A fault detection apparatus, method and storage medium

Abstract

The application discloses a kind of fault detection equipment, method and storage medium, it is related to fault detection technical field, the equipment includes: sampling head activation unit, for activating cyclone sampling head to generate high-speed airflow adsorption equipment inside nanometer micro particle;Surface charge analysis unit is used to obtain particle size information;Difference capacitance detection unit is used to obtain detection result;Lifetime prediction unit is used to determine remaining life prediction result;Fault early warning unit is used to trigger active early warning execution module to carry out fault early warning to equipment according to remaining life prediction result.Solve the technical problem that existing industrial equipment fault detection technology is difficult to accurately capture nanometer micro particle in very early stage, leading to early warning lag, high false alarm rate, reaches the technical effect that through adopting cyclone sampling head and tunnel current sensor real-time monitoring nanometer micro particle, improve the accuracy of early fault identification, reduce false alarm rate.

Description

Technical Field

[0001] This application relates to the field of fault detection technology, specifically to a fault detection device, method, and storage medium. Background Technology

[0002] With the development of industrial automation and intelligence, predictive maintenance of equipment has become an important means to improve production efficiency and reduce maintenance costs. Traditional fault detection technologies mainly focus on monitoring macroscopic anomalies, such as abnormal changes in parameters like temperature, vibration, and pressure. However, these methods often struggle to accurately identify equipment failures in their very early stages, especially in detecting nanoscale particles. Nanoscale particles, such as those generated by metal wear, corrosion, or ionization processes, typically begin to appear in the early stages of equipment failure. If these particles can be detected in real time and provided with early warnings before a failure occurs, the operational reliability of the equipment can be greatly improved and its service life extended. However, existing technologies have significant limitations in this regard. For example, traditional particle counters can only detect larger particles and cannot effectively capture nanoscale particles; semiconductor gas sensors can detect gas decomposition products but cannot detect solid nanoparticles; and vibration and temperature monitoring systems can only reflect the macroscopic state of the equipment and cannot provide early warnings for very early microscopic damage. Therefore, there is an urgent need for detection equipment and methods capable of capturing nanoscale particles in real time and accurately, so as to provide early warnings in the early stages of failure, thereby reducing the probability of equipment failure and ensuring the continuous and stable operation of the equipment. Summary of the Invention

[0003] This application provides a fault detection device, method, and storage medium, which solves the technical problem that existing industrial equipment fault detection technologies are unable to accurately capture nanoscale particles in the very early stages, resulting in delayed early warnings and high false alarm rates. It achieves the technical effect of improving the accuracy of early fault identification and reducing false alarm rates by using a cyclone sampling head and a tunnel current sensor to monitor nanoscale particles in real time.

[0004] This application provides a fault detection device, comprising: a sampling head activation unit, used to activate a cyclone sampling head based on a nano-microparticle collection module to generate a high-speed airflow to adsorb nano-microparticles inside the device; a surface charge analysis unit, used to analyze the surface charge distribution of nano-microparticles using a tunneling current sensor to obtain particle size information; a differential capacitance detection unit, used to synchronize the particle size information to a high-precision detection module for differential capacitance detection to obtain detection results; a lifetime prediction unit, used to perform fault detection and identification by combining the detection results with an intelligent data processing module, generate a device fault type, predict the remaining lifetime based on the device fault type, and determine the remaining lifetime prediction result; and a fault early warning unit, used to trigger an active early warning execution module to issue a fault warning to the device based on the remaining lifetime prediction result.

[0005] This application also provides a fault detection method, the method comprising: activating a cyclone sampling head based on a nano-scale microparticle acquisition module to generate a high-speed airflow to adsorb nano-scale microparticles inside the device; using a tunneling current sensor to detect and analyze the surface charge distribution of the nano-scale microparticles to obtain particle size information; synchronizing the particle size information to a high-precision detection module for differential capacitance detection to obtain detection results; using an intelligent data processing module to combine the detection results for fault detection and identification, generating a device fault type; performing a lifespan prediction based on the device fault type to determine the remaining lifespan prediction result; and triggering an active early warning execution module to issue a fault warning to the device based on the remaining lifespan prediction result.

