A fire protection early warning method and system for a converter station

Abstract

The application discloses a fire protection early warning method and system for a converter station, and belongs to the field of safety protection. The method is as follows: real-time current data is acquired, and Fourier transform operation is performed on the real-time current data to obtain a real-time current waveform; an abnormal current waveform is extracted from the real-time current waveform; waveform parameters of the abnormal current waveform are determined, and a corresponding probability grade value is determined according to the waveform parameters by using a preset support vector machine; when the probability grade value is greater than a preset probability grade value, a plurality of potential fault points are searched according to the size of the waveform parameters, and corresponding safety early warning processing is triggered according to real-time images and temperature time sequence parameters of the potential fault points. The application can accurately determine a fire grade value and trigger an alarm by simultaneously monitoring equipment in the converter station through current data, pressure data, temperature and real-time images, can improve the evaluation accuracy of safety accidents, and can trigger different alarms for different fire grade values, improve the timeliness of early warning and shorten the alarm delay.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal alarm, and in particular to a fire protection early warning method and system for a converter station. BACKGROUND

[0002] With the continuous expansion of the power grid scale and the continuous improvement of the voltage level of high-voltage direct current transmission, the number and capacity of power equipment in the converter station are increasing, and the safety of the converter station as the core hub of the power system is directly related to the safe and stable operation of the power grid.

[0003] Since the equipment in the converter station is highly dense, the operating environment is high temperature and high pressure, and when the signals of any device are abnormal, the signals are prone to generate strong pressure waves under the current coupling effect and ignite the surrounding flammable insulating materials in the high temperature area, causing chain explosion and fire accidents. Therefore, it is necessary to early warn the converter station of fire, and one of the commonly used fire protection early warning methods is to set multiple different detection terminals (including temperature sensors, electrical parameter sensors) in the converter station, monitor the real-time voltage, current and temperature and other parameters of the equipment, and compare the monitored parameters with the corresponding preset threshold values, to make fire judgment and trigger the early warning according to the comparison results.

[0004] However, the current common method has the following technical problems: relying on the comparison of a single parameter and a threshold value, which often triggers the early warning when the numerical deviation is large, and the fire has occurred when the numerical deviation is large, so the early warning method of the prior art is delayed and has poor timeliness. SUMMARY

[0005] The present application provides a fire protection early warning method and system for a converter station, which can solve the technical problem of the prior art that the early warning is delayed and has poor timeliness using the comparison result of a single parameter and a threshold value.

[0006] The first aspect of the embodiment of the present application provides a fire protection early warning method for a converter station, the method comprising:

[0007] obtaining real-time current data and performing Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extracting an abnormal current waveform from the real-time current waveform, wherein the real-time current data is real-time data of equipment in a valve hall of the converter station;

[0008] determining a waveform parameter of the abnormal current waveform, and determining a corresponding probability level value according to the waveform parameter by using a preset support vector machine, wherein the probability level value is a fault level corresponding to the probability of arc caused by equipment overload;

[0009] When the probability level value is greater than a preset probability level value, a plurality of potential fault points are searched according to the waveform parameter size, and a pressure wave superposition characteristic value from the plurality of potential fault points is obtained;

[0010] A position coordinate of each pressure wave superposition characteristic value is determined, a temperature time sequence parameter and a real-time image of each position coordinate are obtained, a fire level value corresponding to each position coordinate is determined according to the real-time image and the temperature time sequence parameter, and when any one of the fire level values is greater than a preset fault level value, a corresponding safety warning process is triggered according to the fire level value.

[0011] In combination with the first aspect, in an implementation manner, the determination of the fire level value corresponding to each position coordinate according to the real-time image and the temperature time sequence parameter comprises:

[0012] A preset first image recognition model is called to recognize an object quantity in the real-time image and an interval distance between each object and the device, and an object quantity value and an interval distance value are obtained respectively;

[0013] A quantity weight value is calculated according to the object quantity value and the interval distance value;

[0014] A temperature change value of the temperature time sequence parameter is calculated, and the temperature change value and the quantity weight value are used to determine the fire level value.

[0015] In combination with the first aspect, in an implementation manner, the determination of the fire level value corresponding to each position coordinate according to the real-time image and the temperature time sequence parameter comprises:

[0016] A preset second image recognition model is called to perform material recognition on each object in each real-time image, and a plurality of material information is obtained;

[0017] A burning point temperature value corresponding to each material is searched based on each material information, a K-means algorithm is called to cluster a plurality of burning point temperature values, and a plurality of burning point categories are obtained;

[0018] A temperature quantity value is obtained by counting the number of burning point temperature values of each burning point category, and a temperature peak value of the temperature time sequence parameter is calculated;

[0019] The temperature quantity value and the temperature peak value are used to determine the fire level value.

[0020] In combination with the first aspect, in an implementation manner, the waveform parameter comprises a frequency value and an amplitude value;

[0021] The determination of the corresponding probability level value according to the waveform parameter by using a preset support vector machine comprises:

[0022] The amplitude change rate is calculated according to the amplitude value, and the frequency value, the amplitude value and the amplitude change rate are converted into a vector to obtain a parameter feature vector;

[0023] The preset support vector machine is called to perform classification processing according to the parameter feature vector to obtain an arc probability value;

[0024] The target interval value is determined from a plurality of preset interval values according to the arc probability value, and a probability grade value corresponding to the target interval value is determined.

[0025] In combination with the first aspect, in an implementation manner, the step of searching for a plurality of potential fault points according to the waveform parameter size and acquiring pressure wave superposition characteristic values from the plurality of potential fault points comprises:

[0026] The position coordinates of the equipment corresponding to each abnormal current waveform are determined to obtain a plurality of potential fault points;

[0027] Real-time pressure waveforms are acquired based on the pressure sensors of the potential fault points, and a wave peak superposition degree value, a wave trough offset value and an energy distribution value are extracted from each real-time pressure waveform;

[0028] The pressure combination values are obtained by combining the wave peak superposition degree value, the wave trough offset value and the energy distribution value, and a plurality of pressure wave superposition characteristic values are screened from a plurality of pressure combination values, the pressure wave superposition characteristic value being a pressure combination value with a value greater than a preset pressure value.

[0029] In combination with the first aspect, in an implementation manner, the step of acquiring real-time current data and performing Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extracting an abnormal current waveform from the real-time current waveform comprises:

[0030] A plurality of real-time current data are acquired, and each real-time current data is preprocessed to obtain a plurality of processing data, and each processing data is converted into a real-time current waveform by fast Fourier transform, wherein the preprocessing comprises noise reduction processing and time domain conversion processing;

[0031] The total harmonic distortion rate of each real-time current waveform is calculated, and an abnormal current waveform is determined based on the total harmonic distortion rate from a plurality of real-time current waveforms.

[0032] In combination with the first aspect, in an implementation manner, after the step of acquiring the temperature time sequence parameter of each position coordinate and the real-time image, the method further comprises:

[0033] If the temperature time sequence parameter meets a threshold value or the real-time image contains a heat source is identified by calling a preset BP model, a heat source coordinate is determined, and an isolation boundary region is constructed with the heat source coordinate as the center.

