A method, device, equipment and storage medium for monitoring water leakage fault of a converter valve

Abstract

The application provides a converter valve water leakage fault monitoring method, device, equipment and storage medium, comprising: acquiring laser water leakage visualization time sequence image data in a to-be-judged period; pre-processing the laser water leakage visualization time sequence image data to obtain binary time sequence image difference result data; and determining whether there is a water leakage fault based on the binary time sequence image difference result data. The application can complete laser water leakage visualization time sequence image processing by using a general computing architecture without increasing the demand for special computing resources such as artificial intelligence chips, and realizes early slight leakage water leakage fault detection of a converter valve.

Description

Technical Field

[0001] The present invention relates to the technical field of power generation and transmission and transformation, and in particular to a method, device, equipment and storage medium for monitoring water leakage faults in converter valves. Background Art

[0002] The converter valve is a core component of power transmission projects, primarily composed of large-scale, multi-stage series-connected power electronic units. During operation, it can continuously flow high currents of up to thousands of amperes, generating significant heat loss in the valve's core power electronic components. To promptly dissipate this heat loss, ensure the core power electronic components maintain their designed operating temperature, and prevent heat loss from affecting the valve's normal operation, a cooling system was specifically designed for the valve. This system uses a water turbine to provide circulating power, and a cooling water circuit to allow cooling water to flow through the heat dissipation surfaces of the power electronic units, rapidly removing heat.

[0003] The internal waterway connections of converter valves utilize polymer sealing rings that achieve sealing through extrusion and deformation. Under long-term engineering stress, the polymer sealing rings deform due to material aging, reducing the extrusion force and degrading sealing performance. Regular maintenance requires disassembly and assembly of the waterway connections, impacting their sealing performance. These two major factors can lead to leakage at the converter valve waterway connections during their service life. Furthermore, converter valve components through which cooling water flows, such as the radiator and valve saturation reactor, can also experience local structural damage and leakage due to corrosion, mechanical vibration, and wear. Leakage in converter valves directly impacts the operational reliability of the equipment itself and, in severe cases, can cause a DC transmission line outage. In the early stages of a converter valve leakage, it typically manifests as a minor seepage. Over time, the leakage becomes increasingly severe, with increasing flow rates, and eventually progresses to a leak, ultimately affecting converter valve operation. Therefore, implementing technical measures to monitor early-stage converter valve leakage is of great engineering significance.

[0004] Currently, there is no technology that can realize the early leakage fault monitoring of the converter valve and put it into engineering application. The flip-type leakage detection device and test paper color change method used in the project cannot detect early leakage faults, and cannot meet the project needs in terms of detection sensitivity and advance time.

[0005] File CN202110680882.0 proposes a system and method for monitoring water leakage in a converter valve based on laser radar. However, since the laser radar scans at an angle, it cannot scan all positions in the space in real time. At the same time, its resolution will decrease rapidly when the distance is far, and it is easy to miss detections. File 1CN202311048594.9 proposes a spatial liquid drop detection system, method and computer equipment, and proposes a method and equipment for using a line laser to make water leakage events in space explicit, and realizing water leakage fault monitoring through image analysis. Due to the use of line laser scanning, it has the ability to scan spatial areas in real time, and the resolution is less affected by distance. Summary of the Invention

[0006] In order to solve the problem that the existing technology cannot detect early leakage faults and cannot meet engineering requirements in terms of detection sensitivity and advance speed, the present invention proposes a method for monitoring and locating water leakage faults in a converter valve, comprising:

[0007] Obtaining laser water leakage explicit time series image data within the time period to be determined;

[0008] Preprocessing the laser water leakage explicit time series image data to obtain binarized time series image difference result data;

[0009] Determine whether there is a water leakage fault based on the binarized time series image difference result data.

[0010] Optionally, the preprocessing of the laser water leakage explicit time series image data to obtain binarized time series image difference result data includes:

[0011] Perform channel conversion on the laser water leakage explicit time series image data to obtain a water leakage time series image represented by a four-dimensional matrix;

[0012] Performing image difference calculation on the water leakage time series image represented by the four-dimensional matrix to obtain a time series image difference result;

[0013] The time series image difference result is subjected to image binarization processing to obtain binarized time series image difference result data.

[0014] Optionally, performing channel conversion on the laser water leakage explicit time series image data to obtain a water leakage time series image represented by a four-dimensional matrix includes:

[0015] Based on the laser water leakage explicit time-series image data format, converting the laser water leakage explicit time-series image data into a plurality of consecutive frames of continuous time-series image data represented by a numerical matrix;

[0016] Converting each frame of the continuous time-series image data into a three-dimensional numerical matrix having multiple color channels;

[0017] A water leakage time series image represented by a four-dimensional matrix is ​​constructed by combining the three-dimensional data matrix with the time series dimension.

[0018] Optionally, the time series image difference result is calculated as follows:

[0019] X fm-s =[x fm(1) ,x fm(2) ,…,x fm(n-s-1) ,x fm(n-s) ]

[0020] =|[x1-x 1+s ,x2-x 2+s ,…,x n-s-1 -x n-1 ,x n-s -x n ]|

[0021] Where, X fm-s is the time series image difference result after the original time series image data X is differentiated according to the step size s, x fm(1) 、x fm(2) 、x fm(n-s-1) 、x fm(n-s) X fm-s The first, second, third, up to the nth element in x fm(i) =x i -x i+s ,i=1,2,…,ns,x1、x 1+s ,x2,x 2+s 、x n-s-1 、x n-1 、x n-s 、x n are elements in the original time series image data X, and n is a positive integer.