[0006] This application also provides a computer-readable storage medium, including: a computer program stored thereon, which, when executed by a processor, implements a fault detection device.

[0007] This application proposes a fault detection device, method, and storage medium, comprising: a sampling head activation unit for activating a cyclone sampling head based on a nanoparticle acquisition module to generate a high-speed airflow to adsorb nanoparticles inside the device; a surface charge analysis unit for analyzing the surface charge distribution of nanoparticles using a tunneling current sensor to obtain particle size information; a differential capacitance detection unit for synchronizing the particle size information to a high-precision detection module for differential capacitance detection to obtain detection results; a lifetime prediction unit for using an intelligent data processing module to combine the detection results for fault detection and identification, generating a device fault type, and predicting the remaining lifetime based on the fault type; and a fault early warning unit for triggering an active early warning execution module to issue a fault warning to the device based on the remaining lifetime prediction result. This addresses the technical problem of existing industrial equipment fault detection technologies being unable to accurately capture nanoparticles in the very early stages, leading to delayed warnings and high false alarm rates. It achieves the technical effect of improving the accuracy of early fault identification and reducing false alarm rates by using a cyclone sampling head and a tunneling current sensor to monitor nanoparticles in real time. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the equipment according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0009] Figure 1 This is a schematic diagram of a fault detection equipment provided in an embodiment of this application.

[0010] Figure 2 This is a schematic flowchart of a fault detection method provided in an embodiment of this application.

[0011] Figure labeling: Sampling head activation unit 11, surface charge analysis unit 12, differential capacitance detection unit 13, lifetime prediction unit 14, fault warning unit 15. Detailed Implementation

[0012] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below.

[0013] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, apparatus, product, or server that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, products, or apparatuses. 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 belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0015] This application provides a fault detection device, such as... Figure 1 As shown, the equipment includes:

[0016] The sampling head activation unit 11 is used to activate the cyclone sampling head based on the nano-microparticle collection module to generate a high-speed airflow to adsorb nano-microparticles inside the device.

[0017] Preferably, in the sampling head activation unit 11, the nanoscale microparticle acquisition module generates a high-speed directional airflow by activating the cyclone sampling head, and forms a strong airflow (e.g., 10 m / s) in the conical airflow channel using the Bernoulli effect. This strong airflow can effectively adsorb nanoscale microparticles present inside the device, and then, through the composite membrane (pore size 20 nm), the metal / non-metal nanoparticles are retained by a dual mechanism of chemical bonding and physical sieving, providing samples for subsequent microparticle analysis.

[0018] Furthermore, the specific configuration of the sampling head activation unit 11 also includes: constructing a conical airflow channel, activating the cyclone sampling head to generate high-speed airflow parameters using the Bernoulli effect; adsorbing particles inside the device according to the high-speed airflow parameters and calculating the particle adsorption rate; and using a graphene-metal-organic framework composite membrane for nanofiltration according to the particle adsorption rate to generate nanoscale microparticles.

[0019] Preferably, a conical airflow channel is first constructed, which is the core structure of the cyclone sampling head. When the cyclone sampling head is activated, a high-speed airflow is generated in the conical airflow channel using the Bernoulli effect, generating a high-speed airflow parameter, which is the airflow velocity, such as 10 m / s. The Bernoulli effect shows that in a fluid, the higher the flow velocity, the lower the pressure. Therefore, the design of the conical airflow channel can effectively generate a high-speed airflow for adsorbing nanoparticles inside the device. After the cyclone sampling head is activated and the high-speed airflow parameter is generated, the microparticles inside the device are adsorbed and guided into the nanoparticle collection module under the action of the high-speed airflow. In this process, the hydroxyl groups modified on the surface of the filter membrane play a key role. By subtracting the concentration of microparticles sampled by the sampling head from the concentration of the original microparticles inside the device, and then dividing the calculated difference by the concentration of the original microparticles inside the device, the adsorption rate of the microparticles is obtained, that is, the proportion of nanoparticles that the airflow can capture under certain time and airflow conditions. While calculating the particle adsorption rate, a graphene-metal-organic framework (MOF) composite membrane is used for nanoparticle filtration. The pore size of this membrane is typically set at 20 nm, effectively trapping nanoparticles larger than 20 nm while allowing gas molecules to pass through. The surface of the MOF membrane is modified with hydroxyl groups, enabling an adsorption rate of over 95% for metal nanoparticles, significantly improving the filtration effect and ensuring efficient capture of nanoparticles generated within the equipment. This precise design enables efficient, real-time capture and filtration of nanoparticles within the equipment, providing crucial data support for subsequent fault analysis and early warning systems.

