CN112001327B - Fault identification method and system for valve hall equipment - Google Patents

Abstract

The invention discloses a valve hall equipment fault identification method and a system, wherein the method comprises the following steps: acquiring a video stream of equipment to be detected, which is shot by image acquisition equipment in real time, wherein the video stream comprises: infrared images and ultraviolet images; preprocessing any frame of image in the video stream, and extracting feature data of the preprocessed image; analyzing the feature data by at least one of a dynamic feature analysis algorithm and a heuristic classification algorithm to judge whether the equipment to be detected has faults or not; when a fault is detected, a fault type is generated and located to a specific location. According to the invention, the infrared image and the ultraviolet image are preprocessed and the characteristic data are extracted, at least one of a dynamic characteristic analysis algorithm and a heuristic classification algorithm is utilized to analyze the characteristic data, the type and the fault area position of the valve hall equipment fault are clear, and the reliability, the scientificity and the intellectualization of the operation and maintenance of the valve hall equipment are improved.

Description

Fault identification method and system for valve hall equipment

technical field

The invention relates to the technical field of electric power inspection, in particular to a fault identification method and system for valve hall equipment.

Background technique

At present, the time-based preventive test and regular maintenance of power equipment maintenance system, due to its own blindness and low level of maintenance, may cause damage to the good operation of the equipment or make the equipment "more repairs and worse". , especially complex and large-scale equipment such as UHV converter valves, converter transformers, and wall bushings. Since the commonly used preventive tests are usually carried out with low voltage applied offline, the low-voltage test cannot simulate the equipment in the UHV situation. Under the operating conditions; offline, it is impossible to simulate the thermal stress and other characteristics of the equipment; at the same time, the structure of the UHV converter valve and other equipment in the valve hall is complex, and there are waterway shuttles and high and low potential intersections inside the valve body. The overall insulation characteristics are reliably guaranteed in the design and production of the product, but with long-term operation, the performance of the components decreases, the operating environment changes, especially water seepage, which poses a potential threat to the reliable operation of the converter station and the high-voltage terminals of the converter transformer. It is difficult to detect abnormalities in time through the current artificial monitoring methods such as abnormal oil seepage, which poses potential risks to the continuous and stable operation of the entire UHV DC project.

The valve hall of the converter valve is the core building unit of the UHV converter station. It operates in a complex environment intertwined with multiple physical fields such as electricity, magnetism, and heat for a long time. The reliability of the operation of the valve hall of the converter valve directly affects the entire DC project. The key point of stability, the current way to eliminate hidden dangers in the operation of the equipment in the valve hall of the converter valve is regular maintenance, and the regular maintenance under the power failure state cannot simulate the actual operating conditions of the equipment such as the converter valve, and at the same time lead to some operational defects of the equipment or Hidden dangers cannot be reproduced, and there are problems that the type of valve hall equipment failure and the location of the failure area cannot be clarified.

Contents of the invention

Therefore, the valve hall equipment fault identification method and system provided by the present invention overcomes the defects in the prior art that the types of valve hall equipment faults and the location of the fault area cannot be clarified.

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

In the first aspect, an embodiment of the present invention provides a valve hall equipment fault identification method, including:

Obtaining a video stream of the device to be detected captured by the image acquisition device in real time, the video stream comprising: an infrared image and an ultraviolet image;

Preprocess any frame of image in the video stream, and extract the feature data of the preprocessed image;

Using at least one of a dynamic feature analysis algorithm and a heuristic classification algorithm to analyze the feature data to determine whether there is a fault in the device to be tested;

When a fault is detected, the fault type is generated and located to a specific location.

In one embodiment, after the step of generating a fault type and locating to a specific location when a fault is detected, it further includes: generating alarm information, and drawing and visually displaying the alarm information, the alarm information includes fault valve hall Equipment failure type, time, image and video information.

In an embodiment, the failure types include: open flame, electric discharge, overheating, and water seepage.

In one embodiment, when the detected equipment fault is open flame, overheating, or water seepage, the fault area is identified in the infrared image; when the equipment fault is detected as a discharge type, the fault area is identified in the ultraviolet image.

In one embodiment, the step of detecting an open flame failure includes: acquiring an infrared image captured by an image acquisition device in real time, determining a suspicious area of the flame through a preset image processing algorithm, extracting the characteristic value of each flame at each moment in the suspicious area, and inputting the dynamic Feature pool, feature values include: average circularity, area average rate of change, perimeter average rate of change, use heuristic classification algorithm to classify feature values into teams, and combine dynamic feature analysis algorithm to judge whether multiple devices to be tested have open flame failures .

In one embodiment, the step of detecting the discharge fault includes: acquiring an ultraviolet image captured by an image acquisition device in real time, detecting suspicious discharge points in a single frame image through a preset image detection algorithm, and counting the distribution of ultraviolet noise according to the discharge negative sample , using the general model fitting, the probability density function of the noise can be obtained, based on the probability density function, the heuristic feature classification algorithm is used to construct the time sequence of the suspicious discharge area, and the probability that the sequence is a noise sequence is calculated. A threshold comparison is set to judge whether there is a discharge fault in the device to be detected.