[0022] Optionally, performing image binarization processing on the time series image difference result to obtain binarized time series image difference result data includes:

[0023] The elements in the time sequence image difference results that are greater than the set binarization threshold are assigned the maximum value, and the elements in the time sequence image difference results that are not greater than the set binarization threshold are assigned the minimum value, to obtain the binarized time sequence image difference result data.

[0024] Optionally, the determining whether a water leakage fault exists based on the binarized time series image difference result data includes:

[0025] Traverse all binarized time series image difference result data, retain the frames with connected image regions, and form binary difference time series image data containing connected regions;

[0026] Traverse the connected areas corresponding to the binary differential time series image data containing connected areas, calculate the distance between the centroids of each connected area in each frame and other connected areas in the frame, and the distance between the centroids of each connected area in each frame and connected areas in other frames;

[0027] Determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold;

[0028] The water leakage characteristic light spot determines that the converter valve has a water leakage fault.

[0029] Optionally, after determining that the converter valve has a water leakage fault, the method further includes:

[0030] The sampling image processing tool obtains the number of pixels in the connected region and the pixel coordinates in the image;

[0031] The mean or median of the pixel coordinates of the water leakage characteristic spots in all connected areas in the image is taken as the water leakage spot position;

[0032] The valve layer where the water leakage occurs is determined by the position of the water leakage light spot.

[0033] Optionally, the determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold includes:

[0034] Determine whether the centroid distance is less than the set coordinate error threshold. If the centroid distance is less than the set coordinate error threshold, and the number of occurrences of the centroid distance being less than the set coordinate error threshold exceeds the set judgment number threshold, then the connected area is a water leakage characteristic light spot; otherwise, the connected area is a non-water leakage characteristic light spot.

[0035] Optionally, after determining whether a converter valve water leakage fault exists based on the binarized time-series image difference result data, the method further includes:

[0036] The characteristic light spots judged as non-leakage in the multi-period time series image data are compared to monitor and locate long-interval leakage faults.

[0037] Optionally, comparing the characteristic light spots determined as non-leakage in the multi-period time series image data to monitor long-interval water leakage faults includes:

[0038] Traverse all connected regions in multiple time series image data and calculate the distance between the centroid of each connected region and other connected regions;

[0039] Traverse the connected areas corresponding to the binary differential time series image data containing connected areas, calculate the distance between the centroids of each connected area in each frame and other connected areas in the frame, and the distance between the centroids of each connected area in each frame and connected areas in other frames;

[0040] Determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold;

[0041] Based on the water leakage characteristic light spot, it is determined that the converter valve has a long-interval water leakage fault.

[0042] Optionally, after determining that the converter valve has a long-interval water leakage fault, the method further includes:

[0043] The sampling image processing tool obtains the number of pixels in the connected region and the pixel coordinates in the image;

[0044] The mean or median of the pixel coordinates of the water leakage feature spots in all connected areas in the image is taken as the water leakage spot position.

[0045] In another aspect, the present invention further provides a device for monitoring a water leakage fault of a converter valve, comprising:

[0046] A communication unit, used for acquiring laser water leakage explicit time-series image data within a period to be determined;

[0047] The calculation unit is used to pre-process the laser water leakage explicit time sequence image data to obtain binarized time sequence image difference result data; and determine whether the converter valve is leaking based on the binarized time sequence image difference result data.

[0048] Optionally, the calculation unit includes:

[0049] a channel conversion subunit, configured to perform channel conversion on the laser water leakage explicit time series image data to obtain a water leakage time series image represented by a four-dimensional matrix;

[0050] A differential processing subunit is used to perform image differential calculation on the water leakage time series image represented by the four-dimensional matrix to obtain a time series image differential result;

[0051] A binarization processing subunit, configured to perform image binarization processing on the time-series image difference result to obtain binarized time-series image difference result data;

[0052] The positioning processing subunit is used to determine whether the converter valve is leaking based on the binary time series image difference result data.

[0053] Optionally, the channel conversion subunit is specifically configured to:

[0054] Based on the laser water leakage explicit time-series image data format, converting the laser water leakage explicit time-series image data into a plurality of consecutive frames of continuous time-series image data represented by a numerical matrix;

[0055] Converting each frame of the continuous time-series image data into a three-dimensional numerical matrix having multiple color channels;

[0056] A water leakage time series image represented by a four-dimensional matrix is ​​constructed by combining the three-dimensional data matrix with the time series dimension.

[0057] Optionally, the time series image difference result in the difference processing subunit is calculated as follows:

[0058] X fm-s =[x fm(1) ,x fm(2) ,…,x fm(n-s-1) ,x fm(n-s) ]

[0059] =|[x1-x 1+s ,x2-x 2+s ,…,x n-s-1 -x n-1 ,x n-s -x n ]|

[0060] Where, X fm-s is the time series image difference result after the original time series image data X is differentiated according to the step size s, x fm(1) 、x fm(2) 、x fm(n-s-1) 、x fm(n-s) X fm-s The first, second, third, up to the nth element in x fm(i) =x i -x i+s ,i=1,2,…,ns,x1、x2、x 2+s 、x n-s-1 、x n-1 、x n-s 、x n are elements in the original time series image data X, and n is a positive integer.

[0061] Optionally, the binarization processing subunit is specifically used to:

[0062] The elements in the time sequence image difference results that are greater than the set binarization threshold are assigned the maximum value, and the elements in the time sequence image difference results that are not greater than the set binarization threshold are assigned the minimum value, to obtain the binarized time sequence image difference result data.

