Numerical Approach for Predicting Corrosion Rate of Buried Earth Electrodes Using Backfill Material

Abstract

The issue of corrosion has been an inevitable canker that keeps affecting all facets of human lives; from industry, utilities, transport, production, to manufacturing fields. Issues of corrosion in sectors like oil, water, gas, etc. have been tackled, leaving behind electricity. The generation, transmission and distribution systems have well-installed backfilled earthing systems for protection against all abnormalities such as fire, shocks and electrocution which are very fatal, but much attention has not been given to the protection at the consumers’ end. There are faulty, improperly installed, corroded or absence of earthing systems at the consumers’ end. Some backfill materials have been applied to some rods to reduce earth resistance and minimise corrosion of these earthing systems. It also seeks to tackle the usage of numerical approach for reducing corrosion of buried earth electrodes using two artificial intelligence techniques. Five copper-coated cast-iron rods were utilised for this investigation; one (reference) was buried without backfill material whereas the other four were buried with tyre ash, palm kernel cake, charred coconut husk and coconut coir as backfill materials for twenty-two months. A visual inspection along with measurement of thickness, analyses by X-Ray Fluorescence Spectroscopy (XRF), and Scanning Electron Microscopy coupled with backscattered Energy Dispersive Spectroscopy (SEM-EDS) were conducted to determine the extent of corrosion. Six parameters, namely, resistance, rainfall, temperature, backfill materials, time (weeks), and corrosion rates were used to compare the results from the use of Mamdani Fuzzy Logic and Artificial Neural Network (ANN) techniques. The results obtained showed a colour change of all the rods from reddish-brown to grey, suggesting that corrosion might have taken place. This was confirmed by the increase in thickness for all the buried rods as against the unburied rod. The XRF analysis recorded an increase in the percentage values of the iron content for all buried rods with charred coconut husk recording the highest value of 99.42% and 88.58% as least for the unburied rod. SEM-EDS results showed a percentage of oxygen in all the buried rods’ analyses. The rod buried in tyre ash backfill recorded 29.73% for the highest oxygen value with no oxygen value for the reference. These were proof of corrosion activity for the two analyses. Lastly, the ANN had a lower Mean Square Error (MSE) of 0.0865 compared to that of the Mamdani Fuzzy Logic Model 0.3583, showing that ANN performed better than the Mamdani Fuzzy Logic Model when both techniques were used to predict the corrosion rates of the rods. Out of the four backfill materials used, charred coconut husk gave the best results in terms of reduced resistance values and lower rate of corrosion value.