[0034] Turning off the equipment in the isolation boundary area, and starting emergency alarm processing.

[0035] The second aspect of the embodiment of the present application provides a fire early warning system of a converter station, the system comprising:

[0036] An acquisition module is configured to acquire real-time current data, perform Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extract an abnormal current waveform from the real-time current waveform, wherein the real-time current data is real-time data of equipment in a valve hall of the converter station;

[0037] A determination module is configured to determine a waveform parameter of the abnormal current waveform, and determine a corresponding probability level value according to the waveform parameter by using a preset support vector machine, wherein the probability level value corresponds to a fault level of a probability of arc caused by equipment overload;

[0038] An extraction module is configured to, when the probability level value is greater than a preset probability level value, search for a plurality of potential fault points according to the waveform parameter size, and acquire pressure wave superposition characteristic values from the plurality of potential fault points.

[0039] An early warning module is configured to determine a position coordinate of each of the pressure wave superposition characteristic values, acquire a temperature time sequence parameter and a real-time image of each of the position coordinates, determine a fire level value corresponding to each of the position coordinates according to the real-time image and the temperature time sequence parameter, and trigger a corresponding safety early warning process according to the fire level value when any one of the fire level values is greater than a preset fault level value.

[0040] Compared with the prior art, the fire protection early warning method and system of the converter station provided by the embodiment of the present application has the beneficial effects that: the present application can obtain real-time current data and perform Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extract an abnormal current waveform from the real-time current waveform; determine the waveform parameters of the abnormal current waveform, and determine the corresponding probability level value according to the waveform parameters by using a preset support vector machine; when the probability level value is greater than a preset probability level value, a plurality of potential fault points are searched according to the size of the waveform parameters, and the pressure wave superposition characteristic values from the plurality of potential fault points are obtained; the position coordinates of each pressure wave superposition characteristic value are determined, and the temperature time sequence parameters and real-time images of each position coordinate are obtained, the fire grade value corresponding to each position coordinate is determined according to the real-time image and the temperature time sequence parameters, and when any one fire grade value is greater than a preset fault grade value, a safety early warning process is triggered. The present application can accurately determine the fire grade value and trigger the alarm by monitoring the equipment in the converter station through current data, pressure data, temperature and real-time images, which not only can improve the evaluation accuracy of safety accidents, but also can trigger different alarms for different fire grade values, so that the technical personnel can be reminded to repair and overhaul in a timely manner, improve the timeliness of early warning, shorten the alarm delay, so that the technical personnel can maintain in a timely manner, further reduce the risk of equipment, and protect the life safety of technical personnel. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a flowchart of a fire protection early warning method of a converter station provided by an embodiment of the present application;

[0042] Figure 2 is a structural schematic diagram of a fire early warning system of a converter station provided by an embodiment of the present application. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0044] With the rapid development of high-voltage technology and the continuous expansion of the power grid scale, the converter station bears higher and higher voltage levels and power densities. One common converter station is an extra-high voltage station. The safety of the extra-high voltage station directly relates to the safe and stable operation of the power grid as the core hub of the power system.

[0045] Because the equipment in the converter station is highly dense, not only the running environment is high temperature and high pressure, but also when the signal of any equipment is abnormal, the signal is easy to produce arc under the current coupling effect, and in the arc expansion stage, the high temperature and high pressure plasma channel is formed due to the rapid injection of external energy, the gas-liquid two-phase interface drives the liquid medium to form a shock wave due to the pressure difference, and the shock waves of different speeds are superimposed to form a primary pressure wave. The strong pressure wave and arc ignite the surrounding flammable insulating materials in the high temperature area, causing a chain explosion and fire accident. For example, when a device has an overload fault, the current coupling effect will cause the nonlinear propagation characteristics of the fault signal between devices, so that the local overload of a single device quickly evolves into a systematic fault involving multiple key nodes. This fault diffusion process will produce a strong pressure wave, and when the pressure wave is refracted and reflected between various components in the valve hall, a multi-point impact superposition effect is formed, further exacerbating the physical damage of the equipment and the destruction of the insulating materials. If the high temperature area generated in the fault process easily ignites the surrounding flammable insulating materials, the fire will show a rapid spreading trend.

[0046] Therefore, it is necessary to early warn the converter station of fire. One commonly used fire protection warning method is to set multiple different detection terminals (including temperature sensors, electrical parameter sensors) in the converter station, monitor the real-time voltage, current and temperature of the equipment, and compare the monitored parameters with the corresponding preset threshold values, to make a fire judgment and trigger a warning according to the comparison results.

[0047] However, the commonly used method has the following technical problems: relying on the comparison of a single parameter and a threshold value, often triggering a warning when the numerical deviation is large, and the numerical deviation is large when the fire has already occurred. Therefore, the warning method of the prior art is delayed and has poor timeliness.

[0048] In order to solve the above problems, the following will introduce and explain in detail a fire protection warning method and system for a converter station provided by the embodiments of the present application through the following specific embodiments.

[0049] In order to solve the technical problem that the prior art uses the comparison result of a single parameter and a threshold value for warning and has delay and poor timeliness, referring to Figure 1 , a flowchart of a fire protection warning method for a converter station provided by an embodiment of the present application is shown.

[0050] In an embodiment, the fire protection early warning method of the converter station can be applied to the central control platform of the converter station, and a plurality of sensing terminals (such as temperature sensors, pressure sensors, cameras, and other different detection instruments) can be arranged in the converter station. The central control platform can be in communication connection with each sensing terminal, collect different data through the sensing terminals, analyze the fire or fault according to the data collected by the sensing terminals, and trigger an early warning or an alarm immediately when there is a fire or fault risk, so as to improve the timeliness of the early warning, early warning, and avoid causing major accidents.

[0051] As an example, the fire protection early warning method of the converter station can include:

[0052] S11, acquiring real-time current data and performing Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extracting an abnormal current waveform from the real-time current waveform, wherein the real-time current data is real-time data of equipment in a valve hall of the converter station.

[0053] In an embodiment, one of the reasons why the converter station is prone to cause a fire accident is that an arc is generated due to equipment failure, and the high temperature of the arc ignites surrounding objects to cause a fire. Therefore, real-time current data of equipment in the valve hall of the converter station can be acquired, and it can be determined whether there is equipment failure or whether there is a faulty equipment according to the real-time current data. For example, it can be determined whether there is an abnormal value in the value of the real-time current data, and if there is, it can be determined that there is a faulty equipment.

[0054] If there is a faulty equipment, Fourier transform operation is performed on the real-time current data, so that the real-time current data can be converted into a real-time current waveform, and an abnormal current waveform can be extracted from the real-time current waveform.

[0055] Alternatively, the current waveform of each equipment can also be detected by an inspection instrument to directly obtain the real-time current waveform, and an abnormal current waveform can be extracted according to the frequency or amplitude of the real-time current waveform.

