Multidirectional monitoring method, device and terminal based on electric power material sampling inspection

Abstract

The invention discloses a multi-azimuth monitoring method, device and terminal based on electric power material sampling inspection, and the method comprises the steps: obtaining material sampling inspection data in real time, the material sampling inspection data being data collected by a multi-azimuth monitoring terminal in the sampling inspection process of electric power materials, and the electric power materials being fixed on the multi-azimuth monitoring terminal in the sampling inspection process; and based on the sampling inspection task state of the electric power materials, analyzing the material sampling inspection data, and outputting the electric power material state and / or alarm information so as to complete monitoring of the electric power materials in the sampling inspection process. By monitoring the material sampling inspection data generated in the sampling inspection process of the electric power materials and combining the sampling inspection task state of the electric power materials, the material sampling inspection data is analyzed, the state of the electric power materials is judged, and the corresponding electric power material state and alarm information are output, so that the monitoring of each link in the sampling inspection process of the electric power materials is realized; the electric power material sampling inspection efficiency is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of material monitoring, and particularly relates to a multi-faceted monitoring method, device and terminal based on random inspection of electric power materials. Background Art

[0002] In the process of power grid construction, the quality of materials is the basis and premise for ensuring the safety and stable operation of the power system. Doing a good job in material quality supervision and special random inspection work is of great importance. Especially in the distribution link, there are many types and a large number of materials, resulting in many suppliers and a large difference in quality control levels. Therefore, the random inspection work of distribution materials is particularly important.

[0003] During the random inspection process of materials, different operation links such as sampling, sample encapsulation, sample delivery and transportation, and receiving samples are involved. Due to the lack of perfect technical control means in the process, ordinary encapsulation methods such as two-dimensional codes and RFID tags are mainly used for storing and protecting sensitive information of samples. During the sample submission process, personnel cannot be controlled, and a series of risk problems such as the seal being damaged or the seal being replaced exist, thus affecting the accuracy and fairness of the random inspection results of materials and leaving hidden dangers for the safe operation of power grid equipment.

[0004] At present, the monitoring of electric power materials focuses on security and warehousing, and does not analyze and detect the problems in the process of random inspection of electric power materials. Patent CN110211318A discloses a full-automatic integrated power material supervision system with functions of power material detection, analysis, ventilation control and automatic warehousing and outwarehousing, including a power supply module, a main control module, an information storage module, a temperature and humidity detection module, a power material transportation module, a temperature and humidity adjustment module, an anti-theft monitoring module, an in-out warehouse module, an alarm module, and a display module. The power supply module provides power for the warehouse supervision system, and the temperature and humidity detection module collects temperature and humidity data and transmits it back to the main control module; the anti-theft monitoring module collects signals and transmits them back to the main control module; the main control module receives, processes data and outputs the data to the information storage module, the display module and the alarm module. By adopting the temperature and humidity detection module, the anti-theft monitoring module, the main control module and the temperature and humidity adjustment module, it is ensured that the electric power materials are in a suitable storage state, the safety of the storage of electric power materials is guaranteed, manpower is saved, and the work efficiency is improved.

[0005] Therefore, how to monitor each link in the process of random inspection of electric power materials and improve the efficiency of random inspection of electric power materials is a problem that needs to be solved currently. Summary of the Invention

[0006] In view of the defects existing in the above-mentioned prior art, the present invention provides a multi-directional monitoring method, device and terminal based on random inspection of electric power materials. The method includes: obtaining random inspection data of materials in real time, where the random inspection data of materials is the data collected by a multi-directional monitoring terminal during the random inspection of electric power materials, and the electric power materials are fixed on the multi-directional monitoring terminal during the random inspection process; analyzing the random inspection data of materials based on the random inspection task status of the electric power materials, and outputting the status of the electric power materials and / or alarm information to complete the monitoring of the electric power materials during the random inspection process. By monitoring the random inspection data of materials generated during the random inspection of electric power materials and combining the random inspection task status of the electric power materials, analyzing the random inspection data of materials, judging the status of the electric power materials and outputting the corresponding status of the electric power materials and alarm information, the monitoring of each link during the random inspection of electric power materials is realized, and the efficiency of random inspection of electric power materials is improved.

[0007] In the first aspect, the present invention provides a multi-directional monitoring method based on random inspection of electric power materials, which specifically includes the following steps:

[0008] Obtaining random inspection data of materials in real time, where the random inspection data of materials is the data collected by a multi-directional monitoring terminal during the random inspection of electric power materials, and the electric power materials are fixed on the multi-directional monitoring terminal during the random inspection process;

[0009] Analyzing the random inspection data of materials based on the random inspection task status of the electric power materials, and outputting the status of the electric power materials and / or alarm information to complete the monitoring of the electric power materials during the random inspection process.

[0010] Further, the random inspection data of materials includes the status of the pressure switch and the light intensity data;

[0011] Analyzing the random inspection data of materials based on the random inspection task status of the electric power materials, and outputting the status of the electric power materials and / or alarm information specifically includes:

[0012] If the random inspection task status of the electric power materials is that the sealing sample is completed, then judge the status of the pressure switch and the light intensity data;

[0013] If the pressure switch status is raised or the light intensity data reaches the light intensity threshold, output the removal alarm information and the location information of the electric power materials;

[0014] If the pressure switch status is lowered and the light intensity data is less than the light intensity threshold, combine the preset sealing sample time interval and output the status of the electric power materials.

[0015] Further, analyzing the random inspection data of materials based on the random inspection task status of the electric power materials, and outputting the status of the electric power materials and / or alarm information specifically includes:

[0016] If the random inspection task status of the electric power materials is the sample delivery stage, then judge the status of the pressure switch and the light intensity data;

[0017] If the pressure switch status is raised or the light intensity data reaches the light intensity threshold, output the removal alarm information and the location information of the power materials.

[0018] If the pressure switch status is lowered and the light intensity data is less than the light intensity threshold, combine the preset sample delivery time interval and output the status of the power materials.

