Intelligent monitoring and early warning system for power distribution cabinet

Abstract

The invention relates to the technical field of power distribution cabinet monitoring, in particular to a power distribution cabinet intelligent monitoring and early warning system, which comprises the following steps: firstly, acquiring a first external image, a second internal image, first data and second data of a power distribution cabinet in real time; secondly, respectively analyzing the first external image and the second internal image to obtain a first external corrosion coefficient and a second internal corrosion coefficient; then, analyzing the first data and the second data to obtain a first fault coefficient, a second fault coefficient, a third fault coefficient and a fourth fault coefficient; then, performing weighted fusion on the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient to obtain a comprehensive abnormal coefficient of the power distribution cabinet; and finally, the comprehensive abnormal coefficient is analyzed, and real-time early warning is carried out. The internal and external corrosion and internal faults of the power distribution cabinet are analyzed, so that the safety and the stability of the power distribution cabinet are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution cabinet monitoring, and in particular to an intelligent monitoring and early warning system for a power distribution cabinet. Background Art

[0002] The power distribution cabinet is an indispensable part of the power system. It can distribute, protect and control electric energy, and is also used for power management and maintenance in the power system. It plays an important role in various fields of application, and its performance and reliability will directly affect the stable operation of the power system.

[0003] Intelligent monitoring and early warning of distribution cabinets play an important role in the power system, including real-time monitoring and early warning, fault prediction and prevention, and automated maintenance of the power system, and effectively reducing the need for manual inspections, improving the safety and stability of the power system, and saving the operation and maintenance costs of the power system.

[0004] With the continuous development of science and technology, some intelligent monitoring and early warning technologies for distribution cabinets have been proposed. However, the existing intelligent monitoring and early warning methods for distribution cabinets still have certain defects. For example, the Chinese patent with publication number CN118097904A provides an intelligent monitoring and early warning system, method and distribution cabinet for distribution cabinets. The monitoring and early warning system includes: a human body detection module, which is used to detect whether there is someone at the distribution cabinet when the distribution cabinet alarms; if so, a person signal is sent; an image acquisition module, after receiving the person signal, collects image information at the distribution cabinet; an identification module is used to identify whether the person in the image information is wearing protective tools, if so, an unlocking signal is sent, if not, a prompt is sent signal; a prompt module for receiving and responding to the prompt signal to output a first prompt message; a lock module for receiving and responding to an unlock signal to unlock the distribution cabinet; for another example, a Chinese patent with publication number CN116106648A provides a method for intelligent monitoring of a distribution cabinet, including: S1, receiving maintenance information; S2, authenticating the equipment to be maintained, and jumping to step S3 in response to the authentication of the equipment to be maintained being passed; S3, obtaining device association information, and calculating a first transformation ratio value based on the device association information to generate first configuration information; S4, sending the first configuration information to the distribution cabinet system; S5, receiving the second configuration information issued by the distribution cabinet system, and performing configuration verification. The present invention authenticates the current transformer, and calculates the transformation ratio value of the current transformer based on the basic information of the current transformer and the multi-function ammeter, and automatically configures the current transformer based on the transformation ratio value; the above prior art fails to fully consider the corrosion of the distribution cabinet on the one hand, and fails to fully consider the fault conditions inside the distribution cabinet on the other hand, which easily affects the safety and stability of the distribution cabinet.

[0005] Therefore, an intelligent monitoring and early warning system for distribution cabinets is proposed. Summary of the invention

[0006] The purpose of the present invention is to provide an intelligent monitoring and early warning system for a distribution cabinet. First, a first external image and a second internal image of the distribution cabinet and first data and second data of the distribution cabinet are obtained in real time; secondly, the first external image and the second internal image are analyzed respectively to obtain a first external corrosion coefficient and a second internal corrosion coefficient; then, the first data and the second data are analyzed to obtain a first fault coefficient, a second fault coefficient, a third fault coefficient and a fourth fault coefficient; then, the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient are weightedly fused to obtain a comprehensive abnormality coefficient of the distribution cabinet; finally, the comprehensive abnormality coefficient of the distribution cabinet is analyzed and a real-time early warning is issued. The present invention effectively improves the safety and stability of the distribution cabinet by analyzing the internal and external corrosion and internal faults of the distribution cabinet.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] An intelligent monitoring and early warning system for a power distribution cabinet, comprising:

[0009] An image acquisition and data acquisition module, used for acquiring a first external image, a second internal image, first data, and second data of the power distribution cabinet in real time;

[0010] A distribution cabinet corrosion coefficient calculation module, used for preprocessing and feature extraction of the first external image and the second internal image respectively to obtain a first external corrosion feature and a second internal corrosion feature, and analyzing the first external corrosion feature and the second internal corrosion feature to obtain a first external corrosion coefficient and a second internal corrosion coefficient;

[0011] A distribution cabinet fault coefficient calculation module is used to perform data preprocessing and feature extraction on the first data and the second data to obtain a comprehensive fault feature vector, wherein the comprehensive fault feature vector includes a first fault feature vector, a second fault feature vector, a third fault feature vector, a fourth fault feature vector and an attention mechanism weighted feature vector, and analyze the five feature vectors to obtain a first fault coefficient, a second fault coefficient, a third fault coefficient and a fourth fault coefficient;

[0012] A comprehensive abnormality coefficient calculation module for a power distribution cabinet, used for weighted fusion of the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient to obtain a comprehensive abnormality coefficient of the power distribution cabinet;

[0013] The intelligent early warning module of the power distribution cabinet is used to compare the comprehensive abnormality coefficient with the preset early warning threshold, and to issue a real-time early warning if the comprehensive abnormality coefficient is greater than the preset early warning threshold.

[0014] Preferably, the first external image includes an external surface image, a door lock image, a hinge image and an external junction box image of the distribution cabinet acquired by a high-resolution camera; the second internal image includes an internal wiring image, a switch image, a contactor image and a relay image of the distribution cabinet acquired by a high-resolution camera; the first data includes current and voltage data of the distribution cabinet acquired by current sensors and voltage sensors respectively; the second data includes temperature data, humidity data and vibration data of the distribution cabinet acquired by temperature sensors, humidity sensors and vibration sensors respectively.

[0015] Preferably, the distribution cabinet corrosion coefficient calculation module includes:

[0016] An image preprocessing unit, configured to perform image preprocessing on the first external image and the second internal image to obtain first external image detection data and second internal image detection data;

[0017] A feature extraction unit, configured to extract features from the first external image detection data and the second internal image detection data by using a pre-trained CNN model to obtain a first external corrosion feature and a second internal corrosion feature;

[0018] The corrosion coefficient calculation unit is used to analyze the first external corrosion characteristic and the second internal corrosion characteristic to obtain a first external corrosion coefficient and a second internal corrosion coefficient.

[0019] Preferably, the image preprocessing unit performs denoising, contrast enhancement and edge detection on the first external image and the second internal image respectively to obtain first external image detection data and second internal image detection data.

[0020] Preferably, the first external corrosion features include external surface corrosion features, external door lock corrosion features, external hinge corrosion features and external junction box corrosion features, and the second internal corrosion features include internal wiring corrosion features, internal switch corrosion features, internal contactor corrosion features and internal relay corrosion features, and the above external corrosion features and internal corrosion features all include area features, perimeter features, shape features, color features and texture features.