[0056] In order to accurately extract the abnormal current waveform and improve the accuracy of subsequent analysis and early warning, as an example, the acquiring real-time current data and performing Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extracting an abnormal current waveform from the real-time current waveform can include the following sub-steps:

[0057] S111, acquiring a plurality of real-time current data and performing preprocessing on each real-time current data to obtain a plurality of processing data, and converting each processing data into a real-time current waveform through fast Fourier transform, wherein the preprocessing includes noise reduction processing and time domain conversion processing.

[0058] S112, calculate total harmonic distortion of each real-time current waveform, determine abnormal current waveform from several real-time current waveforms based on the total harmonic distortion.

[0059] Since there are multiple devices in the valve hall of the converter station, each device can be provided with a detection instrument, and real-time current data of each device is obtained through the detection instrument to obtain several real-time current data.

[0060] The real-time current signal collected in the time domain is preprocessed, and the preprocessing includes noise reduction processing and time domain conversion processing. Through preprocessing, the direct current component and noise interference can be removed to obtain a standardized current signal sequence. The standardized current signal sequence is subjected to fast Fourier transform, and each real-time current signal in the time domain can be converted to a real-time current signal in the frequency domain to obtain a real-time current waveform. Then, the amplitude and phase information of the frequency component of the real-time current waveform can be obtained, and the amplitude ratio of the fundamental wave and the harmonic wave is determined by using the amplitude and phase information of the frequency component to obtain the frequency spectrum distribution data.

[0061] The total harmonic distortion is calculated by using the fundamental wave and the harmonic wave of the frequency spectrum distribution data. Specifically, the ratio of the root mean square value of the harmonic content to the root mean square value of its fundamental wave component (expressed in percentage) can be calculated. If the total harmonic distortion rate exceeds the preset distortion rate threshold, it is judged that there is a waveform distortion abnormal state, and the real-time current waveform whose total harmonic distortion rate exceeds the preset distortion rate threshold can be taken as the abnormal current waveform.

[0062] For the frequency spectrum distribution data of the judgment of the waveform distortion abnormal state, the harmonic component whose amplitude exceeds the preset proportion threshold of the fundamental wave amplitude is extracted, and the frequency value and amplitude value corresponding to the harmonic component are recorded. The maximum amplitude value is identified from the recorded data as the abnormal fluctuation amplitude, and the corresponding frequency value is identified as the abnormal fluctuation frequency.

[0063] Specifically, the collection of real-time current signal data is usually realized through current transformers. These transformers are installed at key positions of various devices in the valve hall and can convert high-voltage large-current into small-current signals suitable for measurement in proportion. The acquisition system can obtain thousands of sampling points per second to form continuous real-time current data.

[0064] In one possible implementation, the preprocessing process mainly targets the interference components in the original signal. The removal of the direct current component is achieved by calculating the average value within a period and subtracting this value from the original signal. The elimination of noise interference adopts the method of moving average or low-pass filtering, retaining the main frequency components of the current signal and filtering out high-frequency noise. The standardized current signal sequence after preprocessing has clearer waveform characteristics, laying the foundation for subsequent frequency domain analysis. Through fast Fourier transform, the current waveform that changes over time can be decomposed into different frequency sinusoidal components. The fundamental wave is usually 50 Hz, while the harmonics are integer multiples of the fundamental frequency, such as the second harmonic of 100 Hz, the third harmonic of 150 Hz, etc. Each frequency component has corresponding amplitude and phase information, the amplitude reflects the strength of the frequency component in the original signal, and the phase indicates the time offset of the component relative to the reference point.

[0065] It should be noted that the total harmonic distortion is an important indicator for measuring the quality of the current waveform. Its calculation process involves comprehensive evaluation of all harmonic components: first, calculate the square of the amplitude of each harmonic, then sum and take the square root, and finally compare with the fundamental wave amplitude. When the device is running normally, the current waveform is close to a sine wave, and the total harmonic distortion is low; when an anomaly occurs, the waveform will be distorted, showing an increase in harmonic content. The preset distortion threshold is usually determined according to the specific requirements of the power system and the characteristics of the device.

[0066] Specifically, the identification of abnormal fluctuations requires in-depth analysis of the frequency spectrum data. When the total harmonic distortion exceeds the threshold, further screening of the harmonic components with larger amplitudes is performed. These significant harmonic components are often related to specific abnormal causes.

[0067] For example, the 5th and 7th harmonics may be due to the nonlinear characteristics of the converter, while the 3rd harmonic may be related to unbalanced loads. By recording the frequencies and amplitudes of these main harmonics, the characteristics of the abnormal fluctuations can be accurately located. The maximum amplitude value directly reflects the severity of the anomaly, while the corresponding frequency value helps to analyze the root cause of the anomaly, providing key evidence for subsequent fault diagnosis and handling.

[0068] S12, determine the waveform parameter of the abnormal current waveform, and determine the corresponding probability level value according to the waveform parameter by using a preset support vector machine, wherein the probability level value is a fault level corresponding to the probability of causing an electric arc due to device overload.

[0069] After determining one or more abnormal current waveforms, the waveform parameter of each abnormal current waveform can be obtained, which can be the amplitude of the abnormal current waveform. Then, the amplitude of the abnormal current waveform can be input to a preset support vector machine to evaluate the probability of causing an electric arc, and then determine the level of causing a fire accident according to the probability, to determine whether subsequent warning analysis is needed.

[0070] In an optional embodiment, the waveform parameters include frequency values and amplitude values; and the step of determining the corresponding probability level value according to the waveform parameters by using the preset support vector machine can include the following sub-steps:

[0071] S121, calculating an amplitude change rate using the amplitude value, and converting the frequency value, the amplitude value and the amplitude change rate into a vector to obtain a parameter feature vector.

[0072] S122, calling a preset support vector machine to perform classification processing according to the parameter feature vector to obtain an arc probability value.

[0073] S123, determining a target interval value from a plurality of preset interval values according to the arc probability value, and determining a probability level value corresponding to the target interval value.

[0074] After obtaining the waveform parameters of the abnormal current waveform, the waveform parameters can include fluctuation frequency and amplitude data, the amplitude value can be obtained, the difference between the amplitude value at the current time and the amplitude value at the adjacent time is calculated, and then the amplitude change rate is obtained by calculating the ratio of the difference to the time interval.

[0075] After converting the frequency value, the amplitude value and the amplitude change rate into a vector to obtain a parameter feature vector, the support vector machine algorithm can be used to perform classification processing according to the parameter feature vector, and the device overload arc occurrence probability value is calculated by the distance between the decision function value and the classification boundary to obtain the arc probability value.

[0076] A plurality of preset interval values can be preset, for example, 0-10 is the first interval value, 10-20 is the second interval value, 20-30 is the third interval value, and so on. Then each interval value corresponds to a probability level value. The interval value probability level value of 0-10 is 1, and the interval value probability level value of 10-20 is 2. Finally, the preset interval value corresponding to the arc probability value can be determined.

