- •Thermal monitoring of low voltage switchgear using thermal ionization detector
- •An intelligent fault diagnosis method for electrical equipment using infrared images
- •The methods in infrared thermal imaging diagnosis technology of power equipment
- •B. Image Segmentation Methods
- •Condition Monitoring Tool for Electrical Equipment — Thermography
- •Equipment monitoring for temperature related failures using thermography cameras
- •About Thermal Stresses Monitoring and Diagnosis of Electrical Equipment
- •An Instrumentation System for Smart Monitoring of Surface Temperature
- •10. Electrical Contact Failure Detection Based on Dynamic Resistance Principle Component Analysis and rbf Neural Network
- •11. A Wireless and Passive Online Temperature Monitoring System for gis Based on Surface-Acoustic-Wave Sensor
- •12. Reliability assessment and field failure predictions – a prognostic model for separable electrical contacts
The methods in infrared thermal imaging diagnosis technology of power equipment
In the applications of electric power industry, the infrared thermal imaging method, which has been occupied a very important position in the predictive and preventive maintenance measure, has the advantages of non-contact, less susceptible to electromagnetic interference, safety and reliability etc. This technology can diagnostic target instant visualization, verify thermal profile, rapid position hot points, determine the severity of the problem, contribute to the establishment of equipment failure database and convenient equipment overhaul, more important is, infrared thermography can diagnosis power system when it keeps running.
Although using infrared thermography in electric equipment is very convenient, it still has many problems to be considered. The images have the characteristics of high noise, low contrast, which require preprocessing; in the acquisition of images, there are interferences about background, experimental simulation conductor, conductor, other equipment and so on, so that we must choose the suitable method to segment target area, then using the appropriate algorithm to extract the image features; finally, using intelligent algorithms to diagnosis faults, if there is no proper algorithm, that will lead to low efficiency of monitoring, judgment error and incorrect diagnosis.
The good infrared diagnosis system must solve the problems about instrument accuracy, image analysis and fault classification. In view of the above, this paper will focus on the description of hotspots, difficulties and future trends in infrared thermal image processing, intelligent fault diagnosis research.
At present, the intelligent diagnosis system of thermal infrared images in electrical equipment usually consists of five main steps, as shown in Figure 1. The first step is to get images by thermal imager for electrical equipment in different experimental conditions, then dealing with denoising noise, segmentation and feature extraction in image processing, finally generalizing analysis by using artificial intelligence algorithms and determining whether it has thermal fault and draw the conclusion.
A. Denoising Noise Methods
Infrared image has the characteristics of high noise, low contrast. The noise sources include infrared focal plane array detector, electronic sensor, background etc. The three types of main noise including shot noise, thermal noise and 1/f noise, their probability statistical characteristics obey Gauss distribution with part of the impulse noise. These noises result in image degradation and image features be covered, which directly affect the accuracy of image segmentation, feature extraction and subsequent work. Therefore, denoising image and enhance the image quality are the premise steps of image processing and analysis.
1) Single Denoising Algorithm
Single denoising algorithm is usually for a certain class of noise removal, such as linear smoothing is better on the removal of Gauss noise, but it's not very effective for removing impulse noise, similarly, median filter is suitable for removing impulse noise, but not applicable to Gauss noise.
2) Multiple Denoising Algorithms
In order to eliminate the confused noise, recently a large number of improved algorithms based on linear filter and median filterhave been emerged, such as weighted mean filtermethod, gradient weighted average method, the fuzzy weighted average filtermethod, weighted median filtermethod, hybrid median filtermethod. Li Wenjie had proposed a median filtering and wavelet adaptive diffusion denoising method. The experimental results show that, the method was better than the median filter, neighborhood average method, Wiener filter, wavelet threshold method, could effectively remove the Gauss noise and pulse noise, but also could preserve the details of images and have good denoising performance.
