the instant a fault is detected, minimizing unplanned downtime caused by waiting for spare parts. For example, if the main motor bearing of a high-cycle stamping production line fails and there are no spare parts in stock, a manual procurement process might take hours or even days to initiate. However, this system generates an emergency purchase order for the bearing simultaneously with the generation of the repair work order, saving valuable time for production resumption.

[0057] The second extension module focuses on optimizing the scheduling of maintenance tasks from a holistic production system perspective. This system categorizes all generated maintenance work orders based on pre-defined production line data. This production line data, sourced from the Manufacturing Execution System (MES), defines the affiliation of each electrical device with a specific production line; for example, "Robot Welding Gun A" belongs to the "Body-in-White Welding Line," and "Packaging Machine B" belongs to the "Finished Product Packaging Line." Using this data, the system groups scattered maintenance work orders according to their respective production lines, obtaining a set of production line maintenance work orders for each line. This elevates the maintenance management perspective from individual devices to the availability level of the entire production line.

[0058] Next, the system performs a production line operation stability correlation assessment on each work order in the production line maintenance work order set based on production line data and the probability distribution of remaining service life. This assessment is a multi-factor decision-making process that comprehensively considers two core dimensions: first, the criticality of the equipment in its production line; for example, on an automated assembly line, equipment at bottleneck stations is usually more critical than equipment at non-bottleneck stations; second, the urgency of the equipment failure, which is quantified by the probability distribution of remaining service life. For example, equipment with an expected remaining service life of only 2 hours is more urgent than equipment with a remaining service life of 200 hours. The system calculates a comprehensive risk score for each work order using a weighted scoring model and generates a corresponding maintenance priority label for each work order based on this score, such as "urgent," "high," "medium," and "low." The significance of this assessment is to prioritize the allocation of limited maintenance resources (personnel, time windows) to failures that pose the greatest threat to production continuity, thereby maximizing maintenance efficiency. For example, for two production lines that both experience pump bearing warnings, one being a unique production line and the other having parallel backups, the system will label the pump maintenance priority of the unique production line as "urgent," while labeling the other as "medium."

[0059] Finally, the system makes maintenance planning decisions for the set of production line maintenance work orders based on preset downtime and maintenance time windows and maintenance priority indicators. The downtime and maintenance time window is a non-production period provided by the production planning system that allows for planned maintenance. The system matches and optimizes the scheduling of maintenance work orders with the estimated working hours, required tools, and personnel qualifications, and the available time windows. Its decision logic is: within a given time window, prioritize work orders with "urgent" and "high" priorities, and attempt to bundle multiple work orders from the same production line or the same physical area to reduce the number of production line start-ups and shutdowns. Through this operational optimization, the system generates the optimal maintenance strategy for each production line. This strategy may be a detailed weekly maintenance plan or a specific task list for the upcoming weekend maintenance window. The significance of this step is to elevate reactive fault response to proactive, preventative maintenance management that is coordinated with production planning. It ensures that maintenance activities are executed efficiently while maximizing production flow, thereby improving the overall efficiency and reliability of the production system.

[0060] On the other hand, please see Figure 2 , Figure 2 This is a schematic diagram of an embodiment of an electronic device provided in this application.

[0061] like Figure 2 As shown, the electronic device 2 in this embodiment includes: at least one processor 21 ( Figure 2 Only one is shown in the diagram), memory 22, and computer program 23 stored in the memory 22 and executable on the at least one processor 21. When the processor 21 executes the computer program 23, it implements the steps in the embodiments of the remote intelligent diagnostic method for electrical equipment of this application.

[0062] Figure 2 The illustrated electronic device 2 may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will understand that... Figure 2 This is merely an example of electronic device 2 and does not constitute a limitation on electronic devices. It may have more or fewer components than shown in the figure, or combine certain components, or have different components. For example, it may also include input / output devices, network access devices, etc.