[0021] Preferably, the first external corrosion feature is:

[0022] F ex =[F surface ,F lock ,F hinge ,F box ];

[0023] F surface,i =[Asurface,i ,P surface,i ,S surface,i ,C surface,i ,T surface,i ];

[0024] F lock,i =[A lock,i ,P lock,i ,S lock,i ,C lock,i ,T lock,i ];

[0025] F hinge,i =[A hinge,i ,P hinge,i ,S hinge,i ,C hinge,i ,T hinge,i ];

[0026] F box,i =[A box,i ,P box,i ,S box,i ,C box,i ,T box,i ];

[0027] Among them, F ex Represented as the first external corrosion feature; F surface Represents the external surface corrosion characteristics; F lock Represents the corrosion characteristics of external door locks; F hinge Represents the external hinge corrosion feature; F box Represents the corrosion characteristics of the external junction box; A surface,i , P surface,i , S surface,i , C surface,i , T surface,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area on the external surface; A lock,i , P lock,i , S lock,i , C lock,i , T lock,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external door lock; A hinge,i , P hinge,i , S hinge,i , C hinge,i , T hinge,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external hinge; A box,i , P box,i , S box,i , C box,i , Tbox,i They represent the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external junction box respectively;

[0028] The second internal corrosion characteristic is:

[0029] F in =[F writing ,F switch ,F contactor ,F relay ];

[0030] F writing,i =[A writing,i ,P writing,i ,S writing,i ,C writing,i ,T writing ,i];

[0031] F switch,i =[A switch,i ,P switch,i ,S switch,i ,C switch,i ,T switch,i ];

[0032] F contactor,i =[A contactor,i ,P contactor,i ,S contactor,i ,C contactor,i ,T contactor,i ];

[0033] F relay,i =[A relay,i ,P relay,i ,S relay,i ,C relay,i ,T relay,i ];

[0034] Among them, F in Represented as the second internal corrosion feature; F writing Indicates the internal wiring corrosion characteristics; F switch Represents the internal switch corrosion characteristics; F contactor Represents the internal contactor corrosion characteristics; F relay Represents the internal relay corrosion characteristics; A writing,i , P writing,i , S writing,i , C writing,i , T writing,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal wiring; A switch,i , P switch,i , S switch,i , Cswitch,i , T switch,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal switch; A contactor,i , P contactor,i , S contactor,i , C contactor,i , T contactor,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal contactor; A relay,i , P relay,i , S relay,i , C relay,i , T relay,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal relay.

[0035] Preferably, the first external corrosion coefficient is:

[0036] CI ex =α surface *CI surface +α lock *CI lock +α hinge *CI hinge +α box *CI box ;

[0037]

[0038] Among them, CI ex Expressed as the first external corrosion coefficient; CI surface Expressed as external surface corrosion coefficient; CI lock Expressed as the external door lock corrosion coefficient; CI hinge Expressed as the external hinge corrosion coefficient; CI box Expressed as the external junction box corrosion coefficient; α surface Expressed as the external surface corrosion coefficient weight; α lock Expressed as the external door lock corrosion coefficient weight; α hinge Expressed as the external hinge corrosion coefficient weight; α box Expressed as the external junction box corrosion coefficient weight; N surface Expressed as the number of external surface box corrosion areas; N lock Expressed as the number of external door lock corrosion areas; N hinge Expressed as the number of corrosion areas of the external hinge box; N box Expressed as the number of corrosion areas in the external junction box;

[0039] The second internal corrosion coefficient is:

[0040] CI in =β writing *CI writing +β switch *CI switch +β contactor *CI contactor +β relay *CI relay ;

[0041]

[0042]

[0043] Among them, CI in Expressed as the second internal corrosion coefficient; CI writing Expressed as the internal wiring corrosion coefficient; CI switch Expressed as the internal switch corrosion coefficient; CI contactor Expressed as the internal contactor corrosion coefficient; CI relay Expressed as the internal relay corrosion coefficient; β writing Expressed as the internal wiring corrosion coefficient weight; β switch Expressed as the internal switch corrosion coefficient weight; β contactor Expressed as the weight of the internal contactor corrosion coefficient; β relay Expressed as the internal relay corrosion coefficient weight; N writing Expressed as the number of internal wiring corrosion areas; N contactor Expressed as the number of internal contactor corrosion areas; N relay Expressed as the number of internal relay corrosion areas.

[0044] Preferably, the distribution cabinet failure coefficient calculation module includes:

[0045] A data preprocessing unit, used for performing data preprocessing on the first data and the second data to obtain first detection data and second detection data; the data preprocessing includes denoising and normalization;

[0046] A fault feature extraction unit, configured to extract features from the first detection data and the second detection data by using an LSTM network combined with an attention mechanism to obtain the comprehensive fault feature vector;

[0047] A fault coefficient calculation unit is used to analyze the comprehensive fault feature vector to obtain the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient.

[0048] Preferably, the comprehensive fault feature vector is:

[0049] F t =[ΔIt ,ΔV t ,T t ,H t ,X t ,c t ];

[0050] Among them, ΔI t represents the first fault characteristic vector, which is used to represent the current fluctuation characteristics; ΔV t represents the second fault feature vector, which is used to represent the voltage fluctuation feature; T t It is represented as the third fault feature vector, which is used to represent the temperature feature; H t It is represented as the fourth fault feature vector, which is used to represent the humidity feature; t is represented as the weighted feature vector of the attention mechanism, which is used to represent the vibration characteristics of the distribution cabinet; c t Represented as the weighted feature vector of the attention mechanism, used to enhance important features;

[0051] The first fault coefficient is used to describe the poor contact fault coefficient of the power distribution cabinet, and its calculation formula is:

[0052]

[0053] Wherein, FI1 represents the first fault coefficient; λ1 represents the first fault coefficient weight; ΔI k is the kth value of the first fault feature vector; ΔV k is the kth value of the second fault feature vector; T k is the kth value of the third fault feature vector; H k is the kth value of the fourth fault feature vector; μ H is the mean of all values ​​in the fourth fault feature vector; μ X is the mean of all values ​​in the attention mechanism weighted feature vector; c k It is represented as the kth value of the weighted feature vector of the attention mechanism; N represents the number of samples;

[0054] The second fault coefficient is used to describe the short-circuit fault coefficient of the power distribution cabinet, and its calculation formula is:

[0055]

[0056] Wherein, FI2 represents the second fault coefficient; λ2 represents the weight of the second fault coefficient; μ T is the mean of all values ​​in the third fault feature vector;

[0057] The third fault coefficient is used to describe the overload fault coefficient of the power distribution cabinet, and its calculation formula is:

[0058]

[0059] Wherein, FI3 represents the third fault coefficient; λ3 represents the third fault coefficient weight;

[0060] The fourth fault coefficient is used to describe the arc discharge fault coefficient of the power distribution cabinet, and its calculation formula is:

[0061]

[0062] Among them, FI4 represents the fourth fault coefficient; λ4 represents the fourth fault coefficient weight.