[0077] Alternatively, the arc probability value can also be compared with a preset probability threshold value. If the probability value is less than the first threshold value, the probability level value is determined to be 1; if the probability value is between the first threshold value and the second threshold value, the probability level value is determined to be 2; and if the probability value is greater than the second threshold value, the probability level value is determined to be 3.

[0078] Specifically, the acquisition of the fluctuation frequency and amplitude data of the abnormal current waveform is the basis for arc fault diagnosis. These data are derived from the previous spectrum analysis results, and contain the specific values of each frequency component and the corresponding amplitude information. The amplitude change rate reflects the dynamic characteristics of the current fluctuation, which is obtained by dividing the amplitude difference between adjacent sampling times by the time interval.

[0079] For example, if the current amplitude is 100A at the current moment, and the amplitude becomes 120A after 0.01s, the amplitude change rate is 2000A / s. Such a change rate can reflect the mutation characteristics of arc occurrence.

[0080] It should be noted that the frequency value is usually concentrated in a certain range, such as the arc fault usually generates a characteristic frequency between 2kHz and 10kHz; the amplitude value reflects the severity of the anomaly; and the amplitude change rate reflects the speed of fault development. The amplitude, frequency and amplitude change rate can be used to construct a feature vector, and a preset support vector machine is called to perform classification processing according to the three feature vectors to obtain the probability of arc occurrence. By comprehensively considering the information in multiple dimensions, the accuracy of subsequent evaluation can be further improved.

[0081] In a possible implementation, when the preset support vector machine calculates the probability value, the distance of the classification boundary is calculated by using a sigmoid function or other probability mapping method, and then the distance is converted into a probability value between 0 and 1. When the distance is positive and large, the probability is close to 1, indicating that the arc is extremely likely to occur; and when the distance is negative and the absolute value is large, the probability is close to 0, indicating that the arc is very unlikely to occur.

[0082] In addition, the probability level can also be divided by using a segmented threshold. The first threshold is usually set to about 0.3, and the second threshold is usually set to about 0.7. Such a grading manner enables the device maintenance personnel to take corresponding measures according to different risk levels. A low probability level means that the device is basically normal and only needs to be monitored regularly; a medium probability level indicates that there is a potential risk and the inspection needs to be strengthened; and a high probability level indicates that the arc fault is about to occur or has occurred, and immediate measures need to be taken.

[0083] S13, when the probability level value is greater than the preset probability level value, a plurality of potential fault points are found according to the waveform parameter size, and a pressure wave superposition feature value of the plurality of potential fault points is obtained.

[0084] When the probability level value is greater than the preset probability level value, it indicates that the probability of causing an arc is high, and the probability of causing a fire and a fault is also high. In order to determine the position where the fault is likely to occur as soon as possible, a plurality of potential fault points can be found according to the waveform parameter size, and the potential fault point is the position of the faulty device. Because each device corresponds to a real-time current data, if the device is faulty, the real-time current data will become an abnormal current waveform, and the probability level value calculated based on the waveform parameter of the abnormal current waveform is greater than the preset probability level value. The position of the device of the abnormal current waveform can be obtained to determine a potential fault point.

[0085] For the probability level value less than the preset probability level value, the device can be investigated or the technician is notified to check it, further reducing the risk of accidents. For example, when the probability level value is less than the preset probability level value, the technician can be notified to investigate the device at a regular position.

[0086] Specifically, the judgment of potential fault points can also rely on the association of frequency characteristics and device physical characteristics. Different positions of the device have different inherent frequencies. The inherent frequency of the converter valve can be near 3 kHz, and the inherent frequency of the bus can be near 5 kHz. When the detected abnormal frequency is close to the inherent frequency of a certain device, the device becomes the key object of investigation. At the same time, the priority and range of investigation can also be determined by the probability level: when the probability level is high, the device with the highest frequency matching degree and its adjacent devices are concentrated for inspection; when the probability level is medium, the inspection range is expanded to the entire area; when the probability level is low, only routine inspection is needed. This hierarchical and zoned fault positioning method greatly improves the efficiency and accuracy of fault diagnosis.

[0087] In one of the embodiments, the step of searching for a plurality of potential fault points according to the waveform parameter size, and obtaining pressure wave superposition characteristic values from the plurality of potential fault points can include the following sub-steps:

[0088] S131, determining the position coordinates of the corresponding device of each abnormal current waveform, to obtain a plurality of potential fault points.

[0089] S132, obtaining real-time pressure waveforms based on the pressure sensors of the potential fault points, and extracting wave peak superposition degree values, wave valley offset values, and energy distribution values from each real-time pressure waveform.

[0090] S133, combining the wave peak superposition degree values, the wave valley offset values, and the energy distribution values to obtain pressure combination values, and screening a plurality of pressure wave superposition characteristic values from a plurality of pressure combination values, the pressure wave superposition characteristic values being pressure combination values with values greater than a preset pressure value.

[0091] Because of a device failure, the generated data can cause several device abnormalities, and several abnormal current waveforms can be screened, corresponding to the position coordinates of several abnormal devices or fault devices, thereby obtaining a plurality of potential fault points.

[0092] Meanwhile, because the device with the abnormal current waveform is a faulty device, the data generated by the backend device according to the erroneous signal transmitted by the device may also be error-free. If there is no error, the waveform corresponding to the device is a normal waveform. Even if the data of the backend device is error-free, the arc may also affect the device, becoming a fault location. In order to further investigate these devices, the size of the waveform parameter can be used as a screening condition to further find several potential fault points. For example, a current waveform that satisfies the waveform parameter (the frequency of the waveform is greater than the frequency value of the waveform parameter, and the amplitude of the waveform is greater than the amplitude value of the waveform parameter) is found from the normal real-time current waveform, and the position of the device with the current waveform is used as a potential fault point. Through the above finding process, several potential fault points can be found. Then the position coordinates of the potential fault points can be connected together to construct a fault propagation path. When connecting, the position coordinates of the potential fault points can be connected together to obtain the fault propagation path according to the transmission direction of the signal or the transmission direction of the current.

[0093] Then, the pressure wave superposition characteristic value corresponding to the potential fault point in the fault propagation path can be extracted, and the real-time pressure is determined through the pressure wave superposition characteristic value to determine whether the arc influence causes the position to generate a pressure wave, so as to facilitate subsequent accident analysis.

[0094] In a specific operation mode, real-time pressure wave data can be obtained from the pressure sensor of the device corresponding to the potential fault point of the fault propagation path. The pressure wave data is a sequence of continuous sampling values of pressure change over time. Time domain analysis is performed on the collected pressure wave data to convert it into a waveform graph to obtain a real-time pressure wave waveform. The maximum value point in the real-time pressure wave waveform is identified as a wave crest position, and the minimum value point is identified as a wave trough position. For the identified wave crest and wave trough positions, the time difference of the same pressure wave crest detected by different sensors is calculated, the wave crest superposition degree value is determined according to the time difference, the wave trough offset value is calculated to obtain the wave trough offset value, and the pressure wave amplitude square in the preset time window is accumulated to obtain the energy distribution value.

