CN-121977169-A - Collaborative diagnosis method and system for leakage of pulverized coal conveying pipeline of thermal power plant

CN121977169ACN 121977169 ACN121977169 ACN 121977169ACN-121977169-A

Abstract

The application discloses a collaborative diagnosis method and a collaborative diagnosis system for leakage of a pulverized coal conveying pipeline of a thermal power plant. The method comprises the steps of obtaining a reference wind pressure value and calculating a wind pressure descending rate threshold value, collecting wind pressure data in a pipeline, calculating a real-time wind pressure change rate, judging that leakage is suspected when the real-time wind pressure change rate exceeds the wind pressure descending rate threshold value, generating a rechecking request, calling video data corresponding to a leakage suspected point to calculate gray standard deviation, transmittance and visibility descending rate of a leakage area in a video segment, constructing a visual feature vector, carrying out time sequence feature matching on a wind pressure change sequence and the visual feature vector sequence in a suspected leakage period, calculating correlation coefficients of the wind pressure change sequence and the visual feature vector sequence, judging that leakage is true if the wind pressure change rate exceeds the wind pressure descending rate threshold value, and judging that leakage is true if the wind pressure change rate exceeds the wind pressure descending rate threshold value, closing a control instruction of a pipeline valve corresponding to the leakage point and triggering an alarm. The method can solve the difficult problems of frequent false alarm and response lag.

Inventors

Jiao Lizhang
LIU YAN
ZHAO WENQUAN
LI ZHENFU
ZHOU YUJIAN
JIANG ZHIHAO
ZHAO PENGCHENG
CHEN KUNMING

Assignees

天津国能津能热电有限公司

Dates

Publication Date: 20260505
Application Date: 20251222

Claims (10)

1. The collaborative diagnosis method for the leakage of the pulverized coal conveying pipeline of the thermal power plant is characterized by comprising the following steps of: Step one, obtaining a reference wind pressure value of a pulverized coal conveying pipeline in a normal running state, and calculating a wind pressure drop rate threshold value according to the reference wind pressure value; step two, acquiring wind pressure data in a pipeline in real time, calculating a real-time wind pressure change rate, and judging that the pipeline is suspected to leak and generating a rechecking request when the real-time wind pressure change rate exceeds the wind pressure reduction rate threshold; step three, responding to the rechecking request, calling video data corresponding to the suspected leakage point, calculating gray standard deviation, transmissivity and visibility reduction rate of a leakage area in the video segment, and constructing a visual feature vector; Step four, carrying out time sequence feature matching on the wind pressure change sequence and the visual feature vector sequence in the suspected leakage period, and calculating the correlation coefficient of the wind pressure change sequence and the visual feature vector sequence; and fifthly, if the real leakage is judged, a control instruction for closing the pipeline valve corresponding to the leakage point is sent out, and an alarm is triggered.
2. The collaborative diagnosis method for leakage of pulverized coal conveying pipelines of a thermal power plant according to claim 1, wherein in the first step, the method for calculating the wind pressure drop rate threshold is as follows: ; Wherein: a wind pressure drop rate threshold value in unit time; For the preset sensitivity coefficient, the value range is taken A preset sampling period; 。
3. the method for collaborative diagnosis of pulverized coal transport pipeline leakage in a thermal power plant according to claim 2, further comprising: Real-time computing sliding window Rate of change of wind pressure in If | |> Triggering a suspected leakage state and generating a rechecking request, wherein the rechecking request comprises a time stamp, a pipeline position and an abnormal feature vector.
4. A thermal power plant pulverized coal transport pipe leakage collaborative diagnosis method according to claim 3, characterized in that step two further comprises: And if the absolute value of the second derivative is larger than a preset change rate acceleration threshold, judging that the pressure is suddenly changed, and triggering a suspected leakage state.
5. The collaborative diagnosis method for leakage of a coal dust conveying pipeline of a thermal power plant according to claim 1, wherein in the third step, gray standard deviation is calculated, and a method for establishing a coal dust concentration and visual characteristic mapping model is as follows: ; c is the gray standard deviation of an image, and the gray standard deviation C positively represents the coal dust concentration in the area; The gray value of the ith pixel point; is the average gray value for that region.
6. The collaborative diagnosis method for leakage of a pulverized coal conveying pipeline of a thermal power plant according to claim 5, wherein in the third step, a dark channel prior model method is adopted to evaluate the transmissivity, and if the transmissivity is less than 0.4 and the visibility decrease rate continuously exceeds a preset threshold, the gray standard deviation C, the transmissivity and the visibility decrease rate are stored in the visual feature vector together.
7. The collaborative diagnosis method for leakage of pulverized coal conveying pipes of a thermal power plant according to claim 6, wherein in the fourth step, the timing feature matching of the suspected leakage data and the visual feature vector is performed by: Selecting the nearest A number of sampling points, wherein, ; Acquiring a wind pressure signal sequence And visual characteristic sequence ; Calculating a wind pressure average value: ; Calculating a visual characteristic mean value: ; Calculating a correlation coefficient : ; If it is Then a true leak is determined.
8. The method for collaborative diagnosis of pulverized coal feed line leakage in a thermal power plant according to claim 2, wherein in step one, the sensitivity coefficient is reset every time the system is restarted or zero daily Is an initial value of 0.1.
9. The collaborative diagnosis method for pulverized coal conveying pipeline leakage of a thermal power plant according to claim 4, further comprising the step of sequentially calling video data for rechecking according to the order of the descending rate of wind pressure of each point when a plurality of suspected leakage points are monitored at the same time.
10. A thermal power plant pulverized coal transport pipe leakage cooperative diagnosis system applied to the thermal power plant pulverized coal transport pipe leakage cooperative diagnosis method according to any one of claims 1 to 9, characterized by comprising: The detection unit comprises a pressure sensor and a data acquisition module which are arranged on a preset node of the pulverized coal conveying pipeline; The video acquisition system comprises a camera, a video storage unit, a purging device and an infrared illumination module, wherein the camera is arranged in a leakage-prone area on a pulverized coal conveying pipeline; The cooperative processing unit is used for being interconnected with the detection unit and the video acquisition system and executing preset control logic; And the execution unit is connected with the cooperative processing unit and is used for receiving related instructions and comprises an isolation valve, an alarm, an execution output module and an operation and maintenance terminal.

