Real-Time Mode of Operation Data Analysis to Catch the Thread-Tip Denotes the Failure Cause of the Grid- Tie PV Central Inverter

Abstract

Abstract The inverter is considered the core of the PV power plant. The inverters failure leads to generation loss and decreasing the plant availability. So, it is required to investigate a clear Root Cause Analysis (RCA) to deduce the failure causes and implement the required corrective action in addition to the preventive action to avoid more inverter failure, hereby maintain the plant availability to the certain value. This paper discusses real-time mode operation data analysis of the PV grid-connected inverter due to real central inverter incidents in Benban solar park located in Egypt. Background: There is no doubt that the central inverter plays an important role in the Mega-Scale PV power plant. The main function of this inverter is to deliver the DC power produced by the PV modules and convert it to AC power to be injected into the utility grid at certain limits of characteristics based on the agreed grid code. The availability of any PV power plant directly depends on the healthy inverter’s operation. The more increasing installed inverters, the less availability loss in the cause of inverter partial or catastrophic failures. So, it is required to focus on the failure causes of the central inverter by implementing a technical analysis using the available operational data. The monitored data of the central inverter in the PV power plant is classified into two types. The first type is the continuous time data stored in the memory. It represents the waveforms of inverter outputs like voltage, current, frequency, …. etc. Unfortunately, in case of a catastrophic failure, the central inverter is completely charred, and the continuous time data is lost due to damage to the storage memory. The second type is the operation data that is recorded by the SCADA system (one minute). Hereby, the operation data is the sole available data in the case of the inverter completely burning. Methodology: This paper discusses real-time mode operation data analysis of the PV grid-connected inverter by using mathematical analysis to deduce the IGBT junction temperature rather than the measured heatsink temperature. The processes start by data gathering, then data representation to capture symptoms indicate to the fault signatures which consider the thread-tip Denotes the Failure Cause. The data processing is performed through the mathematical model to deduce the unmeasured parameters. The expected scenarios for the failure shall be verified and approved to deduce a real RCA. In addition to detecting the signature of the IGBT thermal stress on the voltage balance of the DC link capacitor. Further, interpretation for DC link capacitors explosion. Results: The investigation outcomes include three results. The first result is detectingthe signature of the IGBT thermal stress on the voltage balance of the DC link capacitor. The second result is verifying a scenario for the cause of the IGBT failure by implementing a technical mathematical model based on the detected symptoms that denote the fault signature which is considered the thread-tip for detect the failure cause. The third result is the simulating scenario for the interpretation of a DC link capacitors explosion due to the short circuit fault. Conclusions: The investigation in this paper is performed based on operation data analysis of the PV grid-connected inverter due to a real inverter incident. The analysis methodology is based on mathematical calculation for the IGBT junction temperature using the measured heatsink temperature. The study concludes that after the IGBT failure, it was a short circuit for a while and closed the terminals of the DC link capacitors. So, the DC link capacitors exploded and produces heavy sparks that led to enough fire to burn the inverter container completely.