[0063] Preferably, the comprehensive abnormality coefficient is:

[0064] CAI=α1*CI ex +α2*CI in +β1*FI1+β2*FI2+β3*FI3+β4*FI4;

[0065] Wherein, CAI represents the comprehensive anomaly coefficient; α1 represents the importance of the first external corrosion coefficient; CI ex Expressed as the first external corrosion coefficient; CI in It is represented as the second internal corrosion coefficient; α2 is represented as the importance of the second internal corrosion coefficient; FI1 is represented as the first fault coefficient; β1 is represented as the importance of the first fault coefficient; FI2 is represented as the second fault coefficient; β2 is represented as the importance of the second fault coefficient; FI3 is represented as the third fault coefficient; β3 is represented as the importance of the third fault coefficient; FI4 is represented as the fourth fault coefficient; β4 is represented as the importance of the fourth fault coefficient.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] 1. The present invention obtains the first external corrosion feature (external surface corrosion feature, external door lock corrosion feature, external hinge corrosion feature and external junction box corrosion feature) and the second internal corrosion feature (internal wiring corrosion feature, internal switch corrosion feature, internal contactor corrosion feature and internal relay corrosion feature) by analyzing the external and internal images of the distribution cabinet, and obtains the first external corrosion coefficient and the second internal corrosion coefficient by analyzing the area feature, perimeter feature, shape feature, color feature and texture feature of the first external corrosion feature and the second internal corrosion feature respectively. The corrosion degree of the distribution cabinet can be reasonably evaluated, and potential corrosion risks can be discovered in time, thereby effectively improving the safety and stability of the distribution cabinet.

[0068] 2. The present invention obtains the fault characteristics (current fluctuation characteristics, voltage fluctuation characteristics, temperature characteristics, humidity characteristics and vibration characteristics) of the distribution cabinet by analyzing various sensor data such as current, voltage, temperature, humidity and vibration of the distribution cabinet, and extracts features of the sensor data through an LSTM network combined with an attention mechanism. The fault characteristics are analyzed to obtain the first fault coefficient, second fault coefficient, third fault coefficient and fourth fault coefficient of the distribution cabinet. Through these four fault coefficients, the poor contact, short circuit, overload and arc discharge fault characteristics of the distribution cabinet can be effectively analyzed, which is beneficial to improving the comprehensiveness and accuracy of fault monitoring of the distribution cabinet, thereby effectively improving the safety and stability of the distribution cabinet.

[0069] 3. The present invention comprehensively considers the corrosion and fault conditions of the distribution cabinet, including analyzing the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient of the distribution cabinet, to obtain the comprehensive abnormality coefficient of the distribution cabinet. The comprehensive abnormality coefficient can more comprehensively evaluate the internal operating state of the distribution cabinet by comprehensively analyzing the corrosion and fault characteristics of the distribution cabinet, thereby effectively improving the safety and stability of the distribution cabinet. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 A flow chart of an intelligent monitoring and early warning system for a power distribution cabinet provided by an embodiment of the present invention;

[0071] Figure 2 A schematic diagram of an intelligent monitoring and early warning system for a power distribution cabinet provided by an embodiment of the present invention;

[0072] Figure 3 A schematic diagram of a corrosion coefficient calculation module for a power distribution cabinet provided in an embodiment of the present invention;

[0073] Figure 4 A schematic diagram of a failure coefficient calculation module for a power distribution cabinet provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0074] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0075] Intelligent monitoring and early warning of power distribution cabinets are of great significance for improving the safety and stability of power distribution cabinets. In the prior art, on the one hand, the corrosion of the power distribution cabinet is not fully considered, and on the other hand, the fault conditions inside the power distribution cabinet are not fully considered, which easily affects the safety and stability of the power distribution cabinet. The present invention will be described in detail by relying on the following two embodiments to prove that the system proposed by the present invention has certain effectiveness in the intelligent monitoring and early warning of the power distribution cabinet.

[0076] Embodiment 1

[0077] This embodiment analyzes the intelligent monitoring and early warning process of power distribution cabinet A;

[0078] The flow chart of the intelligent monitoring and early warning system for power distribution cabinet proposed in the embodiment of the present application corresponds to Figure 1 The steps include: S10. Acquire the first external image and the second internal image of the distribution cabinet and the first data and the second data of the distribution cabinet; S20. Analyze the first external image and the second internal image respectively to obtain the first external corrosion coefficient and the second internal corrosion coefficient of the distribution cabinet; S30. Analyze the first data and the second data to obtain the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient; S40. Weightedly fuse the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient to obtain the comprehensive abnormality coefficient of the distribution cabinet; S50. Compare and analyze the comprehensive abnormality coefficient of the distribution cabinet with the preset warning threshold, and issue a real-time warning if the comprehensive abnormality coefficient of the distribution cabinet is greater than the preset warning threshold;

[0079] Further, in Figure 2 The schematic diagram of a power distribution cabinet intelligent monitoring and early warning system includes multiple modules, including an image acquisition and data acquisition module, a power distribution cabinet corrosion coefficient calculation module, a power distribution cabinet fault coefficient calculation module, a power distribution cabinet comprehensive abnormality coefficient calculation module and a power distribution cabinet intelligent early warning module. Figure 1 The following is a detailed implementation process of the embodiment through specific steps.

[0080] The first external image and the second internal image of the A power distribution cabinet and the first data and the second data of the A power distribution cabinet are collected in real time according to the image collection and data acquisition module; corresponding to step S10;

[0081] The first external image includes 3 external surface images, 2 door lock images, 2 hinge images and 2 external junction box images; the second internal image includes 3 internal wiring images, 2 switch images, 2 contactor images and 2 relay images; wherein the external surface images and internal wiring images are taken from the front, side and back respectively; the door lock images, hinge images, external junction box images, switch images, contactor images and relay images are taken from the front and side respectively, and the above shooting processes are all performed by a high-resolution camera;

[0082] The first data includes the voltage and current data of the A distribution cabinet obtained by the current sensor and the voltage sensor, and the second data includes the temperature data, humidity and vibration data of the A distribution cabinet obtained by the temperature sensor, the humidity sensor and the vibration sensor respectively; the first data and the second data finally form time series data, including 32 samples, each sample includes 100 time steps, and each time step has 5 data (current, voltage, temperature, humidity, vibration); each time step interval is 1 minute; the data format is [32,100,5], which is used as the input layer data of LSTM; some data of the A distribution cabinet are shown in Table 1;

[0083] Table 1 Data of some time steps of the distribution cabinet A

[0084] Time step Current (A) Voltage (V) Temperature(℃) humidity(%) <![CDATA[Vibration (m / s 2 )]]> 1 100.1 230.5 25.3 45.0 0.02 2 99.8 230.6 25.3 45.0 0.03 3 100.0 230.7 25.4 45.1 0.02 ... ... ... ... ... ... 100 100.2 230.4 25.2 45.2 0.01

[0085] The distribution cabinet corrosion coefficient calculation module (corresponding to the above step S20) includes an image preprocessing unit, a feature extraction unit and a corrosion coefficient calculation unit, as shown in the following figure: Figure 3 As shown, the image preprocessing unit performs image preprocessing on the first external image and the second internal image to obtain first external image detection data and second internal image detection data; the image preprocessing process includes Stromberg filtering denoising, contrast enhancement and edge detection;

[0086] The feature extraction unit uses the pre-trained CNN model to perform feature extraction on the first external image detection data and the second internal image detection data to obtain a first external corrosion feature and a second internal corrosion feature; including using a threshold segmentation method to extract corrosion regions from each pre-processed image, and calculating the area feature, perimeter feature, shape feature, color feature and texture feature of each corrosion region; obtaining the first external corrosion feature and the second internal corrosion feature;