[0095] It should be noted that the sampling frequency of the pressure wave data reaches tens of thousands per second, which can capture the pressure change at the microsecond level. These continuous sampling values reflect the instantaneous fluctuation of air pressure. When the arc discharges, a sharp pressure rise will form a shock wave that propagates in all directions. Time domain analysis identifies extreme points by comparing the size relationship of adjacent sampling values point by point: when the pressure value of a point is greater than that of the two adjacent sampling points, the point is a wave crest; otherwise, it is a wave trough. The calculation of the wave crest arrival time difference reveals the propagation characteristics of the pressure wave.

[0096] For example, sensor A detects a wave crest at time t1, and sensor B detects a wave crest of the same feature at time t2, the time difference t2-t1 reflects the time required for the pressure wave to propagate from the vicinity of A to the vicinity of B. When multiple sensors detect pressure waves at the same time, the wave crests at certain positions may superimpose on each other, forming a higher pressure peak. The degree of superposition is quantified by comparing the ratio of the theoretical single wave peak value to the actual measured peak value.

[0097] In an embodiment, the energy distribution value can be expressed as follows:

[0098] ;

[0099] In the above formula, E d represents the energy distribution value, M represents the number of sampling points in the preset time window, P k represents the pressure wave amplitude of the kth sampling point, and Δt represents the sampling time interval. The formula calculates the energy distribution characteristic value in the time period by accumulating the squares of all pressure wave amplitudes in the time window. The wave crest superposition degree value, the wave trough offset value, and the energy distribution value are combined to form a pressure combination value.

[0100] According to the analysis above, the potential fault point screened may be a normal device, an abnormal device, or a device affected by an abnormal device. Therefore, the pressure combination values of different devices are different, and a pressure combination value greater than a preset pressure value can be screened from multiple pressure combination values as a pressure wave superposition characteristic value.

[0101] S14, determine the position coordinates of each pressure wave superposition characteristic value, and obtain the temperature time sequence parameter and real-time image of each position coordinate, determine the fire class value corresponding to each position coordinate according to the real-time image and the temperature time sequence parameter, and when any one of the fire class values is greater than a preset fault class value, trigger the corresponding safety warning processing according to the fire class value.

[0102] In an embodiment, the device corresponding to each pressure wave superposition characteristic value can be determined, and then the position coordinates of the device are obtained, and the temperature time sequence parameter is collected by the temperature sensor of the device corresponding to each position coordinate, and the real-time image is obtained by the camera of the device corresponding to each position coordinate. According to the real-time image, it is determined whether there is a fire source, and according to the temperature time sequence parameter, it is determined whether a fire occurs, and then the fire class value corresponding to each position coordinate is determined.

[0103] Because there are several pressure wave superposition characteristic values determined, there are also several fire class values calculated. When any one of the fire class values is greater than a preset fault class value, it indicates that there is a fire risk, and the safety warning processing can be triggered.

[0104] Optionally, the pressure impact source position, i.e. the position of the equipment generating pressure due to failure or arc, can also be determined according to the energy distribution value, and then the temperature time sequence parameter and real-time image can be obtained through the temperature sensor and camera of the equipment at the pressure impact source position.

[0105] Specifically, the three sensor positions with the maximum energy can be identified according to the energy distribution value in the pressure wave superposition characteristic value, the wave peak arrival time difference detected by the three sensors and the known sensor spatial coordinates are used to determine the impact source position by calculating the relationship between the pressure wave propagation time and distance. Wherein, the pressure arrival time difference can be shown as follows:

[0106] ;

[0107] In the above formula, Δt ij represents the arrival time difference of the wave peak detected by sensor i and sensor j, t j represents the wave peak arrival time of sensor j, t i represents the wave peak arrival time of sensor i, d sj represents the distance from the impact source to sensor j, d si represents the distance from the impact source to sensor i, and v represents the pressure wave propagation speed.

[0108] The pressure impact source position can be shown as follows:

[0109] ;

[0110] In the above formula, x s , y s and z s represent the spatial coordinates of the impact source, x i , y i and z i represent the known spatial coordinates of the i-th sensor, v represents the pressure wave propagation speed, t i represents the time when the pressure wave reaches the i-th sensor, and d0 represents the reference distance constant.

[0111] The pressure impact source position can also be determined in the above manner, the coordinates of the pressure impact source position are obtained, the position coordinates are obtained, the temperature time sequence parameter and real-time image of each pressure impact source position are obtained, and the fire class value corresponding to each position coordinate is determined according to the real-time image and the temperature time sequence parameter.

[0112] In addition, the calculation of the trough offset is also important. Normally, the trough of the pressure wave should appear at a fixed time interval after the peak, but due to reflection and interference effects, the trough position will be offset. This offset is obtained by calculating the difference between the actual trough time and the theoretical trough time, and can reflect the degree of interference received by the pressure wave during propagation. The energy distribution value is calculated using the energy integration method. In the preset time window, select the time period containing the complete pressure waveform, square the pressure value of each sampling point and accumulate. The square of the pressure represents the energy density at that time, and the accumulation result reflects the total energy received at that sensor position. The spatial difference of the energy distribution directly indicates the approximate direction of the impact source.

[0113] The determination of the above impact position is based on the principle of acoustic positioning. After identifying the three sensors with the maximum energy, a positioning equation set is constructed using their spatial coordinates and the peak arrival time difference. Assuming that the distances from the impact source to the three sensors are d1, d2, and d3, and the propagation speed of the pressure wave is v, then the distance difference can be calculated by the time difference.

[0114] When the calculated fire level value is greater than the preset fault level value, a safety warning process can be triggered. Different fire level values can correspond to different alarm information, and different safety warning processes can be performed according to the fire level value.

[0115] For example, when the fire level value is 1, the safety warning process can be a warning broadcast, and a warning broadcast can be sent to remind the construction personnel to pay attention.

[0116] When the fire level value is 2, the safety warning process can be an emergency alarm broadcast, a work suspension warning can be sent, and the background personnel can be prompted to suspend work and evacuate the technicians near the potential fault point for emergency measures by the maintenance personnel.

[0117] When the fire level value is 3, the technician is notified to start the emergency response, the power supply of the equipment is cut off, and the evacuation work is arranged, etc.

[0118] In an embodiment, the location has a large number of objects, and once a fault or arc is triggered, a fire is easily triggered. In order to analyze the fire level according to the number of objects, in one of the embodiments, the determination of the fire level value corresponding to each position coordinate according to the real-time image and the temperature time sequence parameter can include the following sub-steps:

[0119] S21, a preset first image recognition model is called to identify the number of objects in the real-time image and the interval distance between each object and the equipment, and the number of objects and the interval distance value are obtained respectively.

