GCB holographic data acquisition and analysis device and method

Abstract

The invention discloses a GCB holographic data acquisition and analysis device and method. The device comprises a power supply module, a data acquisition module, a time sequence synchronization module and a main control module, the power supply module supplies power to the data acquisition module, the time sequence synchronization module and the main control module; the time sequence synchronization module is used for ensuring that the data transmitted from the data module to the main control module is kept in time sequence synchronization; the data acquisition module comprises a stress strain sensor, a current sensor, a stroke sensor, a vibration sensor, a visible light sensor and an SF6 gas sensor; the data acquisition module is used for packing sensor data and sending to the main control module; and the main control module is responsible for processing and storing the transmitted sensor data. According to the GCB state monitoring system, a plurality of physical quantities used for GCB state detection can be collected through standard hardware and a communication interface, the expansibility of the system is improved, the collected data are effectively processed and utilized, and state monitoring of GCB equipment is facilitated.

Description

Technical Field

[0001] The present invention belongs to the technical field of GCB holographic data acquisition and analysis, and in particular relates to a GCB holographic data acquisition and analysis device and method. Background Art

[0002] At present, the demand for safe operation and intelligence of hydropower stations has led to more and more physical quantities that need to be monitored by GCB (generator circuit breaker). How to manage the monitoring data in a unified and standardized manner is particularly important. Due to the large number of monitoring equipment manufacturers, the functions of the monitoring devices are too subdivided, and the degree of software and hardware coupling is high, resulting in a large number of monitoring devices on the project site, high costs, which is not conducive to the promotion of equipment status perception technology and restricts the development of lightweight and miniaturized equipment. In addition, due to the lack of a unified and effective management platform, the various types of data obtained by sensors are simply used, and some data are even simply stored without effective use; at the same time, the system does not explore the connotation of the data during storage, making it difficult to accurately and reasonably judge the status of the equipment. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a GCB holographic data acquisition and analysis device and method, which can acquire a variety of physical quantities used for GCB status detection with standard hardware and communication interfaces, increase the scalability of the system, and effectively process and utilize the acquired data, so as to facilitate the status monitoring of GCB equipment.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A GCB holographic data acquisition and analysis device comprises a power supply module, a data acquisition module, a timing synchronization module and a main control module;

[0006] The power supply module supplies power to the data acquisition module, the timing synchronization module and the main control module;

[0007] The timing synchronization module is used to ensure that the data transmitted from the data module to the main control module maintains timing synchronization;

[0008] The data acquisition module includes a stress strain sensor, a current sensor, a travel sensor, a vibration sensor, a visible light sensor, and an SF6 gas sensor; the data acquisition module is used to package the sensor data and send it to the main control module;

[0009] The main control module is responsible for processing and storing the transmitted sensor data.

[0010] Preferably, the SF6 gas sensor is electrically connected to the gas monitoring module, the visible light sensor is electrically connected to the isolating switch opening and closing monitoring module, the stress strain sensor is electrically connected to the stress strain module, and the current sensor, stroke sensor and vibration sensor are electrically connected to the mechanical characteristic module; the gas monitoring module, the isolating switch opening and closing monitoring module, the stress strain module and the mechanical characteristic module are used to transmit the GCB status data to the main control module.

[0011] Preferably, the disconnector opening and closing monitoring module is used to obtain visible light images, and image preprocessing adopts image graying, histogram correction and filtering noise reduction;

[0012] The stress-strain module is used to draw the received strain information into a strain curve, and to alarm when an abnormal stress value is detected, and to locate the abnormal stress position;

[0013] The mechanical characteristic module is used to obtain signals from current, vibration and stroke sensors.

[0014] An analysis method of a GCB holographic data acquisition and analysis device comprises the following steps:

[0015] Start the GCB holographic data acquisition and analysis device;

[0016] Transmit the data collected by the stress strain sensor, current sensor, stroke sensor, vibration sensor and SF6 gas sensor to the main control module;

[0017] Obtain the status information of the disconnector through image processing and analysis;

[0018] The characteristic parameters of current, vibration and travel sensors are obtained through signal processing and feature extraction;

[0019] Store various sensor data for subsequent analysis and diagnosis;

[0020] The isolating switch status information and characteristic parameters of current, vibration and travel sensors are obtained based on image processing.

