Predictive Probability of Detection Curves for Scaling Monitoring in Geothermal Pipelines using Ultrasonic Testing

Accumulated deposits in pipelines, such as scaling, are a major concern across industries. Numerous approaches are used to monitor scale buildup and to verify the effectiveness of descaling measures. This is particularly challenging in geothermal plants, where testing techniques capable of enduring high-temperature conditions are required. And while developing appropriate non-destructive inspection techniques is crucial, suitable tools to assess the reliability of these techniques are just as important. Probability of detection (POD) curves are standard tools for reliability assessment, but they would require a substantial amount of data from pipes with different scaling degrees. The so-called predictive probability of detection (P-POD) curves are a suitable alternative, as they consider the statistical variability in the measurements and can be generated using only information from a single measurement record from the undamaged state. This paper is a proof-of-concept study, where P-POD is applied to scaling monitoring for the first time. The dataset was recorded using ultrasonic testing in a pulse-echo configuration on a heavily scaled pipeline section from the geothermal power plant in Sauerlach, Germany. The analysis was based on time-of-flight considerations. P-POD curves and confidence intervals were constructed for both increasing (i.e., scaling) and decreasing (i.e., descaling) thickness changes, using a single reference measurement for each case. Subsequently, the predictions were validated with experimental data. The results demonstrate the accuracy of the predicted POD. The successful application of the P-POD method translates into a significant contribution to pipeline monitoring in general, where reliable fouling inspection is key for minimizing operational difficulties.