Health-monitoring of power semiconductor modules in photovoltaic inverters – MEMPHIS

The aim of MEMPHIS project is to develop a set of tools to monitor the health of power modules. The primary effort is to provide tools applicable to the majority of commercial inverters used in photovoltaic power systems.
The path toward a solar-powered world requires efforts both in PV systems cost reduction and energy production optimization. Reliability and predictive maintenance are essential elements to reach these goals. In fact higher reliability leads to lower Operation and Maintenance (O&M) costs as well as longer system lifetime and higher availability, which eventually results in low-cost solar electricity.
Inverter is considered to be the weakest link in photovoltaic systems; as a matter of fact it is the very first element to fail, accordingly photovoltaic inverter reliability is a major issue in PV systems growth.

Our aim is to achieve results that can be quickly transferred to industries and at the same time anticipate further market developments and needs. Therefore the project is composed of two parallel packages involving different activities and different risk levels. Both packages focus on the indicators study of power modules degradation, although they follow different directions: the first involves research on solutions very close to the application-implying a low risk- while the other explores less-known and more complex methods in order to find more efficient solutions –implying a higher risk.

For the first package we use the fact that PV inverters do not work at night; thus it is possible to perform OFF-Line measurements using methods similar to laboratory ones. For example we could measure power devices forward voltage and thermal resistance or impedance, that require localized injections of current. In lab this can be easily done using an external power source, while to perform these measurements on an inverter at night we need a way to take energy from the grid without decreasing the overall efficiency of the inverter. Moreover the designed solution should be applicable to the majority of solar inverters.

In the second package, we aim to study indicators for ON-Line measurements. Possibly these indicators should be measureable by currently integrated sensors in today inverters. The studied indicators are electrical parameters that vary with the ageing of the power device. Besides the problem when using electrical parameters to evaluate ageing is that they usually are temperature dependent. This is why an important task of the project is to develop an ON-line temperature sensor, ideally using available electrical parameters. To validate the selected temperature sensor dedicated test devices (power modules suitable for infrared measurements, instrumented semiconductor chips) will be implemented.

Finally the project will be completed by the implementation and integration of both packages results on a demonstrator suitable for photovoltaic applications.