In one embodiment, the step of detecting an overheating fault includes: acquiring an infrared image captured by an image acquisition device in real time, determining a suspicious heating area through a preset image processing algorithm, extracting a temperature value of the heating area, and analyzing the temperature of the heating area using a dynamic feature analysis algorithm The temperature rise, temperature difference, and relative temperature difference are compared with the corresponding preset thresholds to judge whether there is an overheating fault in the device to be tested.

In one embodiment, the step of detecting the water seepage fault includes: acquiring the infrared image captured by the image acquisition device in real time, using the image segmentation algorithm to find the suspicious area of water seepage in the single frame image, and calculating the average temperature difference between the area inside the area and the area outside the area boundary , preliminarily screen suspicious areas, use the heuristic classification algorithm of contours in time series, divide different contours under multiple frames into different queues, each queue represents the change of a water seepage area in time series, and calculate the time series of each queue The average change rate of the area and the average change rate of the perimeter are compared with the set change rate threshold, and the seepage area is screened out.

In the second aspect, an embodiment of the present invention provides a fault identification system for valve hall equipment, including:

An image acquisition module, configured to acquire a video stream of the device to be detected captured by the image acquisition device in real time, the video stream comprising: an infrared image and an ultraviolet image;

A feature extraction module is used to preprocess any frame of image in the video stream, and extract feature data of the preprocessed image;

The fault detection module is used to analyze the feature data by using at least one of the dynamic feature analysis algorithm and the heuristic classification algorithm, and judge whether there is a fault in the equipment to be detected;

The fault identification module is used to generate a fault type and locate a specific location when a fault is detected.

In a third aspect, an embodiment of the present invention provides a terminal, including: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores information that can be executed by the at least one processor. Instructions, the instructions are executed by the at least one processor, so that the at least one processor executes the fault identification method for valve hall equipment according to the first aspect of the embodiments of the present invention.

In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method described in the first aspect of the embodiment of the present invention. Fault identification method for valve hall equipment.

The technical solution of the present invention has the following advantages:

The valve hall equipment fault identification method and system provided by the present invention obtain the video stream of the equipment to be detected captured by the image acquisition device in real time. The video stream includes: infrared images and ultraviolet images; processing, and extracting the feature data of the preprocessed image; using at least one of the dynamic feature analysis algorithm and the heuristic classification algorithm to analyze the feature data to determine whether there is a fault in the device to be detected; when a fault is detected, generate the fault type and locate it to a specific location. It is proposed to preprocess infrared images and ultraviolet images and extract feature data, and use at least one of dynamic feature analysis algorithm and heuristic classification algorithm to analyze feature data, clarify the type of valve hall equipment failure and the location of the fault area, and improve This ensures the reliability, scientificity and intelligence of valve hall equipment operation and maintenance.

Description of drawings

In order to more clearly illustrate the specific implementation of the present invention or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings that need to be used in the specific implementation or description of the prior art. Obviously, the accompanying drawings in the following description The drawings show some implementations of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

FIG. 1 is a flow chart of a specific example of a valve hall equipment fault identification method provided by an embodiment of the present invention;

Fig. 2 is a flow chart of a specific login example of a valve hall equipment fault identification method provided by an embodiment of the present invention;

Fig. 3 is a schematic diagram of a dynamic feature pool of a specific example of a valve hall equipment fault identification method provided by an embodiment of the present invention;

FIG. 4 is a schematic diagram of a specific example of a fault identification method for valve hall equipment provided by an embodiment of the present invention for solving fault overlap;

Fig. 5 is a block diagram of a fault identification system for valve hall equipment provided by an embodiment of the present invention;

FIG. 6 is a composition diagram of a specific example of a terminal provided by an embodiment of the present invention.

Detailed ways

The technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Apparently, the described embodiments are part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" etc. The indicated orientation or positional relationship is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, or in a specific orientation. construction and operation, therefore, should not be construed as limiting the invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and should not be construed as indicating or implying relative importance.

In the description of the present invention, it should be noted that unless otherwise specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection or a detachable connection. Connected, or integrally connected; it can be mechanically or electrically connected; it can be directly connected, or indirectly connected through an intermediary, or it can be the internal communication of two components, which can be wireless or wired connect. Those of ordinary skill in the art can understand the specific meanings of the above terms in the present invention in specific situations.

In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as there is no conflict with each other.

Example 1

A valve hall equipment fault identification method provided by an embodiment of the present invention, as shown in Figure 1, includes the following steps:

Step S1: Obtain a video stream of the device to be detected captured by the image acquisition device in real time, the video stream including: an infrared image and an ultraviolet image.

In the embodiment of the present invention, as shown in Figure 2, when the user obtains the video stream of the device to be detected in real time taken by the image acquisition device in real time, he needs to input the user name and password to log in. NVR) and other network devices initiate a login request, this is just an example, not limited to this, and the corresponding network device is selected in practical applications. After the device is successfully logged in, data initialization is performed, including: initializing TCP connection, video stream, infrared raw data stream, monitoring interface, fault detection rule parameters, etc., and then updating the device tree according to the network device information. For each device administrator, the alarm threshold can be set and other parameters.