[0120] S22, calculating a quantity weight value according to the number of objects and the interval distance value.

[0121] S23, calculate a temperature change value of the temperature time sequence parameter, and determine a fire grade value by using the temperature change value and the quantity weight value.

[0122] In an embodiment, a preset first image recognition model (e.g., a pre-trained CNN image recognition model) can be called to recognize objects in the real-time image and count the number of the objects. The recognized objects can include the device and objects stacked around the device. Then, the distance between the objects and the device can be calculated to obtain the object quantity value and the distance value, respectively.

[0123] In actual operation, a preset first image recognition model (e.g., a pre-trained CNN image recognition model) can be called to recognize objects in the real-time image and count the number of the objects to obtain the object quantity value. Then, the image distance between each object and the device is recognized, and the image distance is converted into an actual distance according to a conventional image conversion method to obtain the distance value.

[0124] The farther the distance between the objects and the device, the lower the probability of being directly ignited due to a fault, and vice versa. The quantity weight value can be calculated according to the object quantity value and the distance value.

[0125] In an embodiment, the calculation of the quantity weight value can be as follows:

[0126] Q = (D1 + D2 + … + Di) / i;

[0127] wherein Di is the reciprocal of the distance value between each object and the device, and i is the object quantity value.

[0128] Then, the maximum value and the minimum value can be obtained from the temperature time sequence parameter, the difference between the maximum value and the minimum value is calculated to obtain the temperature change value, and finally, the product of the temperature change value and the quantity weight value is calculated to obtain the fire probability value. The fire grade value can be determined according to the size of the fire probability value.

[0129] Alternatively, a plurality of preset interval values can also be set in advance, for example, 0-10 is the first interval value, 10-20 is the second interval value, 20-30 is the third interval value, and so on. Then, each interval value corresponds to a fire grade value. For example, the fire grade value of the 0-10 interval value is 1, and the fire grade value of the 10-20 interval value is 2. Finally, the fire grade value can be determined according to the size of the fire probability value.

[0130] In an embodiment, the objects in the positions are combustible materials, and there are more flammable and explosive materials, and the fire probability is higher. In order to analyze the fire level according to the properties of the materials, in another embodiment, the step of determining the fire level value corresponding to each position coordinate according to the real-time image and the temperature time sequence parameter can include the following sub-steps:

[0131] S31, calling a preset second image recognition model to identify the materials of the objects in each real-time image to obtain a plurality of material information.

[0132] S32, finding the ignition temperature value corresponding to each material based on each material information, and calling a K-means algorithm to cluster a plurality of ignition temperature values to obtain a plurality of ignition categories.

[0133] S33, counting the number of ignition temperature values in each ignition category to obtain a temperature quantity value, and calculating a temperature peak value of the temperature time sequence parameter.

[0134] S34, determining a fire level value using the temperature quantity value and the temperature peak value.

[0135] The preset second image recognition model (such as a pre-trained CNN image recognition model) is called to identify the materials of the objects in each real-time image to obtain a plurality of material information.

[0136] It should be noted that the preset first image recognition model is used to identify objects, object quantity, and object distance, and the preset second image recognition model is used to identify material information. Although both can be CNN model frameworks, the two can be trained using different data.

[0137] Then, the ignition temperature value corresponding to each material can be found based on each material information. The ignition temperature value is the combustible temperature value of the material.

[0138] Since there are many objects, there can be multiple ignition temperature values, and the span of the ignition temperature values is large. For example, the ignition temperature value of some materials is 150 degrees, and the ignition temperature value of some materials is 500 degrees. In order to improve processing efficiency, a K-means algorithm can be called to cluster a plurality of ignition temperature values to obtain a plurality of ignition categories. Each category corresponds to a temperature interval. For example, 100 degrees can be used as the interval of the category, and after clustering, there are three ignition categories. The temperature interval of the first ignition category is 0-100 degrees, the temperature interval of the second ignition category is 100-200 degrees, and the temperature interval of the third ignition category is 200-300 degrees.

[0139] Then, the number of ignition temperature values in each ignition category can be counted to obtain a temperature quantity value, and a temperature peak value of the temperature time sequence parameter can be calculated.

[0140] Finally, the fire probability value is calculated by using the temperature quantity value and the temperature peak value, and the fire grade value is determined according to the fire probability value.

[0141] In the embodiment, the calculation of the fire probability value can be shown in the following formula:

[0142] K = T * (e A1 + e A2 + … + e Aj ).

[0143] Wherein, K is the fire probability value, T is the temperature peak value, e is a calculation constant (specifically, an integer greater than 1, and optionally, can be in the interval of 1-2), Aj is the temperature quantity value, and j is the number of ignition categories.

[0144] Alternatively, a plurality of preset interval values can be set in advance according to the above-mentioned embodiment, for example, 0-10 is the first interval value, 10-20 is the second interval value, 20-30 is the third interval value, and so on. Then each interval value corresponds to a fire grade value. For example, the fire grade value of the 0-10 interval value is 1, and the fire grade value of the 10-20 interval value is 2. Finally, the fire grade value can be determined according to the size of the fire probability value.

[0145] In an embodiment, the failure of a strong current equipment can instantaneously cause an electric arc, and further cause a fire. In order to timely handle the fire, after the step of acquiring the temperature time sequence parameter and the real-time image of each position coordinate, the method further includes the following sub-steps:

[0146] S41, if the temperature time sequence parameter meets a threshold value or the real-time image contains a heat source by calling a preset BP model, determining a heat source coordinate and constructing an isolation boundary region with the heat source coordinate as the center.

[0147] S42, turning off the equipment in the isolation boundary region, and starting emergency alarm processing.

[0148] In an embodiment, if the maximum value of the temperature time sequence parameter is greater than a preset high temperature threshold value or the real-time image contains a heat source by calling a preset BP model, it is determined that a fire has occurred, the heat source coordinate can be determined, the coordinate of the camera can be acquired, and the heat source coordinate is obtained. Finally, an isolation boundary region is constructed with the heat source coordinate as the center.

[0149] In an actual operation mode, the fire grade value can be synchronously evaluated, and the radius is determined according to the fire grade value, for example, the fire grade value is 1, and the radius is 50 meters, the fire grade value is 2, and the radius is 100 meters. A isolation boundary area is constructed with the heat source coordinates as the center and the above distance as the radius. Finally, the power supply of the equipment in the isolation boundary area can be cut off to shut down the equipment, and an emergency alarm process can be started to inform the background personnel in the isolation boundary area of the safety vehicle and inform the maintenance personnel to take emergency measures.

