Converter station valve hall water leakage monitoring method and system based on photon counter

Abstract

The invention relates to a converter station valve hall water leakage monitoring method and system based on photon counters, and the method comprises the following steps: setting a line laser and a plurality of photon counters in a to-be-detected water leakage region, and illuminating the to-be-detected water leakage region of a converter station valve hall through the line laser; the photon number collected by the photon counter is sliced according to the form of time sequence data, and photon number time sequence data slices are obtained; the photon number time sequence data slices collected by each photon counter are analyzed respectively, and a water leakage monitoring judgment result of a single photon counter is obtained; and obtaining a water leakage monitoring judgment result of the to-be-detected water leakage area based on the water leakage monitoring judgment result of the single photon counter. Compared with the prior art, the method has the advantages of expanding the monitoring data dimension, improving the monitoring and judging speed and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of water leakage monitoring, and in particular to a method and system for monitoring water leakage in a valve hall of a converter station based on a photon counter. Background Art

[0002] The valve hall of the converter station is a key area in the high-voltage direct current (HVDC) transmission system, carrying the heavy task of power conversion and regulation. As an important hub in the power system, the valve hall is responsible for converting AC power into DC current through the converter valve, or converting DC current back to AC power. The stability of its function is crucial to the safety and reliability of the power system. The valve hall is equipped with core facilities such as converter valves, DC electrical equipment, and control systems. These devices not only bear high voltage and high current loads, but also need to work for a long time in extreme environments. Therefore, the operation of the valve hall directly affects the overall efficiency of the converter station. If there is a problem with the valve hall, it may trigger a series of chain reactions and even affect large-scale power transmission.

[0003] However, due to the long-term engineering stress and environmental factors of the valve hall of the converter station, its equipment and structure will gradually age and wear, especially the valve hall ceiling. As an important part of the valve hall, the ceiling is subject to great physical pressure and environmental changes, and is easily affected by moisture, temperature changes and air humidity. Long-term temperature difference changes may cause the sealing materials of the ceiling to age or break, which in turn causes water leakage. In addition, during the operation of the equipment in the valve hall, facilities such as pipelines and ventilation systems may produce condensed water, which may also leak along the ceiling to the equipment area. Once the leakage is not discovered and handled in time, the moisture will accumulate near the equipment under the ceiling, affecting the normal operation of the equipment. In particular, high-voltage electrical equipment such as converter valves may cause equipment performance degradation or failure once exposed to a humid environment. Therefore, the existence of ceiling leakage may seriously threaten the safety of the equipment inside the valve hall and the overall stable operation of the converter station.

[0004] The converter valves, control systems and related electrical equipment installed inside the valve hall are critical facilities in the system, and they need to maintain a high degree of reliability and stability. In the event of a water leak in the valve hall, moisture may penetrate into the converter valve and other electrical equipment, which may cause equipment failure, performance degradation, and even affect the entire operating efficiency and safety of the converter station in severe cases. Especially in high-voltage DC systems, any equipment failure may cause instability in the power system, leading to energy supply interruption or equipment damage, causing great economic losses and social impact. Therefore, in order to ensure the normal operation of the equipment in the valve hall of the converter station, effective monitoring measures must be taken to detect and quickly respond to the occurrence of valve hall leakage failures in real time. This can not only protect the equipment from water damage, but also identify potential risks in advance, carry out maintenance in a timely manner, ensure the safety, reliability and stability of the entire converter station system, and thus provide a strong guarantee for the smooth operation of the power system.

[0005] CN117073910A discloses a valve hall roof leakage monitoring system for a converter station, which uses image recognition to monitor the valve hall leakage of the converter station. However, since image processing has high performance requirements on the processing equipment, the processing speed is greatly affected by the performance of the processing equipment, and it is difficult to quickly obtain monitoring results. In addition, the accuracy of the output result is affected by the training data set. When it is difficult to obtain enough training images, it is difficult to ensure the accuracy of the final result. Therefore, a leakage monitoring method that can ensure both monitoring accuracy and monitoring speed is needed. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for monitoring water leakage in the valve hall of a converter station based on a photon counter. The photon counter collects the number of photons in space, and further processes the photon number time series data through slicing and feature calculation to realize water leakage monitoring in the valve hall of the converter station.

[0007] The purpose of the present invention can be achieved by the following technical solutions: A method for monitoring water leakage in a converter station valve hall based on a photon counter comprises the following steps: A line laser is set in the water leakage area to be detected and the water leakage area to be detected in the valve hall of the converter station is illuminated by the line laser, wherein the laser track emitted by the set line laser is a fan-shaped column with a preset thickness, and the laser track covers the water leakage area to be detected; A plurality of photon counters are arranged in the water leakage area to be detected, wherein the arranged photon counters measure the number of photons received within a fixed wavelength range, and the measurable wavelength range of the photon counters overlaps with the wavelength range of the laser emitted by the line laser; Slicing the number of photons collected by the photon counter in the form of time series data to obtain photon number time series data slices, wherein the photon number time series data slices include a one-dimensional time series array with a preset length organized according to a preset sampling rate, a start time of the time series data, and sampling rate information; Analyze the photon number time series data slices collected by each photon counter respectively and obtain the water leakage monitoring and discrimination result of a single photon counter; The water leakage monitoring and discrimination result of the water leakage area to be detected is obtained based on the water leakage monitoring and discrimination result of a single photon counter.