In the embodiment of the present invention, the image acquisition device includes: an RGB camera, an ultraviolet camera, an infrared camera and other image acquisition devices. This is just an example, not limited thereto, and the corresponding image acquisition device is selected in practical applications.

Step S2: Perform preprocessing on any frame of image in the video stream, and extract feature data of the preprocessed image.

In the embodiment of the present invention, by setting the callback function of the video stream, the current frame image in the memory and the infrared raw temperature data are updated in real time, and the current frame and the infrared raw temperature data are read. First, any frame image in the video stream is processed Preprocessing, preprocessing operations include filtering out part of the noise, edge sharpening, and segmenting the device area. This is just an example, not limited to this. In practical applications, select the corresponding preprocessing means to extract feature data of suspicious candidate areas and perform simple filtering.

Step S3: Using at least one of a dynamic feature analysis algorithm and a heuristic classification algorithm to analyze the feature data to determine whether there is a fault in the device to be detected.

In the embodiment of the present invention, the fault types include: open flame, discharge, overheating, water seepage; this is just an example, not limited to this, and the corresponding fault type is selected in practical applications; when the detected equipment fault is open flame, In the case of overheating and water seepage, the fault area is identified in the infrared image; when the equipment fault is detected as a discharge type, the fault area is identified in the ultraviolet image.

In the embodiment of the present invention, the step of open flame fault detection includes: acquiring the infrared image captured by the image acquisition device in real time, determining the suspicious area of the flame through the preset image processing algorithm, extracting the feature value of each flame at each moment in the suspicious area, and inputting Dynamic feature pool, feature values include: average circularity, area average rate of change, perimeter average rate of change, use heuristic classification algorithm to classify feature values into groups, combined with dynamic feature analysis algorithm to judge whether there are open flames in multiple devices to be detected Fault.

Specifically, when a flame burns, a large amount of electromagnetic radiation will be generated. The electromagnetic wave bands are mainly concentrated in the infrared region and the visible light region, and there is also a certain amount of radiation in the ultraviolet band. Because the brightness of the flame region in the infrared image is much higher than that of In other areas, use the preset image processing algorithm to segment the flame area, but heat sources such as incandescent lamps and halogen lamps also emit electromagnetic wave radiation similar to flames, so the area initially identified by the infrared image is regarded as a suspicious area.

In the infrared image, the outline of objects such as incandescent lamps is similar to a circle, and most high-temperature mobile interference sources, such as people, usually have a relatively regular shape with relatively gentle changes, while the shape of the flame is irregular and its circle The circularity is not only low on average, but also changes significantly with time. Therefore, the circularity and its rate of change are selected as important criteria for the flame.

The circularity calculation formula is as follows:

Among them, A is the area of the suspected fire area; P is the perimeter of the suspected fire area.

The formula for calculating the average rate of change of circularity is as follows:

Among them, n is the length of circularity feature vector.

Analyze the flame sample, set the circularity threshold C _t according to the average circularity of the flame, exclude some interference contours, count the dynamic change law of the circularity, set the average change rate threshold V _ct , and exclude the flat circularity change Interference contours.

Since the shape of the flame changes significantly with time, the area change rate and perimeter change rate can also be used as important criteria for eliminating interference. Calculate the area change rate and perimeter change rate respectively by the following formulas:

The formulas for calculating the average rate of change of area and the average rate of change of perimeter are respectively:

Analyze the flame sample, set the change rate thresholds V _at and V _lt according to the average change rate of the area and the average change rate of the circumference of the sample, and exclude the interference contours with gentle changes in the area and circumference.

In the dynamic change of the flame, its contour shape will change with time, and the statistics of the change rate of its contour shape can also be used as the criterion of the flame. Extract the contour V=(x(i),y(i)) of the suspicious area, and the contour can be regarded as a one-dimensional complex number sequence:

V(i)=x(i)+jy(i)

Do discrete cosine transform (DCT) on it:

Among them, the low-frequency part mainly reflects the overall shape of the region, and the coefficient of the low-frequency part is taken to define the shape description coefficient:

The shape change between any two contours S ₁ and S ₂ is calculated using the Euclidean distance by the following formula:

Analyze the flame sample and set the inter-frame shape change threshold D _st . Count the total number of feature elements whose shape change is greater than D _st between two adjacent frames for a feature queue containing M feature elements. If it is less than M/3, it is considered that the contour does not change significantly in time series. This is just an example, not limited to this , select the corresponding value according to the actual needs in practical applications, then the heating target tracked by this feature queue is not considered to be a flame.

In the application scenario of the embodiment of the present invention, multiple heat sources may appear at the same time, while the previous flame recognition algorithm can be summarized into three steps: extraction of suspicious areas; detection of dynamic features; feature data analysis and decision-making, which is only applicable to a single target analyze. The embodiment of the present invention implements a multi-target area tracking and optimization algorithm, calculates feature values for each suspicious area at each moment, and inputs them into the dynamic feature pool as shown in Figure 3, and uses a heuristic classification algorithm to classify the feature values into groups.