[0087] Further, the first external corrosion feature is:

[0088] F ex =[F surface ,F lock ,F hinge ,Fbox ];

[0089] F surface,i =[A surface,i ,P surface,i ,S surface,i ,C surface,i ,T surface,i ];

[0090] F lock,i =[A lock,i ,P lock,i ,S lock,i ,C lock,i ,T lock,i ];

[0091] F hinge,i =[A hinge,i ,P hinge,i ,S hinge,i ,C hinge,i ,T hinge,i ];

[0092] F box,i =[A box,i ,P box,i ,S box,i ,C box,i ,T box,i ];

[0093] Among them, F ex Represented as the first external corrosion feature; F surface Represents the external surface corrosion characteristics; F lock Represents the corrosion characteristics of external door locks; F hinge Represents the external hinge corrosion feature; F box Represents the corrosion characteristics of the external junction box; A surface,i , P surface,i , S surface,i , C surface,i , T surface,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area on the external surface; A lock,i , P lock,i , S lock,i , C lock,i , T lock,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external door lock; A hinge,i , P hinge,i , S hinge,i , C hinge,i , T hinge,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external hinge; Abox,i , P box,i , S box,i , C box,i , T box,i They represent the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external junction box respectively;

[0094] The second internal corrosion characteristic is:

[0095] F in =[F writing ,F switch ,F contactor ,F relay ];

[0096] F writing,i =[A writing,i ,P writing,i ,S writing,i ,C writing,i ,T writing,i ];

[0097] F switch,i =[A switch,i ,P switch,i ,S switch,i ,C switch,i ,T switch,i ];

[0098] F contactor,i =[A contactor,i ,P contactor,i ,S contactor,i ,C contactor,i ,T contactor,i ];

[0099] F relay,i =[A relay,i ,P relay,i ,S relay,i ,C relay,i ,T relay,i ];

[0100] Among them, F in Represented as the second internal corrosion feature; F writing Indicates the internal wiring corrosion characteristics; F switch Represents the internal switch corrosion characteristics; F contactor Represents the internal contactor corrosion characteristics; F relay Represents the internal relay corrosion characteristics; A writing,i , P writing,i , S writing,i , C writing,i , T writing,iare respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal wiring; A switch,i , P switch,i , S switch,i , C switch,i , T switch,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal switch; A contactor,i , P contactor,i , S contactor,i , C contactor,i , T contactor,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal contactor; A relay,i , P relay,i , S relay,i , C relay,i , T relay,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal relay;

[0101] This embodiment performs image processing and analysis on the external and internal components of the distribution cabinet to obtain a first external corrosion feature and a second internal corrosion feature of the distribution cabinet, thereby facilitating intelligent monitoring and early warning of the distribution cabinet, thereby further improving the safety and stability of the distribution cabinet.

[0102] The corrosion coefficient calculation unit is used to analyze the first external corrosion feature and the second internal corrosion feature to obtain a first external corrosion coefficient and a second internal corrosion coefficient of the A power distribution cabinet;

[0103] Further, the first external corrosion coefficient is:

[0104] CI ex =α surface *CI surface +α lock *CI lock +α hinge *CI hinge +α box *CI box ;

[0105]

[0106] Among them, CI ex Expressed as the first external corrosion coefficient; CI surface Expressed as external surface corrosion coefficient; CI lock Expressed as the external door lock corrosion coefficient; CI hinge Expressed as the external hinge corrosion coefficient; CI boxExpressed as the external junction box corrosion coefficient; α surface Expressed as the external surface corrosion coefficient weight; α lock Expressed as the external door lock corrosion coefficient weight; α hinge Expressed as the external hinge corrosion coefficient weight; α box Expressed as the external junction box corrosion coefficient weight; N surface Expressed as the number of external surface box corrosion areas; N lock Expressed as the number of external door lock corrosion areas; N hinge Expressed as the number of corrosion areas of the external hinge box; N box Expressed as the number of corrosion areas in the external junction box;

[0107] The second internal corrosion coefficient is:

[0108] CI in =β writing *CI writing +β switch *CI switch +β contactor *CI contactor +β relay *CI relay ;

[0109]

[0110] Among them, CI in Expressed as the second internal corrosion coefficient; CI writing Expressed as the internal wiring corrosion coefficient; CI switch Expressed as the internal switch corrosion coefficient; CI contactor Expressed as the internal contactor corrosion coefficient; CI relay Expressed as the internal relay corrosion coefficient; β writing Expressed as the internal wiring corrosion coefficient weight; β switch Expressed as the internal switch corrosion coefficient weight; β contactor Expressed as the weight of the internal contactor corrosion coefficient; β relay Expressed as the internal relay corrosion coefficient weight; N writing Expressed as the number of internal wiring corrosion areas; N contactor Expressed as the number of internal contactor corrosion areas; N relay It is expressed as the number of internal relay corrosion areas;

[0111] This embodiment obtains first external corrosion features (external surface corrosion features, external door lock corrosion features, external hinge corrosion features, and external junction box corrosion features) and second internal corrosion features (internal wiring corrosion features, internal switch corrosion features, internal contactor corrosion features, and internal relay corrosion features) by analyzing the external and internal images of the distribution cabinet. The first external corrosion coefficient and the second internal corrosion coefficient are obtained by analyzing the area features, perimeter features, shape features, color features, and texture features of the first external corrosion features and the second internal corrosion features, respectively. This can reasonably evaluate the corrosion degree of the distribution cabinet and promptly discover potential corrosion risks, thereby effectively improving the safety and stability of the distribution cabinet.

[0112] A distribution cabinet fault coefficient calculation module, used for analyzing the first data and the second data to obtain a first fault coefficient, a second fault coefficient, a third fault coefficient and a fourth fault coefficient; corresponding to the above step S30;

[0113] The distribution cabinet fault coefficient calculation module includes a data preprocessing unit, a fault feature extraction unit and a fault coefficient calculation unit. Figure 4 As shown;

[0114] The data preprocessing unit is used to perform data preprocessing on the first data and the second data to obtain first detection data and second detection data;

[0115] Furthermore, the data preprocessing process includes denoising and normalizing the first data and the second data, wherein the denoising is achieved by smoothing the data using a moving average filter;

[0116] The feature extraction unit performs feature extraction through an LSTM network combined with an attention mechanism to obtain the comprehensive fault feature vector, wherein the comprehensive feature vector includes a first fault feature vector, a second fault feature vector, a third fault feature vector, a fourth fault feature vector, a fifth fault feature vector and an attention mechanism weighted feature vector;

[0117] The comprehensive fault feature vector is:

[0118] F t =[ΔI t ,ΔV t ,T t ,H t ,X t ,c t ];

[0119] Among them, ΔI t represents the first fault characteristic vector, which is used to represent the current fluctuation characteristics; ΔV t represents the second fault feature vector, which is used to represent the voltage fluctuation feature; Tt It is represented as the third fault feature vector, which is used to represent the temperature feature; H t It is represented as the fourth fault feature vector, which is used to represent the humidity feature; t is represented as the fifth fault feature vector, used to represent the vibration characteristics of the power distribution cabinet; c t Represented as the weighted feature vector of the attention mechanism, used to enhance important features;

[0120] Among them, the weighted feature vector formula of the attention mechanism is:

[0121]