[0008] When the water leakage area to be detected is a rectangular area, the setting of the line laser includes the following steps: Determine the area to be detected and the location where the line laser can be installed: according to the site conditions, determine the horizontal boundary of the area to be detected for water leakage and the location range where the line laser can be installed. The location where the line laser can be installed is on the wall or frame of the building, represented by one or more line segments; Determine the installation position of the line laser: draw a perpendicular midline of the short side of the rectangular leak area to be detected, and form an intersection with the line segment describing the position where the line laser can be installed, and use the intersection as the position for installing the line laser; Determine the fan angle and effective distance of the line laser: The parameters of the line laser include the fan angle and effective distance L , the parameters of the line laser meet the following conditions: 1) effective distance L The maximum value of the distance from the installation position of the line laser to all positions in the leak area to be detected; 2) Laser fan angle The effective working sector range of the line laser covers the rectangular leak area to be detected, that is ,in, b is the length of the short side of the rectangular leakage area to be detected, d The distance between the installation position of the line laser and the short side of the rectangular leak area to be detected; Determine the installation angle of the line laser: set the vertical installation angle of the line laser to horizontal, that is, the plane of the fan-shaped column formed by the laser trajectory is parallel to the horizontal plane, and at the same time set the horizontal installation angle of the line laser to be perpendicular to the leaking area to be detected, that is, the angle bisector of the fan-shaped section of the fan-shaped column formed by the laser trajectory coincides with the perpendicular bisector of the short side of the rectangular area to be detected.

[0009] The setting of the photon counter comprises the following steps: Determine the installable position of the photon counter: determine the installable position of the photon counter according to the site conditions. The installable position of the photon counter is on the wall or frame of the building, represented by one or more line segments, and is consistent with the horizontal position of the installable line laser; Based on the installable positions of the photon counters, multiple photon counters are configured in the horizontal direction, with the receiving probes of all photon counters facing the leakage area to be detected; in the vertical direction, the height difference between the installation height of the photon counter and the line laser is set to be greater than a preset threshold; the installation angle of the photon counter is set to be parallel to the horizontal plane.

[0010] The difference between the upper and lower limits of the laser wavelength emitted by the line laser does not exceed 50 nm.

[0011] The photon counter can measure a wavelength range that is less than or equal to the wavelength range of the laser emitted by the line laser.

[0012] When slicing the time series data, adjacent time series data slices are made to have partially overlapping data according to a preset overlap rate.

[0013] The method of respectively analyzing the photon number time series data slices collected by each photon counter and obtaining the water leakage monitoring and discrimination result of a single photon counter specifically includes the following steps: A set of multidimensional feature discrimination methods for a one-dimensional time series is set, wherein the set of multidimensional feature discrimination methods includes a discrimination method based on a root mean square value, a kurtosis index, an EMD decomposition feature, and a spectrum energy feature of a sensitive band synchronous compression wavelet transform; Selecting multiple feature discrimination methods from a set of multidimensional feature discrimination methods, respectively calculating corresponding features based on the selected feature discrimination methods, and respectively determining whether a water leakage fault occurs according to the calculated features; The discrimination results of multiple feature discrimination methods for a single photon counter are summarized, and the weighted summation rule is used to determine whether the corresponding photon counter has detected a water leakage fault.

[0014] The discrimination method based on the root mean square value includes the following steps: Extract the root mean square value feature based on each photon number time series slice data of a certain photon counter; The root mean square value discrimination limit is set to be the baseline data mean plus three times the baseline data standard deviation, wherein the baseline data is a plurality of root mean square value features corresponding to a plurality of photon number time series slice data within a preset time; Setting a first discrimination time and a first discrimination number limit and discriminating a water leakage fault: within the set first discrimination time, judging the number of times that the root mean square value characteristic calculated based on multiple photon number time series slice data of the photon counter is greater than or equal to the root mean square value discrimination limit, which is recorded as the first number; if the first number is greater than or equal to the first discrimination number limit, it is judged that the photon counter detects a water leakage fault; The discrimination method based on kurtosis index includes the following steps: Extract the kurtosis index feature based on each photon number time series slice data of a certain photon counter; The kurtosis index determination limit is set to be the baseline data mean plus three times the baseline data standard deviation, wherein the baseline data is a plurality of kurtosis index features corresponding to a plurality of photon number time series slice data within a preset time; Setting a second discrimination time and a second discrimination number limit and discriminating a water leakage fault: within the set second discrimination time, judging the number of times that the kurtosis index feature calculated based on the multiple photon number time series slice data of the photon counter is greater than or equal to the kurtosis index discrimination limit, recorded as the second number, if the second number is greater than or equal to the second discrimination number limit, it is judged that the photon counter detects a water leakage fault; The discrimination method based on EMD decomposition characteristics includes the following steps: Slice the photon number time series data[ x 1 ,x 2 ,…,x N ] to perform EMD decomposition and obtain the corresponding IMF components of each order ( c 1 ( x ), c 2 ( x ),…, c m ( x )) T ,in, N Indicates the number of data in the slice, c i ( x ) indicates the i The order IMF component, m The photon number time series slice data is divided into m IMF components; Set the EMD decomposition discrimination index and EMD decomposition discrimination index limit: Take the first-order IMF component data after EMD decomposition of the photon number time series slice data c 1 ( x ), obtain the first-order IMF component data c 1 ( x ) ,average value and standard deviation , the first-order IMF component data of the photon number time series data c 1 ( x ) As the EMD decomposition indicator, the first-order IMF component data of the photon number time series data isc 1 ( x ) and 3 times the standard deviation The sum is used as the EMD decomposition judgment index limit ,Right now ; Set the third discrimination time and the third discrimination number limit and judge the water leakage fault: within the set third discrimination time, judge the EMD decomposition discrimination index calculated based on the multiple photon number time series slice data of the photon counter Greater than or equal to the EMD decomposition judgment index limit The number of times is recorded as the third number. If the third number is greater than or equal to the third discrimination number limit, it is determined that the photon counter detects a water leakage fault; The method for distinguishing the spectrum energy characteristics based on the sensitive band synchronous compression wavelet transform includes the following steps: Based on the time series slice data of each photon number of a photon counter, the time-frequency spectrum characteristics of the synchronous compression wavelet transform represented by a two-dimensional matrix are calculated: , in, T and F They represent the maximum number of numbers in the time and frequency dimensions respectively. The first dimension of the spectrum obtained by synchronous compression wavelet transform is time, and the second dimension is frequency. Set the sensitive frequency band and complete the spectrum capture: set the lower limit of the sensitive frequency band and upper limit , and the spectral feature matrix of synchronous compression wavelet transform is intercepted according to the sensitive frequency band , that is, only the synchronous compression wavelet transform time-frequency spectrum within the sensitive frequency band is retained; The truncated synchronous compression wavelet transform time spectrum is accumulated in the frequency dimension to obtain the sensitive frequency band energy time series data ,in , obtain the sensitive frequency band energy time series data The maximum value in ,average value and standard deviation , the sensitive frequency band energy time series data The maximum value of As an energy feature discrimination index, the average value and 3 times the standard deviation The sum is used as the limit value of the time-frequency energy characteristic discrimination index ,Right now ; Set the fourth discrimination time and the fourth discrimination number limit and judge the water leakage fault: within the set fourth discrimination time, judge the discrimination index of the energy characteristic of the sensitive frequency band of the synchronous compression wavelet transform calculated based on the multiple photon number time series slice data of the photon counter Greater than or equal to the time-frequency energy characteristic discrimination index limit The number of times is recorded as the fourth number. If the fourth number is greater than or equal to the fourth discrimination number limit, it is determined that the photon counter detects a water leakage fault; The intercepted synchronous compression wavelet transform time spectrum is processed with pseudo color and sent to the pre-trained convolutional neural network model for discrimination, and the water leakage monitoring discrimination result for a single photon number time series data slice is obtained; Set the fifth judgment time and the fifth judgment number limit and judge the water leakage fault: within the set fifth judgment time, judge that the output result of the neural network model calculated based on the multiple photon number time series slice data of the photon counter is the number of times the water leakage fault occurs, recorded as the fifth number; if the fifth number is greater than or equal to the fifth judgment number limit, it is determined that the photon counter has detected a water leakage fault.