[0063] The processor 21 can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0064] In some embodiments, the memory 22 may be an internal storage unit of the electronic device, such as a hard drive or memory. In other embodiments, the memory 22 may be an external storage device of the electronic device, such as a plug-in hard drive, smart media card (SMC), secure digital card (SD), flash card, etc. Furthermore, the memory 22 may include both internal and external storage units of the electronic device. The memory 22 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 22 can also be used to temporarily store data that has been output or will be output.

[0065] Furthermore, in one embodiment, as Figure 3 As shown, this application also provides a remote intelligent diagnostic device for electrical equipment, comprising: The feature extraction module 1100 is used to collect the operating status dataset of the target electrical equipment in real time and perform preprocessing and feature extraction on the operating status dataset to obtain the operating status feature set of each equipment. The status assessment module 1200 is used to perform weighted fusion and equipment health quantification assessment on the operating status feature set based on the detected environmental condition data, and obtain the real-time health index of each device and the corresponding probability distribution of remaining service life. The fault diagnosis module 1300 is used to identify and locate faults in various functional structures of the equipment according to the operating status dataset based on preset fault diagnosis rules when the real-time health index is lower than the preset warning threshold, and to obtain the fault feature data and fault location information of the corresponding equipment. The work order generation module 1400 is used to perform data fusion processing on the remaining service life probability distribution and the fault feature data according to the preset AR data structure and the fault location information, and generate a maintenance work order for the corresponding equipment.

[0066] In some embodiments, the feature extraction module 1100 is specifically used for: The running state dataset is subjected to signal filtering processing to obtain a denoised state time series dataset; The denoised state time series dataset is subjected to analog-to-digital conversion to obtain a digital signal sequence of the running state; The digital signal sequence of the operating state is normalized to obtain a standardized dataset; Based on preset key feature types and preset device identifiers, the standardized dataset is subjected to time-domain and frequency-domain feature extraction and data classification processing to generate operating status feature sets for each device.

[0067] In some embodiments, the state assessment module 1200 is specifically used for: The current ambient temperature, current operating load rate, and current operating mode of each device are obtained to construct environmental operating condition data for each device. Based on the environmental condition data, the set of feature weight coefficients for each device is obtained by using preset working condition and weight mapping rules. The operating status feature set is weighted and fused according to the feature weight coefficient set to generate the real-time health index of each device. Based on the real-time health index, historical decay prediction data sequences for each device are extracted from a preset device lifecycle database. Based on the historical decay prediction data sequence and the preset health index failure threshold, the real-time health index is subjected to nonlinear regression decay fitting quantitative evaluation to generate the probability distribution of the remaining service life of each device.

[0068] In some embodiments, the fault diagnosis module 1300 is specifically used for: By using preset structural prior fault rules, the fault screening of each functional structure in the device is performed on the operating status dataset, and a preliminary fault type list of the corresponding device is generated. Based on the preset fault type and structural feature dataset and the preliminary fault type list, deep signal analysis and comparison are performed on the operating status dataset to identify fault feature data and fault functional structure. Based on the device identifier corresponding to the device, extract the corresponding structural connection diagram from the preset device engineering structure database; Based on the structural connection diagram and the fault function structure, the fault feature data is processed for fault source tracing and component location mapping to obtain fault location information.

[0069] In some embodiments, the work order generation module 1400 is specifically used for: Based on the equipment identifier corresponding to the equipment, extract the corresponding 3D model of the equipment from the equipment engineering structure database; Based on the fault location information, virtual space anchoring and high-brightness annotation processing of the faulty components are performed in the three-dimensional model of the equipment to generate virtual space coordinate data of the fault point. Key prediction information is extracted from the remaining service life probability distribution, and fault description text is extracted from the fault feature data, so as to integrate the key prediction information and the fault description text into maintenance guidance text data; Based on the preset AR data structure, the virtual spatial coordinate data of the fault point and the maintenance guidance text data are fused and encapsulated to generate a maintenance work order for the corresponding equipment.