[0063] Optionally, the positioning processing subunit is specifically configured to:

[0064] Traverse all binarized time series image difference result data, retain the frames with connected image regions, and form binary difference time series image data containing connected regions;

[0065] Traverse the connected areas corresponding to the binary differential time series image data containing connected areas, calculate the distance between the centroids of each connected area in each frame and other connected areas in the frame, and the distance between the centroids of each connected area in each frame and connected areas in other frames;

[0066] Determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold;

[0067] The water leakage characteristic light spot determines that the converter valve has a water leakage fault.

[0068] Optionally, the calculation unit further includes a water leakage locating subunit configured to:

[0069] The sampling image processing tool obtains the number of pixels in the connected region and the pixel coordinates in the image;

[0070] The mean or median of the pixel coordinates of the water leakage characteristic spots in all connected areas in the image is taken as the water leakage spot position;

[0071] The valve layer where the water leakage occurs is determined by the position of the water leakage light spot.

[0072] Optionally, the positioning processing subunit determines whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold as follows:

[0073] Determine whether the centroid distance is less than the set coordinate error threshold. If the centroid distance is less than the set coordinate error threshold, and the number of occurrences of the centroid distance being less than the set coordinate error threshold exceeds the set judgment number threshold, then the connected area is a water leakage characteristic light spot; otherwise, the connected area is a non-water leakage characteristic light spot.

[0074] Optionally, the calculation unit further includes: a long-interval water leakage judgment subunit, configured to:

[0075] The characteristic light spots judged as non-leakage in the multi-period time series image data are compared to monitor long-interval leakage faults.

[0076] Optionally, the long-interval water leakage locating subunit is specifically used for:

[0077] Traverse all connected regions in multiple time series image data and calculate the distance between the centroid of each connected region and other connected regions;

[0078] Traverse the connected areas corresponding to the binary differential time series image data containing connected areas, calculate the distance between the centroids of each connected area in each frame and other connected areas in the frame, and the distance between the centroids of each connected area in each frame and connected areas in other frames;

[0079] Determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold;

[0080] Based on the water leakage characteristic light spot, it is determined that the converter valve has a long-interval water leakage fault.

[0081] Optionally, the calculation unit further includes: a long-interval water leakage fault locating subunit, configured to:

[0082] The sampling image processing tool obtains the number of pixels in the connected region and the pixel coordinates in the image;

[0083] The mean or median of the pixel coordinates of the water leakage feature spots in all connected areas in the image is taken as the water leakage spot position.

[0084] In another aspect, the present application further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0085] The memory is used to store one or more programs;

[0086] When the one or more programs are executed by the at least one processor, the above-mentioned method for monitoring water leakage fault of a converter valve is implemented.

[0087] On the other hand, the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed, the method for monitoring a water leakage fault of a converter valve as described above is implemented.

[0088] Compared with the prior art, the present invention has the following beneficial effects:

[0089] The present invention provides a method for monitoring converter valve water leakage faults, comprising: obtaining laser-generated time-series image data for a period to be determined; preprocessing the laser-generated time-series image data to obtain binary time-series image difference results; and determining whether a water leakage fault exists based on the binary time-series image difference results. This method utilizes a general computing architecture to complete the laser-generated time-series image processing without requiring specialized computing resources such as artificial intelligence chips, enabling early detection of minor water leakage faults in converter valves. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 This is a flow chart of a method for monitoring a water leakage fault of a converter valve according to the present invention;

[0091] Figure 2 This is a flow chart of the laser water leakage explicit time-series image data processing and water leakage identification method for a single-stage-multi-stage combined converter valve of the present invention;

[0092] Figure 3 This is a flow chart of the method for processing laser-generated time-series image data and identifying water leakage in a single-stage converter valve according to the present invention;

[0093] Figure 4 This is a flow chart of the method for processing time-series image data and identifying water leakage by laser-assisted water leakage visualization of a multi-stage converter valve according to the present invention;

[0094] Figure 5 This is a schematic structural diagram of a converter valve water leakage fault monitoring device according to the present invention;

[0095] Figure 6 The figure is a schematic structural diagram of an electronic device of the present invention. DETAILED DESCRIPTION

[0096] The present invention proposes a method for monitoring a converter valve water leakage fault, which realizes the identification and determination of a converter valve water leakage fault by processing and calculating laser water leakage explicit time series image data collected by an image acquisition device.

[0097] In order to better understand the present invention, the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0098] Example 1:

[0099] A method for monitoring a water leakage fault of a converter valve, such as Figure 1 Shown, including:

[0100] Step S1: Acquire laser water leakage explicit time series image data within the time period to be determined;

[0101] Step S2: pre-processing the laser water leakage explicit time series image data to obtain binarized time series image difference result data;

[0102] Step S3: determining whether a water leakage fault exists based on the binarized time series image difference result data.

[0103] like Figure 2 As shown, the technical solution provided by the present invention can monitor converter valve leakage events and locate pixel coordinates within the image by processing time-series image data. Furthermore, a single-segment and multi-segment combined time-series image processing method can be used to visualize time-series image data for converter valve laser leakage. This approach ensures real-time leakage monitoring while preventing missed leak detection due to long intervals between repeated leak events.

[0104] Specifically, such as Figure 3 As shown, by transforming and calculating a segment of time-series image data, the water leakage event of the converter valve corresponding to this segment of time-series image data is identified, and the position coordinates of the laser water leakage explicit light spot in the image data are fed back; when it is determined that no water leakage event has occurred in this segment of time-series image data, the water leakage light spot in each frame of the image and its position coordinates in the image data are recorded.