[0122] Among them, c t Represented as the weighted feature vector of the attention mechanism; α t It is represented as the attention weight of the hidden state at the tth time step; h t Represented as the hidden state of the t-th time step; wherein the hidden state is the output of the LSTM layer of the LSTM at each time step;

[0123] The fault coefficient calculation unit is used to analyze five fault characteristic vectors to obtain the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient;

[0124] The first fault coefficient is used to describe the poor contact fault coefficient of the power distribution cabinet, and its calculation formula is:

[0125]

[0126] Wherein, FI1 represents the first fault coefficient; λ1 represents the first fault coefficient weight; ΔI k is the kth value of the first fault feature vector; ΔV k is the kth value of the second fault feature vector; T k is the kth value of the third fault feature vector; H k is the kth value of the fourth fault feature vector; μ H is the mean of all values ​​in the fourth fault feature vector; μ X is the mean of all values ​​in the fifth fault feature vector; c k It is represented as the kth value of the weighted feature vector of the attention mechanism; N represents the number of samples;

[0127] The second fault coefficient is used to describe the short-circuit fault coefficient of the power distribution cabinet, and its calculation formula is:

[0128]

[0129] Wherein, FI2 represents the second fault coefficient; λ2 represents the weight of the second fault coefficient; μ T is the mean of all values ​​in the third fault feature vector;

[0130] The third fault coefficient is used to describe the overload fault coefficient of the power distribution cabinet, and its calculation formula is:

[0131]

[0132] Wherein, FI3 represents the third fault coefficient; λ3 represents the third fault coefficient weight;

[0133] The fourth fault coefficient is used to describe the arc discharge fault coefficient of the power distribution cabinet, and its calculation formula is:

[0134]

[0135] Among them, FI4 represents the fourth fault coefficient; λ4 represents the fourth fault coefficient weight.

[0136] This embodiment analyzes various sensor data such as current, voltage, temperature, humidity and vibration of the distribution cabinet, and extracts features of the sensor data through an LSTM network combined with an attention mechanism to obtain the fault characteristics of the distribution cabinet (current fluctuation characteristics, voltage fluctuation characteristics, temperature characteristics, humidity characteristics and vibration characteristics), and analyzes the fault characteristics to obtain the first fault coefficient, second fault coefficient, third fault coefficient and fourth fault coefficient of the distribution cabinet. Through these four fault coefficients, the poor contact, short circuit, overload and arc discharge fault characteristics of the distribution cabinet can be effectively analyzed, which is beneficial to improving the comprehensiveness and accuracy of fault monitoring of the distribution cabinet, thereby effectively improving the safety and stability of the distribution cabinet.

[0137] Further, according to the comprehensive abnormality coefficient calculation module of the power distribution cabinet (corresponding to the above step S40), the comprehensive abnormality coefficient is:

[0138] CAI=α1*CI ex +α2*CI in +β1*FI1+β2*FI2+β3*FI3+β4*FI4;

[0139] Wherein, CAI represents the comprehensive anomaly coefficient; α1 represents the importance of the first external corrosion coefficient; CI ex Expressed as the first external corrosion coefficient; CI inIt is represented as the second internal corrosion coefficient; α2 is represented as the importance of the second internal corrosion coefficient; FI1 is represented as the first fault coefficient; β1 is represented as the importance of the first fault coefficient; FI2 is represented as the second fault coefficient; β2 is represented as the importance of the second fault coefficient; FI3 is represented as the third fault coefficient; β3 is represented as the importance of the third fault coefficient; FI4 is represented as the fourth fault coefficient; β4 is represented as the importance of the fourth fault coefficient.

[0140] This embodiment comprehensively considers the corrosion and fault conditions of the distribution cabinet, including analyzing the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient of the distribution cabinet, to obtain the comprehensive abnormality coefficient of the distribution cabinet. The comprehensive abnormality coefficient can more comprehensively evaluate the internal operating state of the distribution cabinet by comprehensively analyzing the corrosion and fault characteristics of the distribution cabinet, thereby effectively improving the safety and stability of the distribution cabinet.

[0141] Furthermore, the intelligent early warning module of the power distribution cabinet includes a preset early warning threshold. When the comprehensive abnormality coefficient is higher than the preset early warning threshold, an early warning will be issued to the power distribution cabinet.

[0142] This implementation first acquires in real time the first external image and the second internal image of the distribution cabinet and the first data and the second data of the distribution cabinet; secondly, analyzes the first external image and the second internal image respectively to obtain the first external corrosion coefficient and the second internal corrosion coefficient; then, analyzes the first data and the second data to obtain the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient; then, performs weighted fusion on the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient to obtain the comprehensive abnormality coefficient of the distribution cabinet; finally, analyzes the comprehensive abnormality coefficient of the distribution cabinet and issues a real-time warning. This embodiment effectively improves the safety and stability of the distribution cabinet by analyzing the internal and external corrosion and internal faults of the distribution cabinet.

[0143] Embodiment 2

[0144] This embodiment analyzes the intelligent monitoring and early warning process of the B distribution cabinet;

[0145] The flow chart of the intelligent monitoring and early warning system for power distribution cabinet proposed in the embodiment of the present application corresponds to Figure 1The steps include: S10. Acquire the first external image and the second internal image of the distribution cabinet and the first data and the second data of the distribution cabinet; S20. Analyze the first external image and the second internal image respectively to obtain the first external corrosion coefficient and the second internal corrosion coefficient of the distribution cabinet; S30. Analyze the first data and the second data to obtain the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient; S40. Weightedly fuse the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient to obtain the comprehensive abnormality coefficient of the distribution cabinet; S50. Compare and analyze the comprehensive abnormality coefficient of the distribution cabinet with the preset warning threshold, and issue a real-time warning if the comprehensive abnormality coefficient of the distribution cabinet is greater than the preset warning threshold;

[0146] Further, in Figure 2 The schematic diagram of a power distribution cabinet intelligent monitoring and early warning system includes multiple modules, including an image acquisition and data acquisition module, a power distribution cabinet corrosion coefficient calculation module, a power distribution cabinet fault coefficient calculation module, a power distribution cabinet comprehensive abnormality coefficient calculation module and a power distribution cabinet intelligent early warning module. Figure 1 The following is a detailed implementation process of the embodiment through specific steps.

[0147] The first external image and the second internal image of the B power distribution cabinet and the first data and the second data of the B power distribution cabinet are collected in real time according to the image collection and data acquisition module; corresponding to step S10;

[0148] The first external image includes 3 external surface images, 2 door lock images, 2 hinge images and 2 external junction box images; the second internal image includes 3 internal wiring images, 2 switch images, 2 contactor images and 2 relay images;

[0149] Among them, the external surface image and the internal wiring image are taken from the front, side and back respectively; the door lock image, hinge image, external junction box image, switch image, contactor image and relay image are taken from the front and side respectively, and the above shooting process is carried out by a high-resolution camera;

[0150] The first data includes the voltage and current data of the B distribution cabinet obtained by the current sensor and the voltage sensor, and the second data includes the temperature data, humidity and vibration data of the B distribution cabinet obtained by the temperature sensor, the humidity sensor and the vibration sensor respectively; the first data and the second data finally form time series data, including 32 samples, each sample includes 100 time steps, and each time step has 5 data (current, voltage, temperature, humidity, vibration); each time step interval is 1 minute; the data format is [32,100,5], which is used as the input layer data of LSTM; some data of the B distribution cabinet are shown in Table 2;