[0021] Specifically, the data collected by the stress strain sensor, the current sensor, the travel sensor, the vibration sensor and the SF6 gas sensor are transmitted to the main control module; comprising the following steps:

[0022] Step 2.1: After the connection is successful, the main control module sends the configuration file to the data acquisition module;

[0023] Step 2.2: The data acquisition module verifies the configuration file;

[0024] If verification succeeds: turn on the semaphore;

[0025] If verification fails: reply to the host with an error code and wait for the configuration file to be resent;

[0026] Step 2.3: The data acquisition module starts the timing synchronization module to prepare for sensor data acquisition;

[0027] Step 2.4: The main control module adopts the polling method, and the default polling interval is set to 1 second, and polling is started once;

[0028] Step 2.5: The data acquisition module reads the addresses of each sensor in the configuration file, arranges them according to the query format, sends a query to the sensor and waits for a reply;

[0029] Step 2.6: The data collection module verifies the reply data;

[0030] If the data is normal: package the data and send it to the main control module;

[0031] If the verification fails: re-initiate the query; if multiple queries fail, the abnormal data and abnormal information are packaged and sent to the main control module, and the next sensor data is accessed until all sensor data are collected.

[0032] Specifically, the state information of the isolating switch is obtained by image processing and analysis, comprising the following steps:

[0033] Step 3.1: The disconnector opening and closing monitoring module obtains a visible light image;

[0034] Step 3.2: Image preprocessing:

[0035] Image grayscale: Convert a color image to a grayscale image; the formula is:

[0036] I gray (x,y)=0.299·R(x,y)+0.587·G(x,y)+0.114·B(x,y);

[0037] Among them, I gray (x,y) represents a grayscale image, R(x,y), G(x,y), and B(x,y) represent the pixel values ​​of the red, green, and blue channels, respectively;

[0038] Histogram correction: Apply the histogram equalization algorithm to make the image details clearer. The formula is:

[0039]

[0040] Where S(k) is the new pixel intensity, p(rj) is the pixel intensity, r j The probability of yes, L is the number of gray levels; k represents an index value from 0 to L-1;

[0041] Filtering and denoising: Filter the image to remove noise. The formula is:

[0042]

[0043] Among them, I filtered (x, y) represents the filtered image, N is the size of the filter window; (x, y) represents the position coordinates of the pixel to be processed;

[0044] Step 3.3: Use computer vision technology for image recognition:

[0045] Edge detection: Use the Sobel operator for edge detection. The formula is:

[0046]

[0047] Among them, G x and G y Represent the gradients in the horizontal and vertical directions respectively, and G represents the gradient amplitude;

[0048] Feature extraction: Use convolutional neural network for feature extraction and classification:

[0049] f(x) = σ(W*x+b);

[0050] Among them, W is the weight matrix, b is the bias term, σ is the activation function, and * represents the convolution operation.

[0051] Specifically, the method of obtaining characteristic parameters of current, vibration and travel sensors by signal processing and feature extraction comprises the following steps:

[0052] Step 4.1: Process the data of current, vibration and travel sensors;

[0053] Current signal processing: zero bias processing, median filtering, and smoothing;

[0054] I corrected( t)=I(t)-μ zero ;

[0055] Among them, I corrected (t) is the corrected current signal, I(t) is the original current signal, μ zero is the average value at zero current;

[0056] Vibration signal processing: Extract the characteristics of the vibration signal; the formula is:

[0057]

[0058] Among them, V rmsis the effective value of the vibration signal, T is the length of the integration interval; t represents the period, that is, the time required for the voltage change to complete a complete cycle; dt represents a small time increment, which is a concept in calculus and is used to represent an infinitely subdivided time period; V(t) represents the instantaneous voltage value at time t;

[0059] Travel signal processing: Extract the characteristics of the travel signal; the formula is:

[0060]

[0061] Among them, D Davg is the average value of the travel signal, d i is the travel value of the i-th sampling point, and N is the total number of sampling points;

[0062] Step 4.2: Store the processed signal features and perform fusion analysis.

[0063] Specifically, various types of sensor data are stored for subsequent analysis and diagnosis;

[0064] Step 5.1: The main control module processes and stores the received data;

[0065] Data parsing: parsing received data packets;

[0066] Data storage: store the parsed data in a database or file;

[0067] Step 5.2: Develop lightweight applications to realize real-time display of sensor data, device parameter setting, historical data viewing, and standard curve setting functions;

[0068] Specifically, the isolating switch status information obtained based on image processing, as well as characteristic parameters of current, vibration, and travel sensors, include the following:

[0069] Step 6.1: Study the current, vibration, and travel sensor data during the circuit breaker operation process in the time domain and frequency domain;

[0070] Frequency domain analysis: perform Fourier transform and extract frequency domain features;

[0071]

[0072] Among them, X(f) is the frequency domain signal, x(t) is the time domain signal, and f is the frequency;