Using the above algorithm for each suspicious area, its various eigenvalues can be obtained. For example, for a suspicious flame area, its circularity, area, perimeter, and shape description coefficient can be calculated. This group of eigenvalues is called the suspicious area. The feature element q of , the position attribute (X _q , Y _q ) of the feature element is calculated by the following formula:

Among them, (x(i), y(i)) are the points forming the outline of the suspicious area, (X _q , Y _q ) is the position of the feature element q, which is the outline center of the suspicious area.

Input the feature elements into the dynamic feature pool. Multiple feature elements are arranged in time order to form a feature queue Q. A feature queue corresponds to a heat source in the infrared image, and defines the center of the heat target area tracked by the current queue in the image. The coordinate value is the position attribute (X _Q , Y _Q ) of the queue; when a new feature element q joins the queue at time t, the position of the queue is updated using the "weighted average movement" by the following formula:

Among them, λ is a smoothing factor. When λ is large, the newly added elements have a greater impact on the position attribute of the queue. The queue can quickly track the position changes of elements, but the jitter is also large, and it is easy to be affected by individual elements and cause tracking failure; When λ is small, newly added elements have little influence on the position attribute of the queue, historical elements have a greater influence on the position of the queue, and the queue position jitter is small, but when λ is too small, the position change of the element cannot be effectively tracked, so λ To choose an appropriate value.

The distance between feature elements or feature queues is calculated as Euclidean distance by the following formula:

Three operations between feature elements and feature queues are defined:

Element return: feature elements are added to the end of a feature queue, and the queue position is updated;

Create queue: Create a new feature queue, that is, a new characteristic time series of hot areas; queue fusion:

The two queues are merged into one queue, and the elements are arranged in chronological order. If multiple elements appear at the same time, the one with the larger area is reserved. When a new element is added to the dynamic feature pool, use the distance formula to calculate its distance D _n from all queues, and take the minimum distance and the corresponding queues D _min and Q _min . Set the maximum return distance threshold D _t , when D _min ≤ D _t , put the element into the Q _min queue, otherwise create a new queue; check the distance between the new queue and other queues, and when there is a distance smaller than D _t , put The two queues are merged.

Finally, the dynamic feature algorithm is used to analyze each queue separately to realize the tracking and detection of multiple targets. This heuristic feature classification algorithm is also used in water seepage detection and discharge detection.

In the embodiment of the present invention, the step of detecting the discharge fault includes: obtaining the ultraviolet image captured by the image acquisition device in real time, detecting the suspicious discharge point in the single frame image through the preset image detection algorithm, and counting the ultraviolet noise according to the discharge negative sample The probability density function of the noise can be obtained by fitting the general model. Based on the probability density function, the heuristic feature classification algorithm is used to construct the sequence of suspicious discharge areas in time, and the probability that the sequence is a noise sequence is calculated. Through the corresponding Preset thresholds are compared to determine whether there is a discharge fault in the device to be detected.

In the embodiment of the present invention, the suspicious area of discharge can be easily obtained in the ultraviolet image by using the pixel brightness information. The key to distinguish it is the discharge point and ultraviolet noise. Using simple image detection technology, the suspicious discharge in a single frame image can be detected The suspicious area where the point is located, and then use statistical algorithms to calculate the probability of continuous discharge points around the suspicious area in a time series. When the probability value is greater than the set prior probability, the suspicious area is determined to be a discharge area. .

The brightness of the discharge point in the ultraviolet image is significantly higher than the brightness of the background environment, the shape is random, and there is no texture information, but the noise of the ultraviolet image also has similar characteristics. In the past, the discharge detection based on ultraviolet imaging focused on the research of discharge intensity and ultraviolet gain. , spot area, shape, distance, photon number and other parameters, the purpose is to quantify the discharge intensity, this is just an example, not limited to this, in practical applications, select the corresponding research parameters according to actual needs; when the discharge intensity is relatively Large, when the detection distance is short, the spot area is large, which can be well distinguished from the background noise, but when the discharge intensity is small, there is occlusion between the discharge point and the UV camera, and the detection distance is long, the discharge point is in the ultraviolet image The area of the spot may be small, and it is impossible to effectively distinguish the spot area from the background noise.

The position of ultraviolet noise in the image is randomly distributed, and the discharge point always appears in the equipment failure area, the position is relatively fixed, and has a certain frequency, and it appears in a certain order in the time sequence of the ultraviolet image. Points appear continuously in adjacent frames in the ultraviolet image, while the probability of noise appearing continuously in a region is low.

First of all, since the discharge point in the ultraviolet image has the characteristics of high brightness, simple image detection technology can be used to detect the suspicious discharge point in the single frame image, and according to the discharge negative sample, the distribution of ultraviolet noise is counted, and the general model is used to fit , the probability density function f(t) of the noise can be obtained.

