Integrating a real-time Tvj calculation into an IPM

Konferenz: PCIM Europe 2016 - International Exhibition and Conference for Power Electronics, Intelligent Motion, Renewable Energy and Energy Management
10.05.2016 - 12.05.2016 in Nürnberg, Deutschland

Tagungsband: PCIM Europe 2016

Seiten: 7Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Autoren:
Schmies, Stefan; Lahl, Peter; Kruschel, Wolfram; Lassmann, Matthias (Infineon Technologies AG, Germany)

Inhalt:
In power semiconductors, the junction temperature is one of the main specification limits for the achievable output power. To maximize the power density, systems are designed corresponding to the maximum junction temperature allowed, taking overload conditions into account. Exceeding the specified maximum junction temperatures can damage the power semiconductor chips, reduce the lifetime or cause an immediate failure. Usually only the DCB temperature in power modules and IPMs is directly measured by using an NTC. As a result, there either is a large uncertainty if the application relies on temperature monitoring based on the NTC temperature only, or the application needs to have detailed information on the thermal characteristics and the capability to perform a junction temperature calculation. Looking at the integration trends, there is a large benefit of integrating a junction temperature determination into IPMs and hereby allowing the application to directly read-out the junction temperatures. Furthermore, the IPM is then able to monitor the junction temperatures and generate warning signals as well as shut-downs prior to exceeding specified limits. Here, it is desirable to perform the junction temperature calculation at a kHz bandwidth to enable observing the specific limits in overload conditions like low-voltage ride-through in wind applications. This paper presents an innovative approach for a precise real-time calculation of the junction temperatures of a half-bridge IPM based on a lateral thermal model and semiconductor loss data gained by applying a Design-of-Experiment approach.