Research And Development Of Remote Monitoring, Early-Warning And Performance Test System For The Reciprocating Air Compressor

Reciprocating air compressor is a main kind of mine used air compressors, which are one of the important powers of coal mine production. They take on the task of supplying power to the whole mine's pneumatic machines and coal-whipper in main mine. The equipment's scale is so large and its price is so high, once it breaks down, it will not only cause huge economic losses but also bring hidden dangerous to coal mine production. Besides, fault diagnosis for reciprocating air compressor is complex because of its complex structure, multi-stimulations, chaos and coupling of signals. So it is neccesarry to monitor and predict for the faults of reciprocating air compressor. On the other hands, in most of native coal mines, condition monitoring, fault diagnosis and performance test for reciprocating air compressors are still in low automation degree, besides, the work condition is bad and the efficiency is low. So, it is necessary to develop a remote monitoring, early-warning and performance test system to insure the compressors in safe, reliable and economic running states.After deep investigation and study on the development of the equipment monitoring system and on its failure mechanism and fault classification, the model of remote monitoring and pre-alarm system for the reciprocating air compressor is established based on computer monitoring technology, industrial Ethernet technology, fault diagnosis technology and support vector machine technology. This system adopted ring network structure based on industry Ethernet, established data acquisition system, acquired states data of reciprocating air compressor, and transferred these data to monitoring and warning center by network, and real time condition monitoring of reciprocating air compressor are implemented. In order to detect potential fault of reciprocating air compressor, the diagnosis method based on Support Vectory Machine is suggested, which taked vibration signal of reciprocating air compressor as analysis object, established fault diagnosis model with AR model and SVM theory. Based on hardware of monitoring system, performance test software for reciprocating air compressor is developed with King View software. This software acquires datas of reciprocating air compressor, and save them to attach database. Then these data are processed and test result are analized by the software in order to finish performance test of reciprocating air compressor.This system has run stably from the October 2008 to present and achieved the expected function of condition monitoring, pre-alarming the fault and testing the performance of reciprocating air compressor, and ensured that reciprocating air compressors work safely, reliably and efficiently.