A test bench for conducting compressor surge experiments is set up, and different system configurations formed by changing the length of compressor outlet pipeline are tested for surge. Dynamic pressure signals relati...A test bench for conducting compressor surge experiments is set up, and different system configurations formed by changing the length of compressor outlet pipeline are tested for surge. Dynamic pressure signals relating to surges are acquired at different locations of the configurations using unsteady measurement ＆ data acquisition system. The sliding window method is adopted to set up quantitative criterion on the surge. Parameters included in the criterion, such as location of data collection, size and step of sliding window, a mathematical quantity surge-judging and its threshold, etc., are given. Flow chart of surge evaluation is shown, and surge frequency was evaluated based on system configurations. With all these, the problem of judging the existence of surge by human experiences in compressor performance experiments can be solved. Hence this new approach may help to achieve intelligent operations on automatic compressor performance testrig.展开更多
基金Sponsored by the National Natural Science Foundation of China (50676011)
文摘A test bench for conducting compressor surge experiments is set up, and different system configurations formed by changing the length of compressor outlet pipeline are tested for surge. Dynamic pressure signals relating to surges are acquired at different locations of the configurations using unsteady measurement ＆ data acquisition system. The sliding window method is adopted to set up quantitative criterion on the surge. Parameters included in the criterion, such as location of data collection, size and step of sliding window, a mathematical quantity surge-judging and its threshold, etc., are given. Flow chart of surge evaluation is shown, and surge frequency was evaluated based on system configurations. With all these, the problem of judging the existence of surge by human experiences in compressor performance experiments can be solved. Hence this new approach may help to achieve intelligent operations on automatic compressor performance testrig.
