Improved inspection data quality using modified corrosion rate method for offshore pipeline assessment

Abstract

Data quality is crucial to any data analysis task. Information collected from many channels prone to disturbance, inconsistent, missing values and redundant information. In our case, these errors arise in metal loss data collected at different point of time using dissimilar sensors and devices in offshore pipeline structure. Furthermore, data collection and analysis are often time consuming and expensive making it undesirable for recollection. Thus, rather than discard the corrupted data, we need to evaluate and enhanced its quality by correcting the errors as much as possible. In this paper, we discuss how data is pre-processed by means of correcting the data to make it ready for further analysis. A modified corrosion rate method was used to enhance the quality of data coupled with a linear prediction method to verify the accuracy of the corrected data. Result shows that the proposed method can minimize the effects of uncertainties on the reliability of the inspection data.Data quality is crucial to any data analysis task. Information collected from many channels prone to disturbance, inconsistent, missing values and redundant information. In our case, these errors arise in metal loss data collected at different point of time using dissimilar sensors and devices in offshore pipeline structure. Furthermore, data collection and analysis are often time consuming and expensive making it undesirable for recollection. Thus, rather than discard the corrupted data, we need to evaluate and enhanced its quality by correcting the errors as much as possible. In this paper, we discuss how data is pre-processed by means of correcting the data to make it ready for further analysis. A modified corrosion rate method was used to enhance the quality of data coupled with a linear prediction method to verify the accuracy of the corrected data. Result shows that the proposed method can minimize the effects of uncertainties on the reliability of the inspection data.