[0073] Step 6.2: Analyze the relationship between circuit breaker status and sensor data, and find explicit and invisible features between multiple parameters;

[0074] Feature fusion: fuse the state information obtained from image processing with the sensor feature parameters; the formula is:

[0075] F fusion =W·[F image ,F current ,F υibration ,F traυel ];

[0076] Where W is the weight matrix, F image is the image feature vector, F current is the current eigenvector, F vibration is the vibration eigenvector, F travel is the trip feature vector;

[0077] Step 6.3: Study the coordination relationship between various information, perform various feature engineering on the signal, and extract the characteristic parameters of the signal;

[0078] Step 6.4: Design a suitable algorithm for evaluation and finally realize GCB status monitoring based on multi-dimensional information fusion;

[0079] Fault diagnosis: Use support vector machine (SVM) for fault diagnosis; the formula is:

[0080]

[0081] Among them, w is the weight vector, b is the bias term, C is the penalty factor, and ξi is the slack variable.

[0082] A computer device comprising: one or more processors;

[0083] The processor is used to store one or more programs;

[0084] When the one or more programs are executed by the one or more processors, the analysis method of the GCB holographic data acquisition and analysis device as described in any one of claims 4 to 9 is implemented.

[0085] The present invention can achieve the following beneficial effects:

[0086] 1. The existing products require a large number of devices to achieve holographic data monitoring and have poor scalability. However, the hardware interface of this device is standardized and the software program is APP-based, which facilitates the expansion of the existing system and realizes a breakthrough change from "one state, one set of equipment, one manufacturer, one set of solutions" to standardization, achieving a breakthrough in lightweight and miniaturized equipment.

[0087] 2. The existing device has low data utilization rate and is only used for display or storage. It is not used for circuit breaker status judgment. Some products have the function of status judgment, but the research methods are mainly focused on the detection and analysis of single-source signals, which are susceptible to environmental interference, prone to misjudgment and missed judgment, and difficult to achieve effective identification during actual operation. This device uses multi-source signals for fusion analysis and diagnosis, and selects optimized machine learning algorithms and optimization algorithms in combination with fault characteristics to achieve accurate identification of fault status. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0089] Figure 1 This is a module connection diagram of a GCB holographic data acquisition and analysis device of the present invention;

[0090] Figure 2 This is the wiring diagram of the sensor of the present invention;

[0091] Figure 3 This is a power supply schematic diagram of the power module of the present invention;

[0092] Figure 4 It is the signal transmission diagram of various sensors of the present invention;

[0093] Figure 5 It is the GA-BP genetic neural network grid diagram of the present invention;

[0094] Figure 6 This is a comparison chart showing the error reduction of the present invention compared to the traditional algorithm. DETAILED DESCRIPTION

[0095] Embodiment 1:

[0096] The preferred solution is Figures 1 to 6 As shown, a GCB holographic data acquisition and analysis device includes a power module, a data acquisition module, a timing synchronization module and a main control module;

[0097] The power module is used to power the data acquisition module, the timing synchronization module and the main control module. The power module is used to convert the industrial frequency voltage into the 24V voltage commonly used in the industry to power the device and realize the electromagnetic compatibility design. The structure of the power module is shown in Figure 3As shown in the figure, the control cabinets of various substations are currently equipped with 220VAC power supply, and the voltage specification of the industrial control interface is generally 24VDC. Therefore, an isolated power module needs to be selected as a voltage conversion tool, and it also has a certain electromagnetic protection function. In order to ensure the long-term healthy operation of the power module, a pre-surge protection design is required for the power frequency voltage. Protection devices such as varistors, gas discharge tubes, transient suppression diodes and fuses have different working principles, but they can all play a role in circuit limitation. When the varistor works below the varistor voltage, the current is extremely small, which is equivalent to a resistor with infinite resistance. When the surge comes, the varistor enters the breakdown region and turns into a low-resistance state, the current increases sharply, and the voltage is clamped in a smaller range. However, when the varistor is used for AC input protection, leakage current will be generated due to the influence of parasitic capacitance. The instability of leakage current is the direct cause of accelerated aging and explosion. In view of the above problems, a gas discharge tube is introduced. When the input end is normal, the gas discharge tube separates the varistor from the power input end to prevent the increase of the varistor leakage current and slow down the decline of the varistor performance. In order to ensure that the added gas discharge tube does not affect the normal operation of the line, the DC breakdown voltage of the discharge tube must be higher than the maximum normal operating voltage of the line. Finally, in order to achieve a higher immunity to electrical fast transient pulse groups, an AC single-phase three-section EMI power filter is introduced to enable the monitoring system to operate normally in a complex electromagnetic interference environment.