Using the heuristic feature classification algorithm to construct the time sequence of suspicious discharge areas, and calculate the probability that the sequence is a noise sequence:

Among them, t is the time difference between the current spot and the previous spot in the same queue. By comparing with the corresponding preset threshold, it is judged whether there is a discharge fault in the device to be detected. When P _Q <0.01, it is judged that this sequence is a discharge sequence. This is just an example and not limited thereto. In practical applications, a corresponding preset threshold is selected according to actual requirements.

In the embodiment of the present invention, the step of detecting an overheating fault includes: acquiring an infrared image captured by an image acquisition device in real time, determining a suspicious heating area through a preset image processing algorithm, extracting a temperature value of the heating area, and analyzing the heating area using a dynamic feature analysis algorithm The temperature rise, temperature difference, and relative temperature difference are compared with the corresponding preset thresholds to judge whether there is an overheating fault in the equipment to be tested.

In the embodiment of the present invention, when the power transmission and transformation equipment is in normal operation, the conductive circuit, insulating medium and iron core all have heat and temperature rise within the normal design range; when the equipment has various bad states such as poor contact, short circuit, and pollutant coverage At the same time, various components outside and inside the power equipment may have different thermal effects that exceed the design standards. The infrared thermal imager is used to detect the infrared energy emitted by the thermal defect of the power equipment. Once there is a defect in the tested equipment, the temperature field of the corresponding part will change. This change can be accurately captured by the infrared thermal imager. The detection equipment is running The distribution of temperature in the state can diagnose the equipment failure in time.

Overheat detection using infrared imaging is very important in power equipment inspections. Many potential accidents have been discovered and prevented. However, at present, infrared detection relies on on-site engineers to hold infrared thermal imaging cameras and conduct artificial analysis on equipment status based on infrared images. The embodiment of the present invention utilizes an online infrared imager and a dynamic feature analysis algorithm to analyze the temperature rise, temperature difference, and relative temperature difference in the heating area, and compares it with the corresponding preset threshold to determine whether there is an overheating fault in the device to be detected, thereby realizing the power monitoring. All-weather monitoring of equipment temperature status, combined with image processing technology, realizes real-time autonomous fault location and accurate analysis of equipment overheating.

The embodiments of the present invention obtain the infrared image of the device in real time, use image processing technology to determine the heating area, read the infrared raw data stream at the same time, perform mathematical transformation on the original data to obtain temperature data, and count the temperature data of the heating area to obtain the temperature T of the heating point. Compare the hot spot temperature T with the normal operating temperature threshold T _t , and send an alarm when T>T _t . For example, the "Jiangsu Province Infrared Temperature Measurement Standardization Work Instructions" stipulates that when the hot spot temperature of the metal wire is > 80 ℃, it is judged as a serious defect. This is just an example, not limited to this, and the corresponding alarm standard should be selected in practical applications.

Calculate the temperature rise based on T and the ambient temperature with reference to the body surface temperature T _e :

θ = TT _e .

Calculate the temperature difference between different DUTs or different parts of the same DUT:

D＝T ₁ -T ₂

Compare the current temperature difference D with the normal operating temperature difference threshold D _t , and send an alarm when D>D _t , for example: when the temperature rise of the metal wire is specified to be >15K, it is judged as a serious defect. Select the corresponding threshold according to actual needs in the application.

Calculate the relative temperature difference according to the temperature rise stomach and temperature difference D:

Compare the current relative temperature difference δ _t with the normal working relative temperature difference threshold δ _t , and send an alarm when δ _t > δ _t . For example: It is stipulated that when the relative temperature difference of the metal wire is ≥95%, it is judged as a critical defect. This is just an example and not limited to this. In practical applications, select the corresponding threshold according to actual needs.

The engineer responsible for the safety of equipment operation formulates the corresponding infrared feature detection standards, uses the infrared thermal imaging camera to collect and calculate the infrared feature values of the equipment in real time, and diagnoses equipment failures according to the standards, so as to realize automatic monitoring of equipment heating conditions around the clock. Once a heating failure occurs, Alerts can be issued in a timely manner.

In the embodiment of the present invention, the step of water seepage fault detection includes: acquiring the infrared image captured by the image acquisition device in real time, using the image segmentation algorithm to find the suspicious area of water seepage in the single frame image, and calculating the average temperature of the area inside the area and the area outside the area boundary Poor, preliminarily screen suspicious areas, use the heuristic classification algorithm of contours in time series, divide different contours under multiple frames into different queues, each queue represents the change of a water seepage area in time series, and calculate the time series of each queue The average change rate of the area and the average change rate of the perimeter are compared with the set change rate threshold, and the water seepage area is screened out.

In the embodiment of the present invention, water seepage refers to the situation that the water in the pipeline permeates and diffuses to the surface of the equipment or even drips due to reasons such as damaged pipelines and poorly sealed joints; when the cooling water seeps out, the water temperature is lower than that of the operating equipment When the temperature is high, a continuous and isolated low-temperature area will appear under the infrared thermal imaging camera. In the past, water seepage detection was carried out by using moisture detection test paper or on-site personnel holding an infrared thermal imaging camera to observe the equipment. The embodiment of the present invention provides an online real-time water seepage detection algorithm in combination with the low-temperature characteristics and dynamic characteristic changes of water seepage, which can detect the seepage fault of cooling water lower than the equipment temperature.