[0020] The surface charge analysis unit 12 is used to analyze the surface charge distribution of nanoscale microparticles by using a tunneling current sensor to obtain particle size information.

[0021] Preferably, in the surface charge analysis unit 12, the tunneling current sensor operates through the quantum tunneling effect. When nanoparticles pass through the sensing area of ​​the probe, the surface charge of the nanoparticles alters the tunneling path of the electrons, thereby causing a change in current. By measuring this current change, the distribution of the particle surface charge can be calculated and then correlated with the particle size. Typically, the change in tunneling current has an exponential relationship with the particle size. By utilizing this relationship, the particle size can be accurately determined, forming particle size data as the basis for subsequent fault identification and analysis.

[0022] Furthermore, the specific configuration of the surface charge analysis unit 12 also includes: detecting the quantum tunneling effect of the surface charge of nanoparticles by extracting probes based on the tunneling current sensor, capturing the surface charge distribution information of the particles; extracting the tunneling current signal from the surface charge distribution information of the particles through a lock-in amplifier; performing particle size correlation analysis based on the tunneling current signal, determining the correlation index, and determining the particle size information based on the correlation index.

[0023] Preferably, the tunneling current sensor first detects the charge distribution on the surface of nanoparticles through the quantum tunneling effect. When a nanoparticle passes under the extraction probe of the tunneling current sensor, the surface charge of the particle causes a change in the electron tunneling path between the probe and the particle due to the quantum tunneling effect, resulting in current fluctuations. The tunneling current sensor then uses these fluctuations to capture the surface charge distribution information of the particle. Subsequently, the captured particle surface charge distribution information is extracted by a lock-in amplifier. The lock-in amplifier enhances the signal-to-noise ratio through synchronous processing, thereby separating the weak tunneling current signal from the noise and extracting the tunneling current signal related to the surface charge of the nanoparticle. Afterwards, a correlation analysis of particle size is performed based on the extracted tunneling current signal. Since tunneling current and particle size are usually exponentially related, a correlation index (ΔI∝d⁻²) related to particle size can be calculated by performing regression analysis on the sample tunneling current signal and the sample particle size. This correlation index is a quantitative relationship between particle size and current change, reflecting the influence of particle size on the tunneling current. Finally, by combining the known tunneling current signal with the correlation index, the particle size is calculated to form particle size information, which further provides particle size data for fault detection and early warning, helping to discover potential equipment faults in a timely manner.

[0024] The differential capacitance detection unit 13 is used to synchronize the particle size information to the high-precision detection module for differential capacitance detection and obtain the detection result.

[0025] Preferably, in the differential capacitance detection unit 13, after obtaining the particle size information, the particle size information is used as an input parameter and synchronized to the high-precision detection module. The high-precision detection module, using its interdigitated dual-electrode structure, analyzes the capacitance characteristics of the particles. When nanoparticles enter the region between the electrodes, their dielectric constant interacts with the electric field between the electrodes, resulting in a change in capacitance. At this time, the high-precision detection module detects the difference in dielectric constant of the nanoparticles by comparing the capacitance difference on both sides of the electrodes. The dielectric constant is a physical property of a substance, and different types of nanoparticles often have different dielectric constants. By detecting this difference, the high-precision detection module can further obtain the characteristic information of the nanoparticles, forming a detection result, and providing key information for subsequent fault diagnosis and lifespan prediction.

[0026] Furthermore, the specific configuration of the differential capacitance detection unit 13 also includes: synchronizing the particle size information to the high-precision detection module and analyzing it using an interdigitated dual-electrode structure to construct a differential capacitance; detecting the difference in dielectric constant of nanoparticles based on the differential capacitance to obtain the difference in dielectric constant of nanoparticles; and eliminating environmental electric field interference by adjusting the electric field change between electrodes based on the difference in dielectric constant of nanoparticles to obtain the detection result.