[0098] The timing synchronization module is used to ensure that the data transmitted from the data module to the main control module maintains timing synchronization; the timing synchronization module in the data acquisition module includes software synchronization and hardware synchronization.

[0099] Software synchronization:

[0100] Based on the MODBUS communication protocol, the embedded software of the data acquisition module synchronizes the time of each lower computer (sensor) through broadcasting at regular intervals. The lower computer modifies the RTC time to the broadcast time and continues to run the RTC. Since the RTC running time is only related to the crystal oscillator and has nothing to do with the program running priority, higher accuracy can be achieved in timing. The timer stores the data information collected by the sensor into the ring storage area at a certain sampling rate. When the value of the nth data in the ring storage area exceeds the set threshold, the sensor records n and the RTC time tp at this moment, because in software synchronization, the timestamp is performed based on the sampling value exceeding the threshold. Due to the performance differences of each sensor, this will cause the RTC time to deviate to a certain extent, thereby affecting the accuracy of the timestamp. Therefore, each sensor can be calibrated for accuracy. After calibration, the time error of soft synchronization can be controlled within 1ms.

[0101] Hardware Sync:

[0102] like Figure 2As shown in the figure, synchronous acquisition is achieved by monitoring analog level signals. All lower computers are connected to the same synchronous bus through optocouplers. When a device detects the action of the circuit breaker, it pulls up the output optocoupler level. Other devices start sampling immediately after detecting the increase in voltage at the input optocoupler, thus achieving synchronous data acquisition.

[0103] The data acquisition module includes low-speed sensors such as stress and strain sensors, current sensors, travel sensors, vibration sensors, visible light sensors, SF6 gas sensors, and high-speed sensors such as visible light image sensors. The low-speed sensors transmit data through the RS-485 serial port using the Modbus protocol, and the high-speed sensors transmit data through the RJ45 network port using the TCP protocol. The data acquisition module also includes a data acquisition board that uses an ARM+FPGA dual-core architecture, in which the FPGA part is responsible for establishing multiple parallel serial ports to collect sensor data, and the ARM part is responsible for establishing TCP communication with the main control, and packaging the collected sensor data to send to the main control module.

[0104] The main control module is responsible for processing and storing the transmitted data.

[0105] The main control module is responsible for processing and storing the transmitted sensor data.

[0106] Furthermore, the SF6 gas sensor is electrically connected to the gas monitoring module, the visible light sensor is electrically connected to the disconnector opening and closing monitoring module, the stress strain sensor is electrically connected to the stress strain module, and the current sensor, stroke sensor and vibration sensor are electrically connected to the mechanical characteristic module; the gas monitoring module, the disconnector opening and closing monitoring module, the stress strain module and the mechanical characteristic module are used to transmit the GCB status data to the main control module.

[0107] Furthermore, the disconnector opening and closing monitoring module is used to obtain visible light images, and image preprocessing adopts image graying, histogram correction and filtering noise reduction;

[0108] The stress-strain module is used to draw the received strain information into a strain curve, and to alarm when an abnormal stress value is detected, and to locate the abnormal stress position;

[0109] The mechanical characteristic module is used to obtain signals from current, vibration and stroke sensors.

[0110] The working principle of this device is:

[0111] Start the device, the programs on the data acquisition module and the main control module start automatically, the data acquisition module establishes a TCP client, the main control module reads the configuration file and establishes the client and server, and then the two establish a connection; after the connection is successful, the main control module transmits the configuration file to the data acquisition module. The configuration file implementation includes the settings of the sensor number, interface bus number, baud rate, custom polling cycle and sensor function attributes of the sensor in the data acquisition module, as well as the settings of the IP address, port number and sensor function attributes of each sensor transmitted using the network port; the data acquisition module verifies the transmitted configuration file, and turns on the semaphore after the verification is successful, otherwise it replies to the main control end with an error code and waits for the configuration file to be sent again ; The data acquisition module starts the startup timing synchronization module to prepare for sensor data collection, and configures the relevant information of each sensor according to the configuration file; the main control module adopts a polling method, and the default polling interval is set to 1s. The data acquisition module reads the sensor addresses in the configuration file, arranges them according to the query format, sends queries to the sensors and waits for replies; the data acquisition module verifies the reply data, and if it is normal, the data transmission is packaged and sent to the main control module; if the verification fails, the query is re-initiated; if multiple queries fail, the abnormal data and abnormal information are packaged and sent to the main control board, and then the next sensor data is accessed according to the configuration file until all sensor data are collected.