Firstly, the image segmentation algorithm is used to obtain the suspicious area of water seepage in a single frame image, and the average temperature difference between the area inside the area and the area outside the area boundary is calculated, and the suspicious area is initially screened. Using the heuristic classification algorithm of contours in time series, different contours under multiple frames are divided into different queues, and each queue represents the change of a water seepage area in time series. Calculate the average change rate of area and the average change rate of perimeter of each cohort in the time series, and compare with the set change rate threshold, so as to screen out the seepage area.

Step S4: When a fault is detected, generate a fault type and locate a specific location.

In the embodiment of the present invention, after the step of generating a fault type and locating to a specific location when a fault is detected, it further includes: generating alarm information, drawing a chart and visually displaying the alarm information, and the alarm information includes fault valves The fault type, time, image and video information of hall equipment.

In the embodiment of the present invention, the embodiment of the present invention detects the four faults of overheating, open flame, water seepage, and discharge synchronously in real time by analyzing the images of the infrared and ultraviolet channels. However, except for the water seepage fault, the other three faults are not independent of each other. When an open flame fault occurs, the high-temperature flame area in the infrared image will also be detected as an overheating fault, and the radiation released by the flame in the ultraviolet band will be detected as a discharge fault. This phenomenon is called overlap between faults. Overlapping faults can raise the wrong type of fault alert. The following table lists the overlap of the four fault detections.

Fault type Involved channel Overlap overheat infrared overlap with an open flame fire UV, IR Overlap with open flame, electric discharge discharge UV overlap with an open flame Seepage infrared independent no overlap

In order to eliminate the wrong detection of fault types caused by fault overlap, it is necessary to combine various types of fault detection for comprehensive analysis and decision-making. When an open flame fault occurs, due to the detection algorithm, the overheating and discharge detection will immediately judge the fault before the open flame detection. If the overheating or discharge alarm is issued immediately, the alarm type error will occur. In order to eliminate this error, it is necessary to wait for a period of time when overheating or discharge is detected, and comprehensively analyze the fault type based on the results of open flame detection. This period of time is called "conflict waiting time". If no open flame is detected after conflict waiting, Raise a response type alert. When an open flame fault is detected, it is necessary to combine the detection results of overheating and discharge faults. If there are overheating and discharge faults waiting, an open flame alarm is issued and the first two alarms are canceled.

Set up a sending buffer. When discharge or overheating is detected, a fault message, fault screenshot, and fault short video will be generated immediately, and the fault information will be added to the buffer to start conflict waiting. After the waiting is over, the fault information will be sent. When an open flame is detected When there is a fault, check the fault information in the sending buffer. When there is discharge and overheating fault information, generate and send the open flame fault information, and clear the buffer. The advantage of this is that when overheating and discharge occur, the fault information can be captured and saved in time, avoiding the loss of corresponding fault images and short videos after the waiting is over, and the fault information can be sent immediately after the waiting is over, reducing the System delay, flame detection combined with comprehensive analysis of overheating and discharge detection results, can further reduce false detection rate.

The valve hall equipment fault identification method provided in the embodiment of the present invention obtains the video stream of the equipment to be detected captured by the image acquisition device in real time, and the video stream includes: infrared images and ultraviolet images; any frame of image in the video stream Perform preprocessing, and extract the feature data of the preprocessed image; use at least one of the dynamic feature analysis algorithm and the heuristic classification algorithm to analyze the feature data, and judge whether there is a fault in the device to be detected; when a fault is detected, generate the fault type and localized to a specific location. It is proposed to analyze the characteristic data by using at least one of the dynamic characteristic analysis algorithm and the heuristic classification algorithm to analyze the characteristic data by preprocessing the infrared image and the ultraviolet image and extracting the characteristic data, so as to clarify the type of valve hall equipment failure and the location of the failure area, and improve the The reliability, scientificity and intelligence of valve hall equipment operation and maintenance.

Example 2

An embodiment of the present invention provides a fault identification system for valve hall equipment, as shown in Figure 5, including:

The image acquisition module 1 is used to acquire the video stream of the device to be detected captured by the image acquisition device in real time, and the video stream includes: infrared images and ultraviolet images; this module executes the method described in step S1 in Embodiment 1, where No longer.

The feature extraction module 2 is used to preprocess any frame of image in the video stream, and extract the feature data of the preprocessed image; this module executes the method described in step S2 in Embodiment 1, and will not repeat them here .

The fault detection module 3 is used to analyze the feature data by using at least one of the dynamic feature analysis algorithm and the heuristic classification algorithm, and judge whether there is a fault in the device to be detected; this module executes the method described in step S3 in Embodiment 1, in This will not be repeated here.

The fault identification module 4 is configured to generate a fault type and locate a specific location when a fault is detected; this module executes the method described in step S4 in Embodiment 1, which will not be repeated here.