[0105] like Figure 4 As shown, by comparing and calculating the positions of the light spots that are not judged as water leakage features in the processing results of multiple single-segment time-series image data, the long-interval water leakage faults distributed in multiple segments of time-series image data can be identified, avoiding the missed detection of water leakage faults caused by water leakage events that do not contain enough repetitions in one segment of time-series image data.

[0106] The technical solution provided by the present invention is a single-segment and multi-segment combined time-series image data processing method for visualizing time-series image data of laser water leakage in a converter valve. The method is characterized by comprising the steps of data segmentation, single-segment data recognition, and multi-segment data joint recognition.

[0107] According to the preset data length value, the continuous laser water leakage explicit time series image data of a certain length is organized into a data segment in a certain format. The data length value can be in time or in the number of time series data frames.

[0108] By calculating and processing a continuous segment of time-series image data, it determines whether a water leak has occurred, provides feedback on the pixel coordinates of the leaking light spot in the image, and outputs a single-segment data leak fault identification result. The system is characterized by four main steps: acquiring time-series image data, preprocessing time-series image data, identifying water leaks and calculating light spot coordinates, and outputting single-segment data identification results.

[0109] Step S1: Obtaining laser water leakage explicit time series image data within the time period to be determined, including:

[0110] The time-series image data acquisition phase involves acquiring a certain length of raw time-series image data from an image acquisition device and simultaneously determining the format of the raw time-series image data. The raw time-series image data is typically a long, continuous sequence represented by numerical values. The format of the raw time-series image data typically includes information such as the image's length, width, color format, and frame rate. The number of frames in the raw time-series image data should be greater than one.

[0111] Step S2: Preprocessing the laser water leakage explicit time series image data to obtain binarized time series image difference result data, including:

[0112] Perform channel conversion on the laser water leakage explicit time series image data to obtain a water leakage time series image represented by a four-dimensional matrix;

[0113] Performing image difference calculation on the water leakage time series image represented by the four-dimensional matrix to obtain a time series image difference result;

[0114] The time series image difference result is subjected to image binarization processing to obtain binarized time series image difference result data.

[0115] Furthermore, the method includes: performing channel conversion on the laser water leakage explicit time series image data to obtain a water leakage time series image represented by a four-dimensional matrix, including:

[0116] Based on the laser water leakage explicit time-series image data format, converting the laser water leakage explicit time-series image data into a plurality of consecutive frames of continuous time-series image data represented by a numerical matrix;

[0117] Converting each frame of the continuous time-series image data into a three-dimensional numerical matrix having multiple color channels;

[0118] A water leakage time series image represented by a four-dimensional matrix is ​​constructed by combining the three-dimensional data matrix with the time series dimension.

[0119] Furthermore, the time series image difference result is calculated as follows:

[0120] X fm-s =[x fm(1) ,x fm(2) ,…,x fm(n-s-1) ,x fm(n-s) ]

[0121] =|[x1-x 1+s ,x2-x 2+s ,…,x n-s-1 -x n-1 ,xn-s -x n ]|

[0122] Where, X fm-s is the time series image difference result after the original time series image data X is differentiated according to the step size s, x fm(1) 、x fm(2) 、x fm(n-s-1) 、x fm(n-s) X fm-s The first, second, third, up to the nth element in x fm(i) =x i -x i+s ,i=1,2,…,ns. x1,x2,x 2+s 、x n-s-1 、x n-1 、x n-s 、x n are elements in the original time series image data X, and n is a positive integer.

[0123] Furthermore, the performing of image binarization processing on the time series image difference result to obtain binarized time series image difference result data includes:

[0124] The elements in the time sequence image difference results that are greater than the set binarization threshold are assigned the maximum value, and the elements in the time sequence image difference results that are not greater than the set binarization threshold are assigned the minimum value, to obtain the binarized time sequence image difference result data.

[0125] Step S2 specifically includes: time series image data preprocessing, including the following main steps:

[0126] (1) Image channel transformation.

[0127] 1) First, the original data is converted into time-series image data of continuous images represented by a numerical matrix for a number of consecutive frames according to the length, width, color format, and frame rate information of the original time-series image data format;

[0128] 2) Channel conversion is used to convert each frame of the continuous time-series image data into a three-dimensional numerical matrix with multiple color channels, where two dimensions represent the width and length of the image data, and the other dimension represents the color channel number. After image channel conversion, the time-series image data is a four-dimensional matrix with a time-series dimension.

[0129] (2) Image difference processing.

[0130] 1) Selection of main color channels. According to the laser wavelength range selected for water leakage visualization, the color channel that is sensitive to the laser in the selected wavelength range is selected as the main color channel for subsequent processing. The number of main color channels can be one or more. After image channel transformation and selection, the water leakage time series image data is in the form of X = [x1, x2, ..., x n-1 ,x n ], where n represents the frame number of the image data in the overall time series data, and x i (i=1,2,…,n) is a three-dimensional matrix containing two length and width dimensions and one color channel dimension, representing a frame of image in the time series image data;

[0131] 2) Image difference calculation. First, determine the image difference time sequence step s, which is an integer greater than or equal to 1; secondly, calculate the time sequence image difference result. The specific calculation formula is X fm-s =[x fm(1) ,x fm(2) ,…,x fm(n-s-1) ,x fm(n-s) ]=|[x1-x 1+s ,x2-x 2+s ,…,x n-s-1 -x n-1 ,x n-s -x n ]|;

[0132] (3) Image binarization

[0133] 1) Binarization threshold setting. Set the image data binarization threshold th b .

