Unlock instant, AI-driven research and patent intelligence for your innovation.

Minimal sufficient statistic mode analysis-based water feeding pump fault detection method

A technology for pattern analysis and fault detection, applied in computing, instrumentation, data processing applications, etc., can solve problems such as performance degradation, failure to detect faults, difficulty in obtaining fault samples, and achieve high detection accuracy.

Active Publication Date: 2018-04-17
SOUTHEAST UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

First, these methods require a large number of fault samples to train the model, and it is particularly difficult to obtain fault samples in the actual process.
Moreover, the performance of these methods drops sharply when the status signal of the feed pump is in a non-Gaussian distribution
More importantly, these methods cannot detect the fault when the fault degree is low, so that the fault can be eliminated in time

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Minimal sufficient statistic mode analysis-based water feeding pump fault detection method
  • Minimal sufficient statistic mode analysis-based water feeding pump fault detection method
  • Minimal sufficient statistic mode analysis-based water feeding pump fault detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0091] The feedwater pump fault detection method based on the analysis of the minimum sufficient statistics model includes the following steps:

[0092] Step 1: Collect steam-driven feedwater pump operation and related status data from the DCS of a 300MW unit in a certain power plant. The sampling time is 1s. The specific status variables are shown in Table 1. Collect 5000 groups of normal state, and then normalize with a mean value of 0 and standard deviation of 1, forming a training data set X∈R m×n . Among them, m=11 is the number of state variables, and n=5000 is the number of training samples. In addition, 900 sets of data on the transition from the normal state to the fault state of the two common faults of fluid cavitation and abnormal wear of the steam-driven feedwater pump were also collected to detect the effect of the method.

[0093] Table 1 Status data

[0094] Feed water pump inlet flow

Feed water pump inlet temperature

X direction of radial shaft vibration of f...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a minimal sufficient statistic mode analysis-based water feeding pump fault detection method. The method comprises the following steps of: establishing a statistic model by utilizing minimal sufficient statistic mode analysis, carrying out orthogonal transformation and minimal sufficient statistic calculation in sequence, and converting water feeding pump operation and related state parameters into statistic indexes; and modeling the statistic indexes by utilizing a principal component analysis method, calculating a T2 statistic in a principle component space, and whenthe T2 statistic exceeds a threshold value, giving a fault alarm. According to the method, the difficulty, of online water feeding pump fault detection, caused by nongaussianity and multi-state coupling of water feeding pump state data can be well solved through minimal sufficient statistic mode analysis, so that the fault detection correctness is improved.

Description

Technical field [0001] The invention relates to the field of thermal automatic control and process monitoring, in particular to a feedwater pump fault detection method based on a minimum sufficient statistical quantity mode analysis. Background technique [0002] Feedwater pumps are widely used in industrial processes such as power stations, chemicals, and pharmaceuticals. Accurate and timely detection of feedwater pump failures is an important means to ensure the safety of equipment and product quality. In recent years, especially in power station boilers, feedwater pumps have gradually become larger and more complex. Model-based fault detection methods have the problem of complicated mechanism modeling; while fault detection methods based on experience knowledge have encountered difficulties in acquiring knowledge. Therefore, data-driven fault detection methods have become the mainstream technology in the field of fault detection. However, in the actual process, the state data...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06Q10/06
CPCG06Q10/0635
Inventor 董顺李益国刘西陲沈炯潘蕾吴啸
Owner SOUTHEAST UNIV