[0070] In some embodiments, the remote intelligent diagnostic device for electrical equipment further includes an inventory query module 1500, which is specifically used for: Extract faulty component information from the maintenance work order, and query the inventory status in the preset spare parts management database based on the faulty component information to obtain the spare parts inventory status information of each material required for the maintenance of the corresponding equipment. When the spare parts inventory status information belongs to the preset inventory sufficient information type, a corresponding material collection instruction is generated according to the preset collection instruction format and the spare parts inventory status information, and the collection instruction is data-bound with the corresponding maintenance work order. When the spare parts inventory status information does not belong to the inventory sufficiency information type, a purchase requisition order for the corresponding material is generated according to the preset purchase order format and the faulty component information.

[0071] In some embodiments, the remote intelligent diagnostic device for electrical equipment further includes a maintenance planning module 1600, which is specifically used for: The maintenance work orders are classified and processed according to the preset production line data to obtain the production line maintenance work order set for each production line. Based on the production line data and the probability distribution of remaining service life, a production line operation stability correlation assessment is performed on each work order in the production line maintenance work order set, and a maintenance priority identifier corresponding to each work order is generated. Based on the preset downtime and maintenance time window and the maintenance priority identifier, maintenance planning decisions are made for the set of production line maintenance work orders to generate the optimal maintenance strategy for each production line.

[0072] It should be noted that the aforementioned remote intelligent diagnostic device for electrical equipment can be understood as a virtual device that can be installed in the electronic device in the aforementioned embodiments. The electronic device calls the remote intelligent diagnostic device for electrical equipment through its processor, thereby running the specific implementation scheme in the above embodiments of the remote intelligent diagnostic method for electrical equipment. The information interaction and execution process between the aforementioned devices / units are based on the same concept as the method embodiments of this application; their specific functions and technical effects can be found in the method embodiments section.

[0073] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0074] This application also provides a storage medium, which is a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps in the above-described method embodiments.

[0075] This application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0076] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0077] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0078] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A remote intelligent diagnostic method for electrical equipment, characterized in that, The method includes: The operating status dataset of the target electrical equipment is collected in real time and synchronously, and the operating status dataset is preprocessed and features are extracted to obtain the operating status feature set of each equipment. Based on the detected environmental condition data, the operating status feature set is weighted and fused, and the equipment health is quantitatively assessed to obtain the real-time health index of each device and the corresponding probability distribution of its remaining service life. When the real-time health index is lower than the preset warning threshold, the fault identification and location processing of each functional structure in the device are performed according to the operating status dataset based on the preset fault diagnosis rules, so as to obtain the fault feature data and fault location information of the corresponding device. Based on the preset AR data structure and the fault location information, the remaining service life probability distribution and the fault feature data are fused together to generate a maintenance work order for the corresponding equipment.

2. The method according to claim 1, characterized in that, The preprocessing and feature extraction of the operational status dataset to obtain the operational status feature set of each device includes: The running state dataset is subjected to signal filtering processing to obtain a denoised state time series dataset; The denoised state time series dataset is subjected to analog-to-digital conversion to obtain a digital signal sequence of the running state; The digital signal sequence of the operating state is normalized to obtain a standardized dataset; Based on preset key feature types and preset device identifiers, the standardized dataset is subjected to time-domain and frequency-domain feature extraction and data classification processing to generate operating status feature sets for each device.

3. The method according to claim 1, characterized in that, The step of performing weighted fusion and quantitative assessment of the operating status feature set based on the detected environmental condition data to obtain the real-time health index and corresponding remaining service life probability distribution of each device includes: The current ambient temperature, current operating load rate, and current operating mode of each device are obtained to construct environmental operating condition data for each device. Based on the environmental condition data, the set of feature weight coefficients for each device is obtained by using preset working condition and weight mapping rules. The operating status feature set is weighted and fused according to the feature weight coefficient set to generate the real-time health index of each device. Based on the real-time health index, historical decay prediction data sequences for each device are extracted from a preset device lifecycle database. Based on the historical decay prediction data sequence and the preset health index failure threshold, the real-time health index is subjected to nonlinear regression decay fitting quantitative evaluation to generate the probability distribution of the remaining service life of each device.

