Some difficulties in fitting dynamic models to experimental transient data in continuous culture