[0151] Table 2B Data of some time steps of the power distribution cabinet

[0152] Time step Current (A) Voltage (V) Temperature(℃) humidity(%) <![CDATA[Vibration (m / s 2 )]]> 1 99.7 228.5 26.3 46.0 0.10 2 98.2 228.6 26.5 47.0 0.11 3 100.0 229.7 26.7 48.1 0.12 ... ... ... ... ... ... 100 100.8 231.0 26.8 48.6 0.15

[0153] The distribution cabinet corrosion coefficient calculation module (corresponding to the above step S20) includes an image preprocessing unit, a feature extraction unit and a corrosion coefficient calculation unit, as shown in the following figure: Figure 3 As shown, the image preprocessing unit performs image preprocessing on the first external image and the second internal image to obtain first external image detection data and second internal image detection data; the image preprocessing process includes Gaussian filtering denoising, contrast enhancement and edge detection;

[0154] The feature extraction unit uses the pre-trained CNN model to perform feature extraction on the first external image detection data and the second internal image detection data to obtain a first external corrosion feature and a second internal corrosion feature; including using a threshold segmentation method to extract corrosion regions from each pre-processed image, and calculating the area feature, perimeter feature, shape feature, color feature and texture feature of each corrosion region; obtaining the first external corrosion feature and the second internal corrosion feature;

[0155] Further, the first external corrosion feature is:

[0156] F ex =[F surface ,F lock ,F hinge ,F box ];

[0157] F surface,i =[A surface,i ,P surface,i ,S surface,i ,C surface,i ,T surface,i ];

[0158] F lock,i =[A lock,i ,P lock,i ,Slock,i ,C lock,i ,T lock,i ];

[0159] F hinge,i =[A hinge,i ,P hinge,i ,S hinge,i ,C hinge,i ,T hinge,i ];

[0160] F box,i =[A box,i ,P box,i ,S box,i ,C box,i ,T box,i ];

[0161] Among them, F ex Represented as the first external corrosion feature; F surface Represents the external surface corrosion characteristics; F lock Represents the corrosion characteristics of external door locks; F hinge Represents the external hinge corrosion feature; F box Represents the corrosion characteristics of the external junction box; A surface,i , P surface,i , S surface,i , C surface,i , T surface,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area on the external surface; A lock,i , P lock,i , S lock,i , C lock,i , T lock,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external door lock; A hinge,i , P hinge,i , S hinge,i , C hinge,i , T hinge,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external hinge; A box,i , P box,i , S box,i , C box,i , T box,i They represent the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external junction box respectively;

[0162] The second internal corrosion characteristic is:

[0163] F in =[F writing ,Fswitch ,F contactor ,F relay ];

[0164] F writing,i =[A writing,i ,P writing,i ,S writing,i ,C writing,i ,T writing,i ];

[0165] F switch,i =[A switch,i ,P switch,i ,S switch,i ,C switch,i ,T switch,i ];

[0166] F contactor,i =[A contactor,i ,P contactor,i ,S contactor,i ,C contactor,i ,T contactor,i ];

[0167] F relay,i =[A relay,i ,P relay,i ,S relay,i ,C relay,i ,T relay,i ];

[0168] Among them, F in Represented as the second internal corrosion feature; F writing Indicates the internal wiring corrosion characteristics; F switch Represents the internal switch corrosion characteristics; F contactor Represents the internal contactor corrosion characteristics; F relay Represents the internal relay corrosion characteristics; A writing,i , P writing,i , S writing,i , C writing,i , T writing,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal wiring; A switch,i , P switch,i , S switch,i , C switch,i , T switch,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal switch; A contactor,i , P contactor,i , S contactor,i , C contactor,i , T contactor,iare respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal contactor; A relay,i , P relay,i , S relay,i , C relay,i , T relay,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal relay;

[0169] The corrosion coefficient calculation unit is used to analyze the first external corrosion feature and the second internal corrosion feature to obtain a first external corrosion coefficient and a second internal corrosion coefficient of the B distribution cabinet;

[0170] Further, the first external corrosion coefficient is:

[0171] CI ex =α surface *CI surface +α lock *CI lock +α hinge *CI hinge +α box *CI box ;

[0172]

[0173] Among them, CI ex Expressed as the first external corrosion coefficient; CI surface Expressed as external surface corrosion coefficient; CI lock Expressed as the external door lock corrosion coefficient; CI hinge Expressed as the external hinge corrosion coefficient; CI box Expressed as the external junction box corrosion coefficient; α surface Expressed as the external surface corrosion coefficient weight; α lock Expressed as the external door lock corrosion coefficient weight; α hinge Expressed as the external hinge corrosion coefficient weight; α box Expressed as the external junction box corrosion coefficient weight; N surface Expressed as the number of external surface box corrosion areas; N lock Expressed as the number of external door lock corrosion areas; N hinge Expressed as the number of corrosion areas of the external hinge box; N box Expressed as the number of corrosion areas in the external junction box;

[0174] The second internal corrosion coefficient is:

[0175] CI in =β writing *CI writing+β switch *CI switch +β contactor *CI contactor +β relay *CI relay ;

[0176]

[0177] Among them, CI in Expressed as the second internal corrosion coefficient; CI writing Expressed as the internal wiring corrosion coefficient; CI switch Expressed as the internal switch corrosion coefficient; CI contactor Expressed as the internal contactor corrosion coefficient; CI relay Expressed as the internal relay corrosion coefficient; β writing Expressed as the internal wiring corrosion coefficient weight; β switch Expressed as the internal switch corrosion coefficient weight; β contactor Expressed as the weight of the internal contactor corrosion coefficient; β relay Expressed as the internal relay corrosion coefficient weight; N writing Expressed as the number of internal wiring corrosion areas; N contactor Expressed as the number of internal contactor corrosion areas; N relay It is expressed as the number of internal relay corrosion areas;

[0178] A distribution cabinet fault coefficient calculation module, used for analyzing the first data and the second data to obtain a first fault coefficient, a second fault coefficient, a third fault coefficient and a fourth fault coefficient; corresponding to the above step S30;

[0179] The distribution cabinet fault coefficient calculation module includes a data preprocessing unit, a fault feature extraction unit and a fault coefficient calculation unit. Figure 4 As shown;

[0180] The data preprocessing unit is used to perform data preprocessing on the first data and the second data to obtain first detection data and second detection data;

[0181] Furthermore, the data preprocessing process includes denoising and normalizing the first data and the second data, wherein the denoising is achieved by smoothing the data using a moving average filter;

[0182] The feature extraction unit performs feature extraction through an LSTM network combined with an attention mechanism to obtain the comprehensive fault feature vector, wherein the comprehensive feature vector includes a first fault feature vector, a second fault feature vector, a third fault feature vector, a fourth fault feature vector, a fifth fault feature vector and an attention mechanism weighted feature vector;

[0183] The comprehensive fault feature vector is:

[0184] F t =[ΔI t ,ΔV t ,T t ,H t ,X t ,c t ];

[0185] Among them, ΔI t represents the first fault characteristic vector, which is used to represent the current fluctuation characteristics; ΔV t represents the second fault feature vector, which is used to represent the voltage fluctuation feature; T t It is represented as the third fault feature vector, which is used to represent the temperature feature; H t It is represented as the fourth fault feature vector, which is used to represent the humidity feature; t is represented as the fifth fault feature vector, used to represent the vibration characteristics of the power distribution cabinet; c t Represented as the weighted feature vector of the attention mechanism, used to enhance important features;