[0019] Furthermore, the material sampling inspection data includes acceleration data and positioning data.

[0020] Based on the sampling inspection task status of the power materials, analyze the material sampling inspection data and output the status of the power materials and / or alarm information, specifically including:

[0021] If the sampling inspection task status of the power materials is sample sealing completed, judge the acceleration data and the positioning data.

[0022] If the acceleration data exceeds the speed threshold or the positioning data exceeds the preset sample sealing area range, output the position deviation alarm information and the location information of the power materials.

[0023] If the acceleration data is less than the speed threshold and the positioning data is within the preset sample sealing area range, combine the preset sample sealing time interval and output the status of the power materials.

[0024] Furthermore, the material sampling inspection data includes vibration timing data.

[0025] Based on the sampling inspection task status of the power materials, analyze the material sampling inspection data and output the status of the power materials and / or alarm information, specifically including:

[0026] According to the sampling inspection task status of the power materials, perform noise reduction processing on the vibration timing data to obtain standard vibration timing data.

[0027] Combine the preset silent threshold, judge the standard vibration timing data, and give the judgment result.

[0028] According to the judgment result, perform window segmentation on the standard vibration timing data to obtain multiple standard vibration timing sub-data.

[0029] Fuse the frequency domain features and time domain features in multiple standard vibration timing sub-data to determine the first target vibration feature.

[0030] Based on the pre-constructed vibration analysis model, combine the first target vibration feature and the second target vibration feature, and give the status of the power materials, where the second target vibration feature is obtained through the pre-constructed timing analysis model.

[0031] Combine the analysis of the status of the power materials and output the alarm information.

[0032] Furthermore, the first target vibration feature includes at least one of a waveform factor, a zero-crossing rate, and a maximum spectral amplitude;

[0033] Fusing the frequency-domain features and time-domain features in multiple standard vibration time-series sub-data to determine the first target vibration feature specifically includes:

[0034] Analyzing the fluctuation conditions of multiple standard time-series sub-data in the time domain to give time-domain features;

[0035] Analyzing the proportion of different frequency bands of multiple standard time-series sub-data in the frequency domain to give frequency-domain features;

[0036] Combining the time-domain features and frequency-domain features to give at least one of a waveform factor, a zero-crossing rate, and a maximum spectral amplitude.

[0037] Furthermore, based on a pre-constructed vibration analysis model, combining the first target vibration feature and the second target vibration feature to give the power material state specifically includes:

[0038] Performing splicing and normalization processing on the first target vibration feature and the second target vibration feature to obtain a comprehensive vibration feature;

[0039] Based on the pre-constructed vibration analysis model, performing time-series analysis on the comprehensive vibration feature to obtain a time-series analysis result;

[0040] According to the time-series analysis result, using the classification layer of the vibration analysis model to classify the comprehensive vibration feature to give the power material state.

[0041] Furthermore, the pre-constructed vibration analysis model is determined by the following method:

[0042] Obtaining a vibration training data set, where each piece of data in the vibration training data set is labeled with a vibration event label;

[0043] Constructing a first memory network layer and a second memory network layer, and combining the vibration training data set to perform pre-training until convergence to obtain a vibration pre-analysis model;

[0044] Freezing the model parameters of the first memory network layer and the second memory network layer in the vibration pre-analysis model, and combining the loss function to train the fully connected layer and the classification layer until convergence to obtain a vibration analysis model.

[0045] In a second aspect, the present invention also provides a multi-faceted monitoring device based on power material sampling inspection, adopting the multi-faceted monitoring method based on power material sampling inspection as described in any one of the above, including:

[0046] A data acquisition unit for obtaining material sampling inspection data in real time, where the material sampling inspection data is data collected by a multi-directional monitoring terminal during the sampling inspection of electrical materials, and the electrical materials are fixed on the multi-directional monitoring terminal during the sampling inspection process;

[0047] A data analysis unit for analyzing the material sampling inspection data based on the sampling inspection task status of the electrical materials, and outputting the electrical material status and / or alarm information to complete the monitoring of the electrical materials during the sampling inspection process.

[0048] Thirdly, the present invention also provides a multi-directional monitoring terminal based on the sampling inspection of electrical materials, including a controller module, a pressure switch, a mobile communication module, a positioning module, a light sensor and a vibration sensor, and the pressure switch, the mobile communication module, the positioning module, the light sensor and the vibration sensor are all connected to the controller module;

[0049] The pressure switch is used to obtain the pressure switch status;

[0050] The mobile communication module is used to establish a communication channel for the controller module;

[0051] The positioning module is used to obtain the position information of the electrical materials and output positioning data;

[0052] The light sensor is used to obtain light intensity data;

[0053] The vibration sensor is used to obtain acceleration data and vibration time series data

[0054] The controller module is used to obtain the pressure switch status, positioning data, light intensity data, acceleration data and vibration time series data, and execute the multi-directional monitoring method based on the sampling inspection of electrical materials as described in any one of the above.

[0055] The multi-directional monitoring method, device and terminal based on the sampling inspection of electrical materials provided by the present invention have at least the following beneficial effects:

[0056] (1) By monitoring the material sampling inspection data generated during the sampling inspection of electrical materials, and combining with the sampling inspection task status of the electrical materials, analyzing the material sampling inspection data, judging the status of the electrical materials and outputting the corresponding electrical material status and alarm information, realizing the monitoring of each link during the sampling inspection of electrical materials, and improving the sampling inspection efficiency of electrical materials.

[0057] (2) By directly splicing the first vibration target feature and the second vibration target feature, the vibration analysis model has a stronger dependence on the comprehensive vibration feature, thereby improving the ability to distinguish similar events (such as sudden braking and human collision).

