Fan condition monitoring system based on physical structure of gearbox

A centralized and unified monitoring center is proposed to complete the functions of real-time data acquisition and remote centralized control. The application of wind turbine fault monitoring system based on cloud computing is proposed. It is proposed that one of the development directions of condition monitoring and fault diagnosis technology in the future is to realize the communication between remote fault diagnosis systems through network connection to form an open fault diagnosis system. Therefore, the establishment of a distributed fan condition monitoring system is a general fan health monitoring method.

The general fan condition monitoring system periodically collects vibration data by arranging multiple physical sensors on the transmission chain of each fan, and transmits them to the wind field control center in real time through optical fiber network. The wind farm control center stores and analyzes the original data locally, judges the health status of the fan drive chain according to the analysis results, and gives fault maintenance suggestions, so as to realize the real-time monitoring of the fan health status.

In the distributed condition monitoring system of wind turbine, the general data processing flow is as follows

(1) The wind farm management center remotely controls the working status and configuration parameters of data acquisition equipment in CMS;

(2) The data acquisition equipment transmits the time domain discrete data generated by the front-end sensors to the wind farm control center, which can be directly transmitted or stored and then transmitted;

(3) The wind farm control center receives the original data, carries out data preprocessing, analysis and management, and judges the health status of the fan according to the analysis and statistics.

Aiming at the common gearbox structure, in order to ensure the integrity and effectiveness of data, generally eight vibration sensors are selected, and the sampling rate of each channel is 51.2 kHz to obtain more complete frequency range information. In order to obtain higher analysis frequency, the acquisition time is set to 60 s, and the data capacity generated by each acquisition is nearly 200 MB. It can be seen that the data generated by the front-end has the characteristics of large volume and high speed, and multiple fans have the characteristics of strong concurrency in data transmission. In addition, in order to accumulate raw data, all storage needs a large enough storage space.

Therefore, in the existing wind farm network environment, an effective data preprocessing method is needed, which can reduce the amount of data and maximize the retention of effective information, so as to reduce the occupation of network bandwidth, reduce the storage space and improve the analysis efficiency.