Principal element degree of association sensor fault detection method and apparatus based on density clustering

A technology of sensor failure and density clustering, which is applied in the direction of instruments, character and pattern recognition, computer components, etc., can solve problems such as difficulty in detecting sensor failure, performance degradation of control system, economic loss, etc., and achieve the effect of fast and accurate fault diagnosis

CN105894027AActive Publication Date: 2016-08-24NORTH CHINA ELECTRICAL POWER RES INST +3
8 Cites 16 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2016-08-24

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention relates to a principal element degree of association sensor fault detection method and apparatus based on density clustering. The method comprises steps of: determining the operating condition data of monitoring sensors by using a unit multi-operating condition model and classifying the monitoring sensors to obtain an operating condition cluster, wherein the operating condition information of all monitoring sensors in the operating condition cluster forms a matrix X; forming a matrix K by using the operating condition information of any two monitoring sensors in the operating condition cluster; analyzing the principal element of the matrix K to obtain the major characteristic T of the matrix K; determining the degree of association of the any two monitoring sensors by using the normalized matrix X and the major characteristic T; and comparing the degree of association with a threshold value and detecting the fault of the monitoring sensor corresponding to the degree of association greater than the threshold value.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical field of fault detection, in particular to a method and device for fault detection of a principal component correlation sensor based on density clustering. Background technique

[0002] As an essential underlying component in thermal power plants, sensors play an important role in the safe and stable operation of the unit. During the normal operation of thermal power units, a large number of various types of sensors are used to measure important thermal process parameters, such as main steam temperature, main steam pressure, steam turbine speed, and drum water level. Once the sensor breaks down, the performance of the control system will be degraded, or it may cause serious accidents and cause major economic losses. There are many sensors in the thermal process of thermal power plants, which are distributed in multiple parts of various equipment. It is very difficult to detect sensor failures by manpower. Theref...

Examples

Embodiment

[0095] Randomly select all measuring points within a certain period of time for clustering, and divide them into several typical working condition clusters. 500 sets of data are used as the fixed time window length, and the last 500 sets of data are used as test data. In this case, the test data is superimposed with 5% deviation fault data, and the average length of the sliding window is selected as 5. After combining the fixed window and the sliding window, real-time Computes the correlation between the measured points within the window.

[0096] exist Figure 7 ~ Figure 11 In , the meaning expressed by the axis of ordinate is: the correlation degree between measuring points, and the meaning expressed by the axis of abscissa is; time series. For example: 1 represents time 1, 2 represents time 2, and so on.

[0097] Figure 7 It is the change curve of the correlation between measuring point 1 and measuring point 2. It can be seen that the correlation between measuring point...