Model-based diagnosis of electric cooling fan drive systems is a contribution to the field of fault detection and diagnosis for electrically driven engine cooling fans. Its main focus is on the online gathering and determination of important parameters and internal states. The developed methods for fault detection and diagnosis are characterized by resource and computing efficient design and by a low application effort, drastically reducing the costs for transferring them to
other applications. Novel algorithms are presented for determination of the winding resistance, the flux linkage over angle and the equivalent series resistance. Based on these algorithms, a new and innovative approach for determination of the magnet temperature is proposed, utilizing the winding temperature, which is derived without requiring an additional temperature sensor. Furthermore, methods are presented for detection of a demagnetization event, detection of an aged DC-link capacitor and...