Shaft vibration alarm strategy for motor-driven centrifugal compressor set based on single-node ant colony algorithm

The motor-driven centrifugal compressor set may have some faults that can cause abnormal change in direct amplitude of shaft vibration during the steady-state operation. However, such a change does not necessarily trigger the alarm, so it is difficult to find the faults in time, which may affect the safe operation of equipment. In order to ensure that the alarm information can be released in time when abnormal change occurs in the direct amplitude of shaft vibration, an alarm strategy based on single-node ant colony algorithm was designed by using the rotation speed and the direct amplitude of shaft vibration as the key parameters, in combination with other parameters such as pheromone, fault tolerance radius and extended range. This alarm strategy was verified by a motor-driven centrifugal compressor set in steady-state operation. The results show that the faults such as friction and surge can be identified with this algorithm rapidly and accurately, which is fault tolerant to the mechanical, electrical and technical jittering while identifying the abnormity of direct amplitude of shaft vibration. In addition, this algorithm has a low rate of false alarm and could effectively guarantee the safe operation of motor-driven centrifugal compressor set by issuing alarm information accurately and timely.