[0150] In the embodiment, the fire protection early warning method of the converter station has the beneficial effects that: the real-time current data can be obtained, and the real-time current data is subjected to Fourier transform operation to obtain a real-time current waveform, and an abnormal current waveform is extracted from the real-time current waveform; the waveform parameters of the abnormal current waveform are determined, and a corresponding probability grade value is determined according to the waveform parameters by using a preset support vector machine; when the probability grade value is greater than a preset probability grade value, a plurality of potential fault points are searched according to the size of the waveform parameters, and pressure wave superposition characteristic values from the plurality of potential fault points are obtained; the position coordinates of each pressure wave superposition characteristic value are determined, and the temperature time sequence parameters and real-time images of each position coordinate are obtained, the fire grade value corresponding to each position coordinate is determined according to the real-time images and the temperature time sequence parameters, and when any one fire grade value is greater than a preset fault grade value, a safety early warning process is triggered. The current data, pressure data, temperature, and real-time images are used to monitor the equipment in the converter station, the fire grade value can be accurately determined and the alarm can be triggered, the evaluation accuracy of safety accidents can be improved, different alarms can be triggered for different fire grade values, the technical personnel can be reminded to repair and maintain early, the timeliness of the early warning can be improved, the alarm delay can be shortened, the technical personnel can maintain early, the risk of the equipment can be further reduced, and the life safety of the technical personnel can be protected.

[0151] The embodiment of the present application further provides a fire early warning system of a converter station, referring to Figure 2 , a structural schematic diagram of a fire early warning system of a converter station provided by an embodiment of the present application is shown.

[0152] For example, the fire early warning system of the converter station can include:

[0153] The acquisition module 201 is configured to acquire real-time current data, perform Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extract an abnormal current waveform from the real-time current waveform, wherein the real-time current data is real-time data of equipment in a valve hall of the converter station.

[0154] The determining module 202 is configured to determine a waveform parameter of the abnormal current waveform, and determine a corresponding probability level value according to the waveform parameter by using a preset support vector machine, wherein the probability level value corresponds to a failure level of an arc caused by equipment overload.

[0155] The extracting module 203 is configured to, when the probability level value is greater than a preset probability level value, find a plurality of potential failure points according to the waveform parameter, and obtain pressure wave superposition characteristic values of the plurality of potential failure points.

[0156] The early warning module 204 is configured to determine a position coordinate of each of the pressure wave superposition characteristic values, and obtain a temperature time sequence parameter and a real-time image of each of the position coordinates, determine a fire level value corresponding to each of the position coordinates according to the real-time image and the temperature time sequence parameter, and trigger a corresponding safety early warning process according to the fire level value when any one of the fire level values is greater than a preset failure level value.

[0157] Optionally, the determining of the fire level value corresponding to each of the position coordinates according to the real-time image and the temperature time sequence parameter comprises:

[0158] calling a preset first image recognition model to recognize a number of objects in the real-time image and a spacing distance between each object and the equipment, and obtaining an object number value and a spacing distance value, respectively;

[0159] calculating a quantity weight value according to the object number value and the spacing distance value;

[0160] calculating a temperature change value of the temperature time sequence parameter, and determining the fire level value by using the temperature change value and the quantity weight value.

[0161] Optionally, the determining of the fire level value corresponding to each of the position coordinates according to the real-time image and the temperature time sequence parameter comprises:

[0162] calling a preset second image recognition model to perform material recognition on each object in each of the real-time images, and obtaining a plurality of material information;

[0163] finding a corresponding ignition temperature value of each material based on each of the material information, calling a K-means algorithm to cluster a plurality of the ignition temperature values, and obtaining a plurality of ignition categories;

[0164] counting a number of ignition temperature values of each of the ignition categories to obtain a temperature number value, and calculating a temperature peak value of the temperature time sequence parameter;

[0165] determining the fire level value by using the temperature number value and the temperature peak value.

[0166] Optionally, the waveform parameter comprises a frequency value and an amplitude value.

[0167] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0168] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0169] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0170] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0171] Optionally, the searching for a plurality of potential fault points according to the waveform parameter size, and acquiring pressure wave superposition characteristic values from the plurality of potential fault points comprises:

[0172] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0173] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0174] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0175] Optionally, the acquiring real-time current data and performing Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extracting an abnormal current waveform from the real-time current waveform comprises:

[0176] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0177] The using the preset support vector machine to determine a corresponding probability level value according to the waveform parameter comprises:

[0178] Optionally, the system further comprises:

[0179] The constructing area module is configured to, after the step of acquiring the temperature time sequence parameter and the real-time image of each position coordinate, determine a heat source coordinate and construct an isolated boundary area with the heat source coordinate as the center, if the temperature time sequence parameter meets a threshold or the real-time image contains a heat source identified by calling a preset BP model.

[0180] The closing module is configured to close the equipment in the isolated boundary area and start an emergency alarm process.

[0181] Those skilled in the art can clearly understand that, for the convenience of description and conciseness, the specific working process of the device described above can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0182] Further, the embodiment of the present application also provides an electronic device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the fire protection and early warning method of the converter station as described in the foregoing embodiments when executing the program.

[0183] Further, the embodiment of the present application also provides a computer readable storage medium, which stores a computer executable program, and the computer executable program is used to make a computer execute the fire protection and early warning method of the converter station as described in the foregoing embodiments.

[0184] In the description of the embodiments of the present application, it should be noted that the positions or position relationships indicated by the terms "upper", "lower" and the like are based on the positions or position relationships shown in the drawings, and are only for the convenience of describing the embodiments of the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular position, be constructed and operated in a particular position, and therefore cannot be understood as a limitation on the present application. When an element such as a layer, a region or a substrate is referred to as "on" or "above" another element, it can be directly on the other element, or there can be an intermediate element. In contrast, when an element is referred to as "directly on" or "directly above" another element, there is no intermediate element. It should also be understood that when an element is referred to as "below" or "under" another element, it can be directly below or under the other element, or there can be an intermediate element. In contrast, when an element is referred to as "directly below" or "directly under" another element, there is no intermediate element. Unless otherwise specified and limited, the terms "mounting", "connection", "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or the connection between two elements inside. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0185] Those skilled in the art will appreciate that embodiments of the application can also provide for computer program products. Accordingly, the present application can be embodied in the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0186] The present application is described in reference to the flowchart illustrations and / or block diagrams of methods, apparatus (systems), means and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing device or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks.

[0187] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks.

[0188] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks.

[0189] The above description is only preferred embodiment of the present application, it should be pointed out that, for those skilled in the art, without departing from the technical principles of the present application, can make a number of improvements and variations, these improvements and variations should be considered as the protection scope of the present application.