[0058] (3) By adopting a random forest model based on a long short-term memory network, LSTM and random forest are combined. At the same time, the first target vibration feature and the second target vibration feature are fused. The first target vibration feature is a statistical feature and is sensitive to certain transient features. The second target vibration feature is obtained through LSTM. LSTM can capture the dynamic changes in time, that is, the order or duration of events. The combination of the first target vibration feature and the second target vibration feature can describe the data more comprehensively, provide a more comprehensive discrimination basis for the random forest, and improve the classification accuracy. Description of the Drawings

[0059] Figure 1 It is a structural block diagram of a multi-directional monitoring terminal for power material sampling inspection provided by an embodiment of the present invention;

[0060] Figure 2 It is a structural schematic diagram of a multi-directional monitoring terminal for power material sampling inspection provided by an embodiment of the present invention;

[0061] Figure 3 It is a schematic diagram of the housing of a multi-directional monitoring terminal for power material sampling inspection provided by an embodiment of the present invention;

[0062] Figure 4 It is a structural schematic diagram of power materials and a multi-directional monitoring terminal including a flat installation surface provided by an embodiment of the present invention;

[0063] Figure 5 It is a structural schematic diagram of power materials and a multi-directional monitoring terminal including a cylindrical surface provided by an embodiment of the present invention;

[0064] Figure 6 It is a structural schematic diagram of power materials and a multi-directional monitoring terminal including a transfer box cover provided by an embodiment of the present invention;

[0065] Figure 7 It is a flowchart of a multi-directional monitoring method for power material sampling inspection provided by an embodiment of the present invention;

[0066] Figure 8 It is a flowchart of determining the state of power materials through vibration time series data provided by an embodiment of the present invention;

[0067] Figure 9 It is a flowchart of determining the first target vibration feature provided by an embodiment of the present invention;

[0068] Figure 10 It is a flowchart of determining the state of power materials provided by an embodiment of the present invention;

[0069] Figure 11 It is a flowchart of determining a vibration analysis model provided by an embodiment of the present invention;

[0070] Figure 12 It is the architecture diagram of the vibration analysis model provided by the embodiment of the present invention;

[0071] Figure 13 It is the structural block diagram of the multi-directional monitoring device based on the sampling inspection of electric power materials provided by the embodiment of the present invention.

[0072] Among them, 201 is the data acquisition unit; 202 is the data analysis unit. Specific embodiments

[0073] In order to better understand the above technical solutions, the above technical solutions will be described in detail below in conjunction with the accompanying drawings of the specification and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0074] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms of "a", "the" and "said" used in the embodiments of the present invention and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. "Plural" generally includes at least two.

[0075] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover a non-exclusive inclusion, so that a commodity or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such commodity or device. Without further limitation, the element defined by the statement "including one..." does not exclude the existence of another identical element in the commodity or device including the said element.

[0076] In order to realize the monitoring of each link in the process of sampling inspection of electric power materials and improve the efficiency of sampling inspection of electric power materials, refer to Figure 1, an embodiment of the present invention provides a multi - azimuth monitoring terminal based on power material sampling inspection, which includes a controller module, a pressure switch, a mobile communication module, a positioning module, a light - sensitive sensor, and a vibration sensor. The pressure switch, the mobile communication module, the positioning module, the light - sensitive sensor, and the vibration sensor are all connected to the controller module. The pressure switch is used to obtain the pressure switch state. The mobile communication module is used to establish a communication channel for the controller module. The positioning module is used to obtain the position information of the power material and output positioning data. The light - sensitive sensor is used to obtain light - sensitive intensity data. The vibration sensor is used to obtain acceleration data and vibration timing data. The controller module is used to obtain the pressure switch state, positioning data, light - sensitive intensity data, acceleration data, and vibration timing data, and execute the multi - azimuth monitoring method based on power material sampling inspection provided by the embodiment of the present invention.

[0077] In a specific example, referring to Figure 2 , the controller module uses a low - power STM32L151 single - chip microcomputer chip. The pressure switch in the multi - azimuth monitoring terminal adopts an XKB5858 pressure key switch, the light - sensitive sensor adopts a GL5506 photosensitive resistor, the vibration sensor adopts an ADXL345 three - axis acceleration sensor, the positioning module adopts a B305 - 5Q single Beidou, and the mobile communication module adopts an AIR780EG 4G mobile communication module. In other examples, in order to ensure the security of the communication channel established by the controller module through the mobile communication module, the multi - azimuth monitoring terminal adopts an NRSEC3000 security encryption communication chip to encrypt the data transmitted through the communication channel of the mobile communication module.

[0078] The GPIO interface in the controller module is used to detect the pressure switch state of the pressure switch in real time, the ADC interface is used to obtain the light - sensitive intensity detected by the light - sensitive sensor, that is, light - sensitive intensity data, in real time, the I 2 C interface is used to collect the three - axis impact acceleration and inclination values obtained by the vibration sensor in real time, the UART interface is used to connect to the positioning module to receive Beidou satellite positioning longitude, latitude, and moving speed information, the SPI interface is used to interact with the security encryption communication chip to realize the encryption and decryption of communication data between the multi - azimuth monitoring terminal and the monitoring management platform, and the UART interface is used to interact with the mobile communication module to realize the HTTP protocol communication between the multi - azimuth monitoring terminal and the monitoring management platform.

[0079] During the operation of the multi - azimuth monitoring terminal, the monitoring management platform will switch the sampling inspection task status according to the sampling inspection business process of the power material. The administrator of the monitoring management platform can manually control the termination of the sampling inspection task at any time. Before each time the multi - azimuth monitoring terminal uploads information to the monitoring management platform, it first requests the sampling inspection task status through the HTTP protocol and makes different logical processing on the detected data according to the current sampling inspection task status.

[0080] In the examples provided by the present invention, referring to Figure 3 , in order to facilitate the binding of the multi-directional monitoring terminal to power materials, the multi-directional monitoring terminal is designed as a housing made of ABS material with dimensions of 60mm * 50mm. The surface of the housing is provided with an NFC identification area, a pressure switch, a power indicator light, and a photosensitive area. In order to make the multi-directional monitoring terminal dustproof and waterproof, a dustproof and waterproof film is pasted on the surface of the housing of the multi-directional monitoring terminal. In other implementation examples, according to different requirements of power materials, the material or volume of the multi-directional monitoring terminal can be adjusted, which is not limited herein.