[0134] 2) Image

[0135] Binarization processing. Will exceed the binarization threshold th b All elements in the time series image difference result data are assigned the maximum value, which will not exceed the binarization processing threshold th b All elements in the time series image difference result data are assigned to the minimum value, and the binary time series image difference result data X is obtained. fmb =[x fmb(1) ,x fmb(2) ,…,x fmb(n-s-1) ,x fmb(n-s) ], x fmb(1) ,x fmb(2) ,x fmb(n-s-1) ,x fmb(n-s) They represent the time series image difference results after binarization with time series step sizes of 1, 2, ns-1 and ns, respectively.

[0136] Step S3: determining whether there is a water leakage fault based on the binarized time series image difference result data, including:

[0137] Traverse all binarized time series image difference result data, retain the frames with connected image regions, and form binary difference time series image data containing connected regions;

[0138] Traverse the connected areas corresponding to the binary differential time series image data containing connected areas, calculate the distance between the centroids of each connected area in each frame and other connected areas in the frame, and the distance between the centroids of each connected area in each frame and connected areas in other frames;

[0139] Determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold;

[0140] The water leakage characteristic light spot determines that the converter valve has a water leakage fault.

[0141] Furthermore, after determining that the converter valve has a water leakage fault, the following steps are also included:

[0142] The sampling image processing tool obtains the number of pixels in the connected region and the pixel coordinates in the image;

[0143] The mean or median of the pixel coordinates of the water leakage characteristic spots in all connected areas in the image is taken as the water leakage spot position;

[0144] The valve layer where the water leakage occurs is determined by the position of the water leakage light spot.

[0145] Furthermore, the determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold includes:

[0146] Determine whether the centroid distance is less than the set coordinate error threshold. If the centroid distance is less than the set coordinate error threshold, and the number of occurrences of the centroid distance being less than the set coordinate error threshold exceeds the set judgment number threshold, then the connected area is a water leakage characteristic light spot; otherwise, the connected area is a non-water leakage characteristic light spot.

[0147] Furthermore, after determining whether a water leakage fault exists based on the binarized time series image difference result data, the method further includes:

[0148] The characteristic light spots judged as non-leakage in the multi-period time series image data are compared to monitor long-interval leakage faults.

[0149] Furthermore, the process of comparing the characteristic light spots determined as non-leakage in the multi-period time series image data to monitor long-interval leakage faults includes:

[0150] Traverse all connected regions in multiple time series image data and calculate the distance between the centroid of each connected region and other connected regions;

[0151] Traverse the connected areas corresponding to the binary differential time series image data containing connected areas, calculate the distance between the centroids of each connected area in each frame and other connected areas in the frame, and the distance between the centroids of each connected area in each frame and connected areas in other frames;

[0152] Determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold;

[0153] Based on the water leakage characteristic light spot, it is determined that the converter valve has a long-interval water leakage fault.

[0154] Furthermore, after determining that there is a long-interval water leakage fault, the following steps are also included:

[0155] The sampling image processing tool obtains the number of pixels in the connected region and the pixel coordinates in the image;

[0156] The mean or median of the pixel coordinates of the water leakage feature spots in all connected areas in the image is taken as the water leakage spot position.

[0157] Step S3 specifically includes: a water leakage determination step, including one or more of the following main steps:

[0158] (1) Image connected region detection

[0159] The binary time series image difference result data X obtained from the time series image data preprocessing step fmb , using mature image processing tools to process each frame of image x in the binary time series image difference result data fmb(i) (i=1,2,…,ns) performs image connected region detection, and obtains several connected image regions A after excluding the background for a single frame image. ik (i=1,2,…,ns,k=1,2,…o) (It should be noted that o is not a fixed value, and the m value corresponding to different frame images may be different), and obtain the number of regional pixels m corresponding to these image regions ik and the coordinates of the center of mass position in the image (u ik ,v ik ).

[0160] (2) Water leakage identification and light spot position calculation

[0161] 1) Filter the time series image frames with connected image regions. Traverse all binary time series image difference result data X from 1 to ns fmb Number, only retain the frames with connected image regions, and form the binary differential time series image data X containing the connected regions fmba =[x fmba(1) ,x fmba(2) ,…,x fmba(m-1) ,z fmba(m) ], contains m image frames with connected regions, z fmba(m) Represents the image difference result frame data of the mth connected area.

[0162] 2) Set the coordinate error threshold th loc and the discrimination threshold th count . Set the coordinate error threshold th loc Is a positive number. This threshold is mainly used to determine whether the light spots detected in two frames of images are in the same position. Set the number of determination thresholds th count is a positive integer. This threshold is mainly used to determine whether the number of repetitions of water droplets at the same position in the processed time series image data reaches a preset value.

[0163] 3) Comparison and judgment of distance and number thresholds. Traverse the connected regions A corresponding to m binary differential time series image frames containing connected regions. jk (j=1,2,…,m,k=1,2,…o), calculate the distance L between the centroid of each connected region in the jth frame and the connected regions in other frames jlk (j, l = 1, 2, ..., m, j ≠ l, k = 1, 2, ... o), when the distance between the centroid of a connected area in a certain frame and the connected areas in other frames is L jlk There are more than th count Satisfy L jlk <th loc If it is established, then this connected area is considered to be a water leakage feature spot, and all the distances between this feature spot and the connected area satisfy L jlk <th loc The connected areas are all identified as water leakage feature spots, and a water leakage feature spot set is formed. According to the above steps, multiple sets of water leakage spot sets may be obtained by analyzing a single segment of time series image data.

[0164] 4) Calculation of leak spot locations. Analyze a single segment of time-series image data to obtain one or more leak spot sets. Statistically calculate the coordinate distribution characteristics of the spots within the set. Use the mean or median of the pixel coordinates of all the spots within the set as the output leak spot location. Preferably, a standard deviation statistical characteristic parameter can be provided to describe the pixel coordinate distribution characteristics of the spots corresponding to a single leak spot set within the image.