[0015] The method of summarizing the discrimination results of a single photon counter using multiple feature discrimination methods and judging whether the corresponding photon counter has detected a water leakage fault according to the weight summation rule is as follows: For the multidimensional feature discrimination method of multiple selected one-dimensional time series, the weights of the discrimination results of different features are set respectively. , and guarantee , P is the number of discrimination results obtained by using multiple feature discrimination methods; For the discrimination result output by a certain feature discrimination method, if the result is a fault detected, the discrimination result is set to 1, and if the result is no fault, the result is set to 0. The weighted sum is performed based on the set weight and the discrimination result output by each feature discrimination method to obtain the summary discrimination result; A limit value for the water leakage fault monitoring judgment result of a single photon counter is set. If the summary judgment result is greater than the limit value for the water leakage fault monitoring judgment result of a single photon counter, the water leakage monitoring result of the photon counter is considered to be faulty.

[0016] A converter station valve hall water leakage monitoring system based on a photon counter is used to implement the method described above. The system includes a line laser, multiple photon counters, a control unit, a data processing unit, a communication unit and a human-machine interface, wherein: The line laser is used to realize non-interval scanning and irradiation of the water leakage area to be detected; The photon counter is used to collect the changes in the number of photons in the leaking area to be detected, and obtain the time series data of the number of photons; The control unit is used to realize the coordinated control of the line laser, the photon counter, and the data processing unit; The data processing unit is used to slice and analyze the photon number time series data collected by the photon counter, and output the water leakage monitoring and discrimination result; The communication unit is used to realize the communication inside and outside the converter station valve hall water leakage monitoring system; The human-machine interface is used to realize the interaction between personnel and the valve hall water leakage monitoring system of the converter station.

[0017] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention proposes a method for monitoring water leakage in the valve hall of a converter station based on a line laser and a photon counter. Compared with the prior art, the present invention innovatively adopts a photon counter as a sensing unit, which expands the sensing dimension compared with the prior art and supplements the existing data dimension. The technology of the present invention can enrich the data used for water leakage monitoring and optimize the monitoring effect of water leakage in the valve hall of a converter station.

[0018] (2) The present invention collects one-dimensional photon count time series data through a photon counter, and uses simple numerical type features such as root mean square value, kurtosis index, maximum value of EMD decomposition Imf component, etc. as discrimination indicators, which can achieve a faster discrimination speed in data judgment than the existing technology.

[0019] (3) The present invention adopts the energy characteristics of the synchronous compression wavelet transform time spectrum as the judgment method of the key characteristics, enriches the application depth of the photon number time series data slice, and further improves the accuracy of water leakage monitoring.

[0020] (4) The present invention proposes a configuration method of a line laser and a photon counter for engineering purposes. Compared with the prior art, the method fully considers the engineering conditions and application effects, and can guide the implementation of the technology of the present invention in engineering.

[0021] (5) The present invention can achieve effective monitoring of valve hall water leakage in a converter station without acquiring a large amount of raw data, and has greater practicality and actual application value than the prior art.

[0022] (6) The data processing method of the present invention is simple, does not need to rely on the performance of the computing device, has low requirements on the computing device, and is easier to deploy in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flow chart of the method of the present invention; Figure 2 A schematic top view of a configuration scheme of lasers and photon counters for a water leakage monitoring line in a converter station valve hall in an embodiment; Figure 3A schematic side view of a configuration scheme of a laser and a photon counter for a water leakage monitoring line in a converter station valve hall in an embodiment; Figure 4 A flow chart of a water leakage monitoring method using a single photon counter in one embodiment; Figure 5 A schematic diagram of the structure of a pre-trained convolutional neural network for processing the time-frequency spectrum of sensitive frequency bands of photon number time-series slice data in one embodiment; Figure 6 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0024] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0025] Example 1 Firstly, the present invention provides a method for monitoring water leakage in the valve hall of a converter station by combining a photon counter and a line laser. The line laser is used to illuminate the water leakage detection area in the space, and the photon counter is used to capture photons of a specific frequency band in the space. The fluctuation of the number of photons is analyzed to determine whether there is a water leakage fault in the space, thereby realizing rapid identification and monitoring of water leakage faults.

