Font Size: a A A

Development Of Performance Test Equipment For Engine Nozzle Based On Computer Vision Technology

Posted on:2008-09-09Degree:MasterType:Thesis
Country:ChinaCandidate:W PengFull Text:PDF
GTID:2178360278469999Subject:Thermal Engineering
Abstract/Summary:PDF Full Text Request
As a key element of engine, the performance of nozzle has directly effects on engine's operating efficiency and security. The atomization angle and uniformity tested under definite pressure and temperature are two most important indexes of nozzle performance. For a long time, these two indexes are always measured by eyeballing, which result in high labor intensity but low efficiency. Therefore, it's extremely important to develop new equipment with high level automation and convenience.In the dissertation, fanjet swirl nozzle was taken as research object. Several modern examination methods were compared and finally an intelligent analysis method based on computer vision technology was used successfully to obtain atomization angle and uniformity.Firstly, the spray outlet field distribution condition was simulated by Fluent 6.2, and the result was consistent with actual situation.Then, the dissertation gave a thorough discussion on how to automatically get atomization angle and uniformity using computer vision technology. After a series of dealing such as image enhancement, edge detection and edge thinning, binary image of atomization angle was obtained. Then Hough transform was used to get image borderline, from which atomization angle was identified; after getting the rough position of liquid-level using video frame, more accurate information was obtained by correlation measure method, and then atomization uniformity was calculated based on an emaciated table. The actual test data indicated this non-contact method has excellent examination precision.In order to guarantee stable system working pressure, PID controller based on BP neural network adjusting was adopted considering the property of controlled parameter and working situation. Then a corresponding simulation was carried out on MATLAB. The simulation result and actual test indicated that new control scheme is better than conventional PID controller.Finally, electronic-controlled injection system, computer control system and image processing system of nozzle tester were designed. Meanwhile, the corresponding monitoring and image processing program were developed combined with the concept of virtual instrument.
Keywords/Search Tags:Atomization angle, Atomization uniformity, BP neural network, Computer vision
PDF Full Text Request
Related items