[0112] like Figure 3 As shown in the figure, after the data of each sensor is transmitted to the main control module, the data is divided into the following four modules: gas monitoring module, disconnector opening and closing monitoring module, stress strain module, and mechanical characteristic module, which together characterize the state of the circuit breaker.

[0113] The gas detection module mainly obtains information such as SF6 gas pressure, moisture, and composition;

[0114] The disconnector opening and closing monitoring module mainly obtains visible light images. Image preprocessing is planned to adopt three steps: image graying, histogram correction and filtering noise reduction. Among them, image graying is used to convert color photos into black and white photos, which is convenient for high-speed data transmission. Histogram correction can make the grayscale spacing of the corrected image wider or the grayscale distribution of the image uniform, thereby increasing the contrast and making the image details clear, thereby achieving the purpose of image enhancement. Finally, the photo is filtered and denoised to make the picture clearer. After that, computer vision technology is used for image recognition and image analysis, and useful areas are selected from the preprocessed image, and automatic analysis and recognition are performed to obtain the status information of the disconnector.

[0115] The stress-strain module plots the received strain information into a strain curve, and issues an alarm when an abnormal stress value is detected, locating the abnormal stress position. In combination with the information from other sensors, the abnormal stress can also be analyzed to diagnose the fault.

[0116] The mechanical characteristic module mainly obtains the signals of current, vibration and stroke sensors for subsequent multi-source signal fusion diagnosis. It is necessary to match the three when reading data. First, for the file names in the read folder, the valid files of the two are filtered out according to the identification codes of current and stroke. Finally, for all files of the two, time matching of the same trigger is required. Given a certain time tolerance, the data is then paired within the error tolerance range. In addition, there is a zero drift current problem in the coil current, which is eliminated by zero offset processing. For the zero offset processing of the coil current, first connect the sensor to the power supply, measure the amplitude at zero current, calculate the average value, and use it as the basic value. Subtract this value from the subsequent measured data to complete the zero offset processing. For the problem that the coil current has many burrs and is prone to small spikes, first use median filtering to cut off the small spikes, and then use sliding average filtering to smooth the curve.

[0117] After data acquisition and processing, the data acquisition program corresponding to each sensor is turned into an APP under the embedded Linux operating system framework of the main control module, realizing lightweight application, facilitating the loading, deletion and upgrading of each sensor program, and facilitating the expansion of this system. Each sensor APP includes real-time display of sensor data, device parameter setting, historical data viewing, and standard curve setting.

[0118] Finally, the current, vibration, and travel sensor data during the circuit breaker operation process are studied in the time domain and frequency domain, the relationship between the circuit breaker status and the sensor data is analyzed, the explicit and invisible characteristics between multiple parameters are found, and the coordination relationship between various information is studied; at the same time, multiple feature engineering is performed on the signal, the characteristic parameters of the signal are extracted, and a suitable algorithm is designed for evaluation, ultimately realizing GCB status monitoring based on multi-dimensional information fusion.

[0119] Specifically, for the current signal, its peak and valley values ​​are extracted as features. For the vibration signal, the short-time energy method and EMD decomposition are used to extract the peak and time of the short-time energy, and the energy of the EMD decomposition signal is used as features. For the travel signal, the inherent opening (closing) time, contact travel, and opening (closing) speed are calculated according to the curve as features. The GA-BP genetic neural network algorithm is used for modeling, and finally the state of the circuit breaker is output. The specific process is as follows: Figure 5 shown.

[0120] This embodiment improves the accuracy by integrating multi-source features. Table 1 is a statistical table of various failure rates of current features, vibration features, stroke features, and multi-source fusion features.

[0121] Table 1 Failure rate statistics of current characteristics, vibration characteristics, travel characteristics, and multi-source fusion characteristics

[0122]

[0123] After the implementation of this embodiment, the accuracy improvement effect is as follows: Figure 6 shown.

[0124] Embodiment 2:

[0125] An analysis method of a GCB holographic data acquisition and analysis device comprises the following steps:

[0126] Step 1, start the GCB holographic data acquisition and analysis device;

[0127] Step 1.1: The power module starts up to supply power to the data acquisition module, timing synchronization module and main control module.

[0128] Step 1.2: The programs on the data acquisition module and the main control module start automatically.

[0129] Step 1.3: The data acquisition module establishes a TCP client, and the main control module reads the configuration file and establishes a connection between the client and the server.