[0026] Specifically, Figure 1 As shown, the following steps are included: Step 1) A line laser is set in the area to be detected for water leakage and the line laser is used to illuminate the area to be detected for water leakage in the converter station valve hall.

[0027] In this embodiment, the line laser has a certain fan angle and effective distance, and the laser track it emits is a fan-shaped column with a certain thickness. It must be ensured that the laser track emitted by the line laser covers the leak area to be detected. The wavelength of the laser emitted by the line laser is distributed in a narrow range. Preferably, the difference between the upper and lower limits of the laser wavelength emitted by the line laser does not exceed 50nm.

[0028] Step 2) Set up multiple photon counters in the area to be detected for water leakage.

[0029] The photon counter is set up to measure the number of photons received within a fixed wavelength range and transmit this data to external devices and systems. The measurable wavelength range of the photon counter must overlap with the wavelength range of the laser emitted by the line laser.

[0030] Step 3) Slice the number of photons collected by the photon counter in the form of time series data to obtain photon number time series data slices.

[0031] The obtained photon number time series data slices are organized according to a certain sampling rate and have a one-dimensional array of a certain length. At the same time, the time series data slices also include the start time and sampling rate information of the time series data.

[0032] In this embodiment, when executing photon number time series data slicing, there is a certain overlap rate between adjacent time series data slices. This can increase the length of a single time series data and improve the accuracy of subsequent analysis on the one hand, and reduce the analysis time interval and increase the calculation frequency of water leakage fault monitoring on the other hand.

[0033] Specifically, the photon number time series data slice length used in this embodiment is 1 s, and the sampling frequency is 100 kHz.

[0034] Step 4) Analyze the photon number time series data slices collected by each photon counter respectively and obtain the water leakage monitoring and discrimination result of a single photon counter.

[0035] like Figure 4 As shown, the specific steps include: Step 41) setting a set of multidimensional feature discrimination methods for one-dimensional time series, wherein the set of multidimensional feature discrimination methods includes discrimination methods based on root mean square value, kurtosis index, EMD decomposition feature and spectral energy feature of sensitive band synchronous compression wavelet transform.

[0036] The root mean square value-based discrimination method includes the following steps: A1, based on the time series slice data of each photon number of a photon counter, the root mean square value feature is extracted, and the calculation method is: , in, x i is the data value, N The amount of data.

[0037] According to this calculation method, a single photon number time series slice data can extract a numeric value root mean square value feature. For multiple photon number time series slice data, multiple ordered numeric value root mean square value features can be extracted.

[0038] A2, set the RMS value judgment limit.

[0039] In this embodiment, the RMS value discrimination limit is set to the baseline data mean plus three times the baseline data standard deviation, where the baseline data is a plurality of RMS value features corresponding to a plurality of photon number time series slice data within a preset time, that is, .in, , They are the average values ​​and standard deviations of multiple root mean square values ​​corresponding to multiple photon number time series slice data within a period of time.

[0040] A3, set the first judgment time and the first judgment number limit and judge the water leakage fault.

[0041] Within the set first discrimination time, it is determined that the root mean square value feature calculated based on the multiple photon number time series slice data of the photon counter is greater than or equal to the root mean square value discrimination limit (i.e. ) is recorded as the first number. If the first number is greater than or equal to the first judgment number limit, it is determined that the photon counter has detected a water leakage fault.

[0042] Among them, the discrimination method based on the kurtosis index includes the following steps: B1, based on each photon number time series slice data of a photon counter, the kurtosis index feature is extracted, and the calculation method is: , in, Indicates the average value.

[0043] According to this calculation method, a single photon number time series slice data can extract a numerical type of kurtosis index feature. For multiple photon number time series slice data, multiple ordered numerical type kurtosis index features can be extracted.

[0044] B2, set the kurtosis index judgment limit.

[0045] In this embodiment, the kurtosis index determination limit is set to the baseline data mean plus three times the baseline data standard deviation, wherein the baseline data is a plurality of kurtosis index features corresponding to a plurality of photon number time series slice data within a preset time; that is, .in, , They are the average values ​​and standard deviations of multiple kurtosis indicators corresponding to multiple photon number time series slice data within a period of time.

[0046] B3, set the second judgment time and the second judgment number limit and judge the water leakage fault.

[0047] Within the set second discrimination time, it is determined that the kurtosis index feature calculated based on the multiple photon number time series slice data of the photon counter is greater than or equal to the kurtosis index discrimination limit value (i.e. ) is recorded as the second number. If the second number is greater than or equal to the second judgment number limit, it is determined that the photon counter has detected a water leakage fault.

[0048] Among them, the discrimination method based on EMD decomposition characteristics includes the following steps: C1, slice data of photon number time series [ x 1,x 2 ,…,x N ] to perform EMD decomposition and obtain the corresponding IMF components of each order ( c 1 ( x ), c 2 ( x ),…, c m ( x )) T ,in, N Indicates the number of data in the slice, c i ( x ) indicates the i The order IMF component, m The photon number time series slice data is divided into m According to this calculation method, the single photon number time series slice data can be calculated to obtain m IMF component data.

[0049] C2, set the EMD decomposition discrimination index and the EMD decomposition discrimination index limit.

[0050] Take the first-order IMF component data after EMD decomposition of the photon number time series slice data c 1 ( x ), obtain the first-order IMF component data c 1 ( x ) ,average value and standard deviation , the first-order IMF component data of the photon number time series data c 1 ( x ) As the EMD decomposition indicator, the first-order IMF component data of the photon number time series data is c 1 ( x ) and 3 times the standard deviation The sum is used as the EMD decomposition judgment index limit ,Right now .