[0081] The multi-directional monitoring terminal combined with auxiliary tooling is applicable to the installation and construction of all distribution power materials, including high and low voltage switch cabinets, JP cabinets, cable branch boxes, disconnectors, ring main units, box-type substations, electric energy metering boxes, distribution transformers, voltage transformers, current transformers, lightning arresters, line insulators, overhead cables, power cables, cable protection pipes, concrete poles, fittings, etc.

[0082] Referring to Figure 4 , for box cabinet-type materials such as high and low voltage switch cabinets, JP cabinets, cable branch boxes, high disconnectors, ring main units, box-type substations, electric energy metering boxes, etc., and power materials with flat installation surfaces such as distribution transformers, voltage transformers, current transformers, etc., the multi-directional monitoring terminal is directly pasted on the sample with double-sided tape to realize the monitoring of power materials during the sampling inspection process.

[0083] Referring to Figure 5 , for power materials with cylindrical surfaces such as lightning arresters, line insulators, overhead cables, power cables, cable protection pipes, concrete poles, etc., they are tied to the sample with a tie strap with a base, and then the multi-directional monitoring terminal is pasted on the tie strap base with 3M double-sided tape.

[0084] Referring to Figure 6 , for various shaped and sized line fittings, after being packed in a customized transfer box, the transfer box lid is locked with a tie strap with a base, and then the multi-directional monitoring terminal is pasted on the tie strap base with 3M double-sided tape.

[0085] A multi-directional monitoring terminal provided by the present invention for power material sampling inspection is applied in the power material sampling inspection process. The monitoring terminal integrates multiple principle monitoring technologies such as pressure switch state detection, light intensity measurement, three-axis shock and vibration measurement, and high-precision Beidou positioning, as well as Internet of Things communication encryption technology. When the sample to be inspected is sealed, the monitoring terminal is bound to the sample to realize real-time monitoring and warning of the sealed state, position, attitude and other information of the sample after sealing, during sample delivery, and before sample reception, preventing illegal acts such as deliberate disassembly and replacement of the sample, solving the irregularities and quality risks existing in the power material sampling inspection process, and at the same time improving the power material sampling inspection efficiency.

[0086] AsFigure 7 As shown in Figure 7 , an embodiment of the present invention provides a multi - aspect monitoring method based on random inspection of power materials, and the specific steps are as follows:

[0087] S101: Obtain random inspection data of materials in real time.

[0088] Specifically, the random inspection data of materials is the data collected by a multi - aspect monitoring terminal during the random inspection of power materials, and the power materials are fixed on the multi - aspect monitoring terminal during the random inspection process.

[0089] S102: Analyze the random inspection data of materials based on the random inspection task status of power materials, and output the status of power materials and / or alarm information to complete the monitoring of power materials during the random inspection process.

[0090] In the first specific implementation manner, the random inspection data of materials includes the status of the pressure switch and the light intensity data. If the random inspection task status of the power material is that the sample sealing is completed, then judge the status of the pressure switch and the light intensity data. If the pressure switch status is raised or the light intensity data reaches the light - sensing threshold, it means that the multi - aspect monitoring terminal may be removed from the power material sample during the sample sealing process, and output the removal alarm information and the location information of the power material, which is used to prompt the staff that an abnormality has occurred and give the location where abnormal handling is required. If the pressure switch status is lowered and the light intensity data is less than the light - sensing threshold, it means that the multi - aspect monitoring terminal is working normally during the sample sealing process without any abnormality. Combining the preset sample - sealing time interval, output the status of the power material. The above - mentioned light - sensing threshold is set according to factors such as actual power - inspected materials and inspection environments. The above - mentioned sample - sealing time interval is set according to the actual scenario and is not limited in this regard. In this example, the sample - sealing time interval is 1 hour.

[0091] In the second specific implementation manner, if the random inspection task status of the power material is the sample - sending stage, then judge the status of the pressure switch and the light intensity data. If the pressure switch status is raised or the light intensity data reaches the light - sensing threshold, it means that the multi - aspect monitoring terminal may be removed from the power material sample during the sample - sending process, and output the removal alarm information and the location information of the power material, which is used to prompt the staff that an abnormality has occurred and give the location where abnormal handling is required. If the pressure switch status is lowered and the light intensity data is less than the light - sensing threshold, it means that the multi - aspect monitoring terminal is working normally during the sample - sending process without any abnormality. Combining the preset sample - sending time interval, output the status of the power material. The above - mentioned sample - sending time interval is set according to the actual scenario and is not limited in this regard. In this example, the sample - sealing time interval is 2 minutes.

[0092] In the third specific implementation manner, the material sampling inspection data includes acceleration data and positioning data. If the sampling inspection task status of the power material is sample sealing completed, then the acceleration data and the positioning data are judged. If the acceleration data exceeds the speed threshold or the positioning data exceeds the preset sample sealing area range, it means that the power material deviates from the sample sealing area during the sample sealing process, which will have a certain impact on the subsequent process, and the position deviation alarm information and the position information of the power material are output to prompt the staff that the position of the power material has shifted and the position information that needs to be processed is given. If the acceleration data is less than the speed threshold and the positioning data is within the preset sample sealing area range, it means that the power material does not deviate from the sample sealing area during the sample sealing process, and the power material status is output in combination with the preset sample sealing time interval.