[0130] Step 2: Transmit the data collected by the stress strain sensor, current sensor, stroke sensor, vibration sensor and SF6 gas sensor to the main control module;

[0131] Step 2.1: After the connection is successful, the main control module sends the configuration file to the data acquisition module;

[0132] Step 2.2: The data acquisition module verifies the configuration file;

[0133] If verification succeeds: turn on the semaphore;

[0134] If verification fails: reply to the host with an error code and wait for the configuration file to be resent;

[0135] Step 2.3: The data acquisition module starts the timing synchronization module to prepare for sensor data acquisition;

[0136] Step 2.4: The main control module adopts the polling method, and the default polling interval is set to 1 second, and polling is started once;

[0137] Step 2.5: The data acquisition module reads the addresses of each sensor in the configuration file, arranges them according to the query format, sends a query to the sensor and waits for a reply;

[0138] Step 2.6: The data collection module verifies the reply data;

[0139] If the data is normal: package the data and send it to the main control module;

[0140] If the verification fails: re-initiate the query; if multiple queries fail, the abnormal data and abnormal information are packaged and sent to the main control module, and the next sensor data is accessed until all sensor data are collected.

[0141] Step 3: Obtain the status information of the disconnector through image processing and analysis;

[0142] Step 3.1: The disconnector opening and closing monitoring module obtains a visible light image;

[0143] Step 3.2: Image preprocessing:

[0144] Image grayscale: Convert a color image to a grayscale image; the formula is:

[0145] I gray (x,y)=0.299·R(x,y)+0.587·G(x,y)+0.114·B(x,y);

[0146] Among them, I gray (x,y) represents a grayscale image, R(x,y), G(x,y), and B(x,y) represent the pixel values ​​of the red, green, and blue channels, respectively;

[0147] Histogram correction: Apply the histogram equalization algorithm to make the image details clearer. The formula is:

[0148]

[0149] Where S(k) is the new pixel intensity, p(rj) is the pixel intensity, r j The probability of yes, L is the number of gray levels; k represents an index value from 0 to L-1;

[0150] Filtering and denoising: Filter the image to remove noise. The formula is:

[0151]

[0152] Among them, I filtered (x, y) represents the filtered image, N is the size of the filter window; (x, y) represents the position coordinates of the pixel to be processed;

[0153] Step 3.3: Use computer vision technology for image recognition:

[0154] Edge detection: Use the Sobel operator for edge detection. The formula is:

[0155]

[0156] Among them, Gx and Gy represent the gradients in the horizontal and vertical directions respectively, and G represents the gradient amplitude;

[0157] Feature extraction: Use convolutional neural network for feature extraction and classification:

[0158] f(x) = σ(W*x+b);

[0159] Among them, W is the weight matrix, b is the bias term, σ is the activation function, and * represents the convolution operation.

[0160] Step 4: Obtain characteristic parameters of current, vibration and travel sensors through signal processing and feature extraction;

[0161] Step 4.1: Process the data of current, vibration and travel sensors;

[0162] Current signal processing: zero bias processing, median filtering, and smoothing;

[0163] I corrected( t)=I(t)-μ zero ;

[0164] Among them, I corrected (t) is the corrected current signal, I(t) is the original current signal, μ zero is the average value at zero current;

[0165] Vibration signal processing: Extract the characteristics of the vibration signal; the formula is:

[0166]

[0167] Among them, V rms is the effective value of the vibration signal, T is the length of the integration interval; t represents the period, that is, the time required for the voltage change to complete a complete cycle; dt represents a small time increment, which is a concept in calculus and is used to represent an infinitely subdivided time period; V(t) represents the instantaneous voltage value at time t;

[0168] Travel signal processing: Extract the characteristics of the travel signal; the formula is:

[0169]

[0170] Among them, D Davg is the average value of the travel signal, d i is the travel value of the i-th sampling point, and N is the total number of sampling points;

[0171] Step 4.2: Store the processed signal features and perform fusion analysis.

[0172] Step 5: Store various sensor data for subsequent analysis and diagnosis;

[0173] Step 5.1: The main control module processes and stores the received data;

[0174] Data parsing: parsing received data packets;

[0175] Data storage: store the parsed data in a database or file;

[0176] Step 5.2: Develop lightweight applications to realize real-time display of sensor data, device parameter setting, historical data viewing, and standard curve setting functions;

[0177] Step 6: The isolating switch status information and characteristic parameters of the current, vibration and travel sensors are obtained based on image processing.

