Development of a Network Architecture for Real-Time Monitoring of the Condition of Centralized Inverters

Abstract

Relevance: in power plants based on renewable energy sources, centralized inverters represent a key energy technological component. During operation, they are frequently exposed to deviations and failures due to load variations, environmental influences, thermal stresses, and aging of power electronic components. Therefore,
comprehensive real-time monitoring of the electrical, thermal, and operational condition of centralized inverters, control of data quality (DQ), identification of risk indicators (FT, FU, FI), and state classification based on the reliability index (HI) are of significant scientific and practical importance. This approach enables improvement of
inverter reliability, early detection of emergency conditions, optimization of maintenance activities, and prompt operational decision-making.
Aim: to develop a network architecture based on real-time acquisition and processing of electrical, thermal, and operational parameters of centralized inverters; to design monitoring and diagnostic algorithms that provide data quality assessment, detection of critical conditions (WARNING/CRITICAL), and classification of inverter states
in the “Green/Warn/Red” format based on the reliability index HI(t).
Methods: in this study, the condition of centralized inverters was evaluated using the following methods: a parameter vector x(t) was formed based on the main electrical, thermal, and operational parameters obtained from the inverter (Udc, Idc, Uac, Iac, P, Q, cosφ, f, Ths, Tigbt, RPM, Fault). The collected data were subjected to quality control through time synchronization, range validation, and anomaly detection.
Results: as a result of the proposed approach, a structural model, logical scheme and algorithm for real-time monitoring of centralized inverters were developed. A data quality control and recovery (filter/recovery) module was implemented, reducing erroneous decisions caused by unreliable data. Based on the analysis of electrical and thermal parameters, early detection of hazardous conditions was achieved, and an automatic transition logic to the “WARNING/CRITICAL” mode was established. The principle of color-based classification (Green/Warn/Red)
and automatic notification based on the HI(t) index was substantiated. Under laboratory conditions, the TEX 6.2KBGN inverter was selected, parameter dynamics were investigated under various operating modes, and an informative set of indicators for monitoring was formed.