[0165] The invented single-section converter valve leakage explicit image data recognition result output link refers to the output of the identified leakage event and the leakage characteristic light spot position or distribution characteristics, and at the same time outputs all other connected areas in this section of data that are not identified as leakage characteristic light spots and the corresponding pixel coordinate positions in the image.

[0166] A method for jointly identifying converter valve water leakage using multi-segment laser-generated, time-series image data has been developed. This method processes a preset number of multi-segment laser-generated, time-series image data and combines these data to further determine whether a water leakage event has occurred. This method includes obtaining multi-segment data identification results, determining the combined identification results, and outputting the combined identification results.

[0167] The invented multi-segment data recognition result acquisition link refers to obtaining all connected areas that are not identified as water leakage feature spots in the recognition results of multi-segment laser water leakage explicit image data according to a preset number and the corresponding pixel coordinate position data in the image.

[0168] The invented multi-segment data joint recognition result determination process is similar to the recognition process for single-segment data, and includes the following steps:

[0169] 1) Set the coordinate error threshold th loc and the discrimination threshold th count . Set the coordinate error threshold th loc Is a positive number. This threshold is mainly used to determine whether the light spots detected in two frames of images are in the same position. Set the number of determination thresholds th count is a positive integer. This threshold is used to determine whether the number of repetitions of water droplets at the same location in the processed time series image data reaches a preset value. It should be noted that the coordinate error threshold and the number of times threshold for joint discrimination of multiple segments of data can be different from the thresholds used for single-segment discrimination.

[0170] 2) Comparison and judgment of distance and number thresholds. Traverse all connected areas A in the acquired multi-segment data jk (j=1,2,…,m,k=1,2,…o), calculate the distance L between the centroid of each connected area and other connected areas jlk (j, l = 1, 2, ..., m, j ≠ l, k = 1, 2, ... o), when the distance between the centroid of a connected area in a certain frame and the connected areas in other frames is L jlk There are more than th count satisfies L jlk <th loc If it is established, then this connected area is considered to be a water leakage feature spot, and all the distances between this feature spot and the connected area satisfy L jlk <th locThe connected areas are all identified as water leakage feature spots, and a water leakage feature spot set is formed. According to the above steps, multiple sets of water leakage spot sets may be obtained by analyzing multiple time series image data.

[0171] 3) Calculation of leak spot locations. Analyze multiple time-series image data to obtain one or more leak spot sets. Statistically calculate the characteristic values ​​of the spot coordinate distribution within each set. The mean or median of the pixel coordinates of all the spots within the set is used as the output leak spot location. Preferably, a standard deviation statistical characteristic parameter can be provided to describe the pixel coordinate distribution characteristics of the spots corresponding to a single leak spot set within the image.

[0172] The invented multi-segment data joint recognition result output link refers to outputting the leakage spot position or the pixel coordinate distribution characteristics of the spot in the image in the joint recognition result of the multi-segment laser leakage explicit time series image data of the converter valve.

[0173] Compared with the prior art, the present invention has the following beneficial effects:

[0174] (1) The prior art does not provide a specific method for processing laser-generated time-series images of water leakage in converter valves. The present invention provides a technical solution that can complete laser-generated time-series image processing of water leakage without increasing the demand for special computing resources such as artificial intelligence chips, using a general computing architecture, thereby realizing early detection of minor water leakage in converter valves.

[0175] (2) The present invention provides a time-series image processing method for laser water leakage in a converter valve that combines simple image processing algorithms such as difference frames, binarization, connected area recognition, and target distance calculation, thereby realizing the detection of water leakage events in single-segment data and the positioning of pixel coordinates in the image.

[0176] (3) The present invention provides a single-segment and multi-segment combined laser leakage explicit time series image processing method for the converter valve. By jointly distinguishing the recognition results of multiple segments of data, the leakage event is further identified and judged, which can avoid the problem of missed detection of leakage faults due to the long interval between two repeated leakage events.

[0177] (4) The present invention provides a device that can implement the method proposed by the present invention and mark the characteristic light spot of the water leakage by displaying the laser water leakage explicit image for the water leakage fault.

[0178] Example 2

[0179] A device for monitoring water leakage fault of a converter valve, such as Figure 5 Shown, including:

[0180] A communication unit, used for acquiring laser water leakage explicit time-series image data within a period to be determined;

[0181] The calculation unit is used to pre-process the laser water leakage explicit time sequence image data to obtain binarized time sequence image difference result data; and determine whether there is a water leakage fault based on the binarized time sequence image difference result data.

[0182] Furthermore, the communication unit is used to communicate with one or more laser-visualized converter valve water leakage monitoring devices to obtain laser-visualized time-series image data. The invented communication unit can also communicate and exchange data with other devices or systems.

[0183] Furthermore, the calculation unit is used to complete the pre-processing of the laser water leakage explicit time series image data as described in Example 1, and to monitor whether there is a water leakage fault, which will not be repeated here.

[0184] The converter valve water leakage fault monitoring device of the present invention further includes a control unit and a human-computer interaction unit;

[0185] The invented control unit is used to connect with other functional units and to coordinate and control the work of other functional units.

[0186] The human-computer interaction unit is used to realize the interaction between the device and management personnel or converter station operators, supporting personnel to view device information and manage the device. At the same time, when a water leakage fault is detected, a laser water leakage explicit image is displayed and the position of the identified water leakage feature light spot is marked in the image for the operator's reference.