[0178] Step 6.1: Study the current, vibration, and travel sensor data during the circuit breaker operation process in the time domain and frequency domain;

[0179] Frequency domain analysis: perform Fourier transform and extract frequency domain features;

[0180]

[0181] Among them, X(f) is the frequency domain signal, x(t) is the time domain signal, and f is the frequency;

[0182] Step 6.2: Analyze the relationship between circuit breaker status and sensor data, and find explicit and invisible features between multiple parameters;

[0183] Feature fusion: fuse the state information obtained from image processing with the sensor feature parameters; the formula is:

[0184] F fusion =W·[F image , F current , F υibration , F traυel ];

[0185] Where W is the weight matrix, F image is the image feature vector, F current is the current eigenvector, F vibration is the vibration eigenvector, F travel is the trip feature vector;

[0186] Step 6.3: Study the coordination relationship between various information, perform various feature engineering on the signal, and extract the characteristic parameters of the signal;

[0187] Step 6.4: Design a suitable algorithm for evaluation and finally realize GCB status monitoring based on multi-dimensional information fusion;

[0188] Fault diagnosis: Use support vector machine (SVM) for fault diagnosis; the formula is:

[0189]

[0190] Among them, w is the weight vector, b is the bias term, C is the penalty factor, and ξi is the slack variable.

[0191] Embodiment 3:

[0192] A computer device comprising: one or more processors;

[0193] The processor is used to store one or more programs;

[0194] When the one or more programs are executed by the one or more processors, the analysis method of the GCB holographic data acquisition and analysis device as described in any one of claims 4 to 9 is implemented.

[0195] The above embodiments are only preferred technical solutions of the present invention and should not be regarded as limiting the present invention. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A GCB holographic data acquisition and analysis device, characterized in that: It includes power supply module, data acquisition module, timing synchronization module and main control module; The power supply module supplies power to the data acquisition module, the timing synchronization module and the main control module; The timing synchronization module is used to ensure that the data transmitted from the data module to the main control module maintains timing synchronization; The data acquisition module includes a stress strain sensor, a current sensor, a travel sensor, a vibration sensor, a visible light sensor, and an SF6 gas sensor; the data acquisition module is used to package the sensor data and send it to the main control module; The main control module is responsible for processing and storing the transmitted sensor data.

2. A GCB holographic data acquisition and analysis device according to claim 1, characterized in that: The SF6 gas sensor is electrically connected to the gas monitoring module, the visible light sensor is electrically connected to the disconnector opening and closing monitoring module, the stress strain sensor is electrically connected to the stress strain module, and the current sensor, stroke sensor and vibration sensor are electrically connected to the mechanical characteristic module; the gas monitoring module, the disconnector opening and closing monitoring module, the stress strain module and the mechanical characteristic module are used to transmit the GCB status data to the main control module.

3. A GCB holographic data acquisition and analysis device according to claim 2, characterized in that: The disconnector opening and closing monitoring module is used to obtain visible light images, and image preprocessing uses image graying, histogram correction and filtering noise reduction; The stress-strain module is used to draw the received strain information into a strain curve, and to alarm when an abnormal stress value is detected, and to locate the abnormal stress position; The mechanical characteristic module is used to obtain signals from current, vibration and stroke sensors.

4. The analysis method of the GCB holographic data acquisition and analysis device according to any one of claims 1 to 3, characterized in that The following steps are involved: Start the GCB holographic data acquisition and analysis device; Transmit the data collected by the stress strain sensor, current sensor, stroke sensor, vibration sensor and SF6 gas sensor to the main control module; Obtain the status information of the disconnector through image processing and analysis; The characteristic parameters of current, vibration and travel sensors are obtained through signal processing and feature extraction; Store various sensor data for subsequent analysis and diagnosis; The isolating switch status information and characteristic parameters of current, vibration and travel sensors are obtained by image processing.

5. The analysis method of the GCB holographic data acquisition and analysis device according to claim 4, characterized in that: The method of transmitting the data collected by the stress strain sensor, the current sensor, the travel sensor, the vibration sensor and the SF6 gas sensor to the main control module comprises the following steps: Step 2.1: After the connection is successful, the main control module sends the configuration file to the data acquisition module; Step 2.2: The data acquisition module verifies the configuration file; If verification succeeds: turn on the semaphore; If verification fails: reply to the host with an error code and wait for the configuration file to be resent; Step 2.3: The data acquisition module starts the timing synchronization module to prepare for sensor data acquisition; Step 2.4: The main control module adopts the polling method, and the default polling interval is set to 1 second, and polling is started once; Step 2.5: The data acquisition module reads the addresses of each sensor in the configuration file, arranges them according to the query format, sends a query to the sensor and waits for a reply; Step 2.6: The data collection module verifies the reply data; If the data is normal: package the data and send it to the main control module; If the verification fails: re-initiate the query; if multiple queries fail, the abnormal data and abnormal information are packaged and sent to the main control module, and the next sensor data is accessed until all sensor data are collected.

