Hybrid diagnostic approach for the diagnosis of district heating networks

Decarbonization of the heating sector is a mandatory target towards climate neutrality by 2050. In this framework, District Heating Networks (DHNs) do play an important role since the heat carrier is dispatched from a heat source to end-users. Reliability of DHNs can be affected by several faults, of which the negative consequences can be prevented by employing diagnostic methodologies to evaluate the health state of the DHN and promptly localize the fault cause. In the literature, DHN faults are usually detected by means of physics-based and data-driven methodologies, but their drawbacks may hinder their application. Thus, this paper proposes a hybrid approach composed of two steps aimed at detecting the most common faults occurring in DHNs, i.e., water leakages as well as anomalous heat and pressure losses. First, a data-driven diagnostic methodology is employed to assess whether a fault is occurring. Then, a physics-based diagnostic approach identifies the health indices of each pipe of the DHN, the fault position, its type and magnitude. In this paper, the hybrid diagnostic approach is applied to the DHN of the campus of the University of Parma, where different faults were artificially implanted. The diagnostic approach proves to correctly detect and identify the implanted faults, by also correctly estimating their magnitude even in the most challenging scenarios.