The Research On Corrosion Detection And Life Prediction For Grounding Grids

Grounding grids is an indispensable establishment in power stations and substations. Itplays a vital role in keeping stablility and safety of power system, and in ensuring the securityof operators and power equipments. In our country, corrosion of the grounding grids whichare made of steel or galvanized steel would occur on the conductor branches after severalyears of operation. When it is serious, it will make grounding parameters worsen. Then thesafety and stability of the power system will be in danger directly. So, it is urgent to develop amethod to detect the corrosion status of grounding grids without excavation or power off, thentake relevant defensive measures to avoid the potential accidents.By introducing corrosion mechanism of grounding body and analyzing weaknesses ofexisting grounding grid detecting approaches, a corrosoion detection method for groundinggrids is proposed based on quasi-steady measurement, linear polarization technology andwavelet de-noising method. Equivalent circuit models are constructed by simulating thecorrosion of grounding grids, corrosion rate and corrosion depth of grid conductors areobtained by using quasi-steady-state measurement and then the corrosion states of groundinggrids are recognised. Furthermore, wavelet transform and hardware filter are used to filter theincidental noises in test signals since there are a mass of strong noises in the test enviroment.Based on the above principle, the corrosion detection system developed by our team isintroduced. PC104is used as the core-module of the hardware parts, and sub-modules consistof resistance measurement module, quasi-steady-state measurement module and signalprocessing module. The whole software parts is developed by C language, and composed ofsignal processing, data processing, data management, human-computer inerface and so on.Meanwhile, the reliability design of this system is studied in both hardware and software.Furthermore, by studying the application of grey forecasting theory, geneticalgorithm and BP neural network in the corrosion rate prediction of groungding grids,a combined model is developed. The practical example results show this model isfeasible and accurate. Based on this, the remaining life prediction of groungding gridsis preliminarily studied, and then the basic flow of remaining life prediction is given,which can provide some reference to the routine maintenance of grounding grids.