[0051] It should be noted that, unlike the root mean square value and kurtosis index feature methods, the discrimination index limit of the EMD decomposition feature is calculated completely independently for each photon number time series slice data, that is, each independent photon number time series slice data has an independent EMD decomposition feature discrimination index limit.

[0052] C3, set the third judgment time and the third judgment number limit and judge the water leakage fault.

[0053] Within the set third discrimination time, the EMD decomposition discrimination index calculated based on the multiple photon number time series slice data of the photon counter is judged. Greater than or equal to the EMD decomposition judgment index limit (Right now ) is recorded as the third number. If the third number is greater than or equal to the third judgment number limit, it is determined that the photon counter has detected a water leakage fault.

[0054] The method for distinguishing the spectrum energy characteristics based on the synchronous compression wavelet transform of the sensitive frequency band includes the following steps: D1, synchronous compression wavelet transform is an improved wavelet transform method, which can squeeze the wavelet factor spectrum along the scale axis and make the energy distribution of the time-frequency spectrum more concentrated. Therefore, the synchronous compression wavelet transform time-frequency spectrum has higher concentration and resolution than the general wavelet factor spectrum. Based on the time series slice data of each photon number of a photon counter, the synchronous compression wavelet transform time-frequency spectrum characteristics represented by a two-dimensional matrix are calculated: , in, T and F They represent the maximum number of numbers in the time and frequency dimensions respectively. The first dimension of the spectrum of the obtained synchronous compression wavelet transform is time, and the second dimension is frequency.

[0055] D2, set the sensitive frequency band and complete the spectrum capture.

[0056] The significance of setting the sensitive frequency band is that the change in the number of photons will cause a strong change in the time-frequency spectrum within a specific frequency band. After setting the sensitive frequency band, the change in the number of photons can be detected more sensitively and accurately. and upper limit , and the spectral feature matrix of synchronous compression wavelet transform is intercepted according to the sensitive frequency band , that is, only the synchronous compressed wavelet transform time-frequency spectrum within the sensitive frequency band is retained.

[0057] D3, accumulate the truncated synchronous compression wavelet transform time spectrum in the frequency dimension to obtain the sensitive frequency band energy time series data ,in , obtain the sensitive frequency band energy time series data The maximum value in ,average value and standard deviation , the sensitive frequency band energy time series data The maximum value of As an energy feature discrimination index, the average value and 3 times the standard deviation The sum is used as the limit value of the time-frequency energy characteristic discrimination index ,Right now .

[0058] The synchronous compression wavelet transform sensitive frequency band energy characteristic discrimination index limit is calculated completely independently for each photon number time series slice data, that is, each independent photon number time series slice data has an independent synchronous compression wavelet transform sensitive frequency band energy discrimination index limit.

[0059] D4, set the fourth judgment time and the fourth judgment number limit and judge the water leakage fault.

[0060] Within the set fourth discrimination time, the synchronous compression wavelet transform sensitive band energy characteristic discrimination index calculated based on the multiple photon number time series slice data of the photon counter is judged. Greater than or equal to the time-frequency energy characteristic discrimination index limit (Right now ) is recorded as the fourth number. If the fourth number is greater than or equal to the fourth judgment number limit, it is determined that the photon counter has detected a water leakage fault.

[0061] D5, the intercepted synchronous compressed wavelet transform time spectrum is processed with pseudo color and sent to the pre-trained convolutional neural network model for discrimination, and the water leakage monitoring discrimination result for the single photon number time series data slice is obtained.

[0062] like Figure 5 The convolutional neural network model structure used in one embodiment is shown. Other structures may also be used in other embodiments, and this embodiment does not make specific restrictions on this.

[0063] D6, set the fifth judgment time and the fifth judgment number limit and judge the water leakage fault.

[0064] Within the set fifth judgment time, the output result of the neural network model calculated based on the multiple photon number time series slice data of the photon counter is judged as the number of water leakage failures, which is recorded as the fifth number. If the fifth number is greater than or equal to the fifth judgment number limit, it is determined that the photon counter has detected a water leakage failure.

[0065] Step 42) Select multiple feature discrimination methods from the multi-dimensional feature discrimination method set, calculate corresponding features based on the selected feature discrimination methods, and determine whether a water leakage fault occurs based on the calculated features.

[0066] In this embodiment, the five results obtained by the above four methods are used to make judgments. For all the different feature judgment methods of the single photon counter, a judgment time of 2 minutes and a number limit of 5 times are uniformly adopted.

[0067] The wavelet transform time-frequency spectrum of the sensitive band is synchronously compressed, and the pseudo-color processing is performed according to the jet color map and then scaled to 512×512 size. Figure 5 The convolutional neural network model in is used to realize water leakage identification, and the sensitive frequency band is set to 10kHz~20kHz.

[0068] For the results of deep learning judgment of these five root mean square values, kurtosis index, EMD decomposition characteristics, spectral energy characteristics of sensitive band synchronous compression wavelet transform, and spectrum deep learning judgment of sensitive band synchronous compression wavelet transform, the weights are set as W=[0.2, 0.2, 0.2, 0.1, 0.3] respectively.

[0069] Step 43) Summarize the discrimination results of multiple feature discrimination methods for a single photon counter, and determine whether the corresponding photon counter detects a water leakage fault according to a weighted summation rule.

[0070] Specifically: For the multidimensional feature discrimination method of multiple selected one-dimensional time series, the weights of the discrimination results of different features are set respectively. , and guarantee , P is the number of discrimination results obtained by using multiple feature discrimination methods; For the discrimination result output by a certain feature discrimination method, if the result is a fault detected, the discrimination result is set to 1, and if the result is no fault, the result is set to 0. The weighted sum is performed based on the set weight and the discrimination result output by each feature discrimination method to obtain the summary discrimination result; Set the limit value of the single photon counter water leakage fault monitoring judgment result , this limit The value is between 0 and 1. If the summary judgment result is greater than the single photon counter leakage fault monitoring judgment result limit , then the water leakage monitoring result of the photon counter is considered to be faulty.