The embodiment of the present invention provides a fault identification system for valve hall equipment, and proposes to obtain the video stream of the equipment to be detected captured by the image acquisition device in real time. The video stream includes: infrared images and ultraviolet images; any one of the video streams The frame image is preprocessed, and the feature data of the preprocessed image is extracted; the feature data is analyzed by using at least one of the dynamic feature analysis algorithm and the heuristic classification algorithm, and it is judged whether there is a fault in the equipment to be detected; when a fault is detected, a generated Fault type and locate to specific location. By preprocessing infrared images and ultraviolet images and extracting feature data, at least one of dynamic feature analysis algorithm and heuristic classification algorithm is used to analyze feature data to clarify the type of valve hall equipment failure and the location of the fault area, and improve the valve hall. The reliability, scientificity and intelligence of equipment operation and maintenance.

Example 3

An embodiment of the present invention provides a terminal, as shown in FIG. 6 , including: at least one processor 401, such as a CPU (Central Processing Unit, central processing unit), at least one communication interface 403, memory 404, and at least one communication bus 402. Wherein, the communication bus 402 is used to realize connection and communication between these components. Wherein, the communication interface 403 may include a display screen (Display) and a keyboard (Keyboard), and the optional communication interface 403 may also include a standard wired interface and a wireless interface. The memory 404 may be a high-speed RAM memory (Random Access Memory, volatile random access memory), or a non-volatile memory (non-volatile memory), such as at least one disk memory. Optionally, the memory 404 may also be at least one storage device located away from the aforementioned processor 401 . Wherein, the processor 401 may execute the valve hall equipment fault identification method in Embodiment 1. A set of program codes are stored in the memory 404, and the processor 401 invokes the program codes stored in the memory 404 to execute the valve hall equipment failure identification method in Embodiment 1. Wherein, the communication bus 402 may be a peripheral component interconnect (PCI for short) bus or an extended industry standard architecture (EISA for short) bus or the like. The communication bus 402 can be divided into address bus, data bus, control bus and so on. For ease of representation, only one line is used in FIG. 6 , but it does not mean that there is only one bus or one type of bus. Wherein, the memory 404 may include a volatile memory (English: volatile memory), such as a random-access memory (English: random-access memory, abbreviation: RAM); the memory may also include a non-volatile memory (English: non-volatile memory), such as flash memory (English: flash memory), hard disk (English: hard disk drive, abbreviation: HDD) or solid-state hard disk (English: solid-state drive, abbreviation: SSD); the memory 404 can also include the above-mentioned types combination of memory. Wherein, the processor 401 may be a central processing unit (English: central processing unit, abbreviated: CPU), a network processor (English: network processor, abbreviated: NP) or a combination of CPU and NP.

Wherein, the memory 404 may include a volatile memory (English: volatile memory), such as a random-access memory (English: random-access memory, abbreviation: RAM); the memory may also include a non-volatile memory (English: non-volatile memory), such as flash memory (English: flash memory), hard disk (English: hard diskdrive, abbreviated: HDD) or solid-state hard drive (English: solid-state drive, abbreviated: SSD); the storage 404 can also include the above-mentioned types of storage The combination.

Wherein, the processor 401 may be a central processing unit (English: central processing unit, abbreviated: CPU), a network processor (English: network processor, abbreviated: NP) or a combination of CPU and NP.

Wherein, the processor 401 may further include a hardware chip. The aforementioned hardware chip may be an application-specific integrated circuit (English: application-specific integrated circuit, abbreviation: ASIC), a programmable logic device (English: programmable logic device, abbreviation: PLD) or a combination thereof. The above-mentioned PLD can be a complex programmable logic device (English: complex programmable logic device, abbreviated: CPLD), field-programmable logic gate array (English: field-programmable gate array, abbreviated: FPGA), general array logic (English: generic array logic , Abbreviation: GAL) or any combination thereof.

Optionally, the memory 404 is also used to store program instructions. The processor 401 can invoke program instructions to implement the fault identification method for valve hall equipment in Embodiment 1 of the present application.

The embodiment of the present invention also provides a computer-readable storage medium, on which computer-executable instructions are stored, and the computer-executable instructions can execute the valve hall equipment fault identification method in Embodiment 1. Wherein, the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a flash memory (Flash Memory), a hard disk (Hard Disk) Disk Drive, abbreviation: HDD) or solid-state drive (Solid-State Drive, SSD), etc.; the storage medium may also include a combination of the above-mentioned types of memory.

Apparently, the above-mentioned embodiments are only examples for clear description, rather than limiting the implementation. For those of ordinary skill in the art, other changes or changes in different forms can be made on the basis of the above description. It is not necessary and impossible to exhaustively list all the implementation manners here. However, the obvious changes or changes derived therefrom are still within the scope of protection of the present invention.