[0093] It is understandable that the status of the spot-check task is successively idle, spot-check start, sample sealing completed, sample delivery preparation, sample delivery start, sample delivery completed, sample receiving start, and sample receiving completed, and finally returns to the idle spot-check task status. When the status of the spot-check task is "idle", the multi-directional monitoring terminal uploads the terminal information and positioning information every 24 hours. When the status of the spot-check task is "spot-check start", if the multi-directional terminal detects that the pressure switch is pressed and rises after 2 seconds, it is determined that the user actively triggers the multi-directional monitoring terminal to upload relevant information. When the status of the spot-check task is "sample sealing completed", the sample of the power material bound by the multi-directional monitoring terminal is stored in the sample sealing area. If the multi-directional monitoring terminal detects that the pressure switch rises or the light intensity data exceeds the light intensity threshold, it is determined that the multi-directional monitoring terminal is removed from the sample of the power material, and the removal alarm information and the position information of the power material are immediately triggered for upload. If the multi-directional monitoring terminal detects that the impact acceleration amplitude or inclination angle exceeds the threshold and the positioning position is far from the sample sealing area, it is determined that the sample has been moved, that is, the acceleration data exceeds the speed threshold or the positioning data exceeds the preset sample sealing area range, and the position deviation alarm information and the position information of the power material are immediately triggered for upload. When the status of the spot-check task is "sample sealing completed", the sample of the power material bound by the multi-directional monitoring terminal is stored in the sample sealing area. If the multi-directional monitoring terminal does not detect any abnormalities, it uploads the terminal information and positioning information every 1 hour. When the status of the spot-check task is "sample delivery preparation", "sample delivery start", "sample delivery completed", and "sample receiving start", the multi-directional monitoring terminal is bound to the sample of the power material. If the multi-directional monitoring terminal detects that the pressure switch rises or the light intensity data exceeds the light intensity threshold, it is determined that the multi-directional monitoring terminal is removed from the sample of the power material, and the removal warning information and the positioning information are immediately triggered for upload. When the status of the spot-check task is "sample delivery preparation", "sample delivery start", "sample delivery completed", and "sample receiving start", the multi-directional monitoring terminal is bound to the sample of the power material for transportation. If the terminal does not detect any abnormalities in the pressure switch and light intensity data, it uploads the terminal information and positioning information every 2 minutes. When the status of the spot-check task is "sample receiving completed", it automatically switches to "idle", indicating that the previous power material has completed the spot-check.

[0094] In order to distinguish the normal vibration of power materials in the spot-check, sample delivery, and sample receiving links (such as bumps during transportation, sharp turns, and vibrations during braking) from abnormal impacts (such as manual disassembly and damage, replacement, dropping, and severe collisions), improve the recognition accuracy, reduce the false alarm rate of the multi-directional monitoring terminal, and improve the reliability, during the process of analyzing the status of power materials, an intelligent analysis module is deployed on the multi-directional monitoring terminal or the monitoring management platform, which works in cooperation with the time-series feature extraction module to analyze and process the vibration time-series data collected by the multi-directional monitoring terminal, and realizes the accurate recognition of the normal vibration and abnormal impact of the spot-checked materials. Refer to Figure 8 , specifically including:

[0095] Noise reduction processing is performed on the vibration time series data according to the sampling inspection task status of the power materials to obtain standard vibration time series data;

[0096] Combined with a preset silent threshold, the standard vibration time series data is judged to give a judgment result;

[0097] According to the judgment result, the standard vibration time series data is windowed to obtain multiple standard vibration time series sub-data;

[0098] The frequency domain features and time domain features in multiple standard vibration time series sub-data are fused to determine the first target vibration feature;

[0099] Based on a pre-constructed vibration analysis model, combined with the first target vibration feature and the second target vibration feature, the power material state is given, where the second target vibration feature is obtained through a pre-constructed time series analysis model;

[0100] Combined with the analysis of the power material state, an alarm message is output.

[0101] In the example provided by the present invention, the time series analysis model is a long short-term memory network model. The LSTM part in the vibration analysis model can also be used to obtain the second target vibration feature, which is not limited thereto.

[0102] Further, referring to Figure 9 , the first target vibration feature includes at least one of a waveform factor, a zero-crossing rate, and a maximum spectral amplitude;

[0103] Fusing the frequency domain features and time domain features in multiple standard vibration time series sub-data to determine the first target vibration feature specifically includes:

[0104] Analyze the fluctuation conditions of multiple standard time series sub-data in the time domain to give time domain features;

[0105] Analyze the proportion of different frequency bands of multiple standard time series sub-data in the frequency domain to give frequency domain features;

[0106] Combined with the time domain features and frequency domain features, at least one of a waveform factor, a zero-crossing rate, and a maximum spectral amplitude is given.

[0107] In the fourth specific embodiment, the material sampling inspection data includes vibration time series data, which is XYZ-axis vibration time series data collected by a triaxial acceleration sensor (ADXL345) in a multi-directional monitoring terminal. In the example provided by the present invention, the sampling frequency of the above triaxial acceleration sensor is 200 Hz, and the measurement range is set to ±16 g. By setting the sampling frequency, the triaxial acceleration sensor can balance high-frequency vibration capture and power consumption. By setting the measurement range, it can cover the possible maximum impact generated during the sample delivery and receiving process. The data generated by the triaxial acceleration sensor is a time series. It is necessary to filter and denoise the vibration time series data, and perform sliding window segmentation, and then extract statistical features to obtain the first target vibration feature.

[0108] In the example provided by the present invention, a low-pass filter is used to remove high-frequency noise in the vibration time series data, such as the jitter of the sensor itself, to obtain standard vibration time series data. The standard vibration time series data is segmented into 5-second windows (500 sampling points / axis) with an overlap rate of 50%, ensuring that continuous events in the standard vibration time series data are not missed and reducing the consumption of computing resources. The window segmentation is only activated when the acceleration peak exceeds the silence threshold (such as 0.5 g), and the window is not segmented in the idle state to save computing resources. When the acceleration peak exceeds the silence threshold, it indicates that an abnormal situation may exist, and more accurate analysis of the data is required. When the acceleration peak does not exceed the silence threshold, it indicates that the sample of the electrical material is in a stationary state or the transportation is relatively stable. The above silence threshold can be set according to the actual scenario and is not limited here.