[0071] Step 5) obtaining a water leakage monitoring identification result of the water leakage area to be detected based on the water leakage monitoring identification result of a single photon counter.

[0072] A K / N voting rule is set, that is, within a certain time range, if the water leakage fault monitoring results provided by more than K photon counters among N photon counters are true, it is considered that a water leakage fault has occurred in the water leakage area to be detected.

[0073] Example 2 Based on Example 1, this embodiment provides an installation configuration scheme of a line laser and a photon counter for regional water leakage monitoring. By properly arranging the line laser and the photon counter, water leakage fault monitoring in a specific area of ​​the converter valve hall is achieved.

[0074] In this configuration of line laser and photon counter, the configuration goal is to achieve regional water leakage monitoring coverage with fewer photon counters. The water leakage area to be detected in this embodiment is a square specific area in space, that is, this area has a horizontal boundary. The configuration method includes two key contents: configuring line laser and configuring photon counter.

[0075] (1) Configuration of line laser The setup of a line laser consists of the following steps: S101, determine the area to be detected and the location where the line laser can be installed: according to the site conditions, determine the horizontal boundary of the area to be detected for water leakage and the location range where the line laser can be installed. The location where the line laser can be installed is generally on the wall or frame of the building, represented by one or more line segments.

[0076] S102, determine the installation position of the line laser: Figure 2 As shown, make the short side of the rectangular leakage area to be detected b The perpendicular bisector of forms an intersection point O with the line segment describing the position where the line laser can be installed, and the intersection point O is used as the position where the line laser is installed.

[0077] S103, determine the fan angle and effective distance of the line laser: the parameters of the line laser include the fan angle and effective distance L , the parameters of the line laser meet the following conditions: 1) effective distance L It must exceed the maximum value of the distance from the line laser installation position O to all positions in the leak area to be detected; 2) Laser fan angle The effective working sector range of the line laser covers the rectangular leak area to be detected, that is ,in, b is the length of the short side of the rectangular leakage area to be detected, d The line laser installation position O is the short side of the rectangular leak area to be detected. b distance.

[0078] S104, determine the installation angle of the line laser: set the vertical installation angle of the line laser to be horizontal, that is, the plane of the fan-shaped column formed by the laser trajectory is parallel to the horizontal plane, and set the horizontal installation angle of the line laser to be perpendicular to the leaking area to be detected, that is, the angle bisector of the fan-shaped cross section of the fan-shaped column formed by the laser trajectory is parallel to the short side of the rectangular area to be detected. b The perpendicular bisectors coincide.

[0079] In a preferred embodiment, Figure 2 , Figure 3 As shown in the figure, a line laser and a photon counter are installed on the wall facing the converter valve area. The distance between the line laser and the leak area to be detected is d =6m. The wavelength range of the line laser used is 635±15nm. The wavelength range of the photon counter used is 635±5nm. The photon counter can detect wavelengths less than or equal to the wavelength range of the line laser. The effective working range of the line laser used is L=30m, and the laser fan angle is =30°.

[0080] In the embodiment provided by the present invention, the photon number time series data slice length used is 1 s and the sampling frequency is 100 kHz.

[0081] (2) Configuration of photon counter The configuration of the photon counter consists of the following steps: S201, determine the location where the photon counter can be installed: determine the location where the photon counter can be installed according to the site conditions. The location where the photon counter can be installed is on the wall or frame of the building, represented by one or more line segments, and is consistent with the horizontal position where the line laser can be installed.

[0082] S202, based on the available installation positions of the photon counters, multiple photon counters are configured in the horizontal direction, and the receiving probes of all the photon counters are oriented toward the leaking area to be detected. In the vertical direction, the installation height of the photon counter should have a certain height difference h with the line laser, either higher than the line laser, or lower than the line laser, such as Figure 3 In this embodiment, the height difference h between the photon counter and the line laser must be no less than 0.5 m. In addition, the installation angle of the photon counter is set to be parallel to the horizontal plane.

[0083] In a preferred embodiment, Figure 2 , Figure 3 As shown, a total of 8 photon counters are installed on the wall. The height difference between the photon counter and the line laser is h=0.5m. The 8 photon counters used are installed in the form of 4 on the upper and lower layers.

[0084] Example 3 This embodiment provides a converter station valve hall water leakage monitoring system based on a photon counter, which is used to implement the method described in the above embodiment 1. Figure 6 As shown, the system includes a line laser, a plurality of photon counters, a control unit, a data processing unit, a communication unit and a human-machine interface, wherein: Line laser is used to realize non-interval scanning and irradiation of the leaking area to be detected; The photon counter is used to collect the changes in the number of photons in the leaking area to be detected and obtain the time series data of the number of photons; The control unit is used to realize the coordinated control of the line laser, the photon counter, and the data processing unit; The data processing unit is used to slice and analyze the photon number time series data collected by the photon counter, and output the water leakage monitoring and judgment results; The communication unit is used to realize the communication within and with the outside of the converter station valve hall water leakage monitoring system; The human-machine interface is used to realize the interaction between personnel and the valve hall leakage monitoring system of the converter station.

[0085] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the described unit / module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0086] In this embodiment, a common microcomputer is used as the control unit and the calculation unit. The communication between the microcomputer and the line laser is realized by RS232 port, and the communication with the photon counter is realized by RJ45 port Ethernet. A graphical software program is used in combination with a display, a mouse, and a keyboard to realize human-computer interaction.