6. The analysis method of the GCB holographic data acquisition and analysis device according to claim 4, characterized in that: The state information of the isolating switch is obtained by image processing and analysis, comprising the following steps: Step 3.1: The disconnector opening and closing monitoring module obtains a visible light image; Step 3.2: Image preprocessing: Image grayscale: Convert a color image to a grayscale image; the formula is: I gray (x,y)=0.299·R(x,y)+0.587·G(x,y)+0.114·B(x,y); Among them, I gray (x,y) represents a grayscale image, R(x,y), G(x,y), and B(x,y) represent the pixel values ​​of the red, green, and blue channels, respectively; Histogram correction: Apply the histogram equalization algorithm to make the image details clearer. The formula is: Where S(k) is the new pixel intensity, p(rj) is the pixel intensity, r j The probability of yes, L is the number of gray levels; k represents an index value from 0 to L-1; Filtering and denoising: Filter the image to remove noise. The formula is: Among them, I filtered (x, y) represents the filtered image, N is the size of the filter window; (x, y) represents the position coordinates of the pixel to be processed; Step 3.3: Use computer vision technology for image recognition: Edge detection: Use the Sobel operator for edge detection. The formula is: Among them, G x and G y Represent the gradients in the horizontal and vertical directions respectively, and G represents the gradient amplitude; Feature extraction: Use convolutional neural network for feature extraction and classification: f(x) = σ(W*x+b); Among them, W is the weight matrix, b is the bias term, σ is the activation function, and * represents the convolution operation.

7. The analysis method of the GCB holographic data acquisition and analysis device according to claim 4, characterized in that: The method of obtaining characteristic parameters of current, vibration and travel sensors by signal processing and feature extraction comprises the following steps: Step 4.1: Process the data of current, vibration and travel sensors; Current signal processing: zero bias processing, median filtering, and smoothing; I corrected( t)=I(t)-μ zero ; Among them, I corrected (t) is the corrected current signal, I(t) is the original current signal, μ zero is the average value at zero current; Vibration signal processing: Extract the characteristics of the vibration signal; the formula is: Among them, V rms is the effective value of the vibration signal, T is the length of the integration interval; t represents the period, that is, the time required for the voltage change to complete a complete cycle; dt represents a small time increment, which is a concept in calculus and is used to represent an infinitely subdivided time period; V(t) represents the instantaneous voltage value at time t; Travel signal processing: Extract the characteristics of the travel signal; the formula is: Among them, D Davg is the average value of the travel signal, d i is the travel value of the i-th sampling point, and N is the total number of sampling points; Step 4.2: Store the processed signal features and perform fusion analysis.

8. The analysis method of the GCB holographic data acquisition and analysis device according to claim 4, characterized in that: Store various sensor data for subsequent analysis and diagnosis; Step 5.1: The main control module processes and stores the received data; Data parsing: parsing received data packets; Data storage: store the parsed data in a database or file; Step 5.2: Develop lightweight applications to realize real-time display of sensor data, device parameter setting, historical data viewing, and standard curve setting functions.

9. The analysis method of the GCB holographic data acquisition and analysis device according to claim 4, characterized in that: The isolating switch status information obtained based on image processing, as well as characteristic parameters of current, vibration, and travel sensors, include the following: Step 6.1: Study the current, vibration, and travel sensor data during the circuit breaker operation process in the time domain and frequency domain; Frequency domain analysis: perform Fourier transform and extract frequency domain features; Among them, X(f) is the frequency domain signal, x(t) is the time domain signal, and f is the frequency; Step 6.2: Analyze the relationship between circuit breaker status and sensor data, and find explicit and invisible features between multiple parameters; Feature fusion: fuse the state information obtained from image processing with the sensor feature parameters; the formula is: F fusion =W·[F image ,F current ,F υibration ,F travel ]; Where W is the weight matrix, F image is the image feature vector, F current is the current eigenvector, F vibration is the vibration eigenvector, F travel is the trip feature vector; Step 6.3: Study the coordination relationship between various information, perform various feature engineering on the signal, and extract the characteristic parameters of the signal; Step 6.4: Design a suitable algorithm for evaluation and finally realize GCB status monitoring based on multi-dimensional information fusion; Fault diagnosis: Use support vector machine (SVM) for fault diagnosis; the formula is: Among them, w is the weight vector, b is the bias term, C is the penalty factor, and ξi is the slack variable.

10. A computer device comprising: one or more processors; The processor is used to store one or more programs; When the one or more programs are executed by the one or more processors, the analysis method of the GCB holographic data acquisition and analysis device as described in any one of claims 4 to 9 is implemented.

Images