[0109] After completing the window segmentation of the standard vibration time series data, multiple standard vibration time series sub-data are obtained. For a certain standard vibration time series sub-data, the following specific processing is carried out:

[0110] For time-domain features, calculate the mean, variance, peak value, peak-to-peak value, root mean square, for the X-axis, Y-axis, and Z-axis, five dimensions for each axis, a total of 15 dimensions;

[0111] For frequency-domain features, calculate the low-frequency band energy ratio and high-frequency band energy ratio respectively after fast Fourier transform. Among them, the low-frequency band energy ratio is used to represent normal vibration, and the high-frequency band energy ratio is used to represent abnormal impact, two dimensions for each axis, a total of 6 dimensions. In this example, the range from 0 Hz to 50 Hz is the low-frequency band, and the range from 50 Hz to 200 Hz is the high-frequency band.

[0112] For the time-frequency joint features, calculate the waveform factor, zero-crossing rate, and maximum spectral amplitude. There are 9 dimensions in total, with 3 dimensions for each axis. Among them, the waveform factor is the ratio of the mean to the root mean square, which is used to reflect the vibration degree of the power material samples. The zero-crossing rate (ZCR) is the number of times the signal passes through zero within a unit time, that is, the frequency at which the signal changes from positive to negative or from negative to positive. The maximum spectral amplitude is the maximum spectral amplitude after the standard vibration time series sub-data undergoes a fast Fourier transform.

[0113] Specifically expressed as:

[0114]

[0115] Peak = max(|X i |)

[0116] PtoP = max(X i ) - min(X i )

[0117]

[0118] where μ is the mean, σ 2 is the variance, Peak is the peak value, PtoP is the peak-to-peak value, RMS is the root mean square, N is the total number of sampling points in each standard vibration time series sub-data, and X i is the acceleration value of the i-th sampling point in each standard vibration time series sub-data.

[0119] The peak value is the maximum value of the absolute value of the acceleration within each window. The peak-to-peak value is the difference between the maximum and minimum values of the acceleration within each window. The root mean square represents the average energy intensity of the signal.

[0120] Perform a fast Fourier transform on each standard vibration time series sub-data to obtain the spectral amplitude |X(f)|, where f is the signal frequency, then

[0121]

[0122] where LE is the low-frequency energy ratio corresponding to each standard vibration time series sub-data, and HE is the high-frequency energy ratio corresponding to each standard vibration time series sub-data.

[0123] The waveform factor ρ is specifically expressed as: ρ = μ / RMS

[0124] The zero-crossing rate ZCR is specifically expressed as:

[0125]

[0126] To eliminate the dimensional differences between various features and at the same time combine the data characteristics of different features, different standardization methods are used to process each feature.

[0127] For the mean, variance, peak value, peak-to-peak value, and root mean square, Min-Max normalization is used for data processing. For the low-frequency band energy ratio, high-frequency band energy ratio, waveform factor, zero-crossing rate, and maximum spectral amplitude, Z-Score normalization is used for data processing.

[0128] Further, referring to Figure 10 , the power material status is determined, specifically including:

[0129] The first target vibration feature and the second target vibration feature are spliced and normalized to obtain a comprehensive vibration feature;

[0130] Based on the pre-constructed vibration analysis model, time series analysis is performed on the comprehensive vibration feature to obtain a time series analysis result;

[0131] According to the time series analysis result, the classification layer of the vibration analysis model is used to classify the comprehensive vibration feature to give the power material status.

[0132] In a specific implementation manner, the vibration pre-analysis model is a random forest model based on the Long Short-Term Memory (LSTM) network. LSTM is used to extract the time series features of the vibration data, that is, the second target vibration feature, which is fused with the first target vibration feature and used as the input of the random forest classifier for classification and recognition of the vibration mode.

[0133] By setting the first memory network layer to return the complete time step sequence, setting the second memory network layer to return the final time step output. Setting the fully connected layer to output the second target vibration feature, where the dropout rate is set in the dropout layer of the fully connected layer to avoid overfitting. In the example provided by the present invention, the first memory network layer is set with 64 LSTM units to return the complete time step sequence. The second memory network layer is set with 128 LSTM units to return the final time step output. The dropout layer sets the dropout rate to 0.3 to prevent overfitting. The fully connected layer outputs the final time series feature vector, which is used to characterize the time series pattern of the vibration signal.

[0134] The LSTM part including the first memory network layer, the second memory network layer, and the fully connected layer can process time series vibration data and capture long-term dependence relationships. Through multiple layers of neurons, non-linear transformation is performed on the vibration time series data of a certain segmentation window of synchronous input, and a fixed-length vector (128-dimensional) is output, with each dimension representing an abstract time series feature.

[0135] LSTM is good at processing time series data and can capture long-term dependencies and dynamic changes. For example, during transportation, sudden braking may appear as a continuous acceleration change process, while a human collision may be a sudden high-frequency vibration event. LSTM can identify these time series patterns through time step processing, thereby providing richer feature representation.

[0136] After the first vibration target feature and the second vibration target feature are concatenated, they are standardized using Z-Score to obtain a comprehensive vibration feature as the input of the classification layer. Standardizing the concatenated first vibration target feature and the second vibration target feature can ensure that the scales of different features are consistent and avoid the bias of random forest due to differences in feature dimensions.

[0137] By directly splicing the first vibration target feature and the second vibration target feature, the vibration analysis model is made more dependent on the comprehensive vibration feature, thereby improving the ability to distinguish similar events (such as sudden braking and human collision).

[0138] The parameters of the random forest classifier, i.e., the number of classification layer trees, maximum depth, category weight, etc., are set, and the classification probability is output in combination with the comprehensive vibration characteristics to obtain the power material status.

[0139] Further, refer to Figure 11 , a pre-built vibration analysis model, is determined by:

[0140] Acquire a vibration training data set, wherein each data in the vibration training data set is marked with a vibration event label;

[0141] Construct the first memory network layer and the second memory network layer, combine them with the vibration training data set, perform pre-training until convergence, and obtain a vibration pre-analysis model;

[0142] The model parameters of the first memory network layer and the second memory network layer in the vibration pre-analysis model are frozen, and the full connection layer and the classification layer are trained with the loss function until convergence to obtain the vibration analysis model.