[0027] Preferably, the obtained particle size information is first synchronously transmitted to a high-precision detection module. This module uses an interdigitated dual-electrode structure for differential capacitance analysis. The interdigitated dual-electrode structure has a spacing of 50 μm and an electrode thickness of 200 nm, which allows it to accurately capture the influence of nanoparticles on the electric field. When nanoparticles enter the region between the electrodes, their charge and physical properties change the electric field, causing a change in capacitance. By measuring the capacitance difference between the two electrodes, a differential capacitance is constructed. This differential capacitance reflects the difference in the dielectric constant of the nanoparticles. Therefore, by detecting the capacitance difference, the dielectric constant information of the particles can be obtained. Subsequently, the difference in the dielectric constant of the nanoparticles is detected using the measured differential capacitance. That is, the difference in the dielectric constant of the nanoparticles is obtained by multiplying the differential capacitance by the distance between the electrodes, and then dividing it by the vacuum dielectric constant and the electrode area. Afterwards, based on the measured difference in the dielectric constant of the nanoparticles, environmental electric field interference is eliminated to mitigate the change in the electric field between the electrodes. Since electric field interference may originate from the electromagnetic environment around the equipment, a dual-electrode structure is used to eliminate environmental electric field noise. Furthermore, temperature drift is ensured to be no more than ±0.1℃ through micro-hotplate temperature control, effectively suppressing capacitance deviation caused by temperature changes. These compensation and optimization measures ultimately yield accurate detection results, ensuring the accuracy and reliability of measurement data and providing high-precision support for equipment fault detection, lifespan prediction, and early warning.

[0028] The life prediction unit 14 is used to perform fault detection and identification by combining the detection results with the intelligent data processing module, generate equipment fault types, perform life prediction based on the equipment fault types, and determine the remaining life prediction result.

[0029] Preferably, in the life prediction unit 14, the intelligent data processing module receives the detection results from the differential capacitance detection unit 13 and uses this data for fault detection and identification. Specifically, the intelligent data processing module extracts current and capacitance signals from the detection results and identifies possible fault types of the equipment through an embedded multimodal input model. These fault types can include different equipment damage states such as metal wear, corrosion, insulation aging, and discharge. Once a fault type is identified, the equipment fault type, along with the current and capacitance signals, are used as input data and transmitted to the life analysis model. This life analysis model is built using a Long Short-Term Memory (LSTM) network for existing life prediction through transfer learning. The transfer learning steps include forward propagation, loss calculation, backpropagation, and parameter optimization. By combining existing fault types and historical data, the life analysis model calculates the remaining lifespan of the equipment and outputs the remaining life prediction result. This remaining life prediction result not only provides fault warnings for equipment managers but also helps optimize equipment maintenance plans, ensuring necessary repairs or replacements are performed before equipment failures actually occur, reducing equipment downtime and maintenance costs, thereby improving the overall operating efficiency and reliability of the equipment.

[0030] Furthermore, the specific configuration of the life prediction unit 14 also includes transmitting the detection results to the intelligent data processing module to extract the original current information and the original capacitance signal; performing wavelet denoising on the original current information and the original capacitance signal to obtain the current information and capacitance signal; performing feature extraction based on the current information and the capacitance signal to obtain current feature data and capacitance feature data; constructing a multimodal input model, synchronizing the current feature data and the capacitance feature data to the multimodal input model for fault type identification, and generating the equipment fault type.