[0187] Example 3

[0188] The embodiment of the present invention provides a time sequence image data processing device for detecting a water leakage fault of a converter valve, such as Figure 4 shown. Figure 4 Only the external communication units are shown, the internal communication units are indicated by arrows.

[0189] In the embodiments provided herein, an Ethernet network with a twisted-pair cable as the physical layer connection and a corresponding network adapter are used as the communication unit to exchange data with a converter valve laser leakage visualization image acquisition device to acquire time-series image data. In practice, the present invention can also employ any other communication method capable of data exchange, such as Wi-Fi, 5G, satellite communications, and Bluetooth.

[0190] In the embodiments provided herein, the computing unit and control unit are implemented by a general-purpose X86 architecture microcomputer, employing a microprocessor as both the computing unit and the control unit, and corresponding software programs are used to implement specific control and calculations. In practice, the present invention may employ single-chip microcomputers, embedded devices, centralized cloud computing devices, and the like as the computing unit and the control unit.

[0191] In the embodiments provided herein, a display, speaker, keyboard, mouse, and microphone, along with corresponding software, serve as the human-computer interaction unit. The display outputs the corresponding laser water leak visualization image data and marks the location of the characteristic light spot of the leak fault; the speaker provides an audible alarm; and the keyboard, mouse, and microphone receive information and instructions input by management and operation personnel and execute the corresponding commands. In specific implementations, this invention can also utilize methods such as VR virtual reality output and somatosensory input to achieve human-computer interaction.

[0192] In an embodiment of the present invention, a single-segment and multi-segment data joint recognition method is specifically adopted to implement converter valve leakage fault monitoring based on time series image data analysis.

[0193] In the embodiment provided by the present invention, the length of a single segment of data is set to 1 minute, and the number of multiple segments of data is set to 1440 segments, corresponding to the length of 24 hours. In the implementation of the present invention, the specific length of a single segment of data and the number of multiple segments of data can be set to different values ​​according to the site conditions.

[0194] In the embodiment provided by the present invention, the resolution of a single segment of sequential image data is 640×480, the data frame rate is 5 frames per second, the number of frames in a single segment is 300, and the original sequential image data format is BayerBG8. In the implementation of the present invention, the specific image data resolution and frame rate can be set to different values ​​according to site conditions, and the original sequential image data format can be any data format that can express image data, such as RGB, BayerGB, BayerGR, Mono, etc.

[0195] In the embodiment provided by the present invention, OpenCV tool software is specifically used to implement image channel conversion of original data in BayerBG8 format, converting it into RGB three-channel format, with each pixel in each channel represented by an 8-bit unsigned integer. In the implementation of the present invention, image channel conversion can also be implemented using other tool software or by personnel's own coding.

[0196] In the embodiment provided by the present invention, the primary color channel used before image difference processing is the R channel. Other color channels may also be selected in the implementation of the present invention. In the embodiment provided by the present invention, the difference timing step size s=1 is set. In the implementation of the present invention, the difference timing step size may also be set to other values.

[0197] In the embodiment provided by the present invention, the binarization threshold th is set b = 50. In the implementation of the present invention, the binarization threshold may also be set to other values.

[0198] In the embodiment provided by the present invention, OpenCV tool software is specifically used to implement the detection and marking of connected regions of the image result data after differential binarization. In the implementation of the present invention, the detection of connected regions of image data can also be implemented by other tool software or by staff's own coding.

[0199] In the embodiment provided by the present invention, the coordinate error threshold th is set loc =20, discrimination times threshold th count = 5. In the implementation of the present invention, the coordinate error threshold and the discrimination times threshold may also be set to other values.

[0200] In the embodiment provided by the present invention, the distance between each connected region is calculated using the Euclidean distance method. In the implementation of the present invention, other distance calculation methods may also be used to calculate the distance between each connected region.

[0201] In the embodiment provided by the present invention, for the set of water leakage characteristic light spots in the single-segment data recognition results, the average of the water leakage characteristic light spot coordinates in the set is used as the pixel position coordinates of the water leakage characteristic spot in the image, and the standard deviation is used as the characteristic value of the pixel position coordinate distribution of the water leakage characteristic spot in the image.

[0202] The present invention provides an embodiment for the problem of distinguishing the joint recognition results of multiple segments of data, specifically setting the coordinate error threshold th loc =15, discrimination threshold th count = 10. In the implementation of the present invention, the coordinate error threshold and the number of times threshold for determining the result of the joint determination of the majority of data may also be set to other values.

[0203] In the embodiment provided by the present invention, for the problem of joint recognition result recognition of multiple segments of data, the Euclidean distance method is used to calculate the distance between each connected region. In the implementation of the present invention, other distance calculation methods can also be used to calculate the distance between each connected region.

[0204] In the embodiment provided by the present invention, for the set of leakage characteristic light spots in the joint identification results of multiple segments of data, the average of the coordinates of the leakage characteristic light spots in the set is used as the pixel position coordinates of the leakage characteristic spots in the image, and the standard deviation is used as the characteristic value of the distribution of the pixel position coordinates of the leakage characteristic spots in the image.

[0205] The embodiments of the present invention also provide storage media such as CDs and USB flash drives, which contain programs that can implement the methods proposed by the present invention and can be distributed to facilitate engineering application implementation.

[0206] Example 4

[0207] like Figure 6As shown, the present invention also provides an electronic device, which may be a computer, a single-chip microcomputer, a smart mobile device, or the like. The electronic device in this embodiment may include a processor, a memory, a transceiver component, and the like. The memory, processor, and transceiver component are connected via a bus; the memory may be used to store an execution program, which may include instructions; and the processor may be used to execute the instructions stored in the memory. The memory may also be used to store data, which may be accessed and / or modified during the execution of the instructions.