Claims (10)

1. A valve hall apparatus fault identification method, comprising:

acquiring a video stream of equipment to be detected, which is shot by image acquisition equipment in real time, wherein the video stream comprises: infrared images and ultraviolet images;

preprocessing any frame of image in the video stream, and extracting feature data of the preprocessed image;

analyzing the feature data by at least one of a dynamic feature analysis algorithm and a heuristic classification algorithm to judge whether the equipment to be detected has faults or not;

generating a fault type and locating to a specific position when a fault is detected, wherein when a device fault is detected as a discharge type, the step of detecting the discharge fault comprises: acquiring an ultraviolet image shot by image acquisition equipment in real time, detecting suspicious discharge points in a single frame image through a preset image detection algorithm, counting ultraviolet noise distribution according to a discharge negative sample, fitting by using a model of a profile, obtaining a probability density function of noise, constructing a sequence of suspicious discharge areas in time by using a heuristic feature classification algorithm based on the probability density function, calculating the probability that the sequence is a noise sequence, judging whether discharge faults exist in equipment to be detected by comparing the probability with a corresponding preset threshold value, wherein the sequence of suspicious discharge areas in time is constructed by using the heuristic feature classification algorithm, and calculating the probability that the sequence is the noise sequence:

wherein ,tis the time difference between the current spot and the last spot in the same queue,and judging whether the equipment to be detected has discharge faults or not by comparing the probability density function with a corresponding preset threshold value.

2. The valve hall device fault identification method of claim 1, wherein when a fault is detected, the step of generating a fault type and locating to a specific location further comprises: generating alarm information, drawing a chart of the alarm information, and visually displaying the alarm information, wherein the alarm information comprises fault type, time, image and video information of the fault valve hall equipment.

3. The valve hall device fault identification method of claim 1, wherein the fault type comprises: open fire, discharge, over-heat and water seepage.

4. A valve hall equipment failure recognition method according to claim 3, wherein when an equipment failure is detected as an open flame, overheat, water seepage type, a failure area is identified in the infrared image; when a device failure is detected as a discharge type, a failure area is identified in the ultraviolet image.

5. The valve hall device fault identification method of claim 4, wherein the step of open flame fault detection comprises: the method comprises the steps of acquiring an infrared image shot by image acquisition equipment in real time, determining a flame suspicious region through a preset image processing algorithm, extracting a characteristic value of each flame at each moment of the suspicious region, inputting a dynamic characteristic pool, wherein the characteristic value comprises average circularity, average area change rate and average perimeter change rate, classifying and grouping the characteristic value by using a heuristic classification algorithm, and judging whether open flame faults exist in a plurality of equipment to be detected or not by combining a dynamic characteristic analysis algorithm.

6. The valve hall device fault identification method of claim 4 wherein the step of overheat fault detection comprises: the method comprises the steps of acquiring an infrared image shot by image acquisition equipment in real time, determining a heating suspicious region through a preset image processing algorithm, extracting a temperature value of the heating region, analyzing temperature rise, temperature difference and relative temperature difference of the heating region through a dynamic characteristic analysis algorithm, and judging whether the equipment to be detected has overheat faults or not through comparison with a corresponding preset threshold value.

7. The valve hall device fault identification method of claim 4, wherein the step of water seepage fault detection comprises: the method comprises the steps of acquiring an infrared image shot by image acquisition equipment in real time, solving a water seepage suspicious region in a single frame image by utilizing an image segmentation algorithm, calculating the average temperature difference between the suspicious region and a region boundary outer region in the region, primarily screening the suspicious region, dividing different contours under a plurality of frames into different queues by utilizing a heuristic classification algorithm of the contours on a time sequence, each queue represents the change of a water seepage region on the time sequence, calculating the average change rate of the area and the average change rate of the circumference of each queue on the time sequence, comparing with a set change rate threshold, and screening the water seepage region.

8. A valve hall device fault identification system, comprising:

the image acquisition module is used for acquiring a video stream of equipment to be detected, which is shot by the image acquisition equipment in real time, wherein the video stream comprises: infrared images and ultraviolet images;

the feature extraction module is used for preprocessing any frame of image in the video stream and extracting feature data of the preprocessed image;

the fault detection module is used for analyzing the feature data by utilizing at least one of a dynamic feature analysis algorithm and a heuristic classification algorithm and judging whether the equipment to be detected has faults or not;

the fault identification module is used for generating a fault type and positioning the fault type to a specific position when a fault is detected, wherein when the equipment fault is detected to be a discharge type, the step of detecting the discharge fault comprises the following steps: acquiring an ultraviolet image shot by image acquisition equipment in real time, detecting suspicious discharge points in a single frame image through a preset image detection algorithm, counting ultraviolet noise distribution according to a discharge negative sample, fitting by using a model of a profile, obtaining a probability density function of noise, constructing a sequence of suspicious discharge areas in time by using a heuristic feature classification algorithm based on the probability density function, calculating the probability that the sequence is a noise sequence, judging whether discharge faults exist in equipment to be detected by comparing the probability with a corresponding preset threshold value, wherein the sequence of suspicious discharge areas in time is constructed by using the heuristic feature classification algorithm, and calculating the probability that the sequence is the noise sequence:

wherein ,tis the time difference between the current spot and the last spot in the same queue,and judging whether the equipment to be detected has discharge faults or not by comparing the probability density function with a corresponding preset threshold value.

9. A terminal, comprising: at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the valve hall device fault identification method of any one of claims 1-7.

10. A computer-readable storage medium storing computer instructions for causing the computer to perform the valve hall apparatus fault identification method of any one of claims 1-7.