Description

Collaborative diagnosis method and system for leakage of pulverized coal conveying pipeline of thermal power plant Technical Field The application relates to the technical field of safety monitoring of thermal power plants, in particular to a thermal power plant pulverized coal conveying pipeline leakage collaborative diagnosis method and system based on time sequence feature matching and dynamic threshold. Background The coal powder conveying system is used as a core link of power plant fuel supply and is responsible for conveying qualified coal powder into a boiler burner in a pneumatic conveying mode, and the operation safety of the coal powder conveying system is directly related to the stable operation of a unit and the safety of personnel and equipment. As the pulverized coal has the characteristics of small particle size, strong fluidity, inflammability, explosiveness and the like, once a conveying pipeline leaks, the leaked pulverized coal forms an explosive mixture in the air, dust explosion is easily caused when the pulverized coal meets an ignition source, and equipment abrasion, environmental pollution and combustion efficiency are reduced, so that the method has important engineering significance and safety value for accurately, timely and reliably monitoring pipeline leakage. Currently, a sensor monitoring scheme based on a Distributed Control System (DCS) is commonly adopted in a thermal power plant, and whether leakage occurs is judged by acquiring parameter changes such as wind pressure, wind speed and the like through a pressure transmitter, a wind speed sensor and the like arranged on a conveying pipeline. In addition, some power plants attempt to introduce video monitoring aids, and the key areas are identified for coal dust leakage by using an AI image identification technology (such as a YOLOv-based visual detection model). However, the prior art scheme has obvious defects that on one hand, a DCS system depends on single type sensor data, a coal dust environment is extremely easy to cause dust accumulation and blockage or signal drift of the sensor, the actual measurement failure rate exceeds 35% and causes monthly false alarm to be dozens of times, on the other hand, a video monitoring system and the DCS are isolated from each other, a video is required to be manually called for video rechecking after alarm triggering, the average response time is as long as 8-12 minutes, and the timeliness and credibility of the monitoring system are seriously weakened by far exceeding a gold response window (3 minutes) required by leakage treatment. The more prominent problem is that the existing monitoring mechanism lacks multisource information fusion and dynamic adaptation capability. Firstly, the full-quantity deployment AI vision analysis needs to process 7X 24 hours video streams of more than 200 cameras, the single-path calculation force requirement is more than or equal to 8TOPS, so that more than 50 edge servers are required to be configured, serious resource waste and calculation force bottlenecks are caused, secondly, the existing method adopts fixed threshold criteria (such as wind pressure drop is more than 20 percent, namely alarm), the existing method cannot adapt to dynamic working conditions such as unit start-stop, variable load and the like, the false alarm rate is as high as 38.5 percent, and in the existing technology, although a leakage detection method based on image recognition is provided, only static frame analysis is carried out, the time sequence evolution characteristics of pulverized coal leakage are not modeled, instantaneous interference and real leakage are difficult to distinguish, and space-time alignment and association analysis between DCS parameters and vision characteristics are not realized, so that the scientific basis is lacked in the rechecking process. Disclosure of Invention The application mainly aims to provide a thermal power plant pulverized coal conveying pipeline leakage collaborative diagnosis method and a system based on time sequence feature matching and dynamic threshold value, which are used for solving the key technical problems of false alarm frequently, response lag, system cracking and resource waste in the prior art. In order to achieve the above purpose, the first aspect of the application provides a collaborative diagnosis method for leakage of a coal dust conveying pipeline of a thermal power plant, which comprises the following steps of firstly obtaining a reference wind pressure value of the coal dust conveying pipeline in a normal operation state and calculating a wind pressure drop rate threshold value according to the reference wind pressure value, secondly collecting wind pressure data in the pipeline in real time, calculating a real-time wind pressure change rate, judging as suspected leakage and generating a rechecking request when the real-time wind pressure change rate exceeds the wind pressure drop rate threshold value, third