4. The method according to claim 2, wherein the fault diagnosis rules include structural prior fault rules and a fault type and structural feature dataset, characterized in that, The step of identifying and locating faults in various functional structures of the equipment based on the operating status dataset using preset fault diagnosis rules, and obtaining corresponding fault feature data and fault location information of the equipment, includes: By using preset structural prior fault rules, the fault screening of each functional structure in the device is performed on the operating status dataset, and a preliminary fault type list of the corresponding device is generated. Based on the preset fault type and structural feature dataset and the preliminary fault type list, deep signal analysis and comparison are performed on the operating status dataset to identify fault feature data and fault functional structure. Based on the device identifier corresponding to the device, extract the corresponding structural connection diagram from the preset device engineering structure database; Based on the structural connection diagram and the fault function structure, the fault feature data is processed for fault source tracing and component location mapping to obtain fault location information.

5. The method according to claim 4, characterized in that, The step of fusing the remaining service life probability distribution and the fault feature data according to the preset AR data structure and the fault location information to generate a maintenance work order for the corresponding equipment includes: Based on the equipment identifier corresponding to the equipment, extract the corresponding 3D model of the equipment from the equipment engineering structure database; Based on the fault location information, virtual space anchoring and high-brightness annotation processing of the faulty components are performed in the three-dimensional model of the equipment to generate virtual space coordinate data of the fault point. Key prediction information is extracted from the remaining service life probability distribution, and fault description text is extracted from the fault feature data, so as to integrate the key prediction information and the fault description text into maintenance guidance text data; Based on the preset AR data structure, the virtual spatial coordinate data of the fault point and the maintenance guidance text data are fused and encapsulated to generate a maintenance work order for the corresponding equipment.

6. The method according to claim 1, characterized in that, The method further includes: Extract faulty component information from the maintenance work order, and query the inventory status in the preset spare parts management database based on the faulty component information to obtain the spare parts inventory status information of each material required for the maintenance of the corresponding equipment. When the spare parts inventory status information belongs to the preset inventory sufficient information type, a corresponding material collection instruction is generated according to the preset collection instruction format and the spare parts inventory status information, and the collection instruction is data-bound with the corresponding maintenance work order. When the spare parts inventory status information does not belong to the inventory sufficiency information type, a purchase requisition order for the corresponding material is generated according to the preset purchase order format and the faulty component information.

7. The method according to claim 1, characterized in that, The method further includes: The maintenance work orders are classified and processed according to the preset production line data to obtain the production line maintenance work order set for each production line. Based on the production line data and the probability distribution of remaining service life, a production line operation stability correlation assessment is performed on each work order in the production line maintenance work order set, and a maintenance priority identifier corresponding to each work order is generated. Based on the preset downtime and maintenance time window and the maintenance priority identifier, maintenance planning decisions are made for the set of maintenance work orders for the production line, and the optimal maintenance strategy for each production line is generated.

8. A remote intelligent diagnostic device for electrical equipment, applied to the remote intelligent diagnostic method for electrical equipment as described in claim 1, characterized in that, The device includes: The feature extraction module is used to collect the operating status dataset of the target electrical equipment in real time and perform preprocessing and feature extraction on the operating status dataset to obtain the operating status feature set of each equipment. The status assessment module is used to perform weighted fusion and quantitative assessment of the operating status feature set based on the detected environmental condition data, and to obtain the real-time health index of each device and the corresponding probability distribution of its remaining service life. The fault diagnosis module is used to identify and locate faults in various functional structures of the equipment according to the operating status dataset based on preset fault diagnosis rules when the real-time health index is lower than the preset warning threshold, and to obtain the corresponding fault feature data and fault location information of the equipment. The work order generation module is used to perform data fusion processing on the remaining service life probability distribution and the fault feature data according to the preset AR data structure and the fault location information, and generate a maintenance work order for the corresponding equipment.

9. An electronic device, characterized in that, The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the remote intelligent diagnostic method for electrical equipment according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the remote intelligent diagnostic method for electrical equipment according to any one of claims 1 to 7.

Images