Parameter estimation in heat transfer through an overhead power cable by using the markov chain monte carlo method
Farith M. Absi Salas, Helcio R. B. Orlande And Luis A. M. C. Domingues
In this work, we apply a Bayesian approach for the estimation of the thermal parameters appearing in the heat conduction problem in overhead power cables. The cables used for the transport of electric energy are composed by layers of stranded metallic wires, resulting in a heterogeneous media (wires and air voids). The inner core is made of steel wires, over which three aluminum wire layers are winded to form the conductor. The physical problem is formulated in terms of transient heat conduction in two homogeneous cylindrical regions. The Markov Chain Monte Carlo (MCMC) method is used to estimate the effective radial thermal conductivities and the volumetric heat capacities of both regions (steel core and aluminum conductor), the combined heat transfer coefficient at the cable surface, as well as the parameters related to the electrical current distribution in the conductor region. Temperature measurements were used in the inverse analysis, taken by thermocouples inserted in the cable together with the conductor wires during its manufacturing process, in experiments under controlled laboratory conditions.
Keywords: Thermal properties, Bayesian framework, MCMC method, ACSR cable, transformer effect