[0031] Preferably, when the lifetime prediction unit 14 performs lifetime prediction, the obtained detection results (including current information, capacitance signals, etc.) are first transmitted to the intelligent data processing module. In this module, the original current signal and capacitance signal are extracted from the detection results. Subsequently, wavelet denoising processing is performed on the original current information and capacitance signal. Wavelet denoising is a signal processing technique designed to separate useful signals from noise. Through wavelet transform, noise from environmental interference, sensor errors, and other factors can be effectively removed while retaining the main characteristics of the signal. Therefore, the denoised current information and capacitance signal will be more accurate and cleaner, providing reliable basic data for subsequent analysis. After denoising, feature extraction is performed on the current and capacitance signals. Feature extraction is a crucial step in data processing, aiming to extract important features reflecting the equipment's fault state from the raw signals. By analyzing the current and capacitance signals across multiple dimensions, including the time and frequency domains, characteristic data of the current and capacitance can be extracted, such as the specific values ​​of current and capacitance, the peak and average values ​​of the current, the change in the dielectric constant of the capacitance, and the spectral characteristics of the capacitance. This characteristic data can better reflect the equipment's operating state and potential fault modes. Subsequently, based on sample data processed in the same way as these current and capacitance characteristic data, a multimodal input model is constructed. This multimodal input model combines various feature data, enabling comprehensive analysis of complex equipment states and more complete identification of potential fault types. Finally, the intelligent data processing module synchronously inputs the current and capacitance characteristic data into the multimodal input model for fault type identification, obtaining the current fault type of the equipment. This identification result will be used for subsequent equipment maintenance and fault early warning, ensuring timely repair or maintenance of the equipment and preventing sudden failures.

[0032] Furthermore, the specific configuration of the lifetime prediction unit 14 also includes: performing multiphysics field feature analysis on nanoscale microparticles to obtain physical features; performing time-series analysis on nanoscale microparticles to obtain time-series data; using an LSTM-GRU neural network to fuse the time-series data with the physical features to construct a multimodal input initial model; constructing a rule library for the association between particle features and fault types, mapping the rule library to the multimodal input initial model for updating, and obtaining the multimodal input model.

[0033] Preferably, multiphysics field (MPF) analysis is first performed on nanoscale microparticles to obtain their physical characteristics. MPF analysis involves multiple fields of physics, such as electricity, magnetism, and mechanics. Detailed analysis is conducted using parameters such as the surface charge, electromagnetic properties, and stress response of the nanoparticles to extract their physical characteristics in these physical fields. These characteristics include the particle's dielectric constant, permeability, and elastic modulus, providing crucial data for subsequent fault identification. Next, time-series analysis is performed on the nanoscale microparticles. This involves analyzing the changes in continuous particle characteristic data (such as current and capacitance) over time to identify the trends in particle characteristics over time, obtaining time-series data to reveal the dynamic characteristics of the equipment during operation and determine the existence of potential fault risks. Finally, an LSTM-GRU neural network is used to fuse the time-series data with the physical characteristics. LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Unit) are deep learning models commonly used for processing time-series data. LSTM effectively captures dependencies in long-term sequences, while GRU accelerates training through a simplified gating mechanism. By combining time-series data with physical features, the LSTM-GRU neural network can learn the patterns of particle feature changes over time, constructing a multimodal input initial model. This model integrates various information such as current, capacitance, and particle physical properties, providing comprehensive data support for subsequent fault type identification. Finally, a rule base for associating particle features with fault types is constructed and mapped to the multimodal input initial model for updates. This rule base extracts the relationship between equipment faults and particle features from historical data, establishing association rules between particle features and fault types. Mapping these rules to the multimodal input model further optimizes its predictive capabilities. Through continuous training and updates, accurate fault type identification capabilities are achieved, enabling more precise fault prediction and maintenance decisions based on particle features. Ultimately, the updated multimodal input model can more accurately identify equipment fault types, providing strong support for fault early warning and maintenance.

[0034] The fault warning unit 15 is used to trigger the active warning execution module to provide fault warnings to the equipment based on the remaining life prediction results.

[0035] Preferably, in the fault early warning unit 15, the health status of the equipment is assessed based on the remaining life prediction results. When the remaining life prediction results indicate that the remaining service time of the equipment is about to reach the warning threshold, the active early warning execution module is automatically triggered. The active early warning execution module generates early warning notifications based on the remaining life prediction results, combined with the equipment operating conditions and fault type. These notifications are promptly transmitted to equipment managers, reminding them to take necessary preventive measures, such as conducting equipment inspections, maintenance, or replacing critical components, to avoid sudden equipment failures. In this way, early warnings can be provided before equipment failures occur, thereby effectively reducing equipment downtime, lowering maintenance costs, and improving equipment reliability and safety.