[0186] Among them, the weighted feature vector formula of the attention mechanism is:

[0187]

[0188] Among them, c t Represented as the weighted feature vector of the attention mechanism; α t It is represented as the attention weight of the hidden state at the tth time step; h t Represented as the hidden state of the t-th time step; wherein the hidden state is the output of the LSTM layer of the LSTM at each time step;

[0189] The fault coefficient calculation unit is used to analyze five fault characteristic vectors to obtain the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient;

[0190] The first fault coefficient is used to describe the poor contact fault coefficient of the power distribution cabinet, and its calculation formula is:

[0191]

[0192] Wherein, FI1 represents the first fault coefficient; λ1 represents the first fault coefficient weight; ΔI k is the kth value of the first fault feature vector; ΔV k is the kth value of the second fault feature vector; T k is the kth value of the third fault feature vector; Hk is the kth value of the fourth fault feature vector; μ H is the mean of all values ​​in the fourth fault feature vector; μ X is the mean of all values ​​in the fifth fault feature vector; c k Represented as the kth value of the weighted feature vector of the attention mechanism;

[0193] The second fault coefficient is used to describe the short-circuit fault coefficient of the power distribution cabinet, and its calculation formula is:

[0194]

[0195] Wherein, FI2 represents the second fault coefficient; λ2 represents the weight of the second fault coefficient; μ T is the mean of all values ​​in the third fault feature vector;

[0196] The third fault coefficient is used to describe the overload fault coefficient of the power distribution cabinet, and its calculation formula is:

[0197]

[0198] Wherein, FI3 represents the third fault coefficient; λ3 represents the third fault coefficient weight;

[0199] The fourth fault coefficient is used to describe the arc discharge fault coefficient of the power distribution cabinet, and its calculation formula is:

[0200]

[0201] Among them, FI4 represents the fourth fault coefficient; λ4 represents the fourth fault coefficient weight.

[0202] Furthermore, the comprehensive abnormality coefficient is:

[0203] CAI=α1*CI ex +α2*CI in +β1*FI1+β2*FI2+β3*FI3+β4*FI4;

[0204] Wherein, CAI represents the comprehensive anomaly coefficient; α1 represents the importance of the first external corrosion coefficient; CI ex Expressed as the first external corrosion coefficient; CI inIt is represented as the second internal corrosion coefficient; α2 is represented as the importance of the second internal corrosion coefficient; FI1 is represented as the first fault coefficient; β1 is represented as the importance of the first fault coefficient; FI2 is represented as the second fault coefficient; β2 is represented as the importance of the second fault coefficient; FI3 is represented as the third fault coefficient; β3 is represented as the importance of the third fault coefficient; FI4 is represented as the fourth fault coefficient; β4 is represented as the importance of the fourth fault coefficient.

[0205] Furthermore, the intelligent early warning module of the power distribution cabinet includes a preset early warning threshold. When the comprehensive abnormality coefficient is higher than the preset early warning threshold, an early warning will be issued to the power distribution cabinet.

[0206] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent monitoring and early warning system for a power distribution cabinet, characterized in that: include: An image acquisition and data acquisition module, used for acquiring a first external image, a second internal image, first data, and second data of the power distribution cabinet in real time; A distribution cabinet corrosion coefficient calculation module, used for performing image preprocessing and feature extraction on the first external image and the second internal image respectively to obtain a first external corrosion feature and a second internal corrosion feature, and analyzing the first external corrosion feature and the second internal corrosion feature to obtain a first external corrosion coefficient and a second internal corrosion coefficient; A distribution cabinet fault coefficient calculation module is used to perform data preprocessing and feature extraction on the first data and the second data to obtain a comprehensive fault feature vector, wherein the comprehensive fault feature vector includes a first fault feature vector, a second fault feature vector, a third fault feature vector, a fourth fault feature vector and an attention mechanism weighted feature vector, and analyze the five feature vectors to obtain a first fault coefficient, a second fault coefficient, a third fault coefficient and a fourth fault coefficient; A comprehensive abnormality coefficient calculation module for a power distribution cabinet, used for weighted fusion of the first external corrosion coefficient, the second internal corrosion coefficient, the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient to obtain a comprehensive abnormality coefficient of the power distribution cabinet; The intelligent early warning module of the power distribution cabinet is used to compare the comprehensive abnormality coefficient with the preset early warning threshold, and to issue an early warning if the comprehensive abnormality coefficient is greater than the preset early warning threshold.

2. According to claim 1, the intelligent monitoring and early warning system for power distribution cabinets is characterized in that: The first external image includes an external surface image, a door lock image, a hinge image and an external junction box image of the distribution cabinet acquired by a high-resolution camera; the second internal image includes an internal wiring image, a switch image, a contactor image and a relay image of the distribution cabinet acquired by a high-resolution camera; the first data includes current and voltage data of the distribution cabinet acquired by current sensors and voltage sensors respectively; the second data includes temperature data, humidity data and vibration data of the distribution cabinet acquired by temperature sensors, humidity sensors and vibration sensors respectively.

3. The intelligent monitoring and early warning system for power distribution cabinet according to claim 1 is characterized in that: The distribution cabinet corrosion coefficient calculation module includes: An image preprocessing unit, configured to perform image preprocessing on the first external image and the second internal image to obtain first external image detection data and second internal image detection data; A feature extraction unit, configured to extract features from the first external image detection data and the second internal image detection data by using a pre-trained CNN model to obtain the first external corrosion feature and the second internal corrosion feature; The corrosion coefficient calculation unit is used to analyze the first external corrosion characteristic and the second internal corrosion characteristic to obtain a first external corrosion coefficient and a second internal corrosion coefficient.

4. The intelligent monitoring and early warning system for power distribution cabinet according to claim 3 is characterized in that: The image preprocessing unit performs denoising, contrast enhancement, and edge detection on the first external image and the second internal image respectively to obtain the first external image detection data and the second internal image detection data.