[0087] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning, or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A method for monitoring water leakage in a converter station valve hall based on a photon counter, characterized in that: The following steps are involved: A line laser and multiple photon counters are set in the leakage area to be detected, and the line laser is used to illuminate the leakage area to be detected in the valve hall of the converter station, and the number of photons received by the photon counter within a fixed wavelength range is measured. The measurable wavelength range of the photon counter overlaps with the wavelength range of the laser emitted by the line laser; Slice the number of photons collected by the photon counter in the form of time series data to obtain photon number time series data slices; The photon number time series data slices collected by each photon counter are analyzed respectively to obtain the water leakage monitoring and discrimination result of a single photon counter, and the water leakage monitoring and discrimination result of the water leakage area to be detected is obtained based on the water leakage monitoring and discrimination result of the single photon counter.

2. According to claim 1, a method for monitoring water leakage in a converter station valve hall based on a photon counter is characterized in that: The laser track emitted by the set line laser is a fan-shaped column with a preset thickness, and the laser track covers the water leakage area to be detected; When the water leakage area to be detected is a rectangular area, the setting of the line laser includes the following steps: Determine the area to be detected and the location where the line laser can be installed: according to the site conditions, determine the horizontal boundary of the area to be detected for water leakage and the location range where the line laser can be installed. The location where the line laser can be installed is on the wall or frame of the building, represented by one or more line segments; Determine the installation position of the line laser: draw a perpendicular midline of the short side of the rectangular leak area to be detected, and form an intersection with the line segment describing the position where the line laser can be installed, and use the intersection as the position for installing the line laser; Determine the fan angle and effective distance of the line laser: The parameters of the line laser include the fan angle and effective distance L , the parameters of the line laser meet the following conditions: 1) effective distance L The maximum value of the distance from the installation position of the line laser to all positions in the leak area to be detected; 2) Laser fan angle The effective working sector range of the line laser covers the rectangular leak area to be detected, that is, ,in, b is the length of the short side of the rectangular leak area to be detected, d The distance between the installation position of the line laser and the short side of the rectangular leak area to be detected; Determine the installation angle of the line laser: set the vertical installation angle of the line laser to horizontal, that is, the plane of the fan-shaped column formed by the laser trajectory is parallel to the horizontal plane, and at the same time set the horizontal installation angle of the line laser to be perpendicular to the leaking area to be detected, that is, the angle bisector of the fan-shaped section of the fan-shaped column formed by the laser trajectory coincides with the perpendicular bisector of the short side of the rectangular area to be detected.

3. According to claim 1, a method for monitoring water leakage in a converter station valve hall based on a photon counter is characterized in that: The setting of the photon counter comprises the following steps: Determine the installable position of the photon counter: determine the installable position of the photon counter according to the site conditions. The installable position of the photon counter is on the wall or frame of the building, represented by one or more line segments, and is consistent with the horizontal position of the installable line laser; Based on the installable positions of the photon counters, multiple photon counters are configured in the horizontal direction, with the receiving probes of all photon counters facing the leakage area to be detected; in the vertical direction, the height difference between the installation height of the photon counter and the line laser is set to be greater than a preset threshold; the installation angle of the photon counter is set to be parallel to the horizontal plane.

4. According to claim 1, a method for monitoring water leakage in a converter station valve hall based on a photon counter is characterized in that: The difference between the upper and lower limits of the laser wavelength emitted by the line laser does not exceed 50nm.

5. The method for monitoring water leakage in the valve hall of a converter station based on a photon counter according to claim 1 is characterized in that: The photon counter can measure a wavelength range that is less than or equal to the wavelength range of the laser emitted by the line laser.

6. The method for monitoring water leakage in the valve hall of a converter station based on a photon counter according to claim 1 is characterized in that: When slicing the time series data, adjacent time series data slices are made to have partially overlapping data according to a preset overlap rate.

7. The method for monitoring water leakage in the valve hall of a converter station based on a photon counter according to claim 1 is characterized in that: The photon number time series data slice includes a one-dimensional time series array with a preset length, organized according to a preset sampling rate, the start time of the time series data, and the sampling rate information; The method of respectively analyzing the photon number time series data slices collected by each photon counter and obtaining the water leakage monitoring and identification result of a single photon counter specifically includes the following steps: A set of multidimensional feature discrimination methods for a one-dimensional time series is set, wherein the set of multidimensional feature discrimination methods includes a discrimination method based on a root mean square value, a kurtosis index, an EMD decomposition feature, and a spectrum energy feature of a sensitive band synchronous compression wavelet transform; Selecting multiple feature discrimination methods from a set of multidimensional feature discrimination methods, calculating corresponding features based on the selected feature discrimination methods, and judging whether a water leakage fault occurs according to the calculated features; The discrimination results of multiple feature discrimination methods for a single photon counter are summarized, and the weighted summation rule is used to determine whether the corresponding photon counter has detected a water leakage fault.