[0036] In the above text, refer to Figure 1 A fault detection device according to an embodiment of the present invention has been described in detail. Next, reference will be made to... Figure 2 A fault detection method according to an embodiment of the present invention is described. A fault detection method, such as... Figure 2 As shown, the method includes: activating a cyclone sampling head based on a nanoscale microparticle acquisition module to generate a high-speed airflow to adsorb nanoscale microparticles inside the device; using a tunneling current sensor to detect and analyze the surface charge distribution of the nanoscale microparticles to obtain particle size information; synchronizing the particle size information to a high-precision detection module for differential capacitance detection to obtain detection results; using an intelligent data processing module to combine the detection results for fault detection and identification, generating device fault types, predicting lifespan based on the device fault types, and determining the remaining lifespan prediction result; and triggering an active early warning execution module to issue a fault warning to the device based on the remaining lifespan prediction result.

[0037] In one possible implementation, the fault detection method further includes: constructing a conical airflow channel, activating the cyclone sampling head to generate high-speed airflow parameters using the Bernoulli effect; adsorbing particles inside the device based on the high-speed airflow parameters and calculating the particle adsorption rate; and performing nanofiltration using a graphene-metal-organic framework composite membrane according to the particle adsorption rate to generate nanoscale microparticles.

[0038] In one possible implementation, the fault detection method further includes: detecting the quantum tunneling effect of the surface charge of nanoparticles using a probe extracted from a tunneling current sensor to capture the surface charge distribution information of the particles; extracting a tunneling current signal from the surface charge distribution information of the particles using a lock-in amplifier; performing particle size correlation analysis based on the tunneling current signal to determine a correlation index; and determining particle size information based on the correlation index.

[0039] In one possible implementation, the fault detection method further includes:

[0040] The particle size information is synchronized to a high-precision detection module and analyzed using an interdigitated dual-electrode structure to construct a differential capacitor. The difference in dielectric constant of nanoparticles is detected based on the differential capacitor to obtain the difference in dielectric constant of nanoparticles. Based on the difference in dielectric constant of nanoparticles, environmental electric field interference is eliminated to eliminate the change in electric field between electrodes, and the detection result is obtained.

[0041] In one possible implementation, the fault detection method further includes:

[0042] The detection results are transmitted to the intelligent data processing module to extract the original current information and original capacitance signal; wavelet denoising is performed on the original current information and original capacitance signal to obtain current information and capacitance signal; feature extraction is performed based on the current information and capacitance signal to obtain current feature data and capacitance feature data; a multimodal input model is constructed, and the current feature data and capacitance feature data are synchronized to the multimodal input model for fault type identification to generate the equipment fault type.

[0043] In one possible implementation, the fault detection method further includes:

[0044] Multiphysics field feature analysis is performed on nanoscale microparticles to obtain physical features; temporal analysis is performed on nanoscale microparticles to obtain temporal data; the temporal data and physical features are fused using an LSTM-GRU neural network to construct a multimodal input initial model; a rule library for the association between particle features and fault types is constructed, and the rule library is mapped to the multimodal input initial model for updating to obtain the multimodal input model.

[0045] The fault detection equipment provided in this embodiment of the invention can execute a fault detection method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0046] Based on the foregoing embodiments, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the fault detection equipment to operate in accordance with the foregoing embodiments.

