Gaussian residual regression method for predicting internal corrosion in pipelines

Abstract

The application discloses a Gauss residual regression method for predicting internal corrosion of a pipeline, and comprises the following steps: step one, constructing a state transfer equation of corrosion time; step two, predicting corrosion depth by using a GPR model to obtain a first corrosion depth prediction value; step three, predicting new corrosion time by using a power law model to obtain a new log corrosion time prediction value; step four, calculating a residual error; step five, calculating a weight sequence by using the residual error; step six, least square optimization, optimizing and updating model parameters by minimizing the residual error square sum between actual observation values and model prediction values, and obtaining a second corrosion depth prediction value by using the GPR model after parameter updating; and step seven, correcting the prediction result to obtain a corrected corrosion depth prediction result. The application effectively fuses a physical process based on a corrosion mechanism and a Gauss process regression based on data driving, and considers the influence of residual errors, so that the precision of pipeline corrosion depth prediction is significantly improved.

Images