Research And Application On The Visionbased Measuring Method For Detecting The Deformation Of The Steam Pipe

To ensure the safe operation of steam-water piping system,the creepage of the boiler's steam pipes and its displacement are most important aspects to supervise.Traditional method to measure the creepage is manually carried out when boiler unit is off-the-line,not only does it call for long period of time,also it is difficult to keep in pace with the creepage.Especially when emergency occurs,the data of creepage is difficult to sample consecutively.Those drawbacks make the manual method hard to supervise and evaluate the steam pipes' running status.To overcome the drawback mentioned above,this paper proposed a vision based methods to supervise the creepage of the boiler's steam pipes,and designed a system to utilize the mentioned vision based method.The mentioned vision based method is described as follow:First weld two mark circles on the bellowing part of the spring hanger,set each center of the circles as mark points.Using the world coordinates of the two mark points to fit the line where bellowing part of the spring hanger locates.Then solve the simultaneous equation,which is combined the pipe's contour surface equation and the mentioned straight line equation,to get the intersection of the curve and the line,and by real-time monitoring the world coordinate of this intersection,we can calculate the creepage of the pipes.Converting the locations and creepages of several spring hangers into the same coordinate,then we can analyze the changes of intersections to get the creepage between the two steam pipes or among several steam pipes.The world coordinates of the mark point is measured by the vision based method.The system is operated around the clock,in order to capture the coordinates of the mark points in real time,this paper proposed a vision based strategy to detect the mark points: First,turn the image captured by camera into grayscale image and equalize its histogram.Then de-noise the image and turn the grayscale image into binary image.Next,extract the edge information of the binary image and then use the Hough transformation to detect the mark circle and fit the world coordinate of the circle's center according the image-forming principle.This paper introduced a method to detect the circle,and several experiments are conducted and illustrate that our method is practicable.For the strong light environment and the dust stratification,which the normal method cannot detect the circle precisely.The HSV color model is usually robust to the strong light and shadow,in this way,this paper proposed an improvement that using the HSV color model to do the detecting work.To do the numerical calculations among the H,S and V components,i.e.to adjust the coefficient of this three channel,the influence from strong light and the dust and be reduced to the utmost.In practical application,our team designed a vision based measuring system to detect the creepage of the steam pipes,and it is installed at the first power unit of Zhu Hai power plant.The software is developed in VC language environment using VS2013,the background algorithm uses the Emgucv algorithms library.This system can realize the all-weather monitoring of the steam pipes' deformations and record the detecting data.