8. The method for monitoring water leakage in the valve hall of a converter station based on a photon counter according to claim 7 is characterized in that: The discrimination method based on the root mean square value includes the following steps: Extract the root mean square value feature based on each photon number time series slice data of a certain photon counter; The root mean square value discrimination limit is set to be the baseline data mean plus three times the baseline data standard deviation, wherein the baseline data is a plurality of root mean square value features corresponding to a plurality of photon number time series slice data within a preset time; Setting a first discrimination time and a first discrimination number limit and discriminating a water leakage fault: within the set first discrimination time, judging the number of times that the root mean square value feature calculated based on multiple photon number time series slice data of the photon counter is greater than or equal to the root mean square value discrimination limit, which is recorded as the first number; if the first number is greater than or equal to the first discrimination number limit, it is judged that the photon counter detects a water leakage fault; The discrimination method based on kurtosis index includes the following steps: Extract the kurtosis index feature based on each photon number time series slice data of a certain photon counter; The kurtosis index determination limit is set to be the baseline data mean plus three times the baseline data standard deviation, wherein the baseline data is a plurality of kurtosis index features corresponding to a plurality of photon number time series slice data within a preset time; Setting a second discrimination time and a second discrimination number limit and discriminating a water leakage fault: within the set second discrimination time, judging the number of times that the kurtosis index feature calculated based on the multiple photon number time series slice data of the photon counter is greater than or equal to the kurtosis index discrimination limit, recorded as the second number, if the second number is greater than or equal to the second discrimination number limit, it is judged that the photon counter detects a water leakage fault; The discrimination method based on EMD decomposition characteristics includes the following steps: Slice the photon number time series data[ x 1 ,x 2 ,…,x N ] to perform EMD decomposition and obtain the corresponding IMF components of each order ( c 1 ( x ), c 2 ( x ),…, c m ( x )) T ,in, N Indicates the number of data in the slice, c i ( x ) indicates the i The order IMF component, m The photon number time series slice data is divided into m IMF components; Set the EMD decomposition discrimination index and EMD decomposition discrimination index limit: Take the first-order IMF component data after EMD decomposition of the photon number time series slice data c 1 ( x ), obtain the first-order IMF component data c 1 ( x ) ,average value and standard deviation , the first-order IMF component data of the photon number time series data c 1 ( x ) As the EMD decomposition indicator, the first-order IMF component data of the photon number time series data is c 1 ( x ) and 3 times the standard deviation The sum is used as the EMD decomposition judgment index limit ,Right now ; Set the third discrimination time and the third discrimination number limit and judge the water leakage fault: within the set third discrimination time, judge the EMD decomposition discrimination index calculated based on the multiple photon number time series slice data of the photon counter Greater than or equal to the EMD decomposition judgment index limit The number of times is recorded as the third number. If the third number is greater than or equal to the third discrimination number limit, it is determined that the photon counter detects a water leakage fault; The method for distinguishing the spectrum energy characteristics based on the sensitive band synchronous compression wavelet transform includes the following steps: Based on the time series slice data of each photon number of a photon counter, the time-frequency spectrum characteristics of the synchronous compression wavelet transform represented by a two-dimensional matrix are calculated: ， in, T and F They represent the maximum number of numbers in the time and frequency dimensions respectively. The first dimension of the spectrum obtained by synchronous compression wavelet transform is time, and the second dimension is frequency. Set the sensitive frequency band and complete the spectrum capture: set the lower limit of the sensitive frequency band and upper limit , and the spectral feature matrix of synchronous compression wavelet transform is intercepted according to the sensitive frequency band , that is, only the synchronous compression wavelet transform time-frequency spectrum within the sensitive frequency band is retained; The truncated synchronous compression wavelet transform time spectrum is accumulated in the frequency dimension to obtain the sensitive frequency band energy time series data ,in , obtain the sensitive frequency band energy time series data The maximum value in ,average value and standard deviation , the sensitive frequency band energy time series data The maximum value of As an energy feature discrimination index, the average value and 3 times the standard deviation The sum is used as the limit value of the time-frequency energy characteristic discrimination index ,Right now ; Set the fourth discrimination time and the fourth discrimination number limit and judge the water leakage fault: within the set fourth discrimination time, judge the discrimination index of the energy characteristic of the sensitive frequency band of the synchronous compression wavelet transform calculated based on the multiple photon number time series slice data of the photon counter Greater than or equal to the time-frequency energy characteristic discrimination index limit The number of times is recorded as the fourth number. If the fourth number is greater than or equal to the fourth discrimination number limit, it is determined that the photon counter detects a water leakage fault; The intercepted synchronous compression wavelet transform time spectrum is processed with pseudo color and sent to the pre-trained convolutional neural network model for discrimination, and the water leakage monitoring discrimination result for a single photon number time series data slice is obtained; Set the fifth judgment time and the fifth judgment number limit and judge the water leakage fault: within the set fifth judgment time, judge that the output result of the neural network model calculated based on the multiple photon number time series slice data of the photon counter is the number of times the water leakage fault occurs, recorded as the fifth number; if the fifth number is greater than or equal to the fifth judgment number limit, it is determined that the photon counter has detected a water leakage fault.

9. The method for monitoring water leakage in the valve hall of a converter station based on a photon counter according to claim 7 is characterized in that: The method of summarizing the discrimination results of a single photon counter using multiple feature discrimination methods and judging whether the corresponding photon counter has detected a water leakage fault according to the weight summation rule is as follows: For the multidimensional feature discrimination method of multiple selected one-dimensional time series, the weights of the discrimination results of different features are set respectively. , and guarantee , P is the number of discrimination results obtained by using multiple feature discrimination methods; For the discrimination result output by a certain feature discrimination method, if the result is a fault detected, the discrimination result is set to 1, and if the result is no fault, the result is set to 0. The weighted sum is performed based on the set weight and the discrimination result output by each feature discrimination method to obtain the summary discrimination result; A limit value for the water leakage fault monitoring judgment result of a single photon counter is set. If the summary judgment result is greater than the limit value for the water leakage fault monitoring judgment result of a single photon counter, the water leakage monitoring result of the photon counter is considered to be faulty.

10. A converter station valve hall water leakage monitoring system based on a photon counter, used to implement the method according to any one of claims 1 to 9, characterized in that: The system includes a line laser, a plurality of photon counters, a control unit, a data processing unit, a communication unit and a human-machine interface, wherein: The line laser is used to realize non-interval scanning and irradiation of the water leakage area to be detected; The photon counter is used to collect the changes in the number of photons in the leaking area to be detected, and obtain the time series data of the number of photons; The control unit is used to realize the coordinated control of the line laser, the photon counter, and the data processing unit; The data processing unit is used to slice and analyze the photon number time series data collected by the photon counter, and output the water leakage monitoring and discrimination result; The communication unit is used to realize the communication inside and outside the converter station valve hall water leakage monitoring system; The human-machine interface is used to realize the interaction between personnel and the valve hall water leakage monitoring system of the converter station.

Images