[0143] In a specific implementation, historical data is used to construct a vibration training data set required for model training. For example, the original vibration data collected by a three-axis acceleration sensor is extracted from historical sampling tasks of electric power materials, and samples containing typical events are preferentially selected, including samples of smooth transportation, sharp turns, and braking in normal vibration, and samples of falls (large instantaneous acceleration in the vertical direction), collisions (multi-axis high-frequency oscillations), and human damage (abnormal continuous periodic bumps, instantaneous impacts, high-frequency vibrations, and severe vibrations during disassembly) in abnormal impacts.

[0144] Label the data in the vibration training dataset, label normal vibrations as "0" and abnormal shocks as "1". Divide the vibration training dataset into a training set, a validation set, and a test set according to a certain proportion based on the time series.

[0145] The vibration analysis model adopts a phased training strategy. First, perform LSTM pre-training on the first memory network layer, the second memory network layer, and the fully connected layer, and then perform joint fine-tuning training in combination with the classification layer.

[0146] In the LSTM pre-training stage, first set the hyperparameters, including batch size, number of training epochs, etc. Use part of the data in the vibration training dataset for LSTM pre-training to learn the long-term dependence relationship and dynamic patterns of the vibration signals in the power material sampling and sample receiving links, and provide a good initial state for the LSTM model parameters.

[0147] After completing the pre-training, freeze the model parameters of the first memory network layer and the second memory network layer in the vibration pre-analysis model, that is, retain the model parameters of the first memory network layer and the second memory network layer, and only train the fully connected layer and the classification layer. The classification layer uses a random forest classifier.

[0148] Construct the loss function of the vibration analysis model. Use Focal Loss as the loss function to reduce the loss weight of easily distinguishable vibration category samples and increase the loss weight of difficult-to-distinguish vibration category samples, so that the model pays more attention to those samples that are difficult to distinguish and optimize the model's ability to identify abnormal shocks.

[0149] The loss function FL(p t ) is specifically expressed as:

[0150] FL(p t ) = -α t (1 - p t ) γ log(p t )

[0151]

[0152] Among them, p t is the predicted probability of the vibration analysis model for the power material state. y = 1 represents an abnormal shock, y = 0 represents a normal vibration, p is the probability that the power material state is an abnormal shock, α t is the balance factor used to adjust the weights of positive and negative samples, and γ is the adjustment factor used to control the weight difference between easy and difficult samples. In this example, the adjustment factor is preset to 2.

[0153] In other embodiments, after the training of the vibration analysis model is completed, the trained vibration analysis model is evaluated using the data in the test set. In this example, the F1 score is used as the model evaluation metric, which is specifically expressed as:

[0154]

[0155] Among them, F1 _ Score is the F1 score, TP is the number of correctly identified abnormal shock samples, FP is the number of false alarm normal vibration samples, and FN is the number of missed abnormal shock samples.

[0156] The larger the F1 score, the better the balance between the precision and recall of the vibration analysis model, and the more accurate the prediction effect for the positive class.

[0157] If the value of the F1 score is not within the preset model evaluation range, it indicates that the training effect of the vibration analysis model is limited and further optimization is required. The optimization of the vibration analysis model can be achieved through multiple trainings, or by combining the loss function and the F1 score, and using Bayesian search to optimize the hyperparameters in the vibration analysis model. For example, optimizing the number of LSTM units (search range 32 - 128), the number of random forest trees (search range 100 - 300), the maximum depth (search range 10 - 50), etc. At the same time, optimize the optimal combination of the balance factor and the adjustment factor in the loss function.

[0158] Referring to Figure 12 , the present invention performs noise reduction filtering, sliding window segmentation, and statistical feature extraction on the collected material sampling inspection data to obtain the first target vibration feature. At the same time, an LSTM model is used to perform time series analysis on the material sampling inspection data, extract time series features, and give the second target vibration feature. The first target vibration feature and the second target vibration feature are fused to obtain a comprehensive vibration feature. Finally, the comprehensive vibration feature is input into the random forest classifier to obtain the corresponding power material state, that is, the power material corresponding to the material sampling inspection data is in a normal vibration state or an abnormal shock state.

[0159] The present invention combines LSTM and random forest by adopting a random forest model based on long short-term memory network. At the same time, the first target vibration feature and the second target vibration feature are fused. The first target vibration feature is a statistical feature and is sensitive to certain transient features. The second target vibration feature is obtained through LSTM, and LSTM can capture the dynamic changes in time, that is, the order or duration of events. The combination of the first target vibration feature and the second target vibration feature can describe the data more comprehensively, provide a more comprehensive discrimination basis for the random forest, and improve the classification accuracy.

[0160] For example, an abnormal impact (such as a drop) may be manifested as a sudden peak (statistical feature) and a continuous pattern of high-frequency oscillation (LSTM time-series feature). The combination of the two can improve the classification accuracy. By combining time-series and statistical features, multi-modal feature fusion is achieved, the accuracy of anomaly detection is improved, and similar events such as sudden braking and human collision can be more accurately distinguished.

[0161] Referring to Figure 13 , an embodiment of the present invention provides a multi-directional monitoring device based on random inspection of electric power materials, including:

[0162] A data acquisition unit 201, configured to acquire random inspection data of materials in real time, where the random inspection data of materials is data collected by a multi-directional monitoring terminal during the random inspection of electric power materials, and the electric power materials are fixed on the multi-directional monitoring terminal during the random inspection process;

[0163] A data analysis unit 202, configured to analyze the random inspection data of materials based on the random inspection task status of the electric power materials, and output the status and / or alarm information of the electric power materials, so as to complete the monitoring of the electric power materials during the random inspection process.

[0164] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the described modules can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0165] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed to include the preferred embodiments and all changes and modifications falling within the scope of the present invention. Obviously, those skilled in the art can make various changes and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A multi-directional monitoring method based on random inspection of power materials, characterized in that: include: Real-time acquisition of material sampling data, where the material sampling data is data collected by the multi-directional monitoring terminal during the sampling process of power materials, and the power materials are fixed on the multi-directional monitoring terminal during the sampling process; Based on the inspection task status of power materials, the material inspection data is analyzed, and the power material status and / or alarm information is output to complete the monitoring of power materials during the inspection process.