Claims

1. A method for early warning of fire protection of a converter station, characterized in that, The method comprises: obtaining real-time current data and performing Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extracting an abnormal current waveform from the real-time current waveform, wherein the real-time current data is real-time data of equipment in a valve hall of a converter station; determining a waveform parameter of the abnormal current waveform, and determining a corresponding probability level value according to the waveform parameter by using a preset support vector machine, wherein the probability level value is a fault level corresponding to a probability of arc caused by equipment overload; when the probability level value is greater than a preset probability level value, searching for a plurality of potential fault points according to the waveform parameter size, and obtaining pressure wave superposition characteristic values from the plurality of potential fault points; determining a position coordinate of each of the pressure wave superposition characteristic values, and obtaining a temperature time sequence parameter and a real-time image of each of the position coordinates, determining a fire level value corresponding to each of the position coordinates according to the real-time image and the temperature time sequence parameter, and triggering a corresponding safety warning process according to the fire level value when any one of the fire level values is greater than a preset fault level value; the searching for a plurality of potential fault points according to the waveform parameter size and obtaining pressure wave superposition characteristic values from the plurality of potential fault points comprises: determining a position coordinate of the equipment corresponding to each of the abnormal current waveforms to obtain a plurality of potential fault points; obtaining real-time pressure waveforms based on pressure sensors of the potential fault points, and extracting a wave peak superposition degree value, a wave trough offset value and an energy distribution value from each of the real-time pressure waveforms; combining the wave peak superposition degree value, the wave trough offset value and the energy distribution value to obtain a pressure combination value, and screening a plurality of pressure wave superposition characteristic values from a plurality of pressure combination values, wherein the pressure wave superposition characteristic value is a pressure combination value with a value greater than a preset pressure value.

2. The fire protection early warning method of a converter station according to claim 1, characterized in that, the determining a fire level value corresponding to each of the position coordinates according to the real-time image and the temperature time sequence parameter comprises: calling a preset first image recognition model to recognize the number of objects in the real-time image and the interval distance between each object and the equipment to obtain an object number value and an interval distance value, respectively; calculating a quantity weight value according to the object number value and the interval distance value; calculating a temperature change value of the temperature time sequence parameter, and determining a fire level value by using the temperature change value and the quantity weight value.

3. The fire protection early warning method of a converter station according to claim 1, characterized in that, the determining a fire level value corresponding to each of the position coordinates according to the real-time image and the temperature time sequence parameter comprises: calling a preset second image recognition model to perform material identification on the objects in each of the real-time images to obtain a plurality of material information; finding a burning point temperature value corresponding to each material based on each of the material information, calling a K-means algorithm to cluster a plurality of the burning point temperature values to obtain a plurality of burning point categories; counting the number of burning point temperature values of each of the burning point categories to obtain a temperature number value, and calculating a temperature peak value of the temperature time sequence parameter; determining a fire level value by using the temperature number value and the temperature peak value.

4. The fire protection early warning method of a converter station according to claim 1, characterized in that, the waveform parameter comprises a frequency value and an amplitude value. The probability level value corresponding to the waveform parameter is determined by using a preset support vector machine, and the method comprises the steps of: The amplitude variation rate is calculated by using the amplitude value, and the frequency value, the amplitude value and the amplitude variation rate are converted into a vector to obtain a parameter feature vector; The parameter feature vector is classified by calling a preset support vector machine to obtain an arc probability value; The target interval value is determined from a plurality of preset interval values according to the arc probability value, and the probability level value corresponding to the target interval value is determined.

5. The fire protection early warning method of a converter station according to claim 1, characterized in that, The real-time current data is obtained, and the real-time current data is subjected to Fourier transform operation to obtain a real-time current waveform, and an abnormal current waveform is extracted from the real-time current waveform, and the method comprises the steps of: A plurality of real-time current data is obtained, and each real-time current data is preprocessed to obtain a plurality of processing data, and each processing data is converted into a real-time current waveform by fast Fourier transform, wherein the preprocessing comprises noise reduction processing and time domain conversion processing; The total harmonic distortion rate of each real-time current waveform is calculated, and the abnormal current waveform is determined from a plurality of real-time current waveforms based on the total harmonic distortion rate.

6. The method of claim 1-5, wherein After the step of obtaining the temperature time sequence parameter and the real-time image of each position coordinate, the method further comprises the steps of: If the temperature time sequence parameter meets a threshold value or the real-time image contains a heat source by calling a preset BP model, a heat source coordinate is determined, and an isolation boundary region is constructed with the heat source coordinate as the center; The equipment in the isolation boundary region is turned off, and an emergency alarm process is started.

7. A fire warning system for a converter station, characterized in that The system comprises: An acquisition module is configured to acquire real-time current data, perform Fourier transform operation on the real-time current data to obtain a real-time current waveform, and extract an abnormal current waveform from the real-time current waveform, wherein the real-time current data is real-time data of equipment in a valve hall of a converter station; A determination module is configured to determine a waveform parameter of the abnormal current waveform, and determine a corresponding probability level value according to the waveform parameter by using a preset support vector machine, wherein the probability level value corresponds to a fault level of an arc caused by equipment overload; An extraction module is configured to, when the probability level value is greater than a preset probability level value, find a plurality of potential fault points according to the waveform parameter size, and acquire pressure wave superposition characteristic values from the plurality of potential fault points; An early warning module is configured to determine a position coordinate of each pressure wave superposition characteristic value, acquire a temperature time sequence parameter and a real-time image of each position coordinate, determine a fire disaster level value corresponding to each position coordinate according to the real-time image and the temperature time sequence parameter, and trigger a corresponding safety early warning process according to the fire disaster level value when any one of the fire disaster level values is greater than a preset fault level value; The method of finding a plurality of potential fault points according to the waveform parameter size and acquiring pressure wave superposition characteristic values from the plurality of potential fault points comprises the steps of: The position coordinates of the corresponding equipment of each abnormal current waveform are determined to obtain a plurality of potential fault points. The pressure sensor based on the potential fault point acquires real-time pressure waveforms, and extracts a wave crest superposition degree value, a wave trough offset value and an energy distribution value from each real-time pressure waveform; The wave crest superposition degree value, the wave trough offset value and the energy distribution value are combined to obtain a pressure combination value, and a plurality of pressure wave superposition characteristic values are screened from a plurality of pressure combination values, wherein the pressure wave superposition characteristic value is a pressure combination value with a value greater than a preset pressure value.

8. The fire warning system of a converter station according to claim 7, characterized in that, The fire grade value corresponding to each position coordinate is determined according to the real-time image and the temperature time sequence parameter, and the method comprises the steps of: A preset first image recognition model is called to recognize the number of objects in the real-time image and the interval distance between each object and the device, and an object quantity value and an interval distance value are obtained respectively; A quantity weight value is calculated according to the object quantity value and the interval distance value; A temperature change value of the temperature time sequence parameter is calculated, and the temperature change value and the quantity weight value are used to determine a fire grade value.

9. The fire warning system of a converter station according to claim 7, characterized in that, The fire grade value corresponding to each position coordinate is determined according to the real-time image and the temperature time sequence parameter, and the method comprises the steps of: A preset second image recognition model is called to perform material identification on the objects in each real-time image, and a plurality of material information is obtained; Based on each material information, the ignition temperature value corresponding to each material is found, and a K-means algorithm is called to cluster a plurality of ignition temperature values to obtain a plurality of ignition categories; The number of ignition temperature values of each ignition category is counted to obtain a temperature quantity value, and a temperature peak value of the temperature time sequence parameter is calculated; The temperature quantity value and the temperature peak value are used to determine a fire grade value.

Images