[0047] Although this application makes various references to certain modules in the equipment according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0048] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A fault detection device, characterized in that, The equipment includes: The sampling head activation unit is used to activate the cyclone sampling head based on the nano-microparticle collection module to generate a high-speed airflow to adsorb nano-microparticles inside the device. The surface charge analysis unit is used to detect and analyze the surface charge distribution of nanoscale microparticles using a tunneling current sensor to obtain particle size information. The differential capacitance detection unit is used to synchronize the particle size information to the high-precision detection module for differential capacitance detection to obtain the detection result; The life prediction unit is used to perform fault detection and identification by combining the detection results with the intelligent data processing module, generate equipment fault types, perform life prediction based on the equipment fault types, and determine the remaining life prediction result. The fault warning unit is used to trigger the active warning execution module to provide fault warnings to the equipment based on the remaining life prediction results. The differential capacitance detection unit includes: The particle size information is synchronized to the high-precision detection module and analyzed using an interdigitated dual-electrode structure to construct a differential capacitor. The difference in dielectric constant of nanoparticles is detected based on the differential capacitance to obtain the difference in dielectric constant of nanoparticles. Based on the difference in dielectric constant of the nanoparticles, the interference from the environmental electric field is eliminated by targeting the change in electric field between electrodes to obtain the detection result; The lifetime prediction unit includes: The detection results are transmitted to the intelligent data processing module, which extracts the raw current information and raw capacitance signal. Wavelet denoising is performed on the original current information and the original capacitance signal to obtain the current information and capacitance signal; Based on the current information and the capacitance signal, feature extraction is performed to obtain current feature data and capacitance feature data. A multimodal input model is constructed, and the current characteristic data and the capacitance characteristic data are synchronized to the multimodal input model for fault type identification to generate the equipment fault type; The lifetime prediction unit includes: Multiphysics field characterization of nanoscale microparticles was performed to obtain their physical characteristics; Time-series analysis of nanoscale particles was performed to obtain time-series data; An LSTM-GRU neural network is used to fuse the time-series data with the physical features to construct a multimodal input initial model; A rule base for the association between particle features and fault types is constructed, and the rule base is mapped to the multimodal input initial model for updating to obtain the multimodal input model.

2. The fault detection equipment as described in claim 1, characterized in that, The sampling head activation unit includes: A conical airflow channel is constructed, and the cyclone sampling head is activated to generate high-speed airflow parameters using the Bernoulli effect; The adsorption rate of particles is calculated by adsorbing particles inside the device based on the high-speed airflow parameters. Nanofiltration is performed using a graphene-metal-organic framework composite membrane based on the stated particle adsorption rate to generate nanoscale microparticles.

3. The fault detection equipment as described in claim 1, characterized in that, The surface charge analysis unit includes: A probe is extracted using a tunneling current sensor to detect the quantum tunneling effect of surface charge on nanoparticles and capture information on the surface charge distribution of the particles. The tunneling current signal is extracted from the surface charge distribution information of the particles using a lock-in amplifier. Particle size correlation analysis is performed based on the tunneling current signal to determine the correlation index, and particle size information is determined based on the correlation index.

4. A fault detection method, characterized in that, The method is applied to a fault detection equipment according to any one of claims 1-3, and the method includes: The nanoscale microparticle collection module activates the cyclone sampling head to generate a high-speed airflow that adsorbs nanoscale microparticles inside the device. The surface charge distribution of nanoscale microparticles was detected and analyzed using a tunneling current sensor to obtain particle size information. The particle size information is synchronized to the high-precision detection module for differential capacitance detection to obtain the detection results; The intelligent data processing module combines the detection results to perform fault detection and identification, generate equipment fault types, and perform life prediction based on the equipment fault types to determine the remaining life prediction results. Based on the remaining lifespan prediction results, the active early warning execution module is triggered to provide fault warnings for the equipment. The fault detection method further includes: The particle size information is synchronized to the high-precision detection module and analyzed using an interdigitated dual-electrode structure to construct a differential capacitor. The difference in dielectric constant of nanoparticles is detected based on the differential capacitance to obtain the difference in dielectric constant of nanoparticles. Based on the difference in dielectric constant of the nanoparticles, the interference from the environmental electric field is eliminated by targeting the change in electric field between electrodes to obtain the detection result; The fault detection method further includes: The detection results are transmitted to the intelligent data processing module, which extracts the raw current information and raw capacitance signal. Wavelet denoising is performed on the original current information and the original capacitance signal to obtain the current information and capacitance signal; Based on the current information and the capacitance signal, feature extraction is performed to obtain current feature data and capacitance feature data. A multimodal input model is constructed, and the current characteristic data and the capacitance characteristic data are synchronized to the multimodal input model for fault type identification to generate the equipment fault type; The fault detection method further includes: Multiphysics field characterization of nanoscale microparticles was performed to obtain their physical characteristics; Time-series analysis of nanoscale particles was performed to obtain time-series data; An LSTM-GRU neural network is used to fuse the time-series data with the physical features to construct a multimodal input initial model; A rule base for the association between particle features and fault types is constructed, and the rule base is mapped to the multimodal input initial model for updating to obtain the multimodal input model.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it is applied to a fault detection device as described in any one of claims 1-3.

Images