5. The intelligent monitoring and early warning system for power distribution cabinet according to claim 1 is characterized in that: The first external corrosion features include external surface corrosion features, external door lock corrosion features, external hinge corrosion features and external junction box corrosion features, the second internal corrosion features include internal wiring corrosion features, internal switch corrosion features, internal contactor corrosion features and internal relay corrosion features, and the external surface corrosion features, external door lock corrosion features, external hinge corrosion features, external junction box corrosion, internal wiring corrosion features, internal switch corrosion features, internal contactor corrosion features and internal relay corrosion features all include area features, perimeter features, shape features, color features and texture features; The first external corrosion feature is: F ex =[F surface, F lock ,F hinge ,F box ]; F surface,i =[A surface,i ,P surface,i ,S surface,i ,C surface,i ,T surface,i ]; F lock,i =[A lock,i ,P lock,i ,S lock,i ,C lock,i ,T lock,i ]; F hinge,i =[A hinge,i ,P hinge,i ,S hinge,i ,C hinge,i ,T hinge,i ]; F box,i =[A box,i ,P box,i ,S box,i ,C box,i ,T box,i ]; Among them, F ex Represented as the first external corrosion feature; F surface Represents the external surface corrosion characteristics; F lock Represents the corrosion characteristics of external door locks; F hinge Represents the external hinge corrosion feature; F box Represents the corrosion characteristics of the external junction box; A surface,i , P surface,i , S surface,i , C surface,i , T surface,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area on the external surface; A lock,i , P lock,i , S lock,i , C lock,i , T lock,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external door lock; A hinge,i , P hinge,i , S hinge,i , C hinge,i , T hinge,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external hinge; A box,i , P box,i , S box,i , C box,i , T box,i They represent the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the external junction box respectively; The second internal corrosion feature is: F in =[F writing, F switch ,F contactor ,F relay ]; F writing,i =[A writing,i ,P writing,i ,S writing,i ,C writing,i ,T writing,i ]; F switch,i =[A switch,i ,P switch,i ,S switch,i ,C switch,i ,T switch,i ]; F contactor,i =[A contactor,i ,P contactor,i ,S contactor,i ,C contactor,i ,T contactor,i ]; F relay,i =[A relay,i ,P relay,i ,S relay,i ,C relay,i ,T relay,i ]; Among them, F in Represented as the second internal corrosion feature; F writing Indicates the internal wiring corrosion characteristics; F switch Represents the internal switch corrosion characteristics; F contactor Represents the internal contactor corrosion characteristics; F relay Represents the internal relay corrosion characteristics; A writing,i , P writing,i , S writing,i , C writing,i , T writing,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal wiring; A switch,i , P switch,i , S switch,i , C switch,i , T switch,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal switch; A contactor,i , P contactor,i , S contactor,i , C contactor,i , T contactor,i are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal contactor; A relay,i , P relay,i , S relay,i , C relay,i , T relay,i They are respectively represented as the area feature, perimeter feature, shape feature, color feature and texture feature of the i-th corrosion area of ​​the internal relay.

6. The intelligent monitoring and early warning system for power distribution cabinet according to claim 1 is characterized in that: The first external corrosion coefficient is: IN ex α surface *IN surface +α lock *IN lock +α hinge *IN hinge +α box *IN box 4 Among them, CI ex Expressed as the first external corrosion coefficient; CI surface Expressed as external surface corrosion coefficient; CI lock Expressed as the external door lock corrosion coefficient; CI hinge Expressed as the external hinge corrosion coefficient; CI box Expressed as the external junction box corrosion coefficient; α surface Expressed as the external surface corrosion coefficient weight; α lock Expressed as the external door lock corrosion coefficient weight; α hinge Expressed as the external hinge corrosion coefficient weight; α box Expressed as the external junction box corrosion coefficient weight; N surface Expressed as the number of external surface box corrosion areas; N lock Expressed as the number of external door lock corrosion areas; N hinge Expressed as the number of corrosion areas of the external hinge box; N box Expressed as the number of corrosion areas in the external junction box; The second internal corrosion coefficient is: CI in =β writing *CI writing +β switch *CI switch +β contactor *CI contactor +β relay *CI relay ; Among them, CI in Expressed as the second internal corrosion coefficient; CI writing Expressed as the internal wiring corrosion coefficient; CI switch Expressed as the internal switch corrosion coefficient; CI contactor Expressed as the internal contactor corrosion coefficient; CI relay Expressed as the internal relay corrosion coefficient; β writing Expressed as the internal wiring corrosion coefficient weight; β switch Expressed as the internal switch corrosion coefficient weight; β contactor Expressed as the weight of the internal contactor corrosion coefficient; β relay Expressed as the internal relay corrosion coefficient weight; N writing Expressed as the number of internal wiring corrosion areas; N switch Expressed as the number of internal switch corrosion areas; N contactor Expressed as the number of internal contactor corrosion areas; N relay Expressed as the number of internal relay corrosion areas.

7. The intelligent monitoring and early warning system for power distribution cabinet according to claim 1 is characterized in that: The distribution cabinet failure coefficient calculation module includes: A data preprocessing unit, used for performing data preprocessing on the first data and the second data to obtain first detection data and second detection data; the data preprocessing process includes denoising and normalization; A fault feature extraction unit, configured to extract features from the first detection data and the second detection data by using an LSTM network combined with an attention mechanism to obtain the comprehensive fault feature vector; A fault coefficient calculation unit is used to analyze the comprehensive fault feature vector to obtain the first fault coefficient, the second fault coefficient, the third fault coefficient and the fourth fault coefficient.

8. The intelligent monitoring and early warning system for power distribution cabinet according to claim 1 is characterized in that: The comprehensive fault feature vector is: F t =[ΔI t ,ΔV t ,T t ,H t ,X t ,c t ]; Among them, ΔI t represents the first fault characteristic vector, which is used to represent the current fluctuation characteristics; ΔV t represents the second fault feature vector, which is used to represent the voltage fluctuation feature; T t represents the third fault feature vector, which is used to represent the temperature feature; H t It is represented as the fourth fault feature vector, which is used to represent the humidity feature; t represents the weighted feature vector of the attention mechanism, which is used to represent the vibration characteristics of the distribution cabinet; c t Represents the weighted feature vector of the attention mechanism, which is used to enhance important features.

9. The intelligent monitoring and early warning system for power distribution cabinet according to claim 1 is characterized in that: The first fault coefficient is used to describe the poor contact fault coefficient of the power distribution cabinet, and its calculation formula is: Wherein, FI1 represents the first fault coefficient; λ1 represents the first fault coefficient weight; ΔI k is the kth value of the first fault feature vector; ΔV k is the kth value of the second fault feature vector; T k is the kth value of the third fault feature vector; H k is the kth value of the fourth fault feature vector; μ H is the mean of all values ​​in the fourth fault feature vector; μ X is the mean of all values ​​in the attention mechanism weighted feature vector; c k It is represented as the kth value of the weighted feature vector of the attention mechanism; N represents the number of samples; The second fault coefficient is used to describe the short-circuit fault coefficient of the power distribution cabinet, and its calculation formula is: Wherein, FI2 represents the second fault coefficient; λ2 represents the weight of the second fault coefficient; μ T The mean of all values ​​in the third fault feature vector; The third fault coefficient is used to describe the overload fault coefficient of the power distribution cabinet, and its calculation formula is: Wherein, FI3 represents the third fault coefficient; λ3 represents the third fault coefficient weight; The fourth fault coefficient is used to describe the arc discharge fault coefficient of the power distribution cabinet, and its calculation formula is: Among them, FI4 represents the fourth fault coefficient; λ4 represents the fourth fault coefficient weight.

10. The intelligent monitoring and early warning system for power distribution cabinet according to claim 1, characterized in that: The comprehensive anomaly coefficient is: CAI=α1*CI ex +α2*CI in +β1*FI1+β2*FI2+β3*FI3+β4*FI4; Wherein, CAI represents the comprehensive anomaly coefficient; α1 represents the importance of the first external corrosion coefficient; CI ex Expressed as the first external corrosion coefficient; CI in It is represented as the second internal corrosion coefficient; α2 is represented as the importance of the second internal corrosion coefficient; FI1 is represented as the first fault coefficient; β1 is represented as the importance of the first fault coefficient; FI2 is represented as the second fault coefficient; β2 is represented as the importance of the second fault coefficient; FI3 is represented as the third fault coefficient; β3 is represented as the importance of the third fault coefficient; FI4 is represented as the fourth fault coefficient; β4 is represented as the importance of the fourth fault coefficient.

Images