2. The multi-directional monitoring method based on random inspection of electric power materials according to claim 1 is characterized in that: Material sampling data includes pressure switch status and light intensity data; Based on the inspection task status of power materials, analyze the material inspection data and output the power material status and / or alarm information, including: If the status of the sampling task of the power materials is that the sample is sealed, the pressure switch status and light intensity data are judged; If the pressure switch is in the raised state or the light intensity data reaches the light threshold, the removal alarm information and the location information of the power materials are output; If the pressure switch state is decreasing and the light intensity data is less than the light threshold, combined with the preset sealing time interval, the power material state is output.

3. The multi-directional monitoring method based on random inspection of electric power materials according to claim 2 is characterized in that: Based on the inspection task status of power materials, analyze the material inspection data and output the power material status and / or alarm information, including: If the sampling task status of the power materials is in the sample delivery stage, the pressure switch status and light intensity data are judged; If the pressure switch is in the raised state or the light intensity data reaches the light threshold, the removal alarm information and the location information of the power materials are output; If the pressure switch state is decreasing and the light intensity data is less than the light threshold, the power material state is output in combination with the preset sample delivery time interval.

4. The multi-directional monitoring method based on random inspection of electric power materials according to claim 1 is characterized in that: Material sampling data includes acceleration data and positioning data; Based on the inspection task status of power materials, analyze the material inspection data and output the power material status and / or alarm information, including: If the status of the sampling task of the power materials is that the sample is sealed, the acceleration data and positioning data are judged; If the acceleration data exceeds the speed threshold or the positioning data exceeds the preset sealing area, the position deviation alarm information and the location information of the power materials are output; If the acceleration data is less than the speed threshold and the positioning data is within the preset sealing area, combined with the preset sealing time interval, the power material status is output.

5. The multi-directional monitoring method based on random inspection of electric power materials according to claim 1 is characterized in that: Material sampling data includes vibration time series data; Based on the inspection task status of power materials, analyze the material inspection data and output the power material status and / or alarm information, including: According to the inspection task status of power materials, the vibration time series data is subjected to noise reduction processing to obtain standard vibration time series data; Combined with the preset silence threshold, the standard vibration time series data is judged and the judgment result is given; According to the judgment result, the standard vibration time series data is window segmented to obtain a plurality of standard vibration time series sub-data; Fusing frequency domain features and time domain features in a plurality of standard vibration time series sub-data to determine a first target vibration feature; Based on the pre-built vibration analysis model, the power material status is given by combining the first target vibration characteristic and the second target vibration characteristic, wherein the second target vibration characteristic is obtained by the pre-built time series analysis model; Combined with the analysis of the status of power materials, alarm information is output.

6. The multi-directional monitoring method based on random inspection of electric power materials according to claim 5 is characterized in that: The first target vibration characteristic includes at least one of a waveform factor, a zero crossing rate, and a maximum spectrum amplitude; The frequency domain features and time domain features in multiple standard vibration time series sub-data are integrated to determine the first target vibration feature, which specifically includes: Analyze the fluctuation of multiple standard time series sub-data in the time domain and give the time domain characteristics; Analyze the proportion of multiple standard time series sub-data in different frequency bands in the frequency domain and give the frequency domain characteristics; At least one of the waveform factor, the zero-crossing rate, and the maximum spectrum amplitude is given by combining the time domain characteristics and the frequency domain characteristics.

7. The multi-directional monitoring method based on random inspection of electric power materials according to claim 5 is characterized in that: Based on the pre-built vibration analysis model, combined with the first target vibration characteristics and the second target vibration characteristics, the power material status is given, including: The first target vibration feature and the second target vibration feature are concatenated and normalized to obtain a comprehensive vibration feature; Based on the pre-built vibration analysis model, the comprehensive vibration characteristics are analyzed in time series to obtain the time series analysis results; According to the results of time series analysis, the classification layer of the vibration analysis model is used to classify the comprehensive vibration characteristics and give the status of power materials.

8. The multi-directional monitoring method based on random inspection of electric power materials according to claim 7 is characterized in that: Pre-built vibration analysis models, determined by: Acquire a vibration training data set, wherein each data in the vibration training data set is marked with a vibration event label; Construct the first memory network layer and the second memory network layer, combine them with the vibration training data set, perform pre-training until convergence, and obtain a vibration pre-analysis model; The model parameters of the first memory network layer and the second memory network layer in the vibration pre-analysis model are frozen, and the full connection layer and the classification layer are trained with the loss function until convergence to obtain the vibration analysis model.

9. A multi-directional monitoring device based on random inspection of power materials, characterized in that: The multi-directional monitoring method based on random inspection of electric power materials as described in any one of claims 1 to 8 is adopted, comprising: A data acquisition unit is used to acquire material sampling data in real time, wherein the material sampling data is data collected by a multi-directional monitoring terminal during the sampling process of electric power materials, and the electric power materials are fixed on the multi-directional monitoring terminal during the sampling process; The data analysis unit is used to analyze the material sampling data based on the sampling task status of the power materials, and output the power material status and / or alarm information to complete the monitoring of the power materials during the sampling process.

10. A multi-directional monitoring terminal based on random inspection of power materials, characterized in that: It includes a controller module, a pressure switch, a mobile communication module, a positioning module, a light sensor and a vibration sensor, and the pressure switch, the mobile communication module, the positioning module, the light sensor and the vibration sensor are all connected to the controller module; Pressure switch, used to obtain the pressure switch status; A mobile communication module, used to establish a communication channel with the controller module; Positioning module, used to obtain the location information of power materials and output positioning data; Light sensor, used to obtain light intensity data; Vibration sensor for acquiring acceleration data and vibration time series data A controller module is used to obtain the pressure switch status, positioning data, light intensity data, acceleration data and vibration timing data, and to execute the multi-directional monitoring method based on random inspection of power materials as described in any one of claims 1-8.

Images