Inverse Conduction Method for Complex Thermal Loading

Abstract

Abstract For U.S. nuclear power plants, transfer functions have typically been utilized in fatigue monitoring applications to characterize a pressure vessel or piping component’s local, inside fluid temperatures, which are the primary forcing functions for stress cycling and the accumulation of fatigue damage, crack nucleation, and crack growth. The transfer function approach involves using simplified, conservative thermal-hydraulic engineering principles to develop algorithms that calculate the local temperatures, using existing plant instruments, such as temperatures, flow rates, valve positions, etc. With the increased ease of adding instrumentation using modern thermocouple and wireless data technology, numerous practical uses are available to plants now, such as: • More accurately and easily characterizing local thermal loading of components for analysis or fatigue monitoring, including flaw tolerance methods. • Determining the inside temperature loading in normally-stagnant branch lines subject to thermal cycling. Accurate knowledge of inside temperatures in these lines can potentially eliminate the need for costly, frequent inspections when a significant temperature range threshold is not exceeded [1] [2]. • Inside temperate loading, based on measured data, can also provide the basis for detailed evaluations in emergent situations when cracks or new loadings are discovered. With thermocouple data, temperatures on the outside surface of components are measured, but inside surface temperatures are needed to compute accurate stresses. There are challenges with attempting to deterministically calculate inside temperatures, because of numerical instability issues caused by temperature fluctuations, measurement inaccuracies, or both. The problem becomes increasingly complex when multiple instruments are required to characterize the loading, such as with thermal stratification or with thermal cycling caused by in-leakage interacting with swirl flow at a branch nozzle. Because of stability challenges and the multivariate nature of the problem, a robust, generalized solution for computing inside temperatures at multiple different thermal zones is nontrivial. Drawing on previous work, this paper discusses the challenge along with a proposed solution.