[0208] The processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of a converter valve water leakage fault monitoring method in the above-mentioned embodiment.

[0209] Example 5

[0210] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in the electronic device for storing programs and data. It can be understood that the storage medium here can include both the built-in storage medium in the electronic device and the extended storage medium supported by the electronic device. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory (non-volatile memory), such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a converter valve water leakage fault monitoring method in the above embodiment.

[0211] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0212] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts 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, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0213] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0214] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0215] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for monitoring a water leakage fault of a converter valve, characterized in that: include: Obtaining laser water leakage explicit time series image data within the time period to be determined; Preprocessing the laser water leakage explicit time series image data to obtain binarized time series image difference result data; Determining whether there is a water leakage fault based on the binarized time series image difference result data; The preprocessing of the laser water leakage explicit time series image data to obtain binarized time series image difference result data includes: Perform channel conversion on the laser water leakage explicit time series image data to obtain a water leakage time series image represented by a four-dimensional matrix; Performing image difference calculation on the water leakage time series image represented by the four-dimensional matrix to obtain a time series image difference result; Performing image binarization processing on the time series image difference result to obtain binarized time series image difference result data; The step of performing channel conversion on the laser water leakage explicit time series image data to obtain a water leakage time series image represented by a four-dimensional matrix includes: Based on the laser water leakage explicit time-series image data format, converting the laser water leakage explicit time-series image data into a plurality of consecutive frames of continuous time-series image data represented by a numerical matrix; Converting each frame of the continuous time-series image data into a three-dimensional numerical matrix having multiple color channels; Constructing a water leakage time series image represented by a four-dimensional matrix by combining the three-dimensional numerical matrix with the time series dimension; The determining whether there is a water leakage fault based on the binarized time series image difference result data includes: Traverse all binarized time series image difference result data, retain the frames with connected image regions, and form binary difference time series image data containing connected regions; Traverse the connected areas corresponding to the binary differential time series image data containing connected areas, calculate the distance between the centroids of each connected area in each frame and other connected areas in the frame, and the distance between the centroids of each connected area in each frame and connected areas in other frames; Determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold; Determining that the converter valve has a water leakage fault based on the water leakage characteristic light spot; The method of determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold includes: Determine whether the centroid distance is less than the set coordinate error threshold. If the centroid distance is less than the set coordinate error threshold, and the number of occurrences of the centroid distance being less than the set coordinate error threshold exceeds the set judgment number threshold, then the connected area is a water leakage characteristic light spot; otherwise, the connected area is a non-water leakage characteristic light spot.

2. The method according to claim 1, wherein The time series image difference result is calculated as follows: Where, The original time series image data The time series image difference result after difference according to the step size s is: 、 、 、 They are The first, second, third, up to the nth element in , 、 、 、 、 、 、 、 The original time series image data The elements in , n is a positive integer.

3. The method according to claim 1, wherein The performing image binarization processing on the time series image difference result to obtain binarized time series image difference result data includes: The elements in the time sequence image difference results that are greater than the set binarization threshold are assigned the maximum value, and the elements in the time sequence image difference results that are not greater than the set binarization threshold are assigned the minimum value, to obtain the binarized time sequence image difference result data.

4. The method according to claim 1, wherein After confirming that there is a water leakage fault in the converter valve, the following steps are also required: Use image processing tools to obtain the number of pixels in the connected area and the pixel coordinates in the image; The mean or median of the pixel coordinates of the water leakage characteristic spots in all connected areas in the image is taken as the water leakage spot position; The valve layer where the water leakage occurs is determined by the position of the water leakage light spot.

5. The method according to claim 1, wherein After determining whether there is a water leakage fault based on the binarized time series image difference result data, the method further includes: The characteristic light spots judged as non-leakage in the multi-period time series image data are compared to monitor long-interval leakage faults.

6. The method according to claim 5, wherein The method of comparing the characteristic light spots determined as non-leakage in the multi-period time series image data to monitor the long-interval leakage fault includes: Traverse all connected regions in multiple time series image data and calculate the distance between the centroid of each connected region and other connected regions; Traverse the connected areas corresponding to the binary differential time series image data containing connected areas, calculate the distance between the centroids of each connected area in each frame and other connected areas in the frame, and the distance between the centroids of each connected area in each frame and connected areas in other frames; Determining whether a connected area is a water leakage characteristic spot based on the centroid distance between each connected area in each frame and the connected areas in the same frame and other frames and a preset coordinate error threshold and a discrimination number threshold; Based on the water leakage characteristic light spot, it is determined that the converter valve has a long-interval water leakage fault.

7. The method according to claim 6, wherein After confirming the existence of a long-interval water leakage fault, the following also need to be done: The sampling image processing tool obtains the number of pixels in the connected region and the pixel coordinates in the image; The mean or median of the pixel coordinates of the water leakage feature spots in all connected areas in the image is taken as the water leakage spot position.

8. A device for implementing the method for monitoring a water leakage fault of a converter valve according to any one of claims 1 to 7, characterized in that: include: A communication unit, used for acquiring laser water leakage explicit time-series image data within a period to be determined; The calculation unit is used to pre-process the laser water leakage explicit time sequence image data to obtain binarized time sequence image difference result data; and determine whether there is a water leakage fault based on the binarized time sequence image difference result data.

9. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a method for monitoring water leakage fault of a converter valve according to any one of claims 1 to 7 is implemented.

10. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, a method for monitoring water leakage fault of a converter valve according to any one of claims 1 to 7 is implemented.

Images